Page MenuHomeFreeBSD
Files F83957260

No OneTemporary
Actions

View Options

This file is larger than 256 KB, so syntax highlighting was skipped.
diff --git a/lib/libbe/be.c b/lib/libbe/be.c
index a5e053cb82df..a27025684305 100644
--- a/lib/libbe/be.c
+++ b/lib/libbe/be.c
@@ -1,1309 +1,1309 @@
/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
* Copyright (c) 2017 Kyle J. Kneitinger <kyle@kneit.in>
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/mount.h>
#include <sys/stat.h>
#include <sys/ucred.h>
#include <sys/queue.h>
#include <sys/zfs_context.h>
#include <sys/mntent.h>
#include <sys/zfs_ioctl.h>
#include <libzutil.h>
#include <ctype.h>
#include <libgen.h>
#include <libzfs_core.h>
#include <libzfs_impl.h>
#include <stdio.h>
#include <stdlib.h>
#include <time.h>
#include <unistd.h>
#include <libzfsbootenv.h>
#include "be.h"
#include "be_impl.h"
struct promote_entry {
char name[BE_MAXPATHLEN];
SLIST_ENTRY(promote_entry) link;
};
struct be_destroy_data {
libbe_handle_t *lbh;
char target_name[BE_MAXPATHLEN];
char *snapname;
SLIST_HEAD(, promote_entry) promotelist;
};
#if SOON
static int be_create_child_noent(libbe_handle_t *lbh, const char *active,
const char *child_path);
static int be_create_child_cloned(libbe_handle_t *lbh, const char *active);
#endif
/* Arbitrary... should tune */
#define BE_SNAP_SERIAL_MAX 1024
/*
* Iterator function for locating the rootfs amongst the children of the
* zfs_be_root set by loader(8). data is expected to be a libbe_handle_t *.
*/
static int
be_locate_rootfs(libbe_handle_t *lbh)
{
struct statfs sfs;
struct mnttab entry;
zfs_handle_t *zfs;
/*
* Check first if root is ZFS; if not, we'll bail on rootfs capture.
* Unfortunately needed because zfs_path_to_zhandle will emit to
* stderr if / isn't actually a ZFS filesystem, which we'd like
* to avoid.
*/
if (statfs("/", &sfs) == 0) {
statfs2mnttab(&sfs, &entry);
if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0)
return (1);
} else
return (1);
zfs = zfs_path_to_zhandle(lbh->lzh, "/", ZFS_TYPE_FILESYSTEM);
if (zfs == NULL)
return (1);
strlcpy(lbh->rootfs, zfs_get_name(zfs), sizeof(lbh->rootfs));
zfs_close(zfs);
return (0);
}
/*
* Initializes the libbe context to operate in the root boot environment
* dataset, for example, zroot/ROOT.
*/
libbe_handle_t *
libbe_init(const char *root)
{
char altroot[MAXPATHLEN];
libbe_handle_t *lbh;
char *poolname, *pos;
int pnamelen;
lbh = NULL;
poolname = pos = NULL;
if ((lbh = calloc(1, sizeof(libbe_handle_t))) == NULL)
goto err;
if ((lbh->lzh = libzfs_init()) == NULL)
goto err;
/*
* Grab rootfs, we'll work backwards from there if an optional BE root
* has not been passed in.
*/
if (be_locate_rootfs(lbh) != 0) {
if (root == NULL)
goto err;
*lbh->rootfs = '\0';
}
if (root == NULL) {
/* Strip off the final slash from rootfs to get the be root */
strlcpy(lbh->root, lbh->rootfs, sizeof(lbh->root));
pos = strrchr(lbh->root, '/');
if (pos == NULL)
goto err;
*pos = '\0';
} else
strlcpy(lbh->root, root, sizeof(lbh->root));
if ((pos = strchr(lbh->root, '/')) == NULL)
goto err;
pnamelen = pos - lbh->root;
poolname = malloc(pnamelen + 1);
if (poolname == NULL)
goto err;
strlcpy(poolname, lbh->root, pnamelen + 1);
if ((lbh->active_phandle = zpool_open(lbh->lzh, poolname)) == NULL)
goto err;
free(poolname);
poolname = NULL;
if (zpool_get_prop(lbh->active_phandle, ZPOOL_PROP_BOOTFS, lbh->bootfs,
sizeof(lbh->bootfs), NULL, true) != 0)
goto err;
if (zpool_get_prop(lbh->active_phandle, ZPOOL_PROP_ALTROOT,
altroot, sizeof(altroot), NULL, true) == 0 &&
strcmp(altroot, "-") != 0)
lbh->altroot_len = strlen(altroot);
return (lbh);
err:
if (lbh != NULL) {
if (lbh->active_phandle != NULL)
zpool_close(lbh->active_phandle);
if (lbh->lzh != NULL)
libzfs_fini(lbh->lzh);
free(lbh);
}
free(poolname);
return (NULL);
}
/*
* Free memory allocated by libbe_init()
*/
void
libbe_close(libbe_handle_t *lbh)
{
if (lbh->active_phandle != NULL)
zpool_close(lbh->active_phandle);
libzfs_fini(lbh->lzh);
free(lbh);
}
/*
* Proxy through to libzfs for the moment.
*/
void
be_nicenum(uint64_t num, char *buf, size_t buflen)
{
zfs_nicenum(num, buf, buflen);
}
static bool
be_should_promote_clones(zfs_handle_t *zfs_hdl, struct be_destroy_data *bdd)
{
char *atpos;
if (zfs_get_type(zfs_hdl) != ZFS_TYPE_SNAPSHOT)
return (false);
/*
* If we're deleting a snapshot, we need to make sure we only promote
* clones that are derived from one of the snapshots we're deleting,
* rather than that of a snapshot we're not touching. This keeps stuff
* in a consistent state, making sure that we don't error out unless
* we really need to.
*/
if (bdd->snapname == NULL)
return (true);
atpos = strchr(zfs_get_name(zfs_hdl), '@');
return (strcmp(atpos + 1, bdd->snapname) == 0);
}
/*
* This is executed from be_promote_dependent_clones via zfs_iter_dependents,
* It checks if the dependent type is a snapshot then attempts to find any
* clones associated with it. Any clones not related to the destroy target are
* added to the promote list.
*/
static int
be_dependent_clone_cb(zfs_handle_t *zfs_hdl, void *data)
{
int err;
bool found;
const char *name;
struct nvlist *nvl;
struct nvpair *nvp;
struct be_destroy_data *bdd;
struct promote_entry *entry, *newentry;
nvp = NULL;
err = 0;
bdd = (struct be_destroy_data *)data;
if (be_should_promote_clones(zfs_hdl, bdd) &&
(nvl = zfs_get_clones_nvl(zfs_hdl)) != NULL) {
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
name = nvpair_name(nvp);
/*
* Skip if the clone is equal to, or a child of, the
* destroy target.
*/
if (strncmp(name, bdd->target_name,
strlen(bdd->target_name)) == 0 ||
strstr(name, bdd->target_name) == name) {
continue;
}
found = false;
SLIST_FOREACH(entry, &bdd->promotelist, link) {
if (strcmp(entry->name, name) == 0) {
found = true;
break;
}
}
if (found)
continue;
newentry = malloc(sizeof(struct promote_entry));
if (newentry == NULL) {
err = ENOMEM;
break;
}
#define BE_COPY_NAME(entry, src) \
strlcpy((entry)->name, (src), sizeof((entry)->name))
if (BE_COPY_NAME(newentry, name) >=
sizeof(newentry->name)) {
/* Shouldn't happen. */
free(newentry);
err = ENAMETOOLONG;
break;
}
#undef BE_COPY_NAME
/*
* We're building up a SLIST here to make sure both that
* we get the order right and so that we don't
* inadvertently observe the wrong state by promoting
* datasets while we're still walking the tree. The
* latter can lead to situations where we promote a BE
* then effectively demote it again.
*/
SLIST_INSERT_HEAD(&bdd->promotelist, newentry, link);
}
nvlist_free(nvl);
}
zfs_close(zfs_hdl);
return (err);
}
/*
* This is called before a destroy, so that any datasets(environments) that are
* dependent on this one get promoted before destroying the target.
*/
static int
be_promote_dependent_clones(zfs_handle_t *zfs_hdl, struct be_destroy_data *bdd)
{
int err;
zfs_handle_t *clone;
struct promote_entry *entry;
snprintf(bdd->target_name, BE_MAXPATHLEN, "%s/", zfs_get_name(zfs_hdl));
- err = zfs_iter_dependents(zfs_hdl, 0, true, be_dependent_clone_cb, bdd);
+ err = zfs_iter_dependents(zfs_hdl, true, be_dependent_clone_cb, bdd);
/*
* Drain the list and walk away from it if we're only deleting a
* snapshot.
*/
if (bdd->snapname != NULL && !SLIST_EMPTY(&bdd->promotelist))
err = BE_ERR_HASCLONES;
while (!SLIST_EMPTY(&bdd->promotelist)) {
entry = SLIST_FIRST(&bdd->promotelist);
SLIST_REMOVE_HEAD(&bdd->promotelist, link);
#define ZFS_GRAB_CLONE() \
zfs_open(bdd->lbh->lzh, entry->name, ZFS_TYPE_FILESYSTEM)
/*
* Just skip this part on error, we still want to clean up the
* promotion list after the first error. We'll then preserve it
* all the way back.
*/
if (err == 0 && (clone = ZFS_GRAB_CLONE()) != NULL) {
err = zfs_promote(clone);
if (err != 0)
err = BE_ERR_DESTROYMNT;
zfs_close(clone);
}
#undef ZFS_GRAB_CLONE
free(entry);
}
return (err);
}
static int
be_destroy_cb(zfs_handle_t *zfs_hdl, void *data)
{
char path[BE_MAXPATHLEN];
struct be_destroy_data *bdd;
zfs_handle_t *snap;
int err;
bdd = (struct be_destroy_data *)data;
if (bdd->snapname == NULL) {
- err = zfs_iter_children(zfs_hdl, 0, be_destroy_cb, data);
+ err = zfs_iter_children(zfs_hdl, be_destroy_cb, data);
if (err != 0)
return (err);
return (zfs_destroy(zfs_hdl, false));
}
/* If we're dealing with snapshots instead, delete that one alone */
- err = zfs_iter_filesystems(zfs_hdl, 0, be_destroy_cb, data);
+ err = zfs_iter_filesystems(zfs_hdl, be_destroy_cb, data);
if (err != 0)
return (err);
/*
* This part is intentionally glossing over any potential errors,
* because there's a lot less potential for errors when we're cleaning
* up snapshots rather than a full deep BE. The primary error case
* here being if the snapshot doesn't exist in the first place, which
* the caller will likely deem insignificant as long as it doesn't
* exist after the call. Thus, such a missing snapshot shouldn't jam
* up the destruction.
*/
snprintf(path, sizeof(path), "%s@%s", zfs_get_name(zfs_hdl),
bdd->snapname);
if (!zfs_dataset_exists(bdd->lbh->lzh, path, ZFS_TYPE_SNAPSHOT))
return (0);
snap = zfs_open(bdd->lbh->lzh, path, ZFS_TYPE_SNAPSHOT);
if (snap != NULL)
zfs_destroy(snap, false);
return (0);
}
#define BE_DESTROY_WANTORIGIN (BE_DESTROY_ORIGIN | BE_DESTROY_AUTOORIGIN)
/*
* Destroy the boot environment or snapshot specified by the name
* parameter. Options are or'd together with the possible values:
* BE_DESTROY_FORCE : forces operation on mounted datasets
* BE_DESTROY_ORIGIN: destroy the origin snapshot as well
*/
static int
be_destroy_internal(libbe_handle_t *lbh, const char *name, int options,
bool odestroyer)
{
struct be_destroy_data bdd;
char origin[BE_MAXPATHLEN], path[BE_MAXPATHLEN];
zfs_handle_t *fs;
char *snapdelim;
int err, force, mounted;
size_t rootlen;
bdd.lbh = lbh;
bdd.snapname = NULL;
SLIST_INIT(&bdd.promotelist);
force = options & BE_DESTROY_FORCE;
*origin = '\0';
be_root_concat(lbh, name, path);
if ((snapdelim = strchr(path, '@')) == NULL) {
if (!zfs_dataset_exists(lbh->lzh, path, ZFS_TYPE_FILESYSTEM))
return (set_error(lbh, BE_ERR_NOENT));
if (strcmp(path, lbh->rootfs) == 0 ||
strcmp(path, lbh->bootfs) == 0)
return (set_error(lbh, BE_ERR_DESTROYACT));
fs = zfs_open(lbh->lzh, path, ZFS_TYPE_FILESYSTEM);
if (fs == NULL)
return (set_error(lbh, BE_ERR_ZFSOPEN));
/* Don't destroy a mounted dataset unless force is specified */
if ((mounted = zfs_is_mounted(fs, NULL)) != 0) {
if (force) {
zfs_unmount(fs, NULL, 0);
} else {
free(bdd.snapname);
return (set_error(lbh, BE_ERR_DESTROYMNT));
}
}
} else {
/*
* If we're initially destroying a snapshot, origin options do
* not make sense. If we're destroying the origin snapshot of
* a BE, we want to maintain the options in case we need to
* fake success after failing to promote.
*/
if (!odestroyer)
options &= ~BE_DESTROY_WANTORIGIN;
if (!zfs_dataset_exists(lbh->lzh, path, ZFS_TYPE_SNAPSHOT))
return (set_error(lbh, BE_ERR_NOENT));
bdd.snapname = strdup(snapdelim + 1);
if (bdd.snapname == NULL)
return (set_error(lbh, BE_ERR_NOMEM));
*snapdelim = '\0';
fs = zfs_open(lbh->lzh, path, ZFS_TYPE_DATASET);
if (fs == NULL) {
free(bdd.snapname);
return (set_error(lbh, BE_ERR_ZFSOPEN));
}
}
/*
* Whether we're destroying a BE or a single snapshot, we need to walk
* the tree of what we're going to destroy and promote everything in our
* path so that we can make it happen.
*/
if ((err = be_promote_dependent_clones(fs, &bdd)) != 0) {
free(bdd.snapname);
/*
* If we're just destroying the origin of some other dataset
* we were invoked to destroy, then we just ignore
* BE_ERR_HASCLONES and return success unless the caller wanted
* to force the issue.
*/
if (odestroyer && err == BE_ERR_HASCLONES &&
(options & BE_DESTROY_AUTOORIGIN) != 0)
return (0);
return (set_error(lbh, err));
}
/*
* This was deferred until after we promote all of the derivatives so
* that we grab the new origin after everything's settled down.
*/
if ((options & BE_DESTROY_WANTORIGIN) != 0 &&
zfs_prop_get(fs, ZFS_PROP_ORIGIN, origin, sizeof(origin),
NULL, NULL, 0, 1) != 0 &&
(options & BE_DESTROY_ORIGIN) != 0)
return (set_error(lbh, BE_ERR_NOORIGIN));
/*
* If the caller wants auto-origin destruction and the origin
* name matches one of our automatically created snapshot names
* (i.e. strftime("%F-%T") with a serial at the end), then
* we'll set the DESTROY_ORIGIN flag and nuke it
* be_is_auto_snapshot_name is exported from libbe(3) so that
* the caller can determine if it needs to warn about the origin
* not being destroyed or not.
*/
if ((options & BE_DESTROY_AUTOORIGIN) != 0 && *origin != '\0' &&
be_is_auto_snapshot_name(lbh, origin))
options |= BE_DESTROY_ORIGIN;
err = be_destroy_cb(fs, &bdd);
zfs_close(fs);
free(bdd.snapname);
if (err != 0) {
/* Children are still present or the mount is referenced */
if (err == EBUSY)
return (set_error(lbh, BE_ERR_DESTROYMNT));
return (set_error(lbh, BE_ERR_UNKNOWN));
}
if ((options & BE_DESTROY_ORIGIN) == 0)
return (0);
/* The origin can't possibly be shorter than the BE root */
rootlen = strlen(lbh->root);
if (*origin == '\0' || strlen(origin) <= rootlen + 1)
return (set_error(lbh, BE_ERR_INVORIGIN));
/*
* We'll be chopping off the BE root and running this back through
* be_destroy, so that we properly handle the origin snapshot whether
* it be that of a deep BE or not.
*/
if (strncmp(origin, lbh->root, rootlen) != 0 || origin[rootlen] != '/')
return (0);
return (be_destroy_internal(lbh, origin + rootlen + 1,
options & ~BE_DESTROY_ORIGIN, true));
}
int
be_destroy(libbe_handle_t *lbh, const char *name, int options)
{
/*
* The consumer must not set both BE_DESTROY_AUTOORIGIN and
* BE_DESTROY_ORIGIN. Internally, we'll set the latter from the former.
* The latter should imply that we must succeed at destroying the
* origin, or complain otherwise.
*/
if ((options & BE_DESTROY_WANTORIGIN) == BE_DESTROY_WANTORIGIN)
return (set_error(lbh, BE_ERR_UNKNOWN));
return (be_destroy_internal(lbh, name, options, false));
}
static void
be_setup_snapshot_name(libbe_handle_t *lbh, char *buf, size_t buflen)
{
time_t rawtime;
int len, serial;
time(&rawtime);
len = strlen(buf);
len += strftime(buf + len, buflen - len, "@%F-%T", localtime(&rawtime));
/* No room for serial... caller will do its best */
if (buflen - len < 2)
return;
for (serial = 0; serial < BE_SNAP_SERIAL_MAX; ++serial) {
snprintf(buf + len, buflen - len, "-%d", serial);
if (!zfs_dataset_exists(lbh->lzh, buf, ZFS_TYPE_SNAPSHOT))
return;
}
}
bool
be_is_auto_snapshot_name(libbe_handle_t *lbh __unused, const char *name)
{
const char *snap;
int day, hour, minute, month, second, serial, year;
if ((snap = strchr(name, '@')) == NULL)
return (false);
++snap;
/* We'll grab the individual components and do some light validation. */
if (sscanf(snap, "%d-%d-%d-%d:%d:%d-%d", &year, &month, &day, &hour,
&minute, &second, &serial) != 7)
return (false);
return (year >= 1970) && (month >= 1 && month <= 12) &&
(day >= 1 && day <= 31) && (hour >= 0 && hour <= 23) &&
(minute >= 0 && minute <= 59) && (second >= 0 && second <= 60) &&
serial >= 0;
}
int
be_snapshot(libbe_handle_t *lbh, const char *source, const char *snap_name,
bool recursive, char *result)
{
char buf[BE_MAXPATHLEN];
int err;
be_root_concat(lbh, source, buf);
if ((err = be_exists(lbh, buf)) != 0)
return (set_error(lbh, err));
if (snap_name != NULL) {
if (strlcat(buf, "@", sizeof(buf)) >= sizeof(buf))
return (set_error(lbh, BE_ERR_INVALIDNAME));
if (strlcat(buf, snap_name, sizeof(buf)) >= sizeof(buf))
return (set_error(lbh, BE_ERR_INVALIDNAME));
if (result != NULL)
snprintf(result, BE_MAXPATHLEN, "%s@%s", source,
snap_name);
} else {
be_setup_snapshot_name(lbh, buf, sizeof(buf));
if (result != NULL && strlcpy(result, strrchr(buf, '/') + 1,
sizeof(buf)) >= sizeof(buf))
return (set_error(lbh, BE_ERR_INVALIDNAME));
}
if ((err = zfs_snapshot(lbh->lzh, buf, recursive, NULL)) != 0) {
switch (err) {
case EZFS_INVALIDNAME:
return (set_error(lbh, BE_ERR_INVALIDNAME));
default:
/*
* The other errors that zfs_ioc_snapshot might return
* shouldn't happen if we've set things up properly, so
* we'll gloss over them and call it UNKNOWN as it will
* require further triage.
*/
if (errno == ENOTSUP)
return (set_error(lbh, BE_ERR_NOPOOL));
return (set_error(lbh, BE_ERR_UNKNOWN));
}
}
return (BE_ERR_SUCCESS);
}
/*
* Create the boot environment specified by the name parameter
*/
int
be_create(libbe_handle_t *lbh, const char *name)
{
int err;
err = be_create_from_existing(lbh, name, be_active_path(lbh));
return (set_error(lbh, err));
}
static int
be_deep_clone_prop(int prop, void *cb)
{
int err;
struct libbe_dccb *dccb;
zprop_source_t src;
char pval[BE_MAXPATHLEN];
char source[BE_MAXPATHLEN];
char *val;
dccb = cb;
/* Skip some properties we don't want to touch */
if (prop == ZFS_PROP_CANMOUNT)
return (ZPROP_CONT);
/* Don't copy readonly properties */
if (zfs_prop_readonly(prop))
return (ZPROP_CONT);
if ((err = zfs_prop_get(dccb->zhp, prop, (char *)&pval,
sizeof(pval), &src, (char *)&source, sizeof(source), false)))
/* Just continue if we fail to read a property */
return (ZPROP_CONT);
/*
* Only copy locally defined or received properties. This continues
* to avoid temporary/default/local properties intentionally without
* breaking received datasets.
*/
if (src != ZPROP_SRC_LOCAL && src != ZPROP_SRC_RECEIVED)
return (ZPROP_CONT);
/* Augment mountpoint with altroot, if needed */
val = pval;
if (prop == ZFS_PROP_MOUNTPOINT)
val = be_mountpoint_augmented(dccb->lbh, val);
nvlist_add_string(dccb->props, zfs_prop_to_name(prop), val);
return (ZPROP_CONT);
}
/*
* Return the corresponding boot environment path for a given
* dataset path, the constructed path is placed in 'result'.
*
* example: say our new boot environment name is 'bootenv' and
* the dataset path is 'zroot/ROOT/default/data/set'.
*
* result should produce: 'zroot/ROOT/bootenv/data/set'
*/
static int
be_get_path(struct libbe_deep_clone *ldc, const char *dspath, char *result, int result_size)
{
char *pos;
char *child_dataset;
/* match the root path for the boot environments */
pos = strstr(dspath, ldc->lbh->root);
/* no match, different pools? */
if (pos == NULL)
return (BE_ERR_BADPATH);
/* root path of the new boot environment */
snprintf(result, result_size, "%s/%s", ldc->lbh->root, ldc->bename);
/* gets us to the parent dataset, the +1 consumes a trailing slash */
pos += strlen(ldc->lbh->root) + 1;
/* skip the parent dataset */
if ((child_dataset = strchr(pos, '/')) != NULL)
strlcat(result, child_dataset, result_size);
return (BE_ERR_SUCCESS);
}
static int
be_clone_cb(zfs_handle_t *ds, void *data)
{
int err;
char be_path[BE_MAXPATHLEN];
char snap_path[BE_MAXPATHLEN];
const char *dspath;
zfs_handle_t *snap_hdl;
nvlist_t *props;
struct libbe_deep_clone *ldc;
struct libbe_dccb dccb;
ldc = (struct libbe_deep_clone *)data;
dspath = zfs_get_name(ds);
snprintf(snap_path, sizeof(snap_path), "%s@%s", dspath, ldc->snapname);
/* construct the boot environment path from the dataset we're cloning */
if (be_get_path(ldc, dspath, be_path, sizeof(be_path)) != BE_ERR_SUCCESS)
return (BE_ERR_UNKNOWN);
/* the dataset to be created (i.e. the boot environment) already exists */
if (zfs_dataset_exists(ldc->lbh->lzh, be_path, ZFS_TYPE_DATASET))
return (BE_ERR_EXISTS);
/* no snapshot found for this dataset, silently skip it */
if (!zfs_dataset_exists(ldc->lbh->lzh, snap_path, ZFS_TYPE_SNAPSHOT))
return (0);
if ((snap_hdl =
zfs_open(ldc->lbh->lzh, snap_path, ZFS_TYPE_SNAPSHOT)) == NULL)
return (BE_ERR_ZFSOPEN);
nvlist_alloc(&props, NV_UNIQUE_NAME, KM_SLEEP);
nvlist_add_string(props, "canmount", "noauto");
dccb.lbh = ldc->lbh;
dccb.zhp = ds;
dccb.props = props;
if (zprop_iter(be_deep_clone_prop, &dccb, B_FALSE, B_FALSE,
ZFS_TYPE_FILESYSTEM) == ZPROP_INVAL)
return (-1);
if ((err = zfs_clone(snap_hdl, be_path, props)) != 0)
return (BE_ERR_ZFSCLONE);
nvlist_free(props);
zfs_close(snap_hdl);
if (ldc->depth_limit == -1 || ldc->depth < ldc->depth_limit) {
ldc->depth++;
- err = zfs_iter_filesystems(ds, 0, be_clone_cb, ldc);
+ err = zfs_iter_filesystems(ds, be_clone_cb, ldc);
ldc->depth--;
}
return (err);
}
/*
* Create a boot environment with a given name from a given snapshot.
* Snapshots can be in the format 'zroot/ROOT/default@snapshot' or
* 'default@snapshot'. In the latter case, 'default@snapshot' will be prepended
* with the root path that libbe was initailized with.
*/
static int
be_clone(libbe_handle_t *lbh, const char *bename, const char *snapshot, int depth)
{
int err;
char snap_path[BE_MAXPATHLEN];
char *parentname, *snapname;
zfs_handle_t *parent_hdl;
struct libbe_deep_clone ldc;
/* ensure the boot environment name is valid */
if ((err = be_validate_name(lbh, bename)) != 0)
return (set_error(lbh, err));
/*
* prepend the boot environment root path if we're
* given a partial snapshot name.
*/
if ((err = be_root_concat(lbh, snapshot, snap_path)) != 0)
return (set_error(lbh, err));
/* ensure the snapshot exists */
if ((err = be_validate_snap(lbh, snap_path)) != 0)
return (set_error(lbh, err));
/* get a copy of the snapshot path so we can disect it */
if ((parentname = strdup(snap_path)) == NULL)
return (set_error(lbh, BE_ERR_UNKNOWN));
/* split dataset name from snapshot name */
snapname = strchr(parentname, '@');
if (snapname == NULL) {
free(parentname);
return (set_error(lbh, BE_ERR_UNKNOWN));
}
*snapname = '\0';
snapname++;
/* set-up the boot environment */
ldc.lbh = lbh;
ldc.bename = bename;
ldc.snapname = snapname;
ldc.depth = 0;
ldc.depth_limit = depth;
/* the boot environment will be cloned from this dataset */
parent_hdl = zfs_open(lbh->lzh, parentname, ZFS_TYPE_DATASET);
/* create the boot environment */
err = be_clone_cb(parent_hdl, &ldc);
free(parentname);
return (set_error(lbh, err));
}
/*
* Create a boot environment from pre-existing snapshot, specifying a depth.
*/
int be_create_depth(libbe_handle_t *lbh, const char *bename,
const char *snap, int depth)
{
return (be_clone(lbh, bename, snap, depth));
}
/*
* Create the boot environment from pre-existing snapshot
*/
int
be_create_from_existing_snap(libbe_handle_t *lbh, const char *bename,
const char *snap)
{
return (be_clone(lbh, bename, snap, -1));
}
/*
* Create a boot environment from an existing boot environment
*/
int
be_create_from_existing(libbe_handle_t *lbh, const char *bename, const char *old)
{
int err;
char snap[BE_MAXPATHLEN];
if ((err = be_snapshot(lbh, old, NULL, true, snap)) != 0)
return (set_error(lbh, err));
err = be_clone(lbh, bename, snap, -1);
return (set_error(lbh, err));
}
/*
* Verifies that a snapshot has a valid name, exists, and has a mountpoint of
* '/'. Returns BE_ERR_SUCCESS (0), upon success, or the relevant BE_ERR_* upon
* failure. Does not set the internal library error state.
*/
int
be_validate_snap(libbe_handle_t *lbh, const char *snap_name)
{
if (strlen(snap_name) >= BE_MAXPATHLEN)
return (BE_ERR_PATHLEN);
if (!zfs_name_valid(snap_name, ZFS_TYPE_SNAPSHOT))
return (BE_ERR_INVALIDNAME);
if (!zfs_dataset_exists(lbh->lzh, snap_name,
ZFS_TYPE_SNAPSHOT))
return (BE_ERR_NOENT);
return (BE_ERR_SUCCESS);
}
/*
* Idempotently appends the name argument to the root boot environment path
* and copies the resulting string into the result buffer (which is assumed
* to be at least BE_MAXPATHLEN characters long. Returns BE_ERR_SUCCESS upon
* success, BE_ERR_PATHLEN if the resulting path is longer than BE_MAXPATHLEN,
* or BE_ERR_INVALIDNAME if the name is a path that does not begin with
* zfs_be_root. Does not set internal library error state.
*/
int
be_root_concat(libbe_handle_t *lbh, const char *name, char *result)
{
size_t name_len, root_len;
name_len = strlen(name);
root_len = strlen(lbh->root);
/* Act idempotently; return be name if it is already a full path */
if (strrchr(name, '/') != NULL) {
if (strstr(name, lbh->root) != name)
return (BE_ERR_INVALIDNAME);
if (name_len >= BE_MAXPATHLEN)
return (BE_ERR_PATHLEN);
strlcpy(result, name, BE_MAXPATHLEN);
return (BE_ERR_SUCCESS);
} else if (name_len + root_len + 1 < BE_MAXPATHLEN) {
snprintf(result, BE_MAXPATHLEN, "%s/%s", lbh->root,
name);
return (BE_ERR_SUCCESS);
}
return (BE_ERR_PATHLEN);
}
/*
* Verifies the validity of a boot environment name (A-Za-z0-9-_.). Returns
* BE_ERR_SUCCESS (0) if name is valid, otherwise returns BE_ERR_INVALIDNAME
* or BE_ERR_PATHLEN.
* Does not set internal library error state.
*/
int
be_validate_name(libbe_handle_t *lbh, const char *name)
{
/*
* Impose the additional restriction that the entire dataset name must
* not exceed the maximum length of a dataset, i.e. MAXNAMELEN.
*/
if (strlen(lbh->root) + 1 + strlen(name) > MAXNAMELEN)
return (BE_ERR_PATHLEN);
if (!zfs_name_valid(name, ZFS_TYPE_DATASET))
return (BE_ERR_INVALIDNAME);
/*
* ZFS allows spaces in boot environment names, but the kernel can't
* handle booting from such a dataset right now. vfs.root.mountfrom
* is defined to be a space-separated list, and there's no protocol for
* escaping whitespace in the path component of a dev:path spec. So
* while loader can handle this situation alright, it can't safely pass
* it on to mountroot.
*/
if (strchr(name, ' ') != NULL)
return (BE_ERR_INVALIDNAME);
return (BE_ERR_SUCCESS);
}
/*
* usage
*/
int
be_rename(libbe_handle_t *lbh, const char *old, const char *new)
{
char full_old[BE_MAXPATHLEN];
char full_new[BE_MAXPATHLEN];
zfs_handle_t *zfs_hdl;
int err;
/*
* be_validate_name is documented not to set error state, so we should
* do so here.
*/
if ((err = be_validate_name(lbh, new)) != 0)
return (set_error(lbh, err));
if ((err = be_root_concat(lbh, old, full_old)) != 0)
return (set_error(lbh, err));
if ((err = be_root_concat(lbh, new, full_new)) != 0)
return (set_error(lbh, err));
if (!zfs_dataset_exists(lbh->lzh, full_old, ZFS_TYPE_DATASET))
return (set_error(lbh, BE_ERR_NOENT));
if (zfs_dataset_exists(lbh->lzh, full_new, ZFS_TYPE_DATASET))
return (set_error(lbh, BE_ERR_EXISTS));
if ((zfs_hdl = zfs_open(lbh->lzh, full_old,
ZFS_TYPE_FILESYSTEM)) == NULL)
return (set_error(lbh, BE_ERR_ZFSOPEN));
/* recurse, nounmount, forceunmount */
struct renameflags flags = {
.nounmount = 1,
};
err = zfs_rename(zfs_hdl, full_new, flags);
zfs_close(zfs_hdl);
if (err != 0)
return (set_error(lbh, BE_ERR_UNKNOWN));
return (0);
}
int
be_export(libbe_handle_t *lbh, const char *bootenv, int fd)
{
char snap_name[BE_MAXPATHLEN];
char buf[BE_MAXPATHLEN];
zfs_handle_t *zfs;
sendflags_t flags = { 0 };
int err;
if ((err = be_snapshot(lbh, bootenv, NULL, true, snap_name)) != 0)
/* Use the error set by be_snapshot */
return (err);
be_root_concat(lbh, snap_name, buf);
if ((zfs = zfs_open(lbh->lzh, buf, ZFS_TYPE_DATASET)) == NULL)
return (set_error(lbh, BE_ERR_ZFSOPEN));
err = zfs_send_one(zfs, NULL, fd, &flags, /* redactbook */ NULL);
zfs_close(zfs);
return (err);
}
int
be_import(libbe_handle_t *lbh, const char *bootenv, int fd)
{
char buf[BE_MAXPATHLEN];
nvlist_t *props;
zfs_handle_t *zfs;
recvflags_t flags = { .nomount = 1 };
int err;
be_root_concat(lbh, bootenv, buf);
if ((err = zfs_receive(lbh->lzh, buf, NULL, &flags, fd, NULL)) != 0) {
switch (err) {
case EINVAL:
return (set_error(lbh, BE_ERR_NOORIGIN));
case ENOENT:
return (set_error(lbh, BE_ERR_NOENT));
case EIO:
return (set_error(lbh, BE_ERR_IO));
default:
return (set_error(lbh, BE_ERR_UNKNOWN));
}
}
if ((zfs = zfs_open(lbh->lzh, buf, ZFS_TYPE_FILESYSTEM)) == NULL)
return (set_error(lbh, BE_ERR_ZFSOPEN));
nvlist_alloc(&props, NV_UNIQUE_NAME, KM_SLEEP);
nvlist_add_string(props, "canmount", "noauto");
nvlist_add_string(props, "mountpoint", "none");
err = zfs_prop_set_list(zfs, props);
nvlist_free(props);
zfs_close(zfs);
if (err != 0)
return (set_error(lbh, BE_ERR_UNKNOWN));
return (0);
}
#if SOON
static int
be_create_child_noent(libbe_handle_t *lbh, const char *active,
const char *child_path)
{
nvlist_t *props;
zfs_handle_t *zfs;
int err;
nvlist_alloc(&props, NV_UNIQUE_NAME, KM_SLEEP);
nvlist_add_string(props, "canmount", "noauto");
nvlist_add_string(props, "mountpoint", child_path);
/* Create */
if ((err = zfs_create(lbh->lzh, active, ZFS_TYPE_DATASET,
props)) != 0) {
switch (err) {
case EZFS_EXISTS:
return (set_error(lbh, BE_ERR_EXISTS));
case EZFS_NOENT:
return (set_error(lbh, BE_ERR_NOENT));
case EZFS_BADTYPE:
case EZFS_BADVERSION:
return (set_error(lbh, BE_ERR_NOPOOL));
case EZFS_BADPROP:
default:
/* We set something up wrong, probably... */
return (set_error(lbh, BE_ERR_UNKNOWN));
}
}
nvlist_free(props);
if ((zfs = zfs_open(lbh->lzh, active, ZFS_TYPE_DATASET)) == NULL)
return (set_error(lbh, BE_ERR_ZFSOPEN));
/* Set props */
if ((err = zfs_prop_set(zfs, "canmount", "noauto")) != 0) {
zfs_close(zfs);
/*
* Similar to other cases, this shouldn't fail unless we've
* done something wrong. This is a new dataset that shouldn't
* have been mounted anywhere between creation and now.
*/
if (err == EZFS_NOMEM)
return (set_error(lbh, BE_ERR_NOMEM));
return (set_error(lbh, BE_ERR_UNKNOWN));
}
zfs_close(zfs);
return (BE_ERR_SUCCESS);
}
static int
be_create_child_cloned(libbe_handle_t *lbh, const char *active)
{
char buf[BE_MAXPATHLEN], tmp[BE_MAXPATHLEN];;
zfs_handle_t *zfs;
int err;
/* XXX TODO ? */
/*
* Establish if the existing path is a zfs dataset or just
* the subdirectory of one
*/
strlcpy(tmp, "tmp/be_snap.XXXXX", sizeof(tmp));
if (mktemp(tmp) == NULL)
return (set_error(lbh, BE_ERR_UNKNOWN));
be_root_concat(lbh, tmp, buf);
printf("Here %s?\n", buf);
if ((err = zfs_snapshot(lbh->lzh, buf, false, NULL)) != 0) {
switch (err) {
case EZFS_INVALIDNAME:
return (set_error(lbh, BE_ERR_INVALIDNAME));
default:
/*
* The other errors that zfs_ioc_snapshot might return
* shouldn't happen if we've set things up properly, so
* we'll gloss over them and call it UNKNOWN as it will
* require further triage.
*/
if (errno == ENOTSUP)
return (set_error(lbh, BE_ERR_NOPOOL));
return (set_error(lbh, BE_ERR_UNKNOWN));
}
}
/* Clone */
if ((zfs = zfs_open(lbh->lzh, buf, ZFS_TYPE_SNAPSHOT)) == NULL)
return (BE_ERR_ZFSOPEN);
if ((err = zfs_clone(zfs, active, NULL)) != 0)
/* XXX TODO correct error */
return (set_error(lbh, BE_ERR_UNKNOWN));
/* set props */
zfs_close(zfs);
return (BE_ERR_SUCCESS);
}
int
be_add_child(libbe_handle_t *lbh, const char *child_path, bool cp_if_exists)
{
struct stat sb;
char active[BE_MAXPATHLEN], buf[BE_MAXPATHLEN];
nvlist_t *props;
const char *s;
/* Require absolute paths */
if (*child_path != '/')
return (set_error(lbh, BE_ERR_BADPATH));
strlcpy(active, be_active_path(lbh), BE_MAXPATHLEN);
strcpy(buf, active);
/* Create non-mountable parent dataset(s) */
s = child_path;
for (char *p; (p = strchr(s+1, '/')) != NULL; s = p) {
size_t len = p - s;
strncat(buf, s, len);
nvlist_alloc(&props, NV_UNIQUE_NAME, KM_SLEEP);
nvlist_add_string(props, "canmount", "off");
nvlist_add_string(props, "mountpoint", "none");
zfs_create(lbh->lzh, buf, ZFS_TYPE_DATASET, props);
nvlist_free(props);
}
/* Path does not exist as a descendent of / yet */
if (strlcat(active, child_path, BE_MAXPATHLEN) >= BE_MAXPATHLEN)
return (set_error(lbh, BE_ERR_PATHLEN));
if (stat(child_path, &sb) != 0) {
/* Verify that error is ENOENT */
if (errno != ENOENT)
return (set_error(lbh, BE_ERR_UNKNOWN));
return (be_create_child_noent(lbh, active, child_path));
} else if (cp_if_exists)
/* Path is already a descendent of / and should be copied */
return (be_create_child_cloned(lbh, active));
return (set_error(lbh, BE_ERR_EXISTS));
}
#endif /* SOON */
/*
* Deactivate old BE dataset; currently just sets canmount=noauto or
* resets boot once configuration.
*/
int
be_deactivate(libbe_handle_t *lbh, const char *ds, bool temporary)
{
zfs_handle_t *zfs;
if (temporary) {
return (lzbe_set_boot_device(
zpool_get_name(lbh->active_phandle), lzbe_add, NULL));
}
if ((zfs = zfs_open(lbh->lzh, ds, ZFS_TYPE_DATASET)) == NULL)
return (1);
if (zfs_prop_set(zfs, "canmount", "noauto") != 0)
return (1);
zfs_close(zfs);
return (0);
}
int
be_activate(libbe_handle_t *lbh, const char *bootenv, bool temporary)
{
char be_path[BE_MAXPATHLEN];
nvlist_t *dsprops;
const char *origin;
zfs_handle_t *zhp;
int err;
be_root_concat(lbh, bootenv, be_path);
/* Note: be_exists fails if mountpoint is not / */
if ((err = be_exists(lbh, be_path)) != 0)
return (set_error(lbh, err));
if (temporary) {
return (lzbe_set_boot_device(
zpool_get_name(lbh->active_phandle), lzbe_add, be_path));
} else {
if (strncmp(lbh->bootfs, "-", 1) != 0 &&
be_deactivate(lbh, lbh->bootfs, false) != 0)
return (-1);
/* Obtain bootenv zpool */
err = zpool_set_prop(lbh->active_phandle, "bootfs", be_path);
if (err)
return (-1);
zhp = zfs_open(lbh->lzh, be_path, ZFS_TYPE_FILESYSTEM);
if (zhp == NULL)
return (-1);
if (be_prop_list_alloc(&dsprops) != 0)
return (-1);
if (be_get_dataset_props(lbh, be_path, dsprops) != 0) {
nvlist_free(dsprops);
return (-1);
}
if (nvlist_lookup_string(dsprops, "origin", &origin) == 0)
err = zfs_promote(zhp);
nvlist_free(dsprops);
zfs_close(zhp);
if (err)
return (-1);
}
return (BE_ERR_SUCCESS);
}
diff --git a/lib/libbe/be_access.c b/lib/libbe/be_access.c
index e6f526de1aca..ec966bc85fb7 100644
--- a/lib/libbe/be_access.c
+++ b/lib/libbe/be_access.c
@@ -1,340 +1,340 @@
/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
* Copyright (c) 2017 Kyle J. Kneitinger <kyle@kneit.in>
* Copyright (c) 2018 Kyle Evans <kevans@FreeBSD.org>
* Copyright (c) 2019 Wes Maag <wes@jwmaag.org>
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/mntent.h>
#include "be.h"
#include "be_impl.h"
struct be_mountcheck_info {
const char *path;
char *name;
};
struct be_mount_info {
libbe_handle_t *lbh;
const char *be;
const char *mountpoint;
int mntflags;
int deepmount;
int depth;
};
static int
be_mountcheck_cb(zfs_handle_t *zfs_hdl, void *data)
{
struct be_mountcheck_info *info;
char *mountpoint;
if (data == NULL)
return (1);
info = (struct be_mountcheck_info *)data;
if (!zfs_is_mounted(zfs_hdl, &mountpoint))
return (0);
if (strcmp(mountpoint, info->path) == 0) {
info->name = strdup(zfs_get_name(zfs_hdl));
free(mountpoint);
return (1);
}
free(mountpoint);
return (0);
}
/*
* Called from be_mount, uses the given zfs_handle and attempts to
* mount it at the passed mountpoint. If the deepmount flag is set, continue
* calling the function for each child dataset.
*/
static int
be_mount_iter(zfs_handle_t *zfs_hdl, void *data)
{
int err;
char *mountpoint;
char tmp[BE_MAXPATHLEN], zfs_mnt[BE_MAXPATHLEN];
struct be_mount_info *info;
info = (struct be_mount_info *)data;
if (zfs_is_mounted(zfs_hdl, &mountpoint)) {
free(mountpoint);
return (0);
}
/*
* canmount and mountpoint are both ignored for the BE dataset, because
* the rest of the system (kernel and loader) will effectively do the
* same.
*/
if (info->depth == 0) {
snprintf(tmp, BE_MAXPATHLEN, "%s", info->mountpoint);
} else {
if (zfs_prop_get_int(zfs_hdl, ZFS_PROP_CANMOUNT) ==
ZFS_CANMOUNT_OFF)
return (0);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_MOUNTPOINT, zfs_mnt,
BE_MAXPATHLEN, NULL, NULL, 0, 1))
return (1);
/*
* We've encountered mountpoint=none at some intermediate
* dataset (e.g. zroot/var) that will have children that may
* need to be mounted. Skip mounting it, but iterate through
* the children.
*/
if (strcmp("none", zfs_mnt) == 0)
goto skipmount;
mountpoint = be_mountpoint_augmented(info->lbh, zfs_mnt);
snprintf(tmp, BE_MAXPATHLEN, "%s%s", info->mountpoint,
mountpoint);
}
if ((err = zfs_mount_at(zfs_hdl, NULL, info->mntflags, tmp)) != 0) {
switch (errno) {
case ENAMETOOLONG:
return (set_error(info->lbh, BE_ERR_PATHLEN));
case ELOOP:
case ENOENT:
case ENOTDIR:
return (set_error(info->lbh, BE_ERR_BADPATH));
case EPERM:
return (set_error(info->lbh, BE_ERR_PERMS));
case EBUSY:
return (set_error(info->lbh, BE_ERR_PATHBUSY));
default:
return (set_error(info->lbh, BE_ERR_UNKNOWN));
}
}
if (!info->deepmount)
return (0);
skipmount:
++info->depth;
- err = zfs_iter_filesystems(zfs_hdl, 0, be_mount_iter, info);
+ err = zfs_iter_filesystems(zfs_hdl, be_mount_iter, info);
--info->depth;
return (err);
}
static int
be_umount_iter(zfs_handle_t *zfs_hdl, void *data)
{
int err;
char *mountpoint;
struct be_mount_info *info;
info = (struct be_mount_info *)data;
++info->depth;
- if((err = zfs_iter_filesystems(zfs_hdl, 0, be_umount_iter, info)) != 0) {
+ if((err = zfs_iter_filesystems(zfs_hdl, be_umount_iter, info)) != 0) {
return (err);
}
--info->depth;
if (!zfs_is_mounted(zfs_hdl, &mountpoint)) {
return (0);
}
free(mountpoint);
if (zfs_unmount(zfs_hdl, NULL, info->mntflags) != 0) {
switch (errno) {
case ENAMETOOLONG:
return (set_error(info->lbh, BE_ERR_PATHLEN));
case ELOOP:
case ENOENT:
case ENOTDIR:
return (set_error(info->lbh, BE_ERR_BADPATH));
case EPERM:
return (set_error(info->lbh, BE_ERR_PERMS));
case EBUSY:
return (set_error(info->lbh, BE_ERR_PATHBUSY));
default:
return (set_error(info->lbh, BE_ERR_UNKNOWN));
}
}
return (0);
}
/*
* usage
*/
int
be_mounted_at(libbe_handle_t *lbh, const char *path, nvlist_t *details)
{
char be[BE_MAXPATHLEN];
zfs_handle_t *root_hdl;
struct be_mountcheck_info info;
prop_data_t propinfo;
bzero(&be, BE_MAXPATHLEN);
if ((root_hdl = zfs_open(lbh->lzh, lbh->root,
ZFS_TYPE_FILESYSTEM)) == NULL)
return (BE_ERR_ZFSOPEN);
info.path = path;
info.name = NULL;
- zfs_iter_filesystems(root_hdl, 0, be_mountcheck_cb, &info);
+ zfs_iter_filesystems(root_hdl, be_mountcheck_cb, &info);
zfs_close(root_hdl);
if (info.name != NULL) {
if (details != NULL) {
if ((root_hdl = zfs_open(lbh->lzh, info.name,
ZFS_TYPE_FILESYSTEM)) == NULL) {
free(info.name);
return (BE_ERR_ZFSOPEN);
}
propinfo.lbh = lbh;
propinfo.list = details;
propinfo.single_object = false;
propinfo.bootonce = NULL;
prop_list_builder_cb(root_hdl, &propinfo);
zfs_close(root_hdl);
}
free(info.name);
return (0);
}
return (1);
}
/*
* usage
*/
int
be_mount(libbe_handle_t *lbh, const char *bootenv, const char *mountpoint,
int flags, char *result_loc)
{
char be[BE_MAXPATHLEN];
char mnt_temp[BE_MAXPATHLEN];
int mntflags, mntdeep;
int err;
struct be_mount_info info;
zfs_handle_t *zhdl;
if ((err = be_root_concat(lbh, bootenv, be)) != 0)
return (set_error(lbh, err));
if ((err = be_exists(lbh, bootenv)) != 0)
return (set_error(lbh, err));
if (is_mounted(lbh->lzh, be, NULL))
return (set_error(lbh, BE_ERR_MOUNTED));
mntdeep = (flags & BE_MNT_DEEP) ? 1 : 0;
mntflags = (flags & BE_MNT_FORCE) ? MNT_FORCE : 0;
/* Create mountpoint if it is not specified */
if (mountpoint == NULL) {
strlcpy(mnt_temp, "/tmp/be_mount.XXXX", sizeof(mnt_temp));
if (mkdtemp(mnt_temp) == NULL)
return (set_error(lbh, BE_ERR_IO));
}
if ((zhdl = zfs_open(lbh->lzh, be, ZFS_TYPE_FILESYSTEM)) == NULL)
return (set_error(lbh, BE_ERR_ZFSOPEN));
info.lbh = lbh;
info.be = be;
info.mountpoint = (mountpoint == NULL) ? mnt_temp : mountpoint;
info.mntflags = mntflags;
info.deepmount = mntdeep;
info.depth = 0;
if((err = be_mount_iter(zhdl, &info) != 0)) {
zfs_close(zhdl);
return (err);
}
zfs_close(zhdl);
if (result_loc != NULL)
strlcpy(result_loc, mountpoint == NULL ? mnt_temp : mountpoint,
BE_MAXPATHLEN);
return (BE_ERR_SUCCESS);
}
/*
* usage
*/
int
be_unmount(libbe_handle_t *lbh, const char *bootenv, int flags)
{
int err;
char be[BE_MAXPATHLEN];
zfs_handle_t *root_hdl;
struct be_mount_info info;
if ((err = be_root_concat(lbh, bootenv, be)) != 0)
return (set_error(lbh, err));
if ((root_hdl = zfs_open(lbh->lzh, be, ZFS_TYPE_FILESYSTEM)) == NULL)
return (set_error(lbh, BE_ERR_ZFSOPEN));
info.lbh = lbh;
info.be = be;
info.mountpoint = NULL;
info.mntflags = (flags & BE_MNT_FORCE) ? MS_FORCE : 0;
info.depth = 0;
if ((err = be_umount_iter(root_hdl, &info)) != 0) {
zfs_close(root_hdl);
return (err);
}
zfs_close(root_hdl);
return (BE_ERR_SUCCESS);
}
/*
* This function will blow away the input buffer as needed if we're discovered
* to be looking at a root-mount. If the mountpoint is naturally beyond the
* root, however, the buffer may be left intact and a pointer to the section
* past altroot will be returned instead for the caller's perusal.
*/
char *
be_mountpoint_augmented(libbe_handle_t *lbh, char *mountpoint)
{
if (lbh->altroot_len == 0)
return (mountpoint);
if (mountpoint == NULL || *mountpoint == '\0')
return (mountpoint);
if (mountpoint[lbh->altroot_len] == '\0') {
*(mountpoint + 1) = '\0';
return (mountpoint);
} else
return (mountpoint + lbh->altroot_len);
}
diff --git a/lib/libbe/be_info.c b/lib/libbe/be_info.c
index c29b4ba06922..81f3d46a7208 100644
--- a/lib/libbe/be_info.c
+++ b/lib/libbe/be_info.c
@@ -1,318 +1,318 @@
/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
* Copyright (c) 2017 Kyle J. Kneitinger <kyle@kneit.in>
* Copyright (c) 2018 Kyle Evans <kevans@FreeBSD.org>
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/zfs_context.h>
#include <libzfsbootenv.h>
#include "be.h"
#include "be_impl.h"
static int snapshot_proplist_update(zfs_handle_t *hdl, prop_data_t *data);
/*
* Returns the name of the active boot environment
*/
const char *
be_active_name(libbe_handle_t *lbh)
{
if (*lbh->rootfs != '\0')
return (strrchr(lbh->rootfs, '/') + sizeof(char));
else
return (lbh->rootfs);
}
/*
* Returns full path of the active boot environment
*/
const char *
be_active_path(libbe_handle_t *lbh)
{
return (lbh->rootfs);
}
/*
* Returns the name of the next active boot environment
*/
const char *
be_nextboot_name(libbe_handle_t *lbh)
{
if (*lbh->bootfs != '\0')
return (strrchr(lbh->bootfs, '/') + sizeof(char));
else
return (lbh->bootfs);
}
/*
* Returns full path of the active boot environment
*/
const char *
be_nextboot_path(libbe_handle_t *lbh)
{
return (lbh->bootfs);
}
/*
* Returns the path of the boot environment root dataset
*/
const char *
be_root_path(libbe_handle_t *lbh)
{
return (lbh->root);
}
/*
* Populates dsnvl with one nvlist per bootenv dataset describing the properties
* of that dataset that we've declared ourselves to care about.
*/
int
be_get_bootenv_props(libbe_handle_t *lbh, nvlist_t *dsnvl)
{
prop_data_t data;
data.lbh = lbh;
data.list = dsnvl;
data.single_object = false;
data.bootonce = NULL;
return (be_proplist_update(&data));
}
int
be_get_dataset_props(libbe_handle_t *lbh, const char *name, nvlist_t *props)
{
zfs_handle_t *snap_hdl;
prop_data_t data;
int ret;
data.lbh = lbh;
data.list = props;
data.single_object = true;
data.bootonce = NULL;
if ((snap_hdl = zfs_open(lbh->lzh, name,
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_SNAPSHOT)) == NULL)
return (BE_ERR_ZFSOPEN);
ret = prop_list_builder_cb(snap_hdl, &data);
zfs_close(snap_hdl);
return (ret);
}
int
be_get_dataset_snapshots(libbe_handle_t *lbh, const char *name, nvlist_t *props)
{
zfs_handle_t *ds_hdl;
prop_data_t data;
int ret;
data.lbh = lbh;
data.list = props;
data.single_object = false;
data.bootonce = NULL;
if ((ds_hdl = zfs_open(lbh->lzh, name,
ZFS_TYPE_FILESYSTEM)) == NULL)
return (BE_ERR_ZFSOPEN);
ret = snapshot_proplist_update(ds_hdl, &data);
zfs_close(ds_hdl);
return (ret);
}
/*
* Internal callback function used by zfs_iter_filesystems. For each dataset in
* the bootenv root, populate an nvlist_t of its relevant properties.
*/
int
prop_list_builder_cb(zfs_handle_t *zfs_hdl, void *data_p)
{
char buf[512], *mountpoint;
prop_data_t *data;
libbe_handle_t *lbh;
nvlist_t *props;
const char *dataset, *name;
boolean_t mounted;
/*
* XXX TODO:
* some system for defining constants for the nvlist keys
* error checking
*/
data = (prop_data_t *)data_p;
lbh = data->lbh;
if (data->single_object)
props = data->list;
else
nvlist_alloc(&props, NV_UNIQUE_NAME, KM_SLEEP);
dataset = zfs_get_name(zfs_hdl);
nvlist_add_string(props, "dataset", dataset);
if (data->bootonce != NULL &&
strcmp(dataset, data->bootonce) == 0)
nvlist_add_boolean_value(props, "bootonce", true);
name = strrchr(dataset, '/') + 1;
nvlist_add_string(props, "name", name);
mounted = zfs_is_mounted(zfs_hdl, &mountpoint);
if (mounted)
nvlist_add_string(props, "mounted", mountpoint);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_MOUNTPOINT, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "mountpoint", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_ORIGIN, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "origin", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_CREATION, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "creation", buf);
nvlist_add_boolean_value(props, "active",
(strcmp(be_active_path(lbh), dataset) == 0));
if (zfs_prop_get(zfs_hdl, ZFS_PROP_USED, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "used", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_USEDDS, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "usedds", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_USEDSNAP, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "usedsnap", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_USEDREFRESERV, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "usedrefreserv", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_REFERENCED, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "referenced", buf);
nvlist_add_boolean_value(props, "nextboot",
(strcmp(be_nextboot_path(lbh), dataset) == 0));
if (!data->single_object)
nvlist_add_nvlist(data->list, name, props);
return (0);
}
/*
* Updates the properties of each bootenv in the libbe handle
* XXX TODO: ensure that this is always consistent (run after adds, deletes,
* renames,etc
*/
int
be_proplist_update(prop_data_t *data)
{
zfs_handle_t *root_hdl;
if ((root_hdl = zfs_open(data->lbh->lzh, data->lbh->root,
ZFS_TYPE_FILESYSTEM)) == NULL)
return (BE_ERR_ZFSOPEN);
(void) lzbe_get_boot_device(zpool_get_name(data->lbh->active_phandle),
&data->bootonce);
/* XXX TODO: some error checking here */
- zfs_iter_filesystems(root_hdl, 0, prop_list_builder_cb, data);
+ zfs_iter_filesystems(root_hdl, prop_list_builder_cb, data);
zfs_close(root_hdl);
return (0);
}
static int
snapshot_proplist_update(zfs_handle_t *hdl, prop_data_t *data)
{
- return (zfs_iter_snapshots_sorted(hdl, 0, prop_list_builder_cb, data,
+ return (zfs_iter_snapshots_sorted(hdl, prop_list_builder_cb, data,
0, 0));
}
int
be_prop_list_alloc(nvlist_t **be_list)
{
return (nvlist_alloc(be_list, NV_UNIQUE_NAME, KM_SLEEP));
}
/*
* frees property list and its children
*/
void
be_prop_list_free(nvlist_t *be_list)
{
nvlist_t *prop_list;
nvpair_t *be_pair;
be_pair = nvlist_next_nvpair(be_list, NULL);
if (nvpair_value_nvlist(be_pair, &prop_list) == 0)
nvlist_free(prop_list);
while ((be_pair = nvlist_next_nvpair(be_list, be_pair)) != NULL) {
if (nvpair_value_nvlist(be_pair, &prop_list) == 0)
nvlist_free(prop_list);
}
}
/*
* Usage
*/
int
be_exists(libbe_handle_t *lbh, const char *be)
{
char buf[BE_MAXPATHLEN];
be_root_concat(lbh, be, buf);
if (!zfs_dataset_exists(lbh->lzh, buf, ZFS_TYPE_DATASET))
return (BE_ERR_NOENT);
return (BE_ERR_SUCCESS);
}
diff --git a/sys/contrib/openzfs/META b/sys/contrib/openzfs/META
index 18b3f3498eeb..8779e512f7be 100644
--- a/sys/contrib/openzfs/META
+++ b/sys/contrib/openzfs/META
@@ -1,10 +1,10 @@
Meta: 1
Name: zfs
Branch: 1.0
Version: 2.1.99
Release: 1
Release-Tags: relext
License: CDDL
Author: OpenZFS
-Linux-Maximum: 6.1
+Linux-Maximum: 6.2
Linux-Minimum: 3.10
diff --git a/sys/contrib/openzfs/cmd/zdb/zdb.c b/sys/contrib/openzfs/cmd/zdb/zdb.c
index e87826f7467b..ec5d1acacf85 100644
--- a/sys/contrib/openzfs/cmd/zdb/zdb.c
+++ b/sys/contrib/openzfs/cmd/zdb/zdb.c
@@ -1,9208 +1,9243 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2019 by Delphix. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2016 Nexenta Systems, Inc.
* Copyright (c) 2017, 2018 Lawrence Livermore National Security, LLC.
* Copyright (c) 2015, 2017, Intel Corporation.
* Copyright (c) 2020 Datto Inc.
* Copyright (c) 2020, The FreeBSD Foundation [1]
*
* [1] Portions of this software were developed by Allan Jude
* under sponsorship from the FreeBSD Foundation.
* Copyright (c) 2021 Allan Jude
* Copyright (c) 2021 Toomas Soome <tsoome@me.com>
*/
#include <stdio.h>
#include <unistd.h>
#include <stdlib.h>
#include <ctype.h>
#include <getopt.h>
#include <openssl/evp.h>
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/dmu.h>
#include <sys/zap.h>
#include <sys/fs/zfs.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_sa.h>
#include <sys/sa.h>
#include <sys/sa_impl.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
#include <sys/metaslab_impl.h>
#include <sys/dmu_objset.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_pool.h>
#include <sys/dsl_bookmark.h>
#include <sys/dbuf.h>
#include <sys/zil.h>
#include <sys/zil_impl.h>
#include <sys/stat.h>
#include <sys/resource.h>
#include <sys/dmu_send.h>
#include <sys/dmu_traverse.h>
#include <sys/zio_checksum.h>
#include <sys/zio_compress.h>
#include <sys/zfs_fuid.h>
#include <sys/arc.h>
#include <sys/arc_impl.h>
#include <sys/ddt.h>
#include <sys/zfeature.h>
#include <sys/abd.h>
#include <sys/blkptr.h>
#include <sys/dsl_crypt.h>
#include <sys/dsl_scan.h>
#include <sys/btree.h>
#include <zfs_comutil.h>
#include <sys/zstd/zstd.h>
#include <libnvpair.h>
#include <libzutil.h>
#include "zdb.h"
#define ZDB_COMPRESS_NAME(idx) ((idx) < ZIO_COMPRESS_FUNCTIONS ? \
zio_compress_table[(idx)].ci_name : "UNKNOWN")
#define ZDB_CHECKSUM_NAME(idx) ((idx) < ZIO_CHECKSUM_FUNCTIONS ? \
zio_checksum_table[(idx)].ci_name : "UNKNOWN")
#define ZDB_OT_TYPE(idx) ((idx) < DMU_OT_NUMTYPES ? (idx) : \
(idx) == DMU_OTN_ZAP_DATA || (idx) == DMU_OTN_ZAP_METADATA ? \
DMU_OT_ZAP_OTHER : \
(idx) == DMU_OTN_UINT64_DATA || (idx) == DMU_OTN_UINT64_METADATA ? \
DMU_OT_UINT64_OTHER : DMU_OT_NUMTYPES)
/* Some platforms require part of inode IDs to be remapped */
#ifdef __APPLE__
#define ZDB_MAP_OBJECT_ID(obj) INO_XNUTOZFS(obj, 2)
#else
#define ZDB_MAP_OBJECT_ID(obj) (obj)
#endif
static const char *
zdb_ot_name(dmu_object_type_t type)
{
if (type < DMU_OT_NUMTYPES)
return (dmu_ot[type].ot_name);
else if ((type & DMU_OT_NEWTYPE) &&
((type & DMU_OT_BYTESWAP_MASK) < DMU_BSWAP_NUMFUNCS))
return (dmu_ot_byteswap[type & DMU_OT_BYTESWAP_MASK].ob_name);
else
return ("UNKNOWN");
}
extern int reference_tracking_enable;
extern int zfs_recover;
extern uint_t zfs_vdev_async_read_max_active;
extern boolean_t spa_load_verify_dryrun;
extern boolean_t spa_mode_readable_spacemaps;
extern uint_t zfs_reconstruct_indirect_combinations_max;
extern uint_t zfs_btree_verify_intensity;
static const char cmdname[] = "zdb";
uint8_t dump_opt[256];
typedef void object_viewer_t(objset_t *, uint64_t, void *data, size_t size);
static uint64_t *zopt_metaslab = NULL;
static unsigned zopt_metaslab_args = 0;
typedef struct zopt_object_range {
uint64_t zor_obj_start;
uint64_t zor_obj_end;
uint64_t zor_flags;
} zopt_object_range_t;
static zopt_object_range_t *zopt_object_ranges = NULL;
static unsigned zopt_object_args = 0;
static int flagbits[256];
#define ZOR_FLAG_PLAIN_FILE 0x0001
#define ZOR_FLAG_DIRECTORY 0x0002
#define ZOR_FLAG_SPACE_MAP 0x0004
#define ZOR_FLAG_ZAP 0x0008
#define ZOR_FLAG_ALL_TYPES -1
#define ZOR_SUPPORTED_FLAGS (ZOR_FLAG_PLAIN_FILE | \
ZOR_FLAG_DIRECTORY | \
ZOR_FLAG_SPACE_MAP | \
ZOR_FLAG_ZAP)
#define ZDB_FLAG_CHECKSUM 0x0001
#define ZDB_FLAG_DECOMPRESS 0x0002
#define ZDB_FLAG_BSWAP 0x0004
#define ZDB_FLAG_GBH 0x0008
#define ZDB_FLAG_INDIRECT 0x0010
#define ZDB_FLAG_RAW 0x0020
#define ZDB_FLAG_PRINT_BLKPTR 0x0040
#define ZDB_FLAG_VERBOSE 0x0080
static uint64_t max_inflight_bytes = 256 * 1024 * 1024; /* 256MB */
static int leaked_objects = 0;
static range_tree_t *mos_refd_objs;
static void snprintf_blkptr_compact(char *, size_t, const blkptr_t *,
boolean_t);
static void mos_obj_refd(uint64_t);
static void mos_obj_refd_multiple(uint64_t);
static int dump_bpobj_cb(void *arg, const blkptr_t *bp, boolean_t free,
dmu_tx_t *tx);
typedef struct sublivelist_verify {
/* FREE's that haven't yet matched to an ALLOC, in one sub-livelist */
zfs_btree_t sv_pair;
/* ALLOC's without a matching FREE, accumulates across sub-livelists */
zfs_btree_t sv_leftover;
} sublivelist_verify_t;
static int
livelist_compare(const void *larg, const void *rarg)
{
const blkptr_t *l = larg;
const blkptr_t *r = rarg;
/* Sort them according to dva[0] */
uint64_t l_dva0_vdev, r_dva0_vdev;
l_dva0_vdev = DVA_GET_VDEV(&l->blk_dva[0]);
r_dva0_vdev = DVA_GET_VDEV(&r->blk_dva[0]);
if (l_dva0_vdev < r_dva0_vdev)
return (-1);
else if (l_dva0_vdev > r_dva0_vdev)
return (+1);
/* if vdevs are equal, sort by offsets. */
uint64_t l_dva0_offset;
uint64_t r_dva0_offset;
l_dva0_offset = DVA_GET_OFFSET(&l->blk_dva[0]);
r_dva0_offset = DVA_GET_OFFSET(&r->blk_dva[0]);
if (l_dva0_offset < r_dva0_offset) {
return (-1);
} else if (l_dva0_offset > r_dva0_offset) {
return (+1);
}
/*
* Since we're storing blkptrs without cancelling FREE/ALLOC pairs,
* it's possible the offsets are equal. In that case, sort by txg
*/
if (l->blk_birth < r->blk_birth) {
return (-1);
} else if (l->blk_birth > r->blk_birth) {
return (+1);
}
return (0);
}
typedef struct sublivelist_verify_block {
dva_t svb_dva;
/*
* We need this to check if the block marked as allocated
* in the livelist was freed (and potentially reallocated)
* in the metaslab spacemaps at a later TXG.
*/
uint64_t svb_allocated_txg;
} sublivelist_verify_block_t;
static void zdb_print_blkptr(const blkptr_t *bp, int flags);
typedef struct sublivelist_verify_block_refcnt {
/* block pointer entry in livelist being verified */
blkptr_t svbr_blk;
/*
* Refcount gets incremented to 1 when we encounter the first
* FREE entry for the svfbr block pointer and a node for it
* is created in our ZDB verification/tracking metadata.
*
* As we encounter more FREE entries we increment this counter
* and similarly decrement it whenever we find the respective
* ALLOC entries for this block.
*
* When the refcount gets to 0 it means that all the FREE and
* ALLOC entries of this block have paired up and we no longer
* need to track it in our verification logic (e.g. the node
* containing this struct in our verification data structure
* should be freed).
*
* [refer to sublivelist_verify_blkptr() for the actual code]
*/
uint32_t svbr_refcnt;
} sublivelist_verify_block_refcnt_t;
static int
sublivelist_block_refcnt_compare(const void *larg, const void *rarg)
{
const sublivelist_verify_block_refcnt_t *l = larg;
const sublivelist_verify_block_refcnt_t *r = rarg;
return (livelist_compare(&l->svbr_blk, &r->svbr_blk));
}
static int
sublivelist_verify_blkptr(void *arg, const blkptr_t *bp, boolean_t free,
dmu_tx_t *tx)
{
ASSERT3P(tx, ==, NULL);
struct sublivelist_verify *sv = arg;
sublivelist_verify_block_refcnt_t current = {
.svbr_blk = *bp,
/*
* Start with 1 in case this is the first free entry.
* This field is not used for our B-Tree comparisons
* anyway.
*/
.svbr_refcnt = 1,
};
zfs_btree_index_t where;
sublivelist_verify_block_refcnt_t *pair =
zfs_btree_find(&sv->sv_pair, &current, &where);
if (free) {
if (pair == NULL) {
/* first free entry for this block pointer */
zfs_btree_add(&sv->sv_pair, &current);
} else {
pair->svbr_refcnt++;
}
} else {
if (pair == NULL) {
/* block that is currently marked as allocated */
for (int i = 0; i < SPA_DVAS_PER_BP; i++) {
if (DVA_IS_EMPTY(&bp->blk_dva[i]))
break;
sublivelist_verify_block_t svb = {
.svb_dva = bp->blk_dva[i],
.svb_allocated_txg = bp->blk_birth
};
if (zfs_btree_find(&sv->sv_leftover, &svb,
&where) == NULL) {
zfs_btree_add_idx(&sv->sv_leftover,
&svb, &where);
}
}
} else {
/* alloc matches a free entry */
pair->svbr_refcnt--;
if (pair->svbr_refcnt == 0) {
/* all allocs and frees have been matched */
zfs_btree_remove_idx(&sv->sv_pair, &where);
}
}
}
return (0);
}
static int
sublivelist_verify_func(void *args, dsl_deadlist_entry_t *dle)
{
int err;
struct sublivelist_verify *sv = args;
zfs_btree_create(&sv->sv_pair, sublivelist_block_refcnt_compare,
sizeof (sublivelist_verify_block_refcnt_t));
err = bpobj_iterate_nofree(&dle->dle_bpobj, sublivelist_verify_blkptr,
sv, NULL);
sublivelist_verify_block_refcnt_t *e;
zfs_btree_index_t *cookie = NULL;
while ((e = zfs_btree_destroy_nodes(&sv->sv_pair, &cookie)) != NULL) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf),
&e->svbr_blk, B_TRUE);
(void) printf("\tERROR: %d unmatched FREE(s): %s\n",
e->svbr_refcnt, blkbuf);
}
zfs_btree_destroy(&sv->sv_pair);
return (err);
}
static int
livelist_block_compare(const void *larg, const void *rarg)
{
const sublivelist_verify_block_t *l = larg;
const sublivelist_verify_block_t *r = rarg;
if (DVA_GET_VDEV(&l->svb_dva) < DVA_GET_VDEV(&r->svb_dva))
return (-1);
else if (DVA_GET_VDEV(&l->svb_dva) > DVA_GET_VDEV(&r->svb_dva))
return (+1);
if (DVA_GET_OFFSET(&l->svb_dva) < DVA_GET_OFFSET(&r->svb_dva))
return (-1);
else if (DVA_GET_OFFSET(&l->svb_dva) > DVA_GET_OFFSET(&r->svb_dva))
return (+1);
if (DVA_GET_ASIZE(&l->svb_dva) < DVA_GET_ASIZE(&r->svb_dva))
return (-1);
else if (DVA_GET_ASIZE(&l->svb_dva) > DVA_GET_ASIZE(&r->svb_dva))
return (+1);
return (0);
}
/*
* Check for errors in a livelist while tracking all unfreed ALLOCs in the
* sublivelist_verify_t: sv->sv_leftover
*/
static void
livelist_verify(dsl_deadlist_t *dl, void *arg)
{
sublivelist_verify_t *sv = arg;
dsl_deadlist_iterate(dl, sublivelist_verify_func, sv);
}
/*
* Check for errors in the livelist entry and discard the intermediary
* data structures
*/
static int
sublivelist_verify_lightweight(void *args, dsl_deadlist_entry_t *dle)
{
(void) args;
sublivelist_verify_t sv;
zfs_btree_create(&sv.sv_leftover, livelist_block_compare,
sizeof (sublivelist_verify_block_t));
int err = sublivelist_verify_func(&sv, dle);
zfs_btree_clear(&sv.sv_leftover);
zfs_btree_destroy(&sv.sv_leftover);
return (err);
}
typedef struct metaslab_verify {
/*
* Tree containing all the leftover ALLOCs from the livelists
* that are part of this metaslab.
*/
zfs_btree_t mv_livelist_allocs;
/*
* Metaslab information.
*/
uint64_t mv_vdid;
uint64_t mv_msid;
uint64_t mv_start;
uint64_t mv_end;
/*
* What's currently allocated for this metaslab.
*/
range_tree_t *mv_allocated;
} metaslab_verify_t;
typedef void ll_iter_t(dsl_deadlist_t *ll, void *arg);
typedef int (*zdb_log_sm_cb_t)(spa_t *spa, space_map_entry_t *sme, uint64_t txg,
void *arg);
typedef struct unflushed_iter_cb_arg {
spa_t *uic_spa;
uint64_t uic_txg;
void *uic_arg;
zdb_log_sm_cb_t uic_cb;
} unflushed_iter_cb_arg_t;
static int
iterate_through_spacemap_logs_cb(space_map_entry_t *sme, void *arg)
{
unflushed_iter_cb_arg_t *uic = arg;
return (uic->uic_cb(uic->uic_spa, sme, uic->uic_txg, uic->uic_arg));
}
static void
iterate_through_spacemap_logs(spa_t *spa, zdb_log_sm_cb_t cb, void *arg)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP))
return;
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
for (spa_log_sm_t *sls = avl_first(&spa->spa_sm_logs_by_txg);
sls; sls = AVL_NEXT(&spa->spa_sm_logs_by_txg, sls)) {
space_map_t *sm = NULL;
VERIFY0(space_map_open(&sm, spa_meta_objset(spa),
sls->sls_sm_obj, 0, UINT64_MAX, SPA_MINBLOCKSHIFT));
unflushed_iter_cb_arg_t uic = {
.uic_spa = spa,
.uic_txg = sls->sls_txg,
.uic_arg = arg,
.uic_cb = cb
};
VERIFY0(space_map_iterate(sm, space_map_length(sm),
iterate_through_spacemap_logs_cb, &uic));
space_map_close(sm);
}
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
static void
verify_livelist_allocs(metaslab_verify_t *mv, uint64_t txg,
uint64_t offset, uint64_t size)
{
sublivelist_verify_block_t svb = {{{0}}};
DVA_SET_VDEV(&svb.svb_dva, mv->mv_vdid);
DVA_SET_OFFSET(&svb.svb_dva, offset);
DVA_SET_ASIZE(&svb.svb_dva, size);
zfs_btree_index_t where;
uint64_t end_offset = offset + size;
/*
* Look for an exact match for spacemap entry in the livelist entries.
* Then, look for other livelist entries that fall within the range
* of the spacemap entry as it may have been condensed
*/
sublivelist_verify_block_t *found =
zfs_btree_find(&mv->mv_livelist_allocs, &svb, &where);
if (found == NULL) {
found = zfs_btree_next(&mv->mv_livelist_allocs, &where, &where);
}
for (; found != NULL && DVA_GET_VDEV(&found->svb_dva) == mv->mv_vdid &&
DVA_GET_OFFSET(&found->svb_dva) < end_offset;
found = zfs_btree_next(&mv->mv_livelist_allocs, &where, &where)) {
if (found->svb_allocated_txg <= txg) {
(void) printf("ERROR: Livelist ALLOC [%llx:%llx] "
"from TXG %llx FREED at TXG %llx\n",
(u_longlong_t)DVA_GET_OFFSET(&found->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&found->svb_dva),
(u_longlong_t)found->svb_allocated_txg,
(u_longlong_t)txg);
}
}
}
static int
metaslab_spacemap_validation_cb(space_map_entry_t *sme, void *arg)
{
metaslab_verify_t *mv = arg;
uint64_t offset = sme->sme_offset;
uint64_t size = sme->sme_run;
uint64_t txg = sme->sme_txg;
if (sme->sme_type == SM_ALLOC) {
if (range_tree_contains(mv->mv_allocated,
offset, size)) {
(void) printf("ERROR: DOUBLE ALLOC: "
"%llu [%llx:%llx] "
"%llu:%llu LOG_SM\n",
(u_longlong_t)txg, (u_longlong_t)offset,
(u_longlong_t)size, (u_longlong_t)mv->mv_vdid,
(u_longlong_t)mv->mv_msid);
} else {
range_tree_add(mv->mv_allocated,
offset, size);
}
} else {
if (!range_tree_contains(mv->mv_allocated,
offset, size)) {
(void) printf("ERROR: DOUBLE FREE: "
"%llu [%llx:%llx] "
"%llu:%llu LOG_SM\n",
(u_longlong_t)txg, (u_longlong_t)offset,
(u_longlong_t)size, (u_longlong_t)mv->mv_vdid,
(u_longlong_t)mv->mv_msid);
} else {
range_tree_remove(mv->mv_allocated,
offset, size);
}
}
if (sme->sme_type != SM_ALLOC) {
/*
* If something is freed in the spacemap, verify that
* it is not listed as allocated in the livelist.
*/
verify_livelist_allocs(mv, txg, offset, size);
}
return (0);
}
static int
spacemap_check_sm_log_cb(spa_t *spa, space_map_entry_t *sme,
uint64_t txg, void *arg)
{
metaslab_verify_t *mv = arg;
uint64_t offset = sme->sme_offset;
uint64_t vdev_id = sme->sme_vdev;
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
/* skip indirect vdevs */
if (!vdev_is_concrete(vd))
return (0);
if (vdev_id != mv->mv_vdid)
return (0);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
if (ms->ms_id != mv->mv_msid)
return (0);
if (txg < metaslab_unflushed_txg(ms))
return (0);
ASSERT3U(txg, ==, sme->sme_txg);
return (metaslab_spacemap_validation_cb(sme, mv));
}
static void
spacemap_check_sm_log(spa_t *spa, metaslab_verify_t *mv)
{
iterate_through_spacemap_logs(spa, spacemap_check_sm_log_cb, mv);
}
static void
spacemap_check_ms_sm(space_map_t *sm, metaslab_verify_t *mv)
{
if (sm == NULL)
return;
VERIFY0(space_map_iterate(sm, space_map_length(sm),
metaslab_spacemap_validation_cb, mv));
}
static void iterate_deleted_livelists(spa_t *spa, ll_iter_t func, void *arg);
/*
* Transfer blocks from sv_leftover tree to the mv_livelist_allocs if
* they are part of that metaslab (mv_msid).
*/
static void
mv_populate_livelist_allocs(metaslab_verify_t *mv, sublivelist_verify_t *sv)
{
zfs_btree_index_t where;
sublivelist_verify_block_t *svb;
ASSERT3U(zfs_btree_numnodes(&mv->mv_livelist_allocs), ==, 0);
for (svb = zfs_btree_first(&sv->sv_leftover, &where);
svb != NULL;
svb = zfs_btree_next(&sv->sv_leftover, &where, &where)) {
if (DVA_GET_VDEV(&svb->svb_dva) != mv->mv_vdid)
continue;
if (DVA_GET_OFFSET(&svb->svb_dva) < mv->mv_start &&
(DVA_GET_OFFSET(&svb->svb_dva) +
DVA_GET_ASIZE(&svb->svb_dva)) > mv->mv_start) {
(void) printf("ERROR: Found block that crosses "
"metaslab boundary: <%llu:%llx:%llx>\n",
(u_longlong_t)DVA_GET_VDEV(&svb->svb_dva),
(u_longlong_t)DVA_GET_OFFSET(&svb->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&svb->svb_dva));
continue;
}
if (DVA_GET_OFFSET(&svb->svb_dva) < mv->mv_start)
continue;
if (DVA_GET_OFFSET(&svb->svb_dva) >= mv->mv_end)
continue;
if ((DVA_GET_OFFSET(&svb->svb_dva) +
DVA_GET_ASIZE(&svb->svb_dva)) > mv->mv_end) {
(void) printf("ERROR: Found block that crosses "
"metaslab boundary: <%llu:%llx:%llx>\n",
(u_longlong_t)DVA_GET_VDEV(&svb->svb_dva),
(u_longlong_t)DVA_GET_OFFSET(&svb->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&svb->svb_dva));
continue;
}
zfs_btree_add(&mv->mv_livelist_allocs, svb);
}
for (svb = zfs_btree_first(&mv->mv_livelist_allocs, &where);
svb != NULL;
svb = zfs_btree_next(&mv->mv_livelist_allocs, &where, &where)) {
zfs_btree_remove(&sv->sv_leftover, svb);
}
}
/*
* [Livelist Check]
* Iterate through all the sublivelists and:
* - report leftover frees (**)
* - record leftover ALLOCs together with their TXG [see Cross Check]
*
* (**) Note: Double ALLOCs are valid in datasets that have dedup
* enabled. Similarly double FREEs are allowed as well but
* only if they pair up with a corresponding ALLOC entry once
* we our done with our sublivelist iteration.
*
* [Spacemap Check]
* for each metaslab:
* - iterate over spacemap and then the metaslab's entries in the
* spacemap log, then report any double FREEs and ALLOCs (do not
* blow up).
*
* [Cross Check]
* After finishing the Livelist Check phase and while being in the
* Spacemap Check phase, we find all the recorded leftover ALLOCs
* of the livelist check that are part of the metaslab that we are
* currently looking at in the Spacemap Check. We report any entries
* that are marked as ALLOCs in the livelists but have been actually
* freed (and potentially allocated again) after their TXG stamp in
* the spacemaps. Also report any ALLOCs from the livelists that
* belong to indirect vdevs (e.g. their vdev completed removal).
*
* Note that this will miss Log Spacemap entries that cancelled each other
* out before being flushed to the metaslab, so we are not guaranteed
* to match all erroneous ALLOCs.
*/
static void
livelist_metaslab_validate(spa_t *spa)
{
(void) printf("Verifying deleted livelist entries\n");
sublivelist_verify_t sv;
zfs_btree_create(&sv.sv_leftover, livelist_block_compare,
sizeof (sublivelist_verify_block_t));
iterate_deleted_livelists(spa, livelist_verify, &sv);
(void) printf("Verifying metaslab entries\n");
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
if (!vdev_is_concrete(vd))
continue;
for (uint64_t mid = 0; mid < vd->vdev_ms_count; mid++) {
metaslab_t *m = vd->vdev_ms[mid];
(void) fprintf(stderr,
"\rverifying concrete vdev %llu, "
"metaslab %llu of %llu ...",
(longlong_t)vd->vdev_id,
(longlong_t)mid,
(longlong_t)vd->vdev_ms_count);
uint64_t shift, start;
range_seg_type_t type =
metaslab_calculate_range_tree_type(vd, m,
&start, &shift);
metaslab_verify_t mv;
mv.mv_allocated = range_tree_create(NULL,
type, NULL, start, shift);
mv.mv_vdid = vd->vdev_id;
mv.mv_msid = m->ms_id;
mv.mv_start = m->ms_start;
mv.mv_end = m->ms_start + m->ms_size;
zfs_btree_create(&mv.mv_livelist_allocs,
livelist_block_compare,
sizeof (sublivelist_verify_block_t));
mv_populate_livelist_allocs(&mv, &sv);
spacemap_check_ms_sm(m->ms_sm, &mv);
spacemap_check_sm_log(spa, &mv);
range_tree_vacate(mv.mv_allocated, NULL, NULL);
range_tree_destroy(mv.mv_allocated);
zfs_btree_clear(&mv.mv_livelist_allocs);
zfs_btree_destroy(&mv.mv_livelist_allocs);
}
}
(void) fprintf(stderr, "\n");
/*
* If there are any segments in the leftover tree after we walked
* through all the metaslabs in the concrete vdevs then this means
* that we have segments in the livelists that belong to indirect
* vdevs and are marked as allocated.
*/
if (zfs_btree_numnodes(&sv.sv_leftover) == 0) {
zfs_btree_destroy(&sv.sv_leftover);
return;
}
(void) printf("ERROR: Found livelist blocks marked as allocated "
"for indirect vdevs:\n");
zfs_btree_index_t *where = NULL;
sublivelist_verify_block_t *svb;
while ((svb = zfs_btree_destroy_nodes(&sv.sv_leftover, &where)) !=
NULL) {
int vdev_id = DVA_GET_VDEV(&svb->svb_dva);
ASSERT3U(vdev_id, <, rvd->vdev_children);
vdev_t *vd = rvd->vdev_child[vdev_id];
ASSERT(!vdev_is_concrete(vd));
(void) printf("<%d:%llx:%llx> TXG %llx\n",
vdev_id, (u_longlong_t)DVA_GET_OFFSET(&svb->svb_dva),
(u_longlong_t)DVA_GET_ASIZE(&svb->svb_dva),
(u_longlong_t)svb->svb_allocated_txg);
}
(void) printf("\n");
zfs_btree_destroy(&sv.sv_leftover);
}
/*
* These libumem hooks provide a reasonable set of defaults for the allocator's
* debugging facilities.
*/
const char *
_umem_debug_init(void)
{
return ("default,verbose"); /* $UMEM_DEBUG setting */
}
const char *
_umem_logging_init(void)
{
return ("fail,contents"); /* $UMEM_LOGGING setting */
}
static void
usage(void)
{
(void) fprintf(stderr,
"Usage:\t%s [-AbcdDFGhikLMPsvXy] [-e [-V] [-p <path> ...]] "
"[-I <inflight I/Os>]\n"
"\t\t[-o <var>=<value>]... [-t <txg>] [-U <cache>] [-x <dumpdir>]\n"
"\t\t[-K <key>]\n"
"\t\t[<poolname>[/<dataset | objset id>] [<object | range> ...]]\n"
"\t%s [-AdiPv] [-e [-V] [-p <path> ...]] [-U <cache>] [-K <key>]\n"
"\t\t[<poolname>[/<dataset | objset id>] [<object | range> ...]\n"
"\t%s [-v] <bookmark>\n"
"\t%s -C [-A] [-U <cache>]\n"
"\t%s -l [-Aqu] <device>\n"
"\t%s -m [-AFLPX] [-e [-V] [-p <path> ...]] [-t <txg>] "
"[-U <cache>]\n\t\t<poolname> [<vdev> [<metaslab> ...]]\n"
"\t%s -O [-K <key>] <dataset> <path>\n"
"\t%s -r [-K <key>] <dataset> <path> <destination>\n"
"\t%s -R [-A] [-e [-V] [-p <path> ...]] [-U <cache>]\n"
"\t\t<poolname> <vdev>:<offset>:<size>[:<flags>]\n"
"\t%s -E [-A] word0:word1:...:word15\n"
"\t%s -S [-AP] [-e [-V] [-p <path> ...]] [-U <cache>] "
"<poolname>\n\n",
cmdname, cmdname, cmdname, cmdname, cmdname, cmdname, cmdname,
cmdname, cmdname, cmdname, cmdname);
(void) fprintf(stderr, " Dataset name must include at least one "
"separator character '/' or '@'\n");
(void) fprintf(stderr, " If dataset name is specified, only that "
"dataset is dumped\n");
(void) fprintf(stderr, " If object numbers or object number "
"ranges are specified, only those\n"
" objects or ranges are dumped.\n\n");
(void) fprintf(stderr,
" Object ranges take the form <start>:<end>[:<flags>]\n"
" start Starting object number\n"
" end Ending object number, or -1 for no upper bound\n"
" flags Optional flags to select object types:\n"
" A All objects (this is the default)\n"
" d ZFS directories\n"
" f ZFS files \n"
" m SPA space maps\n"
" z ZAPs\n"
" - Negate effect of next flag\n\n");
(void) fprintf(stderr, " Options to control amount of output:\n");
(void) fprintf(stderr, " -b --block-stats "
"block statistics\n");
(void) fprintf(stderr, " -c --checksum "
"checksum all metadata (twice for all data) blocks\n");
(void) fprintf(stderr, " -C --config "
"config (or cachefile if alone)\n");
(void) fprintf(stderr, " -d --datasets "
"dataset(s)\n");
(void) fprintf(stderr, " -D --dedup-stats "
"dedup statistics\n");
(void) fprintf(stderr, " -E --embedded-block-pointer=INTEGER\n"
" decode and display block "
"from an embedded block pointer\n");
(void) fprintf(stderr, " -h --history "
"pool history\n");
(void) fprintf(stderr, " -i --intent-logs "
"intent logs\n");
(void) fprintf(stderr, " -l --label "
"read label contents\n");
(void) fprintf(stderr, " -k --checkpointed-state "
"examine the checkpointed state of the pool\n");
(void) fprintf(stderr, " -L --disable-leak-tracking "
"disable leak tracking (do not load spacemaps)\n");
(void) fprintf(stderr, " -m --metaslabs "
"metaslabs\n");
(void) fprintf(stderr, " -M --metaslab-groups "
"metaslab groups\n");
(void) fprintf(stderr, " -O --object-lookups "
"perform object lookups by path\n");
(void) fprintf(stderr, " -r --copy-object "
"copy an object by path to file\n");
(void) fprintf(stderr, " -R --read-block "
"read and display block from a device\n");
(void) fprintf(stderr, " -s --io-stats "
"report stats on zdb's I/O\n");
(void) fprintf(stderr, " -S --simulate-dedup "
"simulate dedup to measure effect\n");
(void) fprintf(stderr, " -v --verbose "
"verbose (applies to all others)\n");
(void) fprintf(stderr, " -y --livelist "
"perform livelist and metaslab validation on any livelists being "
"deleted\n\n");
(void) fprintf(stderr, " Below options are intended for use "
"with other options:\n");
(void) fprintf(stderr, " -A --ignore-assertions "
"ignore assertions (-A), enable panic recovery (-AA) or both "
"(-AAA)\n");
(void) fprintf(stderr, " -e --exported "
"pool is exported/destroyed/has altroot/not in a cachefile\n");
(void) fprintf(stderr, " -F --automatic-rewind "
"attempt automatic rewind within safe range of transaction "
"groups\n");
(void) fprintf(stderr, " -G --dump-debug-msg "
"dump zfs_dbgmsg buffer before exiting\n");
(void) fprintf(stderr, " -I --inflight=INTEGER "
"specify the maximum number of checksumming I/Os "
"[default is 200]\n");
(void) fprintf(stderr, " -K --key=KEY "
"decryption key for encrypted dataset\n");
(void) fprintf(stderr, " -o --option=\"OPTION=INTEGER\" "
"set global variable to an unsigned 32-bit integer\n");
(void) fprintf(stderr, " -p --path==PATH "
"use one or more with -e to specify path to vdev dir\n");
(void) fprintf(stderr, " -P --parseable "
"print numbers in parseable form\n");
(void) fprintf(stderr, " -q --skip-label "
"don't print label contents\n");
(void) fprintf(stderr, " -t --txg=INTEGER "
"highest txg to use when searching for uberblocks\n");
(void) fprintf(stderr, " -u --uberblock "
"uberblock\n");
(void) fprintf(stderr, " -U --cachefile=PATH "
"use alternate cachefile\n");
(void) fprintf(stderr, " -V --verbatim "
"do verbatim import\n");
(void) fprintf(stderr, " -x --dump-blocks=PATH "
"dump all read blocks into specified directory\n");
(void) fprintf(stderr, " -X --extreme-rewind "
"attempt extreme rewind (does not work with dataset)\n");
(void) fprintf(stderr, " -Y --all-reconstruction "
"attempt all reconstruction combinations for split blocks\n");
(void) fprintf(stderr, " -Z --zstd-headers "
"show ZSTD headers \n");
(void) fprintf(stderr, "Specify an option more than once (e.g. -bb) "
"to make only that option verbose\n");
(void) fprintf(stderr, "Default is to dump everything non-verbosely\n");
exit(1);
}
static void
dump_debug_buffer(void)
{
if (dump_opt['G']) {
(void) printf("\n");
(void) fflush(stdout);
zfs_dbgmsg_print("zdb");
}
}
/*
* Called for usage errors that are discovered after a call to spa_open(),
* dmu_bonus_hold(), or pool_match(). abort() is called for other errors.
*/
static void
fatal(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
(void) fprintf(stderr, "%s: ", cmdname);
(void) vfprintf(stderr, fmt, ap);
va_end(ap);
(void) fprintf(stderr, "\n");
dump_debug_buffer();
exit(1);
}
static void
dump_packed_nvlist(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) size;
nvlist_t *nv;
size_t nvsize = *(uint64_t *)data;
char *packed = umem_alloc(nvsize, UMEM_NOFAIL);
VERIFY(0 == dmu_read(os, object, 0, nvsize, packed, DMU_READ_PREFETCH));
VERIFY(nvlist_unpack(packed, nvsize, &nv, 0) == 0);
umem_free(packed, nvsize);
dump_nvlist(nv, 8);
nvlist_free(nv);
}
static void
dump_history_offsets(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) size;
spa_history_phys_t *shp = data;
if (shp == NULL)
return;
(void) printf("\t\tpool_create_len = %llu\n",
(u_longlong_t)shp->sh_pool_create_len);
(void) printf("\t\tphys_max_off = %llu\n",
(u_longlong_t)shp->sh_phys_max_off);
(void) printf("\t\tbof = %llu\n",
(u_longlong_t)shp->sh_bof);
(void) printf("\t\teof = %llu\n",
(u_longlong_t)shp->sh_eof);
(void) printf("\t\trecords_lost = %llu\n",
(u_longlong_t)shp->sh_records_lost);
}
static void
zdb_nicenum(uint64_t num, char *buf, size_t buflen)
{
if (dump_opt['P'])
(void) snprintf(buf, buflen, "%llu", (longlong_t)num);
else
nicenum(num, buf, buflen);
}
static const char histo_stars[] = "****************************************";
static const uint64_t histo_width = sizeof (histo_stars) - 1;
static void
dump_histogram(const uint64_t *histo, int size, int offset)
{
int i;
int minidx = size - 1;
int maxidx = 0;
uint64_t max = 0;
for (i = 0; i < size; i++) {
if (histo[i] == 0)
continue;
if (histo[i] > max)
max = histo[i];
if (i > maxidx)
maxidx = i;
if (i < minidx)
minidx = i;
}
if (max < histo_width)
max = histo_width;
for (i = minidx; i <= maxidx; i++) {
(void) printf("\t\t\t%3u: %6llu %s\n",
i + offset, (u_longlong_t)histo[i],
&histo_stars[(max - histo[i]) * histo_width / max]);
}
}
static void
dump_zap_stats(objset_t *os, uint64_t object)
{
int error;
zap_stats_t zs;
error = zap_get_stats(os, object, &zs);
if (error)
return;
if (zs.zs_ptrtbl_len == 0) {
ASSERT(zs.zs_num_blocks == 1);
(void) printf("\tmicrozap: %llu bytes, %llu entries\n",
(u_longlong_t)zs.zs_blocksize,
(u_longlong_t)zs.zs_num_entries);
return;
}
(void) printf("\tFat ZAP stats:\n");
(void) printf("\t\tPointer table:\n");
(void) printf("\t\t\t%llu elements\n",
(u_longlong_t)zs.zs_ptrtbl_len);
(void) printf("\t\t\tzt_blk: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_zt_blk);
(void) printf("\t\t\tzt_numblks: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_zt_numblks);
(void) printf("\t\t\tzt_shift: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_zt_shift);
(void) printf("\t\t\tzt_blks_copied: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_blks_copied);
(void) printf("\t\t\tzt_nextblk: %llu\n",
(u_longlong_t)zs.zs_ptrtbl_nextblk);
(void) printf("\t\tZAP entries: %llu\n",
(u_longlong_t)zs.zs_num_entries);
(void) printf("\t\tLeaf blocks: %llu\n",
(u_longlong_t)zs.zs_num_leafs);
(void) printf("\t\tTotal blocks: %llu\n",
(u_longlong_t)zs.zs_num_blocks);
(void) printf("\t\tzap_block_type: 0x%llx\n",
(u_longlong_t)zs.zs_block_type);
(void) printf("\t\tzap_magic: 0x%llx\n",
(u_longlong_t)zs.zs_magic);
(void) printf("\t\tzap_salt: 0x%llx\n",
(u_longlong_t)zs.zs_salt);
(void) printf("\t\tLeafs with 2^n pointers:\n");
dump_histogram(zs.zs_leafs_with_2n_pointers, ZAP_HISTOGRAM_SIZE, 0);
(void) printf("\t\tBlocks with n*5 entries:\n");
dump_histogram(zs.zs_blocks_with_n5_entries, ZAP_HISTOGRAM_SIZE, 0);
(void) printf("\t\tBlocks n/10 full:\n");
dump_histogram(zs.zs_blocks_n_tenths_full, ZAP_HISTOGRAM_SIZE, 0);
(void) printf("\t\tEntries with n chunks:\n");
dump_histogram(zs.zs_entries_using_n_chunks, ZAP_HISTOGRAM_SIZE, 0);
(void) printf("\t\tBuckets with n entries:\n");
dump_histogram(zs.zs_buckets_with_n_entries, ZAP_HISTOGRAM_SIZE, 0);
}
static void
dump_none(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static void
dump_unknown(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
(void) printf("\tUNKNOWN OBJECT TYPE\n");
}
static void
dump_uint8(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static void
dump_uint64(objset_t *os, uint64_t object, void *data, size_t size)
{
uint64_t *arr;
uint64_t oursize;
if (dump_opt['d'] < 6)
return;
if (data == NULL) {
dmu_object_info_t doi;
VERIFY0(dmu_object_info(os, object, &doi));
size = doi.doi_max_offset;
/*
* We cap the size at 1 mebibyte here to prevent
* allocation failures and nigh-infinite printing if the
* object is extremely large.
*/
oursize = MIN(size, 1 << 20);
arr = kmem_alloc(oursize, KM_SLEEP);
int err = dmu_read(os, object, 0, oursize, arr, 0);
if (err != 0) {
(void) printf("got error %u from dmu_read\n", err);
kmem_free(arr, oursize);
return;
}
} else {
/*
* Even though the allocation is already done in this code path,
* we still cap the size to prevent excessive printing.
*/
oursize = MIN(size, 1 << 20);
arr = data;
}
if (size == 0) {
if (data == NULL)
kmem_free(arr, oursize);
(void) printf("\t\t[]\n");
return;
}
(void) printf("\t\t[%0llx", (u_longlong_t)arr[0]);
for (size_t i = 1; i * sizeof (uint64_t) < oursize; i++) {
if (i % 4 != 0)
(void) printf(", %0llx", (u_longlong_t)arr[i]);
else
(void) printf(",\n\t\t%0llx", (u_longlong_t)arr[i]);
}
if (oursize != size)
(void) printf(", ... ");
(void) printf("]\n");
if (data == NULL)
kmem_free(arr, oursize);
}
static void
dump_zap(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
zap_cursor_t zc;
zap_attribute_t attr;
void *prop;
unsigned i;
dump_zap_stats(os, object);
(void) printf("\n");
for (zap_cursor_init(&zc, os, object);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
(void) printf("\t\t%s = ", attr.za_name);
if (attr.za_num_integers == 0) {
(void) printf("\n");
continue;
}
prop = umem_zalloc(attr.za_num_integers *
attr.za_integer_length, UMEM_NOFAIL);
(void) zap_lookup(os, object, attr.za_name,
attr.za_integer_length, attr.za_num_integers, prop);
if (attr.za_integer_length == 1) {
if (strcmp(attr.za_name,
DSL_CRYPTO_KEY_MASTER_KEY) == 0 ||
strcmp(attr.za_name,
DSL_CRYPTO_KEY_HMAC_KEY) == 0 ||
strcmp(attr.za_name, DSL_CRYPTO_KEY_IV) == 0 ||
strcmp(attr.za_name, DSL_CRYPTO_KEY_MAC) == 0 ||
strcmp(attr.za_name, DMU_POOL_CHECKSUM_SALT) == 0) {
uint8_t *u8 = prop;
for (i = 0; i < attr.za_num_integers; i++) {
(void) printf("%02x", u8[i]);
}
} else {
(void) printf("%s", (char *)prop);
}
} else {
for (i = 0; i < attr.za_num_integers; i++) {
switch (attr.za_integer_length) {
case 2:
(void) printf("%u ",
((uint16_t *)prop)[i]);
break;
case 4:
(void) printf("%u ",
((uint32_t *)prop)[i]);
break;
case 8:
(void) printf("%lld ",
(u_longlong_t)((int64_t *)prop)[i]);
break;
}
}
}
(void) printf("\n");
umem_free(prop, attr.za_num_integers * attr.za_integer_length);
}
zap_cursor_fini(&zc);
}
static void
dump_bpobj(objset_t *os, uint64_t object, void *data, size_t size)
{
bpobj_phys_t *bpop = data;
uint64_t i;
char bytes[32], comp[32], uncomp[32];
/* make sure the output won't get truncated */
_Static_assert(sizeof (bytes) >= NN_NUMBUF_SZ, "bytes truncated");
_Static_assert(sizeof (comp) >= NN_NUMBUF_SZ, "comp truncated");
_Static_assert(sizeof (uncomp) >= NN_NUMBUF_SZ, "uncomp truncated");
if (bpop == NULL)
return;
zdb_nicenum(bpop->bpo_bytes, bytes, sizeof (bytes));
zdb_nicenum(bpop->bpo_comp, comp, sizeof (comp));
zdb_nicenum(bpop->bpo_uncomp, uncomp, sizeof (uncomp));
(void) printf("\t\tnum_blkptrs = %llu\n",
(u_longlong_t)bpop->bpo_num_blkptrs);
(void) printf("\t\tbytes = %s\n", bytes);
if (size >= BPOBJ_SIZE_V1) {
(void) printf("\t\tcomp = %s\n", comp);
(void) printf("\t\tuncomp = %s\n", uncomp);
}
if (size >= BPOBJ_SIZE_V2) {
(void) printf("\t\tsubobjs = %llu\n",
(u_longlong_t)bpop->bpo_subobjs);
(void) printf("\t\tnum_subobjs = %llu\n",
(u_longlong_t)bpop->bpo_num_subobjs);
}
if (size >= sizeof (*bpop)) {
(void) printf("\t\tnum_freed = %llu\n",
(u_longlong_t)bpop->bpo_num_freed);
}
if (dump_opt['d'] < 5)
return;
for (i = 0; i < bpop->bpo_num_blkptrs; i++) {
char blkbuf[BP_SPRINTF_LEN];
blkptr_t bp;
int err = dmu_read(os, object,
i * sizeof (bp), sizeof (bp), &bp, 0);
if (err != 0) {
(void) printf("got error %u from dmu_read\n", err);
break;
}
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf), &bp,
BP_GET_FREE(&bp));
(void) printf("\t%s\n", blkbuf);
}
}
static void
dump_bpobj_subobjs(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
dmu_object_info_t doi;
int64_t i;
VERIFY0(dmu_object_info(os, object, &doi));
uint64_t *subobjs = kmem_alloc(doi.doi_max_offset, KM_SLEEP);
int err = dmu_read(os, object, 0, doi.doi_max_offset, subobjs, 0);
if (err != 0) {
(void) printf("got error %u from dmu_read\n", err);
kmem_free(subobjs, doi.doi_max_offset);
return;
}
int64_t last_nonzero = -1;
for (i = 0; i < doi.doi_max_offset / 8; i++) {
if (subobjs[i] != 0)
last_nonzero = i;
}
for (i = 0; i <= last_nonzero; i++) {
(void) printf("\t%llu\n", (u_longlong_t)subobjs[i]);
}
kmem_free(subobjs, doi.doi_max_offset);
}
static void
dump_ddt_zap(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
dump_zap_stats(os, object);
/* contents are printed elsewhere, properly decoded */
}
static void
dump_sa_attrs(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
zap_cursor_t zc;
zap_attribute_t attr;
dump_zap_stats(os, object);
(void) printf("\n");
for (zap_cursor_init(&zc, os, object);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
(void) printf("\t\t%s = ", attr.za_name);
if (attr.za_num_integers == 0) {
(void) printf("\n");
continue;
}
(void) printf(" %llx : [%d:%d:%d]\n",
(u_longlong_t)attr.za_first_integer,
(int)ATTR_LENGTH(attr.za_first_integer),
(int)ATTR_BSWAP(attr.za_first_integer),
(int)ATTR_NUM(attr.za_first_integer));
}
zap_cursor_fini(&zc);
}
static void
dump_sa_layouts(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
zap_cursor_t zc;
zap_attribute_t attr;
uint16_t *layout_attrs;
unsigned i;
dump_zap_stats(os, object);
(void) printf("\n");
for (zap_cursor_init(&zc, os, object);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
(void) printf("\t\t%s = [", attr.za_name);
if (attr.za_num_integers == 0) {
(void) printf("\n");
continue;
}
VERIFY(attr.za_integer_length == 2);
layout_attrs = umem_zalloc(attr.za_num_integers *
attr.za_integer_length, UMEM_NOFAIL);
VERIFY(zap_lookup(os, object, attr.za_name,
attr.za_integer_length,
attr.za_num_integers, layout_attrs) == 0);
for (i = 0; i != attr.za_num_integers; i++)
(void) printf(" %d ", (int)layout_attrs[i]);
(void) printf("]\n");
umem_free(layout_attrs,
attr.za_num_integers * attr.za_integer_length);
}
zap_cursor_fini(&zc);
}
static void
dump_zpldir(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
zap_cursor_t zc;
zap_attribute_t attr;
const char *typenames[] = {
/* 0 */ "not specified",
/* 1 */ "FIFO",
/* 2 */ "Character Device",
/* 3 */ "3 (invalid)",
/* 4 */ "Directory",
/* 5 */ "5 (invalid)",
/* 6 */ "Block Device",
/* 7 */ "7 (invalid)",
/* 8 */ "Regular File",
/* 9 */ "9 (invalid)",
/* 10 */ "Symbolic Link",
/* 11 */ "11 (invalid)",
/* 12 */ "Socket",
/* 13 */ "Door",
/* 14 */ "Event Port",
/* 15 */ "15 (invalid)",
};
dump_zap_stats(os, object);
(void) printf("\n");
for (zap_cursor_init(&zc, os, object);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
(void) printf("\t\t%s = %lld (type: %s)\n",
attr.za_name, ZFS_DIRENT_OBJ(attr.za_first_integer),
typenames[ZFS_DIRENT_TYPE(attr.za_first_integer)]);
}
zap_cursor_fini(&zc);
}
static int
get_dtl_refcount(vdev_t *vd)
{
int refcount = 0;
if (vd->vdev_ops->vdev_op_leaf) {
space_map_t *sm = vd->vdev_dtl_sm;
if (sm != NULL &&
sm->sm_dbuf->db_size == sizeof (space_map_phys_t))
return (1);
return (0);
}
for (unsigned c = 0; c < vd->vdev_children; c++)
refcount += get_dtl_refcount(vd->vdev_child[c]);
return (refcount);
}
static int
get_metaslab_refcount(vdev_t *vd)
{
int refcount = 0;
if (vd->vdev_top == vd) {
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
space_map_t *sm = vd->vdev_ms[m]->ms_sm;
if (sm != NULL &&
sm->sm_dbuf->db_size == sizeof (space_map_phys_t))
refcount++;
}
}
for (unsigned c = 0; c < vd->vdev_children; c++)
refcount += get_metaslab_refcount(vd->vdev_child[c]);
return (refcount);
}
static int
get_obsolete_refcount(vdev_t *vd)
{
uint64_t obsolete_sm_object;
int refcount = 0;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (vd->vdev_top == vd && obsolete_sm_object != 0) {
dmu_object_info_t doi;
VERIFY0(dmu_object_info(vd->vdev_spa->spa_meta_objset,
obsolete_sm_object, &doi));
if (doi.doi_bonus_size == sizeof (space_map_phys_t)) {
refcount++;
}
} else {
ASSERT3P(vd->vdev_obsolete_sm, ==, NULL);
ASSERT3U(obsolete_sm_object, ==, 0);
}
for (unsigned c = 0; c < vd->vdev_children; c++) {
refcount += get_obsolete_refcount(vd->vdev_child[c]);
}
return (refcount);
}
static int
get_prev_obsolete_spacemap_refcount(spa_t *spa)
{
uint64_t prev_obj =
spa->spa_condensing_indirect_phys.scip_prev_obsolete_sm_object;
if (prev_obj != 0) {
dmu_object_info_t doi;
VERIFY0(dmu_object_info(spa->spa_meta_objset, prev_obj, &doi));
if (doi.doi_bonus_size == sizeof (space_map_phys_t)) {
return (1);
}
}
return (0);
}
static int
get_checkpoint_refcount(vdev_t *vd)
{
int refcount = 0;
if (vd->vdev_top == vd && vd->vdev_top_zap != 0 &&
zap_contains(spa_meta_objset(vd->vdev_spa),
vd->vdev_top_zap, VDEV_TOP_ZAP_POOL_CHECKPOINT_SM) == 0)
refcount++;
for (uint64_t c = 0; c < vd->vdev_children; c++)
refcount += get_checkpoint_refcount(vd->vdev_child[c]);
return (refcount);
}
static int
get_log_spacemap_refcount(spa_t *spa)
{
return (avl_numnodes(&spa->spa_sm_logs_by_txg));
}
static int
verify_spacemap_refcounts(spa_t *spa)
{
uint64_t expected_refcount = 0;
uint64_t actual_refcount;
(void) feature_get_refcount(spa,
&spa_feature_table[SPA_FEATURE_SPACEMAP_HISTOGRAM],
&expected_refcount);
actual_refcount = get_dtl_refcount(spa->spa_root_vdev);
actual_refcount += get_metaslab_refcount(spa->spa_root_vdev);
actual_refcount += get_obsolete_refcount(spa->spa_root_vdev);
actual_refcount += get_prev_obsolete_spacemap_refcount(spa);
actual_refcount += get_checkpoint_refcount(spa->spa_root_vdev);
actual_refcount += get_log_spacemap_refcount(spa);
if (expected_refcount != actual_refcount) {
(void) printf("space map refcount mismatch: expected %lld != "
"actual %lld\n",
(longlong_t)expected_refcount,
(longlong_t)actual_refcount);
return (2);
}
return (0);
}
static void
dump_spacemap(objset_t *os, space_map_t *sm)
{
const char *ddata[] = { "ALLOC", "FREE", "CONDENSE", "INVALID",
"INVALID", "INVALID", "INVALID", "INVALID" };
if (sm == NULL)
return;
(void) printf("space map object %llu:\n",
(longlong_t)sm->sm_object);
(void) printf(" smp_length = 0x%llx\n",
(longlong_t)sm->sm_phys->smp_length);
(void) printf(" smp_alloc = 0x%llx\n",
(longlong_t)sm->sm_phys->smp_alloc);
if (dump_opt['d'] < 6 && dump_opt['m'] < 4)
return;
/*
* Print out the freelist entries in both encoded and decoded form.
*/
uint8_t mapshift = sm->sm_shift;
int64_t alloc = 0;
uint64_t word, entry_id = 0;
for (uint64_t offset = 0; offset < space_map_length(sm);
offset += sizeof (word)) {
VERIFY0(dmu_read(os, space_map_object(sm), offset,
sizeof (word), &word, DMU_READ_PREFETCH));
if (sm_entry_is_debug(word)) {
uint64_t de_txg = SM_DEBUG_TXG_DECODE(word);
uint64_t de_sync_pass = SM_DEBUG_SYNCPASS_DECODE(word);
if (de_txg == 0) {
(void) printf(
"\t [%6llu] PADDING\n",
(u_longlong_t)entry_id);
} else {
(void) printf(
"\t [%6llu] %s: txg %llu pass %llu\n",
(u_longlong_t)entry_id,
ddata[SM_DEBUG_ACTION_DECODE(word)],
(u_longlong_t)de_txg,
(u_longlong_t)de_sync_pass);
}
entry_id++;
continue;
}
uint8_t words;
char entry_type;
uint64_t entry_off, entry_run, entry_vdev = SM_NO_VDEVID;
if (sm_entry_is_single_word(word)) {
entry_type = (SM_TYPE_DECODE(word) == SM_ALLOC) ?
'A' : 'F';
entry_off = (SM_OFFSET_DECODE(word) << mapshift) +
sm->sm_start;
entry_run = SM_RUN_DECODE(word) << mapshift;
words = 1;
} else {
/* it is a two-word entry so we read another word */
ASSERT(sm_entry_is_double_word(word));
uint64_t extra_word;
offset += sizeof (extra_word);
VERIFY0(dmu_read(os, space_map_object(sm), offset,
sizeof (extra_word), &extra_word,
DMU_READ_PREFETCH));
ASSERT3U(offset, <=, space_map_length(sm));
entry_run = SM2_RUN_DECODE(word) << mapshift;
entry_vdev = SM2_VDEV_DECODE(word);
entry_type = (SM2_TYPE_DECODE(extra_word) == SM_ALLOC) ?
'A' : 'F';
entry_off = (SM2_OFFSET_DECODE(extra_word) <<
mapshift) + sm->sm_start;
words = 2;
}
(void) printf("\t [%6llu] %c range:"
" %010llx-%010llx size: %06llx vdev: %06llu words: %u\n",
(u_longlong_t)entry_id,
entry_type, (u_longlong_t)entry_off,
(u_longlong_t)(entry_off + entry_run),
(u_longlong_t)entry_run,
(u_longlong_t)entry_vdev, words);
if (entry_type == 'A')
alloc += entry_run;
else
alloc -= entry_run;
entry_id++;
}
if (alloc != space_map_allocated(sm)) {
(void) printf("space_map_object alloc (%lld) INCONSISTENT "
"with space map summary (%lld)\n",
(longlong_t)space_map_allocated(sm), (longlong_t)alloc);
}
}
static void
dump_metaslab_stats(metaslab_t *msp)
{
char maxbuf[32];
range_tree_t *rt = msp->ms_allocatable;
zfs_btree_t *t = &msp->ms_allocatable_by_size;
int free_pct = range_tree_space(rt) * 100 / msp->ms_size;
/* max sure nicenum has enough space */
_Static_assert(sizeof (maxbuf) >= NN_NUMBUF_SZ, "maxbuf truncated");
zdb_nicenum(metaslab_largest_allocatable(msp), maxbuf, sizeof (maxbuf));
(void) printf("\t %25s %10lu %7s %6s %4s %4d%%\n",
"segments", zfs_btree_numnodes(t), "maxsize", maxbuf,
"freepct", free_pct);
(void) printf("\tIn-memory histogram:\n");
dump_histogram(rt->rt_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
}
static void
dump_metaslab(metaslab_t *msp)
{
vdev_t *vd = msp->ms_group->mg_vd;
spa_t *spa = vd->vdev_spa;
space_map_t *sm = msp->ms_sm;
char freebuf[32];
zdb_nicenum(msp->ms_size - space_map_allocated(sm), freebuf,
sizeof (freebuf));
(void) printf(
"\tmetaslab %6llu offset %12llx spacemap %6llu free %5s\n",
(u_longlong_t)msp->ms_id, (u_longlong_t)msp->ms_start,
(u_longlong_t)space_map_object(sm), freebuf);
if (dump_opt['m'] > 2 && !dump_opt['L']) {
mutex_enter(&msp->ms_lock);
VERIFY0(metaslab_load(msp));
range_tree_stat_verify(msp->ms_allocatable);
dump_metaslab_stats(msp);
metaslab_unload(msp);
mutex_exit(&msp->ms_lock);
}
if (dump_opt['m'] > 1 && sm != NULL &&
spa_feature_is_active(spa, SPA_FEATURE_SPACEMAP_HISTOGRAM)) {
/*
* The space map histogram represents free space in chunks
* of sm_shift (i.e. bucket 0 refers to 2^sm_shift).
*/
(void) printf("\tOn-disk histogram:\t\tfragmentation %llu\n",
(u_longlong_t)msp->ms_fragmentation);
dump_histogram(sm->sm_phys->smp_histogram,
SPACE_MAP_HISTOGRAM_SIZE, sm->sm_shift);
}
if (vd->vdev_ops == &vdev_draid_ops)
ASSERT3U(msp->ms_size, <=, 1ULL << vd->vdev_ms_shift);
else
ASSERT3U(msp->ms_size, ==, 1ULL << vd->vdev_ms_shift);
dump_spacemap(spa->spa_meta_objset, msp->ms_sm);
if (spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP)) {
(void) printf("\tFlush data:\n\tunflushed txg=%llu\n\n",
(u_longlong_t)metaslab_unflushed_txg(msp));
}
}
static void
print_vdev_metaslab_header(vdev_t *vd)
{
vdev_alloc_bias_t alloc_bias = vd->vdev_alloc_bias;
const char *bias_str = "";
if (alloc_bias == VDEV_BIAS_LOG || vd->vdev_islog) {
bias_str = VDEV_ALLOC_BIAS_LOG;
} else if (alloc_bias == VDEV_BIAS_SPECIAL) {
bias_str = VDEV_ALLOC_BIAS_SPECIAL;
} else if (alloc_bias == VDEV_BIAS_DEDUP) {
bias_str = VDEV_ALLOC_BIAS_DEDUP;
}
uint64_t ms_flush_data_obj = 0;
if (vd->vdev_top_zap != 0) {
int error = zap_lookup(spa_meta_objset(vd->vdev_spa),
vd->vdev_top_zap, VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS,
sizeof (uint64_t), 1, &ms_flush_data_obj);
if (error != ENOENT) {
ASSERT0(error);
}
}
(void) printf("\tvdev %10llu %s",
(u_longlong_t)vd->vdev_id, bias_str);
if (ms_flush_data_obj != 0) {
(void) printf(" ms_unflushed_phys object %llu",
(u_longlong_t)ms_flush_data_obj);
}
(void) printf("\n\t%-10s%5llu %-19s %-15s %-12s\n",
"metaslabs", (u_longlong_t)vd->vdev_ms_count,
"offset", "spacemap", "free");
(void) printf("\t%15s %19s %15s %12s\n",
"---------------", "-------------------",
"---------------", "------------");
}
static void
dump_metaslab_groups(spa_t *spa, boolean_t show_special)
{
vdev_t *rvd = spa->spa_root_vdev;
metaslab_class_t *mc = spa_normal_class(spa);
metaslab_class_t *smc = spa_special_class(spa);
uint64_t fragmentation;
metaslab_class_histogram_verify(mc);
for (unsigned c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = tvd->vdev_mg;
if (mg == NULL || (mg->mg_class != mc &&
(!show_special || mg->mg_class != smc)))
continue;
metaslab_group_histogram_verify(mg);
mg->mg_fragmentation = metaslab_group_fragmentation(mg);
(void) printf("\tvdev %10llu\t\tmetaslabs%5llu\t\t"
"fragmentation",
(u_longlong_t)tvd->vdev_id,
(u_longlong_t)tvd->vdev_ms_count);
if (mg->mg_fragmentation == ZFS_FRAG_INVALID) {
(void) printf("%3s\n", "-");
} else {
(void) printf("%3llu%%\n",
(u_longlong_t)mg->mg_fragmentation);
}
dump_histogram(mg->mg_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
}
(void) printf("\tpool %s\tfragmentation", spa_name(spa));
fragmentation = metaslab_class_fragmentation(mc);
if (fragmentation == ZFS_FRAG_INVALID)
(void) printf("\t%3s\n", "-");
else
(void) printf("\t%3llu%%\n", (u_longlong_t)fragmentation);
dump_histogram(mc->mc_histogram, RANGE_TREE_HISTOGRAM_SIZE, 0);
}
static void
print_vdev_indirect(vdev_t *vd)
{
vdev_indirect_config_t *vic = &vd->vdev_indirect_config;
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
vdev_indirect_births_t *vib = vd->vdev_indirect_births;
if (vim == NULL) {
ASSERT3P(vib, ==, NULL);
return;
}
ASSERT3U(vdev_indirect_mapping_object(vim), ==,
vic->vic_mapping_object);
ASSERT3U(vdev_indirect_births_object(vib), ==,
vic->vic_births_object);
(void) printf("indirect births obj %llu:\n",
(longlong_t)vic->vic_births_object);
(void) printf(" vib_count = %llu\n",
(longlong_t)vdev_indirect_births_count(vib));
for (uint64_t i = 0; i < vdev_indirect_births_count(vib); i++) {
vdev_indirect_birth_entry_phys_t *cur_vibe =
&vib->vib_entries[i];
(void) printf("\toffset %llx -> txg %llu\n",
(longlong_t)cur_vibe->vibe_offset,
(longlong_t)cur_vibe->vibe_phys_birth_txg);
}
(void) printf("\n");
(void) printf("indirect mapping obj %llu:\n",
(longlong_t)vic->vic_mapping_object);
(void) printf(" vim_max_offset = 0x%llx\n",
(longlong_t)vdev_indirect_mapping_max_offset(vim));
(void) printf(" vim_bytes_mapped = 0x%llx\n",
(longlong_t)vdev_indirect_mapping_bytes_mapped(vim));
(void) printf(" vim_count = %llu\n",
(longlong_t)vdev_indirect_mapping_num_entries(vim));
if (dump_opt['d'] <= 5 && dump_opt['m'] <= 3)
return;
uint32_t *counts = vdev_indirect_mapping_load_obsolete_counts(vim);
for (uint64_t i = 0; i < vdev_indirect_mapping_num_entries(vim); i++) {
vdev_indirect_mapping_entry_phys_t *vimep =
&vim->vim_entries[i];
(void) printf("\t<%llx:%llx:%llx> -> "
"<%llx:%llx:%llx> (%x obsolete)\n",
(longlong_t)vd->vdev_id,
(longlong_t)DVA_MAPPING_GET_SRC_OFFSET(vimep),
(longlong_t)DVA_GET_ASIZE(&vimep->vimep_dst),
(longlong_t)DVA_GET_VDEV(&vimep->vimep_dst),
(longlong_t)DVA_GET_OFFSET(&vimep->vimep_dst),
(longlong_t)DVA_GET_ASIZE(&vimep->vimep_dst),
counts[i]);
}
(void) printf("\n");
uint64_t obsolete_sm_object;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (obsolete_sm_object != 0) {
objset_t *mos = vd->vdev_spa->spa_meta_objset;
(void) printf("obsolete space map object %llu:\n",
(u_longlong_t)obsolete_sm_object);
ASSERT(vd->vdev_obsolete_sm != NULL);
ASSERT3U(space_map_object(vd->vdev_obsolete_sm), ==,
obsolete_sm_object);
dump_spacemap(mos, vd->vdev_obsolete_sm);
(void) printf("\n");
}
}
static void
dump_metaslabs(spa_t *spa)
{
vdev_t *vd, *rvd = spa->spa_root_vdev;
uint64_t m, c = 0, children = rvd->vdev_children;
(void) printf("\nMetaslabs:\n");
if (!dump_opt['d'] && zopt_metaslab_args > 0) {
c = zopt_metaslab[0];
if (c >= children)
(void) fatal("bad vdev id: %llu", (u_longlong_t)c);
if (zopt_metaslab_args > 1) {
vd = rvd->vdev_child[c];
print_vdev_metaslab_header(vd);
for (m = 1; m < zopt_metaslab_args; m++) {
if (zopt_metaslab[m] < vd->vdev_ms_count)
dump_metaslab(
vd->vdev_ms[zopt_metaslab[m]]);
else
(void) fprintf(stderr, "bad metaslab "
"number %llu\n",
(u_longlong_t)zopt_metaslab[m]);
}
(void) printf("\n");
return;
}
children = c + 1;
}
for (; c < children; c++) {
vd = rvd->vdev_child[c];
print_vdev_metaslab_header(vd);
print_vdev_indirect(vd);
for (m = 0; m < vd->vdev_ms_count; m++)
dump_metaslab(vd->vdev_ms[m]);
(void) printf("\n");
}
}
static void
dump_log_spacemaps(spa_t *spa)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP))
return;
(void) printf("\nLog Space Maps in Pool:\n");
for (spa_log_sm_t *sls = avl_first(&spa->spa_sm_logs_by_txg);
sls; sls = AVL_NEXT(&spa->spa_sm_logs_by_txg, sls)) {
space_map_t *sm = NULL;
VERIFY0(space_map_open(&sm, spa_meta_objset(spa),
sls->sls_sm_obj, 0, UINT64_MAX, SPA_MINBLOCKSHIFT));
(void) printf("Log Spacemap object %llu txg %llu\n",
(u_longlong_t)sls->sls_sm_obj, (u_longlong_t)sls->sls_txg);
dump_spacemap(spa->spa_meta_objset, sm);
space_map_close(sm);
}
(void) printf("\n");
}
static void
dump_dde(const ddt_t *ddt, const ddt_entry_t *dde, uint64_t index)
{
const ddt_phys_t *ddp = dde->dde_phys;
const ddt_key_t *ddk = &dde->dde_key;
const char *types[4] = { "ditto", "single", "double", "triple" };
char blkbuf[BP_SPRINTF_LEN];
blkptr_t blk;
int p;
for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
if (ddp->ddp_phys_birth == 0)
continue;
ddt_bp_create(ddt->ddt_checksum, ddk, ddp, &blk);
snprintf_blkptr(blkbuf, sizeof (blkbuf), &blk);
(void) printf("index %llx refcnt %llu %s %s\n",
(u_longlong_t)index, (u_longlong_t)ddp->ddp_refcnt,
types[p], blkbuf);
}
}
static void
dump_dedup_ratio(const ddt_stat_t *dds)
{
double rL, rP, rD, D, dedup, compress, copies;
if (dds->dds_blocks == 0)
return;
rL = (double)dds->dds_ref_lsize;
rP = (double)dds->dds_ref_psize;
rD = (double)dds->dds_ref_dsize;
D = (double)dds->dds_dsize;
dedup = rD / D;
compress = rL / rP;
copies = rD / rP;
(void) printf("dedup = %.2f, compress = %.2f, copies = %.2f, "
"dedup * compress / copies = %.2f\n\n",
dedup, compress, copies, dedup * compress / copies);
}
static void
dump_ddt(ddt_t *ddt, enum ddt_type type, enum ddt_class class)
{
char name[DDT_NAMELEN];
ddt_entry_t dde;
uint64_t walk = 0;
dmu_object_info_t doi;
uint64_t count, dspace, mspace;
int error;
error = ddt_object_info(ddt, type, class, &doi);
if (error == ENOENT)
return;
ASSERT(error == 0);
error = ddt_object_count(ddt, type, class, &count);
ASSERT(error == 0);
if (count == 0)
return;
dspace = doi.doi_physical_blocks_512 << 9;
mspace = doi.doi_fill_count * doi.doi_data_block_size;
ddt_object_name(ddt, type, class, name);
(void) printf("%s: %llu entries, size %llu on disk, %llu in core\n",
name,
(u_longlong_t)count,
(u_longlong_t)(dspace / count),
(u_longlong_t)(mspace / count));
if (dump_opt['D'] < 3)
return;
zpool_dump_ddt(NULL, &ddt->ddt_histogram[type][class]);
if (dump_opt['D'] < 4)
return;
if (dump_opt['D'] < 5 && class == DDT_CLASS_UNIQUE)
return;
(void) printf("%s contents:\n\n", name);
while ((error = ddt_object_walk(ddt, type, class, &walk, &dde)) == 0)
dump_dde(ddt, &dde, walk);
ASSERT3U(error, ==, ENOENT);
(void) printf("\n");
}
static void
dump_all_ddts(spa_t *spa)
{
ddt_histogram_t ddh_total = {{{0}}};
ddt_stat_t dds_total = {0};
for (enum zio_checksum c = 0; c < ZIO_CHECKSUM_FUNCTIONS; c++) {
ddt_t *ddt = spa->spa_ddt[c];
for (enum ddt_type type = 0; type < DDT_TYPES; type++) {
for (enum ddt_class class = 0; class < DDT_CLASSES;
class++) {
dump_ddt(ddt, type, class);
}
}
}
ddt_get_dedup_stats(spa, &dds_total);
if (dds_total.dds_blocks == 0) {
(void) printf("All DDTs are empty\n");
return;
}
(void) printf("\n");
if (dump_opt['D'] > 1) {
(void) printf("DDT histogram (aggregated over all DDTs):\n");
ddt_get_dedup_histogram(spa, &ddh_total);
zpool_dump_ddt(&dds_total, &ddh_total);
}
dump_dedup_ratio(&dds_total);
}
static void
dump_dtl_seg(void *arg, uint64_t start, uint64_t size)
{
char *prefix = arg;
(void) printf("%s [%llu,%llu) length %llu\n",
prefix,
(u_longlong_t)start,
(u_longlong_t)(start + size),
(u_longlong_t)(size));
}
static void
dump_dtl(vdev_t *vd, int indent)
{
spa_t *spa = vd->vdev_spa;
boolean_t required;
const char *name[DTL_TYPES] = { "missing", "partial", "scrub",
"outage" };
char prefix[256];
spa_vdev_state_enter(spa, SCL_NONE);
required = vdev_dtl_required(vd);
(void) spa_vdev_state_exit(spa, NULL, 0);
if (indent == 0)
(void) printf("\nDirty time logs:\n\n");
(void) printf("\t%*s%s [%s]\n", indent, "",
vd->vdev_path ? vd->vdev_path :
vd->vdev_parent ? vd->vdev_ops->vdev_op_type : spa_name(spa),
required ? "DTL-required" : "DTL-expendable");
for (int t = 0; t < DTL_TYPES; t++) {
range_tree_t *rt = vd->vdev_dtl[t];
if (range_tree_space(rt) == 0)
continue;
(void) snprintf(prefix, sizeof (prefix), "\t%*s%s",
indent + 2, "", name[t]);
range_tree_walk(rt, dump_dtl_seg, prefix);
if (dump_opt['d'] > 5 && vd->vdev_children == 0)
dump_spacemap(spa->spa_meta_objset,
vd->vdev_dtl_sm);
}
for (unsigned c = 0; c < vd->vdev_children; c++)
dump_dtl(vd->vdev_child[c], indent + 4);
}
static void
dump_history(spa_t *spa)
{
nvlist_t **events = NULL;
char *buf;
uint64_t resid, len, off = 0;
uint_t num = 0;
int error;
char tbuf[30];
if ((buf = malloc(SPA_OLD_MAXBLOCKSIZE)) == NULL) {
(void) fprintf(stderr, "%s: unable to allocate I/O buffer\n",
__func__);
return;
}
do {
len = SPA_OLD_MAXBLOCKSIZE;
if ((error = spa_history_get(spa, &off, &len, buf)) != 0) {
(void) fprintf(stderr, "Unable to read history: "
"error %d\n", error);
free(buf);
return;
}
if (zpool_history_unpack(buf, len, &resid, &events, &num) != 0)
break;
off -= resid;
} while (len != 0);
(void) printf("\nHistory:\n");
for (unsigned i = 0; i < num; i++) {
boolean_t printed = B_FALSE;
if (nvlist_exists(events[i], ZPOOL_HIST_TIME)) {
time_t tsec;
struct tm t;
tsec = fnvlist_lookup_uint64(events[i],
ZPOOL_HIST_TIME);
(void) localtime_r(&tsec, &t);
(void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t);
} else {
tbuf[0] = '\0';
}
if (nvlist_exists(events[i], ZPOOL_HIST_CMD)) {
(void) printf("%s %s\n", tbuf,
fnvlist_lookup_string(events[i], ZPOOL_HIST_CMD));
} else if (nvlist_exists(events[i], ZPOOL_HIST_INT_EVENT)) {
uint64_t ievent;
ievent = fnvlist_lookup_uint64(events[i],
ZPOOL_HIST_INT_EVENT);
if (ievent >= ZFS_NUM_LEGACY_HISTORY_EVENTS)
goto next;
(void) printf(" %s [internal %s txg:%ju] %s\n",
tbuf,
zfs_history_event_names[ievent],
fnvlist_lookup_uint64(events[i],
ZPOOL_HIST_TXG),
fnvlist_lookup_string(events[i],
ZPOOL_HIST_INT_STR));
} else if (nvlist_exists(events[i], ZPOOL_HIST_INT_NAME)) {
(void) printf("%s [txg:%ju] %s", tbuf,
fnvlist_lookup_uint64(events[i],
ZPOOL_HIST_TXG),
fnvlist_lookup_string(events[i],
ZPOOL_HIST_INT_NAME));
if (nvlist_exists(events[i], ZPOOL_HIST_DSNAME)) {
(void) printf(" %s (%llu)",
fnvlist_lookup_string(events[i],
ZPOOL_HIST_DSNAME),
(u_longlong_t)fnvlist_lookup_uint64(
events[i],
ZPOOL_HIST_DSID));
}
(void) printf(" %s\n", fnvlist_lookup_string(events[i],
ZPOOL_HIST_INT_STR));
} else if (nvlist_exists(events[i], ZPOOL_HIST_IOCTL)) {
(void) printf("%s ioctl %s\n", tbuf,
fnvlist_lookup_string(events[i],
ZPOOL_HIST_IOCTL));
if (nvlist_exists(events[i], ZPOOL_HIST_INPUT_NVL)) {
(void) printf(" input:\n");
dump_nvlist(fnvlist_lookup_nvlist(events[i],
ZPOOL_HIST_INPUT_NVL), 8);
}
if (nvlist_exists(events[i], ZPOOL_HIST_OUTPUT_NVL)) {
(void) printf(" output:\n");
dump_nvlist(fnvlist_lookup_nvlist(events[i],
ZPOOL_HIST_OUTPUT_NVL), 8);
}
if (nvlist_exists(events[i], ZPOOL_HIST_ERRNO)) {
(void) printf(" errno: %lld\n",
(longlong_t)fnvlist_lookup_int64(events[i],
ZPOOL_HIST_ERRNO));
}
} else {
goto next;
}
printed = B_TRUE;
next:
if (dump_opt['h'] > 1) {
if (!printed)
(void) printf("unrecognized record:\n");
dump_nvlist(events[i], 2);
}
}
free(buf);
}
static void
dump_dnode(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static uint64_t
blkid2offset(const dnode_phys_t *dnp, const blkptr_t *bp,
const zbookmark_phys_t *zb)
{
if (dnp == NULL) {
ASSERT(zb->zb_level < 0);
if (zb->zb_object == 0)
return (zb->zb_blkid);
return (zb->zb_blkid * BP_GET_LSIZE(bp));
}
ASSERT(zb->zb_level >= 0);
return ((zb->zb_blkid <<
(zb->zb_level * (dnp->dn_indblkshift - SPA_BLKPTRSHIFT))) *
dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT);
}
static void
snprintf_zstd_header(spa_t *spa, char *blkbuf, size_t buflen,
const blkptr_t *bp)
{
abd_t *pabd;
void *buf;
zio_t *zio;
zfs_zstdhdr_t zstd_hdr;
int error;
if (BP_GET_COMPRESS(bp) != ZIO_COMPRESS_ZSTD)
return;
if (BP_IS_HOLE(bp))
return;
if (BP_IS_EMBEDDED(bp)) {
buf = malloc(SPA_MAXBLOCKSIZE);
if (buf == NULL) {
(void) fprintf(stderr, "out of memory\n");
exit(1);
}
decode_embedded_bp_compressed(bp, buf);
memcpy(&zstd_hdr, buf, sizeof (zstd_hdr));
free(buf);
zstd_hdr.c_len = BE_32(zstd_hdr.c_len);
zstd_hdr.raw_version_level = BE_32(zstd_hdr.raw_version_level);
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
" ZSTD:size=%u:version=%u:level=%u:EMBEDDED",
zstd_hdr.c_len, zfs_get_hdrversion(&zstd_hdr),
zfs_get_hdrlevel(&zstd_hdr));
return;
}
pabd = abd_alloc_for_io(SPA_MAXBLOCKSIZE, B_FALSE);
zio = zio_root(spa, NULL, NULL, 0);
/* Decrypt but don't decompress so we can read the compression header */
zio_nowait(zio_read(zio, spa, bp, pabd, BP_GET_PSIZE(bp), NULL, NULL,
ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW_COMPRESS,
NULL));
error = zio_wait(zio);
if (error) {
(void) fprintf(stderr, "read failed: %d\n", error);
return;
}
buf = abd_borrow_buf_copy(pabd, BP_GET_LSIZE(bp));
memcpy(&zstd_hdr, buf, sizeof (zstd_hdr));
zstd_hdr.c_len = BE_32(zstd_hdr.c_len);
zstd_hdr.raw_version_level = BE_32(zstd_hdr.raw_version_level);
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
" ZSTD:size=%u:version=%u:level=%u:NORMAL",
zstd_hdr.c_len, zfs_get_hdrversion(&zstd_hdr),
zfs_get_hdrlevel(&zstd_hdr));
abd_return_buf_copy(pabd, buf, BP_GET_LSIZE(bp));
}
static void
snprintf_blkptr_compact(char *blkbuf, size_t buflen, const blkptr_t *bp,
boolean_t bp_freed)
{
const dva_t *dva = bp->blk_dva;
int ndvas = dump_opt['d'] > 5 ? BP_GET_NDVAS(bp) : 1;
int i;
if (dump_opt['b'] >= 6) {
snprintf_blkptr(blkbuf, buflen, bp);
if (bp_freed) {
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf), " %s", "FREE");
}
return;
}
if (BP_IS_EMBEDDED(bp)) {
(void) sprintf(blkbuf,
"EMBEDDED et=%u %llxL/%llxP B=%llu",
(int)BPE_GET_ETYPE(bp),
(u_longlong_t)BPE_GET_LSIZE(bp),
(u_longlong_t)BPE_GET_PSIZE(bp),
(u_longlong_t)bp->blk_birth);
return;
}
blkbuf[0] = '\0';
for (i = 0; i < ndvas; i++)
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf), "%llu:%llx:%llx ",
(u_longlong_t)DVA_GET_VDEV(&dva[i]),
(u_longlong_t)DVA_GET_OFFSET(&dva[i]),
(u_longlong_t)DVA_GET_ASIZE(&dva[i]));
if (BP_IS_HOLE(bp)) {
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
"%llxL B=%llu",
(u_longlong_t)BP_GET_LSIZE(bp),
(u_longlong_t)bp->blk_birth);
} else {
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
"%llxL/%llxP F=%llu B=%llu/%llu",
(u_longlong_t)BP_GET_LSIZE(bp),
(u_longlong_t)BP_GET_PSIZE(bp),
(u_longlong_t)BP_GET_FILL(bp),
(u_longlong_t)bp->blk_birth,
(u_longlong_t)BP_PHYSICAL_BIRTH(bp));
if (bp_freed)
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf), " %s", "FREE");
(void) snprintf(blkbuf + strlen(blkbuf),
buflen - strlen(blkbuf),
" cksum=%016llx:%016llx:%016llx:%016llx",
(u_longlong_t)bp->blk_cksum.zc_word[0],
(u_longlong_t)bp->blk_cksum.zc_word[1],
(u_longlong_t)bp->blk_cksum.zc_word[2],
(u_longlong_t)bp->blk_cksum.zc_word[3]);
}
}
static void
print_indirect(spa_t *spa, blkptr_t *bp, const zbookmark_phys_t *zb,
const dnode_phys_t *dnp)
{
char blkbuf[BP_SPRINTF_LEN];
int l;
if (!BP_IS_EMBEDDED(bp)) {
ASSERT3U(BP_GET_TYPE(bp), ==, dnp->dn_type);
ASSERT3U(BP_GET_LEVEL(bp), ==, zb->zb_level);
}
(void) printf("%16llx ", (u_longlong_t)blkid2offset(dnp, bp, zb));
ASSERT(zb->zb_level >= 0);
for (l = dnp->dn_nlevels - 1; l >= -1; l--) {
if (l == zb->zb_level) {
(void) printf("L%llx", (u_longlong_t)zb->zb_level);
} else {
(void) printf(" ");
}
}
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf), bp, B_FALSE);
if (dump_opt['Z'] && BP_GET_COMPRESS(bp) == ZIO_COMPRESS_ZSTD)
snprintf_zstd_header(spa, blkbuf, sizeof (blkbuf), bp);
(void) printf("%s\n", blkbuf);
}
static int
visit_indirect(spa_t *spa, const dnode_phys_t *dnp,
blkptr_t *bp, const zbookmark_phys_t *zb)
{
int err = 0;
if (bp->blk_birth == 0)
return (0);
print_indirect(spa, bp, zb, dnp);
if (BP_GET_LEVEL(bp) > 0 && !BP_IS_HOLE(bp)) {
arc_flags_t flags = ARC_FLAG_WAIT;
int i;
blkptr_t *cbp;
int epb = BP_GET_LSIZE(bp) >> SPA_BLKPTRSHIFT;
arc_buf_t *buf;
uint64_t fill = 0;
ASSERT(!BP_IS_REDACTED(bp));
err = arc_read(NULL, spa, bp, arc_getbuf_func, &buf,
ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_CANFAIL, &flags, zb);
if (err)
return (err);
ASSERT(buf->b_data);
/* recursively visit blocks below this */
cbp = buf->b_data;
for (i = 0; i < epb; i++, cbp++) {
zbookmark_phys_t czb;
SET_BOOKMARK(&czb, zb->zb_objset, zb->zb_object,
zb->zb_level - 1,
zb->zb_blkid * epb + i);
err = visit_indirect(spa, dnp, cbp, &czb);
if (err)
break;
fill += BP_GET_FILL(cbp);
}
if (!err)
ASSERT3U(fill, ==, BP_GET_FILL(bp));
arc_buf_destroy(buf, &buf);
}
return (err);
}
static void
dump_indirect(dnode_t *dn)
{
dnode_phys_t *dnp = dn->dn_phys;
zbookmark_phys_t czb;
(void) printf("Indirect blocks:\n");
SET_BOOKMARK(&czb, dmu_objset_id(dn->dn_objset),
dn->dn_object, dnp->dn_nlevels - 1, 0);
for (int j = 0; j < dnp->dn_nblkptr; j++) {
czb.zb_blkid = j;
(void) visit_indirect(dmu_objset_spa(dn->dn_objset), dnp,
&dnp->dn_blkptr[j], &czb);
}
(void) printf("\n");
}
static void
dump_dsl_dir(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object;
dsl_dir_phys_t *dd = data;
time_t crtime;
char nice[32];
/* make sure nicenum has enough space */
_Static_assert(sizeof (nice) >= NN_NUMBUF_SZ, "nice truncated");
if (dd == NULL)
return;
ASSERT3U(size, >=, sizeof (dsl_dir_phys_t));
crtime = dd->dd_creation_time;
(void) printf("\t\tcreation_time = %s", ctime(&crtime));
(void) printf("\t\thead_dataset_obj = %llu\n",
(u_longlong_t)dd->dd_head_dataset_obj);
(void) printf("\t\tparent_dir_obj = %llu\n",
(u_longlong_t)dd->dd_parent_obj);
(void) printf("\t\torigin_obj = %llu\n",
(u_longlong_t)dd->dd_origin_obj);
(void) printf("\t\tchild_dir_zapobj = %llu\n",
(u_longlong_t)dd->dd_child_dir_zapobj);
zdb_nicenum(dd->dd_used_bytes, nice, sizeof (nice));
(void) printf("\t\tused_bytes = %s\n", nice);
zdb_nicenum(dd->dd_compressed_bytes, nice, sizeof (nice));
(void) printf("\t\tcompressed_bytes = %s\n", nice);
zdb_nicenum(dd->dd_uncompressed_bytes, nice, sizeof (nice));
(void) printf("\t\tuncompressed_bytes = %s\n", nice);
zdb_nicenum(dd->dd_quota, nice, sizeof (nice));
(void) printf("\t\tquota = %s\n", nice);
zdb_nicenum(dd->dd_reserved, nice, sizeof (nice));
(void) printf("\t\treserved = %s\n", nice);
(void) printf("\t\tprops_zapobj = %llu\n",
(u_longlong_t)dd->dd_props_zapobj);
(void) printf("\t\tdeleg_zapobj = %llu\n",
(u_longlong_t)dd->dd_deleg_zapobj);
(void) printf("\t\tflags = %llx\n",
(u_longlong_t)dd->dd_flags);
#define DO(which) \
zdb_nicenum(dd->dd_used_breakdown[DD_USED_ ## which], nice, \
sizeof (nice)); \
(void) printf("\t\tused_breakdown[" #which "] = %s\n", nice)
DO(HEAD);
DO(SNAP);
DO(CHILD);
DO(CHILD_RSRV);
DO(REFRSRV);
#undef DO
(void) printf("\t\tclones = %llu\n",
(u_longlong_t)dd->dd_clones);
}
static void
dump_dsl_dataset(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object;
dsl_dataset_phys_t *ds = data;
time_t crtime;
char used[32], compressed[32], uncompressed[32], unique[32];
char blkbuf[BP_SPRINTF_LEN];
/* make sure nicenum has enough space */
_Static_assert(sizeof (used) >= NN_NUMBUF_SZ, "used truncated");
_Static_assert(sizeof (compressed) >= NN_NUMBUF_SZ,
"compressed truncated");
_Static_assert(sizeof (uncompressed) >= NN_NUMBUF_SZ,
"uncompressed truncated");
_Static_assert(sizeof (unique) >= NN_NUMBUF_SZ, "unique truncated");
if (ds == NULL)
return;
ASSERT(size == sizeof (*ds));
crtime = ds->ds_creation_time;
zdb_nicenum(ds->ds_referenced_bytes, used, sizeof (used));
zdb_nicenum(ds->ds_compressed_bytes, compressed, sizeof (compressed));
zdb_nicenum(ds->ds_uncompressed_bytes, uncompressed,
sizeof (uncompressed));
zdb_nicenum(ds->ds_unique_bytes, unique, sizeof (unique));
snprintf_blkptr(blkbuf, sizeof (blkbuf), &ds->ds_bp);
(void) printf("\t\tdir_obj = %llu\n",
(u_longlong_t)ds->ds_dir_obj);
(void) printf("\t\tprev_snap_obj = %llu\n",
(u_longlong_t)ds->ds_prev_snap_obj);
(void) printf("\t\tprev_snap_txg = %llu\n",
(u_longlong_t)ds->ds_prev_snap_txg);
(void) printf("\t\tnext_snap_obj = %llu\n",
(u_longlong_t)ds->ds_next_snap_obj);
(void) printf("\t\tsnapnames_zapobj = %llu\n",
(u_longlong_t)ds->ds_snapnames_zapobj);
(void) printf("\t\tnum_children = %llu\n",
(u_longlong_t)ds->ds_num_children);
(void) printf("\t\tuserrefs_obj = %llu\n",
(u_longlong_t)ds->ds_userrefs_obj);
(void) printf("\t\tcreation_time = %s", ctime(&crtime));
(void) printf("\t\tcreation_txg = %llu\n",
(u_longlong_t)ds->ds_creation_txg);
(void) printf("\t\tdeadlist_obj = %llu\n",
(u_longlong_t)ds->ds_deadlist_obj);
(void) printf("\t\tused_bytes = %s\n", used);
(void) printf("\t\tcompressed_bytes = %s\n", compressed);
(void) printf("\t\tuncompressed_bytes = %s\n", uncompressed);
(void) printf("\t\tunique = %s\n", unique);
(void) printf("\t\tfsid_guid = %llu\n",
(u_longlong_t)ds->ds_fsid_guid);
(void) printf("\t\tguid = %llu\n",
(u_longlong_t)ds->ds_guid);
(void) printf("\t\tflags = %llx\n",
(u_longlong_t)ds->ds_flags);
(void) printf("\t\tnext_clones_obj = %llu\n",
(u_longlong_t)ds->ds_next_clones_obj);
(void) printf("\t\tprops_obj = %llu\n",
(u_longlong_t)ds->ds_props_obj);
(void) printf("\t\tbp = %s\n", blkbuf);
}
static int
dump_bptree_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
(void) arg, (void) tx;
char blkbuf[BP_SPRINTF_LEN];
if (bp->blk_birth != 0) {
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
(void) printf("\t%s\n", blkbuf);
}
return (0);
}
static void
dump_bptree(objset_t *os, uint64_t obj, const char *name)
{
char bytes[32];
bptree_phys_t *bt;
dmu_buf_t *db;
/* make sure nicenum has enough space */
_Static_assert(sizeof (bytes) >= NN_NUMBUF_SZ, "bytes truncated");
if (dump_opt['d'] < 3)
return;
VERIFY3U(0, ==, dmu_bonus_hold(os, obj, FTAG, &db));
bt = db->db_data;
zdb_nicenum(bt->bt_bytes, bytes, sizeof (bytes));
(void) printf("\n %s: %llu datasets, %s\n",
name, (unsigned long long)(bt->bt_end - bt->bt_begin), bytes);
dmu_buf_rele(db, FTAG);
if (dump_opt['d'] < 5)
return;
(void) printf("\n");
(void) bptree_iterate(os, obj, B_FALSE, dump_bptree_cb, NULL, NULL);
}
static int
dump_bpobj_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx)
{
(void) arg, (void) tx;
char blkbuf[BP_SPRINTF_LEN];
ASSERT(bp->blk_birth != 0);
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf), bp, bp_freed);
(void) printf("\t%s\n", blkbuf);
return (0);
}
static void
dump_full_bpobj(bpobj_t *bpo, const char *name, int indent)
{
char bytes[32];
char comp[32];
char uncomp[32];
uint64_t i;
/* make sure nicenum has enough space */
_Static_assert(sizeof (bytes) >= NN_NUMBUF_SZ, "bytes truncated");
_Static_assert(sizeof (comp) >= NN_NUMBUF_SZ, "comp truncated");
_Static_assert(sizeof (uncomp) >= NN_NUMBUF_SZ, "uncomp truncated");
if (dump_opt['d'] < 3)
return;
zdb_nicenum(bpo->bpo_phys->bpo_bytes, bytes, sizeof (bytes));
if (bpo->bpo_havesubobj && bpo->bpo_phys->bpo_subobjs != 0) {
zdb_nicenum(bpo->bpo_phys->bpo_comp, comp, sizeof (comp));
zdb_nicenum(bpo->bpo_phys->bpo_uncomp, uncomp, sizeof (uncomp));
if (bpo->bpo_havefreed) {
(void) printf(" %*s: object %llu, %llu local "
"blkptrs, %llu freed, %llu subobjs in object %llu, "
"%s (%s/%s comp)\n",
indent * 8, name,
(u_longlong_t)bpo->bpo_object,
(u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs,
(u_longlong_t)bpo->bpo_phys->bpo_num_freed,
(u_longlong_t)bpo->bpo_phys->bpo_num_subobjs,
(u_longlong_t)bpo->bpo_phys->bpo_subobjs,
bytes, comp, uncomp);
} else {
(void) printf(" %*s: object %llu, %llu local "
"blkptrs, %llu subobjs in object %llu, "
"%s (%s/%s comp)\n",
indent * 8, name,
(u_longlong_t)bpo->bpo_object,
(u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs,
(u_longlong_t)bpo->bpo_phys->bpo_num_subobjs,
(u_longlong_t)bpo->bpo_phys->bpo_subobjs,
bytes, comp, uncomp);
}
for (i = 0; i < bpo->bpo_phys->bpo_num_subobjs; i++) {
uint64_t subobj;
bpobj_t subbpo;
int error;
VERIFY0(dmu_read(bpo->bpo_os,
bpo->bpo_phys->bpo_subobjs,
i * sizeof (subobj), sizeof (subobj), &subobj, 0));
error = bpobj_open(&subbpo, bpo->bpo_os, subobj);
if (error != 0) {
(void) printf("ERROR %u while trying to open "
"subobj id %llu\n",
error, (u_longlong_t)subobj);
continue;
}
dump_full_bpobj(&subbpo, "subobj", indent + 1);
bpobj_close(&subbpo);
}
} else {
if (bpo->bpo_havefreed) {
(void) printf(" %*s: object %llu, %llu blkptrs, "
"%llu freed, %s\n",
indent * 8, name,
(u_longlong_t)bpo->bpo_object,
(u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs,
(u_longlong_t)bpo->bpo_phys->bpo_num_freed,
bytes);
} else {
(void) printf(" %*s: object %llu, %llu blkptrs, "
"%s\n",
indent * 8, name,
(u_longlong_t)bpo->bpo_object,
(u_longlong_t)bpo->bpo_phys->bpo_num_blkptrs,
bytes);
}
}
if (dump_opt['d'] < 5)
return;
if (indent == 0) {
(void) bpobj_iterate_nofree(bpo, dump_bpobj_cb, NULL, NULL);
(void) printf("\n");
}
}
static int
dump_bookmark(dsl_pool_t *dp, char *name, boolean_t print_redact,
boolean_t print_list)
{
int err = 0;
zfs_bookmark_phys_t prop;
objset_t *mos = dp->dp_spa->spa_meta_objset;
err = dsl_bookmark_lookup(dp, name, NULL, &prop);
if (err != 0) {
return (err);
}
(void) printf("\t#%s: ", strchr(name, '#') + 1);
(void) printf("{guid: %llx creation_txg: %llu creation_time: "
"%llu redaction_obj: %llu}\n", (u_longlong_t)prop.zbm_guid,
(u_longlong_t)prop.zbm_creation_txg,
(u_longlong_t)prop.zbm_creation_time,
(u_longlong_t)prop.zbm_redaction_obj);
IMPLY(print_list, print_redact);
if (!print_redact || prop.zbm_redaction_obj == 0)
return (0);
redaction_list_t *rl;
VERIFY0(dsl_redaction_list_hold_obj(dp,
prop.zbm_redaction_obj, FTAG, &rl));
redaction_list_phys_t *rlp = rl->rl_phys;
(void) printf("\tRedacted:\n\t\tProgress: ");
if (rlp->rlp_last_object != UINT64_MAX ||
rlp->rlp_last_blkid != UINT64_MAX) {
(void) printf("%llu %llu (incomplete)\n",
(u_longlong_t)rlp->rlp_last_object,
(u_longlong_t)rlp->rlp_last_blkid);
} else {
(void) printf("complete\n");
}
(void) printf("\t\tSnapshots: [");
for (unsigned int i = 0; i < rlp->rlp_num_snaps; i++) {
if (i > 0)
(void) printf(", ");
(void) printf("%0llu",
(u_longlong_t)rlp->rlp_snaps[i]);
}
(void) printf("]\n\t\tLength: %llu\n",
(u_longlong_t)rlp->rlp_num_entries);
if (!print_list) {
dsl_redaction_list_rele(rl, FTAG);
return (0);
}
if (rlp->rlp_num_entries == 0) {
dsl_redaction_list_rele(rl, FTAG);
(void) printf("\t\tRedaction List: []\n\n");
return (0);
}
redact_block_phys_t *rbp_buf;
uint64_t size;
dmu_object_info_t doi;
VERIFY0(dmu_object_info(mos, prop.zbm_redaction_obj, &doi));
size = doi.doi_max_offset;
rbp_buf = kmem_alloc(size, KM_SLEEP);
err = dmu_read(mos, prop.zbm_redaction_obj, 0, size,
rbp_buf, 0);
if (err != 0) {
dsl_redaction_list_rele(rl, FTAG);
kmem_free(rbp_buf, size);
return (err);
}
(void) printf("\t\tRedaction List: [{object: %llx, offset: "
"%llx, blksz: %x, count: %llx}",
(u_longlong_t)rbp_buf[0].rbp_object,
(u_longlong_t)rbp_buf[0].rbp_blkid,
(uint_t)(redact_block_get_size(&rbp_buf[0])),
(u_longlong_t)redact_block_get_count(&rbp_buf[0]));
for (size_t i = 1; i < rlp->rlp_num_entries; i++) {
(void) printf(",\n\t\t{object: %llx, offset: %llx, "
"blksz: %x, count: %llx}",
(u_longlong_t)rbp_buf[i].rbp_object,
(u_longlong_t)rbp_buf[i].rbp_blkid,
(uint_t)(redact_block_get_size(&rbp_buf[i])),
(u_longlong_t)redact_block_get_count(&rbp_buf[i]));
}
dsl_redaction_list_rele(rl, FTAG);
kmem_free(rbp_buf, size);
(void) printf("]\n\n");
return (0);
}
static void
dump_bookmarks(objset_t *os, int verbosity)
{
zap_cursor_t zc;
zap_attribute_t attr;
dsl_dataset_t *ds = dmu_objset_ds(os);
dsl_pool_t *dp = spa_get_dsl(os->os_spa);
objset_t *mos = os->os_spa->spa_meta_objset;
if (verbosity < 4)
return;
dsl_pool_config_enter(dp, FTAG);
for (zap_cursor_init(&zc, mos, ds->ds_bookmarks_obj);
zap_cursor_retrieve(&zc, &attr) == 0;
zap_cursor_advance(&zc)) {
char osname[ZFS_MAX_DATASET_NAME_LEN];
char buf[ZFS_MAX_DATASET_NAME_LEN];
int len;
dmu_objset_name(os, osname);
len = snprintf(buf, sizeof (buf), "%s#%s", osname,
attr.za_name);
VERIFY3S(len, <, ZFS_MAX_DATASET_NAME_LEN);
(void) dump_bookmark(dp, buf, verbosity >= 5, verbosity >= 6);
}
zap_cursor_fini(&zc);
dsl_pool_config_exit(dp, FTAG);
}
static void
bpobj_count_refd(bpobj_t *bpo)
{
mos_obj_refd(bpo->bpo_object);
if (bpo->bpo_havesubobj && bpo->bpo_phys->bpo_subobjs != 0) {
mos_obj_refd(bpo->bpo_phys->bpo_subobjs);
for (uint64_t i = 0; i < bpo->bpo_phys->bpo_num_subobjs; i++) {
uint64_t subobj;
bpobj_t subbpo;
int error;
VERIFY0(dmu_read(bpo->bpo_os,
bpo->bpo_phys->bpo_subobjs,
i * sizeof (subobj), sizeof (subobj), &subobj, 0));
error = bpobj_open(&subbpo, bpo->bpo_os, subobj);
if (error != 0) {
(void) printf("ERROR %u while trying to open "
"subobj id %llu\n",
error, (u_longlong_t)subobj);
continue;
}
bpobj_count_refd(&subbpo);
bpobj_close(&subbpo);
}
}
}
static int
dsl_deadlist_entry_count_refd(void *arg, dsl_deadlist_entry_t *dle)
{
spa_t *spa = arg;
uint64_t empty_bpobj = spa->spa_dsl_pool->dp_empty_bpobj;
if (dle->dle_bpobj.bpo_object != empty_bpobj)
bpobj_count_refd(&dle->dle_bpobj);
return (0);
}
static int
dsl_deadlist_entry_dump(void *arg, dsl_deadlist_entry_t *dle)
{
ASSERT(arg == NULL);
if (dump_opt['d'] >= 5) {
char buf[128];
(void) snprintf(buf, sizeof (buf),
"mintxg %llu -> obj %llu",
(longlong_t)dle->dle_mintxg,
(longlong_t)dle->dle_bpobj.bpo_object);
dump_full_bpobj(&dle->dle_bpobj, buf, 0);
} else {
(void) printf("mintxg %llu -> obj %llu\n",
(longlong_t)dle->dle_mintxg,
(longlong_t)dle->dle_bpobj.bpo_object);
}
return (0);
}
static void
dump_blkptr_list(dsl_deadlist_t *dl, const char *name)
{
char bytes[32];
char comp[32];
char uncomp[32];
char entries[32];
spa_t *spa = dmu_objset_spa(dl->dl_os);
uint64_t empty_bpobj = spa->spa_dsl_pool->dp_empty_bpobj;
if (dl->dl_oldfmt) {
if (dl->dl_bpobj.bpo_object != empty_bpobj)
bpobj_count_refd(&dl->dl_bpobj);
} else {
mos_obj_refd(dl->dl_object);
dsl_deadlist_iterate(dl, dsl_deadlist_entry_count_refd, spa);
}
/* make sure nicenum has enough space */
_Static_assert(sizeof (bytes) >= NN_NUMBUF_SZ, "bytes truncated");
_Static_assert(sizeof (comp) >= NN_NUMBUF_SZ, "comp truncated");
_Static_assert(sizeof (uncomp) >= NN_NUMBUF_SZ, "uncomp truncated");
_Static_assert(sizeof (entries) >= NN_NUMBUF_SZ, "entries truncated");
if (dump_opt['d'] < 3)
return;
if (dl->dl_oldfmt) {
dump_full_bpobj(&dl->dl_bpobj, "old-format deadlist", 0);
return;
}
zdb_nicenum(dl->dl_phys->dl_used, bytes, sizeof (bytes));
zdb_nicenum(dl->dl_phys->dl_comp, comp, sizeof (comp));
zdb_nicenum(dl->dl_phys->dl_uncomp, uncomp, sizeof (uncomp));
zdb_nicenum(avl_numnodes(&dl->dl_tree), entries, sizeof (entries));
(void) printf("\n %s: %s (%s/%s comp), %s entries\n",
name, bytes, comp, uncomp, entries);
if (dump_opt['d'] < 4)
return;
(void) putchar('\n');
dsl_deadlist_iterate(dl, dsl_deadlist_entry_dump, NULL);
}
static int
verify_dd_livelist(objset_t *os)
{
uint64_t ll_used, used, ll_comp, comp, ll_uncomp, uncomp;
dsl_pool_t *dp = spa_get_dsl(os->os_spa);
dsl_dir_t *dd = os->os_dsl_dataset->ds_dir;
ASSERT(!dmu_objset_is_snapshot(os));
if (!dsl_deadlist_is_open(&dd->dd_livelist))
return (0);
/* Iterate through the livelist to check for duplicates */
dsl_deadlist_iterate(&dd->dd_livelist, sublivelist_verify_lightweight,
NULL);
dsl_pool_config_enter(dp, FTAG);
dsl_deadlist_space(&dd->dd_livelist, &ll_used,
&ll_comp, &ll_uncomp);
dsl_dataset_t *origin_ds;
ASSERT(dsl_pool_config_held(dp));
VERIFY0(dsl_dataset_hold_obj(dp,
dsl_dir_phys(dd)->dd_origin_obj, FTAG, &origin_ds));
VERIFY0(dsl_dataset_space_written(origin_ds, os->os_dsl_dataset,
&used, &comp, &uncomp));
dsl_dataset_rele(origin_ds, FTAG);
dsl_pool_config_exit(dp, FTAG);
/*
* It's possible that the dataset's uncomp space is larger than the
* livelist's because livelists do not track embedded block pointers
*/
if (used != ll_used || comp != ll_comp || uncomp < ll_uncomp) {
char nice_used[32], nice_comp[32], nice_uncomp[32];
(void) printf("Discrepancy in space accounting:\n");
zdb_nicenum(used, nice_used, sizeof (nice_used));
zdb_nicenum(comp, nice_comp, sizeof (nice_comp));
zdb_nicenum(uncomp, nice_uncomp, sizeof (nice_uncomp));
(void) printf("dir: used %s, comp %s, uncomp %s\n",
nice_used, nice_comp, nice_uncomp);
zdb_nicenum(ll_used, nice_used, sizeof (nice_used));
zdb_nicenum(ll_comp, nice_comp, sizeof (nice_comp));
zdb_nicenum(ll_uncomp, nice_uncomp, sizeof (nice_uncomp));
(void) printf("livelist: used %s, comp %s, uncomp %s\n",
nice_used, nice_comp, nice_uncomp);
return (1);
}
return (0);
}
static char *key_material = NULL;
static boolean_t
zdb_derive_key(dsl_dir_t *dd, uint8_t *key_out)
{
uint64_t keyformat, salt, iters;
int i;
unsigned char c;
VERIFY0(zap_lookup(dd->dd_pool->dp_meta_objset, dd->dd_crypto_obj,
zfs_prop_to_name(ZFS_PROP_KEYFORMAT), sizeof (uint64_t),
1, &keyformat));
switch (keyformat) {
case ZFS_KEYFORMAT_HEX:
for (i = 0; i < WRAPPING_KEY_LEN * 2; i += 2) {
if (!isxdigit(key_material[i]) ||
!isxdigit(key_material[i+1]))
return (B_FALSE);
if (sscanf(&key_material[i], "%02hhx", &c) != 1)
return (B_FALSE);
key_out[i / 2] = c;
}
break;
case ZFS_KEYFORMAT_PASSPHRASE:
VERIFY0(zap_lookup(dd->dd_pool->dp_meta_objset,
dd->dd_crypto_obj, zfs_prop_to_name(ZFS_PROP_PBKDF2_SALT),
sizeof (uint64_t), 1, &salt));
VERIFY0(zap_lookup(dd->dd_pool->dp_meta_objset,
dd->dd_crypto_obj, zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS),
sizeof (uint64_t), 1, &iters));
if (PKCS5_PBKDF2_HMAC_SHA1(key_material, strlen(key_material),
((uint8_t *)&salt), sizeof (uint64_t), iters,
WRAPPING_KEY_LEN, key_out) != 1)
return (B_FALSE);
break;
default:
fatal("no support for key format %u\n",
(unsigned int) keyformat);
}
return (B_TRUE);
}
static char encroot[ZFS_MAX_DATASET_NAME_LEN];
static boolean_t key_loaded = B_FALSE;
static void
zdb_load_key(objset_t *os)
{
dsl_pool_t *dp;
dsl_dir_t *dd, *rdd;
uint8_t key[WRAPPING_KEY_LEN];
uint64_t rddobj;
int err;
dp = spa_get_dsl(os->os_spa);
dd = os->os_dsl_dataset->ds_dir;
dsl_pool_config_enter(dp, FTAG);
VERIFY0(zap_lookup(dd->dd_pool->dp_meta_objset, dd->dd_crypto_obj,
DSL_CRYPTO_KEY_ROOT_DDOBJ, sizeof (uint64_t), 1, &rddobj));
VERIFY0(dsl_dir_hold_obj(dd->dd_pool, rddobj, NULL, FTAG, &rdd));
dsl_dir_name(rdd, encroot);
dsl_dir_rele(rdd, FTAG);
if (!zdb_derive_key(dd, key))
fatal("couldn't derive encryption key");
dsl_pool_config_exit(dp, FTAG);
ASSERT3U(dsl_dataset_get_keystatus(dd), ==, ZFS_KEYSTATUS_UNAVAILABLE);
dsl_crypto_params_t *dcp;
nvlist_t *crypto_args;
crypto_args = fnvlist_alloc();
fnvlist_add_uint8_array(crypto_args, "wkeydata",
(uint8_t *)key, WRAPPING_KEY_LEN);
VERIFY0(dsl_crypto_params_create_nvlist(DCP_CMD_NONE,
NULL, crypto_args, &dcp));
err = spa_keystore_load_wkey(encroot, dcp, B_FALSE);
dsl_crypto_params_free(dcp, (err != 0));
fnvlist_free(crypto_args);
if (err != 0)
fatal(
"couldn't load encryption key for %s: %s",
encroot, err == ZFS_ERR_CRYPTO_NOTSUP ?
"crypto params not supported" : strerror(err));
ASSERT3U(dsl_dataset_get_keystatus(dd), ==, ZFS_KEYSTATUS_AVAILABLE);
printf("Unlocked encryption root: %s\n", encroot);
key_loaded = B_TRUE;
}
static void
zdb_unload_key(void)
{
if (!key_loaded)
return;
VERIFY0(spa_keystore_unload_wkey(encroot));
key_loaded = B_FALSE;
}
static avl_tree_t idx_tree;
static avl_tree_t domain_tree;
static boolean_t fuid_table_loaded;
static objset_t *sa_os = NULL;
static sa_attr_type_t *sa_attr_table = NULL;
static int
open_objset(const char *path, const void *tag, objset_t **osp)
{
int err;
uint64_t sa_attrs = 0;
uint64_t version = 0;
VERIFY3P(sa_os, ==, NULL);
/*
* We can't own an objset if it's redacted. Therefore, we do this
* dance: hold the objset, then acquire a long hold on its dataset, then
* release the pool (which is held as part of holding the objset).
*/
if (dump_opt['K']) {
/* decryption requested, try to load keys */
err = dmu_objset_hold(path, tag, osp);
if (err != 0) {
(void) fprintf(stderr, "failed to hold dataset "
"'%s': %s\n",
path, strerror(err));
return (err);
}
dsl_dataset_long_hold(dmu_objset_ds(*osp), tag);
dsl_pool_rele(dmu_objset_pool(*osp), tag);
/* succeeds or dies */
zdb_load_key(*osp);
/* release it all */
dsl_dataset_long_rele(dmu_objset_ds(*osp), tag);
dsl_dataset_rele(dmu_objset_ds(*osp), tag);
}
int ds_hold_flags = key_loaded ? DS_HOLD_FLAG_DECRYPT : 0;
err = dmu_objset_hold_flags(path, ds_hold_flags, tag, osp);
if (err != 0) {
(void) fprintf(stderr, "failed to hold dataset '%s': %s\n",
path, strerror(err));
return (err);
}
dsl_dataset_long_hold(dmu_objset_ds(*osp), tag);
dsl_pool_rele(dmu_objset_pool(*osp), tag);
if (dmu_objset_type(*osp) == DMU_OST_ZFS &&
(key_loaded || !(*osp)->os_encrypted)) {
(void) zap_lookup(*osp, MASTER_NODE_OBJ, ZPL_VERSION_STR,
8, 1, &version);
if (version >= ZPL_VERSION_SA) {
(void) zap_lookup(*osp, MASTER_NODE_OBJ, ZFS_SA_ATTRS,
8, 1, &sa_attrs);
}
err = sa_setup(*osp, sa_attrs, zfs_attr_table, ZPL_END,
&sa_attr_table);
if (err != 0) {
(void) fprintf(stderr, "sa_setup failed: %s\n",
strerror(err));
dsl_dataset_long_rele(dmu_objset_ds(*osp), tag);
dsl_dataset_rele_flags(dmu_objset_ds(*osp),
ds_hold_flags, tag);
*osp = NULL;
}
}
sa_os = *osp;
return (err);
}
static void
close_objset(objset_t *os, const void *tag)
{
VERIFY3P(os, ==, sa_os);
if (os->os_sa != NULL)
sa_tear_down(os);
dsl_dataset_long_rele(dmu_objset_ds(os), tag);
dsl_dataset_rele_flags(dmu_objset_ds(os),
key_loaded ? DS_HOLD_FLAG_DECRYPT : 0, tag);
sa_attr_table = NULL;
sa_os = NULL;
zdb_unload_key();
}
static void
fuid_table_destroy(void)
{
if (fuid_table_loaded) {
zfs_fuid_table_destroy(&idx_tree, &domain_tree);
fuid_table_loaded = B_FALSE;
}
}
/*
* print uid or gid information.
* For normal POSIX id just the id is printed in decimal format.
* For CIFS files with FUID the fuid is printed in hex followed by
* the domain-rid string.
*/
static void
print_idstr(uint64_t id, const char *id_type)
{
if (FUID_INDEX(id)) {
const char *domain =
zfs_fuid_idx_domain(&idx_tree, FUID_INDEX(id));
(void) printf("\t%s %llx [%s-%d]\n", id_type,
(u_longlong_t)id, domain, (int)FUID_RID(id));
} else {
(void) printf("\t%s %llu\n", id_type, (u_longlong_t)id);
}
}
static void
dump_uidgid(objset_t *os, uint64_t uid, uint64_t gid)
{
uint32_t uid_idx, gid_idx;
uid_idx = FUID_INDEX(uid);
gid_idx = FUID_INDEX(gid);
/* Load domain table, if not already loaded */
if (!fuid_table_loaded && (uid_idx || gid_idx)) {
uint64_t fuid_obj;
/* first find the fuid object. It lives in the master node */
VERIFY(zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES,
8, 1, &fuid_obj) == 0);
zfs_fuid_avl_tree_create(&idx_tree, &domain_tree);
(void) zfs_fuid_table_load(os, fuid_obj,
&idx_tree, &domain_tree);
fuid_table_loaded = B_TRUE;
}
print_idstr(uid, "uid");
print_idstr(gid, "gid");
}
static void
dump_znode_sa_xattr(sa_handle_t *hdl)
{
nvlist_t *sa_xattr;
nvpair_t *elem = NULL;
int sa_xattr_size = 0;
int sa_xattr_entries = 0;
int error;
char *sa_xattr_packed;
error = sa_size(hdl, sa_attr_table[ZPL_DXATTR], &sa_xattr_size);
if (error || sa_xattr_size == 0)
return;
sa_xattr_packed = malloc(sa_xattr_size);
if (sa_xattr_packed == NULL)
return;
error = sa_lookup(hdl, sa_attr_table[ZPL_DXATTR],
sa_xattr_packed, sa_xattr_size);
if (error) {
free(sa_xattr_packed);
return;
}
error = nvlist_unpack(sa_xattr_packed, sa_xattr_size, &sa_xattr, 0);
if (error) {
free(sa_xattr_packed);
return;
}
while ((elem = nvlist_next_nvpair(sa_xattr, elem)) != NULL)
sa_xattr_entries++;
(void) printf("\tSA xattrs: %d bytes, %d entries\n\n",
sa_xattr_size, sa_xattr_entries);
while ((elem = nvlist_next_nvpair(sa_xattr, elem)) != NULL) {
uchar_t *value;
uint_t cnt, idx;
(void) printf("\t\t%s = ", nvpair_name(elem));
nvpair_value_byte_array(elem, &value, &cnt);
for (idx = 0; idx < cnt; ++idx) {
if (isprint(value[idx]))
(void) putchar(value[idx]);
else
(void) printf("\\%3.3o", value[idx]);
}
(void) putchar('\n');
}
nvlist_free(sa_xattr);
free(sa_xattr_packed);
}
static void
dump_znode_symlink(sa_handle_t *hdl)
{
int sa_symlink_size = 0;
char linktarget[MAXPATHLEN];
int error;
error = sa_size(hdl, sa_attr_table[ZPL_SYMLINK], &sa_symlink_size);
if (error || sa_symlink_size == 0) {
return;
}
if (sa_symlink_size >= sizeof (linktarget)) {
(void) printf("symlink size %d is too large\n",
sa_symlink_size);
return;
}
linktarget[sa_symlink_size] = '\0';
if (sa_lookup(hdl, sa_attr_table[ZPL_SYMLINK],
&linktarget, sa_symlink_size) == 0)
(void) printf("\ttarget %s\n", linktarget);
}
static void
dump_znode(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) data, (void) size;
char path[MAXPATHLEN * 2]; /* allow for xattr and failure prefix */
sa_handle_t *hdl;
uint64_t xattr, rdev, gen;
uint64_t uid, gid, mode, fsize, parent, links;
uint64_t pflags;
uint64_t acctm[2], modtm[2], chgtm[2], crtm[2];
time_t z_crtime, z_atime, z_mtime, z_ctime;
sa_bulk_attr_t bulk[12];
int idx = 0;
int error;
VERIFY3P(os, ==, sa_os);
if (sa_handle_get(os, object, NULL, SA_HDL_PRIVATE, &hdl)) {
(void) printf("Failed to get handle for SA znode\n");
return;
}
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_UID], NULL, &uid, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_GID], NULL, &gid, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_LINKS], NULL,
&links, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_GEN], NULL, &gen, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_MODE], NULL,
&mode, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_PARENT],
NULL, &parent, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_SIZE], NULL,
&fsize, 8);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_ATIME], NULL,
acctm, 16);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_MTIME], NULL,
modtm, 16);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_CRTIME], NULL,
crtm, 16);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_CTIME], NULL,
chgtm, 16);
SA_ADD_BULK_ATTR(bulk, idx, sa_attr_table[ZPL_FLAGS], NULL,
&pflags, 8);
if (sa_bulk_lookup(hdl, bulk, idx)) {
(void) sa_handle_destroy(hdl);
return;
}
z_crtime = (time_t)crtm[0];
z_atime = (time_t)acctm[0];
z_mtime = (time_t)modtm[0];
z_ctime = (time_t)chgtm[0];
if (dump_opt['d'] > 4) {
error = zfs_obj_to_path(os, object, path, sizeof (path));
if (error == ESTALE) {
(void) snprintf(path, sizeof (path), "on delete queue");
} else if (error != 0) {
leaked_objects++;
(void) snprintf(path, sizeof (path),
"path not found, possibly leaked");
}
(void) printf("\tpath %s\n", path);
}
if (S_ISLNK(mode))
dump_znode_symlink(hdl);
dump_uidgid(os, uid, gid);
(void) printf("\tatime %s", ctime(&z_atime));
(void) printf("\tmtime %s", ctime(&z_mtime));
(void) printf("\tctime %s", ctime(&z_ctime));
(void) printf("\tcrtime %s", ctime(&z_crtime));
(void) printf("\tgen %llu\n", (u_longlong_t)gen);
(void) printf("\tmode %llo\n", (u_longlong_t)mode);
(void) printf("\tsize %llu\n", (u_longlong_t)fsize);
(void) printf("\tparent %llu\n", (u_longlong_t)parent);
(void) printf("\tlinks %llu\n", (u_longlong_t)links);
(void) printf("\tpflags %llx\n", (u_longlong_t)pflags);
if (dmu_objset_projectquota_enabled(os) && (pflags & ZFS_PROJID)) {
uint64_t projid;
if (sa_lookup(hdl, sa_attr_table[ZPL_PROJID], &projid,
sizeof (uint64_t)) == 0)
(void) printf("\tprojid %llu\n", (u_longlong_t)projid);
}
if (sa_lookup(hdl, sa_attr_table[ZPL_XATTR], &xattr,
sizeof (uint64_t)) == 0)
(void) printf("\txattr %llu\n", (u_longlong_t)xattr);
if (sa_lookup(hdl, sa_attr_table[ZPL_RDEV], &rdev,
sizeof (uint64_t)) == 0)
(void) printf("\trdev 0x%016llx\n", (u_longlong_t)rdev);
dump_znode_sa_xattr(hdl);
sa_handle_destroy(hdl);
}
static void
dump_acl(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static void
dump_dmu_objset(objset_t *os, uint64_t object, void *data, size_t size)
{
(void) os, (void) object, (void) data, (void) size;
}
static object_viewer_t *object_viewer[DMU_OT_NUMTYPES + 1] = {
dump_none, /* unallocated */
dump_zap, /* object directory */
dump_uint64, /* object array */
dump_none, /* packed nvlist */
dump_packed_nvlist, /* packed nvlist size */
dump_none, /* bpobj */
dump_bpobj, /* bpobj header */
dump_none, /* SPA space map header */
dump_none, /* SPA space map */
dump_none, /* ZIL intent log */
dump_dnode, /* DMU dnode */
dump_dmu_objset, /* DMU objset */
dump_dsl_dir, /* DSL directory */
dump_zap, /* DSL directory child map */
dump_zap, /* DSL dataset snap map */
dump_zap, /* DSL props */
dump_dsl_dataset, /* DSL dataset */
dump_znode, /* ZFS znode */
dump_acl, /* ZFS V0 ACL */
dump_uint8, /* ZFS plain file */
dump_zpldir, /* ZFS directory */
dump_zap, /* ZFS master node */
dump_zap, /* ZFS delete queue */
dump_uint8, /* zvol object */
dump_zap, /* zvol prop */
dump_uint8, /* other uint8[] */
dump_uint64, /* other uint64[] */
dump_zap, /* other ZAP */
dump_zap, /* persistent error log */
dump_uint8, /* SPA history */
dump_history_offsets, /* SPA history offsets */
dump_zap, /* Pool properties */
dump_zap, /* DSL permissions */
dump_acl, /* ZFS ACL */
dump_uint8, /* ZFS SYSACL */
dump_none, /* FUID nvlist */
dump_packed_nvlist, /* FUID nvlist size */
dump_zap, /* DSL dataset next clones */
dump_zap, /* DSL scrub queue */
dump_zap, /* ZFS user/group/project used */
dump_zap, /* ZFS user/group/project quota */
dump_zap, /* snapshot refcount tags */
dump_ddt_zap, /* DDT ZAP object */
dump_zap, /* DDT statistics */
dump_znode, /* SA object */
dump_zap, /* SA Master Node */
dump_sa_attrs, /* SA attribute registration */
dump_sa_layouts, /* SA attribute layouts */
dump_zap, /* DSL scrub translations */
dump_none, /* fake dedup BP */
dump_zap, /* deadlist */
dump_none, /* deadlist hdr */
dump_zap, /* dsl clones */
dump_bpobj_subobjs, /* bpobj subobjs */
dump_unknown, /* Unknown type, must be last */
};
static boolean_t
match_object_type(dmu_object_type_t obj_type, uint64_t flags)
{
boolean_t match = B_TRUE;
switch (obj_type) {
case DMU_OT_DIRECTORY_CONTENTS:
if (!(flags & ZOR_FLAG_DIRECTORY))
match = B_FALSE;
break;
case DMU_OT_PLAIN_FILE_CONTENTS:
if (!(flags & ZOR_FLAG_PLAIN_FILE))
match = B_FALSE;
break;
case DMU_OT_SPACE_MAP:
if (!(flags & ZOR_FLAG_SPACE_MAP))
match = B_FALSE;
break;
default:
if (strcmp(zdb_ot_name(obj_type), "zap") == 0) {
if (!(flags & ZOR_FLAG_ZAP))
match = B_FALSE;
break;
}
/*
* If all bits except some of the supported flags are
* set, the user combined the all-types flag (A) with
* a negated flag to exclude some types (e.g. A-f to
* show all object types except plain files).
*/
if ((flags | ZOR_SUPPORTED_FLAGS) != ZOR_FLAG_ALL_TYPES)
match = B_FALSE;
break;
}
return (match);
}
static void
dump_object(objset_t *os, uint64_t object, int verbosity,
boolean_t *print_header, uint64_t *dnode_slots_used, uint64_t flags)
{
dmu_buf_t *db = NULL;
dmu_object_info_t doi;
dnode_t *dn;
boolean_t dnode_held = B_FALSE;
void *bonus = NULL;
size_t bsize = 0;
char iblk[32], dblk[32], lsize[32], asize[32], fill[32], dnsize[32];
char bonus_size[32];
char aux[50];
int error;
/* make sure nicenum has enough space */
_Static_assert(sizeof (iblk) >= NN_NUMBUF_SZ, "iblk truncated");
_Static_assert(sizeof (dblk) >= NN_NUMBUF_SZ, "dblk truncated");
_Static_assert(sizeof (lsize) >= NN_NUMBUF_SZ, "lsize truncated");
_Static_assert(sizeof (asize) >= NN_NUMBUF_SZ, "asize truncated");
_Static_assert(sizeof (bonus_size) >= NN_NUMBUF_SZ,
"bonus_size truncated");
if (*print_header) {
(void) printf("\n%10s %3s %5s %5s %5s %6s %5s %6s %s\n",
"Object", "lvl", "iblk", "dblk", "dsize", "dnsize",
"lsize", "%full", "type");
*print_header = 0;
}
if (object == 0) {
dn = DMU_META_DNODE(os);
dmu_object_info_from_dnode(dn, &doi);
} else {
/*
* Encrypted datasets will have sensitive bonus buffers
* encrypted. Therefore we cannot hold the bonus buffer and
* must hold the dnode itself instead.
*/
error = dmu_object_info(os, object, &doi);
if (error)
fatal("dmu_object_info() failed, errno %u", error);
if (!key_loaded && os->os_encrypted &&
DMU_OT_IS_ENCRYPTED(doi.doi_bonus_type)) {
error = dnode_hold(os, object, FTAG, &dn);
if (error)
fatal("dnode_hold() failed, errno %u", error);
dnode_held = B_TRUE;
} else {
error = dmu_bonus_hold(os, object, FTAG, &db);
if (error)
fatal("dmu_bonus_hold(%llu) failed, errno %u",
object, error);
bonus = db->db_data;
bsize = db->db_size;
dn = DB_DNODE((dmu_buf_impl_t *)db);
}
}
/*
* Default to showing all object types if no flags were specified.
*/
if (flags != 0 && flags != ZOR_FLAG_ALL_TYPES &&
!match_object_type(doi.doi_type, flags))
goto out;
if (dnode_slots_used)
*dnode_slots_used = doi.doi_dnodesize / DNODE_MIN_SIZE;
zdb_nicenum(doi.doi_metadata_block_size, iblk, sizeof (iblk));
zdb_nicenum(doi.doi_data_block_size, dblk, sizeof (dblk));
zdb_nicenum(doi.doi_max_offset, lsize, sizeof (lsize));
zdb_nicenum(doi.doi_physical_blocks_512 << 9, asize, sizeof (asize));
zdb_nicenum(doi.doi_bonus_size, bonus_size, sizeof (bonus_size));
zdb_nicenum(doi.doi_dnodesize, dnsize, sizeof (dnsize));
(void) snprintf(fill, sizeof (fill), "%6.2f", 100.0 *
doi.doi_fill_count * doi.doi_data_block_size / (object == 0 ?
DNODES_PER_BLOCK : 1) / doi.doi_max_offset);
aux[0] = '\0';
if (doi.doi_checksum != ZIO_CHECKSUM_INHERIT || verbosity >= 6) {
(void) snprintf(aux + strlen(aux), sizeof (aux) - strlen(aux),
" (K=%s)", ZDB_CHECKSUM_NAME(doi.doi_checksum));
}
if (doi.doi_compress == ZIO_COMPRESS_INHERIT &&
ZIO_COMPRESS_HASLEVEL(os->os_compress) && verbosity >= 6) {
const char *compname = NULL;
if (zfs_prop_index_to_string(ZFS_PROP_COMPRESSION,
ZIO_COMPRESS_RAW(os->os_compress, os->os_complevel),
&compname) == 0) {
(void) snprintf(aux + strlen(aux),
sizeof (aux) - strlen(aux), " (Z=inherit=%s)",
compname);
} else {
(void) snprintf(aux + strlen(aux),
sizeof (aux) - strlen(aux),
" (Z=inherit=%s-unknown)",
ZDB_COMPRESS_NAME(os->os_compress));
}
} else if (doi.doi_compress == ZIO_COMPRESS_INHERIT && verbosity >= 6) {
(void) snprintf(aux + strlen(aux), sizeof (aux) - strlen(aux),
" (Z=inherit=%s)", ZDB_COMPRESS_NAME(os->os_compress));
} else if (doi.doi_compress != ZIO_COMPRESS_INHERIT || verbosity >= 6) {
(void) snprintf(aux + strlen(aux), sizeof (aux) - strlen(aux),
" (Z=%s)", ZDB_COMPRESS_NAME(doi.doi_compress));
}
(void) printf("%10lld %3u %5s %5s %5s %6s %5s %6s %s%s\n",
(u_longlong_t)object, doi.doi_indirection, iblk, dblk,
asize, dnsize, lsize, fill, zdb_ot_name(doi.doi_type), aux);
if (doi.doi_bonus_type != DMU_OT_NONE && verbosity > 3) {
(void) printf("%10s %3s %5s %5s %5s %5s %5s %6s %s\n",
"", "", "", "", "", "", bonus_size, "bonus",
zdb_ot_name(doi.doi_bonus_type));
}
if (verbosity >= 4) {
(void) printf("\tdnode flags: %s%s%s%s\n",
(dn->dn_phys->dn_flags & DNODE_FLAG_USED_BYTES) ?
"USED_BYTES " : "",
(dn->dn_phys->dn_flags & DNODE_FLAG_USERUSED_ACCOUNTED) ?
"USERUSED_ACCOUNTED " : "",
(dn->dn_phys->dn_flags & DNODE_FLAG_USEROBJUSED_ACCOUNTED) ?
"USEROBJUSED_ACCOUNTED " : "",
(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR) ?
"SPILL_BLKPTR" : "");
(void) printf("\tdnode maxblkid: %llu\n",
(longlong_t)dn->dn_phys->dn_maxblkid);
if (!dnode_held) {
object_viewer[ZDB_OT_TYPE(doi.doi_bonus_type)](os,
object, bonus, bsize);
} else {
(void) printf("\t\t(bonus encrypted)\n");
}
if (key_loaded ||
(!os->os_encrypted || !DMU_OT_IS_ENCRYPTED(doi.doi_type))) {
object_viewer[ZDB_OT_TYPE(doi.doi_type)](os, object,
NULL, 0);
} else {
(void) printf("\t\t(object encrypted)\n");
}
*print_header = B_TRUE;
}
if (verbosity >= 5) {
if (dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr_compact(blkbuf, sizeof (blkbuf),
DN_SPILL_BLKPTR(dn->dn_phys), B_FALSE);
(void) printf("\nSpill block: %s\n", blkbuf);
}
dump_indirect(dn);
}
if (verbosity >= 5) {
/*
* Report the list of segments that comprise the object.
*/
uint64_t start = 0;
uint64_t end;
uint64_t blkfill = 1;
int minlvl = 1;
if (dn->dn_type == DMU_OT_DNODE) {
minlvl = 0;
blkfill = DNODES_PER_BLOCK;
}
for (;;) {
char segsize[32];
/* make sure nicenum has enough space */
_Static_assert(sizeof (segsize) >= NN_NUMBUF_SZ,
"segsize truncated");
error = dnode_next_offset(dn,
0, &start, minlvl, blkfill, 0);
if (error)
break;
end = start;
error = dnode_next_offset(dn,
DNODE_FIND_HOLE, &end, minlvl, blkfill, 0);
zdb_nicenum(end - start, segsize, sizeof (segsize));
(void) printf("\t\tsegment [%016llx, %016llx)"
" size %5s\n", (u_longlong_t)start,
(u_longlong_t)end, segsize);
if (error)
break;
start = end;
}
}
out:
if (db != NULL)
dmu_buf_rele(db, FTAG);
if (dnode_held)
dnode_rele(dn, FTAG);
}
static void
count_dir_mos_objects(dsl_dir_t *dd)
{
mos_obj_refd(dd->dd_object);
mos_obj_refd(dsl_dir_phys(dd)->dd_child_dir_zapobj);
mos_obj_refd(dsl_dir_phys(dd)->dd_deleg_zapobj);
mos_obj_refd(dsl_dir_phys(dd)->dd_props_zapobj);
mos_obj_refd(dsl_dir_phys(dd)->dd_clones);
/*
* The dd_crypto_obj can be referenced by multiple dsl_dir's.
* Ignore the references after the first one.
*/
mos_obj_refd_multiple(dd->dd_crypto_obj);
}
static void
count_ds_mos_objects(dsl_dataset_t *ds)
{
mos_obj_refd(ds->ds_object);
mos_obj_refd(dsl_dataset_phys(ds)->ds_next_clones_obj);
mos_obj_refd(dsl_dataset_phys(ds)->ds_props_obj);
mos_obj_refd(dsl_dataset_phys(ds)->ds_userrefs_obj);
mos_obj_refd(dsl_dataset_phys(ds)->ds_snapnames_zapobj);
mos_obj_refd(ds->ds_bookmarks_obj);
if (!dsl_dataset_is_snapshot(ds)) {
count_dir_mos_objects(ds->ds_dir);
}
}
static const char *const objset_types[DMU_OST_NUMTYPES] = {
"NONE", "META", "ZPL", "ZVOL", "OTHER", "ANY" };
/*
* Parse a string denoting a range of object IDs of the form
* <start>[:<end>[:flags]], and store the results in zor.
* Return 0 on success. On error, return 1 and update the msg
* pointer to point to a descriptive error message.
*/
static int
parse_object_range(char *range, zopt_object_range_t *zor, const char **msg)
{
uint64_t flags = 0;
char *p, *s, *dup, *flagstr, *tmp = NULL;
size_t len;
int i;
int rc = 0;
if (strchr(range, ':') == NULL) {
zor->zor_obj_start = strtoull(range, &p, 0);
if (*p != '\0') {
*msg = "Invalid characters in object ID";
rc = 1;
}
zor->zor_obj_start = ZDB_MAP_OBJECT_ID(zor->zor_obj_start);
zor->zor_obj_end = zor->zor_obj_start;
return (rc);
}
if (strchr(range, ':') == range) {
*msg = "Invalid leading colon";
rc = 1;
return (rc);
}
len = strlen(range);
if (range[len - 1] == ':') {
*msg = "Invalid trailing colon";
rc = 1;
return (rc);
}
dup = strdup(range);
s = strtok_r(dup, ":", &tmp);
zor->zor_obj_start = strtoull(s, &p, 0);
if (*p != '\0') {
*msg = "Invalid characters in start object ID";
rc = 1;
goto out;
}
s = strtok_r(NULL, ":", &tmp);
zor->zor_obj_end = strtoull(s, &p, 0);
if (*p != '\0') {
*msg = "Invalid characters in end object ID";
rc = 1;
goto out;
}
if (zor->zor_obj_start > zor->zor_obj_end) {
*msg = "Start object ID may not exceed end object ID";
rc = 1;
goto out;
}
s = strtok_r(NULL, ":", &tmp);
if (s == NULL) {
zor->zor_flags = ZOR_FLAG_ALL_TYPES;
goto out;
} else if (strtok_r(NULL, ":", &tmp) != NULL) {
*msg = "Invalid colon-delimited field after flags";
rc = 1;
goto out;
}
flagstr = s;
for (i = 0; flagstr[i]; i++) {
int bit;
boolean_t negation = (flagstr[i] == '-');
if (negation) {
i++;
if (flagstr[i] == '\0') {
*msg = "Invalid trailing negation operator";
rc = 1;
goto out;
}
}
bit = flagbits[(uchar_t)flagstr[i]];
if (bit == 0) {
*msg = "Invalid flag";
rc = 1;
goto out;
}
if (negation)
flags &= ~bit;
else
flags |= bit;
}
zor->zor_flags = flags;
zor->zor_obj_start = ZDB_MAP_OBJECT_ID(zor->zor_obj_start);
zor->zor_obj_end = ZDB_MAP_OBJECT_ID(zor->zor_obj_end);
out:
free(dup);
return (rc);
}
static void
dump_objset(objset_t *os)
{
dmu_objset_stats_t dds = { 0 };
uint64_t object, object_count;
uint64_t refdbytes, usedobjs, scratch;
char numbuf[32];
char blkbuf[BP_SPRINTF_LEN + 20];
char osname[ZFS_MAX_DATASET_NAME_LEN];
const char *type = "UNKNOWN";
int verbosity = dump_opt['d'];
boolean_t print_header;
unsigned i;
int error;
uint64_t total_slots_used = 0;
uint64_t max_slot_used = 0;
uint64_t dnode_slots;
uint64_t obj_start;
uint64_t obj_end;
uint64_t flags;
/* make sure nicenum has enough space */
_Static_assert(sizeof (numbuf) >= NN_NUMBUF_SZ, "numbuf truncated");
dsl_pool_config_enter(dmu_objset_pool(os), FTAG);
dmu_objset_fast_stat(os, &dds);
dsl_pool_config_exit(dmu_objset_pool(os), FTAG);
print_header = B_TRUE;
if (dds.dds_type < DMU_OST_NUMTYPES)
type = objset_types[dds.dds_type];
if (dds.dds_type == DMU_OST_META) {
dds.dds_creation_txg = TXG_INITIAL;
usedobjs = BP_GET_FILL(os->os_rootbp);
refdbytes = dsl_dir_phys(os->os_spa->spa_dsl_pool->dp_mos_dir)->
dd_used_bytes;
} else {
dmu_objset_space(os, &refdbytes, &scratch, &usedobjs, &scratch);
}
ASSERT3U(usedobjs, ==, BP_GET_FILL(os->os_rootbp));
zdb_nicenum(refdbytes, numbuf, sizeof (numbuf));
if (verbosity >= 4) {
(void) snprintf(blkbuf, sizeof (blkbuf), ", rootbp ");
(void) snprintf_blkptr(blkbuf + strlen(blkbuf),
sizeof (blkbuf) - strlen(blkbuf), os->os_rootbp);
} else {
blkbuf[0] = '\0';
}
dmu_objset_name(os, osname);
(void) printf("Dataset %s [%s], ID %llu, cr_txg %llu, "
"%s, %llu objects%s%s\n",
osname, type, (u_longlong_t)dmu_objset_id(os),
(u_longlong_t)dds.dds_creation_txg,
numbuf, (u_longlong_t)usedobjs, blkbuf,
(dds.dds_inconsistent) ? " (inconsistent)" : "");
for (i = 0; i < zopt_object_args; i++) {
obj_start = zopt_object_ranges[i].zor_obj_start;
obj_end = zopt_object_ranges[i].zor_obj_end;
flags = zopt_object_ranges[i].zor_flags;
object = obj_start;
if (object == 0 || obj_start == obj_end)
dump_object(os, object, verbosity, &print_header, NULL,
flags);
else
object--;
while ((dmu_object_next(os, &object, B_FALSE, 0) == 0) &&
object <= obj_end) {
dump_object(os, object, verbosity, &print_header, NULL,
flags);
}
}
if (zopt_object_args > 0) {
(void) printf("\n");
return;
}
if (dump_opt['i'] != 0 || verbosity >= 2)
dump_intent_log(dmu_objset_zil(os));
if (dmu_objset_ds(os) != NULL) {
dsl_dataset_t *ds = dmu_objset_ds(os);
dump_blkptr_list(&ds->ds_deadlist, "Deadlist");
if (dsl_deadlist_is_open(&ds->ds_dir->dd_livelist) &&
!dmu_objset_is_snapshot(os)) {
dump_blkptr_list(&ds->ds_dir->dd_livelist, "Livelist");
if (verify_dd_livelist(os) != 0)
fatal("livelist is incorrect");
}
if (dsl_dataset_remap_deadlist_exists(ds)) {
(void) printf("ds_remap_deadlist:\n");
dump_blkptr_list(&ds->ds_remap_deadlist, "Deadlist");
}
count_ds_mos_objects(ds);
}
if (dmu_objset_ds(os) != NULL)
dump_bookmarks(os, verbosity);
if (verbosity < 2)
return;
if (BP_IS_HOLE(os->os_rootbp))
return;
dump_object(os, 0, verbosity, &print_header, NULL, 0);
object_count = 0;
if (DMU_USERUSED_DNODE(os) != NULL &&
DMU_USERUSED_DNODE(os)->dn_type != 0) {
dump_object(os, DMU_USERUSED_OBJECT, verbosity, &print_header,
NULL, 0);
dump_object(os, DMU_GROUPUSED_OBJECT, verbosity, &print_header,
NULL, 0);
}
if (DMU_PROJECTUSED_DNODE(os) != NULL &&
DMU_PROJECTUSED_DNODE(os)->dn_type != 0)
dump_object(os, DMU_PROJECTUSED_OBJECT, verbosity,
&print_header, NULL, 0);
object = 0;
while ((error = dmu_object_next(os, &object, B_FALSE, 0)) == 0) {
dump_object(os, object, verbosity, &print_header, &dnode_slots,
0);
object_count++;
total_slots_used += dnode_slots;
max_slot_used = object + dnode_slots - 1;
}
(void) printf("\n");
(void) printf(" Dnode slots:\n");
(void) printf("\tTotal used: %10llu\n",
(u_longlong_t)total_slots_used);
(void) printf("\tMax used: %10llu\n",
(u_longlong_t)max_slot_used);
(void) printf("\tPercent empty: %10lf\n",
(double)(max_slot_used - total_slots_used)*100 /
(double)max_slot_used);
(void) printf("\n");
if (error != ESRCH) {
(void) fprintf(stderr, "dmu_object_next() = %d\n", error);
abort();
}
ASSERT3U(object_count, ==, usedobjs);
if (leaked_objects != 0) {
(void) printf("%d potentially leaked objects detected\n",
leaked_objects);
leaked_objects = 0;
}
}
static void
dump_uberblock(uberblock_t *ub, const char *header, const char *footer)
{
time_t timestamp = ub->ub_timestamp;
(void) printf("%s", header ? header : "");
(void) printf("\tmagic = %016llx\n", (u_longlong_t)ub->ub_magic);
(void) printf("\tversion = %llu\n", (u_longlong_t)ub->ub_version);
(void) printf("\ttxg = %llu\n", (u_longlong_t)ub->ub_txg);
(void) printf("\tguid_sum = %llu\n", (u_longlong_t)ub->ub_guid_sum);
(void) printf("\ttimestamp = %llu UTC = %s",
(u_longlong_t)ub->ub_timestamp, ctime(&timestamp));
(void) printf("\tmmp_magic = %016llx\n",
(u_longlong_t)ub->ub_mmp_magic);
if (MMP_VALID(ub)) {
(void) printf("\tmmp_delay = %0llu\n",
(u_longlong_t)ub->ub_mmp_delay);
if (MMP_SEQ_VALID(ub))
(void) printf("\tmmp_seq = %u\n",
(unsigned int) MMP_SEQ(ub));
if (MMP_FAIL_INT_VALID(ub))
(void) printf("\tmmp_fail = %u\n",
(unsigned int) MMP_FAIL_INT(ub));
if (MMP_INTERVAL_VALID(ub))
(void) printf("\tmmp_write = %u\n",
(unsigned int) MMP_INTERVAL(ub));
/* After MMP_* to make summarize_uberblock_mmp cleaner */
(void) printf("\tmmp_valid = %x\n",
(unsigned int) ub->ub_mmp_config & 0xFF);
}
if (dump_opt['u'] >= 4) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr(blkbuf, sizeof (blkbuf), &ub->ub_rootbp);
(void) printf("\trootbp = %s\n", blkbuf);
}
(void) printf("\tcheckpoint_txg = %llu\n",
(u_longlong_t)ub->ub_checkpoint_txg);
(void) printf("%s", footer ? footer : "");
}
static void
dump_config(spa_t *spa)
{
dmu_buf_t *db;
size_t nvsize = 0;
int error = 0;
error = dmu_bonus_hold(spa->spa_meta_objset,
spa->spa_config_object, FTAG, &db);
if (error == 0) {
nvsize = *(uint64_t *)db->db_data;
dmu_buf_rele(db, FTAG);
(void) printf("\nMOS Configuration:\n");
dump_packed_nvlist(spa->spa_meta_objset,
spa->spa_config_object, (void *)&nvsize, 1);
} else {
(void) fprintf(stderr, "dmu_bonus_hold(%llu) failed, errno %d",
(u_longlong_t)spa->spa_config_object, error);
}
}
static void
dump_cachefile(const char *cachefile)
{
int fd;
struct stat64 statbuf;
char *buf;
nvlist_t *config;
if ((fd = open64(cachefile, O_RDONLY)) < 0) {
(void) printf("cannot open '%s': %s\n", cachefile,
strerror(errno));
exit(1);
}
if (fstat64(fd, &statbuf) != 0) {
(void) printf("failed to stat '%s': %s\n", cachefile,
strerror(errno));
exit(1);
}
if ((buf = malloc(statbuf.st_size)) == NULL) {
(void) fprintf(stderr, "failed to allocate %llu bytes\n",
(u_longlong_t)statbuf.st_size);
exit(1);
}
if (read(fd, buf, statbuf.st_size) != statbuf.st_size) {
(void) fprintf(stderr, "failed to read %llu bytes\n",
(u_longlong_t)statbuf.st_size);
exit(1);
}
(void) close(fd);
if (nvlist_unpack(buf, statbuf.st_size, &config, 0) != 0) {
(void) fprintf(stderr, "failed to unpack nvlist\n");
exit(1);
}
free(buf);
dump_nvlist(config, 0);
nvlist_free(config);
}
/*
* ZFS label nvlist stats
*/
typedef struct zdb_nvl_stats {
int zns_list_count;
int zns_leaf_count;
size_t zns_leaf_largest;
size_t zns_leaf_total;
nvlist_t *zns_string;
nvlist_t *zns_uint64;
nvlist_t *zns_boolean;
} zdb_nvl_stats_t;
static void
collect_nvlist_stats(nvlist_t *nvl, zdb_nvl_stats_t *stats)
{
nvlist_t *list, **array;
nvpair_t *nvp = NULL;
const char *name;
uint_t i, items;
stats->zns_list_count++;
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
name = nvpair_name(nvp);
switch (nvpair_type(nvp)) {
case DATA_TYPE_STRING:
fnvlist_add_string(stats->zns_string, name,
fnvpair_value_string(nvp));
break;
case DATA_TYPE_UINT64:
fnvlist_add_uint64(stats->zns_uint64, name,
fnvpair_value_uint64(nvp));
break;
case DATA_TYPE_BOOLEAN:
fnvlist_add_boolean(stats->zns_boolean, name);
break;
case DATA_TYPE_NVLIST:
if (nvpair_value_nvlist(nvp, &list) == 0)
collect_nvlist_stats(list, stats);
break;
case DATA_TYPE_NVLIST_ARRAY:
if (nvpair_value_nvlist_array(nvp, &array, &items) != 0)
break;
for (i = 0; i < items; i++) {
collect_nvlist_stats(array[i], stats);
/* collect stats on leaf vdev */
if (strcmp(name, "children") == 0) {
size_t size;
(void) nvlist_size(array[i], &size,
NV_ENCODE_XDR);
stats->zns_leaf_total += size;
if (size > stats->zns_leaf_largest)
stats->zns_leaf_largest = size;
stats->zns_leaf_count++;
}
}
break;
default:
(void) printf("skip type %d!\n", (int)nvpair_type(nvp));
}
}
}
static void
dump_nvlist_stats(nvlist_t *nvl, size_t cap)
{
zdb_nvl_stats_t stats = { 0 };
size_t size, sum = 0, total;
size_t noise;
/* requires nvlist with non-unique names for stat collection */
VERIFY0(nvlist_alloc(&stats.zns_string, 0, 0));
VERIFY0(nvlist_alloc(&stats.zns_uint64, 0, 0));
VERIFY0(nvlist_alloc(&stats.zns_boolean, 0, 0));
VERIFY0(nvlist_size(stats.zns_boolean, &noise, NV_ENCODE_XDR));
(void) printf("\n\nZFS Label NVList Config Stats:\n");
VERIFY0(nvlist_size(nvl, &total, NV_ENCODE_XDR));
(void) printf(" %d bytes used, %d bytes free (using %4.1f%%)\n\n",
(int)total, (int)(cap - total), 100.0 * total / cap);
collect_nvlist_stats(nvl, &stats);
VERIFY0(nvlist_size(stats.zns_uint64, &size, NV_ENCODE_XDR));
size -= noise;
sum += size;
(void) printf("%12s %4d %6d bytes (%5.2f%%)\n", "integers:",
(int)fnvlist_num_pairs(stats.zns_uint64),
(int)size, 100.0 * size / total);
VERIFY0(nvlist_size(stats.zns_string, &size, NV_ENCODE_XDR));
size -= noise;
sum += size;
(void) printf("%12s %4d %6d bytes (%5.2f%%)\n", "strings:",
(int)fnvlist_num_pairs(stats.zns_string),
(int)size, 100.0 * size / total);
VERIFY0(nvlist_size(stats.zns_boolean, &size, NV_ENCODE_XDR));
size -= noise;
sum += size;
(void) printf("%12s %4d %6d bytes (%5.2f%%)\n", "booleans:",
(int)fnvlist_num_pairs(stats.zns_boolean),
(int)size, 100.0 * size / total);
size = total - sum; /* treat remainder as nvlist overhead */
(void) printf("%12s %4d %6d bytes (%5.2f%%)\n\n", "nvlists:",
stats.zns_list_count, (int)size, 100.0 * size / total);
if (stats.zns_leaf_count > 0) {
size_t average = stats.zns_leaf_total / stats.zns_leaf_count;
(void) printf("%12s %4d %6d bytes average\n", "leaf vdevs:",
stats.zns_leaf_count, (int)average);
(void) printf("%24d bytes largest\n",
(int)stats.zns_leaf_largest);
if (dump_opt['l'] >= 3 && average > 0)
(void) printf(" space for %d additional leaf vdevs\n",
(int)((cap - total) / average));
}
(void) printf("\n");
nvlist_free(stats.zns_string);
nvlist_free(stats.zns_uint64);
nvlist_free(stats.zns_boolean);
}
typedef struct cksum_record {
zio_cksum_t cksum;
boolean_t labels[VDEV_LABELS];
avl_node_t link;
} cksum_record_t;
static int
cksum_record_compare(const void *x1, const void *x2)
{
const cksum_record_t *l = (cksum_record_t *)x1;
const cksum_record_t *r = (cksum_record_t *)x2;
int arraysize = ARRAY_SIZE(l->cksum.zc_word);
int difference = 0;
for (int i = 0; i < arraysize; i++) {
difference = TREE_CMP(l->cksum.zc_word[i], r->cksum.zc_word[i]);
if (difference)
break;
}
return (difference);
}
static cksum_record_t *
cksum_record_alloc(zio_cksum_t *cksum, int l)
{
cksum_record_t *rec;
rec = umem_zalloc(sizeof (*rec), UMEM_NOFAIL);
rec->cksum = *cksum;
rec->labels[l] = B_TRUE;
return (rec);
}
static cksum_record_t *
cksum_record_lookup(avl_tree_t *tree, zio_cksum_t *cksum)
{
cksum_record_t lookup = { .cksum = *cksum };
avl_index_t where;
return (avl_find(tree, &lookup, &where));
}
static cksum_record_t *
cksum_record_insert(avl_tree_t *tree, zio_cksum_t *cksum, int l)
{
cksum_record_t *rec;
rec = cksum_record_lookup(tree, cksum);
if (rec) {
rec->labels[l] = B_TRUE;
} else {
rec = cksum_record_alloc(cksum, l);
avl_add(tree, rec);
}
return (rec);
}
static int
first_label(cksum_record_t *rec)
{
for (int i = 0; i < VDEV_LABELS; i++)
if (rec->labels[i])
return (i);
return (-1);
}
static void
print_label_numbers(const char *prefix, const cksum_record_t *rec)
{
fputs(prefix, stdout);
for (int i = 0; i < VDEV_LABELS; i++)
if (rec->labels[i] == B_TRUE)
printf("%d ", i);
putchar('\n');
}
#define MAX_UBERBLOCK_COUNT (VDEV_UBERBLOCK_RING >> UBERBLOCK_SHIFT)
typedef struct zdb_label {
vdev_label_t label;
uint64_t label_offset;
nvlist_t *config_nv;
cksum_record_t *config;
cksum_record_t *uberblocks[MAX_UBERBLOCK_COUNT];
boolean_t header_printed;
boolean_t read_failed;
boolean_t cksum_valid;
} zdb_label_t;
static void
print_label_header(zdb_label_t *label, int l)
{
if (dump_opt['q'])
return;
if (label->header_printed == B_TRUE)
return;
(void) printf("------------------------------------\n");
(void) printf("LABEL %d %s\n", l,
label->cksum_valid ? "" : "(Bad label cksum)");
(void) printf("------------------------------------\n");
label->header_printed = B_TRUE;
}
static void
print_l2arc_header(void)
{
(void) printf("------------------------------------\n");
(void) printf("L2ARC device header\n");
(void) printf("------------------------------------\n");
}
static void
print_l2arc_log_blocks(void)
{
(void) printf("------------------------------------\n");
(void) printf("L2ARC device log blocks\n");
(void) printf("------------------------------------\n");
}
static void
dump_l2arc_log_entries(uint64_t log_entries,
l2arc_log_ent_phys_t *le, uint64_t i)
{
for (int j = 0; j < log_entries; j++) {
dva_t dva = le[j].le_dva;
(void) printf("lb[%4llu]\tle[%4d]\tDVA asize: %llu, "
"vdev: %llu, offset: %llu\n",
(u_longlong_t)i, j + 1,
(u_longlong_t)DVA_GET_ASIZE(&dva),
(u_longlong_t)DVA_GET_VDEV(&dva),
(u_longlong_t)DVA_GET_OFFSET(&dva));
(void) printf("|\t\t\t\tbirth: %llu\n",
(u_longlong_t)le[j].le_birth);
(void) printf("|\t\t\t\tlsize: %llu\n",
(u_longlong_t)L2BLK_GET_LSIZE((&le[j])->le_prop));
(void) printf("|\t\t\t\tpsize: %llu\n",
(u_longlong_t)L2BLK_GET_PSIZE((&le[j])->le_prop));
(void) printf("|\t\t\t\tcompr: %llu\n",
(u_longlong_t)L2BLK_GET_COMPRESS((&le[j])->le_prop));
(void) printf("|\t\t\t\tcomplevel: %llu\n",
(u_longlong_t)(&le[j])->le_complevel);
(void) printf("|\t\t\t\ttype: %llu\n",
(u_longlong_t)L2BLK_GET_TYPE((&le[j])->le_prop));
(void) printf("|\t\t\t\tprotected: %llu\n",
(u_longlong_t)L2BLK_GET_PROTECTED((&le[j])->le_prop));
(void) printf("|\t\t\t\tprefetch: %llu\n",
(u_longlong_t)L2BLK_GET_PREFETCH((&le[j])->le_prop));
(void) printf("|\t\t\t\taddress: %llu\n",
(u_longlong_t)le[j].le_daddr);
(void) printf("|\t\t\t\tARC state: %llu\n",
(u_longlong_t)L2BLK_GET_STATE((&le[j])->le_prop));
(void) printf("|\n");
}
(void) printf("\n");
}
static void
dump_l2arc_log_blkptr(const l2arc_log_blkptr_t *lbps)
{
(void) printf("|\t\tdaddr: %llu\n", (u_longlong_t)lbps->lbp_daddr);
(void) printf("|\t\tpayload_asize: %llu\n",
(u_longlong_t)lbps->lbp_payload_asize);
(void) printf("|\t\tpayload_start: %llu\n",
(u_longlong_t)lbps->lbp_payload_start);
(void) printf("|\t\tlsize: %llu\n",
(u_longlong_t)L2BLK_GET_LSIZE(lbps->lbp_prop));
(void) printf("|\t\tasize: %llu\n",
(u_longlong_t)L2BLK_GET_PSIZE(lbps->lbp_prop));
(void) printf("|\t\tcompralgo: %llu\n",
(u_longlong_t)L2BLK_GET_COMPRESS(lbps->lbp_prop));
(void) printf("|\t\tcksumalgo: %llu\n",
(u_longlong_t)L2BLK_GET_CHECKSUM(lbps->lbp_prop));
(void) printf("|\n\n");
}
static void
dump_l2arc_log_blocks(int fd, const l2arc_dev_hdr_phys_t *l2dhdr,
l2arc_dev_hdr_phys_t *rebuild)
{
l2arc_log_blk_phys_t this_lb;
uint64_t asize;
l2arc_log_blkptr_t lbps[2];
abd_t *abd;
zio_cksum_t cksum;
int failed = 0;
l2arc_dev_t dev;
if (!dump_opt['q'])
print_l2arc_log_blocks();
memcpy(lbps, l2dhdr->dh_start_lbps, sizeof (lbps));
dev.l2ad_evict = l2dhdr->dh_evict;
dev.l2ad_start = l2dhdr->dh_start;
dev.l2ad_end = l2dhdr->dh_end;
if (l2dhdr->dh_start_lbps[0].lbp_daddr == 0) {
/* no log blocks to read */
if (!dump_opt['q']) {
(void) printf("No log blocks to read\n");
(void) printf("\n");
}
return;
} else {
dev.l2ad_hand = lbps[0].lbp_daddr +
L2BLK_GET_PSIZE((&lbps[0])->lbp_prop);
}
dev.l2ad_first = !!(l2dhdr->dh_flags & L2ARC_DEV_HDR_EVICT_FIRST);
for (;;) {
if (!l2arc_log_blkptr_valid(&dev, &lbps[0]))
break;
/* L2BLK_GET_PSIZE returns aligned size for log blocks */
asize = L2BLK_GET_PSIZE((&lbps[0])->lbp_prop);
if (pread64(fd, &this_lb, asize, lbps[0].lbp_daddr) != asize) {
if (!dump_opt['q']) {
(void) printf("Error while reading next log "
"block\n\n");
}
break;
}
fletcher_4_native_varsize(&this_lb, asize, &cksum);
if (!ZIO_CHECKSUM_EQUAL(cksum, lbps[0].lbp_cksum)) {
failed++;
if (!dump_opt['q']) {
(void) printf("Invalid cksum\n");
dump_l2arc_log_blkptr(&lbps[0]);
}
break;
}
switch (L2BLK_GET_COMPRESS((&lbps[0])->lbp_prop)) {
case ZIO_COMPRESS_OFF:
break;
default:
abd = abd_alloc_for_io(asize, B_TRUE);
abd_copy_from_buf_off(abd, &this_lb, 0, asize);
if (zio_decompress_data(L2BLK_GET_COMPRESS(
(&lbps[0])->lbp_prop), abd, &this_lb,
asize, sizeof (this_lb), NULL) != 0) {
(void) printf("L2ARC block decompression "
"failed\n");
abd_free(abd);
goto out;
}
abd_free(abd);
break;
}
if (this_lb.lb_magic == BSWAP_64(L2ARC_LOG_BLK_MAGIC))
byteswap_uint64_array(&this_lb, sizeof (this_lb));
if (this_lb.lb_magic != L2ARC_LOG_BLK_MAGIC) {
if (!dump_opt['q'])
(void) printf("Invalid log block magic\n\n");
break;
}
rebuild->dh_lb_count++;
rebuild->dh_lb_asize += asize;
if (dump_opt['l'] > 1 && !dump_opt['q']) {
(void) printf("lb[%4llu]\tmagic: %llu\n",
(u_longlong_t)rebuild->dh_lb_count,
(u_longlong_t)this_lb.lb_magic);
dump_l2arc_log_blkptr(&lbps[0]);
}
if (dump_opt['l'] > 2 && !dump_opt['q'])
dump_l2arc_log_entries(l2dhdr->dh_log_entries,
this_lb.lb_entries,
rebuild->dh_lb_count);
if (l2arc_range_check_overlap(lbps[1].lbp_payload_start,
lbps[0].lbp_payload_start, dev.l2ad_evict) &&
!dev.l2ad_first)
break;
lbps[0] = lbps[1];
lbps[1] = this_lb.lb_prev_lbp;
}
out:
if (!dump_opt['q']) {
(void) printf("log_blk_count:\t %llu with valid cksum\n",
(u_longlong_t)rebuild->dh_lb_count);
(void) printf("\t\t %d with invalid cksum\n", failed);
(void) printf("log_blk_asize:\t %llu\n\n",
(u_longlong_t)rebuild->dh_lb_asize);
}
}
static int
dump_l2arc_header(int fd)
{
l2arc_dev_hdr_phys_t l2dhdr = {0}, rebuild = {0};
int error = B_FALSE;
if (pread64(fd, &l2dhdr, sizeof (l2dhdr),
VDEV_LABEL_START_SIZE) != sizeof (l2dhdr)) {
error = B_TRUE;
} else {
if (l2dhdr.dh_magic == BSWAP_64(L2ARC_DEV_HDR_MAGIC))
byteswap_uint64_array(&l2dhdr, sizeof (l2dhdr));
if (l2dhdr.dh_magic != L2ARC_DEV_HDR_MAGIC)
error = B_TRUE;
}
if (error) {
(void) printf("L2ARC device header not found\n\n");
/* Do not return an error here for backward compatibility */
return (0);
} else if (!dump_opt['q']) {
print_l2arc_header();
(void) printf(" magic: %llu\n",
(u_longlong_t)l2dhdr.dh_magic);
(void) printf(" version: %llu\n",
(u_longlong_t)l2dhdr.dh_version);
(void) printf(" pool_guid: %llu\n",
(u_longlong_t)l2dhdr.dh_spa_guid);
(void) printf(" flags: %llu\n",
(u_longlong_t)l2dhdr.dh_flags);
(void) printf(" start_lbps[0]: %llu\n",
(u_longlong_t)
l2dhdr.dh_start_lbps[0].lbp_daddr);
(void) printf(" start_lbps[1]: %llu\n",
(u_longlong_t)
l2dhdr.dh_start_lbps[1].lbp_daddr);
(void) printf(" log_blk_ent: %llu\n",
(u_longlong_t)l2dhdr.dh_log_entries);
(void) printf(" start: %llu\n",
(u_longlong_t)l2dhdr.dh_start);
(void) printf(" end: %llu\n",
(u_longlong_t)l2dhdr.dh_end);
(void) printf(" evict: %llu\n",
(u_longlong_t)l2dhdr.dh_evict);
(void) printf(" lb_asize_refcount: %llu\n",
(u_longlong_t)l2dhdr.dh_lb_asize);
(void) printf(" lb_count_refcount: %llu\n",
(u_longlong_t)l2dhdr.dh_lb_count);
(void) printf(" trim_action_time: %llu\n",
(u_longlong_t)l2dhdr.dh_trim_action_time);
(void) printf(" trim_state: %llu\n\n",
(u_longlong_t)l2dhdr.dh_trim_state);
}
dump_l2arc_log_blocks(fd, &l2dhdr, &rebuild);
/*
* The total aligned size of log blocks and the number of log blocks
* reported in the header of the device may be less than what zdb
* reports by dump_l2arc_log_blocks() which emulates l2arc_rebuild().
* This happens because dump_l2arc_log_blocks() lacks the memory
* pressure valve that l2arc_rebuild() has. Thus, if we are on a system
* with low memory, l2arc_rebuild will exit prematurely and dh_lb_asize
* and dh_lb_count will be lower to begin with than what exists on the
* device. This is normal and zdb should not exit with an error. The
* opposite case should never happen though, the values reported in the
* header should never be higher than what dump_l2arc_log_blocks() and
* l2arc_rebuild() report. If this happens there is a leak in the
* accounting of log blocks.
*/
if (l2dhdr.dh_lb_asize > rebuild.dh_lb_asize ||
l2dhdr.dh_lb_count > rebuild.dh_lb_count)
return (1);
return (0);
}
static void
dump_config_from_label(zdb_label_t *label, size_t buflen, int l)
{
if (dump_opt['q'])
return;
if ((dump_opt['l'] < 3) && (first_label(label->config) != l))
return;
print_label_header(label, l);
dump_nvlist(label->config_nv, 4);
print_label_numbers(" labels = ", label->config);
if (dump_opt['l'] >= 2)
dump_nvlist_stats(label->config_nv, buflen);
}
#define ZDB_MAX_UB_HEADER_SIZE 32
static void
dump_label_uberblocks(zdb_label_t *label, uint64_t ashift, int label_num)
{
vdev_t vd;
char header[ZDB_MAX_UB_HEADER_SIZE];
vd.vdev_ashift = ashift;
vd.vdev_top = &vd;
for (int i = 0; i < VDEV_UBERBLOCK_COUNT(&vd); i++) {
uint64_t uoff = VDEV_UBERBLOCK_OFFSET(&vd, i);
uberblock_t *ub = (void *)((char *)&label->label + uoff);
cksum_record_t *rec = label->uberblocks[i];
if (rec == NULL) {
if (dump_opt['u'] >= 2) {
print_label_header(label, label_num);
(void) printf(" Uberblock[%d] invalid\n", i);
}
continue;
}
if ((dump_opt['u'] < 3) && (first_label(rec) != label_num))
continue;
if ((dump_opt['u'] < 4) &&
(ub->ub_mmp_magic == MMP_MAGIC) && ub->ub_mmp_delay &&
(i >= VDEV_UBERBLOCK_COUNT(&vd) - MMP_BLOCKS_PER_LABEL))
continue;
print_label_header(label, label_num);
(void) snprintf(header, ZDB_MAX_UB_HEADER_SIZE,
" Uberblock[%d]\n", i);
dump_uberblock(ub, header, "");
print_label_numbers(" labels = ", rec);
}
}
static char curpath[PATH_MAX];
/*
* Iterate through the path components, recursively passing
* current one's obj and remaining path until we find the obj
* for the last one.
*/
static int
dump_path_impl(objset_t *os, uint64_t obj, char *name, uint64_t *retobj)
{
int err;
boolean_t header = B_TRUE;
uint64_t child_obj;
char *s;
dmu_buf_t *db;
dmu_object_info_t doi;
if ((s = strchr(name, '/')) != NULL)
*s = '\0';
err = zap_lookup(os, obj, name, 8, 1, &child_obj);
(void) strlcat(curpath, name, sizeof (curpath));
if (err != 0) {
(void) fprintf(stderr, "failed to lookup %s: %s\n",
curpath, strerror(err));
return (err);
}
child_obj = ZFS_DIRENT_OBJ(child_obj);
err = sa_buf_hold(os, child_obj, FTAG, &db);
if (err != 0) {
(void) fprintf(stderr,
"failed to get SA dbuf for obj %llu: %s\n",
(u_longlong_t)child_obj, strerror(err));
return (EINVAL);
}
dmu_object_info_from_db(db, &doi);
sa_buf_rele(db, FTAG);
if (doi.doi_bonus_type != DMU_OT_SA &&
doi.doi_bonus_type != DMU_OT_ZNODE) {
(void) fprintf(stderr, "invalid bonus type %d for obj %llu\n",
doi.doi_bonus_type, (u_longlong_t)child_obj);
return (EINVAL);
}
if (dump_opt['v'] > 6) {
(void) printf("obj=%llu %s type=%d bonustype=%d\n",
(u_longlong_t)child_obj, curpath, doi.doi_type,
doi.doi_bonus_type);
}
(void) strlcat(curpath, "/", sizeof (curpath));
switch (doi.doi_type) {
case DMU_OT_DIRECTORY_CONTENTS:
if (s != NULL && *(s + 1) != '\0')
return (dump_path_impl(os, child_obj, s + 1, retobj));
zfs_fallthrough;
case DMU_OT_PLAIN_FILE_CONTENTS:
if (retobj != NULL) {
*retobj = child_obj;
} else {
dump_object(os, child_obj, dump_opt['v'], &header,
NULL, 0);
}
return (0);
default:
(void) fprintf(stderr, "object %llu has non-file/directory "
"type %d\n", (u_longlong_t)obj, doi.doi_type);
break;
}
return (EINVAL);
}
/*
* Dump the blocks for the object specified by path inside the dataset.
*/
static int
dump_path(char *ds, char *path, uint64_t *retobj)
{
int err;
objset_t *os;
uint64_t root_obj;
err = open_objset(ds, FTAG, &os);
if (err != 0)
return (err);
err = zap_lookup(os, MASTER_NODE_OBJ, ZFS_ROOT_OBJ, 8, 1, &root_obj);
if (err != 0) {
(void) fprintf(stderr, "can't lookup root znode: %s\n",
strerror(err));
close_objset(os, FTAG);
return (EINVAL);
}
(void) snprintf(curpath, sizeof (curpath), "dataset=%s path=/", ds);
err = dump_path_impl(os, root_obj, path, retobj);
close_objset(os, FTAG);
return (err);
}
static int
zdb_copy_object(objset_t *os, uint64_t srcobj, char *destfile)
{
int err = 0;
uint64_t size, readsize, oursize, offset;
ssize_t writesize;
sa_handle_t *hdl;
(void) printf("Copying object %" PRIu64 " to file %s\n", srcobj,
destfile);
VERIFY3P(os, ==, sa_os);
if ((err = sa_handle_get(os, srcobj, NULL, SA_HDL_PRIVATE, &hdl))) {
(void) printf("Failed to get handle for SA znode\n");
return (err);
}
if ((err = sa_lookup(hdl, sa_attr_table[ZPL_SIZE], &size, 8))) {
(void) sa_handle_destroy(hdl);
return (err);
}
(void) sa_handle_destroy(hdl);
(void) printf("Object %" PRIu64 " is %" PRIu64 " bytes\n", srcobj,
size);
if (size == 0) {
return (EINVAL);
}
int fd = open(destfile, O_WRONLY | O_CREAT | O_TRUNC, 0644);
if (fd == -1)
return (errno);
/*
* We cap the size at 1 mebibyte here to prevent
* allocation failures and nigh-infinite printing if the
* object is extremely large.
*/
oursize = MIN(size, 1 << 20);
offset = 0;
char *buf = kmem_alloc(oursize, KM_NOSLEEP);
if (buf == NULL) {
(void) close(fd);
return (ENOMEM);
}
while (offset < size) {
readsize = MIN(size - offset, 1 << 20);
err = dmu_read(os, srcobj, offset, readsize, buf, 0);
if (err != 0) {
(void) printf("got error %u from dmu_read\n", err);
kmem_free(buf, oursize);
(void) close(fd);
return (err);
}
if (dump_opt['v'] > 3) {
(void) printf("Read offset=%" PRIu64 " size=%" PRIu64
" error=%d\n", offset, readsize, err);
}
writesize = write(fd, buf, readsize);
if (writesize < 0) {
err = errno;
break;
} else if (writesize != readsize) {
/* Incomplete write */
(void) fprintf(stderr, "Short write, only wrote %llu of"
" %" PRIu64 " bytes, exiting...\n",
(u_longlong_t)writesize, readsize);
break;
}
offset += readsize;
}
(void) close(fd);
if (buf != NULL)
kmem_free(buf, oursize);
return (err);
}
static boolean_t
label_cksum_valid(vdev_label_t *label, uint64_t offset)
{
zio_checksum_info_t *ci = &zio_checksum_table[ZIO_CHECKSUM_LABEL];
zio_cksum_t expected_cksum;
zio_cksum_t actual_cksum;
zio_cksum_t verifier;
zio_eck_t *eck;
int byteswap;
void *data = (char *)label + offsetof(vdev_label_t, vl_vdev_phys);
eck = (zio_eck_t *)((char *)(data) + VDEV_PHYS_SIZE) - 1;
offset += offsetof(vdev_label_t, vl_vdev_phys);
ZIO_SET_CHECKSUM(&verifier, offset, 0, 0, 0);
byteswap = (eck->zec_magic == BSWAP_64(ZEC_MAGIC));
if (byteswap)
byteswap_uint64_array(&verifier, sizeof (zio_cksum_t));
expected_cksum = eck->zec_cksum;
eck->zec_cksum = verifier;
abd_t *abd = abd_get_from_buf(data, VDEV_PHYS_SIZE);
ci->ci_func[byteswap](abd, VDEV_PHYS_SIZE, NULL, &actual_cksum);
abd_free(abd);
if (byteswap)
byteswap_uint64_array(&expected_cksum, sizeof (zio_cksum_t));
if (ZIO_CHECKSUM_EQUAL(actual_cksum, expected_cksum))
return (B_TRUE);
return (B_FALSE);
}
static int
dump_label(const char *dev)
{
char path[MAXPATHLEN];
zdb_label_t labels[VDEV_LABELS] = {{{{0}}}};
uint64_t psize, ashift, l2cache;
struct stat64 statbuf;
boolean_t config_found = B_FALSE;
boolean_t error = B_FALSE;
boolean_t read_l2arc_header = B_FALSE;
avl_tree_t config_tree;
avl_tree_t uberblock_tree;
void *node, *cookie;
int fd;
/*
* Check if we were given absolute path and use it as is.
* Otherwise if the provided vdev name doesn't point to a file,
* try prepending expected disk paths and partition numbers.
*/
(void) strlcpy(path, dev, sizeof (path));
if (dev[0] != '/' && stat64(path, &statbuf) != 0) {
int error;
error = zfs_resolve_shortname(dev, path, MAXPATHLEN);
if (error == 0 && zfs_dev_is_whole_disk(path)) {
if (zfs_append_partition(path, MAXPATHLEN) == -1)
error = ENOENT;
}
if (error || (stat64(path, &statbuf) != 0)) {
(void) printf("failed to find device %s, try "
"specifying absolute path instead\n", dev);
return (1);
}
}
if ((fd = open64(path, O_RDONLY)) < 0) {
(void) printf("cannot open '%s': %s\n", path, strerror(errno));
exit(1);
}
if (fstat64_blk(fd, &statbuf) != 0) {
(void) printf("failed to stat '%s': %s\n", path,
strerror(errno));
(void) close(fd);
exit(1);
}
if (S_ISBLK(statbuf.st_mode) && zfs_dev_flush(fd) != 0)
(void) printf("failed to invalidate cache '%s' : %s\n", path,
strerror(errno));
avl_create(&config_tree, cksum_record_compare,
sizeof (cksum_record_t), offsetof(cksum_record_t, link));
avl_create(&uberblock_tree, cksum_record_compare,
sizeof (cksum_record_t), offsetof(cksum_record_t, link));
psize = statbuf.st_size;
psize = P2ALIGN(psize, (uint64_t)sizeof (vdev_label_t));
ashift = SPA_MINBLOCKSHIFT;
/*
* 1. Read the label from disk
* 2. Verify label cksum
* 3. Unpack the configuration and insert in config tree.
* 4. Traverse all uberblocks and insert in uberblock tree.
*/
for (int l = 0; l < VDEV_LABELS; l++) {
zdb_label_t *label = &labels[l];
char *buf = label->label.vl_vdev_phys.vp_nvlist;
size_t buflen = sizeof (label->label.vl_vdev_phys.vp_nvlist);
nvlist_t *config;
cksum_record_t *rec;
zio_cksum_t cksum;
vdev_t vd;
label->label_offset = vdev_label_offset(psize, l, 0);
if (pread64(fd, &label->label, sizeof (label->label),
label->label_offset) != sizeof (label->label)) {
if (!dump_opt['q'])
(void) printf("failed to read label %d\n", l);
label->read_failed = B_TRUE;
error = B_TRUE;
continue;
}
label->read_failed = B_FALSE;
label->cksum_valid = label_cksum_valid(&label->label,
label->label_offset);
if (nvlist_unpack(buf, buflen, &config, 0) == 0) {
nvlist_t *vdev_tree = NULL;
size_t size;
if ((nvlist_lookup_nvlist(config,
ZPOOL_CONFIG_VDEV_TREE, &vdev_tree) != 0) ||
(nvlist_lookup_uint64(vdev_tree,
ZPOOL_CONFIG_ASHIFT, &ashift) != 0))
ashift = SPA_MINBLOCKSHIFT;
if (nvlist_size(config, &size, NV_ENCODE_XDR) != 0)
size = buflen;
/* If the device is a cache device clear the header. */
if (!read_l2arc_header) {
if (nvlist_lookup_uint64(config,
ZPOOL_CONFIG_POOL_STATE, &l2cache) == 0 &&
l2cache == POOL_STATE_L2CACHE) {
read_l2arc_header = B_TRUE;
}
}
fletcher_4_native_varsize(buf, size, &cksum);
rec = cksum_record_insert(&config_tree, &cksum, l);
label->config = rec;
label->config_nv = config;
config_found = B_TRUE;
} else {
error = B_TRUE;
}
vd.vdev_ashift = ashift;
vd.vdev_top = &vd;
for (int i = 0; i < VDEV_UBERBLOCK_COUNT(&vd); i++) {
uint64_t uoff = VDEV_UBERBLOCK_OFFSET(&vd, i);
uberblock_t *ub = (void *)((char *)label + uoff);
if (uberblock_verify(ub))
continue;
fletcher_4_native_varsize(ub, sizeof (*ub), &cksum);
rec = cksum_record_insert(&uberblock_tree, &cksum, l);
label->uberblocks[i] = rec;
}
}
/*
* Dump the label and uberblocks.
*/
for (int l = 0; l < VDEV_LABELS; l++) {
zdb_label_t *label = &labels[l];
size_t buflen = sizeof (label->label.vl_vdev_phys.vp_nvlist);
if (label->read_failed == B_TRUE)
continue;
if (label->config_nv) {
dump_config_from_label(label, buflen, l);
} else {
if (!dump_opt['q'])
(void) printf("failed to unpack label %d\n", l);
}
if (dump_opt['u'])
dump_label_uberblocks(label, ashift, l);
nvlist_free(label->config_nv);
}
/*
* Dump the L2ARC header, if existent.
*/
if (read_l2arc_header)
error |= dump_l2arc_header(fd);
cookie = NULL;
while ((node = avl_destroy_nodes(&config_tree, &cookie)) != NULL)
umem_free(node, sizeof (cksum_record_t));
cookie = NULL;
while ((node = avl_destroy_nodes(&uberblock_tree, &cookie)) != NULL)
umem_free(node, sizeof (cksum_record_t));
avl_destroy(&config_tree);
avl_destroy(&uberblock_tree);
(void) close(fd);
return (config_found == B_FALSE ? 2 :
(error == B_TRUE ? 1 : 0));
}
static uint64_t dataset_feature_count[SPA_FEATURES];
static uint64_t global_feature_count[SPA_FEATURES];
static uint64_t remap_deadlist_count = 0;
static int
dump_one_objset(const char *dsname, void *arg)
{
(void) arg;
int error;
objset_t *os;
spa_feature_t f;
error = open_objset(dsname, FTAG, &os);
if (error != 0)
return (0);
for (f = 0; f < SPA_FEATURES; f++) {
if (!dsl_dataset_feature_is_active(dmu_objset_ds(os), f))
continue;
ASSERT(spa_feature_table[f].fi_flags &
ZFEATURE_FLAG_PER_DATASET);
dataset_feature_count[f]++;
}
if (dsl_dataset_remap_deadlist_exists(dmu_objset_ds(os))) {
remap_deadlist_count++;
}
for (dsl_bookmark_node_t *dbn =
avl_first(&dmu_objset_ds(os)->ds_bookmarks); dbn != NULL;
dbn = AVL_NEXT(&dmu_objset_ds(os)->ds_bookmarks, dbn)) {
mos_obj_refd(dbn->dbn_phys.zbm_redaction_obj);
if (dbn->dbn_phys.zbm_redaction_obj != 0)
global_feature_count[SPA_FEATURE_REDACTION_BOOKMARKS]++;
if (dbn->dbn_phys.zbm_flags & ZBM_FLAG_HAS_FBN)
global_feature_count[SPA_FEATURE_BOOKMARK_WRITTEN]++;
}
if (dsl_deadlist_is_open(&dmu_objset_ds(os)->ds_dir->dd_livelist) &&
!dmu_objset_is_snapshot(os)) {
global_feature_count[SPA_FEATURE_LIVELIST]++;
}
dump_objset(os);
close_objset(os, FTAG);
fuid_table_destroy();
return (0);
}
/*
* Block statistics.
*/
#define PSIZE_HISTO_SIZE (SPA_OLD_MAXBLOCKSIZE / SPA_MINBLOCKSIZE + 2)
typedef struct zdb_blkstats {
uint64_t zb_asize;
uint64_t zb_lsize;
uint64_t zb_psize;
uint64_t zb_count;
uint64_t zb_gangs;
uint64_t zb_ditto_samevdev;
uint64_t zb_ditto_same_ms;
uint64_t zb_psize_histogram[PSIZE_HISTO_SIZE];
} zdb_blkstats_t;
/*
* Extended object types to report deferred frees and dedup auto-ditto blocks.
*/
#define ZDB_OT_DEFERRED (DMU_OT_NUMTYPES + 0)
#define ZDB_OT_DITTO (DMU_OT_NUMTYPES + 1)
#define ZDB_OT_OTHER (DMU_OT_NUMTYPES + 2)
#define ZDB_OT_TOTAL (DMU_OT_NUMTYPES + 3)
static const char *zdb_ot_extname[] = {
"deferred free",
"dedup ditto",
"other",
"Total",
};
#define ZB_TOTAL DN_MAX_LEVELS
#define SPA_MAX_FOR_16M (SPA_MAXBLOCKSHIFT+1)
typedef struct zdb_cb {
zdb_blkstats_t zcb_type[ZB_TOTAL + 1][ZDB_OT_TOTAL + 1];
uint64_t zcb_removing_size;
uint64_t zcb_checkpoint_size;
uint64_t zcb_dedup_asize;
uint64_t zcb_dedup_blocks;
uint64_t zcb_psize_count[SPA_MAX_FOR_16M];
uint64_t zcb_lsize_count[SPA_MAX_FOR_16M];
uint64_t zcb_asize_count[SPA_MAX_FOR_16M];
uint64_t zcb_psize_len[SPA_MAX_FOR_16M];
uint64_t zcb_lsize_len[SPA_MAX_FOR_16M];
uint64_t zcb_asize_len[SPA_MAX_FOR_16M];
uint64_t zcb_psize_total;
uint64_t zcb_lsize_total;
uint64_t zcb_asize_total;
uint64_t zcb_embedded_blocks[NUM_BP_EMBEDDED_TYPES];
uint64_t zcb_embedded_histogram[NUM_BP_EMBEDDED_TYPES]
[BPE_PAYLOAD_SIZE + 1];
uint64_t zcb_start;
hrtime_t zcb_lastprint;
uint64_t zcb_totalasize;
uint64_t zcb_errors[256];
int zcb_readfails;
int zcb_haderrors;
spa_t *zcb_spa;
uint32_t **zcb_vd_obsolete_counts;
} zdb_cb_t;
/* test if two DVA offsets from same vdev are within the same metaslab */
static boolean_t
same_metaslab(spa_t *spa, uint64_t vdev, uint64_t off1, uint64_t off2)
{
vdev_t *vd = vdev_lookup_top(spa, vdev);
uint64_t ms_shift = vd->vdev_ms_shift;
return ((off1 >> ms_shift) == (off2 >> ms_shift));
}
/*
* Used to simplify reporting of the histogram data.
*/
typedef struct one_histo {
const char *name;
uint64_t *count;
uint64_t *len;
uint64_t cumulative;
} one_histo_t;
/*
* The number of separate histograms processed for psize, lsize and asize.
*/
#define NUM_HISTO 3
/*
* This routine will create a fixed column size output of three different
* histograms showing by blocksize of 512 - 2^ SPA_MAX_FOR_16M
* the count, length and cumulative length of the psize, lsize and
* asize blocks.
*
* All three types of blocks are listed on a single line
*
* By default the table is printed in nicenumber format (e.g. 123K) but
* if the '-P' parameter is specified then the full raw number (parseable)
* is printed out.
*/
static void
dump_size_histograms(zdb_cb_t *zcb)
{
/*
* A temporary buffer that allows us to convert a number into
* a string using zdb_nicenumber to allow either raw or human
* readable numbers to be output.
*/
char numbuf[32];
/*
* Define titles which are used in the headers of the tables
* printed by this routine.
*/
const char blocksize_title1[] = "block";
const char blocksize_title2[] = "size";
const char count_title[] = "Count";
const char length_title[] = "Size";
const char cumulative_title[] = "Cum.";
/*
* Setup the histogram arrays (psize, lsize, and asize).
*/
one_histo_t parm_histo[NUM_HISTO];
parm_histo[0].name = "psize";
parm_histo[0].count = zcb->zcb_psize_count;
parm_histo[0].len = zcb->zcb_psize_len;
parm_histo[0].cumulative = 0;
parm_histo[1].name = "lsize";
parm_histo[1].count = zcb->zcb_lsize_count;
parm_histo[1].len = zcb->zcb_lsize_len;
parm_histo[1].cumulative = 0;
parm_histo[2].name = "asize";
parm_histo[2].count = zcb->zcb_asize_count;
parm_histo[2].len = zcb->zcb_asize_len;
parm_histo[2].cumulative = 0;
(void) printf("\nBlock Size Histogram\n");
/*
* Print the first line titles
*/
if (dump_opt['P'])
(void) printf("\n%s\t", blocksize_title1);
else
(void) printf("\n%7s ", blocksize_title1);
for (int j = 0; j < NUM_HISTO; j++) {
if (dump_opt['P']) {
if (j < NUM_HISTO - 1) {
(void) printf("%s\t\t\t", parm_histo[j].name);
} else {
/* Don't print trailing spaces */
(void) printf(" %s", parm_histo[j].name);
}
} else {
if (j < NUM_HISTO - 1) {
/* Left aligned strings in the output */
(void) printf("%-7s ",
parm_histo[j].name);
} else {
/* Don't print trailing spaces */
(void) printf("%s", parm_histo[j].name);
}
}
}
(void) printf("\n");
/*
* Print the second line titles
*/
if (dump_opt['P']) {
(void) printf("%s\t", blocksize_title2);
} else {
(void) printf("%7s ", blocksize_title2);
}
for (int i = 0; i < NUM_HISTO; i++) {
if (dump_opt['P']) {
(void) printf("%s\t%s\t%s\t",
count_title, length_title, cumulative_title);
} else {
(void) printf("%7s%7s%7s",
count_title, length_title, cumulative_title);
}
}
(void) printf("\n");
/*
* Print the rows
*/
for (int i = SPA_MINBLOCKSHIFT; i < SPA_MAX_FOR_16M; i++) {
/*
* Print the first column showing the blocksize
*/
zdb_nicenum((1ULL << i), numbuf, sizeof (numbuf));
if (dump_opt['P']) {
printf("%s", numbuf);
} else {
printf("%7s:", numbuf);
}
/*
* Print the remaining set of 3 columns per size:
* for psize, lsize and asize
*/
for (int j = 0; j < NUM_HISTO; j++) {
parm_histo[j].cumulative += parm_histo[j].len[i];
zdb_nicenum(parm_histo[j].count[i],
numbuf, sizeof (numbuf));
if (dump_opt['P'])
(void) printf("\t%s", numbuf);
else
(void) printf("%7s", numbuf);
zdb_nicenum(parm_histo[j].len[i],
numbuf, sizeof (numbuf));
if (dump_opt['P'])
(void) printf("\t%s", numbuf);
else
(void) printf("%7s", numbuf);
zdb_nicenum(parm_histo[j].cumulative,
numbuf, sizeof (numbuf));
if (dump_opt['P'])
(void) printf("\t%s", numbuf);
else
(void) printf("%7s", numbuf);
}
(void) printf("\n");
}
}
static void
zdb_count_block(zdb_cb_t *zcb, zilog_t *zilog, const blkptr_t *bp,
dmu_object_type_t type)
{
uint64_t refcnt = 0;
int i;
ASSERT(type < ZDB_OT_TOTAL);
if (zilog && zil_bp_tree_add(zilog, bp) != 0)
return;
spa_config_enter(zcb->zcb_spa, SCL_CONFIG, FTAG, RW_READER);
for (i = 0; i < 4; i++) {
int l = (i < 2) ? BP_GET_LEVEL(bp) : ZB_TOTAL;
int t = (i & 1) ? type : ZDB_OT_TOTAL;
int equal;
zdb_blkstats_t *zb = &zcb->zcb_type[l][t];
zb->zb_asize += BP_GET_ASIZE(bp);
zb->zb_lsize += BP_GET_LSIZE(bp);
zb->zb_psize += BP_GET_PSIZE(bp);
zb->zb_count++;
/*
* The histogram is only big enough to record blocks up to
* SPA_OLD_MAXBLOCKSIZE; larger blocks go into the last,
* "other", bucket.
*/
unsigned idx = BP_GET_PSIZE(bp) >> SPA_MINBLOCKSHIFT;
idx = MIN(idx, SPA_OLD_MAXBLOCKSIZE / SPA_MINBLOCKSIZE + 1);
zb->zb_psize_histogram[idx]++;
zb->zb_gangs += BP_COUNT_GANG(bp);
switch (BP_GET_NDVAS(bp)) {
case 2:
if (DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[1])) {
zb->zb_ditto_samevdev++;
if (same_metaslab(zcb->zcb_spa,
DVA_GET_VDEV(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[1])))
zb->zb_ditto_same_ms++;
}
break;
case 3:
equal = (DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[1])) +
(DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[2])) +
(DVA_GET_VDEV(&bp->blk_dva[1]) ==
DVA_GET_VDEV(&bp->blk_dva[2]));
if (equal != 0) {
zb->zb_ditto_samevdev++;
if (DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[1]) &&
same_metaslab(zcb->zcb_spa,
DVA_GET_VDEV(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[1])))
zb->zb_ditto_same_ms++;
else if (DVA_GET_VDEV(&bp->blk_dva[0]) ==
DVA_GET_VDEV(&bp->blk_dva[2]) &&
same_metaslab(zcb->zcb_spa,
DVA_GET_VDEV(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[0]),
DVA_GET_OFFSET(&bp->blk_dva[2])))
zb->zb_ditto_same_ms++;
else if (DVA_GET_VDEV(&bp->blk_dva[1]) ==
DVA_GET_VDEV(&bp->blk_dva[2]) &&
same_metaslab(zcb->zcb_spa,
DVA_GET_VDEV(&bp->blk_dva[1]),
DVA_GET_OFFSET(&bp->blk_dva[1]),
DVA_GET_OFFSET(&bp->blk_dva[2])))
zb->zb_ditto_same_ms++;
}
break;
}
}
spa_config_exit(zcb->zcb_spa, SCL_CONFIG, FTAG);
if (BP_IS_EMBEDDED(bp)) {
zcb->zcb_embedded_blocks[BPE_GET_ETYPE(bp)]++;
zcb->zcb_embedded_histogram[BPE_GET_ETYPE(bp)]
[BPE_GET_PSIZE(bp)]++;
return;
}
/*
* The binning histogram bins by powers of two up to
* SPA_MAXBLOCKSIZE rather than creating bins for
* every possible blocksize found in the pool.
*/
int bin = highbit64(BP_GET_PSIZE(bp)) - 1;
zcb->zcb_psize_count[bin]++;
zcb->zcb_psize_len[bin] += BP_GET_PSIZE(bp);
zcb->zcb_psize_total += BP_GET_PSIZE(bp);
bin = highbit64(BP_GET_LSIZE(bp)) - 1;
zcb->zcb_lsize_count[bin]++;
zcb->zcb_lsize_len[bin] += BP_GET_LSIZE(bp);
zcb->zcb_lsize_total += BP_GET_LSIZE(bp);
bin = highbit64(BP_GET_ASIZE(bp)) - 1;
zcb->zcb_asize_count[bin]++;
zcb->zcb_asize_len[bin] += BP_GET_ASIZE(bp);
zcb->zcb_asize_total += BP_GET_ASIZE(bp);
if (dump_opt['L'])
return;
if (BP_GET_DEDUP(bp)) {
ddt_t *ddt;
ddt_entry_t *dde;
ddt = ddt_select(zcb->zcb_spa, bp);
ddt_enter(ddt);
dde = ddt_lookup(ddt, bp, B_FALSE);
if (dde == NULL) {
refcnt = 0;
} else {
ddt_phys_t *ddp = ddt_phys_select(dde, bp);
ddt_phys_decref(ddp);
refcnt = ddp->ddp_refcnt;
if (ddt_phys_total_refcnt(dde) == 0)
ddt_remove(ddt, dde);
}
ddt_exit(ddt);
}
VERIFY3U(zio_wait(zio_claim(NULL, zcb->zcb_spa,
refcnt ? 0 : spa_min_claim_txg(zcb->zcb_spa),
bp, NULL, NULL, ZIO_FLAG_CANFAIL)), ==, 0);
}
static void
zdb_blkptr_done(zio_t *zio)
{
spa_t *spa = zio->io_spa;
blkptr_t *bp = zio->io_bp;
int ioerr = zio->io_error;
zdb_cb_t *zcb = zio->io_private;
zbookmark_phys_t *zb = &zio->io_bookmark;
mutex_enter(&spa->spa_scrub_lock);
spa->spa_load_verify_bytes -= BP_GET_PSIZE(bp);
cv_broadcast(&spa->spa_scrub_io_cv);
if (ioerr && !(zio->io_flags & ZIO_FLAG_SPECULATIVE)) {
char blkbuf[BP_SPRINTF_LEN];
zcb->zcb_haderrors = 1;
zcb->zcb_errors[ioerr]++;
if (dump_opt['b'] >= 2)
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
else
blkbuf[0] = '\0';
(void) printf("zdb_blkptr_cb: "
"Got error %d reading "
"<%llu, %llu, %lld, %llx> %s -- skipping\n",
ioerr,
(u_longlong_t)zb->zb_objset,
(u_longlong_t)zb->zb_object,
(u_longlong_t)zb->zb_level,
(u_longlong_t)zb->zb_blkid,
blkbuf);
}
mutex_exit(&spa->spa_scrub_lock);
abd_free(zio->io_abd);
}
static int
zdb_blkptr_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
const zbookmark_phys_t *zb, const dnode_phys_t *dnp, void *arg)
{
zdb_cb_t *zcb = arg;
dmu_object_type_t type;
boolean_t is_metadata;
if (zb->zb_level == ZB_DNODE_LEVEL)
return (0);
if (dump_opt['b'] >= 5 && bp->blk_birth > 0) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
(void) printf("objset %llu object %llu "
"level %lld offset 0x%llx %s\n",
(u_longlong_t)zb->zb_objset,
(u_longlong_t)zb->zb_object,
(longlong_t)zb->zb_level,
(u_longlong_t)blkid2offset(dnp, bp, zb),
blkbuf);
}
if (BP_IS_HOLE(bp) || BP_IS_REDACTED(bp))
return (0);
type = BP_GET_TYPE(bp);
zdb_count_block(zcb, zilog, bp,
(type & DMU_OT_NEWTYPE) ? ZDB_OT_OTHER : type);
is_metadata = (BP_GET_LEVEL(bp) != 0 || DMU_OT_IS_METADATA(type));
if (!BP_IS_EMBEDDED(bp) &&
(dump_opt['c'] > 1 || (dump_opt['c'] && is_metadata))) {
size_t size = BP_GET_PSIZE(bp);
abd_t *abd = abd_alloc(size, B_FALSE);
int flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCRUB | ZIO_FLAG_RAW;
/* If it's an intent log block, failure is expected. */
if (zb->zb_level == ZB_ZIL_LEVEL)
flags |= ZIO_FLAG_SPECULATIVE;
mutex_enter(&spa->spa_scrub_lock);
while (spa->spa_load_verify_bytes > max_inflight_bytes)
cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock);
spa->spa_load_verify_bytes += size;
mutex_exit(&spa->spa_scrub_lock);
zio_nowait(zio_read(NULL, spa, bp, abd, size,
zdb_blkptr_done, zcb, ZIO_PRIORITY_ASYNC_READ, flags, zb));
}
zcb->zcb_readfails = 0;
/* only call gethrtime() every 100 blocks */
static int iters;
if (++iters > 100)
iters = 0;
else
return (0);
if (dump_opt['b'] < 5 && gethrtime() > zcb->zcb_lastprint + NANOSEC) {
uint64_t now = gethrtime();
char buf[10];
uint64_t bytes = zcb->zcb_type[ZB_TOTAL][ZDB_OT_TOTAL].zb_asize;
uint64_t kb_per_sec =
1 + bytes / (1 + ((now - zcb->zcb_start) / 1000 / 1000));
uint64_t sec_remaining =
(zcb->zcb_totalasize - bytes) / 1024 / kb_per_sec;
/* make sure nicenum has enough space */
_Static_assert(sizeof (buf) >= NN_NUMBUF_SZ, "buf truncated");
zfs_nicebytes(bytes, buf, sizeof (buf));
(void) fprintf(stderr,
"\r%5s completed (%4"PRIu64"MB/s) "
"estimated time remaining: "
"%"PRIu64"hr %02"PRIu64"min %02"PRIu64"sec ",
buf, kb_per_sec / 1024,
sec_remaining / 60 / 60,
sec_remaining / 60 % 60,
sec_remaining % 60);
zcb->zcb_lastprint = now;
}
return (0);
}
static void
zdb_leak(void *arg, uint64_t start, uint64_t size)
{
vdev_t *vd = arg;
(void) printf("leaked space: vdev %llu, offset 0x%llx, size %llu\n",
(u_longlong_t)vd->vdev_id, (u_longlong_t)start, (u_longlong_t)size);
}
static metaslab_ops_t zdb_metaslab_ops = {
NULL /* alloc */
};
static int
load_unflushed_svr_segs_cb(spa_t *spa, space_map_entry_t *sme,
uint64_t txg, void *arg)
{
spa_vdev_removal_t *svr = arg;
uint64_t offset = sme->sme_offset;
uint64_t size = sme->sme_run;
/* skip vdevs we don't care about */
if (sme->sme_vdev != svr->svr_vdev_id)
return (0);
vdev_t *vd = vdev_lookup_top(spa, sme->sme_vdev);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
ASSERT(sme->sme_type == SM_ALLOC || sme->sme_type == SM_FREE);
if (txg < metaslab_unflushed_txg(ms))
return (0);
if (sme->sme_type == SM_ALLOC)
range_tree_add(svr->svr_allocd_segs, offset, size);
else
range_tree_remove(svr->svr_allocd_segs, offset, size);
return (0);
}
static void
claim_segment_impl_cb(uint64_t inner_offset, vdev_t *vd, uint64_t offset,
uint64_t size, void *arg)
{
(void) inner_offset, (void) arg;
/*
* This callback was called through a remap from
* a device being removed. Therefore, the vdev that
* this callback is applied to is a concrete
* vdev.
*/
ASSERT(vdev_is_concrete(vd));
VERIFY0(metaslab_claim_impl(vd, offset, size,
spa_min_claim_txg(vd->vdev_spa)));
}
static void
claim_segment_cb(void *arg, uint64_t offset, uint64_t size)
{
vdev_t *vd = arg;
vdev_indirect_ops.vdev_op_remap(vd, offset, size,
claim_segment_impl_cb, NULL);
}
/*
* After accounting for all allocated blocks that are directly referenced,
* we might have missed a reference to a block from a partially complete
* (and thus unused) indirect mapping object. We perform a secondary pass
* through the metaslabs we have already mapped and claim the destination
* blocks.
*/
static void
zdb_claim_removing(spa_t *spa, zdb_cb_t *zcb)
{
if (dump_opt['L'])
return;
if (spa->spa_vdev_removal == NULL)
return;
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
spa_vdev_removal_t *svr = spa->spa_vdev_removal;
vdev_t *vd = vdev_lookup_top(spa, svr->svr_vdev_id);
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
ASSERT0(range_tree_space(svr->svr_allocd_segs));
range_tree_t *allocs = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
for (uint64_t msi = 0; msi < vd->vdev_ms_count; msi++) {
metaslab_t *msp = vd->vdev_ms[msi];
ASSERT0(range_tree_space(allocs));
if (msp->ms_sm != NULL)
VERIFY0(space_map_load(msp->ms_sm, allocs, SM_ALLOC));
range_tree_vacate(allocs, range_tree_add, svr->svr_allocd_segs);
}
range_tree_destroy(allocs);
iterate_through_spacemap_logs(spa, load_unflushed_svr_segs_cb, svr);
/*
* Clear everything past what has been synced,
* because we have not allocated mappings for
* it yet.
*/
range_tree_clear(svr->svr_allocd_segs,
vdev_indirect_mapping_max_offset(vim),
vd->vdev_asize - vdev_indirect_mapping_max_offset(vim));
zcb->zcb_removing_size += range_tree_space(svr->svr_allocd_segs);
range_tree_vacate(svr->svr_allocd_segs, claim_segment_cb, vd);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
static int
increment_indirect_mapping_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
(void) tx;
zdb_cb_t *zcb = arg;
spa_t *spa = zcb->zcb_spa;
vdev_t *vd;
const dva_t *dva = &bp->blk_dva[0];
ASSERT(!bp_freed);
ASSERT(!dump_opt['L']);
ASSERT3U(BP_GET_NDVAS(bp), ==, 1);
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
vd = vdev_lookup_top(zcb->zcb_spa, DVA_GET_VDEV(dva));
ASSERT3P(vd, !=, NULL);
spa_config_exit(spa, SCL_VDEV, FTAG);
ASSERT(vd->vdev_indirect_config.vic_mapping_object != 0);
ASSERT3P(zcb->zcb_vd_obsolete_counts[vd->vdev_id], !=, NULL);
vdev_indirect_mapping_increment_obsolete_count(
vd->vdev_indirect_mapping,
DVA_GET_OFFSET(dva), DVA_GET_ASIZE(dva),
zcb->zcb_vd_obsolete_counts[vd->vdev_id]);
return (0);
}
static uint32_t *
zdb_load_obsolete_counts(vdev_t *vd)
{
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
spa_t *spa = vd->vdev_spa;
spa_condensing_indirect_phys_t *scip =
&spa->spa_condensing_indirect_phys;
uint64_t obsolete_sm_object;
uint32_t *counts;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
EQUIV(obsolete_sm_object != 0, vd->vdev_obsolete_sm != NULL);
counts = vdev_indirect_mapping_load_obsolete_counts(vim);
if (vd->vdev_obsolete_sm != NULL) {
vdev_indirect_mapping_load_obsolete_spacemap(vim, counts,
vd->vdev_obsolete_sm);
}
if (scip->scip_vdev == vd->vdev_id &&
scip->scip_prev_obsolete_sm_object != 0) {
space_map_t *prev_obsolete_sm = NULL;
VERIFY0(space_map_open(&prev_obsolete_sm, spa->spa_meta_objset,
scip->scip_prev_obsolete_sm_object, 0, vd->vdev_asize, 0));
vdev_indirect_mapping_load_obsolete_spacemap(vim, counts,
prev_obsolete_sm);
space_map_close(prev_obsolete_sm);
}
return (counts);
}
static void
zdb_ddt_leak_init(spa_t *spa, zdb_cb_t *zcb)
{
ddt_bookmark_t ddb = {0};
ddt_entry_t dde;
int error;
int p;
ASSERT(!dump_opt['L']);
while ((error = ddt_walk(spa, &ddb, &dde)) == 0) {
blkptr_t blk;
ddt_phys_t *ddp = dde.dde_phys;
if (ddb.ddb_class == DDT_CLASS_UNIQUE)
return;
ASSERT(ddt_phys_total_refcnt(&dde) > 1);
for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++) {
if (ddp->ddp_phys_birth == 0)
continue;
ddt_bp_create(ddb.ddb_checksum,
&dde.dde_key, ddp, &blk);
if (p == DDT_PHYS_DITTO) {
zdb_count_block(zcb, NULL, &blk, ZDB_OT_DITTO);
} else {
zcb->zcb_dedup_asize +=
BP_GET_ASIZE(&blk) * (ddp->ddp_refcnt - 1);
zcb->zcb_dedup_blocks++;
}
}
ddt_t *ddt = spa->spa_ddt[ddb.ddb_checksum];
ddt_enter(ddt);
VERIFY(ddt_lookup(ddt, &blk, B_TRUE) != NULL);
ddt_exit(ddt);
}
ASSERT(error == ENOENT);
}
typedef struct checkpoint_sm_exclude_entry_arg {
vdev_t *cseea_vd;
uint64_t cseea_checkpoint_size;
} checkpoint_sm_exclude_entry_arg_t;
static int
checkpoint_sm_exclude_entry_cb(space_map_entry_t *sme, void *arg)
{
checkpoint_sm_exclude_entry_arg_t *cseea = arg;
vdev_t *vd = cseea->cseea_vd;
metaslab_t *ms = vd->vdev_ms[sme->sme_offset >> vd->vdev_ms_shift];
uint64_t end = sme->sme_offset + sme->sme_run;
ASSERT(sme->sme_type == SM_FREE);
/*
* Since the vdev_checkpoint_sm exists in the vdev level
* and the ms_sm space maps exist in the metaslab level,
* an entry in the checkpoint space map could theoretically
* cross the boundaries of the metaslab that it belongs.
*
* In reality, because of the way that we populate and
* manipulate the checkpoint's space maps currently,
* there shouldn't be any entries that cross metaslabs.
* Hence the assertion below.
*
* That said, there is no fundamental requirement that
* the checkpoint's space map entries should not cross
* metaslab boundaries. So if needed we could add code
* that handles metaslab-crossing segments in the future.
*/
VERIFY3U(sme->sme_offset, >=, ms->ms_start);
VERIFY3U(end, <=, ms->ms_start + ms->ms_size);
/*
* By removing the entry from the allocated segments we
* also verify that the entry is there to begin with.
*/
mutex_enter(&ms->ms_lock);
range_tree_remove(ms->ms_allocatable, sme->sme_offset, sme->sme_run);
mutex_exit(&ms->ms_lock);
cseea->cseea_checkpoint_size += sme->sme_run;
return (0);
}
static void
zdb_leak_init_vdev_exclude_checkpoint(vdev_t *vd, zdb_cb_t *zcb)
{
spa_t *spa = vd->vdev_spa;
space_map_t *checkpoint_sm = NULL;
uint64_t checkpoint_sm_obj;
/*
* If there is no vdev_top_zap, we are in a pool whose
* version predates the pool checkpoint feature.
*/
if (vd->vdev_top_zap == 0)
return;
/*
* If there is no reference of the vdev_checkpoint_sm in
* the vdev_top_zap, then one of the following scenarios
* is true:
*
* 1] There is no checkpoint
* 2] There is a checkpoint, but no checkpointed blocks
* have been freed yet
* 3] The current vdev is indirect
*
* In these cases we return immediately.
*/
if (zap_contains(spa_meta_objset(spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM) != 0)
return;
VERIFY0(zap_lookup(spa_meta_objset(spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM, sizeof (uint64_t), 1,
&checkpoint_sm_obj));
checkpoint_sm_exclude_entry_arg_t cseea;
cseea.cseea_vd = vd;
cseea.cseea_checkpoint_size = 0;
VERIFY0(space_map_open(&checkpoint_sm, spa_meta_objset(spa),
checkpoint_sm_obj, 0, vd->vdev_asize, vd->vdev_ashift));
VERIFY0(space_map_iterate(checkpoint_sm,
space_map_length(checkpoint_sm),
checkpoint_sm_exclude_entry_cb, &cseea));
space_map_close(checkpoint_sm);
zcb->zcb_checkpoint_size += cseea.cseea_checkpoint_size;
}
static void
zdb_leak_init_exclude_checkpoint(spa_t *spa, zdb_cb_t *zcb)
{
ASSERT(!dump_opt['L']);
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
ASSERT3U(c, ==, rvd->vdev_child[c]->vdev_id);
zdb_leak_init_vdev_exclude_checkpoint(rvd->vdev_child[c], zcb);
}
}
static int
count_unflushed_space_cb(spa_t *spa, space_map_entry_t *sme,
uint64_t txg, void *arg)
{
int64_t *ualloc_space = arg;
uint64_t offset = sme->sme_offset;
uint64_t vdev_id = sme->sme_vdev;
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
if (!vdev_is_concrete(vd))
return (0);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
ASSERT(sme->sme_type == SM_ALLOC || sme->sme_type == SM_FREE);
if (txg < metaslab_unflushed_txg(ms))
return (0);
if (sme->sme_type == SM_ALLOC)
*ualloc_space += sme->sme_run;
else
*ualloc_space -= sme->sme_run;
return (0);
}
static int64_t
get_unflushed_alloc_space(spa_t *spa)
{
if (dump_opt['L'])
return (0);
int64_t ualloc_space = 0;
iterate_through_spacemap_logs(spa, count_unflushed_space_cb,
&ualloc_space);
return (ualloc_space);
}
static int
load_unflushed_cb(spa_t *spa, space_map_entry_t *sme, uint64_t txg, void *arg)
{
maptype_t *uic_maptype = arg;
uint64_t offset = sme->sme_offset;
uint64_t size = sme->sme_run;
uint64_t vdev_id = sme->sme_vdev;
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
/* skip indirect vdevs */
if (!vdev_is_concrete(vd))
return (0);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
ASSERT(sme->sme_type == SM_ALLOC || sme->sme_type == SM_FREE);
ASSERT(*uic_maptype == SM_ALLOC || *uic_maptype == SM_FREE);
if (txg < metaslab_unflushed_txg(ms))
return (0);
if (*uic_maptype == sme->sme_type)
range_tree_add(ms->ms_allocatable, offset, size);
else
range_tree_remove(ms->ms_allocatable, offset, size);
return (0);
}
static void
load_unflushed_to_ms_allocatables(spa_t *spa, maptype_t maptype)
{
iterate_through_spacemap_logs(spa, load_unflushed_cb, &maptype);
}
static void
load_concrete_ms_allocatable_trees(spa_t *spa, maptype_t maptype)
{
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t i = 0; i < rvd->vdev_children; i++) {
vdev_t *vd = rvd->vdev_child[i];
ASSERT3U(i, ==, vd->vdev_id);
if (vd->vdev_ops == &vdev_indirect_ops)
continue;
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
metaslab_t *msp = vd->vdev_ms[m];
(void) fprintf(stderr,
"\rloading concrete vdev %llu, "
"metaslab %llu of %llu ...",
(longlong_t)vd->vdev_id,
(longlong_t)msp->ms_id,
(longlong_t)vd->vdev_ms_count);
mutex_enter(&msp->ms_lock);
range_tree_vacate(msp->ms_allocatable, NULL, NULL);
/*
* We don't want to spend the CPU manipulating the
* size-ordered tree, so clear the range_tree ops.
*/
msp->ms_allocatable->rt_ops = NULL;
if (msp->ms_sm != NULL) {
VERIFY0(space_map_load(msp->ms_sm,
msp->ms_allocatable, maptype));
}
if (!msp->ms_loaded)
msp->ms_loaded = B_TRUE;
mutex_exit(&msp->ms_lock);
}
}
load_unflushed_to_ms_allocatables(spa, maptype);
}
/*
* vm_idxp is an in-out parameter which (for indirect vdevs) is the
* index in vim_entries that has the first entry in this metaslab.
* On return, it will be set to the first entry after this metaslab.
*/
static void
load_indirect_ms_allocatable_tree(vdev_t *vd, metaslab_t *msp,
uint64_t *vim_idxp)
{
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
mutex_enter(&msp->ms_lock);
range_tree_vacate(msp->ms_allocatable, NULL, NULL);
/*
* We don't want to spend the CPU manipulating the
* size-ordered tree, so clear the range_tree ops.
*/
msp->ms_allocatable->rt_ops = NULL;
for (; *vim_idxp < vdev_indirect_mapping_num_entries(vim);
(*vim_idxp)++) {
vdev_indirect_mapping_entry_phys_t *vimep =
&vim->vim_entries[*vim_idxp];
uint64_t ent_offset = DVA_MAPPING_GET_SRC_OFFSET(vimep);
uint64_t ent_len = DVA_GET_ASIZE(&vimep->vimep_dst);
ASSERT3U(ent_offset, >=, msp->ms_start);
if (ent_offset >= msp->ms_start + msp->ms_size)
break;
/*
* Mappings do not cross metaslab boundaries,
* because we create them by walking the metaslabs.
*/
ASSERT3U(ent_offset + ent_len, <=,
msp->ms_start + msp->ms_size);
range_tree_add(msp->ms_allocatable, ent_offset, ent_len);
}
if (!msp->ms_loaded)
msp->ms_loaded = B_TRUE;
mutex_exit(&msp->ms_lock);
}
static void
zdb_leak_init_prepare_indirect_vdevs(spa_t *spa, zdb_cb_t *zcb)
{
ASSERT(!dump_opt['L']);
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
ASSERT3U(c, ==, vd->vdev_id);
if (vd->vdev_ops != &vdev_indirect_ops)
continue;
/*
* Note: we don't check for mapping leaks on
* removing vdevs because their ms_allocatable's
* are used to look for leaks in allocated space.
*/
zcb->zcb_vd_obsolete_counts[c] = zdb_load_obsolete_counts(vd);
/*
* Normally, indirect vdevs don't have any
* metaslabs. We want to set them up for
* zio_claim().
*/
vdev_metaslab_group_create(vd);
VERIFY0(vdev_metaslab_init(vd, 0));
vdev_indirect_mapping_t *vim __maybe_unused =
vd->vdev_indirect_mapping;
uint64_t vim_idx = 0;
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
(void) fprintf(stderr,
"\rloading indirect vdev %llu, "
"metaslab %llu of %llu ...",
(longlong_t)vd->vdev_id,
(longlong_t)vd->vdev_ms[m]->ms_id,
(longlong_t)vd->vdev_ms_count);
load_indirect_ms_allocatable_tree(vd, vd->vdev_ms[m],
&vim_idx);
}
ASSERT3U(vim_idx, ==, vdev_indirect_mapping_num_entries(vim));
}
}
static void
zdb_leak_init(spa_t *spa, zdb_cb_t *zcb)
{
zcb->zcb_spa = spa;
if (dump_opt['L'])
return;
dsl_pool_t *dp = spa->spa_dsl_pool;
vdev_t *rvd = spa->spa_root_vdev;
/*
* We are going to be changing the meaning of the metaslab's
* ms_allocatable. Ensure that the allocator doesn't try to
* use the tree.
*/
spa->spa_normal_class->mc_ops = &zdb_metaslab_ops;
spa->spa_log_class->mc_ops = &zdb_metaslab_ops;
spa->spa_embedded_log_class->mc_ops = &zdb_metaslab_ops;
zcb->zcb_vd_obsolete_counts =
umem_zalloc(rvd->vdev_children * sizeof (uint32_t *),
UMEM_NOFAIL);
/*
* For leak detection, we overload the ms_allocatable trees
* to contain allocated segments instead of free segments.
* As a result, we can't use the normal metaslab_load/unload
* interfaces.
*/
zdb_leak_init_prepare_indirect_vdevs(spa, zcb);
load_concrete_ms_allocatable_trees(spa, SM_ALLOC);
/*
* On load_concrete_ms_allocatable_trees() we loaded all the
* allocated entries from the ms_sm to the ms_allocatable for
* each metaslab. If the pool has a checkpoint or is in the
* middle of discarding a checkpoint, some of these blocks
* may have been freed but their ms_sm may not have been
* updated because they are referenced by the checkpoint. In
* order to avoid false-positives during leak-detection, we
* go through the vdev's checkpoint space map and exclude all
* its entries from their relevant ms_allocatable.
*
* We also aggregate the space held by the checkpoint and add
* it to zcb_checkpoint_size.
*
* Note that at this point we are also verifying that all the
* entries on the checkpoint_sm are marked as allocated in
* the ms_sm of their relevant metaslab.
* [see comment in checkpoint_sm_exclude_entry_cb()]
*/
zdb_leak_init_exclude_checkpoint(spa, zcb);
ASSERT3U(zcb->zcb_checkpoint_size, ==, spa_get_checkpoint_space(spa));
/* for cleaner progress output */
(void) fprintf(stderr, "\n");
if (bpobj_is_open(&dp->dp_obsolete_bpobj)) {
ASSERT(spa_feature_is_enabled(spa,
SPA_FEATURE_DEVICE_REMOVAL));
(void) bpobj_iterate_nofree(&dp->dp_obsolete_bpobj,
increment_indirect_mapping_cb, zcb, NULL);
}
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
zdb_ddt_leak_init(spa, zcb);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
static boolean_t
zdb_check_for_obsolete_leaks(vdev_t *vd, zdb_cb_t *zcb)
{
boolean_t leaks = B_FALSE;
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
uint64_t total_leaked = 0;
boolean_t are_precise = B_FALSE;
ASSERT(vim != NULL);
for (uint64_t i = 0; i < vdev_indirect_mapping_num_entries(vim); i++) {
vdev_indirect_mapping_entry_phys_t *vimep =
&vim->vim_entries[i];
uint64_t obsolete_bytes = 0;
uint64_t offset = DVA_MAPPING_GET_SRC_OFFSET(vimep);
metaslab_t *msp = vd->vdev_ms[offset >> vd->vdev_ms_shift];
/*
* This is not very efficient but it's easy to
* verify correctness.
*/
for (uint64_t inner_offset = 0;
inner_offset < DVA_GET_ASIZE(&vimep->vimep_dst);
inner_offset += 1ULL << vd->vdev_ashift) {
if (range_tree_contains(msp->ms_allocatable,
offset + inner_offset, 1ULL << vd->vdev_ashift)) {
obsolete_bytes += 1ULL << vd->vdev_ashift;
}
}
int64_t bytes_leaked = obsolete_bytes -
zcb->zcb_vd_obsolete_counts[vd->vdev_id][i];
ASSERT3U(DVA_GET_ASIZE(&vimep->vimep_dst), >=,
zcb->zcb_vd_obsolete_counts[vd->vdev_id][i]);
VERIFY0(vdev_obsolete_counts_are_precise(vd, &are_precise));
if (bytes_leaked != 0 && (are_precise || dump_opt['d'] >= 5)) {
(void) printf("obsolete indirect mapping count "
"mismatch on %llu:%llx:%llx : %llx bytes leaked\n",
(u_longlong_t)vd->vdev_id,
(u_longlong_t)DVA_MAPPING_GET_SRC_OFFSET(vimep),
(u_longlong_t)DVA_GET_ASIZE(&vimep->vimep_dst),
(u_longlong_t)bytes_leaked);
}
total_leaked += ABS(bytes_leaked);
}
VERIFY0(vdev_obsolete_counts_are_precise(vd, &are_precise));
if (!are_precise && total_leaked > 0) {
int pct_leaked = total_leaked * 100 /
vdev_indirect_mapping_bytes_mapped(vim);
(void) printf("cannot verify obsolete indirect mapping "
"counts of vdev %llu because precise feature was not "
"enabled when it was removed: %d%% (%llx bytes) of mapping"
"unreferenced\n",
(u_longlong_t)vd->vdev_id, pct_leaked,
(u_longlong_t)total_leaked);
} else if (total_leaked > 0) {
(void) printf("obsolete indirect mapping count mismatch "
"for vdev %llu -- %llx total bytes mismatched\n",
(u_longlong_t)vd->vdev_id,
(u_longlong_t)total_leaked);
leaks |= B_TRUE;
}
vdev_indirect_mapping_free_obsolete_counts(vim,
zcb->zcb_vd_obsolete_counts[vd->vdev_id]);
zcb->zcb_vd_obsolete_counts[vd->vdev_id] = NULL;
return (leaks);
}
static boolean_t
zdb_leak_fini(spa_t *spa, zdb_cb_t *zcb)
{
if (dump_opt['L'])
return (B_FALSE);
boolean_t leaks = B_FALSE;
vdev_t *rvd = spa->spa_root_vdev;
for (unsigned c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
if (zcb->zcb_vd_obsolete_counts[c] != NULL) {
leaks |= zdb_check_for_obsolete_leaks(vd, zcb);
}
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
metaslab_t *msp = vd->vdev_ms[m];
ASSERT3P(msp->ms_group, ==, (msp->ms_group->mg_class ==
spa_embedded_log_class(spa)) ?
vd->vdev_log_mg : vd->vdev_mg);
/*
* ms_allocatable has been overloaded
* to contain allocated segments. Now that
* we finished traversing all blocks, any
* block that remains in the ms_allocatable
* represents an allocated block that we
* did not claim during the traversal.
* Claimed blocks would have been removed
* from the ms_allocatable. For indirect
* vdevs, space remaining in the tree
* represents parts of the mapping that are
* not referenced, which is not a bug.
*/
if (vd->vdev_ops == &vdev_indirect_ops) {
range_tree_vacate(msp->ms_allocatable,
NULL, NULL);
} else {
range_tree_vacate(msp->ms_allocatable,
zdb_leak, vd);
}
if (msp->ms_loaded) {
msp->ms_loaded = B_FALSE;
}
}
}
umem_free(zcb->zcb_vd_obsolete_counts,
rvd->vdev_children * sizeof (uint32_t *));
zcb->zcb_vd_obsolete_counts = NULL;
return (leaks);
}
static int
count_block_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
(void) tx;
zdb_cb_t *zcb = arg;
if (dump_opt['b'] >= 5) {
char blkbuf[BP_SPRINTF_LEN];
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
(void) printf("[%s] %s\n",
"deferred free", blkbuf);
}
zdb_count_block(zcb, NULL, bp, ZDB_OT_DEFERRED);
return (0);
}
/*
* Iterate over livelists which have been destroyed by the user but
* are still present in the MOS, waiting to be freed
*/
static void
iterate_deleted_livelists(spa_t *spa, ll_iter_t func, void *arg)
{
objset_t *mos = spa->spa_meta_objset;
uint64_t zap_obj;
int err = zap_lookup(mos, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_DELETED_CLONES, sizeof (uint64_t), 1, &zap_obj);
if (err == ENOENT)
return;
ASSERT0(err);
zap_cursor_t zc;
zap_attribute_t attr;
dsl_deadlist_t ll;
/* NULL out os prior to dsl_deadlist_open in case it's garbage */
ll.dl_os = NULL;
for (zap_cursor_init(&zc, mos, zap_obj);
zap_cursor_retrieve(&zc, &attr) == 0;
(void) zap_cursor_advance(&zc)) {
dsl_deadlist_open(&ll, mos, attr.za_first_integer);
func(&ll, arg);
dsl_deadlist_close(&ll);
}
zap_cursor_fini(&zc);
}
static int
bpobj_count_block_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
ASSERT(!bp_freed);
return (count_block_cb(arg, bp, tx));
}
static int
livelist_entry_count_blocks_cb(void *args, dsl_deadlist_entry_t *dle)
{
zdb_cb_t *zbc = args;
bplist_t blks;
bplist_create(&blks);
/* determine which blocks have been alloc'd but not freed */
VERIFY0(dsl_process_sub_livelist(&dle->dle_bpobj, &blks, NULL, NULL));
/* count those blocks */
(void) bplist_iterate(&blks, count_block_cb, zbc, NULL);
bplist_destroy(&blks);
return (0);
}
static void
livelist_count_blocks(dsl_deadlist_t *ll, void *arg)
{
dsl_deadlist_iterate(ll, livelist_entry_count_blocks_cb, arg);
}
/*
* Count the blocks in the livelists that have been destroyed by the user
* but haven't yet been freed.
*/
static void
deleted_livelists_count_blocks(spa_t *spa, zdb_cb_t *zbc)
{
iterate_deleted_livelists(spa, livelist_count_blocks, zbc);
}
static void
dump_livelist_cb(dsl_deadlist_t *ll, void *arg)
{
ASSERT3P(arg, ==, NULL);
global_feature_count[SPA_FEATURE_LIVELIST]++;
dump_blkptr_list(ll, "Deleted Livelist");
dsl_deadlist_iterate(ll, sublivelist_verify_lightweight, NULL);
}
/*
* Print out, register object references to, and increment feature counts for
* livelists that have been destroyed by the user but haven't yet been freed.
*/
static void
deleted_livelists_dump_mos(spa_t *spa)
{
uint64_t zap_obj;
objset_t *mos = spa->spa_meta_objset;
int err = zap_lookup(mos, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_DELETED_CLONES, sizeof (uint64_t), 1, &zap_obj);
if (err == ENOENT)
return;
mos_obj_refd(zap_obj);
iterate_deleted_livelists(spa, dump_livelist_cb, NULL);
}
static int
dump_block_stats(spa_t *spa)
{
zdb_cb_t *zcb;
zdb_blkstats_t *zb, *tzb;
uint64_t norm_alloc, norm_space, total_alloc, total_found;
int flags = TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA |
TRAVERSE_NO_DECRYPT | TRAVERSE_HARD;
boolean_t leaks = B_FALSE;
int e, c, err;
bp_embedded_type_t i;
zcb = umem_zalloc(sizeof (zdb_cb_t), UMEM_NOFAIL);
(void) printf("\nTraversing all blocks %s%s%s%s%s...\n\n",
(dump_opt['c'] || !dump_opt['L']) ? "to verify " : "",
(dump_opt['c'] == 1) ? "metadata " : "",
dump_opt['c'] ? "checksums " : "",
(dump_opt['c'] && !dump_opt['L']) ? "and verify " : "",
!dump_opt['L'] ? "nothing leaked " : "");
/*
* When leak detection is enabled we load all space maps as SM_ALLOC
* maps, then traverse the pool claiming each block we discover. If
* the pool is perfectly consistent, the segment trees will be empty
* when we're done. Anything left over is a leak; any block we can't
* claim (because it's not part of any space map) is a double
* allocation, reference to a freed block, or an unclaimed log block.
*
* When leak detection is disabled (-L option) we still traverse the
* pool claiming each block we discover, but we skip opening any space
* maps.
*/
zdb_leak_init(spa, zcb);
/*
* If there's a deferred-free bplist, process that first.
*/
(void) bpobj_iterate_nofree(&spa->spa_deferred_bpobj,
bpobj_count_block_cb, zcb, NULL);
if (spa_version(spa) >= SPA_VERSION_DEADLISTS) {
(void) bpobj_iterate_nofree(&spa->spa_dsl_pool->dp_free_bpobj,
bpobj_count_block_cb, zcb, NULL);
}
zdb_claim_removing(spa, zcb);
if (spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY)) {
VERIFY3U(0, ==, bptree_iterate(spa->spa_meta_objset,
spa->spa_dsl_pool->dp_bptree_obj, B_FALSE, count_block_cb,
zcb, NULL));
}
deleted_livelists_count_blocks(spa, zcb);
if (dump_opt['c'] > 1)
flags |= TRAVERSE_PREFETCH_DATA;
zcb->zcb_totalasize = metaslab_class_get_alloc(spa_normal_class(spa));
zcb->zcb_totalasize += metaslab_class_get_alloc(spa_special_class(spa));
zcb->zcb_totalasize += metaslab_class_get_alloc(spa_dedup_class(spa));
zcb->zcb_totalasize +=
metaslab_class_get_alloc(spa_embedded_log_class(spa));
zcb->zcb_start = zcb->zcb_lastprint = gethrtime();
err = traverse_pool(spa, 0, flags, zdb_blkptr_cb, zcb);
/*
* If we've traversed the data blocks then we need to wait for those
* I/Os to complete. We leverage "The Godfather" zio to wait on
* all async I/Os to complete.
*/
if (dump_opt['c']) {
for (c = 0; c < max_ncpus; c++) {
(void) zio_wait(spa->spa_async_zio_root[c]);
spa->spa_async_zio_root[c] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE |
ZIO_FLAG_GODFATHER);
}
}
ASSERT0(spa->spa_load_verify_bytes);
/*
* Done after zio_wait() since zcb_haderrors is modified in
* zdb_blkptr_done()
*/
zcb->zcb_haderrors |= err;
if (zcb->zcb_haderrors) {
(void) printf("\nError counts:\n\n");
(void) printf("\t%5s %s\n", "errno", "count");
for (e = 0; e < 256; e++) {
if (zcb->zcb_errors[e] != 0) {
(void) printf("\t%5d %llu\n",
e, (u_longlong_t)zcb->zcb_errors[e]);
}
}
}
/*
* Report any leaked segments.
*/
leaks |= zdb_leak_fini(spa, zcb);
tzb = &zcb->zcb_type[ZB_TOTAL][ZDB_OT_TOTAL];
norm_alloc = metaslab_class_get_alloc(spa_normal_class(spa));
norm_space = metaslab_class_get_space(spa_normal_class(spa));
total_alloc = norm_alloc +
metaslab_class_get_alloc(spa_log_class(spa)) +
metaslab_class_get_alloc(spa_embedded_log_class(spa)) +
metaslab_class_get_alloc(spa_special_class(spa)) +
metaslab_class_get_alloc(spa_dedup_class(spa)) +
get_unflushed_alloc_space(spa);
total_found = tzb->zb_asize - zcb->zcb_dedup_asize +
zcb->zcb_removing_size + zcb->zcb_checkpoint_size;
if (total_found == total_alloc && !dump_opt['L']) {
(void) printf("\n\tNo leaks (block sum matches space"
" maps exactly)\n");
} else if (!dump_opt['L']) {
(void) printf("block traversal size %llu != alloc %llu "
"(%s %lld)\n",
(u_longlong_t)total_found,
(u_longlong_t)total_alloc,
(dump_opt['L']) ? "unreachable" : "leaked",
(longlong_t)(total_alloc - total_found));
leaks = B_TRUE;
}
if (tzb->zb_count == 0) {
umem_free(zcb, sizeof (zdb_cb_t));
return (2);
}
(void) printf("\n");
(void) printf("\t%-16s %14llu\n", "bp count:",
(u_longlong_t)tzb->zb_count);
(void) printf("\t%-16s %14llu\n", "ganged count:",
(longlong_t)tzb->zb_gangs);
(void) printf("\t%-16s %14llu avg: %6llu\n", "bp logical:",
(u_longlong_t)tzb->zb_lsize,
(u_longlong_t)(tzb->zb_lsize / tzb->zb_count));
(void) printf("\t%-16s %14llu avg: %6llu compression: %6.2f\n",
"bp physical:", (u_longlong_t)tzb->zb_psize,
(u_longlong_t)(tzb->zb_psize / tzb->zb_count),
(double)tzb->zb_lsize / tzb->zb_psize);
(void) printf("\t%-16s %14llu avg: %6llu compression: %6.2f\n",
"bp allocated:", (u_longlong_t)tzb->zb_asize,
(u_longlong_t)(tzb->zb_asize / tzb->zb_count),
(double)tzb->zb_lsize / tzb->zb_asize);
(void) printf("\t%-16s %14llu ref>1: %6llu deduplication: %6.2f\n",
"bp deduped:", (u_longlong_t)zcb->zcb_dedup_asize,
(u_longlong_t)zcb->zcb_dedup_blocks,
(double)zcb->zcb_dedup_asize / tzb->zb_asize + 1.0);
(void) printf("\t%-16s %14llu used: %5.2f%%\n", "Normal class:",
(u_longlong_t)norm_alloc, 100.0 * norm_alloc / norm_space);
if (spa_special_class(spa)->mc_allocator[0].mca_rotor != NULL) {
uint64_t alloc = metaslab_class_get_alloc(
spa_special_class(spa));
uint64_t space = metaslab_class_get_space(
spa_special_class(spa));
(void) printf("\t%-16s %14llu used: %5.2f%%\n",
"Special class", (u_longlong_t)alloc,
100.0 * alloc / space);
}
if (spa_dedup_class(spa)->mc_allocator[0].mca_rotor != NULL) {
uint64_t alloc = metaslab_class_get_alloc(
spa_dedup_class(spa));
uint64_t space = metaslab_class_get_space(
spa_dedup_class(spa));
(void) printf("\t%-16s %14llu used: %5.2f%%\n",
"Dedup class", (u_longlong_t)alloc,
100.0 * alloc / space);
}
if (spa_embedded_log_class(spa)->mc_allocator[0].mca_rotor != NULL) {
uint64_t alloc = metaslab_class_get_alloc(
spa_embedded_log_class(spa));
uint64_t space = metaslab_class_get_space(
spa_embedded_log_class(spa));
(void) printf("\t%-16s %14llu used: %5.2f%%\n",
"Embedded log class", (u_longlong_t)alloc,
100.0 * alloc / space);
}
for (i = 0; i < NUM_BP_EMBEDDED_TYPES; i++) {
if (zcb->zcb_embedded_blocks[i] == 0)
continue;
(void) printf("\n");
(void) printf("\tadditional, non-pointer bps of type %u: "
"%10llu\n",
i, (u_longlong_t)zcb->zcb_embedded_blocks[i]);
if (dump_opt['b'] >= 3) {
(void) printf("\t number of (compressed) bytes: "
"number of bps\n");
dump_histogram(zcb->zcb_embedded_histogram[i],
sizeof (zcb->zcb_embedded_histogram[i]) /
sizeof (zcb->zcb_embedded_histogram[i][0]), 0);
}
}
if (tzb->zb_ditto_samevdev != 0) {
(void) printf("\tDittoed blocks on same vdev: %llu\n",
(longlong_t)tzb->zb_ditto_samevdev);
}
if (tzb->zb_ditto_same_ms != 0) {
(void) printf("\tDittoed blocks in same metaslab: %llu\n",
(longlong_t)tzb->zb_ditto_same_ms);
}
for (uint64_t v = 0; v < spa->spa_root_vdev->vdev_children; v++) {
vdev_t *vd = spa->spa_root_vdev->vdev_child[v];
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
if (vim == NULL) {
continue;
}
char mem[32];
zdb_nicenum(vdev_indirect_mapping_num_entries(vim),
mem, vdev_indirect_mapping_size(vim));
(void) printf("\tindirect vdev id %llu has %llu segments "
"(%s in memory)\n",
(longlong_t)vd->vdev_id,
(longlong_t)vdev_indirect_mapping_num_entries(vim), mem);
}
if (dump_opt['b'] >= 2) {
int l, t, level;
+ char csize[32], lsize[32], psize[32], asize[32];
+ char avg[32], gang[32];
(void) printf("\nBlocks\tLSIZE\tPSIZE\tASIZE"
"\t avg\t comp\t%%Total\tType\n");
+ zfs_blkstat_t *mdstats = umem_zalloc(sizeof (zfs_blkstat_t),
+ UMEM_NOFAIL);
+
for (t = 0; t <= ZDB_OT_TOTAL; t++) {
- char csize[32], lsize[32], psize[32], asize[32];
- char avg[32], gang[32];
const char *typename;
/* make sure nicenum has enough space */
_Static_assert(sizeof (csize) >= NN_NUMBUF_SZ,
"csize truncated");
_Static_assert(sizeof (lsize) >= NN_NUMBUF_SZ,
"lsize truncated");
_Static_assert(sizeof (psize) >= NN_NUMBUF_SZ,
"psize truncated");
_Static_assert(sizeof (asize) >= NN_NUMBUF_SZ,
"asize truncated");
_Static_assert(sizeof (avg) >= NN_NUMBUF_SZ,
"avg truncated");
_Static_assert(sizeof (gang) >= NN_NUMBUF_SZ,
"gang truncated");
if (t < DMU_OT_NUMTYPES)
typename = dmu_ot[t].ot_name;
else
typename = zdb_ot_extname[t - DMU_OT_NUMTYPES];
if (zcb->zcb_type[ZB_TOTAL][t].zb_asize == 0) {
(void) printf("%6s\t%5s\t%5s\t%5s"
"\t%5s\t%5s\t%6s\t%s\n",
"-",
"-",
"-",
"-",
"-",
"-",
"-",
typename);
continue;
}
for (l = ZB_TOTAL - 1; l >= -1; l--) {
level = (l == -1 ? ZB_TOTAL : l);
zb = &zcb->zcb_type[level][t];
if (zb->zb_asize == 0)
continue;
+ if (level != ZB_TOTAL && t < DMU_OT_NUMTYPES &&
+ (level > 0 || DMU_OT_IS_METADATA(t))) {
+ mdstats->zb_count += zb->zb_count;
+ mdstats->zb_lsize += zb->zb_lsize;
+ mdstats->zb_psize += zb->zb_psize;
+ mdstats->zb_asize += zb->zb_asize;
+ mdstats->zb_gangs += zb->zb_gangs;
+ }
+
if (dump_opt['b'] < 3 && level != ZB_TOTAL)
continue;
if (level == 0 && zb->zb_asize ==
zcb->zcb_type[ZB_TOTAL][t].zb_asize)
continue;
zdb_nicenum(zb->zb_count, csize,
sizeof (csize));
zdb_nicenum(zb->zb_lsize, lsize,
sizeof (lsize));
zdb_nicenum(zb->zb_psize, psize,
sizeof (psize));
zdb_nicenum(zb->zb_asize, asize,
sizeof (asize));
zdb_nicenum(zb->zb_asize / zb->zb_count, avg,
sizeof (avg));
zdb_nicenum(zb->zb_gangs, gang, sizeof (gang));
(void) printf("%6s\t%5s\t%5s\t%5s\t%5s"
"\t%5.2f\t%6.2f\t",
csize, lsize, psize, asize, avg,
(double)zb->zb_lsize / zb->zb_psize,
100.0 * zb->zb_asize / tzb->zb_asize);
if (level == ZB_TOTAL)
(void) printf("%s\n", typename);
else
(void) printf(" L%d %s\n",
level, typename);
if (dump_opt['b'] >= 3 && zb->zb_gangs > 0) {
(void) printf("\t number of ganged "
"blocks: %s\n", gang);
}
if (dump_opt['b'] >= 4) {
(void) printf("psize "
"(in 512-byte sectors): "
"number of blocks\n");
dump_histogram(zb->zb_psize_histogram,
PSIZE_HISTO_SIZE, 0);
}
}
}
+ zdb_nicenum(mdstats->zb_count, csize,
+ sizeof (csize));
+ zdb_nicenum(mdstats->zb_lsize, lsize,
+ sizeof (lsize));
+ zdb_nicenum(mdstats->zb_psize, psize,
+ sizeof (psize));
+ zdb_nicenum(mdstats->zb_asize, asize,
+ sizeof (asize));
+ zdb_nicenum(mdstats->zb_asize / mdstats->zb_count, avg,
+ sizeof (avg));
+ zdb_nicenum(mdstats->zb_gangs, gang, sizeof (gang));
+
+ (void) printf("%6s\t%5s\t%5s\t%5s\t%5s"
+ "\t%5.2f\t%6.2f\t",
+ csize, lsize, psize, asize, avg,
+ (double)mdstats->zb_lsize / mdstats->zb_psize,
+ 100.0 * mdstats->zb_asize / tzb->zb_asize);
+ (void) printf("%s\n", "Metadata Total");
/* Output a table summarizing block sizes in the pool */
if (dump_opt['b'] >= 2) {
dump_size_histograms(zcb);
}
+
+ umem_free(mdstats, sizeof (zfs_blkstat_t));
}
(void) printf("\n");
if (leaks) {
umem_free(zcb, sizeof (zdb_cb_t));
return (2);
}
if (zcb->zcb_haderrors) {
umem_free(zcb, sizeof (zdb_cb_t));
return (3);
}
umem_free(zcb, sizeof (zdb_cb_t));
return (0);
}
typedef struct zdb_ddt_entry {
ddt_key_t zdde_key;
uint64_t zdde_ref_blocks;
uint64_t zdde_ref_lsize;
uint64_t zdde_ref_psize;
uint64_t zdde_ref_dsize;
avl_node_t zdde_node;
} zdb_ddt_entry_t;
static int
zdb_ddt_add_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
const zbookmark_phys_t *zb, const dnode_phys_t *dnp, void *arg)
{
(void) zilog, (void) dnp;
avl_tree_t *t = arg;
avl_index_t where;
zdb_ddt_entry_t *zdde, zdde_search;
if (zb->zb_level == ZB_DNODE_LEVEL || BP_IS_HOLE(bp) ||
BP_IS_EMBEDDED(bp))
return (0);
if (dump_opt['S'] > 1 && zb->zb_level == ZB_ROOT_LEVEL) {
(void) printf("traversing objset %llu, %llu objects, "
"%lu blocks so far\n",
(u_longlong_t)zb->zb_objset,
(u_longlong_t)BP_GET_FILL(bp),
avl_numnodes(t));
}
if (BP_IS_HOLE(bp) || BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_OFF ||
BP_GET_LEVEL(bp) > 0 || DMU_OT_IS_METADATA(BP_GET_TYPE(bp)))
return (0);
ddt_key_fill(&zdde_search.zdde_key, bp);
zdde = avl_find(t, &zdde_search, &where);
if (zdde == NULL) {
zdde = umem_zalloc(sizeof (*zdde), UMEM_NOFAIL);
zdde->zdde_key = zdde_search.zdde_key;
avl_insert(t, zdde, where);
}
zdde->zdde_ref_blocks += 1;
zdde->zdde_ref_lsize += BP_GET_LSIZE(bp);
zdde->zdde_ref_psize += BP_GET_PSIZE(bp);
zdde->zdde_ref_dsize += bp_get_dsize_sync(spa, bp);
return (0);
}
static void
dump_simulated_ddt(spa_t *spa)
{
avl_tree_t t;
void *cookie = NULL;
zdb_ddt_entry_t *zdde;
ddt_histogram_t ddh_total = {{{0}}};
ddt_stat_t dds_total = {0};
avl_create(&t, ddt_entry_compare,
sizeof (zdb_ddt_entry_t), offsetof(zdb_ddt_entry_t, zdde_node));
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
(void) traverse_pool(spa, 0, TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA |
TRAVERSE_NO_DECRYPT, zdb_ddt_add_cb, &t);
spa_config_exit(spa, SCL_CONFIG, FTAG);
while ((zdde = avl_destroy_nodes(&t, &cookie)) != NULL) {
ddt_stat_t dds;
uint64_t refcnt = zdde->zdde_ref_blocks;
ASSERT(refcnt != 0);
dds.dds_blocks = zdde->zdde_ref_blocks / refcnt;
dds.dds_lsize = zdde->zdde_ref_lsize / refcnt;
dds.dds_psize = zdde->zdde_ref_psize / refcnt;
dds.dds_dsize = zdde->zdde_ref_dsize / refcnt;
dds.dds_ref_blocks = zdde->zdde_ref_blocks;
dds.dds_ref_lsize = zdde->zdde_ref_lsize;
dds.dds_ref_psize = zdde->zdde_ref_psize;
dds.dds_ref_dsize = zdde->zdde_ref_dsize;
ddt_stat_add(&ddh_total.ddh_stat[highbit64(refcnt) - 1],
&dds, 0);
umem_free(zdde, sizeof (*zdde));
}
avl_destroy(&t);
ddt_histogram_stat(&dds_total, &ddh_total);
(void) printf("Simulated DDT histogram:\n");
zpool_dump_ddt(&dds_total, &ddh_total);
dump_dedup_ratio(&dds_total);
}
static int
verify_device_removal_feature_counts(spa_t *spa)
{
uint64_t dr_feature_refcount = 0;
uint64_t oc_feature_refcount = 0;
uint64_t indirect_vdev_count = 0;
uint64_t precise_vdev_count = 0;
uint64_t obsolete_counts_object_count = 0;
uint64_t obsolete_sm_count = 0;
uint64_t obsolete_counts_count = 0;
uint64_t scip_count = 0;
uint64_t obsolete_bpobj_count = 0;
int ret = 0;
spa_condensing_indirect_phys_t *scip =
&spa->spa_condensing_indirect_phys;
if (scip->scip_next_mapping_object != 0) {
vdev_t *vd = spa->spa_root_vdev->vdev_child[scip->scip_vdev];
ASSERT(scip->scip_prev_obsolete_sm_object != 0);
ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops);
(void) printf("Condensing indirect vdev %llu: new mapping "
"object %llu, prev obsolete sm %llu\n",
(u_longlong_t)scip->scip_vdev,
(u_longlong_t)scip->scip_next_mapping_object,
(u_longlong_t)scip->scip_prev_obsolete_sm_object);
if (scip->scip_prev_obsolete_sm_object != 0) {
space_map_t *prev_obsolete_sm = NULL;
VERIFY0(space_map_open(&prev_obsolete_sm,
spa->spa_meta_objset,
scip->scip_prev_obsolete_sm_object,
0, vd->vdev_asize, 0));
dump_spacemap(spa->spa_meta_objset, prev_obsolete_sm);
(void) printf("\n");
space_map_close(prev_obsolete_sm);
}
scip_count += 2;
}
for (uint64_t i = 0; i < spa->spa_root_vdev->vdev_children; i++) {
vdev_t *vd = spa->spa_root_vdev->vdev_child[i];
vdev_indirect_config_t *vic = &vd->vdev_indirect_config;
if (vic->vic_mapping_object != 0) {
ASSERT(vd->vdev_ops == &vdev_indirect_ops ||
vd->vdev_removing);
indirect_vdev_count++;
if (vd->vdev_indirect_mapping->vim_havecounts) {
obsolete_counts_count++;
}
}
boolean_t are_precise;
VERIFY0(vdev_obsolete_counts_are_precise(vd, &are_precise));
if (are_precise) {
ASSERT(vic->vic_mapping_object != 0);
precise_vdev_count++;
}
uint64_t obsolete_sm_object;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (obsolete_sm_object != 0) {
ASSERT(vic->vic_mapping_object != 0);
obsolete_sm_count++;
}
}
(void) feature_get_refcount(spa,
&spa_feature_table[SPA_FEATURE_DEVICE_REMOVAL],
&dr_feature_refcount);
(void) feature_get_refcount(spa,
&spa_feature_table[SPA_FEATURE_OBSOLETE_COUNTS],
&oc_feature_refcount);
if (dr_feature_refcount != indirect_vdev_count) {
ret = 1;
(void) printf("Number of indirect vdevs (%llu) " \
"does not match feature count (%llu)\n",
(u_longlong_t)indirect_vdev_count,
(u_longlong_t)dr_feature_refcount);
} else {
(void) printf("Verified device_removal feature refcount " \
"of %llu is correct\n",
(u_longlong_t)dr_feature_refcount);
}
if (zap_contains(spa_meta_objset(spa), DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_OBSOLETE_BPOBJ) == 0) {
obsolete_bpobj_count++;
}
obsolete_counts_object_count = precise_vdev_count;
obsolete_counts_object_count += obsolete_sm_count;
obsolete_counts_object_count += obsolete_counts_count;
obsolete_counts_object_count += scip_count;
obsolete_counts_object_count += obsolete_bpobj_count;
obsolete_counts_object_count += remap_deadlist_count;
if (oc_feature_refcount != obsolete_counts_object_count) {
ret = 1;
(void) printf("Number of obsolete counts objects (%llu) " \
"does not match feature count (%llu)\n",
(u_longlong_t)obsolete_counts_object_count,
(u_longlong_t)oc_feature_refcount);
(void) printf("pv:%llu os:%llu oc:%llu sc:%llu "
"ob:%llu rd:%llu\n",
(u_longlong_t)precise_vdev_count,
(u_longlong_t)obsolete_sm_count,
(u_longlong_t)obsolete_counts_count,
(u_longlong_t)scip_count,
(u_longlong_t)obsolete_bpobj_count,
(u_longlong_t)remap_deadlist_count);
} else {
(void) printf("Verified indirect_refcount feature refcount " \
"of %llu is correct\n",
(u_longlong_t)oc_feature_refcount);
}
return (ret);
}
static void
zdb_set_skip_mmp(char *target)
{
spa_t *spa;
/*
* Disable the activity check to allow examination of
* active pools.
*/
mutex_enter(&spa_namespace_lock);
if ((spa = spa_lookup(target)) != NULL) {
spa->spa_import_flags |= ZFS_IMPORT_SKIP_MMP;
}
mutex_exit(&spa_namespace_lock);
}
#define BOGUS_SUFFIX "_CHECKPOINTED_UNIVERSE"
/*
* Import the checkpointed state of the pool specified by the target
* parameter as readonly. The function also accepts a pool config
* as an optional parameter, else it attempts to infer the config by
* the name of the target pool.
*
* Note that the checkpointed state's pool name will be the name of
* the original pool with the above suffix appended to it. In addition,
* if the target is not a pool name (e.g. a path to a dataset) then
* the new_path parameter is populated with the updated path to
* reflect the fact that we are looking into the checkpointed state.
*
* The function returns a newly-allocated copy of the name of the
* pool containing the checkpointed state. When this copy is no
* longer needed it should be freed with free(3C). Same thing
* applies to the new_path parameter if allocated.
*/
static char *
import_checkpointed_state(char *target, nvlist_t *cfg, char **new_path)
{
int error = 0;
char *poolname, *bogus_name = NULL;
boolean_t freecfg = B_FALSE;
/* If the target is not a pool, the extract the pool name */
char *path_start = strchr(target, '/');
if (path_start != NULL) {
size_t poolname_len = path_start - target;
poolname = strndup(target, poolname_len);
} else {
poolname = target;
}
if (cfg == NULL) {
zdb_set_skip_mmp(poolname);
error = spa_get_stats(poolname, &cfg, NULL, 0);
if (error != 0) {
fatal("Tried to read config of pool \"%s\" but "
"spa_get_stats() failed with error %d\n",
poolname, error);
}
freecfg = B_TRUE;
}
if (asprintf(&bogus_name, "%s%s", poolname, BOGUS_SUFFIX) == -1) {
if (target != poolname)
free(poolname);
return (NULL);
}
fnvlist_add_string(cfg, ZPOOL_CONFIG_POOL_NAME, bogus_name);
error = spa_import(bogus_name, cfg, NULL,
ZFS_IMPORT_MISSING_LOG | ZFS_IMPORT_CHECKPOINT |
ZFS_IMPORT_SKIP_MMP);
if (freecfg)
nvlist_free(cfg);
if (error != 0) {
fatal("Tried to import pool \"%s\" but spa_import() failed "
"with error %d\n", bogus_name, error);
}
if (new_path != NULL && path_start != NULL) {
if (asprintf(new_path, "%s%s", bogus_name, path_start) == -1) {
free(bogus_name);
if (path_start != NULL)
free(poolname);
return (NULL);
}
}
if (target != poolname)
free(poolname);
return (bogus_name);
}
typedef struct verify_checkpoint_sm_entry_cb_arg {
vdev_t *vcsec_vd;
/* the following fields are only used for printing progress */
uint64_t vcsec_entryid;
uint64_t vcsec_num_entries;
} verify_checkpoint_sm_entry_cb_arg_t;
#define ENTRIES_PER_PROGRESS_UPDATE 10000
static int
verify_checkpoint_sm_entry_cb(space_map_entry_t *sme, void *arg)
{
verify_checkpoint_sm_entry_cb_arg_t *vcsec = arg;
vdev_t *vd = vcsec->vcsec_vd;
metaslab_t *ms = vd->vdev_ms[sme->sme_offset >> vd->vdev_ms_shift];
uint64_t end = sme->sme_offset + sme->sme_run;
ASSERT(sme->sme_type == SM_FREE);
if ((vcsec->vcsec_entryid % ENTRIES_PER_PROGRESS_UPDATE) == 0) {
(void) fprintf(stderr,
"\rverifying vdev %llu, space map entry %llu of %llu ...",
(longlong_t)vd->vdev_id,
(longlong_t)vcsec->vcsec_entryid,
(longlong_t)vcsec->vcsec_num_entries);
}
vcsec->vcsec_entryid++;
/*
* See comment in checkpoint_sm_exclude_entry_cb()
*/
VERIFY3U(sme->sme_offset, >=, ms->ms_start);
VERIFY3U(end, <=, ms->ms_start + ms->ms_size);
/*
* The entries in the vdev_checkpoint_sm should be marked as
* allocated in the checkpointed state of the pool, therefore
* their respective ms_allocateable trees should not contain them.
*/
mutex_enter(&ms->ms_lock);
range_tree_verify_not_present(ms->ms_allocatable,
sme->sme_offset, sme->sme_run);
mutex_exit(&ms->ms_lock);
return (0);
}
/*
* Verify that all segments in the vdev_checkpoint_sm are allocated
* according to the checkpoint's ms_sm (i.e. are not in the checkpoint's
* ms_allocatable).
*
* Do so by comparing the checkpoint space maps (vdev_checkpoint_sm) of
* each vdev in the current state of the pool to the metaslab space maps
* (ms_sm) of the checkpointed state of the pool.
*
* Note that the function changes the state of the ms_allocatable
* trees of the current spa_t. The entries of these ms_allocatable
* trees are cleared out and then repopulated from with the free
* entries of their respective ms_sm space maps.
*/
static void
verify_checkpoint_vdev_spacemaps(spa_t *checkpoint, spa_t *current)
{
vdev_t *ckpoint_rvd = checkpoint->spa_root_vdev;
vdev_t *current_rvd = current->spa_root_vdev;
load_concrete_ms_allocatable_trees(checkpoint, SM_FREE);
for (uint64_t c = 0; c < ckpoint_rvd->vdev_children; c++) {
vdev_t *ckpoint_vd = ckpoint_rvd->vdev_child[c];
vdev_t *current_vd = current_rvd->vdev_child[c];
space_map_t *checkpoint_sm = NULL;
uint64_t checkpoint_sm_obj;
if (ckpoint_vd->vdev_ops == &vdev_indirect_ops) {
/*
* Since we don't allow device removal in a pool
* that has a checkpoint, we expect that all removed
* vdevs were removed from the pool before the
* checkpoint.
*/
ASSERT3P(current_vd->vdev_ops, ==, &vdev_indirect_ops);
continue;
}
/*
* If the checkpoint space map doesn't exist, then nothing
* here is checkpointed so there's nothing to verify.
*/
if (current_vd->vdev_top_zap == 0 ||
zap_contains(spa_meta_objset(current),
current_vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM) != 0)
continue;
VERIFY0(zap_lookup(spa_meta_objset(current),
current_vd->vdev_top_zap, VDEV_TOP_ZAP_POOL_CHECKPOINT_SM,
sizeof (uint64_t), 1, &checkpoint_sm_obj));
VERIFY0(space_map_open(&checkpoint_sm, spa_meta_objset(current),
checkpoint_sm_obj, 0, current_vd->vdev_asize,
current_vd->vdev_ashift));
verify_checkpoint_sm_entry_cb_arg_t vcsec;
vcsec.vcsec_vd = ckpoint_vd;
vcsec.vcsec_entryid = 0;
vcsec.vcsec_num_entries =
space_map_length(checkpoint_sm) / sizeof (uint64_t);
VERIFY0(space_map_iterate(checkpoint_sm,
space_map_length(checkpoint_sm),
verify_checkpoint_sm_entry_cb, &vcsec));
if (dump_opt['m'] > 3)
dump_spacemap(current->spa_meta_objset, checkpoint_sm);
space_map_close(checkpoint_sm);
}
/*
* If we've added vdevs since we took the checkpoint, ensure
* that their checkpoint space maps are empty.
*/
if (ckpoint_rvd->vdev_children < current_rvd->vdev_children) {
for (uint64_t c = ckpoint_rvd->vdev_children;
c < current_rvd->vdev_children; c++) {
vdev_t *current_vd = current_rvd->vdev_child[c];
VERIFY3P(current_vd->vdev_checkpoint_sm, ==, NULL);
}
}
/* for cleaner progress output */
(void) fprintf(stderr, "\n");
}
/*
* Verifies that all space that's allocated in the checkpoint is
* still allocated in the current version, by checking that everything
* in checkpoint's ms_allocatable (which is actually allocated, not
* allocatable/free) is not present in current's ms_allocatable.
*
* Note that the function changes the state of the ms_allocatable
* trees of both spas when called. The entries of all ms_allocatable
* trees are cleared out and then repopulated from their respective
* ms_sm space maps. In the checkpointed state we load the allocated
* entries, and in the current state we load the free entries.
*/
static void
verify_checkpoint_ms_spacemaps(spa_t *checkpoint, spa_t *current)
{
vdev_t *ckpoint_rvd = checkpoint->spa_root_vdev;
vdev_t *current_rvd = current->spa_root_vdev;
load_concrete_ms_allocatable_trees(checkpoint, SM_ALLOC);
load_concrete_ms_allocatable_trees(current, SM_FREE);
for (uint64_t i = 0; i < ckpoint_rvd->vdev_children; i++) {
vdev_t *ckpoint_vd = ckpoint_rvd->vdev_child[i];
vdev_t *current_vd = current_rvd->vdev_child[i];
if (ckpoint_vd->vdev_ops == &vdev_indirect_ops) {
/*
* See comment in verify_checkpoint_vdev_spacemaps()
*/
ASSERT3P(current_vd->vdev_ops, ==, &vdev_indirect_ops);
continue;
}
for (uint64_t m = 0; m < ckpoint_vd->vdev_ms_count; m++) {
metaslab_t *ckpoint_msp = ckpoint_vd->vdev_ms[m];
metaslab_t *current_msp = current_vd->vdev_ms[m];
(void) fprintf(stderr,
"\rverifying vdev %llu of %llu, "
"metaslab %llu of %llu ...",
(longlong_t)current_vd->vdev_id,
(longlong_t)current_rvd->vdev_children,
(longlong_t)current_vd->vdev_ms[m]->ms_id,
(longlong_t)current_vd->vdev_ms_count);
/*
* We walk through the ms_allocatable trees that
* are loaded with the allocated blocks from the
* ms_sm spacemaps of the checkpoint. For each
* one of these ranges we ensure that none of them
* exists in the ms_allocatable trees of the
* current state which are loaded with the ranges
* that are currently free.
*
* This way we ensure that none of the blocks that
* are part of the checkpoint were freed by mistake.
*/
range_tree_walk(ckpoint_msp->ms_allocatable,
(range_tree_func_t *)range_tree_verify_not_present,
current_msp->ms_allocatable);
}
}
/* for cleaner progress output */
(void) fprintf(stderr, "\n");
}
static void
verify_checkpoint_blocks(spa_t *spa)
{
ASSERT(!dump_opt['L']);
spa_t *checkpoint_spa;
char *checkpoint_pool;
int error = 0;
/*
* We import the checkpointed state of the pool (under a different
* name) so we can do verification on it against the current state
* of the pool.
*/
checkpoint_pool = import_checkpointed_state(spa->spa_name, NULL,
NULL);
ASSERT(strcmp(spa->spa_name, checkpoint_pool) != 0);
error = spa_open(checkpoint_pool, &checkpoint_spa, FTAG);
if (error != 0) {
fatal("Tried to open pool \"%s\" but spa_open() failed with "
"error %d\n", checkpoint_pool, error);
}
/*
* Ensure that ranges in the checkpoint space maps of each vdev
* are allocated according to the checkpointed state's metaslab
* space maps.
*/
verify_checkpoint_vdev_spacemaps(checkpoint_spa, spa);
/*
* Ensure that allocated ranges in the checkpoint's metaslab
* space maps remain allocated in the metaslab space maps of
* the current state.
*/
verify_checkpoint_ms_spacemaps(checkpoint_spa, spa);
/*
* Once we are done, we get rid of the checkpointed state.
*/
spa_close(checkpoint_spa, FTAG);
free(checkpoint_pool);
}
static void
dump_leftover_checkpoint_blocks(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t i = 0; i < rvd->vdev_children; i++) {
vdev_t *vd = rvd->vdev_child[i];
space_map_t *checkpoint_sm = NULL;
uint64_t checkpoint_sm_obj;
if (vd->vdev_top_zap == 0)
continue;
if (zap_contains(spa_meta_objset(spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM) != 0)
continue;
VERIFY0(zap_lookup(spa_meta_objset(spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM,
sizeof (uint64_t), 1, &checkpoint_sm_obj));
VERIFY0(space_map_open(&checkpoint_sm, spa_meta_objset(spa),
checkpoint_sm_obj, 0, vd->vdev_asize, vd->vdev_ashift));
dump_spacemap(spa->spa_meta_objset, checkpoint_sm);
space_map_close(checkpoint_sm);
}
}
static int
verify_checkpoint(spa_t *spa)
{
uberblock_t checkpoint;
int error;
if (!spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT))
return (0);
error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ZPOOL_CHECKPOINT, sizeof (uint64_t),
sizeof (uberblock_t) / sizeof (uint64_t), &checkpoint);
if (error == ENOENT && !dump_opt['L']) {
/*
* If the feature is active but the uberblock is missing
* then we must be in the middle of discarding the
* checkpoint.
*/
(void) printf("\nPartially discarded checkpoint "
"state found:\n");
if (dump_opt['m'] > 3)
dump_leftover_checkpoint_blocks(spa);
return (0);
} else if (error != 0) {
(void) printf("lookup error %d when looking for "
"checkpointed uberblock in MOS\n", error);
return (error);
}
dump_uberblock(&checkpoint, "\nCheckpointed uberblock found:\n", "\n");
if (checkpoint.ub_checkpoint_txg == 0) {
(void) printf("\nub_checkpoint_txg not set in checkpointed "
"uberblock\n");
error = 3;
}
if (error == 0 && !dump_opt['L'])
verify_checkpoint_blocks(spa);
return (error);
}
static void
mos_leaks_cb(void *arg, uint64_t start, uint64_t size)
{
(void) arg;
for (uint64_t i = start; i < size; i++) {
(void) printf("MOS object %llu referenced but not allocated\n",
(u_longlong_t)i);
}
}
static void
mos_obj_refd(uint64_t obj)
{
if (obj != 0 && mos_refd_objs != NULL)
range_tree_add(mos_refd_objs, obj, 1);
}
/*
* Call on a MOS object that may already have been referenced.
*/
static void
mos_obj_refd_multiple(uint64_t obj)
{
if (obj != 0 && mos_refd_objs != NULL &&
!range_tree_contains(mos_refd_objs, obj, 1))
range_tree_add(mos_refd_objs, obj, 1);
}
static void
mos_leak_vdev_top_zap(vdev_t *vd)
{
uint64_t ms_flush_data_obj;
int error = zap_lookup(spa_meta_objset(vd->vdev_spa),
vd->vdev_top_zap, VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS,
sizeof (ms_flush_data_obj), 1, &ms_flush_data_obj);
if (error == ENOENT)
return;
ASSERT0(error);
mos_obj_refd(ms_flush_data_obj);
}
static void
mos_leak_vdev(vdev_t *vd)
{
mos_obj_refd(vd->vdev_dtl_object);
mos_obj_refd(vd->vdev_ms_array);
mos_obj_refd(vd->vdev_indirect_config.vic_births_object);
mos_obj_refd(vd->vdev_indirect_config.vic_mapping_object);
mos_obj_refd(vd->vdev_leaf_zap);
if (vd->vdev_checkpoint_sm != NULL)
mos_obj_refd(vd->vdev_checkpoint_sm->sm_object);
if (vd->vdev_indirect_mapping != NULL) {
mos_obj_refd(vd->vdev_indirect_mapping->
vim_phys->vimp_counts_object);
}
if (vd->vdev_obsolete_sm != NULL)
mos_obj_refd(vd->vdev_obsolete_sm->sm_object);
for (uint64_t m = 0; m < vd->vdev_ms_count; m++) {
metaslab_t *ms = vd->vdev_ms[m];
mos_obj_refd(space_map_object(ms->ms_sm));
}
+ if (vd->vdev_root_zap != 0)
+ mos_obj_refd(vd->vdev_root_zap);
+
if (vd->vdev_top_zap != 0) {
mos_obj_refd(vd->vdev_top_zap);
mos_leak_vdev_top_zap(vd);
}
for (uint64_t c = 0; c < vd->vdev_children; c++) {
mos_leak_vdev(vd->vdev_child[c]);
}
}
static void
mos_leak_log_spacemaps(spa_t *spa)
{
uint64_t spacemap_zap;
int error = zap_lookup(spa_meta_objset(spa),
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_LOG_SPACEMAP_ZAP,
sizeof (spacemap_zap), 1, &spacemap_zap);
if (error == ENOENT)
return;
ASSERT0(error);
mos_obj_refd(spacemap_zap);
for (spa_log_sm_t *sls = avl_first(&spa->spa_sm_logs_by_txg);
sls; sls = AVL_NEXT(&spa->spa_sm_logs_by_txg, sls))
mos_obj_refd(sls->sls_sm_obj);
}
static void
errorlog_count_refd(objset_t *mos, uint64_t errlog)
{
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, mos, errlog);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
mos_obj_refd(za.za_first_integer);
}
zap_cursor_fini(&zc);
}
static int
dump_mos_leaks(spa_t *spa)
{
int rv = 0;
objset_t *mos = spa->spa_meta_objset;
dsl_pool_t *dp = spa->spa_dsl_pool;
/* Visit and mark all referenced objects in the MOS */
mos_obj_refd(DMU_POOL_DIRECTORY_OBJECT);
mos_obj_refd(spa->spa_pool_props_object);
mos_obj_refd(spa->spa_config_object);
mos_obj_refd(spa->spa_ddt_stat_object);
mos_obj_refd(spa->spa_feat_desc_obj);
mos_obj_refd(spa->spa_feat_enabled_txg_obj);
mos_obj_refd(spa->spa_feat_for_read_obj);
mos_obj_refd(spa->spa_feat_for_write_obj);
mos_obj_refd(spa->spa_history);
mos_obj_refd(spa->spa_errlog_last);
mos_obj_refd(spa->spa_errlog_scrub);
if (spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
errorlog_count_refd(mos, spa->spa_errlog_last);
errorlog_count_refd(mos, spa->spa_errlog_scrub);
}
mos_obj_refd(spa->spa_all_vdev_zaps);
mos_obj_refd(spa->spa_dsl_pool->dp_bptree_obj);
mos_obj_refd(spa->spa_dsl_pool->dp_tmp_userrefs_obj);
mos_obj_refd(spa->spa_dsl_pool->dp_scan->scn_phys.scn_queue_obj);
bpobj_count_refd(&spa->spa_deferred_bpobj);
mos_obj_refd(dp->dp_empty_bpobj);
bpobj_count_refd(&dp->dp_obsolete_bpobj);
bpobj_count_refd(&dp->dp_free_bpobj);
mos_obj_refd(spa->spa_l2cache.sav_object);
mos_obj_refd(spa->spa_spares.sav_object);
if (spa->spa_syncing_log_sm != NULL)
mos_obj_refd(spa->spa_syncing_log_sm->sm_object);
mos_leak_log_spacemaps(spa);
mos_obj_refd(spa->spa_condensing_indirect_phys.
scip_next_mapping_object);
mos_obj_refd(spa->spa_condensing_indirect_phys.
scip_prev_obsolete_sm_object);
if (spa->spa_condensing_indirect_phys.scip_next_mapping_object != 0) {
vdev_indirect_mapping_t *vim =
vdev_indirect_mapping_open(mos,
spa->spa_condensing_indirect_phys.scip_next_mapping_object);
mos_obj_refd(vim->vim_phys->vimp_counts_object);
vdev_indirect_mapping_close(vim);
}
deleted_livelists_dump_mos(spa);
if (dp->dp_origin_snap != NULL) {
dsl_dataset_t *ds;
dsl_pool_config_enter(dp, FTAG);
VERIFY0(dsl_dataset_hold_obj(dp,
dsl_dataset_phys(dp->dp_origin_snap)->ds_next_snap_obj,
FTAG, &ds));
count_ds_mos_objects(ds);
dump_blkptr_list(&ds->ds_deadlist, "Deadlist");
dsl_dataset_rele(ds, FTAG);
dsl_pool_config_exit(dp, FTAG);
count_ds_mos_objects(dp->dp_origin_snap);
dump_blkptr_list(&dp->dp_origin_snap->ds_deadlist, "Deadlist");
}
count_dir_mos_objects(dp->dp_mos_dir);
if (dp->dp_free_dir != NULL)
count_dir_mos_objects(dp->dp_free_dir);
if (dp->dp_leak_dir != NULL)
count_dir_mos_objects(dp->dp_leak_dir);
mos_leak_vdev(spa->spa_root_vdev);
for (uint64_t class = 0; class < DDT_CLASSES; class++) {
for (uint64_t type = 0; type < DDT_TYPES; type++) {
for (uint64_t cksum = 0;
cksum < ZIO_CHECKSUM_FUNCTIONS; cksum++) {
ddt_t *ddt = spa->spa_ddt[cksum];
mos_obj_refd(ddt->ddt_object[type][class]);
}
}
}
/*
* Visit all allocated objects and make sure they are referenced.
*/
uint64_t object = 0;
while (dmu_object_next(mos, &object, B_FALSE, 0) == 0) {
if (range_tree_contains(mos_refd_objs, object, 1)) {
range_tree_remove(mos_refd_objs, object, 1);
} else {
dmu_object_info_t doi;
const char *name;
VERIFY0(dmu_object_info(mos, object, &doi));
if (doi.doi_type & DMU_OT_NEWTYPE) {
dmu_object_byteswap_t bswap =
DMU_OT_BYTESWAP(doi.doi_type);
name = dmu_ot_byteswap[bswap].ob_name;
} else {
name = dmu_ot[doi.doi_type].ot_name;
}
(void) printf("MOS object %llu (%s) leaked\n",
(u_longlong_t)object, name);
rv = 2;
}
}
(void) range_tree_walk(mos_refd_objs, mos_leaks_cb, NULL);
if (!range_tree_is_empty(mos_refd_objs))
rv = 2;
range_tree_vacate(mos_refd_objs, NULL, NULL);
range_tree_destroy(mos_refd_objs);
return (rv);
}
typedef struct log_sm_obsolete_stats_arg {
uint64_t lsos_current_txg;
uint64_t lsos_total_entries;
uint64_t lsos_valid_entries;
uint64_t lsos_sm_entries;
uint64_t lsos_valid_sm_entries;
} log_sm_obsolete_stats_arg_t;
static int
log_spacemap_obsolete_stats_cb(spa_t *spa, space_map_entry_t *sme,
uint64_t txg, void *arg)
{
log_sm_obsolete_stats_arg_t *lsos = arg;
uint64_t offset = sme->sme_offset;
uint64_t vdev_id = sme->sme_vdev;
if (lsos->lsos_current_txg == 0) {
/* this is the first log */
lsos->lsos_current_txg = txg;
} else if (lsos->lsos_current_txg < txg) {
/* we just changed log - print stats and reset */
(void) printf("%-8llu valid entries out of %-8llu - txg %llu\n",
(u_longlong_t)lsos->lsos_valid_sm_entries,
(u_longlong_t)lsos->lsos_sm_entries,
(u_longlong_t)lsos->lsos_current_txg);
lsos->lsos_valid_sm_entries = 0;
lsos->lsos_sm_entries = 0;
lsos->lsos_current_txg = txg;
}
ASSERT3U(lsos->lsos_current_txg, ==, txg);
lsos->lsos_sm_entries++;
lsos->lsos_total_entries++;
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
if (!vdev_is_concrete(vd))
return (0);
metaslab_t *ms = vd->vdev_ms[offset >> vd->vdev_ms_shift];
ASSERT(sme->sme_type == SM_ALLOC || sme->sme_type == SM_FREE);
if (txg < metaslab_unflushed_txg(ms))
return (0);
lsos->lsos_valid_sm_entries++;
lsos->lsos_valid_entries++;
return (0);
}
static void
dump_log_spacemap_obsolete_stats(spa_t *spa)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP))
return;
log_sm_obsolete_stats_arg_t lsos = {0};
(void) printf("Log Space Map Obsolete Entry Statistics:\n");
iterate_through_spacemap_logs(spa,
log_spacemap_obsolete_stats_cb, &lsos);
/* print stats for latest log */
(void) printf("%-8llu valid entries out of %-8llu - txg %llu\n",
(u_longlong_t)lsos.lsos_valid_sm_entries,
(u_longlong_t)lsos.lsos_sm_entries,
(u_longlong_t)lsos.lsos_current_txg);
(void) printf("%-8llu valid entries out of %-8llu - total\n\n",
(u_longlong_t)lsos.lsos_valid_entries,
(u_longlong_t)lsos.lsos_total_entries);
}
static void
dump_zpool(spa_t *spa)
{
dsl_pool_t *dp = spa_get_dsl(spa);
int rc = 0;
if (dump_opt['y']) {
livelist_metaslab_validate(spa);
}
if (dump_opt['S']) {
dump_simulated_ddt(spa);
return;
}
if (!dump_opt['e'] && dump_opt['C'] > 1) {
(void) printf("\nCached configuration:\n");
dump_nvlist(spa->spa_config, 8);
}
if (dump_opt['C'])
dump_config(spa);
if (dump_opt['u'])
dump_uberblock(&spa->spa_uberblock, "\nUberblock:\n", "\n");
if (dump_opt['D'])
dump_all_ddts(spa);
if (dump_opt['d'] > 2 || dump_opt['m'])
dump_metaslabs(spa);
if (dump_opt['M'])
dump_metaslab_groups(spa, dump_opt['M'] > 1);
if (dump_opt['d'] > 2 || dump_opt['m']) {
dump_log_spacemaps(spa);
dump_log_spacemap_obsolete_stats(spa);
}
if (dump_opt['d'] || dump_opt['i']) {
spa_feature_t f;
mos_refd_objs = range_tree_create(NULL, RANGE_SEG64, NULL, 0,
0);
dump_objset(dp->dp_meta_objset);
if (dump_opt['d'] >= 3) {
dsl_pool_t *dp = spa->spa_dsl_pool;
dump_full_bpobj(&spa->spa_deferred_bpobj,
"Deferred frees", 0);
if (spa_version(spa) >= SPA_VERSION_DEADLISTS) {
dump_full_bpobj(&dp->dp_free_bpobj,
"Pool snapshot frees", 0);
}
if (bpobj_is_open(&dp->dp_obsolete_bpobj)) {
ASSERT(spa_feature_is_enabled(spa,
SPA_FEATURE_DEVICE_REMOVAL));
dump_full_bpobj(&dp->dp_obsolete_bpobj,
"Pool obsolete blocks", 0);
}
if (spa_feature_is_active(spa,
SPA_FEATURE_ASYNC_DESTROY)) {
dump_bptree(spa->spa_meta_objset,
dp->dp_bptree_obj,
"Pool dataset frees");
}
dump_dtl(spa->spa_root_vdev, 0);
}
for (spa_feature_t f = 0; f < SPA_FEATURES; f++)
global_feature_count[f] = UINT64_MAX;
global_feature_count[SPA_FEATURE_REDACTION_BOOKMARKS] = 0;
global_feature_count[SPA_FEATURE_BOOKMARK_WRITTEN] = 0;
global_feature_count[SPA_FEATURE_LIVELIST] = 0;
(void) dmu_objset_find(spa_name(spa), dump_one_objset,
NULL, DS_FIND_SNAPSHOTS | DS_FIND_CHILDREN);
if (rc == 0 && !dump_opt['L'])
rc = dump_mos_leaks(spa);
for (f = 0; f < SPA_FEATURES; f++) {
uint64_t refcount;
uint64_t *arr;
if (!(spa_feature_table[f].fi_flags &
ZFEATURE_FLAG_PER_DATASET)) {
if (global_feature_count[f] == UINT64_MAX)
continue;
if (!spa_feature_is_enabled(spa, f)) {
ASSERT0(global_feature_count[f]);
continue;
}
arr = global_feature_count;
} else {
if (!spa_feature_is_enabled(spa, f)) {
ASSERT0(dataset_feature_count[f]);
continue;
}
arr = dataset_feature_count;
}
if (feature_get_refcount(spa, &spa_feature_table[f],
&refcount) == ENOTSUP)
continue;
if (arr[f] != refcount) {
(void) printf("%s feature refcount mismatch: "
"%lld consumers != %lld refcount\n",
spa_feature_table[f].fi_uname,
(longlong_t)arr[f], (longlong_t)refcount);
rc = 2;
} else {
(void) printf("Verified %s feature refcount "
"of %llu is correct\n",
spa_feature_table[f].fi_uname,
(longlong_t)refcount);
}
}
if (rc == 0)
rc = verify_device_removal_feature_counts(spa);
}
if (rc == 0 && (dump_opt['b'] || dump_opt['c']))
rc = dump_block_stats(spa);
if (rc == 0)
rc = verify_spacemap_refcounts(spa);
if (dump_opt['s'])
show_pool_stats(spa);
if (dump_opt['h'])
dump_history(spa);
if (rc == 0)
rc = verify_checkpoint(spa);
if (rc != 0) {
dump_debug_buffer();
exit(rc);
}
}
#define ZDB_FLAG_CHECKSUM 0x0001
#define ZDB_FLAG_DECOMPRESS 0x0002
#define ZDB_FLAG_BSWAP 0x0004
#define ZDB_FLAG_GBH 0x0008
#define ZDB_FLAG_INDIRECT 0x0010
#define ZDB_FLAG_RAW 0x0020
#define ZDB_FLAG_PRINT_BLKPTR 0x0040
#define ZDB_FLAG_VERBOSE 0x0080
static int flagbits[256];
static char flagbitstr[16];
static void
zdb_print_blkptr(const blkptr_t *bp, int flags)
{
char blkbuf[BP_SPRINTF_LEN];
if (flags & ZDB_FLAG_BSWAP)
byteswap_uint64_array((void *)bp, sizeof (blkptr_t));
snprintf_blkptr(blkbuf, sizeof (blkbuf), bp);
(void) printf("%s\n", blkbuf);
}
static void
zdb_dump_indirect(blkptr_t *bp, int nbps, int flags)
{
int i;
for (i = 0; i < nbps; i++)
zdb_print_blkptr(&bp[i], flags);
}
static void
zdb_dump_gbh(void *buf, int flags)
{
zdb_dump_indirect((blkptr_t *)buf, SPA_GBH_NBLKPTRS, flags);
}
static void
zdb_dump_block_raw(void *buf, uint64_t size, int flags)
{
if (flags & ZDB_FLAG_BSWAP)
byteswap_uint64_array(buf, size);
VERIFY(write(fileno(stdout), buf, size) == size);
}
static void
zdb_dump_block(char *label, void *buf, uint64_t size, int flags)
{
uint64_t *d = (uint64_t *)buf;
unsigned nwords = size / sizeof (uint64_t);
int do_bswap = !!(flags & ZDB_FLAG_BSWAP);
unsigned i, j;
const char *hdr;
char *c;
if (do_bswap)
hdr = " 7 6 5 4 3 2 1 0 f e d c b a 9 8";
else
hdr = " 0 1 2 3 4 5 6 7 8 9 a b c d e f";
(void) printf("\n%s\n%6s %s 0123456789abcdef\n", label, "", hdr);
#ifdef _LITTLE_ENDIAN
/* correct the endianness */
do_bswap = !do_bswap;
#endif
for (i = 0; i < nwords; i += 2) {
(void) printf("%06llx: %016llx %016llx ",
(u_longlong_t)(i * sizeof (uint64_t)),
(u_longlong_t)(do_bswap ? BSWAP_64(d[i]) : d[i]),
(u_longlong_t)(do_bswap ? BSWAP_64(d[i + 1]) : d[i + 1]));
c = (char *)&d[i];
for (j = 0; j < 2 * sizeof (uint64_t); j++)
(void) printf("%c", isprint(c[j]) ? c[j] : '.');
(void) printf("\n");
}
}
/*
* There are two acceptable formats:
* leaf_name - For example: c1t0d0 or /tmp/ztest.0a
* child[.child]* - For example: 0.1.1
*
* The second form can be used to specify arbitrary vdevs anywhere
* in the hierarchy. For example, in a pool with a mirror of
* RAID-Zs, you can specify either RAID-Z vdev with 0.0 or 0.1 .
*/
static vdev_t *
zdb_vdev_lookup(vdev_t *vdev, const char *path)
{
char *s, *p, *q;
unsigned i;
if (vdev == NULL)
return (NULL);
/* First, assume the x.x.x.x format */
i = strtoul(path, &s, 10);
if (s == path || (s && *s != '.' && *s != '\0'))
goto name;
if (i >= vdev->vdev_children)
return (NULL);
vdev = vdev->vdev_child[i];
if (s && *s == '\0')
return (vdev);
return (zdb_vdev_lookup(vdev, s+1));
name:
for (i = 0; i < vdev->vdev_children; i++) {
vdev_t *vc = vdev->vdev_child[i];
if (vc->vdev_path == NULL) {
vc = zdb_vdev_lookup(vc, path);
if (vc == NULL)
continue;
else
return (vc);
}
p = strrchr(vc->vdev_path, '/');
p = p ? p + 1 : vc->vdev_path;
q = &vc->vdev_path[strlen(vc->vdev_path) - 2];
if (strcmp(vc->vdev_path, path) == 0)
return (vc);
if (strcmp(p, path) == 0)
return (vc);
if (strcmp(q, "s0") == 0 && strncmp(p, path, q - p) == 0)
return (vc);
}
return (NULL);
}
static int
name_from_objset_id(spa_t *spa, uint64_t objset_id, char *outstr)
{
dsl_dataset_t *ds;
dsl_pool_config_enter(spa->spa_dsl_pool, FTAG);
int error = dsl_dataset_hold_obj(spa->spa_dsl_pool, objset_id,
NULL, &ds);
if (error != 0) {
(void) fprintf(stderr, "failed to hold objset %llu: %s\n",
(u_longlong_t)objset_id, strerror(error));
dsl_pool_config_exit(spa->spa_dsl_pool, FTAG);
return (error);
}
dsl_dataset_name(ds, outstr);
dsl_dataset_rele(ds, NULL);
dsl_pool_config_exit(spa->spa_dsl_pool, FTAG);
return (0);
}
static boolean_t
zdb_parse_block_sizes(char *sizes, uint64_t *lsize, uint64_t *psize)
{
char *s0, *s1, *tmp = NULL;
if (sizes == NULL)
return (B_FALSE);
s0 = strtok_r(sizes, "/", &tmp);
if (s0 == NULL)
return (B_FALSE);
s1 = strtok_r(NULL, "/", &tmp);
*lsize = strtoull(s0, NULL, 16);
*psize = s1 ? strtoull(s1, NULL, 16) : *lsize;
return (*lsize >= *psize && *psize > 0);
}
#define ZIO_COMPRESS_MASK(alg) (1ULL << (ZIO_COMPRESS_##alg))
static boolean_t
zdb_decompress_block(abd_t *pabd, void *buf, void *lbuf, uint64_t lsize,
uint64_t psize, int flags)
{
(void) buf;
boolean_t exceeded = B_FALSE;
/*
* We don't know how the data was compressed, so just try
* every decompress function at every inflated blocksize.
*/
void *lbuf2 = umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL);
int cfuncs[ZIO_COMPRESS_FUNCTIONS] = { 0 };
int *cfuncp = cfuncs;
uint64_t maxlsize = SPA_MAXBLOCKSIZE;
uint64_t mask = ZIO_COMPRESS_MASK(ON) | ZIO_COMPRESS_MASK(OFF) |
ZIO_COMPRESS_MASK(INHERIT) | ZIO_COMPRESS_MASK(EMPTY) |
(getenv("ZDB_NO_ZLE") ? ZIO_COMPRESS_MASK(ZLE) : 0);
*cfuncp++ = ZIO_COMPRESS_LZ4;
*cfuncp++ = ZIO_COMPRESS_LZJB;
mask |= ZIO_COMPRESS_MASK(LZ4) | ZIO_COMPRESS_MASK(LZJB);
for (int c = 0; c < ZIO_COMPRESS_FUNCTIONS; c++)
if (((1ULL << c) & mask) == 0)
*cfuncp++ = c;
/*
* On the one hand, with SPA_MAXBLOCKSIZE at 16MB, this
* could take a while and we should let the user know
* we are not stuck. On the other hand, printing progress
* info gets old after a while. User can specify 'v' flag
* to see the progression.
*/
if (lsize == psize)
lsize += SPA_MINBLOCKSIZE;
else
maxlsize = lsize;
for (; lsize <= maxlsize; lsize += SPA_MINBLOCKSIZE) {
for (cfuncp = cfuncs; *cfuncp; cfuncp++) {
if (flags & ZDB_FLAG_VERBOSE) {
(void) fprintf(stderr,
"Trying %05llx -> %05llx (%s)\n",
(u_longlong_t)psize,
(u_longlong_t)lsize,
zio_compress_table[*cfuncp].\
ci_name);
}
/*
* We randomize lbuf2, and decompress to both
* lbuf and lbuf2. This way, we will know if
* decompression fill exactly to lsize.
*/
VERIFY0(random_get_pseudo_bytes(lbuf2, lsize));
if (zio_decompress_data(*cfuncp, pabd,
lbuf, psize, lsize, NULL) == 0 &&
zio_decompress_data(*cfuncp, pabd,
lbuf2, psize, lsize, NULL) == 0 &&
memcmp(lbuf, lbuf2, lsize) == 0)
break;
}
if (*cfuncp != 0)
break;
}
umem_free(lbuf2, SPA_MAXBLOCKSIZE);
if (lsize > maxlsize) {
exceeded = B_TRUE;
}
if (*cfuncp == ZIO_COMPRESS_ZLE) {
printf("\nZLE decompression was selected. If you "
"suspect the results are wrong,\ntry avoiding ZLE "
"by setting and exporting ZDB_NO_ZLE=\"true\"\n");
}
return (exceeded);
}
/*
* Read a block from a pool and print it out. The syntax of the
* block descriptor is:
*
* pool:vdev_specifier:offset:[lsize/]psize[:flags]
*
* pool - The name of the pool you wish to read from
* vdev_specifier - Which vdev (see comment for zdb_vdev_lookup)
* offset - offset, in hex, in bytes
* size - Amount of data to read, in hex, in bytes
* flags - A string of characters specifying options
* b: Decode a blkptr at given offset within block
* c: Calculate and display checksums
* d: Decompress data before dumping
* e: Byteswap data before dumping
* g: Display data as a gang block header
* i: Display as an indirect block
* r: Dump raw data to stdout
* v: Verbose
*
*/
static void
zdb_read_block(char *thing, spa_t *spa)
{
blkptr_t blk, *bp = &blk;
dva_t *dva = bp->blk_dva;
int flags = 0;
uint64_t offset = 0, psize = 0, lsize = 0, blkptr_offset = 0;
zio_t *zio;
vdev_t *vd;
abd_t *pabd;
void *lbuf, *buf;
char *s, *p, *dup, *flagstr, *sizes, *tmp = NULL;
const char *vdev, *errmsg = NULL;
int i, error;
boolean_t borrowed = B_FALSE, found = B_FALSE;
dup = strdup(thing);
s = strtok_r(dup, ":", &tmp);
vdev = s ?: "";
s = strtok_r(NULL, ":", &tmp);
offset = strtoull(s ? s : "", NULL, 16);
sizes = strtok_r(NULL, ":", &tmp);
s = strtok_r(NULL, ":", &tmp);
flagstr = strdup(s ?: "");
if (!zdb_parse_block_sizes(sizes, &lsize, &psize))
errmsg = "invalid size(s)";
if (!IS_P2ALIGNED(psize, DEV_BSIZE) || !IS_P2ALIGNED(lsize, DEV_BSIZE))
errmsg = "size must be a multiple of sector size";
if (!IS_P2ALIGNED(offset, DEV_BSIZE))
errmsg = "offset must be a multiple of sector size";
if (errmsg) {
(void) printf("Invalid block specifier: %s - %s\n",
thing, errmsg);
goto done;
}
tmp = NULL;
for (s = strtok_r(flagstr, ":", &tmp);
s != NULL;
s = strtok_r(NULL, ":", &tmp)) {
for (i = 0; i < strlen(flagstr); i++) {
int bit = flagbits[(uchar_t)flagstr[i]];
if (bit == 0) {
(void) printf("***Ignoring flag: %c\n",
(uchar_t)flagstr[i]);
continue;
}
found = B_TRUE;
flags |= bit;
p = &flagstr[i + 1];
if (*p != ':' && *p != '\0') {
int j = 0, nextbit = flagbits[(uchar_t)*p];
char *end, offstr[8] = { 0 };
if ((bit == ZDB_FLAG_PRINT_BLKPTR) &&
(nextbit == 0)) {
/* look ahead to isolate the offset */
while (nextbit == 0 &&
strchr(flagbitstr, *p) == NULL) {
offstr[j] = *p;
j++;
if (i + j > strlen(flagstr))
break;
p++;
nextbit = flagbits[(uchar_t)*p];
}
blkptr_offset = strtoull(offstr, &end,
16);
i += j;
} else if (nextbit == 0) {
(void) printf("***Ignoring flag arg:"
" '%c'\n", (uchar_t)*p);
}
}
}
}
if (blkptr_offset % sizeof (blkptr_t)) {
printf("Block pointer offset 0x%llx "
"must be divisible by 0x%x\n",
(longlong_t)blkptr_offset, (int)sizeof (blkptr_t));
goto done;
}
if (found == B_FALSE && strlen(flagstr) > 0) {
printf("Invalid flag arg: '%s'\n", flagstr);
goto done;
}
vd = zdb_vdev_lookup(spa->spa_root_vdev, vdev);
if (vd == NULL) {
(void) printf("***Invalid vdev: %s\n", vdev);
goto done;
} else {
if (vd->vdev_path)
(void) fprintf(stderr, "Found vdev: %s\n",
vd->vdev_path);
else
(void) fprintf(stderr, "Found vdev type: %s\n",
vd->vdev_ops->vdev_op_type);
}
pabd = abd_alloc_for_io(SPA_MAXBLOCKSIZE, B_FALSE);
lbuf = umem_alloc(SPA_MAXBLOCKSIZE, UMEM_NOFAIL);
BP_ZERO(bp);
DVA_SET_VDEV(&dva[0], vd->vdev_id);
DVA_SET_OFFSET(&dva[0], offset);
DVA_SET_GANG(&dva[0], !!(flags & ZDB_FLAG_GBH));
DVA_SET_ASIZE(&dva[0], vdev_psize_to_asize(vd, psize));
BP_SET_BIRTH(bp, TXG_INITIAL, TXG_INITIAL);
BP_SET_LSIZE(bp, lsize);
BP_SET_PSIZE(bp, psize);
BP_SET_COMPRESS(bp, ZIO_COMPRESS_OFF);
BP_SET_CHECKSUM(bp, ZIO_CHECKSUM_OFF);
BP_SET_TYPE(bp, DMU_OT_NONE);
BP_SET_LEVEL(bp, 0);
BP_SET_DEDUP(bp, 0);
BP_SET_BYTEORDER(bp, ZFS_HOST_BYTEORDER);
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
zio = zio_root(spa, NULL, NULL, 0);
if (vd == vd->vdev_top) {
/*
* Treat this as a normal block read.
*/
zio_nowait(zio_read(zio, spa, bp, pabd, psize, NULL, NULL,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW, NULL));
} else {
/*
* Treat this as a vdev child I/O.
*/
zio_nowait(zio_vdev_child_io(zio, bp, vd, offset, pabd,
psize, ZIO_TYPE_READ, ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_PROPAGATE |
ZIO_FLAG_DONT_RETRY | ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW |
ZIO_FLAG_OPTIONAL, NULL, NULL));
}
error = zio_wait(zio);
spa_config_exit(spa, SCL_STATE, FTAG);
if (error) {
(void) printf("Read of %s failed, error: %d\n", thing, error);
goto out;
}
uint64_t orig_lsize = lsize;
buf = lbuf;
if (flags & ZDB_FLAG_DECOMPRESS) {
boolean_t failed = zdb_decompress_block(pabd, buf, lbuf,
lsize, psize, flags);
if (failed) {
(void) printf("Decompress of %s failed\n", thing);
goto out;
}
} else {
buf = abd_borrow_buf_copy(pabd, lsize);
borrowed = B_TRUE;
}
/*
* Try to detect invalid block pointer. If invalid, try
* decompressing.
*/
if ((flags & ZDB_FLAG_PRINT_BLKPTR || flags & ZDB_FLAG_INDIRECT) &&
!(flags & ZDB_FLAG_DECOMPRESS)) {
const blkptr_t *b = (const blkptr_t *)(void *)
((uintptr_t)buf + (uintptr_t)blkptr_offset);
if (zfs_blkptr_verify(spa, b, B_FALSE, BLK_VERIFY_ONLY) ==
B_FALSE) {
abd_return_buf_copy(pabd, buf, lsize);
borrowed = B_FALSE;
buf = lbuf;
boolean_t failed = zdb_decompress_block(pabd, buf,
lbuf, lsize, psize, flags);
b = (const blkptr_t *)(void *)
((uintptr_t)buf + (uintptr_t)blkptr_offset);
if (failed || zfs_blkptr_verify(spa, b, B_FALSE,
BLK_VERIFY_LOG) == B_FALSE) {
printf("invalid block pointer at this DVA\n");
goto out;
}
}
}
if (flags & ZDB_FLAG_PRINT_BLKPTR)
zdb_print_blkptr((blkptr_t *)(void *)
((uintptr_t)buf + (uintptr_t)blkptr_offset), flags);
else if (flags & ZDB_FLAG_RAW)
zdb_dump_block_raw(buf, lsize, flags);
else if (flags & ZDB_FLAG_INDIRECT)
zdb_dump_indirect((blkptr_t *)buf,
orig_lsize / sizeof (blkptr_t), flags);
else if (flags & ZDB_FLAG_GBH)
zdb_dump_gbh(buf, flags);
else
zdb_dump_block(thing, buf, lsize, flags);
/*
* If :c was specified, iterate through the checksum table to
* calculate and display each checksum for our specified
* DVA and length.
*/
if ((flags & ZDB_FLAG_CHECKSUM) && !(flags & ZDB_FLAG_RAW) &&
!(flags & ZDB_FLAG_GBH)) {
zio_t *czio;
(void) printf("\n");
for (enum zio_checksum ck = ZIO_CHECKSUM_LABEL;
ck < ZIO_CHECKSUM_FUNCTIONS; ck++) {
if ((zio_checksum_table[ck].ci_flags &
ZCHECKSUM_FLAG_EMBEDDED) ||
ck == ZIO_CHECKSUM_NOPARITY) {
continue;
}
BP_SET_CHECKSUM(bp, ck);
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
czio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL);
czio->io_bp = bp;
if (vd == vd->vdev_top) {
zio_nowait(zio_read(czio, spa, bp, pabd, psize,
NULL, NULL,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW |
ZIO_FLAG_DONT_RETRY, NULL));
} else {
zio_nowait(zio_vdev_child_io(czio, bp, vd,
offset, pabd, psize, ZIO_TYPE_READ,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_DONT_CACHE |
ZIO_FLAG_DONT_PROPAGATE |
ZIO_FLAG_DONT_RETRY |
ZIO_FLAG_CANFAIL | ZIO_FLAG_RAW |
ZIO_FLAG_SPECULATIVE |
ZIO_FLAG_OPTIONAL, NULL, NULL));
}
error = zio_wait(czio);
if (error == 0 || error == ECKSUM) {
zio_t *ck_zio = zio_root(spa, NULL, NULL, 0);
ck_zio->io_offset =
DVA_GET_OFFSET(&bp->blk_dva[0]);
ck_zio->io_bp = bp;
zio_checksum_compute(ck_zio, ck, pabd, lsize);
printf(
"%12s\t"
"cksum=%016llx:%016llx:%016llx:%016llx\n",
zio_checksum_table[ck].ci_name,
(u_longlong_t)bp->blk_cksum.zc_word[0],
(u_longlong_t)bp->blk_cksum.zc_word[1],
(u_longlong_t)bp->blk_cksum.zc_word[2],
(u_longlong_t)bp->blk_cksum.zc_word[3]);
zio_wait(ck_zio);
} else {
printf("error %d reading block\n", error);
}
spa_config_exit(spa, SCL_STATE, FTAG);
}
}
if (borrowed)
abd_return_buf_copy(pabd, buf, lsize);
out:
abd_free(pabd);
umem_free(lbuf, SPA_MAXBLOCKSIZE);
done:
free(flagstr);
free(dup);
}
static void
zdb_embedded_block(char *thing)
{
blkptr_t bp = {{{{0}}}};
unsigned long long *words = (void *)&bp;
char *buf;
int err;
err = sscanf(thing, "%llx:%llx:%llx:%llx:%llx:%llx:%llx:%llx:"
"%llx:%llx:%llx:%llx:%llx:%llx:%llx:%llx",
words + 0, words + 1, words + 2, words + 3,
words + 4, words + 5, words + 6, words + 7,
words + 8, words + 9, words + 10, words + 11,
words + 12, words + 13, words + 14, words + 15);
if (err != 16) {
(void) fprintf(stderr, "invalid input format\n");
exit(1);
}
ASSERT3U(BPE_GET_LSIZE(&bp), <=, SPA_MAXBLOCKSIZE);
buf = malloc(SPA_MAXBLOCKSIZE);
if (buf == NULL) {
(void) fprintf(stderr, "out of memory\n");
exit(1);
}
err = decode_embedded_bp(&bp, buf, BPE_GET_LSIZE(&bp));
if (err != 0) {
(void) fprintf(stderr, "decode failed: %u\n", err);
exit(1);
}
zdb_dump_block_raw(buf, BPE_GET_LSIZE(&bp), 0);
free(buf);
}
/* check for valid hex or decimal numeric string */
static boolean_t
zdb_numeric(char *str)
{
int i = 0;
if (strlen(str) == 0)
return (B_FALSE);
if (strncmp(str, "0x", 2) == 0 || strncmp(str, "0X", 2) == 0)
i = 2;
for (; i < strlen(str); i++) {
if (!isxdigit(str[i]))
return (B_FALSE);
}
return (B_TRUE);
}
int
main(int argc, char **argv)
{
int c;
spa_t *spa = NULL;
objset_t *os = NULL;
int dump_all = 1;
int verbose = 0;
int error = 0;
char **searchdirs = NULL;
int nsearch = 0;
char *target, *target_pool, dsname[ZFS_MAX_DATASET_NAME_LEN];
nvlist_t *policy = NULL;
uint64_t max_txg = UINT64_MAX;
int64_t objset_id = -1;
uint64_t object;
int flags = ZFS_IMPORT_MISSING_LOG;
int rewind = ZPOOL_NEVER_REWIND;
char *spa_config_path_env, *objset_str;
boolean_t target_is_spa = B_TRUE, dataset_lookup = B_FALSE;
nvlist_t *cfg = NULL;
dprintf_setup(&argc, argv);
/*
* If there is an environment variable SPA_CONFIG_PATH it overrides
* default spa_config_path setting. If -U flag is specified it will
* override this environment variable settings once again.
*/
spa_config_path_env = getenv("SPA_CONFIG_PATH");
if (spa_config_path_env != NULL)
spa_config_path = spa_config_path_env;
/*
* For performance reasons, we set this tunable down. We do so before
* the arg parsing section so that the user can override this value if
* they choose.
*/
zfs_btree_verify_intensity = 3;
struct option long_options[] = {
{"ignore-assertions", no_argument, NULL, 'A'},
{"block-stats", no_argument, NULL, 'b'},
{"checksum", no_argument, NULL, 'c'},
{"config", no_argument, NULL, 'C'},
{"datasets", no_argument, NULL, 'd'},
{"dedup-stats", no_argument, NULL, 'D'},
{"exported", no_argument, NULL, 'e'},
{"embedded-block-pointer", no_argument, NULL, 'E'},
{"automatic-rewind", no_argument, NULL, 'F'},
{"dump-debug-msg", no_argument, NULL, 'G'},
{"history", no_argument, NULL, 'h'},
{"intent-logs", no_argument, NULL, 'i'},
{"inflight", required_argument, NULL, 'I'},
{"checkpointed-state", no_argument, NULL, 'k'},
{"key", required_argument, NULL, 'K'},
{"label", no_argument, NULL, 'l'},
{"disable-leak-tracking", no_argument, NULL, 'L'},
{"metaslabs", no_argument, NULL, 'm'},
{"metaslab-groups", no_argument, NULL, 'M'},
{"numeric", no_argument, NULL, 'N'},
{"option", required_argument, NULL, 'o'},
{"object-lookups", no_argument, NULL, 'O'},
{"path", required_argument, NULL, 'p'},
{"parseable", no_argument, NULL, 'P'},
{"skip-label", no_argument, NULL, 'q'},
{"copy-object", no_argument, NULL, 'r'},
{"read-block", no_argument, NULL, 'R'},
{"io-stats", no_argument, NULL, 's'},
{"simulate-dedup", no_argument, NULL, 'S'},
{"txg", required_argument, NULL, 't'},
{"uberblock", no_argument, NULL, 'u'},
{"cachefile", required_argument, NULL, 'U'},
{"verbose", no_argument, NULL, 'v'},
{"verbatim", no_argument, NULL, 'V'},
{"dump-blocks", required_argument, NULL, 'x'},
{"extreme-rewind", no_argument, NULL, 'X'},
{"all-reconstruction", no_argument, NULL, 'Y'},
{"livelist", no_argument, NULL, 'y'},
{"zstd-headers", no_argument, NULL, 'Z'},
{0, 0, 0, 0}
};
while ((c = getopt_long(argc, argv,
"AbcCdDeEFGhiI:kK:lLmMNo:Op:PqrRsSt:uU:vVx:XYyZ",
long_options, NULL)) != -1) {
switch (c) {
case 'b':
case 'c':
case 'C':
case 'd':
case 'D':
case 'E':
case 'G':
case 'h':
case 'i':
case 'l':
case 'm':
case 'M':
case 'N':
case 'O':
case 'r':
case 'R':
case 's':
case 'S':
case 'u':
case 'y':
case 'Z':
dump_opt[c]++;
dump_all = 0;
break;
case 'A':
case 'e':
case 'F':
case 'k':
case 'L':
case 'P':
case 'q':
case 'X':
dump_opt[c]++;
break;
case 'Y':
zfs_reconstruct_indirect_combinations_max = INT_MAX;
zfs_deadman_enabled = 0;
break;
/* NB: Sort single match options below. */
case 'I':
max_inflight_bytes = strtoull(optarg, NULL, 0);
if (max_inflight_bytes == 0) {
(void) fprintf(stderr, "maximum number "
"of inflight bytes must be greater "
"than 0\n");
usage();
}
break;
case 'K':
dump_opt[c]++;
key_material = strdup(optarg);
/* redact key material in process table */
while (*optarg != '\0') { *optarg++ = '*'; }
break;
case 'o':
error = set_global_var(optarg);
if (error != 0)
usage();
break;
case 'p':
if (searchdirs == NULL) {
searchdirs = umem_alloc(sizeof (char *),
UMEM_NOFAIL);
} else {
char **tmp = umem_alloc((nsearch + 1) *
sizeof (char *), UMEM_NOFAIL);
memcpy(tmp, searchdirs, nsearch *
sizeof (char *));
umem_free(searchdirs,
nsearch * sizeof (char *));
searchdirs = tmp;
}
searchdirs[nsearch++] = optarg;
break;
case 't':
max_txg = strtoull(optarg, NULL, 0);
if (max_txg < TXG_INITIAL) {
(void) fprintf(stderr, "incorrect txg "
"specified: %s\n", optarg);
usage();
}
break;
case 'U':
spa_config_path = optarg;
if (spa_config_path[0] != '/') {
(void) fprintf(stderr,
"cachefile must be an absolute path "
"(i.e. start with a slash)\n");
usage();
}
break;
case 'v':
verbose++;
break;
case 'V':
flags = ZFS_IMPORT_VERBATIM;
break;
case 'x':
vn_dumpdir = optarg;
break;
default:
usage();
break;
}
}
if (!dump_opt['e'] && searchdirs != NULL) {
(void) fprintf(stderr, "-p option requires use of -e\n");
usage();
}
#if defined(_LP64)
/*
* ZDB does not typically re-read blocks; therefore limit the ARC
* to 256 MB, which can be used entirely for metadata.
*/
zfs_arc_min = 2ULL << SPA_MAXBLOCKSHIFT;
zfs_arc_max = 256 * 1024 * 1024;
#endif
/*
* "zdb -c" uses checksum-verifying scrub i/os which are async reads.
* "zdb -b" uses traversal prefetch which uses async reads.
* For good performance, let several of them be active at once.
*/
zfs_vdev_async_read_max_active = 10;
/*
* Disable reference tracking for better performance.
*/
reference_tracking_enable = B_FALSE;
/*
* Do not fail spa_load when spa_load_verify fails. This is needed
* to load non-idle pools.
*/
spa_load_verify_dryrun = B_TRUE;
/*
* ZDB should have ability to read spacemaps.
*/
spa_mode_readable_spacemaps = B_TRUE;
kernel_init(SPA_MODE_READ);
if (dump_all)
verbose = MAX(verbose, 1);
for (c = 0; c < 256; c++) {
if (dump_all && strchr("AeEFkKlLNOPrRSXy", c) == NULL)
dump_opt[c] = 1;
if (dump_opt[c])
dump_opt[c] += verbose;
}
libspl_set_assert_ok((dump_opt['A'] == 1) || (dump_opt['A'] > 2));
zfs_recover = (dump_opt['A'] > 1);
argc -= optind;
argv += optind;
if (argc < 2 && dump_opt['R'])
usage();
if (dump_opt['E']) {
if (argc != 1)
usage();
zdb_embedded_block(argv[0]);
return (0);
}
if (argc < 1) {
if (!dump_opt['e'] && dump_opt['C']) {
dump_cachefile(spa_config_path);
return (0);
}
usage();
}
if (dump_opt['l'])
return (dump_label(argv[0]));
if (dump_opt['O']) {
if (argc != 2)
usage();
dump_opt['v'] = verbose + 3;
return (dump_path(argv[0], argv[1], NULL));
}
if (dump_opt['r']) {
target_is_spa = B_FALSE;
if (argc != 3)
usage();
dump_opt['v'] = verbose;
error = dump_path(argv[0], argv[1], &object);
if (error != 0)
fatal("internal error: %s", strerror(error));
}
if (dump_opt['X'] || dump_opt['F'])
rewind = ZPOOL_DO_REWIND |
(dump_opt['X'] ? ZPOOL_EXTREME_REWIND : 0);
/* -N implies -d */
if (dump_opt['N'] && dump_opt['d'] == 0)
dump_opt['d'] = dump_opt['N'];
if (nvlist_alloc(&policy, NV_UNIQUE_NAME_TYPE, 0) != 0 ||
nvlist_add_uint64(policy, ZPOOL_LOAD_REQUEST_TXG, max_txg) != 0 ||
nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY, rewind) != 0)
fatal("internal error: %s", strerror(ENOMEM));
error = 0;
target = argv[0];
if (strpbrk(target, "/@") != NULL) {
size_t targetlen;
target_pool = strdup(target);
*strpbrk(target_pool, "/@") = '\0';
target_is_spa = B_FALSE;
targetlen = strlen(target);
if (targetlen && target[targetlen - 1] == '/')
target[targetlen - 1] = '\0';
/*
* See if an objset ID was supplied (-d <pool>/<objset ID>).
* To disambiguate tank/100, consider the 100 as objsetID
* if -N was given, otherwise 100 is an objsetID iff
* tank/100 as a named dataset fails on lookup.
*/
objset_str = strchr(target, '/');
if (objset_str && strlen(objset_str) > 1 &&
zdb_numeric(objset_str + 1)) {
char *endptr;
errno = 0;
objset_str++;
objset_id = strtoull(objset_str, &endptr, 0);
/* dataset 0 is the same as opening the pool */
if (errno == 0 && endptr != objset_str &&
objset_id != 0) {
if (dump_opt['N'])
dataset_lookup = B_TRUE;
}
/* normal dataset name not an objset ID */
if (endptr == objset_str) {
objset_id = -1;
}
} else if (objset_str && !zdb_numeric(objset_str + 1) &&
dump_opt['N']) {
printf("Supply a numeric objset ID with -N\n");
exit(1);
}
} else {
target_pool = target;
}
if (dump_opt['e']) {
importargs_t args = { 0 };
args.paths = nsearch;
args.path = searchdirs;
args.can_be_active = B_TRUE;
libpc_handle_t lpch = {
.lpc_lib_handle = NULL,
.lpc_ops = &libzpool_config_ops,
.lpc_printerr = B_TRUE
};
error = zpool_find_config(&lpch, target_pool, &cfg, &args);
if (error == 0) {
if (nvlist_add_nvlist(cfg,
ZPOOL_LOAD_POLICY, policy) != 0) {
fatal("can't open '%s': %s",
target, strerror(ENOMEM));
}
if (dump_opt['C'] > 1) {
(void) printf("\nConfiguration for import:\n");
dump_nvlist(cfg, 8);
}
/*
* Disable the activity check to allow examination of
* active pools.
*/
error = spa_import(target_pool, cfg, NULL,
flags | ZFS_IMPORT_SKIP_MMP);
}
}
if (searchdirs != NULL) {
umem_free(searchdirs, nsearch * sizeof (char *));
searchdirs = NULL;
}
/*
* import_checkpointed_state makes the assumption that the
* target pool that we pass it is already part of the spa
* namespace. Because of that we need to make sure to call
* it always after the -e option has been processed, which
* imports the pool to the namespace if it's not in the
* cachefile.
*/
char *checkpoint_pool = NULL;
char *checkpoint_target = NULL;
if (dump_opt['k']) {
checkpoint_pool = import_checkpointed_state(target, cfg,
&checkpoint_target);
if (checkpoint_target != NULL)
target = checkpoint_target;
}
if (cfg != NULL) {
nvlist_free(cfg);
cfg = NULL;
}
if (target_pool != target)
free(target_pool);
if (error == 0) {
if (dump_opt['k'] && (target_is_spa || dump_opt['R'])) {
ASSERT(checkpoint_pool != NULL);
ASSERT(checkpoint_target == NULL);
error = spa_open(checkpoint_pool, &spa, FTAG);
if (error != 0) {
fatal("Tried to open pool \"%s\" but "
"spa_open() failed with error %d\n",
checkpoint_pool, error);
}
} else if (target_is_spa || dump_opt['R'] || objset_id == 0) {
zdb_set_skip_mmp(target);
error = spa_open_rewind(target, &spa, FTAG, policy,
NULL);
if (error) {
/*
* If we're missing the log device then
* try opening the pool after clearing the
* log state.
*/
mutex_enter(&spa_namespace_lock);
if ((spa = spa_lookup(target)) != NULL &&
spa->spa_log_state == SPA_LOG_MISSING) {
spa->spa_log_state = SPA_LOG_CLEAR;
error = 0;
}
mutex_exit(&spa_namespace_lock);
if (!error) {
error = spa_open_rewind(target, &spa,
FTAG, policy, NULL);
}
}
} else if (strpbrk(target, "#") != NULL) {
dsl_pool_t *dp;
error = dsl_pool_hold(target, FTAG, &dp);
if (error != 0) {
fatal("can't dump '%s': %s", target,
strerror(error));
}
error = dump_bookmark(dp, target, B_TRUE, verbose > 1);
dsl_pool_rele(dp, FTAG);
if (error != 0) {
fatal("can't dump '%s': %s", target,
strerror(error));
}
return (error);
} else {
target_pool = strdup(target);
if (strpbrk(target, "/@") != NULL)
*strpbrk(target_pool, "/@") = '\0';
zdb_set_skip_mmp(target);
/*
* If -N was supplied, the user has indicated that
* zdb -d <pool>/<objsetID> is in effect. Otherwise
* we first assume that the dataset string is the
* dataset name. If dmu_objset_hold fails with the
* dataset string, and we have an objset_id, retry the
* lookup with the objsetID.
*/
boolean_t retry = B_TRUE;
retry_lookup:
if (dataset_lookup == B_TRUE) {
/*
* Use the supplied id to get the name
* for open_objset.
*/
error = spa_open(target_pool, &spa, FTAG);
if (error == 0) {
error = name_from_objset_id(spa,
objset_id, dsname);
spa_close(spa, FTAG);
if (error == 0)
target = dsname;
}
}
if (error == 0) {
if (objset_id > 0 && retry) {
int err = dmu_objset_hold(target, FTAG,
&os);
if (err) {
dataset_lookup = B_TRUE;
retry = B_FALSE;
goto retry_lookup;
} else {
dmu_objset_rele(os, FTAG);
}
}
error = open_objset(target, FTAG, &os);
}
if (error == 0)
spa = dmu_objset_spa(os);
free(target_pool);
}
}
nvlist_free(policy);
if (error)
fatal("can't open '%s': %s", target, strerror(error));
/*
* Set the pool failure mode to panic in order to prevent the pool
* from suspending. A suspended I/O will have no way to resume and
* can prevent the zdb(8) command from terminating as expected.
*/
if (spa != NULL)
spa->spa_failmode = ZIO_FAILURE_MODE_PANIC;
argv++;
argc--;
if (dump_opt['r']) {
error = zdb_copy_object(os, object, argv[1]);
} else if (!dump_opt['R']) {
flagbits['d'] = ZOR_FLAG_DIRECTORY;
flagbits['f'] = ZOR_FLAG_PLAIN_FILE;
flagbits['m'] = ZOR_FLAG_SPACE_MAP;
flagbits['z'] = ZOR_FLAG_ZAP;
flagbits['A'] = ZOR_FLAG_ALL_TYPES;
if (argc > 0 && dump_opt['d']) {
zopt_object_args = argc;
zopt_object_ranges = calloc(zopt_object_args,
sizeof (zopt_object_range_t));
for (unsigned i = 0; i < zopt_object_args; i++) {
int err;
const char *msg = NULL;
err = parse_object_range(argv[i],
&zopt_object_ranges[i], &msg);
if (err != 0)
fatal("Bad object or range: '%s': %s\n",
argv[i], msg ?: "");
}
} else if (argc > 0 && dump_opt['m']) {
zopt_metaslab_args = argc;
zopt_metaslab = calloc(zopt_metaslab_args,
sizeof (uint64_t));
for (unsigned i = 0; i < zopt_metaslab_args; i++) {
errno = 0;
zopt_metaslab[i] = strtoull(argv[i], NULL, 0);
if (zopt_metaslab[i] == 0 && errno != 0)
fatal("bad number %s: %s", argv[i],
strerror(errno));
}
}
if (os != NULL) {
dump_objset(os);
} else if (zopt_object_args > 0 && !dump_opt['m']) {
dump_objset(spa->spa_meta_objset);
} else {
dump_zpool(spa);
}
} else {
flagbits['b'] = ZDB_FLAG_PRINT_BLKPTR;
flagbits['c'] = ZDB_FLAG_CHECKSUM;
flagbits['d'] = ZDB_FLAG_DECOMPRESS;
flagbits['e'] = ZDB_FLAG_BSWAP;
flagbits['g'] = ZDB_FLAG_GBH;
flagbits['i'] = ZDB_FLAG_INDIRECT;
flagbits['r'] = ZDB_FLAG_RAW;
flagbits['v'] = ZDB_FLAG_VERBOSE;
for (int i = 0; i < argc; i++)
zdb_read_block(argv[i], spa);
}
if (dump_opt['k']) {
free(checkpoint_pool);
if (!target_is_spa)
free(checkpoint_target);
}
if (os != NULL) {
close_objset(os, FTAG);
} else {
spa_close(spa, FTAG);
}
fuid_table_destroy();
dump_debug_buffer();
kernel_fini();
return (error);
}
diff --git a/sys/contrib/openzfs/cmd/zfs/zfs_iter.c b/sys/contrib/openzfs/cmd/zfs/zfs_iter.c
index 6665627d43e3..202cb0e82b5a 100644
--- a/sys/contrib/openzfs/cmd/zfs/zfs_iter.c
+++ b/sys/contrib/openzfs/cmd/zfs/zfs_iter.c
@@ -1,554 +1,554 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012 Pawel Jakub Dawidek <pawel@dawidek.net>.
* Copyright 2013 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2013 by Delphix. All rights reserved.
*/
#include <libintl.h>
#include <libuutil.h>
#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <libzfs.h>
#include "zfs_util.h"
#include "zfs_iter.h"
/*
* This is a private interface used to gather up all the datasets specified on
* the command line so that we can iterate over them in order.
*
* First, we iterate over all filesystems, gathering them together into an
* AVL tree. We report errors for any explicitly specified datasets
* that we couldn't open.
*
* When finished, we have an AVL tree of ZFS handles. We go through and execute
* the provided callback for each one, passing whatever data the user supplied.
*/
typedef struct zfs_node {
zfs_handle_t *zn_handle;
uu_avl_node_t zn_avlnode;
} zfs_node_t;
typedef struct callback_data {
uu_avl_t *cb_avl;
int cb_flags;
zfs_type_t cb_types;
zfs_sort_column_t *cb_sortcol;
zprop_list_t **cb_proplist;
int cb_depth_limit;
int cb_depth;
uint8_t cb_props_table[ZFS_NUM_PROPS];
} callback_data_t;
uu_avl_pool_t *avl_pool;
/*
* Include snaps if they were requested or if this a zfs list where types
* were not specified and the "listsnapshots" property is set on this pool.
*/
static boolean_t
zfs_include_snapshots(zfs_handle_t *zhp, callback_data_t *cb)
{
zpool_handle_t *zph;
if ((cb->cb_flags & ZFS_ITER_PROP_LISTSNAPS) == 0)
return (cb->cb_types & ZFS_TYPE_SNAPSHOT);
zph = zfs_get_pool_handle(zhp);
return (zpool_get_prop_int(zph, ZPOOL_PROP_LISTSNAPS, NULL));
}
/*
* Called for each dataset. If the object is of an appropriate type,
* add it to the avl tree and recurse over any children as necessary.
*/
static int
zfs_callback(zfs_handle_t *zhp, void *data)
{
callback_data_t *cb = data;
boolean_t should_close = B_TRUE;
boolean_t include_snaps = zfs_include_snapshots(zhp, cb);
boolean_t include_bmarks = (cb->cb_types & ZFS_TYPE_BOOKMARK);
if ((zfs_get_type(zhp) & cb->cb_types) ||
((zfs_get_type(zhp) == ZFS_TYPE_SNAPSHOT) && include_snaps)) {
uu_avl_index_t idx;
zfs_node_t *node = safe_malloc(sizeof (zfs_node_t));
node->zn_handle = zhp;
uu_avl_node_init(node, &node->zn_avlnode, avl_pool);
if (uu_avl_find(cb->cb_avl, node, cb->cb_sortcol,
&idx) == NULL) {
if (cb->cb_proplist) {
if ((*cb->cb_proplist) &&
!(*cb->cb_proplist)->pl_all)
zfs_prune_proplist(zhp,
cb->cb_props_table);
if (zfs_expand_proplist(zhp, cb->cb_proplist,
(cb->cb_flags & ZFS_ITER_RECVD_PROPS),
(cb->cb_flags & ZFS_ITER_LITERAL_PROPS))
!= 0) {
free(node);
return (-1);
}
}
uu_avl_insert(cb->cb_avl, node, idx);
should_close = B_FALSE;
} else {
free(node);
}
}
/*
* Recurse if necessary.
*/
if (cb->cb_flags & ZFS_ITER_RECURSE &&
((cb->cb_flags & ZFS_ITER_DEPTH_LIMIT) == 0 ||
cb->cb_depth < cb->cb_depth_limit)) {
cb->cb_depth++;
/*
* If we are not looking for filesystems, we don't need to
* recurse into filesystems when we are at our depth limit.
*/
if ((cb->cb_depth < cb->cb_depth_limit ||
(cb->cb_flags & ZFS_ITER_DEPTH_LIMIT) == 0 ||
(cb->cb_types &
(ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME))) &&
zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) {
- (void) zfs_iter_filesystems(zhp, cb->cb_flags,
+ (void) zfs_iter_filesystems_v2(zhp, cb->cb_flags,
zfs_callback, data);
}
if (((zfs_get_type(zhp) & (ZFS_TYPE_SNAPSHOT |
ZFS_TYPE_BOOKMARK)) == 0) && include_snaps) {
- (void) zfs_iter_snapshots(zhp, cb->cb_flags,
+ (void) zfs_iter_snapshots_v2(zhp, cb->cb_flags,
zfs_callback, data, 0, 0);
}
if (((zfs_get_type(zhp) & (ZFS_TYPE_SNAPSHOT |
ZFS_TYPE_BOOKMARK)) == 0) && include_bmarks) {
- (void) zfs_iter_bookmarks(zhp, cb->cb_flags,
+ (void) zfs_iter_bookmarks_v2(zhp, cb->cb_flags,
zfs_callback, data);
}
cb->cb_depth--;
}
if (should_close)
zfs_close(zhp);
return (0);
}
int
zfs_add_sort_column(zfs_sort_column_t **sc, const char *name,
boolean_t reverse)
{
zfs_sort_column_t *col;
zfs_prop_t prop;
if ((prop = zfs_name_to_prop(name)) == ZPROP_USERPROP &&
!zfs_prop_user(name))
return (-1);
col = safe_malloc(sizeof (zfs_sort_column_t));
col->sc_prop = prop;
col->sc_reverse = reverse;
if (prop == ZPROP_USERPROP) {
col->sc_user_prop = safe_malloc(strlen(name) + 1);
(void) strcpy(col->sc_user_prop, name);
}
if (*sc == NULL) {
col->sc_last = col;
*sc = col;
} else {
(*sc)->sc_last->sc_next = col;
(*sc)->sc_last = col;
}
return (0);
}
void
zfs_free_sort_columns(zfs_sort_column_t *sc)
{
zfs_sort_column_t *col;
while (sc != NULL) {
col = sc->sc_next;
free(sc->sc_user_prop);
free(sc);
sc = col;
}
}
/*
* Return true if all of the properties to be sorted are populated by
* dsl_dataset_fast_stat(). Note that sc == NULL (no sort) means we
* don't need any extra properties, so returns true.
*/
boolean_t
zfs_sort_only_by_fast(const zfs_sort_column_t *sc)
{
while (sc != NULL) {
switch (sc->sc_prop) {
case ZFS_PROP_NAME:
case ZFS_PROP_GUID:
case ZFS_PROP_CREATETXG:
case ZFS_PROP_NUMCLONES:
case ZFS_PROP_INCONSISTENT:
case ZFS_PROP_REDACTED:
case ZFS_PROP_ORIGIN:
break;
default:
return (B_FALSE);
}
sc = sc->sc_next;
}
return (B_TRUE);
}
boolean_t
zfs_list_only_by_fast(const zprop_list_t *p)
{
if (p == NULL) {
/* NULL means 'all' so we can't use simple mode */
return (B_FALSE);
}
while (p != NULL) {
switch (p->pl_prop) {
case ZFS_PROP_NAME:
case ZFS_PROP_GUID:
case ZFS_PROP_CREATETXG:
case ZFS_PROP_NUMCLONES:
case ZFS_PROP_INCONSISTENT:
case ZFS_PROP_REDACTED:
case ZFS_PROP_ORIGIN:
break;
default:
return (B_FALSE);
}
p = p->pl_next;
}
return (B_TRUE);
}
static int
zfs_compare(const void *larg, const void *rarg)
{
zfs_handle_t *l = ((zfs_node_t *)larg)->zn_handle;
zfs_handle_t *r = ((zfs_node_t *)rarg)->zn_handle;
const char *lname = zfs_get_name(l);
const char *rname = zfs_get_name(r);
char *lat, *rat;
uint64_t lcreate, rcreate;
int ret;
lat = (char *)strchr(lname, '@');
rat = (char *)strchr(rname, '@');
if (lat != NULL)
*lat = '\0';
if (rat != NULL)
*rat = '\0';
ret = strcmp(lname, rname);
if (ret == 0 && (lat != NULL || rat != NULL)) {
/*
* If we're comparing a dataset to one of its snapshots, we
* always make the full dataset first.
*/
if (lat == NULL) {
ret = -1;
} else if (rat == NULL) {
ret = 1;
} else {
/*
* If we have two snapshots from the same dataset, then
* we want to sort them according to creation time. We
* use the hidden CREATETXG property to get an absolute
* ordering of snapshots.
*/
lcreate = zfs_prop_get_int(l, ZFS_PROP_CREATETXG);
rcreate = zfs_prop_get_int(r, ZFS_PROP_CREATETXG);
/*
* Both lcreate and rcreate being 0 means we don't have
* properties and we should compare full name.
*/
if (lcreate == 0 && rcreate == 0)
ret = strcmp(lat + 1, rat + 1);
else if (lcreate < rcreate)
ret = -1;
else if (lcreate > rcreate)
ret = 1;
}
}
if (lat != NULL)
*lat = '@';
if (rat != NULL)
*rat = '@';
return (ret);
}
/*
* Sort datasets by specified columns.
*
* o Numeric types sort in ascending order.
* o String types sort in alphabetical order.
* o Types inappropriate for a row sort that row to the literal
* bottom, regardless of the specified ordering.
*
* If no sort columns are specified, or two datasets compare equally
* across all specified columns, they are sorted alphabetically by name
* with snapshots grouped under their parents.
*/
static int
zfs_sort(const void *larg, const void *rarg, void *data)
{
zfs_handle_t *l = ((zfs_node_t *)larg)->zn_handle;
zfs_handle_t *r = ((zfs_node_t *)rarg)->zn_handle;
zfs_sort_column_t *sc = (zfs_sort_column_t *)data;
zfs_sort_column_t *psc;
for (psc = sc; psc != NULL; psc = psc->sc_next) {
char lbuf[ZFS_MAXPROPLEN], rbuf[ZFS_MAXPROPLEN];
const char *lstr, *rstr;
uint64_t lnum = 0, rnum = 0;
boolean_t lvalid, rvalid;
int ret = 0;
/*
* We group the checks below the generic code. If 'lstr' and
* 'rstr' are non-NULL, then we do a string based comparison.
* Otherwise, we compare 'lnum' and 'rnum'.
*/
lstr = rstr = NULL;
if (psc->sc_prop == ZPROP_USERPROP) {
nvlist_t *luser, *ruser;
nvlist_t *lval, *rval;
luser = zfs_get_user_props(l);
ruser = zfs_get_user_props(r);
lvalid = (nvlist_lookup_nvlist(luser,
psc->sc_user_prop, &lval) == 0);
rvalid = (nvlist_lookup_nvlist(ruser,
psc->sc_user_prop, &rval) == 0);
if (lvalid)
verify(nvlist_lookup_string(lval,
ZPROP_VALUE, &lstr) == 0);
if (rvalid)
verify(nvlist_lookup_string(rval,
ZPROP_VALUE, &rstr) == 0);
} else if (psc->sc_prop == ZFS_PROP_NAME) {
lvalid = rvalid = B_TRUE;
(void) strlcpy(lbuf, zfs_get_name(l), sizeof (lbuf));
(void) strlcpy(rbuf, zfs_get_name(r), sizeof (rbuf));
lstr = lbuf;
rstr = rbuf;
} else if (zfs_prop_is_string(psc->sc_prop)) {
lvalid = (zfs_prop_get(l, psc->sc_prop, lbuf,
sizeof (lbuf), NULL, NULL, 0, B_TRUE) == 0);
rvalid = (zfs_prop_get(r, psc->sc_prop, rbuf,
sizeof (rbuf), NULL, NULL, 0, B_TRUE) == 0);
lstr = lbuf;
rstr = rbuf;
} else {
lvalid = zfs_prop_valid_for_type(psc->sc_prop,
zfs_get_type(l), B_FALSE);
rvalid = zfs_prop_valid_for_type(psc->sc_prop,
zfs_get_type(r), B_FALSE);
if (lvalid)
lnum = zfs_prop_get_int(l, psc->sc_prop);
if (rvalid)
rnum = zfs_prop_get_int(r, psc->sc_prop);
}
if (!lvalid && !rvalid)
continue;
else if (!lvalid)
return (1);
else if (!rvalid)
return (-1);
if (lstr)
ret = strcmp(lstr, rstr);
else if (lnum < rnum)
ret = -1;
else if (lnum > rnum)
ret = 1;
if (ret != 0) {
if (psc->sc_reverse == B_TRUE)
ret = (ret < 0) ? 1 : -1;
return (ret);
}
}
return (zfs_compare(larg, rarg));
}
int
zfs_for_each(int argc, char **argv, int flags, zfs_type_t types,
zfs_sort_column_t *sortcol, zprop_list_t **proplist, int limit,
zfs_iter_f callback, void *data)
{
callback_data_t cb = {0};
int ret = 0;
zfs_node_t *node;
uu_avl_walk_t *walk;
avl_pool = uu_avl_pool_create("zfs_pool", sizeof (zfs_node_t),
offsetof(zfs_node_t, zn_avlnode), zfs_sort, UU_DEFAULT);
if (avl_pool == NULL)
nomem();
cb.cb_sortcol = sortcol;
cb.cb_flags = flags;
cb.cb_proplist = proplist;
cb.cb_types = types;
cb.cb_depth_limit = limit;
/*
* If cb_proplist is provided then in the zfs_handles created we
* retain only those properties listed in cb_proplist and sortcol.
* The rest are pruned. So, the caller should make sure that no other
* properties other than those listed in cb_proplist/sortcol are
* accessed.
*
* If cb_proplist is NULL then we retain all the properties. We
* always retain the zoned property, which some other properties
* need (userquota & friends), and the createtxg property, which
* we need to sort snapshots.
*/
if (cb.cb_proplist && *cb.cb_proplist) {
zprop_list_t *p = *cb.cb_proplist;
while (p) {
if (p->pl_prop >= ZFS_PROP_TYPE &&
p->pl_prop < ZFS_NUM_PROPS) {
cb.cb_props_table[p->pl_prop] = B_TRUE;
}
p = p->pl_next;
}
while (sortcol) {
if (sortcol->sc_prop >= ZFS_PROP_TYPE &&
sortcol->sc_prop < ZFS_NUM_PROPS) {
cb.cb_props_table[sortcol->sc_prop] = B_TRUE;
}
sortcol = sortcol->sc_next;
}
cb.cb_props_table[ZFS_PROP_ZONED] = B_TRUE;
cb.cb_props_table[ZFS_PROP_CREATETXG] = B_TRUE;
} else {
(void) memset(cb.cb_props_table, B_TRUE,
sizeof (cb.cb_props_table));
}
if ((cb.cb_avl = uu_avl_create(avl_pool, NULL, UU_DEFAULT)) == NULL)
nomem();
if (argc == 0) {
/*
* If given no arguments, iterate over all datasets.
*/
cb.cb_flags |= ZFS_ITER_RECURSE;
ret = zfs_iter_root(g_zfs, zfs_callback, &cb);
} else {
zfs_handle_t *zhp = NULL;
zfs_type_t argtype = types;
/*
* If we're recursive, then we always allow filesystems as
* arguments. If we also are interested in snapshots or
* bookmarks, then we can take volumes as well.
*/
if (flags & ZFS_ITER_RECURSE) {
argtype |= ZFS_TYPE_FILESYSTEM;
if (types & (ZFS_TYPE_SNAPSHOT | ZFS_TYPE_BOOKMARK))
argtype |= ZFS_TYPE_VOLUME;
}
for (int i = 0; i < argc; i++) {
if (flags & ZFS_ITER_ARGS_CAN_BE_PATHS) {
zhp = zfs_path_to_zhandle(g_zfs, argv[i],
argtype);
} else {
zhp = zfs_open(g_zfs, argv[i], argtype);
}
if (zhp != NULL)
ret |= zfs_callback(zhp, &cb);
else
ret = 1;
}
}
/*
* At this point we've got our AVL tree full of zfs handles, so iterate
* over each one and execute the real user callback.
*/
for (node = uu_avl_first(cb.cb_avl); node != NULL;
node = uu_avl_next(cb.cb_avl, node))
ret |= callback(node->zn_handle, data);
/*
* Finally, clean up the AVL tree.
*/
if ((walk = uu_avl_walk_start(cb.cb_avl, UU_WALK_ROBUST)) == NULL)
nomem();
while ((node = uu_avl_walk_next(walk)) != NULL) {
uu_avl_remove(cb.cb_avl, node);
zfs_close(node->zn_handle);
free(node);
}
uu_avl_walk_end(walk);
uu_avl_destroy(cb.cb_avl);
uu_avl_pool_destroy(avl_pool);
return (ret);
}
diff --git a/sys/contrib/openzfs/cmd/zfs/zfs_main.c b/sys/contrib/openzfs/cmd/zfs/zfs_main.c
index 4d11042e5cbb..f26adeae70d5 100644
--- a/sys/contrib/openzfs/cmd/zfs/zfs_main.c
+++ b/sys/contrib/openzfs/cmd/zfs/zfs_main.c
@@ -1,8931 +1,8944 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright 2012 Milan Jurik. All rights reserved.
* Copyright (c) 2012, Joyent, Inc. All rights reserved.
* Copyright (c) 2013 Steven Hartland. All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>.
* Copyright 2016 Nexenta Systems, Inc.
* Copyright (c) 2019 Datto Inc.
* Copyright (c) 2019, loli10K <ezomori.nozomu@gmail.com>
* Copyright 2019 Joyent, Inc.
* Copyright (c) 2019, 2020 by Christian Schwarz. All rights reserved.
*/
#include <assert.h>
#include <ctype.h>
#include <sys/debug.h>
#include <errno.h>
#include <getopt.h>
#include <libgen.h>
#include <libintl.h>
#include <libuutil.h>
#include <libnvpair.h>
#include <locale.h>
#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <fcntl.h>
#include <zone.h>
#include <grp.h>
#include <pwd.h>
#include <umem.h>
#include <pthread.h>
#include <signal.h>
#include <sys/list.h>
#include <sys/mkdev.h>
#include <sys/mntent.h>
#include <sys/mnttab.h>
#include <sys/mount.h>
#include <sys/stat.h>
#include <sys/fs/zfs.h>
#include <sys/systeminfo.h>
#include <sys/types.h>
#include <time.h>
#include <sys/zfs_project.h>
#include <libzfs.h>
#include <libzfs_core.h>
#include <zfs_prop.h>
#include <zfs_deleg.h>
#include <libzutil.h>
#ifdef HAVE_IDMAP
#include <aclutils.h>
#include <directory.h>
#endif /* HAVE_IDMAP */
#include "zfs_iter.h"
#include "zfs_util.h"
#include "zfs_comutil.h"
#include "zfs_projectutil.h"
libzfs_handle_t *g_zfs;
static char history_str[HIS_MAX_RECORD_LEN];
static boolean_t log_history = B_TRUE;
static int zfs_do_clone(int argc, char **argv);
static int zfs_do_create(int argc, char **argv);
static int zfs_do_destroy(int argc, char **argv);
static int zfs_do_get(int argc, char **argv);
static int zfs_do_inherit(int argc, char **argv);
static int zfs_do_list(int argc, char **argv);
static int zfs_do_mount(int argc, char **argv);
static int zfs_do_rename(int argc, char **argv);
static int zfs_do_rollback(int argc, char **argv);
static int zfs_do_set(int argc, char **argv);
static int zfs_do_upgrade(int argc, char **argv);
static int zfs_do_snapshot(int argc, char **argv);
static int zfs_do_unmount(int argc, char **argv);
static int zfs_do_share(int argc, char **argv);
static int zfs_do_unshare(int argc, char **argv);
static int zfs_do_send(int argc, char **argv);
static int zfs_do_receive(int argc, char **argv);
static int zfs_do_promote(int argc, char **argv);
static int zfs_do_userspace(int argc, char **argv);
static int zfs_do_allow(int argc, char **argv);
static int zfs_do_unallow(int argc, char **argv);
static int zfs_do_hold(int argc, char **argv);
static int zfs_do_holds(int argc, char **argv);
static int zfs_do_release(int argc, char **argv);
static int zfs_do_diff(int argc, char **argv);
static int zfs_do_bookmark(int argc, char **argv);
static int zfs_do_channel_program(int argc, char **argv);
static int zfs_do_load_key(int argc, char **argv);
static int zfs_do_unload_key(int argc, char **argv);
static int zfs_do_change_key(int argc, char **argv);
static int zfs_do_project(int argc, char **argv);
static int zfs_do_version(int argc, char **argv);
static int zfs_do_redact(int argc, char **argv);
static int zfs_do_wait(int argc, char **argv);
#ifdef __FreeBSD__
static int zfs_do_jail(int argc, char **argv);
static int zfs_do_unjail(int argc, char **argv);
#endif
#ifdef __linux__
static int zfs_do_zone(int argc, char **argv);
static int zfs_do_unzone(int argc, char **argv);
#endif
/*
* Enable a reasonable set of defaults for libumem debugging on DEBUG builds.
*/
#ifdef DEBUG
const char *
_umem_debug_init(void)
{
return ("default,verbose"); /* $UMEM_DEBUG setting */
}
const char *
_umem_logging_init(void)
{
return ("fail,contents"); /* $UMEM_LOGGING setting */
}
#endif
typedef enum {
HELP_CLONE,
HELP_CREATE,
HELP_DESTROY,
HELP_GET,
HELP_INHERIT,
HELP_UPGRADE,
HELP_LIST,
HELP_MOUNT,
HELP_PROMOTE,
HELP_RECEIVE,
HELP_RENAME,
HELP_ROLLBACK,
HELP_SEND,
HELP_SET,
HELP_SHARE,
HELP_SNAPSHOT,
HELP_UNMOUNT,
HELP_UNSHARE,
HELP_ALLOW,
HELP_UNALLOW,
HELP_USERSPACE,
HELP_GROUPSPACE,
HELP_PROJECTSPACE,
HELP_PROJECT,
HELP_HOLD,
HELP_HOLDS,
HELP_RELEASE,
HELP_DIFF,
HELP_BOOKMARK,
HELP_CHANNEL_PROGRAM,
HELP_LOAD_KEY,
HELP_UNLOAD_KEY,
HELP_CHANGE_KEY,
HELP_VERSION,
HELP_REDACT,
HELP_JAIL,
HELP_UNJAIL,
HELP_WAIT,
HELP_ZONE,
HELP_UNZONE,
} zfs_help_t;
typedef struct zfs_command {
const char *name;
int (*func)(int argc, char **argv);
zfs_help_t usage;
} zfs_command_t;
/*
* Master command table. Each ZFS command has a name, associated function, and
* usage message. The usage messages need to be internationalized, so we have
* to have a function to return the usage message based on a command index.
*
* These commands are organized according to how they are displayed in the usage
* message. An empty command (one with a NULL name) indicates an empty line in
* the generic usage message.
*/
static zfs_command_t command_table[] = {
{ "version", zfs_do_version, HELP_VERSION },
{ NULL },
{ "create", zfs_do_create, HELP_CREATE },
{ "destroy", zfs_do_destroy, HELP_DESTROY },
{ NULL },
{ "snapshot", zfs_do_snapshot, HELP_SNAPSHOT },
{ "rollback", zfs_do_rollback, HELP_ROLLBACK },
{ "clone", zfs_do_clone, HELP_CLONE },
{ "promote", zfs_do_promote, HELP_PROMOTE },
{ "rename", zfs_do_rename, HELP_RENAME },
{ "bookmark", zfs_do_bookmark, HELP_BOOKMARK },
{ "program", zfs_do_channel_program, HELP_CHANNEL_PROGRAM },
{ NULL },
{ "list", zfs_do_list, HELP_LIST },
{ NULL },
{ "set", zfs_do_set, HELP_SET },
{ "get", zfs_do_get, HELP_GET },
{ "inherit", zfs_do_inherit, HELP_INHERIT },
{ "upgrade", zfs_do_upgrade, HELP_UPGRADE },
{ NULL },
{ "userspace", zfs_do_userspace, HELP_USERSPACE },
{ "groupspace", zfs_do_userspace, HELP_GROUPSPACE },
{ "projectspace", zfs_do_userspace, HELP_PROJECTSPACE },
{ NULL },
{ "project", zfs_do_project, HELP_PROJECT },
{ NULL },
{ "mount", zfs_do_mount, HELP_MOUNT },
{ "unmount", zfs_do_unmount, HELP_UNMOUNT },
{ "share", zfs_do_share, HELP_SHARE },
{ "unshare", zfs_do_unshare, HELP_UNSHARE },
{ NULL },
{ "send", zfs_do_send, HELP_SEND },
{ "receive", zfs_do_receive, HELP_RECEIVE },
{ NULL },
{ "allow", zfs_do_allow, HELP_ALLOW },
{ NULL },
{ "unallow", zfs_do_unallow, HELP_UNALLOW },
{ NULL },
{ "hold", zfs_do_hold, HELP_HOLD },
{ "holds", zfs_do_holds, HELP_HOLDS },
{ "release", zfs_do_release, HELP_RELEASE },
{ "diff", zfs_do_diff, HELP_DIFF },
{ "load-key", zfs_do_load_key, HELP_LOAD_KEY },
{ "unload-key", zfs_do_unload_key, HELP_UNLOAD_KEY },
{ "change-key", zfs_do_change_key, HELP_CHANGE_KEY },
{ "redact", zfs_do_redact, HELP_REDACT },
{ "wait", zfs_do_wait, HELP_WAIT },
#ifdef __FreeBSD__
{ "jail", zfs_do_jail, HELP_JAIL },
{ "unjail", zfs_do_unjail, HELP_UNJAIL },
#endif
#ifdef __linux__
{ "zone", zfs_do_zone, HELP_ZONE },
{ "unzone", zfs_do_unzone, HELP_UNZONE },
#endif
};
#define NCOMMAND (sizeof (command_table) / sizeof (command_table[0]))
zfs_command_t *current_command;
static const char *
get_usage(zfs_help_t idx)
{
switch (idx) {
case HELP_CLONE:
return (gettext("\tclone [-p] [-o property=value] ... "
"<snapshot> <filesystem|volume>\n"));
case HELP_CREATE:
return (gettext("\tcreate [-Pnpuv] [-o property=value] ... "
"<filesystem>\n"
"\tcreate [-Pnpsv] [-b blocksize] [-o property=value] ... "
"-V <size> <volume>\n"));
case HELP_DESTROY:
return (gettext("\tdestroy [-fnpRrv] <filesystem|volume>\n"
"\tdestroy [-dnpRrv] "
"<filesystem|volume>@<snap>[%<snap>][,...]\n"
"\tdestroy <filesystem|volume>#<bookmark>\n"));
case HELP_GET:
return (gettext("\tget [-rHp] [-d max] "
"[-o \"all\" | field[,...]]\n"
"\t [-t type[,...]] [-s source[,...]]\n"
"\t <\"all\" | property[,...]> "
"[filesystem|volume|snapshot|bookmark] ...\n"));
case HELP_INHERIT:
return (gettext("\tinherit [-rS] <property> "
"<filesystem|volume|snapshot> ...\n"));
case HELP_UPGRADE:
return (gettext("\tupgrade [-v]\n"
"\tupgrade [-r] [-V version] <-a | filesystem ...>\n"));
case HELP_LIST:
return (gettext("\tlist [-Hp] [-r|-d max] [-o property[,...]] "
"[-s property]...\n\t [-S property]... [-t type[,...]] "
"[filesystem|volume|snapshot] ...\n"));
case HELP_MOUNT:
return (gettext("\tmount\n"
"\tmount [-flvO] [-o opts] <-a | filesystem>\n"));
case HELP_PROMOTE:
return (gettext("\tpromote <clone-filesystem>\n"));
case HELP_RECEIVE:
return (gettext("\treceive [-vMnsFhu] "
"[-o <property>=<value>] ... [-x <property>] ...\n"
"\t <filesystem|volume|snapshot>\n"
"\treceive [-vMnsFhu] [-o <property>=<value>] ... "
"[-x <property>] ... \n"
"\t [-d | -e] <filesystem>\n"
"\treceive -A <filesystem|volume>\n"));
case HELP_RENAME:
return (gettext("\trename [-f] <filesystem|volume|snapshot> "
"<filesystem|volume|snapshot>\n"
"\trename -p [-f] <filesystem|volume> <filesystem|volume>\n"
"\trename -u [-f] <filesystem> <filesystem>\n"
"\trename -r <snapshot> <snapshot>\n"));
case HELP_ROLLBACK:
return (gettext("\trollback [-rRf] <snapshot>\n"));
case HELP_SEND:
return (gettext("\tsend [-DLPbcehnpsVvw] "
"[-i|-I snapshot]\n"
"\t [-R [-X dataset[,dataset]...]] <snapshot>\n"
"\tsend [-DnVvPLecw] [-i snapshot|bookmark] "
"<filesystem|volume|snapshot>\n"
"\tsend [-DnPpVvLec] [-i bookmark|snapshot] "
"--redact <bookmark> <snapshot>\n"
"\tsend [-nVvPe] -t <receive_resume_token>\n"
"\tsend [-PnVv] --saved filesystem\n"));
case HELP_SET:
return (gettext("\tset <property=value> ... "
"<filesystem|volume|snapshot> ...\n"));
case HELP_SHARE:
return (gettext("\tshare [-l] <-a [nfs|smb] | filesystem>\n"));
case HELP_SNAPSHOT:
return (gettext("\tsnapshot [-r] [-o property=value] ... "
"<filesystem|volume>@<snap> ...\n"));
case HELP_UNMOUNT:
return (gettext("\tunmount [-fu] "
"<-a | filesystem|mountpoint>\n"));
case HELP_UNSHARE:
return (gettext("\tunshare "
"<-a [nfs|smb] | filesystem|mountpoint>\n"));
case HELP_ALLOW:
return (gettext("\tallow <filesystem|volume>\n"
"\tallow [-ldug] "
"<\"everyone\"|user|group>[,...] <perm|@setname>[,...]\n"
"\t <filesystem|volume>\n"
"\tallow [-ld] -e <perm|@setname>[,...] "
"<filesystem|volume>\n"
"\tallow -c <perm|@setname>[,...] <filesystem|volume>\n"
"\tallow -s @setname <perm|@setname>[,...] "
"<filesystem|volume>\n"));
case HELP_UNALLOW:
return (gettext("\tunallow [-rldug] "
"<\"everyone\"|user|group>[,...]\n"
"\t [<perm|@setname>[,...]] <filesystem|volume>\n"
"\tunallow [-rld] -e [<perm|@setname>[,...]] "
"<filesystem|volume>\n"
"\tunallow [-r] -c [<perm|@setname>[,...]] "
"<filesystem|volume>\n"
"\tunallow [-r] -s @setname [<perm|@setname>[,...]] "
"<filesystem|volume>\n"));
case HELP_USERSPACE:
return (gettext("\tuserspace [-Hinp] [-o field[,...]] "
"[-s field] ...\n"
"\t [-S field] ... [-t type[,...]] "
"<filesystem|snapshot|path>\n"));
case HELP_GROUPSPACE:
return (gettext("\tgroupspace [-Hinp] [-o field[,...]] "
"[-s field] ...\n"
"\t [-S field] ... [-t type[,...]] "
"<filesystem|snapshot|path>\n"));
case HELP_PROJECTSPACE:
return (gettext("\tprojectspace [-Hp] [-o field[,...]] "
"[-s field] ... \n"
"\t [-S field] ... <filesystem|snapshot|path>\n"));
case HELP_PROJECT:
return (gettext("\tproject [-d|-r] <directory|file ...>\n"
"\tproject -c [-0] [-d|-r] [-p id] <directory|file ...>\n"
"\tproject -C [-k] [-r] <directory ...>\n"
"\tproject [-p id] [-r] [-s] <directory ...>\n"));
case HELP_HOLD:
return (gettext("\thold [-r] <tag> <snapshot> ...\n"));
case HELP_HOLDS:
return (gettext("\tholds [-rHp] <snapshot> ...\n"));
case HELP_RELEASE:
return (gettext("\trelease [-r] <tag> <snapshot> ...\n"));
case HELP_DIFF:
return (gettext("\tdiff [-FHth] <snapshot> "
"[snapshot|filesystem]\n"));
case HELP_BOOKMARK:
return (gettext("\tbookmark <snapshot|bookmark> "
"<newbookmark>\n"));
case HELP_CHANNEL_PROGRAM:
return (gettext("\tprogram [-jn] [-t <instruction limit>] "
"[-m <memory limit (b)>]\n"
"\t <pool> <program file> [lua args...]\n"));
case HELP_LOAD_KEY:
return (gettext("\tload-key [-rn] [-L <keylocation>] "
"<-a | filesystem|volume>\n"));
case HELP_UNLOAD_KEY:
return (gettext("\tunload-key [-r] "
"<-a | filesystem|volume>\n"));
case HELP_CHANGE_KEY:
return (gettext("\tchange-key [-l] [-o keyformat=<value>]\n"
"\t [-o keylocation=<value>] [-o pbkdf2iters=<value>]\n"
"\t <filesystem|volume>\n"
"\tchange-key -i [-l] <filesystem|volume>\n"));
case HELP_VERSION:
return (gettext("\tversion\n"));
case HELP_REDACT:
return (gettext("\tredact <snapshot> <bookmark> "
"<redaction_snapshot> ...\n"));
case HELP_JAIL:
return (gettext("\tjail <jailid|jailname> <filesystem>\n"));
case HELP_UNJAIL:
return (gettext("\tunjail <jailid|jailname> <filesystem>\n"));
case HELP_WAIT:
return (gettext("\twait [-t <activity>] <filesystem>\n"));
case HELP_ZONE:
return (gettext("\tzone <nsfile> <filesystem>\n"));
case HELP_UNZONE:
return (gettext("\tunzone <nsfile> <filesystem>\n"));
default:
__builtin_unreachable();
}
}
void
nomem(void)
{
(void) fprintf(stderr, gettext("internal error: out of memory\n"));
exit(1);
}
/*
* Utility function to guarantee malloc() success.
*/
void *
safe_malloc(size_t size)
{
void *data;
if ((data = calloc(1, size)) == NULL)
nomem();
return (data);
}
static void *
safe_realloc(void *data, size_t size)
{
void *newp;
if ((newp = realloc(data, size)) == NULL) {
free(data);
nomem();
}
return (newp);
}
static char *
safe_strdup(const char *str)
{
char *dupstr = strdup(str);
if (dupstr == NULL)
nomem();
return (dupstr);
}
/*
* Callback routine that will print out information for each of
* the properties.
*/
static int
usage_prop_cb(int prop, void *cb)
{
FILE *fp = cb;
(void) fprintf(fp, "\t%-15s ", zfs_prop_to_name(prop));
if (zfs_prop_readonly(prop))
(void) fprintf(fp, " NO ");
else
(void) fprintf(fp, "YES ");
if (zfs_prop_inheritable(prop))
(void) fprintf(fp, " YES ");
else
(void) fprintf(fp, " NO ");
(void) fprintf(fp, "%s\n", zfs_prop_values(prop) ?: "-");
return (ZPROP_CONT);
}
/*
* Display usage message. If we're inside a command, display only the usage for
* that command. Otherwise, iterate over the entire command table and display
* a complete usage message.
*/
static __attribute__((noreturn)) void
usage(boolean_t requested)
{
int i;
boolean_t show_properties = B_FALSE;
FILE *fp = requested ? stdout : stderr;
if (current_command == NULL) {
(void) fprintf(fp, gettext("usage: zfs command args ...\n"));
(void) fprintf(fp,
gettext("where 'command' is one of the following:\n\n"));
for (i = 0; i < NCOMMAND; i++) {
if (command_table[i].name == NULL)
(void) fprintf(fp, "\n");
else
(void) fprintf(fp, "%s",
get_usage(command_table[i].usage));
}
(void) fprintf(fp, gettext("\nEach dataset is of the form: "
"pool/[dataset/]*dataset[@name]\n"));
} else {
(void) fprintf(fp, gettext("usage:\n"));
(void) fprintf(fp, "%s", get_usage(current_command->usage));
}
if (current_command != NULL &&
(strcmp(current_command->name, "set") == 0 ||
strcmp(current_command->name, "get") == 0 ||
strcmp(current_command->name, "inherit") == 0 ||
strcmp(current_command->name, "list") == 0))
show_properties = B_TRUE;
if (show_properties) {
(void) fprintf(fp, "%s",
gettext("\nThe following properties are supported:\n"));
(void) fprintf(fp, "\n\t%-14s %s %s %s\n\n",
"PROPERTY", "EDIT", "INHERIT", "VALUES");
/* Iterate over all properties */
(void) zprop_iter(usage_prop_cb, fp, B_FALSE, B_TRUE,
ZFS_TYPE_DATASET);
(void) fprintf(fp, "\t%-15s ", "userused@...");
(void) fprintf(fp, " NO NO <size>\n");
(void) fprintf(fp, "\t%-15s ", "groupused@...");
(void) fprintf(fp, " NO NO <size>\n");
(void) fprintf(fp, "\t%-15s ", "projectused@...");
(void) fprintf(fp, " NO NO <size>\n");
(void) fprintf(fp, "\t%-15s ", "userobjused@...");
(void) fprintf(fp, " NO NO <size>\n");
(void) fprintf(fp, "\t%-15s ", "groupobjused@...");
(void) fprintf(fp, " NO NO <size>\n");
(void) fprintf(fp, "\t%-15s ", "projectobjused@...");
(void) fprintf(fp, " NO NO <size>\n");
(void) fprintf(fp, "\t%-15s ", "userquota@...");
(void) fprintf(fp, "YES NO <size> | none\n");
(void) fprintf(fp, "\t%-15s ", "groupquota@...");
(void) fprintf(fp, "YES NO <size> | none\n");
(void) fprintf(fp, "\t%-15s ", "projectquota@...");
(void) fprintf(fp, "YES NO <size> | none\n");
(void) fprintf(fp, "\t%-15s ", "userobjquota@...");
(void) fprintf(fp, "YES NO <size> | none\n");
(void) fprintf(fp, "\t%-15s ", "groupobjquota@...");
(void) fprintf(fp, "YES NO <size> | none\n");
(void) fprintf(fp, "\t%-15s ", "projectobjquota@...");
(void) fprintf(fp, "YES NO <size> | none\n");
(void) fprintf(fp, "\t%-15s ", "written@<snap>");
(void) fprintf(fp, " NO NO <size>\n");
(void) fprintf(fp, "\t%-15s ", "written#<bookmark>");
(void) fprintf(fp, " NO NO <size>\n");
(void) fprintf(fp, gettext("\nSizes are specified in bytes "
"with standard units such as K, M, G, etc.\n"));
(void) fprintf(fp, "%s", gettext("\nUser-defined properties "
"can be specified by using a name containing a colon "
"(:).\n"));
(void) fprintf(fp, gettext("\nThe {user|group|project}"
"[obj]{used|quota}@ properties must be appended with\n"
"a user|group|project specifier of one of these forms:\n"
" POSIX name (eg: \"matt\")\n"
" POSIX id (eg: \"126829\")\n"
" SMB name@domain (eg: \"matt@sun\")\n"
" SMB SID (eg: \"S-1-234-567-89\")\n"));
} else {
(void) fprintf(fp,
gettext("\nFor the property list, run: %s\n"),
"zfs set|get");
(void) fprintf(fp,
gettext("\nFor the delegated permission list, run: %s\n"),
"zfs allow|unallow");
}
/*
* See comments at end of main().
*/
if (getenv("ZFS_ABORT") != NULL) {
(void) printf("dumping core by request\n");
abort();
}
exit(requested ? 0 : 2);
}
/*
* Take a property=value argument string and add it to the given nvlist.
* Modifies the argument inplace.
*/
static boolean_t
parseprop(nvlist_t *props, char *propname)
{
char *propval;
if ((propval = strchr(propname, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for property=value argument\n"));
return (B_FALSE);
}
*propval = '\0';
propval++;
if (nvlist_exists(props, propname)) {
(void) fprintf(stderr, gettext("property '%s' "
"specified multiple times\n"), propname);
return (B_FALSE);
}
if (nvlist_add_string(props, propname, propval) != 0)
nomem();
return (B_TRUE);
}
/*
* Take a property name argument and add it to the given nvlist.
* Modifies the argument inplace.
*/
static boolean_t
parsepropname(nvlist_t *props, char *propname)
{
if (strchr(propname, '=') != NULL) {
(void) fprintf(stderr, gettext("invalid character "
"'=' in property argument\n"));
return (B_FALSE);
}
if (nvlist_exists(props, propname)) {
(void) fprintf(stderr, gettext("property '%s' "
"specified multiple times\n"), propname);
return (B_FALSE);
}
if (nvlist_add_boolean(props, propname) != 0)
nomem();
return (B_TRUE);
}
static int
parse_depth(char *opt, int *flags)
{
char *tmp;
int depth;
depth = (int)strtol(opt, &tmp, 0);
if (*tmp) {
(void) fprintf(stderr,
gettext("%s is not an integer\n"), optarg);
usage(B_FALSE);
}
if (depth < 0) {
(void) fprintf(stderr,
gettext("Depth can not be negative.\n"));
usage(B_FALSE);
}
*flags |= (ZFS_ITER_DEPTH_LIMIT|ZFS_ITER_RECURSE);
return (depth);
}
#define PROGRESS_DELAY 2 /* seconds */
static const char *pt_reverse =
"\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b";
static time_t pt_begin;
static char *pt_header = NULL;
static boolean_t pt_shown;
static void
start_progress_timer(void)
{
pt_begin = time(NULL) + PROGRESS_DELAY;
pt_shown = B_FALSE;
}
static void
set_progress_header(const char *header)
{
assert(pt_header == NULL);
pt_header = safe_strdup(header);
if (pt_shown) {
(void) printf("%s: ", header);
(void) fflush(stdout);
}
}
static void
update_progress(const char *update)
{
if (!pt_shown && time(NULL) > pt_begin) {
int len = strlen(update);
(void) printf("%s: %s%*.*s", pt_header, update, len, len,
pt_reverse);
(void) fflush(stdout);
pt_shown = B_TRUE;
} else if (pt_shown) {
int len = strlen(update);
(void) printf("%s%*.*s", update, len, len, pt_reverse);
(void) fflush(stdout);
}
}
static void
finish_progress(const char *done)
{
if (pt_shown) {
(void) puts(done);
(void) fflush(stdout);
}
free(pt_header);
pt_header = NULL;
}
static int
zfs_mount_and_share(libzfs_handle_t *hdl, const char *dataset, zfs_type_t type)
{
zfs_handle_t *zhp = NULL;
int ret = 0;
zhp = zfs_open(hdl, dataset, type);
if (zhp == NULL)
return (1);
/*
* Volumes may neither be mounted or shared. Potentially in the
* future filesystems detected on these volumes could be mounted.
*/
if (zfs_get_type(zhp) == ZFS_TYPE_VOLUME) {
zfs_close(zhp);
return (0);
}
/*
* Mount and/or share the new filesystem as appropriate. We provide a
* verbose error message to let the user know that their filesystem was
* in fact created, even if we failed to mount or share it.
*
* If the user doesn't want the dataset automatically mounted, then
* skip the mount/share step
*/
if (zfs_prop_valid_for_type(ZFS_PROP_CANMOUNT, type, B_FALSE) &&
zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) == ZFS_CANMOUNT_ON) {
if (zfs_mount_delegation_check()) {
(void) fprintf(stderr, gettext("filesystem "
"successfully created, but it may only be "
"mounted by root\n"));
ret = 1;
} else if (zfs_mount(zhp, NULL, 0) != 0) {
(void) fprintf(stderr, gettext("filesystem "
"successfully created, but not mounted\n"));
ret = 1;
} else if (zfs_share(zhp, NULL) != 0) {
(void) fprintf(stderr, gettext("filesystem "
"successfully created, but not shared\n"));
ret = 1;
}
zfs_commit_shares(NULL);
}
zfs_close(zhp);
return (ret);
}
/*
* zfs clone [-p] [-o prop=value] ... <snap> <fs | vol>
*
* Given an existing dataset, create a writable copy whose initial contents
* are the same as the source. The newly created dataset maintains a
* dependency on the original; the original cannot be destroyed so long as
* the clone exists.
*
* The '-p' flag creates all the non-existing ancestors of the target first.
*/
static int
zfs_do_clone(int argc, char **argv)
{
zfs_handle_t *zhp = NULL;
boolean_t parents = B_FALSE;
nvlist_t *props;
int ret = 0;
int c;
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
nomem();
/* check options */
while ((c = getopt(argc, argv, "o:p")) != -1) {
switch (c) {
case 'o':
if (!parseprop(props, optarg)) {
nvlist_free(props);
return (1);
}
break;
case 'p':
parents = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
goto usage;
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing source dataset "
"argument\n"));
goto usage;
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing target dataset "
"argument\n"));
goto usage;
}
if (argc > 2) {
(void) fprintf(stderr, gettext("too many arguments\n"));
goto usage;
}
/* open the source dataset */
if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_SNAPSHOT)) == NULL) {
nvlist_free(props);
return (1);
}
if (parents && zfs_name_valid(argv[1], ZFS_TYPE_FILESYSTEM |
ZFS_TYPE_VOLUME)) {
/*
* Now create the ancestors of the target dataset. If the
* target already exists and '-p' option was used we should not
* complain.
*/
if (zfs_dataset_exists(g_zfs, argv[1], ZFS_TYPE_FILESYSTEM |
ZFS_TYPE_VOLUME)) {
zfs_close(zhp);
nvlist_free(props);
return (0);
}
if (zfs_create_ancestors(g_zfs, argv[1]) != 0) {
zfs_close(zhp);
nvlist_free(props);
return (1);
}
}
/* pass to libzfs */
ret = zfs_clone(zhp, argv[1], props);
/* create the mountpoint if necessary */
if (ret == 0) {
if (log_history) {
(void) zpool_log_history(g_zfs, history_str);
log_history = B_FALSE;
}
ret = zfs_mount_and_share(g_zfs, argv[1], ZFS_TYPE_DATASET);
}
zfs_close(zhp);
nvlist_free(props);
return (!!ret);
usage:
ASSERT3P(zhp, ==, NULL);
nvlist_free(props);
usage(B_FALSE);
return (-1);
}
/*
* Return a default volblocksize for the pool which always uses more than
* half of the data sectors. This primarily applies to dRAID which always
* writes full stripe widths.
*/
static uint64_t
default_volblocksize(zpool_handle_t *zhp, nvlist_t *props)
{
uint64_t volblocksize, asize = SPA_MINBLOCKSIZE;
nvlist_t *tree, **vdevs;
uint_t nvdevs;
nvlist_t *config = zpool_get_config(zhp, NULL);
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &tree) != 0 ||
nvlist_lookup_nvlist_array(tree, ZPOOL_CONFIG_CHILDREN,
&vdevs, &nvdevs) != 0) {
return (ZVOL_DEFAULT_BLOCKSIZE);
}
for (int i = 0; i < nvdevs; i++) {
nvlist_t *nv = vdevs[i];
uint64_t ashift, ndata, nparity;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASHIFT, &ashift) != 0)
continue;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DRAID_NDATA,
&ndata) == 0) {
/* dRAID minimum allocation width */
asize = MAX(asize, ndata * (1ULL << ashift));
} else if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY,
&nparity) == 0) {
/* raidz minimum allocation width */
if (nparity == 1)
asize = MAX(asize, 2 * (1ULL << ashift));
else
asize = MAX(asize, 4 * (1ULL << ashift));
} else {
/* mirror or (non-redundant) leaf vdev */
asize = MAX(asize, 1ULL << ashift);
}
}
/*
* Calculate the target volblocksize such that more than half
* of the asize is used. The following table is for 4k sectors.
*
* n asize blksz used | n asize blksz used
* -------------------------+---------------------------------
* 1 4,096 8,192 100% | 9 36,864 32,768 88%
* 2 8,192 8,192 100% | 10 40,960 32,768 80%
* 3 12,288 8,192 66% | 11 45,056 32,768 72%
* 4 16,384 16,384 100% | 12 49,152 32,768 66%
* 5 20,480 16,384 80% | 13 53,248 32,768 61%
* 6 24,576 16,384 66% | 14 57,344 32,768 57%
* 7 28,672 16,384 57% | 15 61,440 32,768 53%
* 8 32,768 32,768 100% | 16 65,536 65,636 100%
*
* This is primarily a concern for dRAID which always allocates
* a full stripe width. For dRAID the default stripe width is
* n=8 in which case the volblocksize is set to 32k. Ignoring
* compression there are no unused sectors. This same reasoning
* applies to raidz[2,3] so target 4 sectors to minimize waste.
*/
uint64_t tgt_volblocksize = ZVOL_DEFAULT_BLOCKSIZE;
while (tgt_volblocksize * 2 <= asize)
tgt_volblocksize *= 2;
const char *prop = zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE);
if (nvlist_lookup_uint64(props, prop, &volblocksize) == 0) {
/* Issue a warning when a non-optimal size is requested. */
if (volblocksize < ZVOL_DEFAULT_BLOCKSIZE) {
(void) fprintf(stderr, gettext("Warning: "
"volblocksize (%llu) is less than the default "
"minimum block size (%llu).\nTo reduce wasted "
"space a volblocksize of %llu is recommended.\n"),
(u_longlong_t)volblocksize,
(u_longlong_t)ZVOL_DEFAULT_BLOCKSIZE,
(u_longlong_t)tgt_volblocksize);
} else if (volblocksize < tgt_volblocksize) {
(void) fprintf(stderr, gettext("Warning: "
"volblocksize (%llu) is much less than the "
"minimum allocation\nunit (%llu), which wastes "
"at least %llu%% of space. To reduce wasted "
"space,\nuse a larger volblocksize (%llu is "
"recommended), fewer dRAID data disks\n"
"per group, or smaller sector size (ashift).\n"),
(u_longlong_t)volblocksize, (u_longlong_t)asize,
(u_longlong_t)((100 * (asize - volblocksize)) /
asize), (u_longlong_t)tgt_volblocksize);
}
} else {
volblocksize = tgt_volblocksize;
fnvlist_add_uint64(props, prop, volblocksize);
}
return (volblocksize);
}
/*
* zfs create [-Pnpv] [-o prop=value] ... fs
* zfs create [-Pnpsv] [-b blocksize] [-o prop=value] ... -V vol size
*
* Create a new dataset. This command can be used to create filesystems
* and volumes. Snapshot creation is handled by 'zfs snapshot'.
* For volumes, the user must specify a size to be used.
*
* The '-s' flag applies only to volumes, and indicates that we should not try
* to set the reservation for this volume. By default we set a reservation
* equal to the size for any volume. For pools with SPA_VERSION >=
* SPA_VERSION_REFRESERVATION, we set a refreservation instead.
*
* The '-p' flag creates all the non-existing ancestors of the target first.
*
* The '-n' flag is no-op (dry run) mode. This will perform a user-space sanity
* check of arguments and properties, but does not check for permissions,
* available space, etc.
*
* The '-u' flag prevents the newly created file system from being mounted.
*
* The '-v' flag is for verbose output.
*
* The '-P' flag is used for parseable output. It implies '-v'.
*/
static int
zfs_do_create(int argc, char **argv)
{
zfs_type_t type = ZFS_TYPE_FILESYSTEM;
zpool_handle_t *zpool_handle = NULL;
nvlist_t *real_props = NULL;
uint64_t volsize = 0;
int c;
boolean_t noreserve = B_FALSE;
boolean_t bflag = B_FALSE;
boolean_t parents = B_FALSE;
boolean_t dryrun = B_FALSE;
boolean_t nomount = B_FALSE;
boolean_t verbose = B_FALSE;
boolean_t parseable = B_FALSE;
int ret = 1;
nvlist_t *props;
uint64_t intval;
const char *strval;
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
nomem();
/* check options */
while ((c = getopt(argc, argv, ":PV:b:nso:puv")) != -1) {
switch (c) {
case 'V':
type = ZFS_TYPE_VOLUME;
if (zfs_nicestrtonum(g_zfs, optarg, &intval) != 0) {
(void) fprintf(stderr, gettext("bad volume "
"size '%s': %s\n"), optarg,
libzfs_error_description(g_zfs));
goto error;
}
if (nvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_VOLSIZE), intval) != 0)
nomem();
volsize = intval;
break;
case 'P':
verbose = B_TRUE;
parseable = B_TRUE;
break;
case 'p':
parents = B_TRUE;
break;
case 'b':
bflag = B_TRUE;
if (zfs_nicestrtonum(g_zfs, optarg, &intval) != 0) {
(void) fprintf(stderr, gettext("bad volume "
"block size '%s': %s\n"), optarg,
libzfs_error_description(g_zfs));
goto error;
}
if (nvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE),
intval) != 0)
nomem();
break;
case 'n':
dryrun = B_TRUE;
break;
case 'o':
if (!parseprop(props, optarg))
goto error;
break;
case 's':
noreserve = B_TRUE;
break;
case 'u':
nomount = B_TRUE;
break;
case 'v':
verbose = B_TRUE;
break;
case ':':
(void) fprintf(stderr, gettext("missing size "
"argument\n"));
goto badusage;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
goto badusage;
}
}
if ((bflag || noreserve) && type != ZFS_TYPE_VOLUME) {
(void) fprintf(stderr, gettext("'-s' and '-b' can only be "
"used when creating a volume\n"));
goto badusage;
}
if (nomount && type != ZFS_TYPE_FILESYSTEM) {
(void) fprintf(stderr, gettext("'-u' can only be "
"used when creating a filesystem\n"));
goto badusage;
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc == 0) {
(void) fprintf(stderr, gettext("missing %s argument\n"),
zfs_type_to_name(type));
goto badusage;
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
goto badusage;
}
if (dryrun || type == ZFS_TYPE_VOLUME) {
char msg[ZFS_MAX_DATASET_NAME_LEN * 2];
char *p;
if ((p = strchr(argv[0], '/')) != NULL)
*p = '\0';
zpool_handle = zpool_open(g_zfs, argv[0]);
if (p != NULL)
*p = '/';
if (zpool_handle == NULL)
goto error;
(void) snprintf(msg, sizeof (msg),
dryrun ? gettext("cannot verify '%s'") :
gettext("cannot create '%s'"), argv[0]);
if (props && (real_props = zfs_valid_proplist(g_zfs, type,
props, 0, NULL, zpool_handle, B_TRUE, msg)) == NULL) {
zpool_close(zpool_handle);
goto error;
}
}
if (type == ZFS_TYPE_VOLUME) {
const char *prop = zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE);
uint64_t volblocksize = default_volblocksize(zpool_handle,
real_props);
if (volblocksize != ZVOL_DEFAULT_BLOCKSIZE &&
nvlist_lookup_string(props, prop, &strval) != 0) {
char *tmp;
if (asprintf(&tmp, "%llu",
(u_longlong_t)volblocksize) == -1)
nomem();
nvlist_add_string(props, prop, tmp);
free(tmp);
}
/*
* If volsize is not a multiple of volblocksize, round it
* up to the nearest multiple of the volblocksize.
*/
if (volsize % volblocksize) {
volsize = P2ROUNDUP_TYPED(volsize, volblocksize,
uint64_t);
if (nvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_VOLSIZE), volsize) != 0) {
nvlist_free(props);
nomem();
}
}
}
if (type == ZFS_TYPE_VOLUME && !noreserve) {
uint64_t spa_version;
zfs_prop_t resv_prop;
spa_version = zpool_get_prop_int(zpool_handle,
ZPOOL_PROP_VERSION, NULL);
if (spa_version >= SPA_VERSION_REFRESERVATION)
resv_prop = ZFS_PROP_REFRESERVATION;
else
resv_prop = ZFS_PROP_RESERVATION;
volsize = zvol_volsize_to_reservation(zpool_handle, volsize,
real_props);
if (nvlist_lookup_string(props, zfs_prop_to_name(resv_prop),
&strval) != 0) {
if (nvlist_add_uint64(props,
zfs_prop_to_name(resv_prop), volsize) != 0) {
nvlist_free(props);
nomem();
}
}
}
if (zpool_handle != NULL) {
zpool_close(zpool_handle);
nvlist_free(real_props);
}
if (parents && zfs_name_valid(argv[0], type)) {
/*
* Now create the ancestors of target dataset. If the target
* already exists and '-p' option was used we should not
* complain.
*/
if (zfs_dataset_exists(g_zfs, argv[0], type)) {
ret = 0;
goto error;
}
if (verbose) {
(void) printf(parseable ? "create_ancestors\t%s\n" :
dryrun ? "would create ancestors of %s\n" :
"create ancestors of %s\n", argv[0]);
}
if (!dryrun) {
if (zfs_create_ancestors(g_zfs, argv[0]) != 0) {
goto error;
}
}
}
if (verbose) {
nvpair_t *nvp = NULL;
(void) printf(parseable ? "create\t%s\n" :
dryrun ? "would create %s\n" : "create %s\n", argv[0]);
while ((nvp = nvlist_next_nvpair(props, nvp)) != NULL) {
uint64_t uval;
const char *sval;
switch (nvpair_type(nvp)) {
case DATA_TYPE_UINT64:
VERIFY0(nvpair_value_uint64(nvp, &uval));
(void) printf(parseable ?
"property\t%s\t%llu\n" : "\t%s=%llu\n",
nvpair_name(nvp), (u_longlong_t)uval);
break;
case DATA_TYPE_STRING:
VERIFY0(nvpair_value_string(nvp, &sval));
(void) printf(parseable ?
"property\t%s\t%s\n" : "\t%s=%s\n",
nvpair_name(nvp), sval);
break;
default:
(void) fprintf(stderr, "property '%s' "
"has illegal type %d\n",
nvpair_name(nvp), nvpair_type(nvp));
abort();
}
}
}
if (dryrun) {
ret = 0;
goto error;
}
/* pass to libzfs */
if (zfs_create(g_zfs, argv[0], type, props) != 0)
goto error;
if (log_history) {
(void) zpool_log_history(g_zfs, history_str);
log_history = B_FALSE;
}
if (nomount) {
ret = 0;
goto error;
}
ret = zfs_mount_and_share(g_zfs, argv[0], ZFS_TYPE_DATASET);
error:
nvlist_free(props);
return (ret);
badusage:
nvlist_free(props);
usage(B_FALSE);
return (2);
}
/*
* zfs destroy [-rRf] <fs, vol>
* zfs destroy [-rRd] <snap>
*
* -r Recursively destroy all children
* -R Recursively destroy all dependents, including clones
* -f Force unmounting of any dependents
* -d If we can't destroy now, mark for deferred destruction
*
* Destroys the given dataset. By default, it will unmount any filesystems,
* and refuse to destroy a dataset that has any dependents. A dependent can
* either be a child, or a clone of a child.
*/
typedef struct destroy_cbdata {
boolean_t cb_first;
boolean_t cb_force;
boolean_t cb_recurse;
boolean_t cb_error;
boolean_t cb_doclones;
zfs_handle_t *cb_target;
boolean_t cb_defer_destroy;
boolean_t cb_verbose;
boolean_t cb_parsable;
boolean_t cb_dryrun;
nvlist_t *cb_nvl;
nvlist_t *cb_batchedsnaps;
/* first snap in contiguous run */
char *cb_firstsnap;
/* previous snap in contiguous run */
char *cb_prevsnap;
int64_t cb_snapused;
char *cb_snapspec;
char *cb_bookmark;
uint64_t cb_snap_count;
} destroy_cbdata_t;
/*
* Check for any dependents based on the '-r' or '-R' flags.
*/
static int
destroy_check_dependent(zfs_handle_t *zhp, void *data)
{
destroy_cbdata_t *cbp = data;
const char *tname = zfs_get_name(cbp->cb_target);
const char *name = zfs_get_name(zhp);
if (strncmp(tname, name, strlen(tname)) == 0 &&
(name[strlen(tname)] == '/' || name[strlen(tname)] == '@')) {
/*
* This is a direct descendant, not a clone somewhere else in
* the hierarchy.
*/
if (cbp->cb_recurse)
goto out;
if (cbp->cb_first) {
(void) fprintf(stderr, gettext("cannot destroy '%s': "
"%s has children\n"),
zfs_get_name(cbp->cb_target),
zfs_type_to_name(zfs_get_type(cbp->cb_target)));
(void) fprintf(stderr, gettext("use '-r' to destroy "
"the following datasets:\n"));
cbp->cb_first = B_FALSE;
cbp->cb_error = B_TRUE;
}
(void) fprintf(stderr, "%s\n", zfs_get_name(zhp));
} else {
/*
* This is a clone. We only want to report this if the '-r'
* wasn't specified, or the target is a snapshot.
*/
if (!cbp->cb_recurse &&
zfs_get_type(cbp->cb_target) != ZFS_TYPE_SNAPSHOT)
goto out;
if (cbp->cb_first) {
(void) fprintf(stderr, gettext("cannot destroy '%s': "
"%s has dependent clones\n"),
zfs_get_name(cbp->cb_target),
zfs_type_to_name(zfs_get_type(cbp->cb_target)));
(void) fprintf(stderr, gettext("use '-R' to destroy "
"the following datasets:\n"));
cbp->cb_first = B_FALSE;
cbp->cb_error = B_TRUE;
cbp->cb_dryrun = B_TRUE;
}
(void) fprintf(stderr, "%s\n", zfs_get_name(zhp));
}
out:
zfs_close(zhp);
return (0);
}
static int
destroy_batched(destroy_cbdata_t *cb)
{
int error = zfs_destroy_snaps_nvl(g_zfs,
cb->cb_batchedsnaps, B_FALSE);
fnvlist_free(cb->cb_batchedsnaps);
cb->cb_batchedsnaps = fnvlist_alloc();
return (error);
}
static int
destroy_callback(zfs_handle_t *zhp, void *data)
{
destroy_cbdata_t *cb = data;
const char *name = zfs_get_name(zhp);
int error;
if (cb->cb_verbose) {
if (cb->cb_parsable) {
(void) printf("destroy\t%s\n", name);
} else if (cb->cb_dryrun) {
(void) printf(gettext("would destroy %s\n"),
name);
} else {
(void) printf(gettext("will destroy %s\n"),
name);
}
}
/*
* Ignore pools (which we've already flagged as an error before getting
* here).
*/
if (strchr(zfs_get_name(zhp), '/') == NULL &&
zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) {
zfs_close(zhp);
return (0);
}
if (cb->cb_dryrun) {
zfs_close(zhp);
return (0);
}
/*
* We batch up all contiguous snapshots (even of different
* filesystems) and destroy them with one ioctl. We can't
* simply do all snap deletions and then all fs deletions,
* because we must delete a clone before its origin.
*/
if (zfs_get_type(zhp) == ZFS_TYPE_SNAPSHOT) {
cb->cb_snap_count++;
fnvlist_add_boolean(cb->cb_batchedsnaps, name);
if (cb->cb_snap_count % 10 == 0 && cb->cb_defer_destroy) {
error = destroy_batched(cb);
if (error != 0) {
zfs_close(zhp);
return (-1);
}
}
} else {
error = destroy_batched(cb);
if (error != 0 ||
zfs_unmount(zhp, NULL, cb->cb_force ? MS_FORCE : 0) != 0 ||
zfs_destroy(zhp, cb->cb_defer_destroy) != 0) {
zfs_close(zhp);
/*
* When performing a recursive destroy we ignore errors
* so that the recursive destroy could continue
* destroying past problem datasets
*/
if (cb->cb_recurse) {
cb->cb_error = B_TRUE;
return (0);
}
return (-1);
}
}
zfs_close(zhp);
return (0);
}
static int
destroy_print_cb(zfs_handle_t *zhp, void *arg)
{
destroy_cbdata_t *cb = arg;
const char *name = zfs_get_name(zhp);
int err = 0;
if (nvlist_exists(cb->cb_nvl, name)) {
if (cb->cb_firstsnap == NULL)
cb->cb_firstsnap = strdup(name);
if (cb->cb_prevsnap != NULL)
free(cb->cb_prevsnap);
/* this snap continues the current range */
cb->cb_prevsnap = strdup(name);
if (cb->cb_firstsnap == NULL || cb->cb_prevsnap == NULL)
nomem();
if (cb->cb_verbose) {
if (cb->cb_parsable) {
(void) printf("destroy\t%s\n", name);
} else if (cb->cb_dryrun) {
(void) printf(gettext("would destroy %s\n"),
name);
} else {
(void) printf(gettext("will destroy %s\n"),
name);
}
}
} else if (cb->cb_firstsnap != NULL) {
/* end of this range */
uint64_t used = 0;
err = lzc_snaprange_space(cb->cb_firstsnap,
cb->cb_prevsnap, &used);
cb->cb_snapused += used;
free(cb->cb_firstsnap);
cb->cb_firstsnap = NULL;
free(cb->cb_prevsnap);
cb->cb_prevsnap = NULL;
}
zfs_close(zhp);
return (err);
}
static int
destroy_print_snapshots(zfs_handle_t *fs_zhp, destroy_cbdata_t *cb)
{
int err;
assert(cb->cb_firstsnap == NULL);
assert(cb->cb_prevsnap == NULL);
- err = zfs_iter_snapshots_sorted(fs_zhp, 0, destroy_print_cb, cb, 0, 0);
+ err = zfs_iter_snapshots_sorted_v2(fs_zhp, 0, destroy_print_cb, cb, 0,
+ 0);
if (cb->cb_firstsnap != NULL) {
uint64_t used = 0;
if (err == 0) {
err = lzc_snaprange_space(cb->cb_firstsnap,
cb->cb_prevsnap, &used);
}
cb->cb_snapused += used;
free(cb->cb_firstsnap);
cb->cb_firstsnap = NULL;
free(cb->cb_prevsnap);
cb->cb_prevsnap = NULL;
}
return (err);
}
static int
snapshot_to_nvl_cb(zfs_handle_t *zhp, void *arg)
{
destroy_cbdata_t *cb = arg;
int err = 0;
/* Check for clones. */
if (!cb->cb_doclones && !cb->cb_defer_destroy) {
cb->cb_target = zhp;
cb->cb_first = B_TRUE;
- err = zfs_iter_dependents(zhp, 0, B_TRUE,
+ err = zfs_iter_dependents_v2(zhp, 0, B_TRUE,
destroy_check_dependent, cb);
}
if (err == 0) {
if (nvlist_add_boolean(cb->cb_nvl, zfs_get_name(zhp)))
nomem();
}
zfs_close(zhp);
return (err);
}
static int
gather_snapshots(zfs_handle_t *zhp, void *arg)
{
destroy_cbdata_t *cb = arg;
int err = 0;
- err = zfs_iter_snapspec(zhp, 0, cb->cb_snapspec,
+ err = zfs_iter_snapspec_v2(zhp, 0, cb->cb_snapspec,
snapshot_to_nvl_cb, cb);
if (err == ENOENT)
err = 0;
if (err != 0)
goto out;
if (cb->cb_verbose) {
err = destroy_print_snapshots(zhp, cb);
if (err != 0)
goto out;
}
if (cb->cb_recurse)
- err = zfs_iter_filesystems(zhp, 0, gather_snapshots, cb);
+ err = zfs_iter_filesystems_v2(zhp, 0, gather_snapshots, cb);
out:
zfs_close(zhp);
return (err);
}
static int
destroy_clones(destroy_cbdata_t *cb)
{
nvpair_t *pair;
for (pair = nvlist_next_nvpair(cb->cb_nvl, NULL);
pair != NULL;
pair = nvlist_next_nvpair(cb->cb_nvl, pair)) {
zfs_handle_t *zhp = zfs_open(g_zfs, nvpair_name(pair),
ZFS_TYPE_SNAPSHOT);
if (zhp != NULL) {
boolean_t defer = cb->cb_defer_destroy;
int err;
/*
* We can't defer destroy non-snapshots, so set it to
* false while destroying the clones.
*/
cb->cb_defer_destroy = B_FALSE;
- err = zfs_iter_dependents(zhp, 0, B_FALSE,
+ err = zfs_iter_dependents_v2(zhp, 0, B_FALSE,
destroy_callback, cb);
cb->cb_defer_destroy = defer;
zfs_close(zhp);
if (err != 0)
return (err);
}
}
return (0);
}
static int
zfs_do_destroy(int argc, char **argv)
{
destroy_cbdata_t cb = { 0 };
int rv = 0;
int err = 0;
int c;
zfs_handle_t *zhp = NULL;
char *at, *pound;
zfs_type_t type = ZFS_TYPE_DATASET;
/* check options */
while ((c = getopt(argc, argv, "vpndfrR")) != -1) {
switch (c) {
case 'v':
cb.cb_verbose = B_TRUE;
break;
case 'p':
cb.cb_verbose = B_TRUE;
cb.cb_parsable = B_TRUE;
break;
case 'n':
cb.cb_dryrun = B_TRUE;
break;
case 'd':
cb.cb_defer_destroy = B_TRUE;
type = ZFS_TYPE_SNAPSHOT;
break;
case 'f':
cb.cb_force = B_TRUE;
break;
case 'r':
cb.cb_recurse = B_TRUE;
break;
case 'R':
cb.cb_recurse = B_TRUE;
cb.cb_doclones = B_TRUE;
break;
case '?':
default:
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc == 0) {
(void) fprintf(stderr, gettext("missing dataset argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
at = strchr(argv[0], '@');
pound = strchr(argv[0], '#');
if (at != NULL) {
/* Build the list of snaps to destroy in cb_nvl. */
cb.cb_nvl = fnvlist_alloc();
*at = '\0';
zhp = zfs_open(g_zfs, argv[0],
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL) {
nvlist_free(cb.cb_nvl);
return (1);
}
cb.cb_snapspec = at + 1;
if (gather_snapshots(zfs_handle_dup(zhp), &cb) != 0 ||
cb.cb_error) {
rv = 1;
goto out;
}
if (nvlist_empty(cb.cb_nvl)) {
(void) fprintf(stderr, gettext("could not find any "
"snapshots to destroy; check snapshot names.\n"));
rv = 1;
goto out;
}
if (cb.cb_verbose) {
char buf[16];
zfs_nicebytes(cb.cb_snapused, buf, sizeof (buf));
if (cb.cb_parsable) {
(void) printf("reclaim\t%llu\n",
(u_longlong_t)cb.cb_snapused);
} else if (cb.cb_dryrun) {
(void) printf(gettext("would reclaim %s\n"),
buf);
} else {
(void) printf(gettext("will reclaim %s\n"),
buf);
}
}
if (!cb.cb_dryrun) {
if (cb.cb_doclones) {
cb.cb_batchedsnaps = fnvlist_alloc();
err = destroy_clones(&cb);
if (err == 0) {
err = zfs_destroy_snaps_nvl(g_zfs,
cb.cb_batchedsnaps, B_FALSE);
}
if (err != 0) {
rv = 1;
goto out;
}
}
if (err == 0) {
err = zfs_destroy_snaps_nvl(g_zfs, cb.cb_nvl,
cb.cb_defer_destroy);
}
}
if (err != 0)
rv = 1;
} else if (pound != NULL) {
int err;
nvlist_t *nvl;
if (cb.cb_dryrun) {
(void) fprintf(stderr,
"dryrun is not supported with bookmark\n");
return (-1);
}
if (cb.cb_defer_destroy) {
(void) fprintf(stderr,
"defer destroy is not supported with bookmark\n");
return (-1);
}
if (cb.cb_recurse) {
(void) fprintf(stderr,
"recursive is not supported with bookmark\n");
return (-1);
}
/*
* Unfortunately, zfs_bookmark() doesn't honor the
* casesensitivity setting. However, we can't simply
* remove this check, because lzc_destroy_bookmarks()
* ignores non-existent bookmarks, so this is necessary
* to get a proper error message.
*/
if (!zfs_bookmark_exists(argv[0])) {
(void) fprintf(stderr, gettext("bookmark '%s' "
"does not exist.\n"), argv[0]);
return (1);
}
nvl = fnvlist_alloc();
fnvlist_add_boolean(nvl, argv[0]);
err = lzc_destroy_bookmarks(nvl, NULL);
if (err != 0) {
(void) zfs_standard_error(g_zfs, err,
"cannot destroy bookmark");
}
nvlist_free(nvl);
return (err);
} else {
/* Open the given dataset */
if ((zhp = zfs_open(g_zfs, argv[0], type)) == NULL)
return (1);
cb.cb_target = zhp;
/*
* Perform an explicit check for pools before going any further.
*/
if (!cb.cb_recurse && strchr(zfs_get_name(zhp), '/') == NULL &&
zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) {
(void) fprintf(stderr, gettext("cannot destroy '%s': "
"operation does not apply to pools\n"),
zfs_get_name(zhp));
(void) fprintf(stderr, gettext("use 'zfs destroy -r "
"%s' to destroy all datasets in the pool\n"),
zfs_get_name(zhp));
(void) fprintf(stderr, gettext("use 'zpool destroy %s' "
"to destroy the pool itself\n"), zfs_get_name(zhp));
rv = 1;
goto out;
}
/*
* Check for any dependents and/or clones.
*/
cb.cb_first = B_TRUE;
- if (!cb.cb_doclones &&
- zfs_iter_dependents(zhp, 0, B_TRUE, destroy_check_dependent,
- &cb) != 0) {
+ if (!cb.cb_doclones && zfs_iter_dependents_v2(zhp, 0, B_TRUE,
+ destroy_check_dependent, &cb) != 0) {
rv = 1;
goto out;
}
if (cb.cb_error) {
rv = 1;
goto out;
}
cb.cb_batchedsnaps = fnvlist_alloc();
- if (zfs_iter_dependents(zhp, 0, B_FALSE, destroy_callback,
+ if (zfs_iter_dependents_v2(zhp, 0, B_FALSE, destroy_callback,
&cb) != 0) {
rv = 1;
goto out;
}
/*
* Do the real thing. The callback will close the
* handle regardless of whether it succeeds or not.
*/
err = destroy_callback(zhp, &cb);
zhp = NULL;
if (err == 0) {
err = zfs_destroy_snaps_nvl(g_zfs,
cb.cb_batchedsnaps, cb.cb_defer_destroy);
}
if (err != 0 || cb.cb_error == B_TRUE)
rv = 1;
}
out:
fnvlist_free(cb.cb_batchedsnaps);
fnvlist_free(cb.cb_nvl);
if (zhp != NULL)
zfs_close(zhp);
return (rv);
}
static boolean_t
is_recvd_column(zprop_get_cbdata_t *cbp)
{
int i;
zfs_get_column_t col;
for (i = 0; i < ZFS_GET_NCOLS &&
(col = cbp->cb_columns[i]) != GET_COL_NONE; i++)
if (col == GET_COL_RECVD)
return (B_TRUE);
return (B_FALSE);
}
/*
* zfs get [-rHp] [-o all | field[,field]...] [-s source[,source]...]
* < all | property[,property]... > < fs | snap | vol > ...
*
* -r recurse over any child datasets
* -H scripted mode. Headers are stripped, and fields are separated
* by tabs instead of spaces.
* -o Set of fields to display. One of "name,property,value,
* received,source". Default is "name,property,value,source".
* "all" is an alias for all five.
* -s Set of sources to allow. One of
* "local,default,inherited,received,temporary,none". Default is
* all six.
* -p Display values in parsable (literal) format.
*
* Prints properties for the given datasets. The user can control which
* columns to display as well as which property types to allow.
*/
/*
* Invoked to display the properties for a single dataset.
*/
static int
get_callback(zfs_handle_t *zhp, void *data)
{
char buf[ZFS_MAXPROPLEN];
char rbuf[ZFS_MAXPROPLEN];
zprop_source_t sourcetype;
char source[ZFS_MAX_DATASET_NAME_LEN];
zprop_get_cbdata_t *cbp = data;
nvlist_t *user_props = zfs_get_user_props(zhp);
zprop_list_t *pl = cbp->cb_proplist;
nvlist_t *propval;
const char *strval;
const char *sourceval;
boolean_t received = is_recvd_column(cbp);
for (; pl != NULL; pl = pl->pl_next) {
char *recvdval = NULL;
/*
* Skip the special fake placeholder. This will also skip over
* the name property when 'all' is specified.
*/
if (pl->pl_prop == ZFS_PROP_NAME &&
pl == cbp->cb_proplist)
continue;
if (pl->pl_prop != ZPROP_USERPROP) {
if (zfs_prop_get(zhp, pl->pl_prop, buf,
sizeof (buf), &sourcetype, source,
sizeof (source),
cbp->cb_literal) != 0) {
if (pl->pl_all)
continue;
if (!zfs_prop_valid_for_type(pl->pl_prop,
ZFS_TYPE_DATASET, B_FALSE)) {
(void) fprintf(stderr,
gettext("No such property '%s'\n"),
zfs_prop_to_name(pl->pl_prop));
continue;
}
sourcetype = ZPROP_SRC_NONE;
(void) strlcpy(buf, "-", sizeof (buf));
}
if (received && (zfs_prop_get_recvd(zhp,
zfs_prop_to_name(pl->pl_prop), rbuf, sizeof (rbuf),
cbp->cb_literal) == 0))
recvdval = rbuf;
zprop_print_one_property(zfs_get_name(zhp), cbp,
zfs_prop_to_name(pl->pl_prop),
buf, sourcetype, source, recvdval);
} else if (zfs_prop_userquota(pl->pl_user_prop)) {
sourcetype = ZPROP_SRC_LOCAL;
if (zfs_prop_get_userquota(zhp, pl->pl_user_prop,
buf, sizeof (buf), cbp->cb_literal) != 0) {
sourcetype = ZPROP_SRC_NONE;
(void) strlcpy(buf, "-", sizeof (buf));
}
zprop_print_one_property(zfs_get_name(zhp), cbp,
pl->pl_user_prop, buf, sourcetype, source, NULL);
} else if (zfs_prop_written(pl->pl_user_prop)) {
sourcetype = ZPROP_SRC_LOCAL;
if (zfs_prop_get_written(zhp, pl->pl_user_prop,
buf, sizeof (buf), cbp->cb_literal) != 0) {
sourcetype = ZPROP_SRC_NONE;
(void) strlcpy(buf, "-", sizeof (buf));
}
zprop_print_one_property(zfs_get_name(zhp), cbp,
pl->pl_user_prop, buf, sourcetype, source, NULL);
} else {
if (nvlist_lookup_nvlist(user_props,
pl->pl_user_prop, &propval) != 0) {
if (pl->pl_all)
continue;
sourcetype = ZPROP_SRC_NONE;
strval = "-";
} else {
strval = fnvlist_lookup_string(propval,
ZPROP_VALUE);
sourceval = fnvlist_lookup_string(propval,
ZPROP_SOURCE);
if (strcmp(sourceval,
zfs_get_name(zhp)) == 0) {
sourcetype = ZPROP_SRC_LOCAL;
} else if (strcmp(sourceval,
ZPROP_SOURCE_VAL_RECVD) == 0) {
sourcetype = ZPROP_SRC_RECEIVED;
} else {
sourcetype = ZPROP_SRC_INHERITED;
(void) strlcpy(source,
sourceval, sizeof (source));
}
}
if (received && (zfs_prop_get_recvd(zhp,
pl->pl_user_prop, rbuf, sizeof (rbuf),
cbp->cb_literal) == 0))
recvdval = rbuf;
zprop_print_one_property(zfs_get_name(zhp), cbp,
pl->pl_user_prop, strval, sourcetype,
source, recvdval);
}
}
return (0);
}
static int
zfs_do_get(int argc, char **argv)
{
zprop_get_cbdata_t cb = { 0 };
int i, c, flags = ZFS_ITER_ARGS_CAN_BE_PATHS;
int types = ZFS_TYPE_DATASET | ZFS_TYPE_BOOKMARK;
char *fields;
int ret = 0;
int limit = 0;
zprop_list_t fake_name = { 0 };
/*
* Set up default columns and sources.
*/
cb.cb_sources = ZPROP_SRC_ALL;
cb.cb_columns[0] = GET_COL_NAME;
cb.cb_columns[1] = GET_COL_PROPERTY;
cb.cb_columns[2] = GET_COL_VALUE;
cb.cb_columns[3] = GET_COL_SOURCE;
cb.cb_type = ZFS_TYPE_DATASET;
/* check options */
while ((c = getopt(argc, argv, ":d:o:s:rt:Hp")) != -1) {
switch (c) {
case 'p':
cb.cb_literal = B_TRUE;
break;
case 'd':
limit = parse_depth(optarg, &flags);
break;
case 'r':
flags |= ZFS_ITER_RECURSE;
break;
case 'H':
cb.cb_scripted = B_TRUE;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case 'o':
/*
* Process the set of columns to display. We zero out
* the structure to give us a blank slate.
*/
memset(&cb.cb_columns, 0, sizeof (cb.cb_columns));
i = 0;
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const col_subopts[] =
{ "name", "property", "value",
"received", "source", "all" };
static const zfs_get_column_t col_subopt_col[] =
{ GET_COL_NAME, GET_COL_PROPERTY, GET_COL_VALUE,
GET_COL_RECVD, GET_COL_SOURCE };
static const int col_subopt_flags[] =
{ 0, 0, 0, ZFS_ITER_RECVD_PROPS, 0 };
if (i == ZFS_GET_NCOLS) {
(void) fprintf(stderr, gettext("too "
"many fields given to -o "
"option\n"));
usage(B_FALSE);
}
for (c = 0; c < ARRAY_SIZE(col_subopts); ++c)
if (strcmp(tok, col_subopts[c]) == 0)
goto found;
(void) fprintf(stderr,
gettext("invalid column name '%s'\n"), tok);
usage(B_FALSE);
found:
if (c >= 5) {
if (i > 0) {
(void) fprintf(stderr,
gettext("\"all\" conflicts "
"with specific fields "
"given to -o option\n"));
usage(B_FALSE);
}
memcpy(cb.cb_columns, col_subopt_col,
sizeof (col_subopt_col));
flags |= ZFS_ITER_RECVD_PROPS;
i = ZFS_GET_NCOLS;
} else {
cb.cb_columns[i++] = col_subopt_col[c];
flags |= col_subopt_flags[c];
}
}
break;
case 's':
cb.cb_sources = 0;
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const source_opt[] = {
"local", "default",
"inherited", "received",
"temporary", "none" };
static const int source_flg[] = {
ZPROP_SRC_LOCAL, ZPROP_SRC_DEFAULT,
ZPROP_SRC_INHERITED, ZPROP_SRC_RECEIVED,
ZPROP_SRC_TEMPORARY, ZPROP_SRC_NONE };
for (i = 0; i < ARRAY_SIZE(source_opt); ++i)
if (strcmp(tok, source_opt[i]) == 0) {
cb.cb_sources |= source_flg[i];
goto found2;
}
(void) fprintf(stderr,
gettext("invalid source '%s'\n"), tok);
usage(B_FALSE);
found2:;
}
break;
case 't':
types = 0;
flags &= ~ZFS_ITER_PROP_LISTSNAPS;
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const type_opts[] = {
"filesystem", "volume",
"snapshot", "snap",
"bookmark",
"all" };
static const int type_types[] = {
ZFS_TYPE_FILESYSTEM, ZFS_TYPE_VOLUME,
ZFS_TYPE_SNAPSHOT, ZFS_TYPE_SNAPSHOT,
ZFS_TYPE_BOOKMARK,
ZFS_TYPE_DATASET | ZFS_TYPE_BOOKMARK };
for (i = 0; i < ARRAY_SIZE(type_opts); ++i)
if (strcmp(tok, type_opts[i]) == 0) {
types |= type_types[i];
goto found3;
}
(void) fprintf(stderr,
gettext("invalid type '%s'\n"), tok);
usage(B_FALSE);
found3:;
}
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing property "
"argument\n"));
usage(B_FALSE);
}
fields = argv[0];
/*
* Handle users who want to get all snapshots or bookmarks
* of a dataset (ex. 'zfs get -t snapshot refer <dataset>').
*/
if ((types == ZFS_TYPE_SNAPSHOT || types == ZFS_TYPE_BOOKMARK) &&
argc > 1 && (flags & ZFS_ITER_RECURSE) == 0 && limit == 0) {
flags |= (ZFS_ITER_DEPTH_LIMIT | ZFS_ITER_RECURSE);
limit = 1;
}
if (zprop_get_list(g_zfs, fields, &cb.cb_proplist, ZFS_TYPE_DATASET)
!= 0)
usage(B_FALSE);
argc--;
argv++;
/*
* As part of zfs_expand_proplist(), we keep track of the maximum column
* width for each property. For the 'NAME' (and 'SOURCE') columns, we
* need to know the maximum name length. However, the user likely did
* not specify 'name' as one of the properties to fetch, so we need to
* make sure we always include at least this property for
* print_get_headers() to work properly.
*/
if (cb.cb_proplist != NULL) {
fake_name.pl_prop = ZFS_PROP_NAME;
fake_name.pl_width = strlen(gettext("NAME"));
fake_name.pl_next = cb.cb_proplist;
cb.cb_proplist = &fake_name;
}
cb.cb_first = B_TRUE;
/* run for each object */
ret = zfs_for_each(argc, argv, flags, types, NULL,
&cb.cb_proplist, limit, get_callback, &cb);
if (cb.cb_proplist == &fake_name)
zprop_free_list(fake_name.pl_next);
else
zprop_free_list(cb.cb_proplist);
return (ret);
}
/*
* inherit [-rS] <property> <fs|vol> ...
*
* -r Recurse over all children
* -S Revert to received value, if any
*
* For each dataset specified on the command line, inherit the given property
* from its parent. Inheriting a property at the pool level will cause it to
* use the default value. The '-r' flag will recurse over all children, and is
* useful for setting a property on a hierarchy-wide basis, regardless of any
* local modifications for each dataset.
*/
typedef struct inherit_cbdata {
const char *cb_propname;
boolean_t cb_received;
} inherit_cbdata_t;
static int
inherit_recurse_cb(zfs_handle_t *zhp, void *data)
{
inherit_cbdata_t *cb = data;
zfs_prop_t prop = zfs_name_to_prop(cb->cb_propname);
/*
* If we're doing it recursively, then ignore properties that
* are not valid for this type of dataset.
*/
if (prop != ZPROP_INVAL &&
!zfs_prop_valid_for_type(prop, zfs_get_type(zhp), B_FALSE))
return (0);
return (zfs_prop_inherit(zhp, cb->cb_propname, cb->cb_received) != 0);
}
static int
inherit_cb(zfs_handle_t *zhp, void *data)
{
inherit_cbdata_t *cb = data;
return (zfs_prop_inherit(zhp, cb->cb_propname, cb->cb_received) != 0);
}
static int
zfs_do_inherit(int argc, char **argv)
{
int c;
zfs_prop_t prop;
inherit_cbdata_t cb = { 0 };
char *propname;
int ret = 0;
int flags = 0;
boolean_t received = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "rS")) != -1) {
switch (c) {
case 'r':
flags |= ZFS_ITER_RECURSE;
break;
case 'S':
received = B_TRUE;
break;
case '?':
default:
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing property argument\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing dataset argument\n"));
usage(B_FALSE);
}
propname = argv[0];
argc--;
argv++;
if ((prop = zfs_name_to_prop(propname)) != ZPROP_USERPROP) {
if (zfs_prop_readonly(prop)) {
(void) fprintf(stderr, gettext(
"%s property is read-only\n"),
propname);
return (1);
}
if (!zfs_prop_inheritable(prop) && !received) {
(void) fprintf(stderr, gettext("'%s' property cannot "
"be inherited\n"), propname);
if (prop == ZFS_PROP_QUOTA ||
prop == ZFS_PROP_RESERVATION ||
prop == ZFS_PROP_REFQUOTA ||
prop == ZFS_PROP_REFRESERVATION) {
(void) fprintf(stderr, gettext("use 'zfs set "
"%s=none' to clear\n"), propname);
(void) fprintf(stderr, gettext("use 'zfs "
"inherit -S %s' to revert to received "
"value\n"), propname);
}
return (1);
}
if (received && (prop == ZFS_PROP_VOLSIZE ||
prop == ZFS_PROP_VERSION)) {
(void) fprintf(stderr, gettext("'%s' property cannot "
"be reverted to a received value\n"), propname);
return (1);
}
} else if (!zfs_prop_user(propname)) {
(void) fprintf(stderr, gettext("invalid property '%s'\n"),
propname);
usage(B_FALSE);
}
cb.cb_propname = propname;
cb.cb_received = received;
if (flags & ZFS_ITER_RECURSE) {
ret = zfs_for_each(argc, argv, flags, ZFS_TYPE_DATASET,
NULL, NULL, 0, inherit_recurse_cb, &cb);
} else {
ret = zfs_for_each(argc, argv, flags, ZFS_TYPE_DATASET,
NULL, NULL, 0, inherit_cb, &cb);
}
return (ret);
}
typedef struct upgrade_cbdata {
uint64_t cb_numupgraded;
uint64_t cb_numsamegraded;
uint64_t cb_numfailed;
uint64_t cb_version;
boolean_t cb_newer;
boolean_t cb_foundone;
char cb_lastfs[ZFS_MAX_DATASET_NAME_LEN];
} upgrade_cbdata_t;
static int
same_pool(zfs_handle_t *zhp, const char *name)
{
int len1 = strcspn(name, "/@");
const char *zhname = zfs_get_name(zhp);
int len2 = strcspn(zhname, "/@");
if (len1 != len2)
return (B_FALSE);
return (strncmp(name, zhname, len1) == 0);
}
static int
upgrade_list_callback(zfs_handle_t *zhp, void *data)
{
upgrade_cbdata_t *cb = data;
int version = zfs_prop_get_int(zhp, ZFS_PROP_VERSION);
/* list if it's old/new */
if ((!cb->cb_newer && version < ZPL_VERSION) ||
(cb->cb_newer && version > ZPL_VERSION)) {
char *str;
if (cb->cb_newer) {
str = gettext("The following filesystems are "
"formatted using a newer software version and\n"
"cannot be accessed on the current system.\n\n");
} else {
str = gettext("The following filesystems are "
"out of date, and can be upgraded. After being\n"
"upgraded, these filesystems (and any 'zfs send' "
"streams generated from\n"
"subsequent snapshots) will no longer be "
"accessible by older software versions.\n\n");
}
if (!cb->cb_foundone) {
(void) puts(str);
(void) printf(gettext("VER FILESYSTEM\n"));
(void) printf(gettext("--- ------------\n"));
cb->cb_foundone = B_TRUE;
}
(void) printf("%2u %s\n", version, zfs_get_name(zhp));
}
return (0);
}
static int
upgrade_set_callback(zfs_handle_t *zhp, void *data)
{
upgrade_cbdata_t *cb = data;
int version = zfs_prop_get_int(zhp, ZFS_PROP_VERSION);
int needed_spa_version;
int spa_version;
if (zfs_spa_version(zhp, &spa_version) < 0)
return (-1);
needed_spa_version = zfs_spa_version_map(cb->cb_version);
if (needed_spa_version < 0)
return (-1);
if (spa_version < needed_spa_version) {
/* can't upgrade */
(void) printf(gettext("%s: can not be "
"upgraded; the pool version needs to first "
"be upgraded\nto version %d\n\n"),
zfs_get_name(zhp), needed_spa_version);
cb->cb_numfailed++;
return (0);
}
/* upgrade */
if (version < cb->cb_version) {
char verstr[24];
(void) snprintf(verstr, sizeof (verstr),
"%llu", (u_longlong_t)cb->cb_version);
if (cb->cb_lastfs[0] && !same_pool(zhp, cb->cb_lastfs)) {
/*
* If they did "zfs upgrade -a", then we could
* be doing ioctls to different pools. We need
* to log this history once to each pool, and bypass
* the normal history logging that happens in main().
*/
(void) zpool_log_history(g_zfs, history_str);
log_history = B_FALSE;
}
if (zfs_prop_set(zhp, "version", verstr) == 0)
cb->cb_numupgraded++;
else
cb->cb_numfailed++;
(void) strlcpy(cb->cb_lastfs, zfs_get_name(zhp),
sizeof (cb->cb_lastfs));
} else if (version > cb->cb_version) {
/* can't downgrade */
(void) printf(gettext("%s: can not be downgraded; "
"it is already at version %u\n"),
zfs_get_name(zhp), version);
cb->cb_numfailed++;
} else {
cb->cb_numsamegraded++;
}
return (0);
}
/*
* zfs upgrade
* zfs upgrade -v
* zfs upgrade [-r] [-V <version>] <-a | filesystem>
*/
static int
zfs_do_upgrade(int argc, char **argv)
{
boolean_t all = B_FALSE;
boolean_t showversions = B_FALSE;
int ret = 0;
upgrade_cbdata_t cb = { 0 };
int c;
int flags = ZFS_ITER_ARGS_CAN_BE_PATHS;
/* check options */
while ((c = getopt(argc, argv, "rvV:a")) != -1) {
switch (c) {
case 'r':
flags |= ZFS_ITER_RECURSE;
break;
case 'v':
showversions = B_TRUE;
break;
case 'V':
if (zfs_prop_string_to_index(ZFS_PROP_VERSION,
optarg, &cb.cb_version) != 0) {
(void) fprintf(stderr,
gettext("invalid version %s\n"), optarg);
usage(B_FALSE);
}
break;
case 'a':
all = B_TRUE;
break;
case '?':
default:
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if ((!all && !argc) && ((flags & ZFS_ITER_RECURSE) | cb.cb_version))
usage(B_FALSE);
if (showversions && (flags & ZFS_ITER_RECURSE || all ||
cb.cb_version || argc))
usage(B_FALSE);
if ((all || argc) && (showversions))
usage(B_FALSE);
if (all && argc)
usage(B_FALSE);
if (showversions) {
/* Show info on available versions. */
(void) printf(gettext("The following filesystem versions are "
"supported:\n\n"));
(void) printf(gettext("VER DESCRIPTION\n"));
(void) printf("--- -----------------------------------------"
"---------------\n");
(void) printf(gettext(" 1 Initial ZFS filesystem version\n"));
(void) printf(gettext(" 2 Enhanced directory entries\n"));
(void) printf(gettext(" 3 Case insensitive and filesystem "
"user identifier (FUID)\n"));
(void) printf(gettext(" 4 userquota, groupquota "
"properties\n"));
(void) printf(gettext(" 5 System attributes\n"));
(void) printf(gettext("\nFor more information on a particular "
"version, including supported releases,\n"));
(void) printf("see the ZFS Administration Guide.\n\n");
ret = 0;
} else if (argc || all) {
/* Upgrade filesystems */
if (cb.cb_version == 0)
cb.cb_version = ZPL_VERSION;
ret = zfs_for_each(argc, argv, flags, ZFS_TYPE_FILESYSTEM,
NULL, NULL, 0, upgrade_set_callback, &cb);
(void) printf(gettext("%llu filesystems upgraded\n"),
(u_longlong_t)cb.cb_numupgraded);
if (cb.cb_numsamegraded) {
(void) printf(gettext("%llu filesystems already at "
"this version\n"),
(u_longlong_t)cb.cb_numsamegraded);
}
if (cb.cb_numfailed != 0)
ret = 1;
} else {
/* List old-version filesystems */
boolean_t found;
(void) printf(gettext("This system is currently running "
"ZFS filesystem version %llu.\n\n"), ZPL_VERSION);
flags |= ZFS_ITER_RECURSE;
ret = zfs_for_each(0, NULL, flags, ZFS_TYPE_FILESYSTEM,
NULL, NULL, 0, upgrade_list_callback, &cb);
found = cb.cb_foundone;
cb.cb_foundone = B_FALSE;
cb.cb_newer = B_TRUE;
ret |= zfs_for_each(0, NULL, flags, ZFS_TYPE_FILESYSTEM,
NULL, NULL, 0, upgrade_list_callback, &cb);
if (!cb.cb_foundone && !found) {
(void) printf(gettext("All filesystems are "
"formatted with the current version.\n"));
}
}
return (ret);
}
/*
* zfs userspace [-Hinp] [-o field[,...]] [-s field [-s field]...]
* [-S field [-S field]...] [-t type[,...]]
* filesystem | snapshot | path
* zfs groupspace [-Hinp] [-o field[,...]] [-s field [-s field]...]
* [-S field [-S field]...] [-t type[,...]]
* filesystem | snapshot | path
* zfs projectspace [-Hp] [-o field[,...]] [-s field [-s field]...]
* [-S field [-S field]...] filesystem | snapshot | path
*
* -H Scripted mode; elide headers and separate columns by tabs.
* -i Translate SID to POSIX ID.
* -n Print numeric ID instead of user/group name.
* -o Control which fields to display.
* -p Use exact (parsable) numeric output.
* -s Specify sort columns, descending order.
* -S Specify sort columns, ascending order.
* -t Control which object types to display.
*
* Displays space consumed by, and quotas on, each user in the specified
* filesystem or snapshot.
*/
/* us_field_types, us_field_hdr and us_field_names should be kept in sync */
enum us_field_types {
USFIELD_TYPE,
USFIELD_NAME,
USFIELD_USED,
USFIELD_QUOTA,
USFIELD_OBJUSED,
USFIELD_OBJQUOTA
};
static const char *const us_field_hdr[] = { "TYPE", "NAME", "USED", "QUOTA",
"OBJUSED", "OBJQUOTA" };
static const char *const us_field_names[] = { "type", "name", "used", "quota",
"objused", "objquota" };
#define USFIELD_LAST (sizeof (us_field_names) / sizeof (char *))
#define USTYPE_PSX_GRP (1 << 0)
#define USTYPE_PSX_USR (1 << 1)
#define USTYPE_SMB_GRP (1 << 2)
#define USTYPE_SMB_USR (1 << 3)
#define USTYPE_PROJ (1 << 4)
#define USTYPE_ALL \
(USTYPE_PSX_GRP | USTYPE_PSX_USR | USTYPE_SMB_GRP | USTYPE_SMB_USR | \
USTYPE_PROJ)
static int us_type_bits[] = {
USTYPE_PSX_GRP,
USTYPE_PSX_USR,
USTYPE_SMB_GRP,
USTYPE_SMB_USR,
USTYPE_ALL
};
static const char *const us_type_names[] = { "posixgroup", "posixuser",
"smbgroup", "smbuser", "all" };
typedef struct us_node {
nvlist_t *usn_nvl;
uu_avl_node_t usn_avlnode;
uu_list_node_t usn_listnode;
} us_node_t;
typedef struct us_cbdata {
nvlist_t **cb_nvlp;
uu_avl_pool_t *cb_avl_pool;
uu_avl_t *cb_avl;
boolean_t cb_numname;
boolean_t cb_nicenum;
boolean_t cb_sid2posix;
zfs_userquota_prop_t cb_prop;
zfs_sort_column_t *cb_sortcol;
size_t cb_width[USFIELD_LAST];
} us_cbdata_t;
static boolean_t us_populated = B_FALSE;
typedef struct {
zfs_sort_column_t *si_sortcol;
boolean_t si_numname;
} us_sort_info_t;
static int
us_field_index(const char *field)
{
for (int i = 0; i < USFIELD_LAST; i++) {
if (strcmp(field, us_field_names[i]) == 0)
return (i);
}
return (-1);
}
static int
us_compare(const void *larg, const void *rarg, void *unused)
{
const us_node_t *l = larg;
const us_node_t *r = rarg;
us_sort_info_t *si = (us_sort_info_t *)unused;
zfs_sort_column_t *sortcol = si->si_sortcol;
boolean_t numname = si->si_numname;
nvlist_t *lnvl = l->usn_nvl;
nvlist_t *rnvl = r->usn_nvl;
int rc = 0;
boolean_t lvb, rvb;
for (; sortcol != NULL; sortcol = sortcol->sc_next) {
const char *lvstr = "";
const char *rvstr = "";
uint32_t lv32 = 0;
uint32_t rv32 = 0;
uint64_t lv64 = 0;
uint64_t rv64 = 0;
zfs_prop_t prop = sortcol->sc_prop;
const char *propname = NULL;
boolean_t reverse = sortcol->sc_reverse;
switch (prop) {
case ZFS_PROP_TYPE:
propname = "type";
(void) nvlist_lookup_uint32(lnvl, propname, &lv32);
(void) nvlist_lookup_uint32(rnvl, propname, &rv32);
if (rv32 != lv32)
rc = (rv32 < lv32) ? 1 : -1;
break;
case ZFS_PROP_NAME:
propname = "name";
if (numname) {
compare_nums:
(void) nvlist_lookup_uint64(lnvl, propname,
&lv64);
(void) nvlist_lookup_uint64(rnvl, propname,
&rv64);
if (rv64 != lv64)
rc = (rv64 < lv64) ? 1 : -1;
} else {
if ((nvlist_lookup_string(lnvl, propname,
&lvstr) == ENOENT) ||
(nvlist_lookup_string(rnvl, propname,
&rvstr) == ENOENT)) {
goto compare_nums;
}
rc = strcmp(lvstr, rvstr);
}
break;
case ZFS_PROP_USED:
case ZFS_PROP_QUOTA:
if (!us_populated)
break;
if (prop == ZFS_PROP_USED)
propname = "used";
else
propname = "quota";
(void) nvlist_lookup_uint64(lnvl, propname, &lv64);
(void) nvlist_lookup_uint64(rnvl, propname, &rv64);
if (rv64 != lv64)
rc = (rv64 < lv64) ? 1 : -1;
break;
default:
break;
}
if (rc != 0) {
if (rc < 0)
return (reverse ? 1 : -1);
else
return (reverse ? -1 : 1);
}
}
/*
* If entries still seem to be the same, check if they are of the same
* type (smbentity is added only if we are doing SID to POSIX ID
* translation where we can have duplicate type/name combinations).
*/
if (nvlist_lookup_boolean_value(lnvl, "smbentity", &lvb) == 0 &&
nvlist_lookup_boolean_value(rnvl, "smbentity", &rvb) == 0 &&
lvb != rvb)
return (lvb < rvb ? -1 : 1);
return (0);
}
static boolean_t
zfs_prop_is_user(unsigned p)
{
return (p == ZFS_PROP_USERUSED || p == ZFS_PROP_USERQUOTA ||
p == ZFS_PROP_USEROBJUSED || p == ZFS_PROP_USEROBJQUOTA);
}
static boolean_t
zfs_prop_is_group(unsigned p)
{
return (p == ZFS_PROP_GROUPUSED || p == ZFS_PROP_GROUPQUOTA ||
p == ZFS_PROP_GROUPOBJUSED || p == ZFS_PROP_GROUPOBJQUOTA);
}
static boolean_t
zfs_prop_is_project(unsigned p)
{
return (p == ZFS_PROP_PROJECTUSED || p == ZFS_PROP_PROJECTQUOTA ||
p == ZFS_PROP_PROJECTOBJUSED || p == ZFS_PROP_PROJECTOBJQUOTA);
}
static inline const char *
us_type2str(unsigned field_type)
{
switch (field_type) {
case USTYPE_PSX_USR:
return ("POSIX User");
case USTYPE_PSX_GRP:
return ("POSIX Group");
case USTYPE_SMB_USR:
return ("SMB User");
case USTYPE_SMB_GRP:
return ("SMB Group");
case USTYPE_PROJ:
return ("Project");
default:
return ("Undefined");
}
}
static int
userspace_cb(void *arg, const char *domain, uid_t rid, uint64_t space)
{
us_cbdata_t *cb = (us_cbdata_t *)arg;
zfs_userquota_prop_t prop = cb->cb_prop;
char *name = NULL;
const char *propname;
char sizebuf[32];
us_node_t *node;
uu_avl_pool_t *avl_pool = cb->cb_avl_pool;
uu_avl_t *avl = cb->cb_avl;
uu_avl_index_t idx;
nvlist_t *props;
us_node_t *n;
zfs_sort_column_t *sortcol = cb->cb_sortcol;
unsigned type = 0;
const char *typestr;
size_t namelen;
size_t typelen;
size_t sizelen;
int typeidx, nameidx, sizeidx;
us_sort_info_t sortinfo = { sortcol, cb->cb_numname };
boolean_t smbentity = B_FALSE;
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
nomem();
node = safe_malloc(sizeof (us_node_t));
uu_avl_node_init(node, &node->usn_avlnode, avl_pool);
node->usn_nvl = props;
if (domain != NULL && domain[0] != '\0') {
#ifdef HAVE_IDMAP
/* SMB */
char sid[MAXNAMELEN + 32];
uid_t id;
uint64_t classes;
int err;
directory_error_t e;
smbentity = B_TRUE;
(void) snprintf(sid, sizeof (sid), "%s-%u", domain, rid);
if (prop == ZFS_PROP_GROUPUSED || prop == ZFS_PROP_GROUPQUOTA) {
type = USTYPE_SMB_GRP;
err = sid_to_id(sid, B_FALSE, &id);
} else {
type = USTYPE_SMB_USR;
err = sid_to_id(sid, B_TRUE, &id);
}
if (err == 0) {
rid = id;
if (!cb->cb_sid2posix) {
e = directory_name_from_sid(NULL, sid, &name,
&classes);
if (e != NULL)
directory_error_free(e);
if (name == NULL)
name = sid;
}
}
#else
nvlist_free(props);
free(node);
return (-1);
#endif /* HAVE_IDMAP */
}
if (cb->cb_sid2posix || domain == NULL || domain[0] == '\0') {
/* POSIX or -i */
if (zfs_prop_is_group(prop)) {
type = USTYPE_PSX_GRP;
if (!cb->cb_numname) {
struct group *g;
if ((g = getgrgid(rid)) != NULL)
name = g->gr_name;
}
} else if (zfs_prop_is_user(prop)) {
type = USTYPE_PSX_USR;
if (!cb->cb_numname) {
struct passwd *p;
if ((p = getpwuid(rid)) != NULL)
name = p->pw_name;
}
} else {
type = USTYPE_PROJ;
}
}
/*
* Make sure that the type/name combination is unique when doing
* SID to POSIX ID translation (hence changing the type from SMB to
* POSIX).
*/
if (cb->cb_sid2posix &&
nvlist_add_boolean_value(props, "smbentity", smbentity) != 0)
nomem();
/* Calculate/update width of TYPE field */
typestr = us_type2str(type);
typelen = strlen(gettext(typestr));
typeidx = us_field_index("type");
if (typelen > cb->cb_width[typeidx])
cb->cb_width[typeidx] = typelen;
if (nvlist_add_uint32(props, "type", type) != 0)
nomem();
/* Calculate/update width of NAME field */
if ((cb->cb_numname && cb->cb_sid2posix) || name == NULL) {
if (nvlist_add_uint64(props, "name", rid) != 0)
nomem();
namelen = snprintf(NULL, 0, "%u", rid);
} else {
if (nvlist_add_string(props, "name", name) != 0)
nomem();
namelen = strlen(name);
}
nameidx = us_field_index("name");
if (nameidx >= 0 && namelen > cb->cb_width[nameidx])
cb->cb_width[nameidx] = namelen;
/*
* Check if this type/name combination is in the list and update it;
* otherwise add new node to the list.
*/
if ((n = uu_avl_find(avl, node, &sortinfo, &idx)) == NULL) {
uu_avl_insert(avl, node, idx);
} else {
nvlist_free(props);
free(node);
node = n;
props = node->usn_nvl;
}
/* Calculate/update width of USED/QUOTA fields */
if (cb->cb_nicenum) {
if (prop == ZFS_PROP_USERUSED || prop == ZFS_PROP_GROUPUSED ||
prop == ZFS_PROP_USERQUOTA || prop == ZFS_PROP_GROUPQUOTA ||
prop == ZFS_PROP_PROJECTUSED ||
prop == ZFS_PROP_PROJECTQUOTA) {
zfs_nicebytes(space, sizebuf, sizeof (sizebuf));
} else {
zfs_nicenum(space, sizebuf, sizeof (sizebuf));
}
} else {
(void) snprintf(sizebuf, sizeof (sizebuf), "%llu",
(u_longlong_t)space);
}
sizelen = strlen(sizebuf);
if (prop == ZFS_PROP_USERUSED || prop == ZFS_PROP_GROUPUSED ||
prop == ZFS_PROP_PROJECTUSED) {
propname = "used";
if (!nvlist_exists(props, "quota"))
(void) nvlist_add_uint64(props, "quota", 0);
} else if (prop == ZFS_PROP_USERQUOTA || prop == ZFS_PROP_GROUPQUOTA ||
prop == ZFS_PROP_PROJECTQUOTA) {
propname = "quota";
if (!nvlist_exists(props, "used"))
(void) nvlist_add_uint64(props, "used", 0);
} else if (prop == ZFS_PROP_USEROBJUSED ||
prop == ZFS_PROP_GROUPOBJUSED || prop == ZFS_PROP_PROJECTOBJUSED) {
propname = "objused";
if (!nvlist_exists(props, "objquota"))
(void) nvlist_add_uint64(props, "objquota", 0);
} else if (prop == ZFS_PROP_USEROBJQUOTA ||
prop == ZFS_PROP_GROUPOBJQUOTA ||
prop == ZFS_PROP_PROJECTOBJQUOTA) {
propname = "objquota";
if (!nvlist_exists(props, "objused"))
(void) nvlist_add_uint64(props, "objused", 0);
} else {
return (-1);
}
sizeidx = us_field_index(propname);
if (sizeidx >= 0 && sizelen > cb->cb_width[sizeidx])
cb->cb_width[sizeidx] = sizelen;
if (nvlist_add_uint64(props, propname, space) != 0)
nomem();
return (0);
}
static void
print_us_node(boolean_t scripted, boolean_t parsable, int *fields, int types,
size_t *width, us_node_t *node)
{
nvlist_t *nvl = node->usn_nvl;
char valstr[MAXNAMELEN];
boolean_t first = B_TRUE;
int cfield = 0;
int field;
uint32_t ustype;
/* Check type */
(void) nvlist_lookup_uint32(nvl, "type", &ustype);
if (!(ustype & types))
return;
while ((field = fields[cfield]) != USFIELD_LAST) {
nvpair_t *nvp = NULL;
data_type_t type;
uint32_t val32 = -1;
uint64_t val64 = -1;
const char *strval = "-";
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL)
if (strcmp(nvpair_name(nvp),
us_field_names[field]) == 0)
break;
type = nvp == NULL ? DATA_TYPE_UNKNOWN : nvpair_type(nvp);
switch (type) {
case DATA_TYPE_UINT32:
val32 = fnvpair_value_uint32(nvp);
break;
case DATA_TYPE_UINT64:
val64 = fnvpair_value_uint64(nvp);
break;
case DATA_TYPE_STRING:
strval = fnvpair_value_string(nvp);
break;
case DATA_TYPE_UNKNOWN:
break;
default:
(void) fprintf(stderr, "invalid data type\n");
}
switch (field) {
case USFIELD_TYPE:
if (type == DATA_TYPE_UINT32)
strval = us_type2str(val32);
break;
case USFIELD_NAME:
if (type == DATA_TYPE_UINT64) {
(void) sprintf(valstr, "%llu",
(u_longlong_t)val64);
strval = valstr;
}
break;
case USFIELD_USED:
case USFIELD_QUOTA:
if (type == DATA_TYPE_UINT64) {
if (parsable) {
(void) sprintf(valstr, "%llu",
(u_longlong_t)val64);
strval = valstr;
} else if (field == USFIELD_QUOTA &&
val64 == 0) {
strval = "none";
} else {
zfs_nicebytes(val64, valstr,
sizeof (valstr));
strval = valstr;
}
}
break;
case USFIELD_OBJUSED:
case USFIELD_OBJQUOTA:
if (type == DATA_TYPE_UINT64) {
if (parsable) {
(void) sprintf(valstr, "%llu",
(u_longlong_t)val64);
strval = valstr;
} else if (field == USFIELD_OBJQUOTA &&
val64 == 0) {
strval = "none";
} else {
zfs_nicenum(val64, valstr,
sizeof (valstr));
strval = valstr;
}
}
break;
}
if (!first) {
if (scripted)
(void) putchar('\t');
else
(void) fputs(" ", stdout);
}
if (scripted)
(void) fputs(strval, stdout);
else if (field == USFIELD_TYPE || field == USFIELD_NAME)
(void) printf("%-*s", (int)width[field], strval);
else
(void) printf("%*s", (int)width[field], strval);
first = B_FALSE;
cfield++;
}
(void) putchar('\n');
}
static void
print_us(boolean_t scripted, boolean_t parsable, int *fields, int types,
size_t *width, boolean_t rmnode, uu_avl_t *avl)
{
us_node_t *node;
const char *col;
int cfield = 0;
int field;
if (!scripted) {
boolean_t first = B_TRUE;
while ((field = fields[cfield]) != USFIELD_LAST) {
col = gettext(us_field_hdr[field]);
if (field == USFIELD_TYPE || field == USFIELD_NAME) {
(void) printf(first ? "%-*s" : " %-*s",
(int)width[field], col);
} else {
(void) printf(first ? "%*s" : " %*s",
(int)width[field], col);
}
first = B_FALSE;
cfield++;
}
(void) printf("\n");
}
for (node = uu_avl_first(avl); node; node = uu_avl_next(avl, node)) {
print_us_node(scripted, parsable, fields, types, width, node);
if (rmnode)
nvlist_free(node->usn_nvl);
}
}
static int
zfs_do_userspace(int argc, char **argv)
{
zfs_handle_t *zhp;
zfs_userquota_prop_t p;
uu_avl_pool_t *avl_pool;
uu_avl_t *avl_tree;
uu_avl_walk_t *walk;
char *delim;
char deffields[] = "type,name,used,quota,objused,objquota";
char *ofield = NULL;
char *tfield = NULL;
int cfield = 0;
int fields[256];
int i;
boolean_t scripted = B_FALSE;
boolean_t prtnum = B_FALSE;
boolean_t parsable = B_FALSE;
boolean_t sid2posix = B_FALSE;
int ret = 0;
int c;
zfs_sort_column_t *sortcol = NULL;
int types = USTYPE_PSX_USR | USTYPE_SMB_USR;
us_cbdata_t cb;
us_node_t *node;
us_node_t *rmnode;
uu_list_pool_t *listpool;
uu_list_t *list;
uu_avl_index_t idx = 0;
uu_list_index_t idx2 = 0;
if (argc < 2)
usage(B_FALSE);
if (strcmp(argv[0], "groupspace") == 0) {
/* Toggle default group types */
types = USTYPE_PSX_GRP | USTYPE_SMB_GRP;
} else if (strcmp(argv[0], "projectspace") == 0) {
types = USTYPE_PROJ;
prtnum = B_TRUE;
}
while ((c = getopt(argc, argv, "nHpo:s:S:t:i")) != -1) {
switch (c) {
case 'n':
if (types == USTYPE_PROJ) {
(void) fprintf(stderr,
gettext("invalid option 'n'\n"));
usage(B_FALSE);
}
prtnum = B_TRUE;
break;
case 'H':
scripted = B_TRUE;
break;
case 'p':
parsable = B_TRUE;
break;
case 'o':
ofield = optarg;
break;
case 's':
case 'S':
if (zfs_add_sort_column(&sortcol, optarg,
c == 's' ? B_FALSE : B_TRUE) != 0) {
(void) fprintf(stderr,
gettext("invalid field '%s'\n"), optarg);
usage(B_FALSE);
}
break;
case 't':
if (types == USTYPE_PROJ) {
(void) fprintf(stderr,
gettext("invalid option 't'\n"));
usage(B_FALSE);
}
tfield = optarg;
break;
case 'i':
if (types == USTYPE_PROJ) {
(void) fprintf(stderr,
gettext("invalid option 'i'\n"));
usage(B_FALSE);
}
sid2posix = B_TRUE;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing dataset name\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
/* Use default output fields if not specified using -o */
if (ofield == NULL)
ofield = deffields;
do {
if ((delim = strchr(ofield, ',')) != NULL)
*delim = '\0';
if ((fields[cfield++] = us_field_index(ofield)) == -1) {
(void) fprintf(stderr, gettext("invalid type '%s' "
"for -o option\n"), ofield);
return (-1);
}
if (delim != NULL)
ofield = delim + 1;
} while (delim != NULL);
fields[cfield] = USFIELD_LAST;
/* Override output types (-t option) */
if (tfield != NULL) {
types = 0;
do {
boolean_t found = B_FALSE;
if ((delim = strchr(tfield, ',')) != NULL)
*delim = '\0';
for (i = 0; i < sizeof (us_type_bits) / sizeof (int);
i++) {
if (strcmp(tfield, us_type_names[i]) == 0) {
found = B_TRUE;
types |= us_type_bits[i];
break;
}
}
if (!found) {
(void) fprintf(stderr, gettext("invalid type "
"'%s' for -t option\n"), tfield);
return (-1);
}
if (delim != NULL)
tfield = delim + 1;
} while (delim != NULL);
}
if ((zhp = zfs_path_to_zhandle(g_zfs, argv[0], ZFS_TYPE_FILESYSTEM |
ZFS_TYPE_SNAPSHOT)) == NULL)
return (1);
if (zfs_get_underlying_type(zhp) != ZFS_TYPE_FILESYSTEM) {
(void) fprintf(stderr, gettext("operation is only applicable "
"to filesystems and their snapshots\n"));
zfs_close(zhp);
return (1);
}
if ((avl_pool = uu_avl_pool_create("us_avl_pool", sizeof (us_node_t),
offsetof(us_node_t, usn_avlnode), us_compare, UU_DEFAULT)) == NULL)
nomem();
if ((avl_tree = uu_avl_create(avl_pool, NULL, UU_DEFAULT)) == NULL)
nomem();
/* Always add default sorting columns */
(void) zfs_add_sort_column(&sortcol, "type", B_FALSE);
(void) zfs_add_sort_column(&sortcol, "name", B_FALSE);
cb.cb_sortcol = sortcol;
cb.cb_numname = prtnum;
cb.cb_nicenum = !parsable;
cb.cb_avl_pool = avl_pool;
cb.cb_avl = avl_tree;
cb.cb_sid2posix = sid2posix;
for (i = 0; i < USFIELD_LAST; i++)
cb.cb_width[i] = strlen(gettext(us_field_hdr[i]));
for (p = 0; p < ZFS_NUM_USERQUOTA_PROPS; p++) {
if ((zfs_prop_is_user(p) &&
!(types & (USTYPE_PSX_USR | USTYPE_SMB_USR))) ||
(zfs_prop_is_group(p) &&
!(types & (USTYPE_PSX_GRP | USTYPE_SMB_GRP))) ||
(zfs_prop_is_project(p) && types != USTYPE_PROJ))
continue;
cb.cb_prop = p;
if ((ret = zfs_userspace(zhp, p, userspace_cb, &cb)) != 0) {
zfs_close(zhp);
return (ret);
}
}
zfs_close(zhp);
/* Sort the list */
if ((node = uu_avl_first(avl_tree)) == NULL)
return (0);
us_populated = B_TRUE;
listpool = uu_list_pool_create("tmplist", sizeof (us_node_t),
offsetof(us_node_t, usn_listnode), NULL, UU_DEFAULT);
list = uu_list_create(listpool, NULL, UU_DEFAULT);
uu_list_node_init(node, &node->usn_listnode, listpool);
while (node != NULL) {
rmnode = node;
node = uu_avl_next(avl_tree, node);
uu_avl_remove(avl_tree, rmnode);
if (uu_list_find(list, rmnode, NULL, &idx2) == NULL)
uu_list_insert(list, rmnode, idx2);
}
for (node = uu_list_first(list); node != NULL;
node = uu_list_next(list, node)) {
us_sort_info_t sortinfo = { sortcol, cb.cb_numname };
if (uu_avl_find(avl_tree, node, &sortinfo, &idx) == NULL)
uu_avl_insert(avl_tree, node, idx);
}
uu_list_destroy(list);
uu_list_pool_destroy(listpool);
/* Print and free node nvlist memory */
print_us(scripted, parsable, fields, types, cb.cb_width, B_TRUE,
cb.cb_avl);
zfs_free_sort_columns(sortcol);
/* Clean up the AVL tree */
if ((walk = uu_avl_walk_start(cb.cb_avl, UU_WALK_ROBUST)) == NULL)
nomem();
while ((node = uu_avl_walk_next(walk)) != NULL) {
uu_avl_remove(cb.cb_avl, node);
free(node);
}
uu_avl_walk_end(walk);
uu_avl_destroy(avl_tree);
uu_avl_pool_destroy(avl_pool);
return (ret);
}
/*
* list [-Hp][-r|-d max] [-o property[,...]] [-s property] ... [-S property]
* [-t type[,...]] [filesystem|volume|snapshot] ...
*
* -H Scripted mode; elide headers and separate columns by tabs
* -p Display values in parsable (literal) format.
* -r Recurse over all children
* -d Limit recursion by depth.
* -o Control which fields to display.
* -s Specify sort columns, descending order.
* -S Specify sort columns, ascending order.
* -t Control which object types to display.
*
* When given no arguments, list all filesystems in the system.
* Otherwise, list the specified datasets, optionally recursing down them if
* '-r' is specified.
*/
typedef struct list_cbdata {
boolean_t cb_first;
boolean_t cb_literal;
boolean_t cb_scripted;
zprop_list_t *cb_proplist;
} list_cbdata_t;
/*
* Given a list of columns to display, output appropriate headers for each one.
*/
static void
print_header(list_cbdata_t *cb)
{
zprop_list_t *pl = cb->cb_proplist;
char headerbuf[ZFS_MAXPROPLEN];
const char *header;
int i;
boolean_t first = B_TRUE;
boolean_t right_justify;
color_start(ANSI_BOLD);
for (; pl != NULL; pl = pl->pl_next) {
if (!first) {
(void) printf(" ");
} else {
first = B_FALSE;
}
right_justify = B_FALSE;
if (pl->pl_prop != ZPROP_USERPROP) {
header = zfs_prop_column_name(pl->pl_prop);
right_justify = zfs_prop_align_right(pl->pl_prop);
} else {
for (i = 0; pl->pl_user_prop[i] != '\0'; i++)
headerbuf[i] = toupper(pl->pl_user_prop[i]);
headerbuf[i] = '\0';
header = headerbuf;
}
if (pl->pl_next == NULL && !right_justify)
(void) printf("%s", header);
else if (right_justify)
(void) printf("%*s", (int)pl->pl_width, header);
else
(void) printf("%-*s", (int)pl->pl_width, header);
}
color_end();
(void) printf("\n");
}
/*
* Decides on the color that the avail value should be printed in.
* > 80% used = yellow
* > 90% used = red
*/
static const char *
zfs_list_avail_color(zfs_handle_t *zhp)
{
uint64_t used = zfs_prop_get_int(zhp, ZFS_PROP_USED);
uint64_t avail = zfs_prop_get_int(zhp, ZFS_PROP_AVAILABLE);
int percentage = (int)((double)avail / MAX(avail + used, 1) * 100);
if (percentage > 20)
return (NULL);
else if (percentage > 10)
return (ANSI_YELLOW);
else
return (ANSI_RED);
}
/*
* Given a dataset and a list of fields, print out all the properties according
* to the described layout.
*/
static void
print_dataset(zfs_handle_t *zhp, list_cbdata_t *cb)
{
zprop_list_t *pl = cb->cb_proplist;
boolean_t first = B_TRUE;
char property[ZFS_MAXPROPLEN];
nvlist_t *userprops = zfs_get_user_props(zhp);
nvlist_t *propval;
const char *propstr;
boolean_t right_justify;
for (; pl != NULL; pl = pl->pl_next) {
if (!first) {
if (cb->cb_scripted)
(void) putchar('\t');
else
(void) fputs(" ", stdout);
} else {
first = B_FALSE;
}
if (pl->pl_prop == ZFS_PROP_NAME) {
(void) strlcpy(property, zfs_get_name(zhp),
sizeof (property));
propstr = property;
right_justify = zfs_prop_align_right(pl->pl_prop);
} else if (pl->pl_prop != ZPROP_USERPROP) {
if (zfs_prop_get(zhp, pl->pl_prop, property,
sizeof (property), NULL, NULL, 0,
cb->cb_literal) != 0)
propstr = "-";
else
propstr = property;
right_justify = zfs_prop_align_right(pl->pl_prop);
} else if (zfs_prop_userquota(pl->pl_user_prop)) {
if (zfs_prop_get_userquota(zhp, pl->pl_user_prop,
property, sizeof (property), cb->cb_literal) != 0)
propstr = "-";
else
propstr = property;
right_justify = B_TRUE;
} else if (zfs_prop_written(pl->pl_user_prop)) {
if (zfs_prop_get_written(zhp, pl->pl_user_prop,
property, sizeof (property), cb->cb_literal) != 0)
propstr = "-";
else
propstr = property;
right_justify = B_TRUE;
} else {
if (nvlist_lookup_nvlist(userprops,
pl->pl_user_prop, &propval) != 0)
propstr = "-";
else
propstr = fnvlist_lookup_string(propval,
ZPROP_VALUE);
right_justify = B_FALSE;
}
- if (pl->pl_prop == ZFS_PROP_AVAILABLE)
- color_start(zfs_list_avail_color(zhp));
+ /*
+ * zfs_list_avail_color() needs ZFS_PROP_AVAILABLE + USED
+ * - so we need another for() search for the USED part
+ * - when no colors wanted, we can skip the whole thing
+ */
+ if (use_color() && pl->pl_prop == ZFS_PROP_AVAILABLE) {
+ zprop_list_t *pl2 = cb->cb_proplist;
+ for (; pl2 != NULL; pl2 = pl2->pl_next) {
+ if (pl2->pl_prop == ZFS_PROP_USED) {
+ color_start(zfs_list_avail_color(zhp));
+ /* found it, no need for more loops */
+ break;
+ }
+ }
+ }
/*
* If this is being called in scripted mode, or if this is the
* last column and it is left-justified, don't include a width
* format specifier.
*/
if (cb->cb_scripted || (pl->pl_next == NULL && !right_justify))
(void) fputs(propstr, stdout);
else if (right_justify)
(void) printf("%*s", (int)pl->pl_width, propstr);
else
(void) printf("%-*s", (int)pl->pl_width, propstr);
if (pl->pl_prop == ZFS_PROP_AVAILABLE)
color_end();
}
(void) putchar('\n');
}
/*
* Generic callback function to list a dataset or snapshot.
*/
static int
list_callback(zfs_handle_t *zhp, void *data)
{
list_cbdata_t *cbp = data;
if (cbp->cb_first) {
if (!cbp->cb_scripted)
print_header(cbp);
cbp->cb_first = B_FALSE;
}
print_dataset(zhp, cbp);
return (0);
}
static int
zfs_do_list(int argc, char **argv)
{
int c;
char default_fields[] =
"name,used,available,referenced,mountpoint";
int types = ZFS_TYPE_DATASET;
boolean_t types_specified = B_FALSE;
char *fields = default_fields;
list_cbdata_t cb = { 0 };
int limit = 0;
int ret = 0;
zfs_sort_column_t *sortcol = NULL;
int flags = ZFS_ITER_PROP_LISTSNAPS | ZFS_ITER_ARGS_CAN_BE_PATHS;
/* check options */
while ((c = getopt(argc, argv, "HS:d:o:prs:t:")) != -1) {
switch (c) {
case 'o':
fields = optarg;
break;
case 'p':
cb.cb_literal = B_TRUE;
flags |= ZFS_ITER_LITERAL_PROPS;
break;
case 'd':
limit = parse_depth(optarg, &flags);
break;
case 'r':
flags |= ZFS_ITER_RECURSE;
break;
case 'H':
cb.cb_scripted = B_TRUE;
break;
case 's':
if (zfs_add_sort_column(&sortcol, optarg,
B_FALSE) != 0) {
(void) fprintf(stderr,
gettext("invalid property '%s'\n"), optarg);
usage(B_FALSE);
}
break;
case 'S':
if (zfs_add_sort_column(&sortcol, optarg,
B_TRUE) != 0) {
(void) fprintf(stderr,
gettext("invalid property '%s'\n"), optarg);
usage(B_FALSE);
}
break;
case 't':
types = 0;
types_specified = B_TRUE;
flags &= ~ZFS_ITER_PROP_LISTSNAPS;
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const type_subopts[] = {
"filesystem", "volume",
"snapshot", "snap",
"bookmark",
"all" };
static const int type_types[] = {
ZFS_TYPE_FILESYSTEM, ZFS_TYPE_VOLUME,
ZFS_TYPE_SNAPSHOT, ZFS_TYPE_SNAPSHOT,
ZFS_TYPE_BOOKMARK,
ZFS_TYPE_DATASET | ZFS_TYPE_BOOKMARK };
for (c = 0; c < ARRAY_SIZE(type_subopts); ++c)
if (strcmp(tok, type_subopts[c]) == 0) {
types |= type_types[c];
goto found3;
}
(void) fprintf(stderr,
gettext("invalid type '%s'\n"), tok);
usage(B_FALSE);
found3:;
}
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/*
* If "-o space" and no types were specified, don't display snapshots.
*/
if (strcmp(fields, "space") == 0 && types_specified == B_FALSE)
types &= ~ZFS_TYPE_SNAPSHOT;
/*
* Handle users who want to list all snapshots or bookmarks
* of the current dataset (ex. 'zfs list -t snapshot <dataset>').
*/
if ((types == ZFS_TYPE_SNAPSHOT || types == ZFS_TYPE_BOOKMARK) &&
argc > 0 && (flags & ZFS_ITER_RECURSE) == 0 && limit == 0) {
flags |= (ZFS_ITER_DEPTH_LIMIT | ZFS_ITER_RECURSE);
limit = 1;
}
/*
* If the user specifies '-o all', the zprop_get_list() doesn't
* normally include the name of the dataset. For 'zfs list', we always
* want this property to be first.
*/
if (zprop_get_list(g_zfs, fields, &cb.cb_proplist, ZFS_TYPE_DATASET)
!= 0)
usage(B_FALSE);
cb.cb_first = B_TRUE;
/*
* If we are only going to list and sort by properties that are "fast"
* then we can use "simple" mode and avoid populating the properties
* nvlist.
*/
if (zfs_list_only_by_fast(cb.cb_proplist) &&
zfs_sort_only_by_fast(sortcol))
flags |= ZFS_ITER_SIMPLE;
ret = zfs_for_each(argc, argv, flags, types, sortcol, &cb.cb_proplist,
limit, list_callback, &cb);
zprop_free_list(cb.cb_proplist);
zfs_free_sort_columns(sortcol);
if (ret == 0 && cb.cb_first && !cb.cb_scripted)
(void) fprintf(stderr, gettext("no datasets available\n"));
return (ret);
}
/*
* zfs rename [-fu] <fs | snap | vol> <fs | snap | vol>
* zfs rename [-f] -p <fs | vol> <fs | vol>
* zfs rename [-u] -r <snap> <snap>
*
* Renames the given dataset to another of the same type.
*
* The '-p' flag creates all the non-existing ancestors of the target first.
* The '-u' flag prevents file systems from being remounted during rename.
*/
static int
zfs_do_rename(int argc, char **argv)
{
zfs_handle_t *zhp;
renameflags_t flags = { 0 };
int c;
int ret = 0;
int types;
boolean_t parents = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "pruf")) != -1) {
switch (c) {
case 'p':
parents = B_TRUE;
break;
case 'r':
flags.recursive = B_TRUE;
break;
case 'u':
flags.nounmount = B_TRUE;
break;
case 'f':
flags.forceunmount = B_TRUE;
break;
case '?':
default:
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing source dataset "
"argument\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing target dataset "
"argument\n"));
usage(B_FALSE);
}
if (argc > 2) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if (flags.recursive && parents) {
(void) fprintf(stderr, gettext("-p and -r options are mutually "
"exclusive\n"));
usage(B_FALSE);
}
if (flags.nounmount && parents) {
(void) fprintf(stderr, gettext("-u and -p options are mutually "
"exclusive\n"));
usage(B_FALSE);
}
if (flags.recursive && strchr(argv[0], '@') == 0) {
(void) fprintf(stderr, gettext("source dataset for recursive "
"rename must be a snapshot\n"));
usage(B_FALSE);
}
if (flags.nounmount)
types = ZFS_TYPE_FILESYSTEM;
else if (parents)
types = ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME;
else
types = ZFS_TYPE_DATASET;
if ((zhp = zfs_open(g_zfs, argv[0], types)) == NULL)
return (1);
/* If we were asked and the name looks good, try to create ancestors. */
if (parents && zfs_name_valid(argv[1], zfs_get_type(zhp)) &&
zfs_create_ancestors(g_zfs, argv[1]) != 0) {
zfs_close(zhp);
return (1);
}
ret = (zfs_rename(zhp, argv[1], flags) != 0);
zfs_close(zhp);
return (ret);
}
/*
* zfs promote <fs>
*
* Promotes the given clone fs to be the parent
*/
static int
zfs_do_promote(int argc, char **argv)
{
zfs_handle_t *zhp;
int ret = 0;
/* check options */
if (argc > 1 && argv[1][0] == '-') {
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
argv[1][1]);
usage(B_FALSE);
}
/* check number of arguments */
if (argc < 2) {
(void) fprintf(stderr, gettext("missing clone filesystem"
" argument\n"));
usage(B_FALSE);
}
if (argc > 2) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
zhp = zfs_open(g_zfs, argv[1], ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL)
return (1);
ret = (zfs_promote(zhp) != 0);
zfs_close(zhp);
return (ret);
}
static int
zfs_do_redact(int argc, char **argv)
{
char *snap = NULL;
char *bookname = NULL;
char **rsnaps = NULL;
int numrsnaps = 0;
argv++;
argc--;
if (argc < 3) {
(void) fprintf(stderr, gettext("too few arguments\n"));
usage(B_FALSE);
}
snap = argv[0];
bookname = argv[1];
rsnaps = argv + 2;
numrsnaps = argc - 2;
nvlist_t *rsnapnv = fnvlist_alloc();
for (int i = 0; i < numrsnaps; i++) {
fnvlist_add_boolean(rsnapnv, rsnaps[i]);
}
int err = lzc_redact(snap, bookname, rsnapnv);
fnvlist_free(rsnapnv);
switch (err) {
case 0:
break;
case ENOENT: {
zfs_handle_t *zhp = zfs_open(g_zfs, snap, ZFS_TYPE_SNAPSHOT);
if (zhp == NULL) {
(void) fprintf(stderr, gettext("provided snapshot %s "
"does not exist\n"), snap);
} else {
zfs_close(zhp);
}
for (int i = 0; i < numrsnaps; i++) {
zhp = zfs_open(g_zfs, rsnaps[i], ZFS_TYPE_SNAPSHOT);
if (zhp == NULL) {
(void) fprintf(stderr, gettext("provided "
"snapshot %s does not exist\n"), rsnaps[i]);
} else {
zfs_close(zhp);
}
}
break;
}
case EEXIST:
(void) fprintf(stderr, gettext("specified redaction bookmark "
"(%s) provided already exists\n"), bookname);
break;
case ENAMETOOLONG:
(void) fprintf(stderr, gettext("provided bookmark name cannot "
"be used, final name would be too long\n"));
break;
case E2BIG:
(void) fprintf(stderr, gettext("too many redaction snapshots "
"specified\n"));
break;
case EINVAL:
if (strchr(bookname, '#') != NULL)
(void) fprintf(stderr, gettext(
"redaction bookmark name must not contain '#'\n"));
else
(void) fprintf(stderr, gettext(
"redaction snapshot must be descendent of "
"snapshot being redacted\n"));
break;
case EALREADY:
(void) fprintf(stderr, gettext("attempted to redact redacted "
"dataset or with respect to redacted dataset\n"));
break;
case ENOTSUP:
(void) fprintf(stderr, gettext("redaction bookmarks feature "
"not enabled\n"));
break;
case EXDEV:
(void) fprintf(stderr, gettext("potentially invalid redaction "
"snapshot; full dataset names required\n"));
break;
default:
(void) fprintf(stderr, gettext("internal error: %s\n"),
strerror(errno));
}
return (err);
}
/*
* zfs rollback [-rRf] <snapshot>
*
* -r Delete any intervening snapshots before doing rollback
* -R Delete any snapshots and their clones
* -f ignored for backwards compatibility
*
* Given a filesystem, rollback to a specific snapshot, discarding any changes
* since then and making it the active dataset. If more recent snapshots exist,
* the command will complain unless the '-r' flag is given.
*/
typedef struct rollback_cbdata {
uint64_t cb_create;
uint8_t cb_younger_ds_printed;
boolean_t cb_first;
int cb_doclones;
char *cb_target;
int cb_error;
boolean_t cb_recurse;
} rollback_cbdata_t;
static int
rollback_check_dependent(zfs_handle_t *zhp, void *data)
{
rollback_cbdata_t *cbp = data;
if (cbp->cb_first && cbp->cb_recurse) {
(void) fprintf(stderr, gettext("cannot rollback to "
"'%s': clones of previous snapshots exist\n"),
cbp->cb_target);
(void) fprintf(stderr, gettext("use '-R' to "
"force deletion of the following clones and "
"dependents:\n"));
cbp->cb_first = 0;
cbp->cb_error = 1;
}
(void) fprintf(stderr, "%s\n", zfs_get_name(zhp));
zfs_close(zhp);
return (0);
}
/*
* Report some snapshots/bookmarks more recent than the one specified.
* Used when '-r' is not specified. We reuse this same callback for the
* snapshot dependents - if 'cb_dependent' is set, then this is a
* dependent and we should report it without checking the transaction group.
*/
static int
rollback_check(zfs_handle_t *zhp, void *data)
{
rollback_cbdata_t *cbp = data;
/*
* Max number of younger snapshots and/or bookmarks to display before
* we stop the iteration.
*/
const uint8_t max_younger = 32;
if (cbp->cb_doclones) {
zfs_close(zhp);
return (0);
}
if (zfs_prop_get_int(zhp, ZFS_PROP_CREATETXG) > cbp->cb_create) {
if (cbp->cb_first && !cbp->cb_recurse) {
(void) fprintf(stderr, gettext("cannot "
"rollback to '%s': more recent snapshots "
"or bookmarks exist\n"),
cbp->cb_target);
(void) fprintf(stderr, gettext("use '-r' to "
"force deletion of the following "
"snapshots and bookmarks:\n"));
cbp->cb_first = 0;
cbp->cb_error = 1;
}
if (cbp->cb_recurse) {
- if (zfs_iter_dependents(zhp, 0, B_TRUE,
+ if (zfs_iter_dependents_v2(zhp, 0, B_TRUE,
rollback_check_dependent, cbp) != 0) {
zfs_close(zhp);
return (-1);
}
} else {
(void) fprintf(stderr, "%s\n",
zfs_get_name(zhp));
cbp->cb_younger_ds_printed++;
}
}
zfs_close(zhp);
if (cbp->cb_younger_ds_printed == max_younger) {
/*
* This non-recursive rollback is going to fail due to the
* presence of snapshots and/or bookmarks that are younger than
* the rollback target.
* We printed some of the offending objects, now we stop
* zfs_iter_snapshot/bookmark iteration so we can fail fast and
* avoid iterating over the rest of the younger objects
*/
(void) fprintf(stderr, gettext("Output limited to %d "
"snapshots/bookmarks\n"), max_younger);
return (-1);
}
return (0);
}
static int
zfs_do_rollback(int argc, char **argv)
{
int ret = 0;
int c;
boolean_t force = B_FALSE;
rollback_cbdata_t cb = { 0 };
zfs_handle_t *zhp, *snap;
char parentname[ZFS_MAX_DATASET_NAME_LEN];
char *delim;
uint64_t min_txg = 0;
/* check options */
while ((c = getopt(argc, argv, "rRf")) != -1) {
switch (c) {
case 'r':
cb.cb_recurse = 1;
break;
case 'R':
cb.cb_recurse = 1;
cb.cb_doclones = 1;
break;
case 'f':
force = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing dataset argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
/* open the snapshot */
if ((snap = zfs_open(g_zfs, argv[0], ZFS_TYPE_SNAPSHOT)) == NULL)
return (1);
/* open the parent dataset */
(void) strlcpy(parentname, argv[0], sizeof (parentname));
verify((delim = strrchr(parentname, '@')) != NULL);
*delim = '\0';
if ((zhp = zfs_open(g_zfs, parentname, ZFS_TYPE_DATASET)) == NULL) {
zfs_close(snap);
return (1);
}
/*
* Check for more recent snapshots and/or clones based on the presence
* of '-r' and '-R'.
*/
cb.cb_target = argv[0];
cb.cb_create = zfs_prop_get_int(snap, ZFS_PROP_CREATETXG);
cb.cb_first = B_TRUE;
cb.cb_error = 0;
if (cb.cb_create > 0)
min_txg = cb.cb_create;
- if ((ret = zfs_iter_snapshots(zhp, 0, rollback_check, &cb,
+ if ((ret = zfs_iter_snapshots_v2(zhp, 0, rollback_check, &cb,
min_txg, 0)) != 0)
goto out;
- if ((ret = zfs_iter_bookmarks(zhp, 0, rollback_check, &cb)) != 0)
+ if ((ret = zfs_iter_bookmarks_v2(zhp, 0, rollback_check, &cb)) != 0)
goto out;
if ((ret = cb.cb_error) != 0)
goto out;
/*
* Rollback parent to the given snapshot.
*/
ret = zfs_rollback(zhp, snap, force);
out:
zfs_close(snap);
zfs_close(zhp);
if (ret == 0)
return (0);
else
return (1);
}
/*
* zfs set property=value ... { fs | snap | vol } ...
*
* Sets the given properties for all datasets specified on the command line.
*/
static int
set_callback(zfs_handle_t *zhp, void *data)
{
nvlist_t *props = data;
if (zfs_prop_set_list(zhp, props) != 0) {
switch (libzfs_errno(g_zfs)) {
case EZFS_MOUNTFAILED:
(void) fprintf(stderr, gettext("property may be set "
"but unable to remount filesystem\n"));
break;
case EZFS_SHARENFSFAILED:
(void) fprintf(stderr, gettext("property may be set "
"but unable to reshare filesystem\n"));
break;
}
return (1);
}
return (0);
}
static int
zfs_do_set(int argc, char **argv)
{
nvlist_t *props = NULL;
int ds_start = -1; /* argv idx of first dataset arg */
int ret = 0;
int i;
/* check for options */
if (argc > 1 && argv[1][0] == '-') {
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
argv[1][1]);
usage(B_FALSE);
}
/* check number of arguments */
if (argc < 2) {
(void) fprintf(stderr, gettext("missing arguments\n"));
usage(B_FALSE);
}
if (argc < 3) {
if (strchr(argv[1], '=') == NULL) {
(void) fprintf(stderr, gettext("missing property=value "
"argument(s)\n"));
} else {
(void) fprintf(stderr, gettext("missing dataset "
"name(s)\n"));
}
usage(B_FALSE);
}
/* validate argument order: prop=val args followed by dataset args */
for (i = 1; i < argc; i++) {
if (strchr(argv[i], '=') != NULL) {
if (ds_start > 0) {
/* out-of-order prop=val argument */
(void) fprintf(stderr, gettext("invalid "
"argument order\n"));
usage(B_FALSE);
}
} else if (ds_start < 0) {
ds_start = i;
}
}
if (ds_start < 0) {
(void) fprintf(stderr, gettext("missing dataset name(s)\n"));
usage(B_FALSE);
}
/* Populate a list of property settings */
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
nomem();
for (i = 1; i < ds_start; i++) {
if (!parseprop(props, argv[i])) {
ret = -1;
goto error;
}
}
ret = zfs_for_each(argc - ds_start, argv + ds_start, 0,
ZFS_TYPE_DATASET, NULL, NULL, 0, set_callback, props);
error:
nvlist_free(props);
return (ret);
}
typedef struct snap_cbdata {
nvlist_t *sd_nvl;
boolean_t sd_recursive;
const char *sd_snapname;
} snap_cbdata_t;
static int
zfs_snapshot_cb(zfs_handle_t *zhp, void *arg)
{
snap_cbdata_t *sd = arg;
char *name;
int rv = 0;
int error;
if (sd->sd_recursive &&
zfs_prop_get_int(zhp, ZFS_PROP_INCONSISTENT) != 0) {
zfs_close(zhp);
return (0);
}
error = asprintf(&name, "%s@%s", zfs_get_name(zhp), sd->sd_snapname);
if (error == -1)
nomem();
fnvlist_add_boolean(sd->sd_nvl, name);
free(name);
if (sd->sd_recursive)
- rv = zfs_iter_filesystems(zhp, 0, zfs_snapshot_cb, sd);
+ rv = zfs_iter_filesystems_v2(zhp, 0, zfs_snapshot_cb, sd);
zfs_close(zhp);
return (rv);
}
/*
* zfs snapshot [-r] [-o prop=value] ... <fs@snap>
*
* Creates a snapshot with the given name. While functionally equivalent to
* 'zfs create', it is a separate command to differentiate intent.
*/
static int
zfs_do_snapshot(int argc, char **argv)
{
int ret = 0;
int c;
nvlist_t *props;
snap_cbdata_t sd = { 0 };
boolean_t multiple_snaps = B_FALSE;
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
nomem();
if (nvlist_alloc(&sd.sd_nvl, NV_UNIQUE_NAME, 0) != 0)
nomem();
/* check options */
while ((c = getopt(argc, argv, "ro:")) != -1) {
switch (c) {
case 'o':
if (!parseprop(props, optarg)) {
nvlist_free(sd.sd_nvl);
nvlist_free(props);
return (1);
}
break;
case 'r':
sd.sd_recursive = B_TRUE;
multiple_snaps = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
goto usage;
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing snapshot argument\n"));
goto usage;
}
if (argc > 1)
multiple_snaps = B_TRUE;
for (; argc > 0; argc--, argv++) {
char *atp;
zfs_handle_t *zhp;
atp = strchr(argv[0], '@');
if (atp == NULL)
goto usage;
*atp = '\0';
sd.sd_snapname = atp + 1;
zhp = zfs_open(g_zfs, argv[0],
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL)
goto usage;
if (zfs_snapshot_cb(zhp, &sd) != 0)
goto usage;
}
ret = zfs_snapshot_nvl(g_zfs, sd.sd_nvl, props);
nvlist_free(sd.sd_nvl);
nvlist_free(props);
if (ret != 0 && multiple_snaps)
(void) fprintf(stderr, gettext("no snapshots were created\n"));
return (ret != 0);
usage:
nvlist_free(sd.sd_nvl);
nvlist_free(props);
usage(B_FALSE);
return (-1);
}
/*
* Array of prefixes to exclude –
* a linear search, even if executed for each dataset,
* is plenty good enough.
*/
typedef struct zfs_send_exclude_arg {
size_t count;
const char **list;
} zfs_send_exclude_arg_t;
static boolean_t
zfs_do_send_exclude(zfs_handle_t *zhp, void *context)
{
zfs_send_exclude_arg_t *excludes = context;
const char *name = zfs_get_name(zhp);
for (size_t i = 0; i < excludes->count; ++i) {
size_t len = strlen(excludes->list[i]);
if (strncmp(name, excludes->list[i], len) == 0 &&
memchr("/@", name[len], sizeof ("/@")))
return (B_FALSE);
}
return (B_TRUE);
}
/*
* Send a backup stream to stdout.
*/
static int
zfs_do_send(int argc, char **argv)
{
char *fromname = NULL;
char *toname = NULL;
char *resume_token = NULL;
char *cp;
zfs_handle_t *zhp;
sendflags_t flags = { 0 };
int c, err;
nvlist_t *dbgnv = NULL;
char *redactbook = NULL;
zfs_send_exclude_arg_t excludes = { 0 };
struct option long_options[] = {
{"replicate", no_argument, NULL, 'R'},
{"skip-missing", no_argument, NULL, 's'},
{"redact", required_argument, NULL, 'd'},
{"props", no_argument, NULL, 'p'},
{"parsable", no_argument, NULL, 'P'},
{"dedup", no_argument, NULL, 'D'},
{"proctitle", no_argument, NULL, 'V'},
{"verbose", no_argument, NULL, 'v'},
{"dryrun", no_argument, NULL, 'n'},
{"large-block", no_argument, NULL, 'L'},
{"embed", no_argument, NULL, 'e'},
{"resume", required_argument, NULL, 't'},
{"compressed", no_argument, NULL, 'c'},
{"raw", no_argument, NULL, 'w'},
{"backup", no_argument, NULL, 'b'},
{"holds", no_argument, NULL, 'h'},
{"saved", no_argument, NULL, 'S'},
{"exclude", required_argument, NULL, 'X'},
{0, 0, 0, 0}
};
/* check options */
while ((c = getopt_long(argc, argv, ":i:I:RsDpVvnPLeht:cwbd:SX:",
long_options, NULL)) != -1) {
switch (c) {
case 'X':
for (char *ds; (ds = strsep(&optarg, ",")) != NULL; ) {
if (!zfs_name_valid(ds, ZFS_TYPE_DATASET) ||
strchr(ds, '/') == NULL) {
(void) fprintf(stderr, gettext("-X %s: "
"not a valid non-root dataset name"
".\n"), ds);
usage(B_FALSE);
}
excludes.list = safe_realloc(excludes.list,
sizeof (char *) * (excludes.count + 1));
excludes.list[excludes.count++] = ds;
}
break;
case 'i':
if (fromname)
usage(B_FALSE);
fromname = optarg;
break;
case 'I':
if (fromname)
usage(B_FALSE);
fromname = optarg;
flags.doall = B_TRUE;
break;
case 'R':
flags.replicate = B_TRUE;
break;
case 's':
flags.skipmissing = B_TRUE;
break;
case 'd':
redactbook = optarg;
break;
case 'p':
flags.props = B_TRUE;
break;
case 'b':
flags.backup = B_TRUE;
break;
case 'h':
flags.holds = B_TRUE;
break;
case 'P':
flags.parsable = B_TRUE;
break;
case 'V':
flags.progressastitle = B_TRUE;
break;
case 'v':
flags.verbosity++;
flags.progress = B_TRUE;
break;
case 'D':
(void) fprintf(stderr,
gettext("WARNING: deduplicated send is no "
"longer supported. A regular,\n"
"non-deduplicated stream will be generated.\n\n"));
break;
case 'n':
flags.dryrun = B_TRUE;
break;
case 'L':
flags.largeblock = B_TRUE;
break;
case 'e':
flags.embed_data = B_TRUE;
break;
case 't':
resume_token = optarg;
break;
case 'c':
flags.compress = B_TRUE;
break;
case 'w':
flags.raw = B_TRUE;
flags.compress = B_TRUE;
flags.embed_data = B_TRUE;
flags.largeblock = B_TRUE;
break;
case 'S':
flags.saved = B_TRUE;
break;
case ':':
/*
* If a parameter was not passed, optopt contains the
* value that would normally lead us into the
* appropriate case statement. If it's > 256, then this
* must be a longopt and we should look at argv to get
* the string. Otherwise it's just the character, so we
* should use it directly.
*/
if (optopt <= UINT8_MAX) {
(void) fprintf(stderr,
gettext("missing argument for '%c' "
"option\n"), optopt);
} else {
(void) fprintf(stderr,
gettext("missing argument for '%s' "
"option\n"), argv[optind - 1]);
}
free(excludes.list);
usage(B_FALSE);
break;
case '?':
default:
/*
* If an invalid flag was passed, optopt contains the
* character if it was a short flag, or 0 if it was a
* longopt.
*/
if (optopt != 0) {
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
} else {
(void) fprintf(stderr,
gettext("invalid option '%s'\n"),
argv[optind - 1]);
}
free(excludes.list);
usage(B_FALSE);
}
}
if ((flags.parsable || flags.progressastitle) && flags.verbosity == 0)
flags.verbosity = 1;
if (excludes.count > 0 && !flags.replicate) {
free(excludes.list);
(void) fprintf(stderr, gettext("Cannot specify "
"dataset exclusion (-X) on a non-recursive "
"send.\n"));
return (1);
}
argc -= optind;
argv += optind;
if (resume_token != NULL) {
if (fromname != NULL || flags.replicate || flags.props ||
flags.backup || flags.holds ||
flags.saved || redactbook != NULL) {
free(excludes.list);
(void) fprintf(stderr,
gettext("invalid flags combined with -t\n"));
usage(B_FALSE);
}
if (argc > 0) {
free(excludes.list);
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
} else {
if (argc < 1) {
free(excludes.list);
(void) fprintf(stderr,
gettext("missing snapshot argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
free(excludes.list);
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
}
if (flags.saved) {
if (fromname != NULL || flags.replicate || flags.props ||
flags.doall || flags.backup ||
flags.holds || flags.largeblock || flags.embed_data ||
flags.compress || flags.raw || redactbook != NULL) {
free(excludes.list);
(void) fprintf(stderr, gettext("incompatible flags "
"combined with saved send flag\n"));
usage(B_FALSE);
}
if (strchr(argv[0], '@') != NULL) {
free(excludes.list);
(void) fprintf(stderr, gettext("saved send must "
"specify the dataset with partially-received "
"state\n"));
usage(B_FALSE);
}
}
if (flags.raw && redactbook != NULL) {
free(excludes.list);
(void) fprintf(stderr,
gettext("Error: raw sends may not be redacted.\n"));
return (1);
}
if (!flags.dryrun && isatty(STDOUT_FILENO)) {
free(excludes.list);
(void) fprintf(stderr,
gettext("Error: Stream can not be written to a terminal.\n"
"You must redirect standard output.\n"));
return (1);
}
if (flags.saved) {
zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_DATASET);
if (zhp == NULL) {
free(excludes.list);
return (1);
}
err = zfs_send_saved(zhp, &flags, STDOUT_FILENO,
resume_token);
free(excludes.list);
zfs_close(zhp);
return (err != 0);
} else if (resume_token != NULL) {
free(excludes.list);
return (zfs_send_resume(g_zfs, &flags, STDOUT_FILENO,
resume_token));
}
if (flags.skipmissing && !flags.replicate) {
free(excludes.list);
(void) fprintf(stderr,
gettext("skip-missing flag can only be used in "
"conjunction with replicate\n"));
usage(B_FALSE);
}
/*
* For everything except -R and -I, use the new, cleaner code path.
*/
if (!(flags.replicate || flags.doall)) {
char frombuf[ZFS_MAX_DATASET_NAME_LEN];
if (fromname != NULL && (strchr(fromname, '#') == NULL &&
strchr(fromname, '@') == NULL)) {
/*
* Neither bookmark or snapshot was specified. Print a
* warning, and assume snapshot.
*/
(void) fprintf(stderr, "Warning: incremental source "
"didn't specify type, assuming snapshot. Use '@' "
"or '#' prefix to avoid ambiguity.\n");
(void) snprintf(frombuf, sizeof (frombuf), "@%s",
fromname);
fromname = frombuf;
}
if (fromname != NULL &&
(fromname[0] == '#' || fromname[0] == '@')) {
/*
* Incremental source name begins with # or @.
* Default to same fs as target.
*/
char tmpbuf[ZFS_MAX_DATASET_NAME_LEN];
(void) strlcpy(tmpbuf, fromname, sizeof (tmpbuf));
(void) strlcpy(frombuf, argv[0], sizeof (frombuf));
cp = strchr(frombuf, '@');
if (cp != NULL)
*cp = '\0';
(void) strlcat(frombuf, tmpbuf, sizeof (frombuf));
fromname = frombuf;
}
zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_DATASET);
if (zhp == NULL) {
free(excludes.list);
return (1);
}
err = zfs_send_one(zhp, fromname, STDOUT_FILENO, &flags,
redactbook);
free(excludes.list);
zfs_close(zhp);
return (err != 0);
}
if (fromname != NULL && strchr(fromname, '#')) {
(void) fprintf(stderr,
gettext("Error: multiple snapshots cannot be "
"sent from a bookmark.\n"));
free(excludes.list);
return (1);
}
if (redactbook != NULL) {
(void) fprintf(stderr, gettext("Error: multiple snapshots "
"cannot be sent redacted.\n"));
free(excludes.list);
return (1);
}
if ((cp = strchr(argv[0], '@')) == NULL) {
(void) fprintf(stderr, gettext("Error: "
"Unsupported flag with filesystem or bookmark.\n"));
free(excludes.list);
return (1);
}
*cp = '\0';
toname = cp + 1;
zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL) {
free(excludes.list);
return (1);
}
/*
* If they specified the full path to the snapshot, chop off
* everything except the short name of the snapshot, but special
* case if they specify the origin.
*/
if (fromname && (cp = strchr(fromname, '@')) != NULL) {
char origin[ZFS_MAX_DATASET_NAME_LEN];
zprop_source_t src;
(void) zfs_prop_get(zhp, ZFS_PROP_ORIGIN,
origin, sizeof (origin), &src, NULL, 0, B_FALSE);
if (strcmp(origin, fromname) == 0) {
fromname = NULL;
flags.fromorigin = B_TRUE;
} else {
*cp = '\0';
if (cp != fromname && strcmp(argv[0], fromname)) {
zfs_close(zhp);
free(excludes.list);
(void) fprintf(stderr,
gettext("incremental source must be "
"in same filesystem\n"));
usage(B_FALSE);
}
fromname = cp + 1;
if (strchr(fromname, '@') || strchr(fromname, '/')) {
zfs_close(zhp);
free(excludes.list);
(void) fprintf(stderr,
gettext("invalid incremental source\n"));
usage(B_FALSE);
}
}
}
if (flags.replicate && fromname == NULL)
flags.doall = B_TRUE;
err = zfs_send(zhp, fromname, toname, &flags, STDOUT_FILENO,
excludes.count > 0 ? zfs_do_send_exclude : NULL,
&excludes, flags.verbosity >= 3 ? &dbgnv : NULL);
if (flags.verbosity >= 3 && dbgnv != NULL) {
/*
* dump_nvlist prints to stdout, but that's been
* redirected to a file. Make it print to stderr
* instead.
*/
(void) dup2(STDERR_FILENO, STDOUT_FILENO);
dump_nvlist(dbgnv, 0);
nvlist_free(dbgnv);
}
zfs_close(zhp);
free(excludes.list);
return (err != 0);
}
/*
* Restore a backup stream from stdin.
*/
static int
zfs_do_receive(int argc, char **argv)
{
int c, err = 0;
recvflags_t flags = { 0 };
boolean_t abort_resumable = B_FALSE;
nvlist_t *props;
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
nomem();
/* check options */
while ((c = getopt(argc, argv, ":o:x:dehMnuvFsAc")) != -1) {
switch (c) {
case 'o':
if (!parseprop(props, optarg)) {
nvlist_free(props);
usage(B_FALSE);
}
break;
case 'x':
if (!parsepropname(props, optarg)) {
nvlist_free(props);
usage(B_FALSE);
}
break;
case 'd':
if (flags.istail) {
(void) fprintf(stderr, gettext("invalid option "
"combination: -d and -e are mutually "
"exclusive\n"));
usage(B_FALSE);
}
flags.isprefix = B_TRUE;
break;
case 'e':
if (flags.isprefix) {
(void) fprintf(stderr, gettext("invalid option "
"combination: -d and -e are mutually "
"exclusive\n"));
usage(B_FALSE);
}
flags.istail = B_TRUE;
break;
case 'h':
flags.skipholds = B_TRUE;
break;
case 'M':
flags.forceunmount = B_TRUE;
break;
case 'n':
flags.dryrun = B_TRUE;
break;
case 'u':
flags.nomount = B_TRUE;
break;
case 'v':
flags.verbose = B_TRUE;
break;
case 's':
flags.resumable = B_TRUE;
break;
case 'F':
flags.force = B_TRUE;
break;
case 'A':
abort_resumable = B_TRUE;
break;
case 'c':
flags.heal = B_TRUE;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* zfs recv -e (use "tail" name) implies -d (remove dataset "head") */
if (flags.istail)
flags.isprefix = B_TRUE;
/* check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing snapshot argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if (abort_resumable) {
if (flags.isprefix || flags.istail || flags.dryrun ||
flags.resumable || flags.nomount) {
(void) fprintf(stderr, gettext("invalid option\n"));
usage(B_FALSE);
}
char namebuf[ZFS_MAX_DATASET_NAME_LEN];
(void) snprintf(namebuf, sizeof (namebuf),
"%s/%%recv", argv[0]);
if (zfs_dataset_exists(g_zfs, namebuf,
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME)) {
zfs_handle_t *zhp = zfs_open(g_zfs,
namebuf, ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL) {
nvlist_free(props);
return (1);
}
err = zfs_destroy(zhp, B_FALSE);
zfs_close(zhp);
} else {
zfs_handle_t *zhp = zfs_open(g_zfs,
argv[0], ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL)
usage(B_FALSE);
if (!zfs_prop_get_int(zhp, ZFS_PROP_INCONSISTENT) ||
zfs_prop_get(zhp, ZFS_PROP_RECEIVE_RESUME_TOKEN,
NULL, 0, NULL, NULL, 0, B_TRUE) == -1) {
(void) fprintf(stderr,
gettext("'%s' does not have any "
"resumable receive state to abort\n"),
argv[0]);
nvlist_free(props);
zfs_close(zhp);
return (1);
}
err = zfs_destroy(zhp, B_FALSE);
zfs_close(zhp);
}
nvlist_free(props);
return (err != 0);
}
if (isatty(STDIN_FILENO)) {
(void) fprintf(stderr,
gettext("Error: Backup stream can not be read "
"from a terminal.\n"
"You must redirect standard input.\n"));
nvlist_free(props);
return (1);
}
err = zfs_receive(g_zfs, argv[0], props, &flags, STDIN_FILENO, NULL);
nvlist_free(props);
return (err != 0);
}
/*
* allow/unallow stuff
*/
/* copied from zfs/sys/dsl_deleg.h */
#define ZFS_DELEG_PERM_CREATE "create"
#define ZFS_DELEG_PERM_DESTROY "destroy"
#define ZFS_DELEG_PERM_SNAPSHOT "snapshot"
#define ZFS_DELEG_PERM_ROLLBACK "rollback"
#define ZFS_DELEG_PERM_CLONE "clone"
#define ZFS_DELEG_PERM_PROMOTE "promote"
#define ZFS_DELEG_PERM_RENAME "rename"
#define ZFS_DELEG_PERM_MOUNT "mount"
#define ZFS_DELEG_PERM_SHARE "share"
#define ZFS_DELEG_PERM_SEND "send"
#define ZFS_DELEG_PERM_RECEIVE "receive"
#define ZFS_DELEG_PERM_ALLOW "allow"
#define ZFS_DELEG_PERM_USERPROP "userprop"
#define ZFS_DELEG_PERM_VSCAN "vscan" /* ??? */
#define ZFS_DELEG_PERM_USERQUOTA "userquota"
#define ZFS_DELEG_PERM_GROUPQUOTA "groupquota"
#define ZFS_DELEG_PERM_USERUSED "userused"
#define ZFS_DELEG_PERM_GROUPUSED "groupused"
#define ZFS_DELEG_PERM_USEROBJQUOTA "userobjquota"
#define ZFS_DELEG_PERM_GROUPOBJQUOTA "groupobjquota"
#define ZFS_DELEG_PERM_USEROBJUSED "userobjused"
#define ZFS_DELEG_PERM_GROUPOBJUSED "groupobjused"
#define ZFS_DELEG_PERM_HOLD "hold"
#define ZFS_DELEG_PERM_RELEASE "release"
#define ZFS_DELEG_PERM_DIFF "diff"
#define ZFS_DELEG_PERM_BOOKMARK "bookmark"
#define ZFS_DELEG_PERM_LOAD_KEY "load-key"
#define ZFS_DELEG_PERM_CHANGE_KEY "change-key"
#define ZFS_DELEG_PERM_PROJECTUSED "projectused"
#define ZFS_DELEG_PERM_PROJECTQUOTA "projectquota"
#define ZFS_DELEG_PERM_PROJECTOBJUSED "projectobjused"
#define ZFS_DELEG_PERM_PROJECTOBJQUOTA "projectobjquota"
#define ZFS_NUM_DELEG_NOTES ZFS_DELEG_NOTE_NONE
static zfs_deleg_perm_tab_t zfs_deleg_perm_tbl[] = {
{ ZFS_DELEG_PERM_ALLOW, ZFS_DELEG_NOTE_ALLOW },
{ ZFS_DELEG_PERM_CLONE, ZFS_DELEG_NOTE_CLONE },
{ ZFS_DELEG_PERM_CREATE, ZFS_DELEG_NOTE_CREATE },
{ ZFS_DELEG_PERM_DESTROY, ZFS_DELEG_NOTE_DESTROY },
{ ZFS_DELEG_PERM_DIFF, ZFS_DELEG_NOTE_DIFF},
{ ZFS_DELEG_PERM_HOLD, ZFS_DELEG_NOTE_HOLD },
{ ZFS_DELEG_PERM_MOUNT, ZFS_DELEG_NOTE_MOUNT },
{ ZFS_DELEG_PERM_PROMOTE, ZFS_DELEG_NOTE_PROMOTE },
{ ZFS_DELEG_PERM_RECEIVE, ZFS_DELEG_NOTE_RECEIVE },
{ ZFS_DELEG_PERM_RELEASE, ZFS_DELEG_NOTE_RELEASE },
{ ZFS_DELEG_PERM_RENAME, ZFS_DELEG_NOTE_RENAME },
{ ZFS_DELEG_PERM_ROLLBACK, ZFS_DELEG_NOTE_ROLLBACK },
{ ZFS_DELEG_PERM_SEND, ZFS_DELEG_NOTE_SEND },
{ ZFS_DELEG_PERM_SHARE, ZFS_DELEG_NOTE_SHARE },
{ ZFS_DELEG_PERM_SNAPSHOT, ZFS_DELEG_NOTE_SNAPSHOT },
{ ZFS_DELEG_PERM_BOOKMARK, ZFS_DELEG_NOTE_BOOKMARK },
{ ZFS_DELEG_PERM_LOAD_KEY, ZFS_DELEG_NOTE_LOAD_KEY },
{ ZFS_DELEG_PERM_CHANGE_KEY, ZFS_DELEG_NOTE_CHANGE_KEY },
{ ZFS_DELEG_PERM_GROUPQUOTA, ZFS_DELEG_NOTE_GROUPQUOTA },
{ ZFS_DELEG_PERM_GROUPUSED, ZFS_DELEG_NOTE_GROUPUSED },
{ ZFS_DELEG_PERM_USERPROP, ZFS_DELEG_NOTE_USERPROP },
{ ZFS_DELEG_PERM_USERQUOTA, ZFS_DELEG_NOTE_USERQUOTA },
{ ZFS_DELEG_PERM_USERUSED, ZFS_DELEG_NOTE_USERUSED },
{ ZFS_DELEG_PERM_USEROBJQUOTA, ZFS_DELEG_NOTE_USEROBJQUOTA },
{ ZFS_DELEG_PERM_USEROBJUSED, ZFS_DELEG_NOTE_USEROBJUSED },
{ ZFS_DELEG_PERM_GROUPOBJQUOTA, ZFS_DELEG_NOTE_GROUPOBJQUOTA },
{ ZFS_DELEG_PERM_GROUPOBJUSED, ZFS_DELEG_NOTE_GROUPOBJUSED },
{ ZFS_DELEG_PERM_PROJECTUSED, ZFS_DELEG_NOTE_PROJECTUSED },
{ ZFS_DELEG_PERM_PROJECTQUOTA, ZFS_DELEG_NOTE_PROJECTQUOTA },
{ ZFS_DELEG_PERM_PROJECTOBJUSED, ZFS_DELEG_NOTE_PROJECTOBJUSED },
{ ZFS_DELEG_PERM_PROJECTOBJQUOTA, ZFS_DELEG_NOTE_PROJECTOBJQUOTA },
{ NULL, ZFS_DELEG_NOTE_NONE }
};
/* permission structure */
typedef struct deleg_perm {
zfs_deleg_who_type_t dp_who_type;
const char *dp_name;
boolean_t dp_local;
boolean_t dp_descend;
} deleg_perm_t;
/* */
typedef struct deleg_perm_node {
deleg_perm_t dpn_perm;
uu_avl_node_t dpn_avl_node;
} deleg_perm_node_t;
typedef struct fs_perm fs_perm_t;
/* permissions set */
typedef struct who_perm {
zfs_deleg_who_type_t who_type;
const char *who_name; /* id */
char who_ug_name[256]; /* user/group name */
fs_perm_t *who_fsperm; /* uplink */
uu_avl_t *who_deleg_perm_avl; /* permissions */
} who_perm_t;
/* */
typedef struct who_perm_node {
who_perm_t who_perm;
uu_avl_node_t who_avl_node;
} who_perm_node_t;
typedef struct fs_perm_set fs_perm_set_t;
/* fs permissions */
struct fs_perm {
const char *fsp_name;
uu_avl_t *fsp_sc_avl; /* sets,create */
uu_avl_t *fsp_uge_avl; /* user,group,everyone */
fs_perm_set_t *fsp_set; /* uplink */
};
/* */
typedef struct fs_perm_node {
fs_perm_t fspn_fsperm;
uu_avl_t *fspn_avl;
uu_list_node_t fspn_list_node;
} fs_perm_node_t;
/* top level structure */
struct fs_perm_set {
uu_list_pool_t *fsps_list_pool;
uu_list_t *fsps_list; /* list of fs_perms */
uu_avl_pool_t *fsps_named_set_avl_pool;
uu_avl_pool_t *fsps_who_perm_avl_pool;
uu_avl_pool_t *fsps_deleg_perm_avl_pool;
};
static inline const char *
deleg_perm_type(zfs_deleg_note_t note)
{
/* subcommands */
switch (note) {
/* SUBCOMMANDS */
/* OTHER */
case ZFS_DELEG_NOTE_GROUPQUOTA:
case ZFS_DELEG_NOTE_GROUPUSED:
case ZFS_DELEG_NOTE_USERPROP:
case ZFS_DELEG_NOTE_USERQUOTA:
case ZFS_DELEG_NOTE_USERUSED:
case ZFS_DELEG_NOTE_USEROBJQUOTA:
case ZFS_DELEG_NOTE_USEROBJUSED:
case ZFS_DELEG_NOTE_GROUPOBJQUOTA:
case ZFS_DELEG_NOTE_GROUPOBJUSED:
case ZFS_DELEG_NOTE_PROJECTUSED:
case ZFS_DELEG_NOTE_PROJECTQUOTA:
case ZFS_DELEG_NOTE_PROJECTOBJUSED:
case ZFS_DELEG_NOTE_PROJECTOBJQUOTA:
/* other */
return (gettext("other"));
default:
return (gettext("subcommand"));
}
}
static int
who_type2weight(zfs_deleg_who_type_t who_type)
{
int res;
switch (who_type) {
case ZFS_DELEG_NAMED_SET_SETS:
case ZFS_DELEG_NAMED_SET:
res = 0;
break;
case ZFS_DELEG_CREATE_SETS:
case ZFS_DELEG_CREATE:
res = 1;
break;
case ZFS_DELEG_USER_SETS:
case ZFS_DELEG_USER:
res = 2;
break;
case ZFS_DELEG_GROUP_SETS:
case ZFS_DELEG_GROUP:
res = 3;
break;
case ZFS_DELEG_EVERYONE_SETS:
case ZFS_DELEG_EVERYONE:
res = 4;
break;
default:
res = -1;
}
return (res);
}
static int
who_perm_compare(const void *larg, const void *rarg, void *unused)
{
(void) unused;
const who_perm_node_t *l = larg;
const who_perm_node_t *r = rarg;
zfs_deleg_who_type_t ltype = l->who_perm.who_type;
zfs_deleg_who_type_t rtype = r->who_perm.who_type;
int lweight = who_type2weight(ltype);
int rweight = who_type2weight(rtype);
int res = lweight - rweight;
if (res == 0)
res = strncmp(l->who_perm.who_name, r->who_perm.who_name,
ZFS_MAX_DELEG_NAME-1);
if (res == 0)
return (0);
if (res > 0)
return (1);
else
return (-1);
}
static int
deleg_perm_compare(const void *larg, const void *rarg, void *unused)
{
(void) unused;
const deleg_perm_node_t *l = larg;
const deleg_perm_node_t *r = rarg;
int res = strncmp(l->dpn_perm.dp_name, r->dpn_perm.dp_name,
ZFS_MAX_DELEG_NAME-1);
if (res == 0)
return (0);
if (res > 0)
return (1);
else
return (-1);
}
static inline void
fs_perm_set_init(fs_perm_set_t *fspset)
{
memset(fspset, 0, sizeof (fs_perm_set_t));
if ((fspset->fsps_list_pool = uu_list_pool_create("fsps_list_pool",
sizeof (fs_perm_node_t), offsetof(fs_perm_node_t, fspn_list_node),
NULL, UU_DEFAULT)) == NULL)
nomem();
if ((fspset->fsps_list = uu_list_create(fspset->fsps_list_pool, NULL,
UU_DEFAULT)) == NULL)
nomem();
if ((fspset->fsps_named_set_avl_pool = uu_avl_pool_create(
"named_set_avl_pool", sizeof (who_perm_node_t), offsetof(
who_perm_node_t, who_avl_node), who_perm_compare,
UU_DEFAULT)) == NULL)
nomem();
if ((fspset->fsps_who_perm_avl_pool = uu_avl_pool_create(
"who_perm_avl_pool", sizeof (who_perm_node_t), offsetof(
who_perm_node_t, who_avl_node), who_perm_compare,
UU_DEFAULT)) == NULL)
nomem();
if ((fspset->fsps_deleg_perm_avl_pool = uu_avl_pool_create(
"deleg_perm_avl_pool", sizeof (deleg_perm_node_t), offsetof(
deleg_perm_node_t, dpn_avl_node), deleg_perm_compare, UU_DEFAULT))
== NULL)
nomem();
}
static inline void fs_perm_fini(fs_perm_t *);
static inline void who_perm_fini(who_perm_t *);
static inline void
fs_perm_set_fini(fs_perm_set_t *fspset)
{
fs_perm_node_t *node = uu_list_first(fspset->fsps_list);
while (node != NULL) {
fs_perm_node_t *next_node =
uu_list_next(fspset->fsps_list, node);
fs_perm_t *fsperm = &node->fspn_fsperm;
fs_perm_fini(fsperm);
uu_list_remove(fspset->fsps_list, node);
free(node);
node = next_node;
}
uu_avl_pool_destroy(fspset->fsps_named_set_avl_pool);
uu_avl_pool_destroy(fspset->fsps_who_perm_avl_pool);
uu_avl_pool_destroy(fspset->fsps_deleg_perm_avl_pool);
}
static inline void
deleg_perm_init(deleg_perm_t *deleg_perm, zfs_deleg_who_type_t type,
const char *name)
{
deleg_perm->dp_who_type = type;
deleg_perm->dp_name = name;
}
static inline void
who_perm_init(who_perm_t *who_perm, fs_perm_t *fsperm,
zfs_deleg_who_type_t type, const char *name)
{
uu_avl_pool_t *pool;
pool = fsperm->fsp_set->fsps_deleg_perm_avl_pool;
memset(who_perm, 0, sizeof (who_perm_t));
if ((who_perm->who_deleg_perm_avl = uu_avl_create(pool, NULL,
UU_DEFAULT)) == NULL)
nomem();
who_perm->who_type = type;
who_perm->who_name = name;
who_perm->who_fsperm = fsperm;
}
static inline void
who_perm_fini(who_perm_t *who_perm)
{
deleg_perm_node_t *node = uu_avl_first(who_perm->who_deleg_perm_avl);
while (node != NULL) {
deleg_perm_node_t *next_node =
uu_avl_next(who_perm->who_deleg_perm_avl, node);
uu_avl_remove(who_perm->who_deleg_perm_avl, node);
free(node);
node = next_node;
}
uu_avl_destroy(who_perm->who_deleg_perm_avl);
}
static inline void
fs_perm_init(fs_perm_t *fsperm, fs_perm_set_t *fspset, const char *fsname)
{
uu_avl_pool_t *nset_pool = fspset->fsps_named_set_avl_pool;
uu_avl_pool_t *who_pool = fspset->fsps_who_perm_avl_pool;
memset(fsperm, 0, sizeof (fs_perm_t));
if ((fsperm->fsp_sc_avl = uu_avl_create(nset_pool, NULL, UU_DEFAULT))
== NULL)
nomem();
if ((fsperm->fsp_uge_avl = uu_avl_create(who_pool, NULL, UU_DEFAULT))
== NULL)
nomem();
fsperm->fsp_set = fspset;
fsperm->fsp_name = fsname;
}
static inline void
fs_perm_fini(fs_perm_t *fsperm)
{
who_perm_node_t *node = uu_avl_first(fsperm->fsp_sc_avl);
while (node != NULL) {
who_perm_node_t *next_node = uu_avl_next(fsperm->fsp_sc_avl,
node);
who_perm_t *who_perm = &node->who_perm;
who_perm_fini(who_perm);
uu_avl_remove(fsperm->fsp_sc_avl, node);
free(node);
node = next_node;
}
node = uu_avl_first(fsperm->fsp_uge_avl);
while (node != NULL) {
who_perm_node_t *next_node = uu_avl_next(fsperm->fsp_uge_avl,
node);
who_perm_t *who_perm = &node->who_perm;
who_perm_fini(who_perm);
uu_avl_remove(fsperm->fsp_uge_avl, node);
free(node);
node = next_node;
}
uu_avl_destroy(fsperm->fsp_sc_avl);
uu_avl_destroy(fsperm->fsp_uge_avl);
}
static void
set_deleg_perm_node(uu_avl_t *avl, deleg_perm_node_t *node,
zfs_deleg_who_type_t who_type, const char *name, char locality)
{
uu_avl_index_t idx = 0;
deleg_perm_node_t *found_node = NULL;
deleg_perm_t *deleg_perm = &node->dpn_perm;
deleg_perm_init(deleg_perm, who_type, name);
if ((found_node = uu_avl_find(avl, node, NULL, &idx))
== NULL)
uu_avl_insert(avl, node, idx);
else {
node = found_node;
deleg_perm = &node->dpn_perm;
}
switch (locality) {
case ZFS_DELEG_LOCAL:
deleg_perm->dp_local = B_TRUE;
break;
case ZFS_DELEG_DESCENDENT:
deleg_perm->dp_descend = B_TRUE;
break;
case ZFS_DELEG_NA:
break;
default:
assert(B_FALSE); /* invalid locality */
}
}
static inline int
parse_who_perm(who_perm_t *who_perm, nvlist_t *nvl, char locality)
{
nvpair_t *nvp = NULL;
fs_perm_set_t *fspset = who_perm->who_fsperm->fsp_set;
uu_avl_t *avl = who_perm->who_deleg_perm_avl;
zfs_deleg_who_type_t who_type = who_perm->who_type;
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
const char *name = nvpair_name(nvp);
data_type_t type = nvpair_type(nvp);
uu_avl_pool_t *avl_pool = fspset->fsps_deleg_perm_avl_pool;
deleg_perm_node_t *node =
safe_malloc(sizeof (deleg_perm_node_t));
VERIFY(type == DATA_TYPE_BOOLEAN);
uu_avl_node_init(node, &node->dpn_avl_node, avl_pool);
set_deleg_perm_node(avl, node, who_type, name, locality);
}
return (0);
}
static inline int
parse_fs_perm(fs_perm_t *fsperm, nvlist_t *nvl)
{
nvpair_t *nvp = NULL;
fs_perm_set_t *fspset = fsperm->fsp_set;
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
nvlist_t *nvl2 = NULL;
const char *name = nvpair_name(nvp);
uu_avl_t *avl = NULL;
uu_avl_pool_t *avl_pool = NULL;
zfs_deleg_who_type_t perm_type = name[0];
char perm_locality = name[1];
const char *perm_name = name + 3;
who_perm_t *who_perm = NULL;
assert('$' == name[2]);
if (nvpair_value_nvlist(nvp, &nvl2) != 0)
return (-1);
switch (perm_type) {
case ZFS_DELEG_CREATE:
case ZFS_DELEG_CREATE_SETS:
case ZFS_DELEG_NAMED_SET:
case ZFS_DELEG_NAMED_SET_SETS:
avl_pool = fspset->fsps_named_set_avl_pool;
avl = fsperm->fsp_sc_avl;
break;
case ZFS_DELEG_USER:
case ZFS_DELEG_USER_SETS:
case ZFS_DELEG_GROUP:
case ZFS_DELEG_GROUP_SETS:
case ZFS_DELEG_EVERYONE:
case ZFS_DELEG_EVERYONE_SETS:
avl_pool = fspset->fsps_who_perm_avl_pool;
avl = fsperm->fsp_uge_avl;
break;
default:
assert(!"unhandled zfs_deleg_who_type_t");
}
who_perm_node_t *found_node = NULL;
who_perm_node_t *node = safe_malloc(
sizeof (who_perm_node_t));
who_perm = &node->who_perm;
uu_avl_index_t idx = 0;
uu_avl_node_init(node, &node->who_avl_node, avl_pool);
who_perm_init(who_perm, fsperm, perm_type, perm_name);
if ((found_node = uu_avl_find(avl, node, NULL, &idx))
== NULL) {
if (avl == fsperm->fsp_uge_avl) {
uid_t rid = 0;
struct passwd *p = NULL;
struct group *g = NULL;
const char *nice_name = NULL;
switch (perm_type) {
case ZFS_DELEG_USER_SETS:
case ZFS_DELEG_USER:
rid = atoi(perm_name);
p = getpwuid(rid);
if (p)
nice_name = p->pw_name;
break;
case ZFS_DELEG_GROUP_SETS:
case ZFS_DELEG_GROUP:
rid = atoi(perm_name);
g = getgrgid(rid);
if (g)
nice_name = g->gr_name;
break;
default:
break;
}
if (nice_name != NULL) {
(void) strlcpy(
node->who_perm.who_ug_name,
nice_name, 256);
} else {
/* User or group unknown */
(void) snprintf(
node->who_perm.who_ug_name,
sizeof (node->who_perm.who_ug_name),
"(unknown: %d)", rid);
}
}
uu_avl_insert(avl, node, idx);
} else {
node = found_node;
who_perm = &node->who_perm;
}
assert(who_perm != NULL);
(void) parse_who_perm(who_perm, nvl2, perm_locality);
}
return (0);
}
static inline int
parse_fs_perm_set(fs_perm_set_t *fspset, nvlist_t *nvl)
{
nvpair_t *nvp = NULL;
uu_avl_index_t idx = 0;
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
nvlist_t *nvl2 = NULL;
const char *fsname = nvpair_name(nvp);
data_type_t type = nvpair_type(nvp);
fs_perm_t *fsperm = NULL;
fs_perm_node_t *node = safe_malloc(sizeof (fs_perm_node_t));
fsperm = &node->fspn_fsperm;
VERIFY(DATA_TYPE_NVLIST == type);
uu_list_node_init(node, &node->fspn_list_node,
fspset->fsps_list_pool);
idx = uu_list_numnodes(fspset->fsps_list);
fs_perm_init(fsperm, fspset, fsname);
if (nvpair_value_nvlist(nvp, &nvl2) != 0)
return (-1);
(void) parse_fs_perm(fsperm, nvl2);
uu_list_insert(fspset->fsps_list, node, idx);
}
return (0);
}
static inline const char *
deleg_perm_comment(zfs_deleg_note_t note)
{
const char *str = "";
/* subcommands */
switch (note) {
/* SUBCOMMANDS */
case ZFS_DELEG_NOTE_ALLOW:
str = gettext("Must also have the permission that is being"
"\n\t\t\t\tallowed");
break;
case ZFS_DELEG_NOTE_CLONE:
str = gettext("Must also have the 'create' ability and 'mount'"
"\n\t\t\t\tability in the origin file system");
break;
case ZFS_DELEG_NOTE_CREATE:
str = gettext("Must also have the 'mount' ability");
break;
case ZFS_DELEG_NOTE_DESTROY:
str = gettext("Must also have the 'mount' ability");
break;
case ZFS_DELEG_NOTE_DIFF:
str = gettext("Allows lookup of paths within a dataset;"
"\n\t\t\t\tgiven an object number. Ordinary users need this"
"\n\t\t\t\tin order to use zfs diff");
break;
case ZFS_DELEG_NOTE_HOLD:
str = gettext("Allows adding a user hold to a snapshot");
break;
case ZFS_DELEG_NOTE_MOUNT:
str = gettext("Allows mount/umount of ZFS datasets");
break;
case ZFS_DELEG_NOTE_PROMOTE:
str = gettext("Must also have the 'mount'\n\t\t\t\tand"
" 'promote' ability in the origin file system");
break;
case ZFS_DELEG_NOTE_RECEIVE:
str = gettext("Must also have the 'mount' and 'create'"
" ability");
break;
case ZFS_DELEG_NOTE_RELEASE:
str = gettext("Allows releasing a user hold which\n\t\t\t\t"
"might destroy the snapshot");
break;
case ZFS_DELEG_NOTE_RENAME:
str = gettext("Must also have the 'mount' and 'create'"
"\n\t\t\t\tability in the new parent");
break;
case ZFS_DELEG_NOTE_ROLLBACK:
str = gettext("");
break;
case ZFS_DELEG_NOTE_SEND:
str = gettext("");
break;
case ZFS_DELEG_NOTE_SHARE:
str = gettext("Allows sharing file systems over NFS or SMB"
"\n\t\t\t\tprotocols");
break;
case ZFS_DELEG_NOTE_SNAPSHOT:
str = gettext("");
break;
case ZFS_DELEG_NOTE_LOAD_KEY:
str = gettext("Allows loading or unloading an encryption key");
break;
case ZFS_DELEG_NOTE_CHANGE_KEY:
str = gettext("Allows changing or adding an encryption key");
break;
/*
* case ZFS_DELEG_NOTE_VSCAN:
* str = gettext("");
* break;
*/
/* OTHER */
case ZFS_DELEG_NOTE_GROUPQUOTA:
str = gettext("Allows accessing any groupquota@... property");
break;
case ZFS_DELEG_NOTE_GROUPUSED:
str = gettext("Allows reading any groupused@... property");
break;
case ZFS_DELEG_NOTE_USERPROP:
str = gettext("Allows changing any user property");
break;
case ZFS_DELEG_NOTE_USERQUOTA:
str = gettext("Allows accessing any userquota@... property");
break;
case ZFS_DELEG_NOTE_USERUSED:
str = gettext("Allows reading any userused@... property");
break;
case ZFS_DELEG_NOTE_USEROBJQUOTA:
str = gettext("Allows accessing any userobjquota@... property");
break;
case ZFS_DELEG_NOTE_GROUPOBJQUOTA:
str = gettext("Allows accessing any \n\t\t\t\t"
"groupobjquota@... property");
break;
case ZFS_DELEG_NOTE_GROUPOBJUSED:
str = gettext("Allows reading any groupobjused@... property");
break;
case ZFS_DELEG_NOTE_USEROBJUSED:
str = gettext("Allows reading any userobjused@... property");
break;
case ZFS_DELEG_NOTE_PROJECTQUOTA:
str = gettext("Allows accessing any projectquota@... property");
break;
case ZFS_DELEG_NOTE_PROJECTOBJQUOTA:
str = gettext("Allows accessing any \n\t\t\t\t"
"projectobjquota@... property");
break;
case ZFS_DELEG_NOTE_PROJECTUSED:
str = gettext("Allows reading any projectused@... property");
break;
case ZFS_DELEG_NOTE_PROJECTOBJUSED:
str = gettext("Allows accessing any \n\t\t\t\t"
"projectobjused@... property");
break;
/* other */
default:
str = "";
}
return (str);
}
struct allow_opts {
boolean_t local;
boolean_t descend;
boolean_t user;
boolean_t group;
boolean_t everyone;
boolean_t create;
boolean_t set;
boolean_t recursive; /* unallow only */
boolean_t prt_usage;
boolean_t prt_perms;
char *who;
char *perms;
const char *dataset;
};
static inline int
prop_cmp(const void *a, const void *b)
{
const char *str1 = *(const char **)a;
const char *str2 = *(const char **)b;
return (strcmp(str1, str2));
}
static void
allow_usage(boolean_t un, boolean_t requested, const char *msg)
{
const char *opt_desc[] = {
"-h", gettext("show this help message and exit"),
"-l", gettext("set permission locally"),
"-d", gettext("set permission for descents"),
"-u", gettext("set permission for user"),
"-g", gettext("set permission for group"),
"-e", gettext("set permission for everyone"),
"-c", gettext("set create time permission"),
"-s", gettext("define permission set"),
/* unallow only */
"-r", gettext("remove permissions recursively"),
};
size_t unallow_size = sizeof (opt_desc) / sizeof (char *);
size_t allow_size = unallow_size - 2;
const char *props[ZFS_NUM_PROPS];
int i;
size_t count = 0;
FILE *fp = requested ? stdout : stderr;
zprop_desc_t *pdtbl = zfs_prop_get_table();
const char *fmt = gettext("%-16s %-14s\t%s\n");
(void) fprintf(fp, gettext("Usage: %s\n"), get_usage(un ? HELP_UNALLOW :
HELP_ALLOW));
(void) fprintf(fp, gettext("Options:\n"));
for (i = 0; i < (un ? unallow_size : allow_size); i += 2) {
const char *opt = opt_desc[i];
const char *optdsc = opt_desc[i + 1];
(void) fprintf(fp, gettext(" %-10s %s\n"), opt, optdsc);
}
(void) fprintf(fp, gettext("\nThe following permissions are "
"supported:\n\n"));
(void) fprintf(fp, fmt, gettext("NAME"), gettext("TYPE"),
gettext("NOTES"));
for (i = 0; i < ZFS_NUM_DELEG_NOTES; i++) {
const char *perm_name = zfs_deleg_perm_tbl[i].z_perm;
zfs_deleg_note_t perm_note = zfs_deleg_perm_tbl[i].z_note;
const char *perm_type = deleg_perm_type(perm_note);
const char *perm_comment = deleg_perm_comment(perm_note);
(void) fprintf(fp, fmt, perm_name, perm_type, perm_comment);
}
for (i = 0; i < ZFS_NUM_PROPS; i++) {
zprop_desc_t *pd = &pdtbl[i];
if (pd->pd_visible != B_TRUE)
continue;
if (pd->pd_attr == PROP_READONLY)
continue;
props[count++] = pd->pd_name;
}
props[count] = NULL;
qsort(props, count, sizeof (char *), prop_cmp);
for (i = 0; i < count; i++)
(void) fprintf(fp, fmt, props[i], gettext("property"), "");
if (msg != NULL)
(void) fprintf(fp, gettext("\nzfs: error: %s"), msg);
exit(requested ? 0 : 2);
}
static inline const char *
munge_args(int argc, char **argv, boolean_t un, size_t expected_argc,
char **permsp)
{
if (un && argc == expected_argc - 1)
*permsp = NULL;
else if (argc == expected_argc)
*permsp = argv[argc - 2];
else
allow_usage(un, B_FALSE,
gettext("wrong number of parameters\n"));
return (argv[argc - 1]);
}
static void
parse_allow_args(int argc, char **argv, boolean_t un, struct allow_opts *opts)
{
int uge_sum = opts->user + opts->group + opts->everyone;
int csuge_sum = opts->create + opts->set + uge_sum;
int ldcsuge_sum = csuge_sum + opts->local + opts->descend;
int all_sum = un ? ldcsuge_sum + opts->recursive : ldcsuge_sum;
if (uge_sum > 1)
allow_usage(un, B_FALSE,
gettext("-u, -g, and -e are mutually exclusive\n"));
if (opts->prt_usage) {
if (argc == 0 && all_sum == 0)
allow_usage(un, B_TRUE, NULL);
else
usage(B_FALSE);
}
if (opts->set) {
if (csuge_sum > 1)
allow_usage(un, B_FALSE,
gettext("invalid options combined with -s\n"));
opts->dataset = munge_args(argc, argv, un, 3, &opts->perms);
if (argv[0][0] != '@')
allow_usage(un, B_FALSE,
gettext("invalid set name: missing '@' prefix\n"));
opts->who = argv[0];
} else if (opts->create) {
if (ldcsuge_sum > 1)
allow_usage(un, B_FALSE,
gettext("invalid options combined with -c\n"));
opts->dataset = munge_args(argc, argv, un, 2, &opts->perms);
} else if (opts->everyone) {
if (csuge_sum > 1)
allow_usage(un, B_FALSE,
gettext("invalid options combined with -e\n"));
opts->dataset = munge_args(argc, argv, un, 2, &opts->perms);
} else if (uge_sum == 0 && argc > 0 && strcmp(argv[0], "everyone")
== 0) {
opts->everyone = B_TRUE;
argc--;
argv++;
opts->dataset = munge_args(argc, argv, un, 2, &opts->perms);
} else if (argc == 1 && !un) {
opts->prt_perms = B_TRUE;
opts->dataset = argv[argc-1];
} else {
opts->dataset = munge_args(argc, argv, un, 3, &opts->perms);
opts->who = argv[0];
}
if (!opts->local && !opts->descend) {
opts->local = B_TRUE;
opts->descend = B_TRUE;
}
}
static void
store_allow_perm(zfs_deleg_who_type_t type, boolean_t local, boolean_t descend,
const char *who, char *perms, nvlist_t *top_nvl)
{
int i;
char ld[2] = { '\0', '\0' };
char who_buf[MAXNAMELEN + 32];
char base_type = '\0';
char set_type = '\0';
nvlist_t *base_nvl = NULL;
nvlist_t *set_nvl = NULL;
nvlist_t *nvl;
if (nvlist_alloc(&base_nvl, NV_UNIQUE_NAME, 0) != 0)
nomem();
if (nvlist_alloc(&set_nvl, NV_UNIQUE_NAME, 0) != 0)
nomem();
switch (type) {
case ZFS_DELEG_NAMED_SET_SETS:
case ZFS_DELEG_NAMED_SET:
set_type = ZFS_DELEG_NAMED_SET_SETS;
base_type = ZFS_DELEG_NAMED_SET;
ld[0] = ZFS_DELEG_NA;
break;
case ZFS_DELEG_CREATE_SETS:
case ZFS_DELEG_CREATE:
set_type = ZFS_DELEG_CREATE_SETS;
base_type = ZFS_DELEG_CREATE;
ld[0] = ZFS_DELEG_NA;
break;
case ZFS_DELEG_USER_SETS:
case ZFS_DELEG_USER:
set_type = ZFS_DELEG_USER_SETS;
base_type = ZFS_DELEG_USER;
if (local)
ld[0] = ZFS_DELEG_LOCAL;
if (descend)
ld[1] = ZFS_DELEG_DESCENDENT;
break;
case ZFS_DELEG_GROUP_SETS:
case ZFS_DELEG_GROUP:
set_type = ZFS_DELEG_GROUP_SETS;
base_type = ZFS_DELEG_GROUP;
if (local)
ld[0] = ZFS_DELEG_LOCAL;
if (descend)
ld[1] = ZFS_DELEG_DESCENDENT;
break;
case ZFS_DELEG_EVERYONE_SETS:
case ZFS_DELEG_EVERYONE:
set_type = ZFS_DELEG_EVERYONE_SETS;
base_type = ZFS_DELEG_EVERYONE;
if (local)
ld[0] = ZFS_DELEG_LOCAL;
if (descend)
ld[1] = ZFS_DELEG_DESCENDENT;
break;
default:
assert(set_type != '\0' && base_type != '\0');
}
if (perms != NULL) {
char *curr = perms;
char *end = curr + strlen(perms);
while (curr < end) {
char *delim = strchr(curr, ',');
if (delim == NULL)
delim = end;
else
*delim = '\0';
if (curr[0] == '@')
nvl = set_nvl;
else
nvl = base_nvl;
(void) nvlist_add_boolean(nvl, curr);
if (delim != end)
*delim = ',';
curr = delim + 1;
}
for (i = 0; i < 2; i++) {
char locality = ld[i];
if (locality == 0)
continue;
if (!nvlist_empty(base_nvl)) {
if (who != NULL)
(void) snprintf(who_buf,
sizeof (who_buf), "%c%c$%s",
base_type, locality, who);
else
(void) snprintf(who_buf,
sizeof (who_buf), "%c%c$",
base_type, locality);
(void) nvlist_add_nvlist(top_nvl, who_buf,
base_nvl);
}
if (!nvlist_empty(set_nvl)) {
if (who != NULL)
(void) snprintf(who_buf,
sizeof (who_buf), "%c%c$%s",
set_type, locality, who);
else
(void) snprintf(who_buf,
sizeof (who_buf), "%c%c$",
set_type, locality);
(void) nvlist_add_nvlist(top_nvl, who_buf,
set_nvl);
}
}
} else {
for (i = 0; i < 2; i++) {
char locality = ld[i];
if (locality == 0)
continue;
if (who != NULL)
(void) snprintf(who_buf, sizeof (who_buf),
"%c%c$%s", base_type, locality, who);
else
(void) snprintf(who_buf, sizeof (who_buf),
"%c%c$", base_type, locality);
(void) nvlist_add_boolean(top_nvl, who_buf);
if (who != NULL)
(void) snprintf(who_buf, sizeof (who_buf),
"%c%c$%s", set_type, locality, who);
else
(void) snprintf(who_buf, sizeof (who_buf),
"%c%c$", set_type, locality);
(void) nvlist_add_boolean(top_nvl, who_buf);
}
}
}
static int
construct_fsacl_list(boolean_t un, struct allow_opts *opts, nvlist_t **nvlp)
{
if (nvlist_alloc(nvlp, NV_UNIQUE_NAME, 0) != 0)
nomem();
if (opts->set) {
store_allow_perm(ZFS_DELEG_NAMED_SET, opts->local,
opts->descend, opts->who, opts->perms, *nvlp);
} else if (opts->create) {
store_allow_perm(ZFS_DELEG_CREATE, opts->local,
opts->descend, NULL, opts->perms, *nvlp);
} else if (opts->everyone) {
store_allow_perm(ZFS_DELEG_EVERYONE, opts->local,
opts->descend, NULL, opts->perms, *nvlp);
} else {
char *curr = opts->who;
char *end = curr + strlen(curr);
while (curr < end) {
const char *who;
zfs_deleg_who_type_t who_type = ZFS_DELEG_WHO_UNKNOWN;
char *endch;
char *delim = strchr(curr, ',');
char errbuf[256];
char id[64];
struct passwd *p = NULL;
struct group *g = NULL;
uid_t rid;
if (delim == NULL)
delim = end;
else
*delim = '\0';
rid = (uid_t)strtol(curr, &endch, 0);
if (opts->user) {
who_type = ZFS_DELEG_USER;
if (*endch != '\0')
p = getpwnam(curr);
else
p = getpwuid(rid);
if (p != NULL)
rid = p->pw_uid;
else if (*endch != '\0') {
(void) snprintf(errbuf, 256, gettext(
"invalid user %s\n"), curr);
allow_usage(un, B_TRUE, errbuf);
}
} else if (opts->group) {
who_type = ZFS_DELEG_GROUP;
if (*endch != '\0')
g = getgrnam(curr);
else
g = getgrgid(rid);
if (g != NULL)
rid = g->gr_gid;
else if (*endch != '\0') {
(void) snprintf(errbuf, 256, gettext(
"invalid group %s\n"), curr);
allow_usage(un, B_TRUE, errbuf);
}
} else {
if (*endch != '\0') {
p = getpwnam(curr);
} else {
p = getpwuid(rid);
}
if (p == NULL) {
if (*endch != '\0') {
g = getgrnam(curr);
} else {
g = getgrgid(rid);
}
}
if (p != NULL) {
who_type = ZFS_DELEG_USER;
rid = p->pw_uid;
} else if (g != NULL) {
who_type = ZFS_DELEG_GROUP;
rid = g->gr_gid;
} else {
(void) snprintf(errbuf, 256, gettext(
"invalid user/group %s\n"), curr);
allow_usage(un, B_TRUE, errbuf);
}
}
(void) sprintf(id, "%u", rid);
who = id;
store_allow_perm(who_type, opts->local,
opts->descend, who, opts->perms, *nvlp);
curr = delim + 1;
}
}
return (0);
}
static void
print_set_creat_perms(uu_avl_t *who_avl)
{
const char *sc_title[] = {
gettext("Permission sets:\n"),
gettext("Create time permissions:\n"),
NULL
};
who_perm_node_t *who_node = NULL;
int prev_weight = -1;
for (who_node = uu_avl_first(who_avl); who_node != NULL;
who_node = uu_avl_next(who_avl, who_node)) {
uu_avl_t *avl = who_node->who_perm.who_deleg_perm_avl;
zfs_deleg_who_type_t who_type = who_node->who_perm.who_type;
const char *who_name = who_node->who_perm.who_name;
int weight = who_type2weight(who_type);
boolean_t first = B_TRUE;
deleg_perm_node_t *deleg_node;
if (prev_weight != weight) {
(void) printf("%s", sc_title[weight]);
prev_weight = weight;
}
if (who_name == NULL || strnlen(who_name, 1) == 0)
(void) printf("\t");
else
(void) printf("\t%s ", who_name);
for (deleg_node = uu_avl_first(avl); deleg_node != NULL;
deleg_node = uu_avl_next(avl, deleg_node)) {
if (first) {
(void) printf("%s",
deleg_node->dpn_perm.dp_name);
first = B_FALSE;
} else
(void) printf(",%s",
deleg_node->dpn_perm.dp_name);
}
(void) printf("\n");
}
}
static void
print_uge_deleg_perms(uu_avl_t *who_avl, boolean_t local, boolean_t descend,
const char *title)
{
who_perm_node_t *who_node = NULL;
boolean_t prt_title = B_TRUE;
uu_avl_walk_t *walk;
if ((walk = uu_avl_walk_start(who_avl, UU_WALK_ROBUST)) == NULL)
nomem();
while ((who_node = uu_avl_walk_next(walk)) != NULL) {
const char *who_name = who_node->who_perm.who_name;
const char *nice_who_name = who_node->who_perm.who_ug_name;
uu_avl_t *avl = who_node->who_perm.who_deleg_perm_avl;
zfs_deleg_who_type_t who_type = who_node->who_perm.who_type;
char delim = ' ';
deleg_perm_node_t *deleg_node;
boolean_t prt_who = B_TRUE;
for (deleg_node = uu_avl_first(avl);
deleg_node != NULL;
deleg_node = uu_avl_next(avl, deleg_node)) {
if (local != deleg_node->dpn_perm.dp_local ||
descend != deleg_node->dpn_perm.dp_descend)
continue;
if (prt_who) {
const char *who = NULL;
if (prt_title) {
prt_title = B_FALSE;
(void) printf("%s", title);
}
switch (who_type) {
case ZFS_DELEG_USER_SETS:
case ZFS_DELEG_USER:
who = gettext("user");
if (nice_who_name)
who_name = nice_who_name;
break;
case ZFS_DELEG_GROUP_SETS:
case ZFS_DELEG_GROUP:
who = gettext("group");
if (nice_who_name)
who_name = nice_who_name;
break;
case ZFS_DELEG_EVERYONE_SETS:
case ZFS_DELEG_EVERYONE:
who = gettext("everyone");
who_name = NULL;
break;
default:
assert(who != NULL);
}
prt_who = B_FALSE;
if (who_name == NULL)
(void) printf("\t%s", who);
else
(void) printf("\t%s %s", who, who_name);
}
(void) printf("%c%s", delim,
deleg_node->dpn_perm.dp_name);
delim = ',';
}
if (!prt_who)
(void) printf("\n");
}
uu_avl_walk_end(walk);
}
static void
print_fs_perms(fs_perm_set_t *fspset)
{
fs_perm_node_t *node = NULL;
char buf[MAXNAMELEN + 32];
const char *dsname = buf;
for (node = uu_list_first(fspset->fsps_list); node != NULL;
node = uu_list_next(fspset->fsps_list, node)) {
uu_avl_t *sc_avl = node->fspn_fsperm.fsp_sc_avl;
uu_avl_t *uge_avl = node->fspn_fsperm.fsp_uge_avl;
int left = 0;
(void) snprintf(buf, sizeof (buf),
gettext("---- Permissions on %s "),
node->fspn_fsperm.fsp_name);
(void) printf("%s", dsname);
left = 70 - strlen(buf);
while (left-- > 0)
(void) printf("-");
(void) printf("\n");
print_set_creat_perms(sc_avl);
print_uge_deleg_perms(uge_avl, B_TRUE, B_FALSE,
gettext("Local permissions:\n"));
print_uge_deleg_perms(uge_avl, B_FALSE, B_TRUE,
gettext("Descendent permissions:\n"));
print_uge_deleg_perms(uge_avl, B_TRUE, B_TRUE,
gettext("Local+Descendent permissions:\n"));
}
}
static fs_perm_set_t fs_perm_set = { NULL, NULL, NULL, NULL };
struct deleg_perms {
boolean_t un;
nvlist_t *nvl;
};
static int
set_deleg_perms(zfs_handle_t *zhp, void *data)
{
struct deleg_perms *perms = (struct deleg_perms *)data;
zfs_type_t zfs_type = zfs_get_type(zhp);
if (zfs_type != ZFS_TYPE_FILESYSTEM && zfs_type != ZFS_TYPE_VOLUME)
return (0);
return (zfs_set_fsacl(zhp, perms->un, perms->nvl));
}
static int
zfs_do_allow_unallow_impl(int argc, char **argv, boolean_t un)
{
zfs_handle_t *zhp;
nvlist_t *perm_nvl = NULL;
nvlist_t *update_perm_nvl = NULL;
int error = 1;
int c;
struct allow_opts opts = { 0 };
const char *optstr = un ? "ldugecsrh" : "ldugecsh";
/* check opts */
while ((c = getopt(argc, argv, optstr)) != -1) {
switch (c) {
case 'l':
opts.local = B_TRUE;
break;
case 'd':
opts.descend = B_TRUE;
break;
case 'u':
opts.user = B_TRUE;
break;
case 'g':
opts.group = B_TRUE;
break;
case 'e':
opts.everyone = B_TRUE;
break;
case 's':
opts.set = B_TRUE;
break;
case 'c':
opts.create = B_TRUE;
break;
case 'r':
opts.recursive = B_TRUE;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case 'h':
opts.prt_usage = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check arguments */
parse_allow_args(argc, argv, un, &opts);
/* try to open the dataset */
if ((zhp = zfs_open(g_zfs, opts.dataset, ZFS_TYPE_FILESYSTEM |
ZFS_TYPE_VOLUME)) == NULL) {
(void) fprintf(stderr, "Failed to open dataset: %s\n",
opts.dataset);
return (-1);
}
if (zfs_get_fsacl(zhp, &perm_nvl) != 0)
goto cleanup2;
fs_perm_set_init(&fs_perm_set);
if (parse_fs_perm_set(&fs_perm_set, perm_nvl) != 0) {
(void) fprintf(stderr, "Failed to parse fsacl permissions\n");
goto cleanup1;
}
if (opts.prt_perms)
print_fs_perms(&fs_perm_set);
else {
(void) construct_fsacl_list(un, &opts, &update_perm_nvl);
if (zfs_set_fsacl(zhp, un, update_perm_nvl) != 0)
goto cleanup0;
if (un && opts.recursive) {
struct deleg_perms data = { un, update_perm_nvl };
- if (zfs_iter_filesystems(zhp, 0, set_deleg_perms,
+ if (zfs_iter_filesystems_v2(zhp, 0, set_deleg_perms,
&data) != 0)
goto cleanup0;
}
}
error = 0;
cleanup0:
nvlist_free(perm_nvl);
nvlist_free(update_perm_nvl);
cleanup1:
fs_perm_set_fini(&fs_perm_set);
cleanup2:
zfs_close(zhp);
return (error);
}
static int
zfs_do_allow(int argc, char **argv)
{
return (zfs_do_allow_unallow_impl(argc, argv, B_FALSE));
}
static int
zfs_do_unallow(int argc, char **argv)
{
return (zfs_do_allow_unallow_impl(argc, argv, B_TRUE));
}
static int
zfs_do_hold_rele_impl(int argc, char **argv, boolean_t holding)
{
int errors = 0;
int i;
const char *tag;
boolean_t recursive = B_FALSE;
const char *opts = holding ? "rt" : "r";
int c;
/* check options */
while ((c = getopt(argc, argv, opts)) != -1) {
switch (c) {
case 'r':
recursive = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc < 2)
usage(B_FALSE);
tag = argv[0];
--argc;
++argv;
if (holding && tag[0] == '.') {
/* tags starting with '.' are reserved for libzfs */
(void) fprintf(stderr, gettext("tag may not start with '.'\n"));
usage(B_FALSE);
}
for (i = 0; i < argc; ++i) {
zfs_handle_t *zhp;
char parent[ZFS_MAX_DATASET_NAME_LEN];
const char *delim;
char *path = argv[i];
delim = strchr(path, '@');
if (delim == NULL) {
(void) fprintf(stderr,
gettext("'%s' is not a snapshot\n"), path);
++errors;
continue;
}
(void) strlcpy(parent, path, MIN(sizeof (parent),
delim - path + 1));
zhp = zfs_open(g_zfs, parent,
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL) {
++errors;
continue;
}
if (holding) {
if (zfs_hold(zhp, delim+1, tag, recursive, -1) != 0)
++errors;
} else {
if (zfs_release(zhp, delim+1, tag, recursive) != 0)
++errors;
}
zfs_close(zhp);
}
return (errors != 0);
}
/*
* zfs hold [-r] [-t] <tag> <snap> ...
*
* -r Recursively hold
*
* Apply a user-hold with the given tag to the list of snapshots.
*/
static int
zfs_do_hold(int argc, char **argv)
{
return (zfs_do_hold_rele_impl(argc, argv, B_TRUE));
}
/*
* zfs release [-r] <tag> <snap> ...
*
* -r Recursively release
*
* Release a user-hold with the given tag from the list of snapshots.
*/
static int
zfs_do_release(int argc, char **argv)
{
return (zfs_do_hold_rele_impl(argc, argv, B_FALSE));
}
typedef struct holds_cbdata {
boolean_t cb_recursive;
const char *cb_snapname;
nvlist_t **cb_nvlp;
size_t cb_max_namelen;
size_t cb_max_taglen;
} holds_cbdata_t;
#define STRFTIME_FMT_STR "%a %b %e %H:%M %Y"
#define DATETIME_BUF_LEN (32)
/*
*
*/
static void
print_holds(boolean_t scripted, int nwidth, int tagwidth, nvlist_t *nvl,
boolean_t parsable)
{
int i;
nvpair_t *nvp = NULL;
const char *const hdr_cols[] = { "NAME", "TAG", "TIMESTAMP" };
const char *col;
if (!scripted) {
for (i = 0; i < 3; i++) {
col = gettext(hdr_cols[i]);
if (i < 2)
(void) printf("%-*s ", i ? tagwidth : nwidth,
col);
else
(void) printf("%s\n", col);
}
}
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
const char *zname = nvpair_name(nvp);
nvlist_t *nvl2;
nvpair_t *nvp2 = NULL;
(void) nvpair_value_nvlist(nvp, &nvl2);
while ((nvp2 = nvlist_next_nvpair(nvl2, nvp2)) != NULL) {
char tsbuf[DATETIME_BUF_LEN];
const char *tagname = nvpair_name(nvp2);
uint64_t val = 0;
time_t time;
struct tm t;
(void) nvpair_value_uint64(nvp2, &val);
time = (time_t)val;
(void) localtime_r(&time, &t);
(void) strftime(tsbuf, DATETIME_BUF_LEN,
gettext(STRFTIME_FMT_STR), &t);
if (scripted) {
if (parsable) {
(void) printf("%s\t%s\t%ld\n", zname,
tagname, (unsigned long)time);
} else {
(void) printf("%s\t%s\t%s\n", zname,
tagname, tsbuf);
}
} else {
if (parsable) {
(void) printf("%-*s %-*s %ld\n",
nwidth, zname, tagwidth,
tagname, (unsigned long)time);
} else {
(void) printf("%-*s %-*s %s\n",
nwidth, zname, tagwidth,
tagname, tsbuf);
}
}
}
}
}
/*
* Generic callback function to list a dataset or snapshot.
*/
static int
holds_callback(zfs_handle_t *zhp, void *data)
{
holds_cbdata_t *cbp = data;
nvlist_t *top_nvl = *cbp->cb_nvlp;
nvlist_t *nvl = NULL;
nvpair_t *nvp = NULL;
const char *zname = zfs_get_name(zhp);
size_t znamelen = strlen(zname);
if (cbp->cb_recursive) {
const char *snapname;
char *delim = strchr(zname, '@');
if (delim == NULL)
return (0);
snapname = delim + 1;
if (strcmp(cbp->cb_snapname, snapname))
return (0);
}
if (zfs_get_holds(zhp, &nvl) != 0)
return (-1);
if (znamelen > cbp->cb_max_namelen)
cbp->cb_max_namelen = znamelen;
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
const char *tag = nvpair_name(nvp);
size_t taglen = strlen(tag);
if (taglen > cbp->cb_max_taglen)
cbp->cb_max_taglen = taglen;
}
return (nvlist_add_nvlist(top_nvl, zname, nvl));
}
/*
* zfs holds [-rHp] <snap> ...
*
* -r Lists holds that are set on the named snapshots recursively.
* -H Scripted mode; elide headers and separate columns by tabs.
* -p Display values in parsable (literal) format.
*/
static int
zfs_do_holds(int argc, char **argv)
{
int c;
boolean_t errors = B_FALSE;
boolean_t scripted = B_FALSE;
boolean_t recursive = B_FALSE;
boolean_t parsable = B_FALSE;
int types = ZFS_TYPE_SNAPSHOT;
holds_cbdata_t cb = { 0 };
int limit = 0;
int ret = 0;
int flags = 0;
/* check options */
while ((c = getopt(argc, argv, "rHp")) != -1) {
switch (c) {
case 'r':
recursive = B_TRUE;
break;
case 'H':
scripted = B_TRUE;
break;
case 'p':
parsable = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
if (recursive) {
types |= ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME;
flags |= ZFS_ITER_RECURSE;
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc < 1)
usage(B_FALSE);
nvlist_t *nvl = fnvlist_alloc();
for (int i = 0; i < argc; ++i) {
char *snapshot = argv[i];
const char *delim;
const char *snapname;
delim = strchr(snapshot, '@');
if (delim == NULL) {
(void) fprintf(stderr,
gettext("'%s' is not a snapshot\n"), snapshot);
errors = B_TRUE;
continue;
}
snapname = delim + 1;
if (recursive)
snapshot[delim - snapshot] = '\0';
cb.cb_recursive = recursive;
cb.cb_snapname = snapname;
cb.cb_nvlp = &nvl;
/*
* 1. collect holds data, set format options
*/
ret = zfs_for_each(1, argv + i, flags, types, NULL, NULL, limit,
holds_callback, &cb);
if (ret != 0)
errors = B_TRUE;
}
/*
* 2. print holds data
*/
print_holds(scripted, cb.cb_max_namelen, cb.cb_max_taglen, nvl,
parsable);
if (nvlist_empty(nvl))
(void) fprintf(stderr, gettext("no datasets available\n"));
nvlist_free(nvl);
return (errors);
}
#define CHECK_SPINNER 30
#define SPINNER_TIME 3 /* seconds */
#define MOUNT_TIME 1 /* seconds */
typedef struct get_all_state {
boolean_t ga_verbose;
get_all_cb_t *ga_cbp;
} get_all_state_t;
static int
get_one_dataset(zfs_handle_t *zhp, void *data)
{
static const char *const spin[] = { "-", "\\", "|", "/" };
static int spinval = 0;
static int spincheck = 0;
static time_t last_spin_time = (time_t)0;
get_all_state_t *state = data;
zfs_type_t type = zfs_get_type(zhp);
if (state->ga_verbose) {
if (--spincheck < 0) {
time_t now = time(NULL);
if (last_spin_time + SPINNER_TIME < now) {
update_progress(spin[spinval++ % 4]);
last_spin_time = now;
}
spincheck = CHECK_SPINNER;
}
}
/*
* Iterate over any nested datasets.
*/
- if (zfs_iter_filesystems(zhp, 0, get_one_dataset, data) != 0) {
+ if (zfs_iter_filesystems_v2(zhp, 0, get_one_dataset, data) != 0) {
zfs_close(zhp);
return (1);
}
/*
* Skip any datasets whose type does not match.
*/
if ((type & ZFS_TYPE_FILESYSTEM) == 0) {
zfs_close(zhp);
return (0);
}
libzfs_add_handle(state->ga_cbp, zhp);
assert(state->ga_cbp->cb_used <= state->ga_cbp->cb_alloc);
return (0);
}
static void
get_all_datasets(get_all_cb_t *cbp, boolean_t verbose)
{
get_all_state_t state = {
.ga_verbose = verbose,
.ga_cbp = cbp
};
if (verbose)
set_progress_header(gettext("Reading ZFS config"));
(void) zfs_iter_root(g_zfs, get_one_dataset, &state);
if (verbose)
finish_progress(gettext("done."));
}
/*
* Generic callback for sharing or mounting filesystems. Because the code is so
* similar, we have a common function with an extra parameter to determine which
* mode we are using.
*/
typedef enum { OP_SHARE, OP_MOUNT } share_mount_op_t;
typedef struct share_mount_state {
share_mount_op_t sm_op;
boolean_t sm_verbose;
int sm_flags;
char *sm_options;
enum sa_protocol sm_proto; /* only valid for OP_SHARE */
pthread_mutex_t sm_lock; /* protects the remaining fields */
uint_t sm_total; /* number of filesystems to process */
uint_t sm_done; /* number of filesystems processed */
int sm_status; /* -1 if any of the share/mount operations failed */
} share_mount_state_t;
/*
* Share or mount a dataset.
*/
static int
share_mount_one(zfs_handle_t *zhp, int op, int flags, enum sa_protocol protocol,
boolean_t explicit, const char *options)
{
char mountpoint[ZFS_MAXPROPLEN];
char shareopts[ZFS_MAXPROPLEN];
char smbshareopts[ZFS_MAXPROPLEN];
const char *cmdname = op == OP_SHARE ? "share" : "mount";
struct mnttab mnt;
uint64_t zoned, canmount;
boolean_t shared_nfs, shared_smb;
assert(zfs_get_type(zhp) & ZFS_TYPE_FILESYSTEM);
/*
* Check to make sure we can mount/share this dataset. If we
* are in the global zone and the filesystem is exported to a
* local zone, or if we are in a local zone and the
* filesystem is not exported, then it is an error.
*/
zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
if (zoned && getzoneid() == GLOBAL_ZONEID) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot %s '%s': "
"dataset is exported to a local zone\n"), cmdname,
zfs_get_name(zhp));
return (1);
} else if (!zoned && getzoneid() != GLOBAL_ZONEID) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot %s '%s': "
"permission denied\n"), cmdname,
zfs_get_name(zhp));
return (1);
}
/*
* Ignore any filesystems which don't apply to us. This
* includes those with a legacy mountpoint, or those with
* legacy share options.
*/
verify(zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, mountpoint,
sizeof (mountpoint), NULL, NULL, 0, B_FALSE) == 0);
verify(zfs_prop_get(zhp, ZFS_PROP_SHARENFS, shareopts,
sizeof (shareopts), NULL, NULL, 0, B_FALSE) == 0);
verify(zfs_prop_get(zhp, ZFS_PROP_SHARESMB, smbshareopts,
sizeof (smbshareopts), NULL, NULL, 0, B_FALSE) == 0);
if (op == OP_SHARE && strcmp(shareopts, "off") == 0 &&
strcmp(smbshareopts, "off") == 0) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot share '%s': "
"legacy share\n"), zfs_get_name(zhp));
(void) fprintf(stderr, gettext("use exports(5) or "
"smb.conf(5) to share this filesystem, or set "
"the sharenfs or sharesmb property\n"));
return (1);
}
/*
* We cannot share or mount legacy filesystems. If the
* shareopts is non-legacy but the mountpoint is legacy, we
* treat it as a legacy share.
*/
if (strcmp(mountpoint, "legacy") == 0) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot %s '%s': "
"legacy mountpoint\n"), cmdname, zfs_get_name(zhp));
(void) fprintf(stderr, gettext("use %s(8) to "
"%s this filesystem\n"), cmdname, cmdname);
return (1);
}
if (strcmp(mountpoint, "none") == 0) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot %s '%s': no "
"mountpoint set\n"), cmdname, zfs_get_name(zhp));
return (1);
}
/*
* canmount explicit outcome
* on no pass through
* on yes pass through
* off no return 0
* off yes display error, return 1
* noauto no return 0
* noauto yes pass through
*/
canmount = zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT);
if (canmount == ZFS_CANMOUNT_OFF) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot %s '%s': "
"'canmount' property is set to 'off'\n"), cmdname,
zfs_get_name(zhp));
return (1);
} else if (canmount == ZFS_CANMOUNT_NOAUTO && !explicit) {
/*
* When performing a 'zfs mount -a', we skip any mounts for
* datasets that have 'noauto' set. Sharing a dataset with
* 'noauto' set is only allowed if it's mounted.
*/
if (op == OP_MOUNT)
return (0);
if (op == OP_SHARE && !zfs_is_mounted(zhp, NULL)) {
/* also purge it from existing exports */
zfs_unshare(zhp, mountpoint, NULL);
return (0);
}
}
/*
* If this filesystem is encrypted and does not have
* a loaded key, we can not mount it.
*/
if ((flags & MS_CRYPT) == 0 &&
zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION) != ZIO_CRYPT_OFF &&
zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS) ==
ZFS_KEYSTATUS_UNAVAILABLE) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot %s '%s': "
"encryption key not loaded\n"), cmdname, zfs_get_name(zhp));
return (1);
}
/*
* If this filesystem is inconsistent and has a receive resume
* token, we can not mount it.
*/
if (zfs_prop_get_int(zhp, ZFS_PROP_INCONSISTENT) &&
zfs_prop_get(zhp, ZFS_PROP_RECEIVE_RESUME_TOKEN,
NULL, 0, NULL, NULL, 0, B_TRUE) == 0) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot %s '%s': "
"Contains partially-completed state from "
"\"zfs receive -s\", which can be resumed with "
"\"zfs send -t\"\n"),
cmdname, zfs_get_name(zhp));
return (1);
}
if (zfs_prop_get_int(zhp, ZFS_PROP_REDACTED) && !(flags & MS_FORCE)) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot %s '%s': "
"Dataset is not complete, was created by receiving "
"a redacted zfs send stream.\n"), cmdname,
zfs_get_name(zhp));
return (1);
}
/*
* At this point, we have verified that the mountpoint and/or
* shareopts are appropriate for auto management. If the
* filesystem is already mounted or shared, return (failing
* for explicit requests); otherwise mount or share the
* filesystem.
*/
switch (op) {
case OP_SHARE: {
enum sa_protocol prot[] = {SA_PROTOCOL_NFS, SA_NO_PROTOCOL};
shared_nfs = zfs_is_shared(zhp, NULL, prot);
*prot = SA_PROTOCOL_SMB;
shared_smb = zfs_is_shared(zhp, NULL, prot);
if ((shared_nfs && shared_smb) ||
(shared_nfs && strcmp(shareopts, "on") == 0 &&
strcmp(smbshareopts, "off") == 0) ||
(shared_smb && strcmp(smbshareopts, "on") == 0 &&
strcmp(shareopts, "off") == 0)) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot share "
"'%s': filesystem already shared\n"),
zfs_get_name(zhp));
return (1);
}
if (!zfs_is_mounted(zhp, NULL) &&
zfs_mount(zhp, NULL, flags) != 0)
return (1);
*prot = protocol;
if (zfs_share(zhp, protocol == SA_NO_PROTOCOL ? NULL : prot))
return (1);
}
break;
case OP_MOUNT:
mnt.mnt_mntopts = (char *)(options ?: "");
if (!hasmntopt(&mnt, MNTOPT_REMOUNT) &&
zfs_is_mounted(zhp, NULL)) {
if (!explicit)
return (0);
(void) fprintf(stderr, gettext("cannot mount "
"'%s': filesystem already mounted\n"),
zfs_get_name(zhp));
return (1);
}
if (zfs_mount(zhp, options, flags) != 0)
return (1);
break;
}
return (0);
}
/*
* Reports progress in the form "(current/total)". Not thread-safe.
*/
static void
report_mount_progress(int current, int total)
{
static time_t last_progress_time = 0;
time_t now = time(NULL);
char info[32];
/* display header if we're here for the first time */
if (current == 1) {
set_progress_header(gettext("Mounting ZFS filesystems"));
} else if (current != total && last_progress_time + MOUNT_TIME >= now) {
/* too soon to report again */
return;
}
last_progress_time = now;
(void) sprintf(info, "(%d/%d)", current, total);
if (current == total)
finish_progress(info);
else
update_progress(info);
}
/*
* zfs_foreach_mountpoint() callback that mounts or shares one filesystem and
* updates the progress meter.
*/
static int
share_mount_one_cb(zfs_handle_t *zhp, void *arg)
{
share_mount_state_t *sms = arg;
int ret;
ret = share_mount_one(zhp, sms->sm_op, sms->sm_flags, sms->sm_proto,
B_FALSE, sms->sm_options);
pthread_mutex_lock(&sms->sm_lock);
if (ret != 0)
sms->sm_status = ret;
sms->sm_done++;
if (sms->sm_verbose)
report_mount_progress(sms->sm_done, sms->sm_total);
pthread_mutex_unlock(&sms->sm_lock);
return (ret);
}
static void
append_options(char *mntopts, char *newopts)
{
int len = strlen(mntopts);
/* original length plus new string to append plus 1 for the comma */
if (len + 1 + strlen(newopts) >= MNT_LINE_MAX) {
(void) fprintf(stderr, gettext("the opts argument for "
"'%s' option is too long (more than %d chars)\n"),
"-o", MNT_LINE_MAX);
usage(B_FALSE);
}
if (*mntopts)
mntopts[len++] = ',';
(void) strcpy(&mntopts[len], newopts);
}
static enum sa_protocol
sa_protocol_decode(const char *protocol)
{
for (enum sa_protocol i = 0; i < ARRAY_SIZE(sa_protocol_names); ++i)
if (strcmp(protocol, sa_protocol_names[i]) == 0)
return (i);
(void) fputs(gettext("share type must be one of: "), stderr);
for (enum sa_protocol i = 0;
i < ARRAY_SIZE(sa_protocol_names); ++i)
(void) fprintf(stderr, "%s%s",
i != 0 ? ", " : "", sa_protocol_names[i]);
(void) fputc('\n', stderr);
usage(B_FALSE);
}
static int
share_mount(int op, int argc, char **argv)
{
int do_all = 0;
boolean_t verbose = B_FALSE;
int c, ret = 0;
char *options = NULL;
int flags = 0;
/* check options */
while ((c = getopt(argc, argv, op == OP_MOUNT ? ":alvo:Of" : "al"))
!= -1) {
switch (c) {
case 'a':
do_all = 1;
break;
case 'v':
verbose = B_TRUE;
break;
case 'l':
flags |= MS_CRYPT;
break;
case 'o':
if (*optarg == '\0') {
(void) fprintf(stderr, gettext("empty mount "
"options (-o) specified\n"));
usage(B_FALSE);
}
if (options == NULL)
options = safe_malloc(MNT_LINE_MAX + 1);
/* option validation is done later */
append_options(options, optarg);
break;
case 'O':
flags |= MS_OVERLAY;
break;
case 'f':
flags |= MS_FORCE;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (do_all) {
enum sa_protocol protocol = SA_NO_PROTOCOL;
if (op == OP_SHARE && argc > 0) {
protocol = sa_protocol_decode(argv[0]);
argc--;
argv++;
}
if (argc != 0) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
start_progress_timer();
get_all_cb_t cb = { 0 };
get_all_datasets(&cb, verbose);
if (cb.cb_used == 0) {
free(options);
return (0);
}
share_mount_state_t share_mount_state = { 0 };
share_mount_state.sm_op = op;
share_mount_state.sm_verbose = verbose;
share_mount_state.sm_flags = flags;
share_mount_state.sm_options = options;
share_mount_state.sm_proto = protocol;
share_mount_state.sm_total = cb.cb_used;
pthread_mutex_init(&share_mount_state.sm_lock, NULL);
/* For a 'zfs share -a' operation start with a clean slate. */
zfs_truncate_shares(NULL);
/*
* libshare isn't mt-safe, so only do the operation in parallel
* if we're mounting. Additionally, the key-loading option must
* be serialized so that we can prompt the user for their keys
* in a consistent manner.
*/
zfs_foreach_mountpoint(g_zfs, cb.cb_handles, cb.cb_used,
share_mount_one_cb, &share_mount_state,
op == OP_MOUNT && !(flags & MS_CRYPT));
zfs_commit_shares(NULL);
ret = share_mount_state.sm_status;
for (int i = 0; i < cb.cb_used; i++)
zfs_close(cb.cb_handles[i]);
free(cb.cb_handles);
} else if (argc == 0) {
FILE *mnttab;
struct mnttab entry;
if ((op == OP_SHARE) || (options != NULL)) {
(void) fprintf(stderr, gettext("missing filesystem "
"argument (specify -a for all)\n"));
usage(B_FALSE);
}
/*
* When mount is given no arguments, go through
* /proc/self/mounts and display any active ZFS mounts.
* We hide any snapshots, since they are controlled
* automatically.
*/
if ((mnttab = fopen(MNTTAB, "re")) == NULL) {
free(options);
return (ENOENT);
}
while (getmntent(mnttab, &entry) == 0) {
if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0 ||
strchr(entry.mnt_special, '@') != NULL)
continue;
(void) printf("%-30s %s\n", entry.mnt_special,
entry.mnt_mountp);
}
(void) fclose(mnttab);
} else {
zfs_handle_t *zhp;
if (argc > 1) {
(void) fprintf(stderr,
gettext("too many arguments\n"));
usage(B_FALSE);
}
if ((zhp = zfs_open(g_zfs, argv[0],
ZFS_TYPE_FILESYSTEM)) == NULL) {
ret = 1;
} else {
ret = share_mount_one(zhp, op, flags, SA_NO_PROTOCOL,
B_TRUE, options);
zfs_commit_shares(NULL);
zfs_close(zhp);
}
}
free(options);
return (ret);
}
/*
* zfs mount -a
* zfs mount filesystem
*
* Mount all filesystems, or mount the given filesystem.
*/
static int
zfs_do_mount(int argc, char **argv)
{
return (share_mount(OP_MOUNT, argc, argv));
}
/*
* zfs share -a [nfs | smb]
* zfs share filesystem
*
* Share all filesystems, or share the given filesystem.
*/
static int
zfs_do_share(int argc, char **argv)
{
return (share_mount(OP_SHARE, argc, argv));
}
typedef struct unshare_unmount_node {
zfs_handle_t *un_zhp;
char *un_mountp;
uu_avl_node_t un_avlnode;
} unshare_unmount_node_t;
static int
unshare_unmount_compare(const void *larg, const void *rarg, void *unused)
{
(void) unused;
const unshare_unmount_node_t *l = larg;
const unshare_unmount_node_t *r = rarg;
return (strcmp(l->un_mountp, r->un_mountp));
}
/*
* Convenience routine used by zfs_do_umount() and manual_unmount(). Given an
* absolute path, find the entry /proc/self/mounts, verify that it's a
* ZFS filesystem, and unmount it appropriately.
*/
static int
unshare_unmount_path(int op, char *path, int flags, boolean_t is_manual)
{
zfs_handle_t *zhp;
int ret = 0;
struct stat64 statbuf;
struct extmnttab entry;
const char *cmdname = (op == OP_SHARE) ? "unshare" : "unmount";
ino_t path_inode;
/*
* Search for the given (major,minor) pair in the mount table.
*/
if (getextmntent(path, &entry, &statbuf) != 0) {
if (op == OP_SHARE) {
(void) fprintf(stderr, gettext("cannot %s '%s': not "
"currently mounted\n"), cmdname, path);
return (1);
}
(void) fprintf(stderr, gettext("warning: %s not in"
"/proc/self/mounts\n"), path);
if ((ret = umount2(path, flags)) != 0)
(void) fprintf(stderr, gettext("%s: %s\n"), path,
strerror(errno));
return (ret != 0);
}
path_inode = statbuf.st_ino;
if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0) {
(void) fprintf(stderr, gettext("cannot %s '%s': not a ZFS "
"filesystem\n"), cmdname, path);
return (1);
}
if ((zhp = zfs_open(g_zfs, entry.mnt_special,
ZFS_TYPE_FILESYSTEM)) == NULL)
return (1);
ret = 1;
if (stat64(entry.mnt_mountp, &statbuf) != 0) {
(void) fprintf(stderr, gettext("cannot %s '%s': %s\n"),
cmdname, path, strerror(errno));
goto out;
} else if (statbuf.st_ino != path_inode) {
(void) fprintf(stderr, gettext("cannot "
"%s '%s': not a mountpoint\n"), cmdname, path);
goto out;
}
if (op == OP_SHARE) {
char nfs_mnt_prop[ZFS_MAXPROPLEN];
char smbshare_prop[ZFS_MAXPROPLEN];
verify(zfs_prop_get(zhp, ZFS_PROP_SHARENFS, nfs_mnt_prop,
sizeof (nfs_mnt_prop), NULL, NULL, 0, B_FALSE) == 0);
verify(zfs_prop_get(zhp, ZFS_PROP_SHARESMB, smbshare_prop,
sizeof (smbshare_prop), NULL, NULL, 0, B_FALSE) == 0);
if (strcmp(nfs_mnt_prop, "off") == 0 &&
strcmp(smbshare_prop, "off") == 0) {
(void) fprintf(stderr, gettext("cannot unshare "
"'%s': legacy share\n"), path);
(void) fprintf(stderr, gettext("use exportfs(8) "
"or smbcontrol(1) to unshare this filesystem\n"));
} else if (!zfs_is_shared(zhp, NULL, NULL)) {
(void) fprintf(stderr, gettext("cannot unshare '%s': "
"not currently shared\n"), path);
} else {
ret = zfs_unshare(zhp, path, NULL);
zfs_commit_shares(NULL);
}
} else {
char mtpt_prop[ZFS_MAXPROPLEN];
verify(zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, mtpt_prop,
sizeof (mtpt_prop), NULL, NULL, 0, B_FALSE) == 0);
if (is_manual) {
ret = zfs_unmount(zhp, NULL, flags);
} else if (strcmp(mtpt_prop, "legacy") == 0) {
(void) fprintf(stderr, gettext("cannot unmount "
"'%s': legacy mountpoint\n"),
zfs_get_name(zhp));
(void) fprintf(stderr, gettext("use umount(8) "
"to unmount this filesystem\n"));
} else {
ret = zfs_unmountall(zhp, flags);
}
}
out:
zfs_close(zhp);
return (ret != 0);
}
/*
* Generic callback for unsharing or unmounting a filesystem.
*/
static int
unshare_unmount(int op, int argc, char **argv)
{
int do_all = 0;
int flags = 0;
int ret = 0;
int c;
zfs_handle_t *zhp;
char nfs_mnt_prop[ZFS_MAXPROPLEN];
char sharesmb[ZFS_MAXPROPLEN];
/* check options */
while ((c = getopt(argc, argv, op == OP_SHARE ? ":a" : "afu")) != -1) {
switch (c) {
case 'a':
do_all = 1;
break;
case 'f':
flags |= MS_FORCE;
break;
case 'u':
flags |= MS_CRYPT;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (do_all) {
/*
* We could make use of zfs_for_each() to walk all datasets in
* the system, but this would be very inefficient, especially
* since we would have to linearly search /proc/self/mounts for
* each one. Instead, do one pass through /proc/self/mounts
* looking for zfs entries and call zfs_unmount() for each one.
*
* Things get a little tricky if the administrator has created
* mountpoints beneath other ZFS filesystems. In this case, we
* have to unmount the deepest filesystems first. To accomplish
* this, we place all the mountpoints in an AVL tree sorted by
* the special type (dataset name), and walk the result in
* reverse to make sure to get any snapshots first.
*/
FILE *mnttab;
struct mnttab entry;
uu_avl_pool_t *pool;
uu_avl_t *tree = NULL;
unshare_unmount_node_t *node;
uu_avl_index_t idx;
uu_avl_walk_t *walk;
enum sa_protocol *protocol = NULL,
single_protocol[] = {SA_NO_PROTOCOL, SA_NO_PROTOCOL};
if (op == OP_SHARE && argc > 0) {
*single_protocol = sa_protocol_decode(argv[0]);
protocol = single_protocol;
argc--;
argv++;
}
if (argc != 0) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if (((pool = uu_avl_pool_create("unmount_pool",
sizeof (unshare_unmount_node_t),
offsetof(unshare_unmount_node_t, un_avlnode),
unshare_unmount_compare, UU_DEFAULT)) == NULL) ||
((tree = uu_avl_create(pool, NULL, UU_DEFAULT)) == NULL))
nomem();
if ((mnttab = fopen(MNTTAB, "re")) == NULL) {
uu_avl_destroy(tree);
uu_avl_pool_destroy(pool);
return (ENOENT);
}
while (getmntent(mnttab, &entry) == 0) {
/* ignore non-ZFS entries */
if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0)
continue;
/* ignore snapshots */
if (strchr(entry.mnt_special, '@') != NULL)
continue;
if ((zhp = zfs_open(g_zfs, entry.mnt_special,
ZFS_TYPE_FILESYSTEM)) == NULL) {
ret = 1;
continue;
}
/*
* Ignore datasets that are excluded/restricted by
* parent pool name.
*/
if (zpool_skip_pool(zfs_get_pool_name(zhp))) {
zfs_close(zhp);
continue;
}
switch (op) {
case OP_SHARE:
verify(zfs_prop_get(zhp, ZFS_PROP_SHARENFS,
nfs_mnt_prop,
sizeof (nfs_mnt_prop),
NULL, NULL, 0, B_FALSE) == 0);
if (strcmp(nfs_mnt_prop, "off") != 0)
break;
verify(zfs_prop_get(zhp, ZFS_PROP_SHARESMB,
nfs_mnt_prop,
sizeof (nfs_mnt_prop),
NULL, NULL, 0, B_FALSE) == 0);
if (strcmp(nfs_mnt_prop, "off") == 0)
continue;
break;
case OP_MOUNT:
/* Ignore legacy mounts */
verify(zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT,
nfs_mnt_prop,
sizeof (nfs_mnt_prop),
NULL, NULL, 0, B_FALSE) == 0);
if (strcmp(nfs_mnt_prop, "legacy") == 0)
continue;
/* Ignore canmount=noauto mounts */
if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) ==
ZFS_CANMOUNT_NOAUTO)
continue;
break;
default:
break;
}
node = safe_malloc(sizeof (unshare_unmount_node_t));
node->un_zhp = zhp;
node->un_mountp = safe_strdup(entry.mnt_mountp);
uu_avl_node_init(node, &node->un_avlnode, pool);
if (uu_avl_find(tree, node, NULL, &idx) == NULL) {
uu_avl_insert(tree, node, idx);
} else {
zfs_close(node->un_zhp);
free(node->un_mountp);
free(node);
}
}
(void) fclose(mnttab);
/*
* Walk the AVL tree in reverse, unmounting each filesystem and
* removing it from the AVL tree in the process.
*/
if ((walk = uu_avl_walk_start(tree,
UU_WALK_REVERSE | UU_WALK_ROBUST)) == NULL)
nomem();
while ((node = uu_avl_walk_next(walk)) != NULL) {
const char *mntarg = NULL;
uu_avl_remove(tree, node);
switch (op) {
case OP_SHARE:
if (zfs_unshare(node->un_zhp,
node->un_mountp, protocol) != 0)
ret = 1;
break;
case OP_MOUNT:
if (zfs_unmount(node->un_zhp,
mntarg, flags) != 0)
ret = 1;
break;
}
zfs_close(node->un_zhp);
free(node->un_mountp);
free(node);
}
if (op == OP_SHARE)
zfs_commit_shares(protocol);
uu_avl_walk_end(walk);
uu_avl_destroy(tree);
uu_avl_pool_destroy(pool);
} else {
if (argc != 1) {
if (argc == 0)
(void) fprintf(stderr,
gettext("missing filesystem argument\n"));
else
(void) fprintf(stderr,
gettext("too many arguments\n"));
usage(B_FALSE);
}
/*
* We have an argument, but it may be a full path or a ZFS
* filesystem. Pass full paths off to unmount_path() (shared by
* manual_unmount), otherwise open the filesystem and pass to
* zfs_unmount().
*/
if (argv[0][0] == '/')
return (unshare_unmount_path(op, argv[0],
flags, B_FALSE));
if ((zhp = zfs_open(g_zfs, argv[0],
ZFS_TYPE_FILESYSTEM)) == NULL)
return (1);
verify(zfs_prop_get(zhp, op == OP_SHARE ?
ZFS_PROP_SHARENFS : ZFS_PROP_MOUNTPOINT,
nfs_mnt_prop, sizeof (nfs_mnt_prop), NULL,
NULL, 0, B_FALSE) == 0);
switch (op) {
case OP_SHARE:
verify(zfs_prop_get(zhp, ZFS_PROP_SHARENFS,
nfs_mnt_prop,
sizeof (nfs_mnt_prop),
NULL, NULL, 0, B_FALSE) == 0);
verify(zfs_prop_get(zhp, ZFS_PROP_SHARESMB,
sharesmb, sizeof (sharesmb), NULL, NULL,
0, B_FALSE) == 0);
if (strcmp(nfs_mnt_prop, "off") == 0 &&
strcmp(sharesmb, "off") == 0) {
(void) fprintf(stderr, gettext("cannot "
"unshare '%s': legacy share\n"),
zfs_get_name(zhp));
(void) fprintf(stderr, gettext("use "
"exports(5) or smb.conf(5) to unshare "
"this filesystem\n"));
ret = 1;
} else if (!zfs_is_shared(zhp, NULL, NULL)) {
(void) fprintf(stderr, gettext("cannot "
"unshare '%s': not currently "
"shared\n"), zfs_get_name(zhp));
ret = 1;
} else if (zfs_unshareall(zhp, NULL) != 0) {
ret = 1;
}
break;
case OP_MOUNT:
if (strcmp(nfs_mnt_prop, "legacy") == 0) {
(void) fprintf(stderr, gettext("cannot "
"unmount '%s': legacy "
"mountpoint\n"), zfs_get_name(zhp));
(void) fprintf(stderr, gettext("use "
"umount(8) to unmount this "
"filesystem\n"));
ret = 1;
} else if (!zfs_is_mounted(zhp, NULL)) {
(void) fprintf(stderr, gettext("cannot "
"unmount '%s': not currently "
"mounted\n"),
zfs_get_name(zhp));
ret = 1;
} else if (zfs_unmountall(zhp, flags) != 0) {
ret = 1;
}
break;
}
zfs_close(zhp);
}
return (ret);
}
/*
* zfs unmount [-fu] -a
* zfs unmount [-fu] filesystem
*
* Unmount all filesystems, or a specific ZFS filesystem.
*/
static int
zfs_do_unmount(int argc, char **argv)
{
return (unshare_unmount(OP_MOUNT, argc, argv));
}
/*
* zfs unshare -a
* zfs unshare filesystem
*
* Unshare all filesystems, or a specific ZFS filesystem.
*/
static int
zfs_do_unshare(int argc, char **argv)
{
return (unshare_unmount(OP_SHARE, argc, argv));
}
static int
find_command_idx(const char *command, int *idx)
{
int i;
for (i = 0; i < NCOMMAND; i++) {
if (command_table[i].name == NULL)
continue;
if (strcmp(command, command_table[i].name) == 0) {
*idx = i;
return (0);
}
}
return (1);
}
static int
zfs_do_diff(int argc, char **argv)
{
zfs_handle_t *zhp;
int flags = 0;
char *tosnap = NULL;
char *fromsnap = NULL;
char *atp, *copy;
int err = 0;
int c;
struct sigaction sa;
while ((c = getopt(argc, argv, "FHth")) != -1) {
switch (c) {
case 'F':
flags |= ZFS_DIFF_CLASSIFY;
break;
case 'H':
flags |= ZFS_DIFF_PARSEABLE;
break;
case 't':
flags |= ZFS_DIFF_TIMESTAMP;
break;
case 'h':
flags |= ZFS_DIFF_NO_MANGLE;
break;
default:
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr,
gettext("must provide at least one snapshot name\n"));
usage(B_FALSE);
}
if (argc > 2) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
fromsnap = argv[0];
tosnap = (argc == 2) ? argv[1] : NULL;
copy = NULL;
if (*fromsnap != '@')
copy = strdup(fromsnap);
else if (tosnap)
copy = strdup(tosnap);
if (copy == NULL)
usage(B_FALSE);
if ((atp = strchr(copy, '@')) != NULL)
*atp = '\0';
if ((zhp = zfs_open(g_zfs, copy, ZFS_TYPE_FILESYSTEM)) == NULL) {
free(copy);
return (1);
}
free(copy);
/*
* Ignore SIGPIPE so that the library can give us
* information on any failure
*/
if (sigemptyset(&sa.sa_mask) == -1) {
err = errno;
goto out;
}
sa.sa_flags = 0;
sa.sa_handler = SIG_IGN;
if (sigaction(SIGPIPE, &sa, NULL) == -1) {
err = errno;
goto out;
}
err = zfs_show_diffs(zhp, STDOUT_FILENO, fromsnap, tosnap, flags);
out:
zfs_close(zhp);
return (err != 0);
}
/*
* zfs bookmark <fs@source>|<fs#source> <fs#bookmark>
*
* Creates a bookmark with the given name from the source snapshot
* or creates a copy of an existing source bookmark.
*/
static int
zfs_do_bookmark(int argc, char **argv)
{
char *source, *bookname;
char expbuf[ZFS_MAX_DATASET_NAME_LEN];
int source_type;
nvlist_t *nvl;
int ret = 0;
int c;
/* check options */
while ((c = getopt(argc, argv, "")) != -1) {
switch (c) {
case '?':
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
goto usage;
}
}
argc -= optind;
argv += optind;
/* check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing source argument\n"));
goto usage;
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing bookmark argument\n"));
goto usage;
}
source = argv[0];
bookname = argv[1];
if (strchr(source, '@') == NULL && strchr(source, '#') == NULL) {
(void) fprintf(stderr,
gettext("invalid source name '%s': "
"must contain a '@' or '#'\n"), source);
goto usage;
}
if (strchr(bookname, '#') == NULL) {
(void) fprintf(stderr,
gettext("invalid bookmark name '%s': "
"must contain a '#'\n"), bookname);
goto usage;
}
/*
* expand source or bookname to full path:
* one of them may be specified as short name
*/
{
char **expand;
char *source_short, *bookname_short;
source_short = strpbrk(source, "@#");
bookname_short = strpbrk(bookname, "#");
if (source_short == source &&
bookname_short == bookname) {
(void) fprintf(stderr, gettext(
"either source or bookmark must be specified as "
"full dataset paths"));
goto usage;
} else if (source_short != source &&
bookname_short != bookname) {
expand = NULL;
} else if (source_short != source) {
strlcpy(expbuf, source, sizeof (expbuf));
expand = &bookname;
} else if (bookname_short != bookname) {
strlcpy(expbuf, bookname, sizeof (expbuf));
expand = &source;
} else {
abort();
}
if (expand != NULL) {
*strpbrk(expbuf, "@#") = '\0'; /* dataset name in buf */
(void) strlcat(expbuf, *expand, sizeof (expbuf));
*expand = expbuf;
}
}
/* determine source type */
switch (*strpbrk(source, "@#")) {
case '@': source_type = ZFS_TYPE_SNAPSHOT; break;
case '#': source_type = ZFS_TYPE_BOOKMARK; break;
default: abort();
}
/* test the source exists */
zfs_handle_t *zhp;
zhp = zfs_open(g_zfs, source, source_type);
if (zhp == NULL)
goto usage;
zfs_close(zhp);
nvl = fnvlist_alloc();
fnvlist_add_string(nvl, bookname, source);
ret = lzc_bookmark(nvl, NULL);
fnvlist_free(nvl);
if (ret != 0) {
const char *err_msg = NULL;
char errbuf[1024];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot create bookmark '%s'"), bookname);
switch (ret) {
case EXDEV:
err_msg = "bookmark is in a different pool";
break;
case ZFS_ERR_BOOKMARK_SOURCE_NOT_ANCESTOR:
err_msg = "source is not an ancestor of the "
"new bookmark's dataset";
break;
case EEXIST:
err_msg = "bookmark exists";
break;
case EINVAL:
err_msg = "invalid argument";
break;
case ENOTSUP:
err_msg = "bookmark feature not enabled";
break;
case ENOSPC:
err_msg = "out of space";
break;
case ENOENT:
err_msg = "dataset does not exist";
break;
default:
(void) zfs_standard_error(g_zfs, ret, errbuf);
break;
}
if (err_msg != NULL) {
(void) fprintf(stderr, "%s: %s\n", errbuf,
dgettext(TEXT_DOMAIN, err_msg));
}
}
return (ret != 0);
usage:
usage(B_FALSE);
return (-1);
}
static int
zfs_do_channel_program(int argc, char **argv)
{
int ret, fd, c;
size_t progsize, progread;
nvlist_t *outnvl = NULL;
uint64_t instrlimit = ZCP_DEFAULT_INSTRLIMIT;
uint64_t memlimit = ZCP_DEFAULT_MEMLIMIT;
boolean_t sync_flag = B_TRUE, json_output = B_FALSE;
zpool_handle_t *zhp;
/* check options */
while ((c = getopt(argc, argv, "nt:m:j")) != -1) {
switch (c) {
case 't':
case 'm': {
uint64_t arg;
char *endp;
errno = 0;
arg = strtoull(optarg, &endp, 0);
if (errno != 0 || *endp != '\0') {
(void) fprintf(stderr, gettext(
"invalid argument "
"'%s': expected integer\n"), optarg);
goto usage;
}
if (c == 't') {
instrlimit = arg;
} else {
ASSERT3U(c, ==, 'm');
memlimit = arg;
}
break;
}
case 'n': {
sync_flag = B_FALSE;
break;
}
case 'j': {
json_output = B_TRUE;
break;
}
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
goto usage;
}
}
argc -= optind;
argv += optind;
if (argc < 2) {
(void) fprintf(stderr,
gettext("invalid number of arguments\n"));
goto usage;
}
const char *poolname = argv[0];
const char *filename = argv[1];
if (strcmp(filename, "-") == 0) {
fd = 0;
filename = "standard input";
} else if ((fd = open(filename, O_RDONLY)) < 0) {
(void) fprintf(stderr, gettext("cannot open '%s': %s\n"),
filename, strerror(errno));
return (1);
}
if ((zhp = zpool_open(g_zfs, poolname)) == NULL) {
(void) fprintf(stderr, gettext("cannot open pool '%s'\n"),
poolname);
if (fd != 0)
(void) close(fd);
return (1);
}
zpool_close(zhp);
/*
* Read in the channel program, expanding the program buffer as
* necessary.
*/
progread = 0;
progsize = 1024;
char *progbuf = safe_malloc(progsize);
do {
ret = read(fd, progbuf + progread, progsize - progread);
progread += ret;
if (progread == progsize && ret > 0) {
progsize *= 2;
progbuf = safe_realloc(progbuf, progsize);
}
} while (ret > 0);
if (fd != 0)
(void) close(fd);
if (ret < 0) {
free(progbuf);
(void) fprintf(stderr,
gettext("cannot read '%s': %s\n"),
filename, strerror(errno));
return (1);
}
progbuf[progread] = '\0';
/*
* Any remaining arguments are passed as arguments to the lua script as
* a string array:
* {
* "argv" -> [ "arg 1", ... "arg n" ],
* }
*/
nvlist_t *argnvl = fnvlist_alloc();
fnvlist_add_string_array(argnvl, ZCP_ARG_CLIARGV,
(const char **)argv + 2, argc - 2);
if (sync_flag) {
ret = lzc_channel_program(poolname, progbuf,
instrlimit, memlimit, argnvl, &outnvl);
} else {
ret = lzc_channel_program_nosync(poolname, progbuf,
instrlimit, memlimit, argnvl, &outnvl);
}
if (ret != 0) {
/*
* On error, report the error message handed back by lua if one
* exists. Otherwise, generate an appropriate error message,
* falling back on strerror() for an unexpected return code.
*/
const char *errstring = NULL;
const char *msg = gettext("Channel program execution failed");
uint64_t instructions = 0;
if (outnvl != NULL && nvlist_exists(outnvl, ZCP_RET_ERROR)) {
const char *es = NULL;
(void) nvlist_lookup_string(outnvl,
ZCP_RET_ERROR, &es);
if (es == NULL)
errstring = strerror(ret);
else
errstring = es;
if (ret == ETIME) {
(void) nvlist_lookup_uint64(outnvl,
ZCP_ARG_INSTRLIMIT, &instructions);
}
} else {
switch (ret) {
case EINVAL:
errstring =
"Invalid instruction or memory limit.";
break;
case ENOMEM:
errstring = "Return value too large.";
break;
case ENOSPC:
errstring = "Memory limit exhausted.";
break;
case ETIME:
errstring = "Timed out.";
break;
case EPERM:
errstring = "Permission denied. Channel "
"programs must be run as root.";
break;
default:
(void) zfs_standard_error(g_zfs, ret, msg);
}
}
if (errstring != NULL)
(void) fprintf(stderr, "%s:\n%s\n", msg, errstring);
if (ret == ETIME && instructions != 0)
(void) fprintf(stderr,
gettext("%llu Lua instructions\n"),
(u_longlong_t)instructions);
} else {
if (json_output) {
(void) nvlist_print_json(stdout, outnvl);
} else if (nvlist_empty(outnvl)) {
(void) fprintf(stdout, gettext("Channel program fully "
"executed and did not produce output.\n"));
} else {
(void) fprintf(stdout, gettext("Channel program fully "
"executed and produced output:\n"));
dump_nvlist(outnvl, 4);
}
}
free(progbuf);
fnvlist_free(outnvl);
fnvlist_free(argnvl);
return (ret != 0);
usage:
usage(B_FALSE);
return (-1);
}
typedef struct loadkey_cbdata {
boolean_t cb_loadkey;
boolean_t cb_recursive;
boolean_t cb_noop;
char *cb_keylocation;
uint64_t cb_numfailed;
uint64_t cb_numattempted;
} loadkey_cbdata_t;
static int
load_key_callback(zfs_handle_t *zhp, void *data)
{
int ret;
boolean_t is_encroot;
loadkey_cbdata_t *cb = data;
uint64_t keystatus = zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS);
/*
* If we are working recursively, we want to skip loading / unloading
* keys for non-encryption roots and datasets whose keys are already
* in the desired end-state.
*/
if (cb->cb_recursive) {
ret = zfs_crypto_get_encryption_root(zhp, &is_encroot, NULL);
if (ret != 0)
return (ret);
if (!is_encroot)
return (0);
if ((cb->cb_loadkey && keystatus == ZFS_KEYSTATUS_AVAILABLE) ||
(!cb->cb_loadkey && keystatus == ZFS_KEYSTATUS_UNAVAILABLE))
return (0);
}
cb->cb_numattempted++;
if (cb->cb_loadkey)
ret = zfs_crypto_load_key(zhp, cb->cb_noop, cb->cb_keylocation);
else
ret = zfs_crypto_unload_key(zhp);
if (ret != 0) {
cb->cb_numfailed++;
return (ret);
}
return (0);
}
static int
load_unload_keys(int argc, char **argv, boolean_t loadkey)
{
int c, ret = 0, flags = 0;
boolean_t do_all = B_FALSE;
loadkey_cbdata_t cb = { 0 };
cb.cb_loadkey = loadkey;
while ((c = getopt(argc, argv, "anrL:")) != -1) {
/* noop and alternate keylocations only apply to zfs load-key */
if (loadkey) {
switch (c) {
case 'n':
cb.cb_noop = B_TRUE;
continue;
case 'L':
cb.cb_keylocation = optarg;
continue;
default:
break;
}
}
switch (c) {
case 'a':
do_all = B_TRUE;
cb.cb_recursive = B_TRUE;
break;
case 'r':
flags |= ZFS_ITER_RECURSE;
cb.cb_recursive = B_TRUE;
break;
default:
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (!do_all && argc == 0) {
(void) fprintf(stderr,
gettext("Missing dataset argument or -a option\n"));
usage(B_FALSE);
}
if (do_all && argc != 0) {
(void) fprintf(stderr,
gettext("Cannot specify dataset with -a option\n"));
usage(B_FALSE);
}
if (cb.cb_recursive && cb.cb_keylocation != NULL &&
strcmp(cb.cb_keylocation, "prompt") != 0) {
(void) fprintf(stderr, gettext("alternate keylocation may only "
"be 'prompt' with -r or -a\n"));
usage(B_FALSE);
}
ret = zfs_for_each(argc, argv, flags,
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME, NULL, NULL, 0,
load_key_callback, &cb);
if (cb.cb_noop || (cb.cb_recursive && cb.cb_numattempted != 0)) {
(void) printf(gettext("%llu / %llu key(s) successfully %s\n"),
(u_longlong_t)(cb.cb_numattempted - cb.cb_numfailed),
(u_longlong_t)cb.cb_numattempted,
loadkey ? (cb.cb_noop ? "verified" : "loaded") :
"unloaded");
}
if (cb.cb_numfailed != 0)
ret = -1;
return (ret);
}
static int
zfs_do_load_key(int argc, char **argv)
{
return (load_unload_keys(argc, argv, B_TRUE));
}
static int
zfs_do_unload_key(int argc, char **argv)
{
return (load_unload_keys(argc, argv, B_FALSE));
}
static int
zfs_do_change_key(int argc, char **argv)
{
int c, ret;
uint64_t keystatus;
boolean_t loadkey = B_FALSE, inheritkey = B_FALSE;
zfs_handle_t *zhp = NULL;
nvlist_t *props = fnvlist_alloc();
while ((c = getopt(argc, argv, "lio:")) != -1) {
switch (c) {
case 'l':
loadkey = B_TRUE;
break;
case 'i':
inheritkey = B_TRUE;
break;
case 'o':
if (!parseprop(props, optarg)) {
nvlist_free(props);
return (1);
}
break;
default:
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
usage(B_FALSE);
}
}
if (inheritkey && !nvlist_empty(props)) {
(void) fprintf(stderr,
gettext("Properties not allowed for inheriting\n"));
usage(B_FALSE);
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("Missing dataset argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("Too many arguments\n"));
usage(B_FALSE);
}
zhp = zfs_open(g_zfs, argv[argc - 1],
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL)
usage(B_FALSE);
if (loadkey) {
keystatus = zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS);
if (keystatus != ZFS_KEYSTATUS_AVAILABLE) {
ret = zfs_crypto_load_key(zhp, B_FALSE, NULL);
if (ret != 0) {
nvlist_free(props);
zfs_close(zhp);
return (-1);
}
}
/* refresh the properties so the new keystatus is visible */
zfs_refresh_properties(zhp);
}
ret = zfs_crypto_rewrap(zhp, props, inheritkey);
if (ret != 0) {
nvlist_free(props);
zfs_close(zhp);
return (-1);
}
nvlist_free(props);
zfs_close(zhp);
return (0);
}
/*
* 1) zfs project [-d|-r] <file|directory ...>
* List project ID and inherit flag of file(s) or directories.
* -d: List the directory itself, not its children.
* -r: List subdirectories recursively.
*
* 2) zfs project -C [-k] [-r] <file|directory ...>
* Clear project inherit flag and/or ID on the file(s) or directories.
* -k: Keep the project ID unchanged. If not specified, the project ID
* will be reset as zero.
* -r: Clear on subdirectories recursively.
*
* 3) zfs project -c [-0] [-d|-r] [-p id] <file|directory ...>
* Check project ID and inherit flag on the file(s) or directories,
* report the outliers.
* -0: Print file name followed by a NUL instead of newline.
* -d: Check the directory itself, not its children.
* -p: Specify the referenced ID for comparing with the target file(s)
* or directories' project IDs. If not specified, the target (top)
* directory's project ID will be used as the referenced one.
* -r: Check subdirectories recursively.
*
* 4) zfs project [-p id] [-r] [-s] <file|directory ...>
* Set project ID and/or inherit flag on the file(s) or directories.
* -p: Set the project ID as the given id.
* -r: Set on subdirectories recursively. If not specify "-p" option,
* it will use top-level directory's project ID as the given id,
* then set both project ID and inherit flag on all descendants
* of the top-level directory.
* -s: Set project inherit flag.
*/
static int
zfs_do_project(int argc, char **argv)
{
zfs_project_control_t zpc = {
.zpc_expected_projid = ZFS_INVALID_PROJID,
.zpc_op = ZFS_PROJECT_OP_DEFAULT,
.zpc_dironly = B_FALSE,
.zpc_keep_projid = B_FALSE,
.zpc_newline = B_TRUE,
.zpc_recursive = B_FALSE,
.zpc_set_flag = B_FALSE,
};
int ret = 0, c;
if (argc < 2)
usage(B_FALSE);
while ((c = getopt(argc, argv, "0Ccdkp:rs")) != -1) {
switch (c) {
case '0':
zpc.zpc_newline = B_FALSE;
break;
case 'C':
if (zpc.zpc_op != ZFS_PROJECT_OP_DEFAULT) {
(void) fprintf(stderr, gettext("cannot "
"specify '-C' '-c' '-s' together\n"));
usage(B_FALSE);
}
zpc.zpc_op = ZFS_PROJECT_OP_CLEAR;
break;
case 'c':
if (zpc.zpc_op != ZFS_PROJECT_OP_DEFAULT) {
(void) fprintf(stderr, gettext("cannot "
"specify '-C' '-c' '-s' together\n"));
usage(B_FALSE);
}
zpc.zpc_op = ZFS_PROJECT_OP_CHECK;
break;
case 'd':
zpc.zpc_dironly = B_TRUE;
/* overwrite "-r" option */
zpc.zpc_recursive = B_FALSE;
break;
case 'k':
zpc.zpc_keep_projid = B_TRUE;
break;
case 'p': {
char *endptr;
errno = 0;
zpc.zpc_expected_projid = strtoull(optarg, &endptr, 0);
if (errno != 0 || *endptr != '\0') {
(void) fprintf(stderr,
gettext("project ID must be less than "
"%u\n"), UINT32_MAX);
usage(B_FALSE);
}
if (zpc.zpc_expected_projid >= UINT32_MAX) {
(void) fprintf(stderr,
gettext("invalid project ID\n"));
usage(B_FALSE);
}
break;
}
case 'r':
zpc.zpc_recursive = B_TRUE;
/* overwrite "-d" option */
zpc.zpc_dironly = B_FALSE;
break;
case 's':
if (zpc.zpc_op != ZFS_PROJECT_OP_DEFAULT) {
(void) fprintf(stderr, gettext("cannot "
"specify '-C' '-c' '-s' together\n"));
usage(B_FALSE);
}
zpc.zpc_set_flag = B_TRUE;
zpc.zpc_op = ZFS_PROJECT_OP_SET;
break;
default:
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
if (zpc.zpc_op == ZFS_PROJECT_OP_DEFAULT) {
if (zpc.zpc_expected_projid != ZFS_INVALID_PROJID)
zpc.zpc_op = ZFS_PROJECT_OP_SET;
else
zpc.zpc_op = ZFS_PROJECT_OP_LIST;
}
switch (zpc.zpc_op) {
case ZFS_PROJECT_OP_LIST:
if (zpc.zpc_keep_projid) {
(void) fprintf(stderr,
gettext("'-k' is only valid together with '-C'\n"));
usage(B_FALSE);
}
if (!zpc.zpc_newline) {
(void) fprintf(stderr,
gettext("'-0' is only valid together with '-c'\n"));
usage(B_FALSE);
}
break;
case ZFS_PROJECT_OP_CHECK:
if (zpc.zpc_keep_projid) {
(void) fprintf(stderr,
gettext("'-k' is only valid together with '-C'\n"));
usage(B_FALSE);
}
break;
case ZFS_PROJECT_OP_CLEAR:
if (zpc.zpc_dironly) {
(void) fprintf(stderr,
gettext("'-d' is useless together with '-C'\n"));
usage(B_FALSE);
}
if (!zpc.zpc_newline) {
(void) fprintf(stderr,
gettext("'-0' is only valid together with '-c'\n"));
usage(B_FALSE);
}
if (zpc.zpc_expected_projid != ZFS_INVALID_PROJID) {
(void) fprintf(stderr,
gettext("'-p' is useless together with '-C'\n"));
usage(B_FALSE);
}
break;
case ZFS_PROJECT_OP_SET:
if (zpc.zpc_dironly) {
(void) fprintf(stderr,
gettext("'-d' is useless for set project ID and/or "
"inherit flag\n"));
usage(B_FALSE);
}
if (zpc.zpc_keep_projid) {
(void) fprintf(stderr,
gettext("'-k' is only valid together with '-C'\n"));
usage(B_FALSE);
}
if (!zpc.zpc_newline) {
(void) fprintf(stderr,
gettext("'-0' is only valid together with '-c'\n"));
usage(B_FALSE);
}
break;
default:
ASSERT(0);
break;
}
argv += optind;
argc -= optind;
if (argc == 0) {
(void) fprintf(stderr,
gettext("missing file or directory target(s)\n"));
usage(B_FALSE);
}
for (int i = 0; i < argc; i++) {
int err;
err = zfs_project_handle(argv[i], &zpc);
if (err && !ret)
ret = err;
}
return (ret);
}
static int
zfs_do_wait(int argc, char **argv)
{
boolean_t enabled[ZFS_WAIT_NUM_ACTIVITIES];
int error = 0, i;
int c;
/* By default, wait for all types of activity. */
for (i = 0; i < ZFS_WAIT_NUM_ACTIVITIES; i++)
enabled[i] = B_TRUE;
while ((c = getopt(argc, argv, "t:")) != -1) {
switch (c) {
case 't':
/* Reset activities array */
memset(&enabled, 0, sizeof (enabled));
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const col_subopts[
ZFS_WAIT_NUM_ACTIVITIES] = { "deleteq" };
for (i = 0; i < ARRAY_SIZE(col_subopts); ++i)
if (strcmp(tok, col_subopts[i]) == 0) {
enabled[i] = B_TRUE;
goto found;
}
(void) fprintf(stderr,
gettext("invalid activity '%s'\n"), tok);
usage(B_FALSE);
found:;
}
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argv += optind;
argc -= optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing 'filesystem' "
"argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
zfs_handle_t *zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_FILESYSTEM);
if (zhp == NULL)
return (1);
for (;;) {
boolean_t missing = B_FALSE;
boolean_t any_waited = B_FALSE;
for (int i = 0; i < ZFS_WAIT_NUM_ACTIVITIES; i++) {
boolean_t waited;
if (!enabled[i])
continue;
error = zfs_wait_status(zhp, i, &missing, &waited);
if (error != 0 || missing)
break;
any_waited = (any_waited || waited);
}
if (error != 0 || missing || !any_waited)
break;
}
zfs_close(zhp);
return (error);
}
/*
* Display version message
*/
static int
zfs_do_version(int argc, char **argv)
{
(void) argc, (void) argv;
return (zfs_version_print() != 0);
}
int
main(int argc, char **argv)
{
int ret = 0;
int i = 0;
const char *cmdname;
char **newargv;
(void) setlocale(LC_ALL, "");
(void) setlocale(LC_NUMERIC, "C");
(void) textdomain(TEXT_DOMAIN);
opterr = 0;
/*
* Make sure the user has specified some command.
*/
if (argc < 2) {
(void) fprintf(stderr, gettext("missing command\n"));
usage(B_FALSE);
}
cmdname = argv[1];
/*
* The 'umount' command is an alias for 'unmount'
*/
if (strcmp(cmdname, "umount") == 0)
cmdname = "unmount";
/*
* The 'recv' command is an alias for 'receive'
*/
if (strcmp(cmdname, "recv") == 0)
cmdname = "receive";
/*
* The 'snap' command is an alias for 'snapshot'
*/
if (strcmp(cmdname, "snap") == 0)
cmdname = "snapshot";
/*
* Special case '-?'
*/
if ((strcmp(cmdname, "-?") == 0) ||
(strcmp(cmdname, "--help") == 0))
usage(B_TRUE);
/*
* Special case '-V|--version'
*/
if ((strcmp(cmdname, "-V") == 0) || (strcmp(cmdname, "--version") == 0))
return (zfs_do_version(argc, argv));
if ((g_zfs = libzfs_init()) == NULL) {
(void) fprintf(stderr, "%s\n", libzfs_error_init(errno));
return (1);
}
zfs_save_arguments(argc, argv, history_str, sizeof (history_str));
libzfs_print_on_error(g_zfs, B_TRUE);
zfs_setproctitle_init(argc, argv, environ);
/*
* Many commands modify input strings for string parsing reasons.
* We create a copy to protect the original argv.
*/
newargv = safe_malloc((argc + 1) * sizeof (newargv[0]));
for (i = 0; i < argc; i++)
newargv[i] = strdup(argv[i]);
newargv[argc] = NULL;
/*
* Run the appropriate command.
*/
libzfs_mnttab_cache(g_zfs, B_TRUE);
if (find_command_idx(cmdname, &i) == 0) {
current_command = &command_table[i];
ret = command_table[i].func(argc - 1, newargv + 1);
} else if (strchr(cmdname, '=') != NULL) {
verify(find_command_idx("set", &i) == 0);
current_command = &command_table[i];
ret = command_table[i].func(argc, newargv);
} else {
(void) fprintf(stderr, gettext("unrecognized "
"command '%s'\n"), cmdname);
usage(B_FALSE);
ret = 1;
}
for (i = 0; i < argc; i++)
free(newargv[i]);
free(newargv);
if (ret == 0 && log_history)
(void) zpool_log_history(g_zfs, history_str);
libzfs_fini(g_zfs);
/*
* The 'ZFS_ABORT' environment variable causes us to dump core on exit
* for the purposes of running ::findleaks.
*/
if (getenv("ZFS_ABORT") != NULL) {
(void) printf("dumping core by request\n");
abort();
}
return (ret);
}
/*
* zfs zone nsfile filesystem
*
* Add or delete the given dataset to/from the namespace.
*/
#ifdef __linux__
static int
zfs_do_zone_impl(int argc, char **argv, boolean_t attach)
{
zfs_handle_t *zhp;
int ret;
if (argc < 3) {
(void) fprintf(stderr, gettext("missing argument(s)\n"));
usage(B_FALSE);
}
if (argc > 3) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
zhp = zfs_open(g_zfs, argv[2], ZFS_TYPE_FILESYSTEM);
if (zhp == NULL)
return (1);
ret = (zfs_userns(zhp, argv[1], attach) != 0);
zfs_close(zhp);
return (ret);
}
static int
zfs_do_zone(int argc, char **argv)
{
return (zfs_do_zone_impl(argc, argv, B_TRUE));
}
static int
zfs_do_unzone(int argc, char **argv)
{
return (zfs_do_zone_impl(argc, argv, B_FALSE));
}
#endif
#ifdef __FreeBSD__
#include <sys/jail.h>
#include <jail.h>
/*
* Attach/detach the given dataset to/from the given jail
*/
static int
zfs_do_jail_impl(int argc, char **argv, boolean_t attach)
{
zfs_handle_t *zhp;
int jailid, ret;
/* check number of arguments */
if (argc < 3) {
(void) fprintf(stderr, gettext("missing argument(s)\n"));
usage(B_FALSE);
}
if (argc > 3) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
jailid = jail_getid(argv[1]);
if (jailid < 0) {
(void) fprintf(stderr, gettext("invalid jail id or name\n"));
usage(B_FALSE);
}
zhp = zfs_open(g_zfs, argv[2], ZFS_TYPE_FILESYSTEM);
if (zhp == NULL)
return (1);
ret = (zfs_jail(zhp, jailid, attach) != 0);
zfs_close(zhp);
return (ret);
}
/*
* zfs jail jailid filesystem
*
* Attach the given dataset to the given jail
*/
static int
zfs_do_jail(int argc, char **argv)
{
return (zfs_do_jail_impl(argc, argv, B_TRUE));
}
/*
* zfs unjail jailid filesystem
*
* Detach the given dataset from the given jail
*/
static int
zfs_do_unjail(int argc, char **argv)
{
return (zfs_do_jail_impl(argc, argv, B_FALSE));
}
#endif
diff --git a/sys/contrib/openzfs/cmd/zpool/zpool_main.c b/sys/contrib/openzfs/cmd/zpool/zpool_main.c
index bf70f68fb431..301c5f4bfc6f 100644
--- a/sys/contrib/openzfs/cmd/zpool/zpool_main.c
+++ b/sys/contrib/openzfs/cmd/zpool/zpool_main.c
@@ -1,11015 +1,11060 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2012 by Frederik Wessels. All rights reserved.
* Copyright (c) 2012 by Cyril Plisko. All rights reserved.
* Copyright (c) 2013 by Prasad Joshi (sTec). All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>.
* Copyright (c) 2017 Datto Inc.
* Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, loli10K <ezomori.nozomu@gmail.com>
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
* Copyright (c) 2021, Klara Inc.
* Copyright [2021] Hewlett Packard Enterprise Development LP
*/
#include <assert.h>
#include <ctype.h>
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include <getopt.h>
#include <libgen.h>
#include <libintl.h>
#include <libuutil.h>
#include <locale.h>
#include <pthread.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>
#include <unistd.h>
#include <pwd.h>
#include <zone.h>
#include <sys/wait.h>
#include <zfs_prop.h>
#include <sys/fs/zfs.h>
#include <sys/stat.h>
#include <sys/systeminfo.h>
#include <sys/fm/fs/zfs.h>
#include <sys/fm/util.h>
#include <sys/fm/protocol.h>
#include <sys/zfs_ioctl.h>
#include <sys/mount.h>
#include <sys/sysmacros.h>
#include <math.h>
#include <libzfs.h>
#include <libzutil.h>
#include "zpool_util.h"
#include "zfs_comutil.h"
#include "zfeature_common.h"
#include "statcommon.h"
libzfs_handle_t *g_zfs;
static int zpool_do_create(int, char **);
static int zpool_do_destroy(int, char **);
static int zpool_do_add(int, char **);
static int zpool_do_remove(int, char **);
static int zpool_do_labelclear(int, char **);
static int zpool_do_checkpoint(int, char **);
static int zpool_do_list(int, char **);
static int zpool_do_iostat(int, char **);
static int zpool_do_status(int, char **);
static int zpool_do_online(int, char **);
static int zpool_do_offline(int, char **);
static int zpool_do_clear(int, char **);
static int zpool_do_reopen(int, char **);
static int zpool_do_reguid(int, char **);
static int zpool_do_attach(int, char **);
static int zpool_do_detach(int, char **);
static int zpool_do_replace(int, char **);
static int zpool_do_split(int, char **);
static int zpool_do_initialize(int, char **);
static int zpool_do_scrub(int, char **);
static int zpool_do_resilver(int, char **);
static int zpool_do_trim(int, char **);
static int zpool_do_import(int, char **);
static int zpool_do_export(int, char **);
static int zpool_do_upgrade(int, char **);
static int zpool_do_history(int, char **);
static int zpool_do_events(int, char **);
static int zpool_do_get(int, char **);
static int zpool_do_set(int, char **);
static int zpool_do_sync(int, char **);
static int zpool_do_version(int, char **);
static int zpool_do_wait(int, char **);
static zpool_compat_status_t zpool_do_load_compat(
const char *, boolean_t *);
/*
* These libumem hooks provide a reasonable set of defaults for the allocator's
* debugging facilities.
*/
#ifdef DEBUG
const char *
_umem_debug_init(void)
{
return ("default,verbose"); /* $UMEM_DEBUG setting */
}
const char *
_umem_logging_init(void)
{
return ("fail,contents"); /* $UMEM_LOGGING setting */
}
#endif
typedef enum {
HELP_ADD,
HELP_ATTACH,
HELP_CLEAR,
HELP_CREATE,
HELP_CHECKPOINT,
HELP_DESTROY,
HELP_DETACH,
HELP_EXPORT,
HELP_HISTORY,
HELP_IMPORT,
HELP_IOSTAT,
HELP_LABELCLEAR,
HELP_LIST,
HELP_OFFLINE,
HELP_ONLINE,
HELP_REPLACE,
HELP_REMOVE,
HELP_INITIALIZE,
HELP_SCRUB,
HELP_RESILVER,
HELP_TRIM,
HELP_STATUS,
HELP_UPGRADE,
HELP_EVENTS,
HELP_GET,
HELP_SET,
HELP_SPLIT,
HELP_SYNC,
HELP_REGUID,
HELP_REOPEN,
HELP_VERSION,
HELP_WAIT
} zpool_help_t;
/*
* Flags for stats to display with "zpool iostats"
*/
enum iostat_type {
IOS_DEFAULT = 0,
IOS_LATENCY = 1,
IOS_QUEUES = 2,
IOS_L_HISTO = 3,
IOS_RQ_HISTO = 4,
IOS_COUNT, /* always last element */
};
/* iostat_type entries as bitmasks */
#define IOS_DEFAULT_M (1ULL << IOS_DEFAULT)
#define IOS_LATENCY_M (1ULL << IOS_LATENCY)
#define IOS_QUEUES_M (1ULL << IOS_QUEUES)
#define IOS_L_HISTO_M (1ULL << IOS_L_HISTO)
#define IOS_RQ_HISTO_M (1ULL << IOS_RQ_HISTO)
/* Mask of all the histo bits */
#define IOS_ANYHISTO_M (IOS_L_HISTO_M | IOS_RQ_HISTO_M)
/*
* Lookup table for iostat flags to nvlist names. Basically a list
* of all the nvlists a flag requires. Also specifies the order in
* which data gets printed in zpool iostat.
*/
static const char *vsx_type_to_nvlist[IOS_COUNT][15] = {
[IOS_L_HISTO] = {
ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO,
ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO,
NULL},
[IOS_LATENCY] = {
ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO,
ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO,
NULL},
[IOS_QUEUES] = {
ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE,
NULL},
[IOS_RQ_HISTO] = {
ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO,
ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO,
ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO,
ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO,
ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO,
ZPOOL_CONFIG_VDEV_IND_REBUILD_HISTO,
ZPOOL_CONFIG_VDEV_AGG_REBUILD_HISTO,
NULL},
};
/*
* Given a cb->cb_flags with a histogram bit set, return the iostat_type.
* Right now, only one histo bit is ever set at one time, so we can
* just do a highbit64(a)
*/
#define IOS_HISTO_IDX(a) (highbit64(a & IOS_ANYHISTO_M) - 1)
typedef struct zpool_command {
const char *name;
int (*func)(int, char **);
zpool_help_t usage;
} zpool_command_t;
/*
* Master command table. Each ZFS command has a name, associated function, and
* usage message. The usage messages need to be internationalized, so we have
* to have a function to return the usage message based on a command index.
*
* These commands are organized according to how they are displayed in the usage
* message. An empty command (one with a NULL name) indicates an empty line in
* the generic usage message.
*/
static zpool_command_t command_table[] = {
{ "version", zpool_do_version, HELP_VERSION },
{ NULL },
{ "create", zpool_do_create, HELP_CREATE },
{ "destroy", zpool_do_destroy, HELP_DESTROY },
{ NULL },
{ "add", zpool_do_add, HELP_ADD },
{ "remove", zpool_do_remove, HELP_REMOVE },
{ NULL },
{ "labelclear", zpool_do_labelclear, HELP_LABELCLEAR },
{ NULL },
{ "checkpoint", zpool_do_checkpoint, HELP_CHECKPOINT },
{ NULL },
{ "list", zpool_do_list, HELP_LIST },
{ "iostat", zpool_do_iostat, HELP_IOSTAT },
{ "status", zpool_do_status, HELP_STATUS },
{ NULL },
{ "online", zpool_do_online, HELP_ONLINE },
{ "offline", zpool_do_offline, HELP_OFFLINE },
{ "clear", zpool_do_clear, HELP_CLEAR },
{ "reopen", zpool_do_reopen, HELP_REOPEN },
{ NULL },
{ "attach", zpool_do_attach, HELP_ATTACH },
{ "detach", zpool_do_detach, HELP_DETACH },
{ "replace", zpool_do_replace, HELP_REPLACE },
{ "split", zpool_do_split, HELP_SPLIT },
{ NULL },
{ "initialize", zpool_do_initialize, HELP_INITIALIZE },
{ "resilver", zpool_do_resilver, HELP_RESILVER },
{ "scrub", zpool_do_scrub, HELP_SCRUB },
{ "trim", zpool_do_trim, HELP_TRIM },
{ NULL },
{ "import", zpool_do_import, HELP_IMPORT },
{ "export", zpool_do_export, HELP_EXPORT },
{ "upgrade", zpool_do_upgrade, HELP_UPGRADE },
{ "reguid", zpool_do_reguid, HELP_REGUID },
{ NULL },
{ "history", zpool_do_history, HELP_HISTORY },
{ "events", zpool_do_events, HELP_EVENTS },
{ NULL },
{ "get", zpool_do_get, HELP_GET },
{ "set", zpool_do_set, HELP_SET },
{ "sync", zpool_do_sync, HELP_SYNC },
{ NULL },
{ "wait", zpool_do_wait, HELP_WAIT },
};
#define NCOMMAND (ARRAY_SIZE(command_table))
#define VDEV_ALLOC_CLASS_LOGS "logs"
static zpool_command_t *current_command;
static zfs_type_t current_prop_type = (ZFS_TYPE_POOL | ZFS_TYPE_VDEV);
static char history_str[HIS_MAX_RECORD_LEN];
static boolean_t log_history = B_TRUE;
static uint_t timestamp_fmt = NODATE;
static const char *
get_usage(zpool_help_t idx)
{
switch (idx) {
case HELP_ADD:
return (gettext("\tadd [-fgLnP] [-o property=value] "
"<pool> <vdev> ...\n"));
case HELP_ATTACH:
return (gettext("\tattach [-fsw] [-o property=value] "
"<pool> <device> <new-device>\n"));
case HELP_CLEAR:
return (gettext("\tclear [-nF] <pool> [device]\n"));
case HELP_CREATE:
return (gettext("\tcreate [-fnd] [-o property=value] ... \n"
"\t [-O file-system-property=value] ... \n"
"\t [-m mountpoint] [-R root] <pool> <vdev> ...\n"));
case HELP_CHECKPOINT:
return (gettext("\tcheckpoint [-d [-w]] <pool> ...\n"));
case HELP_DESTROY:
return (gettext("\tdestroy [-f] <pool>\n"));
case HELP_DETACH:
return (gettext("\tdetach <pool> <device>\n"));
case HELP_EXPORT:
return (gettext("\texport [-af] <pool> ...\n"));
case HELP_HISTORY:
return (gettext("\thistory [-il] [<pool>] ...\n"));
case HELP_IMPORT:
return (gettext("\timport [-d dir] [-D]\n"
"\timport [-o mntopts] [-o property=value] ... \n"
"\t [-d dir | -c cachefile] [-D] [-l] [-f] [-m] [-N] "
"[-R root] [-F [-n]] -a\n"
"\timport [-o mntopts] [-o property=value] ... \n"
"\t [-d dir | -c cachefile] [-D] [-l] [-f] [-m] [-N] "
"[-R root] [-F [-n]]\n"
"\t [--rewind-to-checkpoint] <pool | id> [newpool]\n"));
case HELP_IOSTAT:
return (gettext("\tiostat [[[-c [script1,script2,...]"
"[-lq]]|[-rw]] [-T d | u] [-ghHLpPvy]\n"
"\t [[pool ...]|[pool vdev ...]|[vdev ...]]"
" [[-n] interval [count]]\n"));
case HELP_LABELCLEAR:
return (gettext("\tlabelclear [-f] <vdev>\n"));
case HELP_LIST:
return (gettext("\tlist [-gHLpPv] [-o property[,...]] "
"[-T d|u] [pool] ... \n"
"\t [interval [count]]\n"));
case HELP_OFFLINE:
return (gettext("\toffline [-f] [-t] <pool> <device> ...\n"));
case HELP_ONLINE:
return (gettext("\tonline [-e] <pool> <device> ...\n"));
case HELP_REPLACE:
return (gettext("\treplace [-fsw] [-o property=value] "
"<pool> <device> [new-device]\n"));
case HELP_REMOVE:
return (gettext("\tremove [-npsw] <pool> <device> ...\n"));
case HELP_REOPEN:
return (gettext("\treopen [-n] <pool>\n"));
case HELP_INITIALIZE:
return (gettext("\tinitialize [-c | -s] [-w] <pool> "
"[<device> ...]\n"));
case HELP_SCRUB:
return (gettext("\tscrub [-s | -p] [-w] <pool> ...\n"));
case HELP_RESILVER:
return (gettext("\tresilver <pool> ...\n"));
case HELP_TRIM:
return (gettext("\ttrim [-dw] [-r <rate>] [-c | -s] <pool> "
"[<device> ...]\n"));
case HELP_STATUS:
return (gettext("\tstatus [-c [script1,script2,...]] "
"[-igLpPstvxD] [-T d|u] [pool] ... \n"
"\t [interval [count]]\n"));
case HELP_UPGRADE:
return (gettext("\tupgrade\n"
"\tupgrade -v\n"
"\tupgrade [-V version] <-a | pool ...>\n"));
case HELP_EVENTS:
return (gettext("\tevents [-vHf [pool] | -c]\n"));
case HELP_GET:
return (gettext("\tget [-Hp] [-o \"all\" | field[,...]] "
"<\"all\" | property[,...]> <pool> ...\n"));
case HELP_SET:
return (gettext("\tset <property=value> <pool>\n"
"\tset <vdev_property=value> <pool> <vdev>\n"));
case HELP_SPLIT:
return (gettext("\tsplit [-gLnPl] [-R altroot] [-o mntopts]\n"
"\t [-o property=value] <pool> <newpool> "
"[<device> ...]\n"));
case HELP_REGUID:
return (gettext("\treguid <pool>\n"));
case HELP_SYNC:
return (gettext("\tsync [pool] ...\n"));
case HELP_VERSION:
return (gettext("\tversion\n"));
case HELP_WAIT:
return (gettext("\twait [-Hp] [-T d|u] [-t <activity>[,...]] "
"<pool> [interval]\n"));
default:
__builtin_unreachable();
}
}
static void
zpool_collect_leaves(zpool_handle_t *zhp, nvlist_t *nvroot, nvlist_t *res)
{
uint_t children = 0;
nvlist_t **child;
uint_t i;
(void) nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children);
if (children == 0) {
char *path = zpool_vdev_name(g_zfs, zhp, nvroot,
VDEV_NAME_PATH);
if (strcmp(path, VDEV_TYPE_INDIRECT) != 0 &&
strcmp(path, VDEV_TYPE_HOLE) != 0)
fnvlist_add_boolean(res, path);
free(path);
return;
}
for (i = 0; i < children; i++) {
zpool_collect_leaves(zhp, child[i], res);
}
}
/*
* Callback routine that will print out a pool property value.
*/
static int
print_pool_prop_cb(int prop, void *cb)
{
FILE *fp = cb;
(void) fprintf(fp, "\t%-19s ", zpool_prop_to_name(prop));
if (zpool_prop_readonly(prop))
(void) fprintf(fp, " NO ");
else
(void) fprintf(fp, " YES ");
if (zpool_prop_values(prop) == NULL)
(void) fprintf(fp, "-\n");
else
(void) fprintf(fp, "%s\n", zpool_prop_values(prop));
return (ZPROP_CONT);
}
/*
* Callback routine that will print out a vdev property value.
*/
static int
print_vdev_prop_cb(int prop, void *cb)
{
FILE *fp = cb;
(void) fprintf(fp, "\t%-19s ", vdev_prop_to_name(prop));
if (vdev_prop_readonly(prop))
(void) fprintf(fp, " NO ");
else
(void) fprintf(fp, " YES ");
if (vdev_prop_values(prop) == NULL)
(void) fprintf(fp, "-\n");
else
(void) fprintf(fp, "%s\n", vdev_prop_values(prop));
return (ZPROP_CONT);
}
/*
* Display usage message. If we're inside a command, display only the usage for
* that command. Otherwise, iterate over the entire command table and display
* a complete usage message.
*/
static __attribute__((noreturn)) void
usage(boolean_t requested)
{
FILE *fp = requested ? stdout : stderr;
if (current_command == NULL) {
int i;
(void) fprintf(fp, gettext("usage: zpool command args ...\n"));
(void) fprintf(fp,
gettext("where 'command' is one of the following:\n\n"));
for (i = 0; i < NCOMMAND; i++) {
if (command_table[i].name == NULL)
(void) fprintf(fp, "\n");
else
(void) fprintf(fp, "%s",
get_usage(command_table[i].usage));
}
} else {
(void) fprintf(fp, gettext("usage:\n"));
(void) fprintf(fp, "%s", get_usage(current_command->usage));
}
if (current_command != NULL &&
current_prop_type != (ZFS_TYPE_POOL | ZFS_TYPE_VDEV) &&
((strcmp(current_command->name, "set") == 0) ||
(strcmp(current_command->name, "get") == 0) ||
(strcmp(current_command->name, "list") == 0))) {
(void) fprintf(fp, "%s",
gettext("\nthe following properties are supported:\n"));
(void) fprintf(fp, "\n\t%-19s %s %s\n\n",
"PROPERTY", "EDIT", "VALUES");
/* Iterate over all properties */
if (current_prop_type == ZFS_TYPE_POOL) {
(void) zprop_iter(print_pool_prop_cb, fp, B_FALSE,
B_TRUE, current_prop_type);
(void) fprintf(fp, "\t%-19s ", "feature@...");
(void) fprintf(fp, "YES "
"disabled | enabled | active\n");
(void) fprintf(fp, gettext("\nThe feature@ properties "
"must be appended with a feature name.\n"
"See zpool-features(7).\n"));
} else if (current_prop_type == ZFS_TYPE_VDEV) {
(void) zprop_iter(print_vdev_prop_cb, fp, B_FALSE,
B_TRUE, current_prop_type);
}
}
/*
* See comments at end of main().
*/
if (getenv("ZFS_ABORT") != NULL) {
(void) printf("dumping core by request\n");
abort();
}
exit(requested ? 0 : 2);
}
/*
* zpool initialize [-c | -s] [-w] <pool> [<vdev> ...]
* Initialize all unused blocks in the specified vdevs, or all vdevs in the pool
* if none specified.
*
* -c Cancel. Ends active initializing.
* -s Suspend. Initializing can then be restarted with no flags.
* -w Wait. Blocks until initializing has completed.
*/
int
zpool_do_initialize(int argc, char **argv)
{
int c;
char *poolname;
zpool_handle_t *zhp;
nvlist_t *vdevs;
int err = 0;
boolean_t wait = B_FALSE;
struct option long_options[] = {
{"cancel", no_argument, NULL, 'c'},
{"suspend", no_argument, NULL, 's'},
{"wait", no_argument, NULL, 'w'},
{0, 0, 0, 0}
};
pool_initialize_func_t cmd_type = POOL_INITIALIZE_START;
while ((c = getopt_long(argc, argv, "csw", long_options, NULL)) != -1) {
switch (c) {
case 'c':
if (cmd_type != POOL_INITIALIZE_START &&
cmd_type != POOL_INITIALIZE_CANCEL) {
(void) fprintf(stderr, gettext("-c cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_INITIALIZE_CANCEL;
break;
case 's':
if (cmd_type != POOL_INITIALIZE_START &&
cmd_type != POOL_INITIALIZE_SUSPEND) {
(void) fprintf(stderr, gettext("-s cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_INITIALIZE_SUSPEND;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
if (optopt != 0) {
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
} else {
(void) fprintf(stderr,
gettext("invalid option '%s'\n"),
argv[optind - 1]);
}
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
return (-1);
}
if (wait && (cmd_type != POOL_INITIALIZE_START)) {
(void) fprintf(stderr, gettext("-w cannot be used with -c or "
"-s\n"));
usage(B_FALSE);
}
poolname = argv[0];
zhp = zpool_open(g_zfs, poolname);
if (zhp == NULL)
return (-1);
vdevs = fnvlist_alloc();
if (argc == 1) {
/* no individual leaf vdevs specified, so add them all */
nvlist_t *config = zpool_get_config(zhp, NULL);
nvlist_t *nvroot = fnvlist_lookup_nvlist(config,
ZPOOL_CONFIG_VDEV_TREE);
zpool_collect_leaves(zhp, nvroot, vdevs);
} else {
for (int i = 1; i < argc; i++) {
fnvlist_add_boolean(vdevs, argv[i]);
}
}
if (wait)
err = zpool_initialize_wait(zhp, cmd_type, vdevs);
else
err = zpool_initialize(zhp, cmd_type, vdevs);
fnvlist_free(vdevs);
zpool_close(zhp);
return (err);
}
/*
* print a pool vdev config for dry runs
*/
static void
print_vdev_tree(zpool_handle_t *zhp, const char *name, nvlist_t *nv, int indent,
const char *match, int name_flags)
{
nvlist_t **child;
uint_t c, children;
char *vname;
boolean_t printed = B_FALSE;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0) {
if (name != NULL)
(void) printf("\t%*s%s\n", indent, "", name);
return;
}
for (c = 0; c < children; c++) {
uint64_t is_log = B_FALSE, is_hole = B_FALSE;
const char *class = "";
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE,
&is_hole);
if (is_hole == B_TRUE) {
continue;
}
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&is_log);
if (is_log)
class = VDEV_ALLOC_BIAS_LOG;
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &class);
if (strcmp(match, class) != 0)
continue;
if (!printed && name != NULL) {
(void) printf("\t%*s%s\n", indent, "", name);
printed = B_TRUE;
}
vname = zpool_vdev_name(g_zfs, zhp, child[c], name_flags);
print_vdev_tree(zhp, vname, child[c], indent + 2, "",
name_flags);
free(vname);
}
}
/*
* Print the list of l2cache devices for dry runs.
*/
static void
print_cache_list(nvlist_t *nv, int indent)
{
nvlist_t **child;
uint_t c, children;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE,
&child, &children) == 0 && children > 0) {
(void) printf("\t%*s%s\n", indent, "", "cache");
} else {
return;
}
for (c = 0; c < children; c++) {
char *vname;
vname = zpool_vdev_name(g_zfs, NULL, child[c], 0);
(void) printf("\t%*s%s\n", indent + 2, "", vname);
free(vname);
}
}
/*
* Print the list of spares for dry runs.
*/
static void
print_spare_list(nvlist_t *nv, int indent)
{
nvlist_t **child;
uint_t c, children;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
&child, &children) == 0 && children > 0) {
(void) printf("\t%*s%s\n", indent, "", "spares");
} else {
return;
}
for (c = 0; c < children; c++) {
char *vname;
vname = zpool_vdev_name(g_zfs, NULL, child[c], 0);
(void) printf("\t%*s%s\n", indent + 2, "", vname);
free(vname);
}
}
static boolean_t
prop_list_contains_feature(nvlist_t *proplist)
{
nvpair_t *nvp;
for (nvp = nvlist_next_nvpair(proplist, NULL); NULL != nvp;
nvp = nvlist_next_nvpair(proplist, nvp)) {
if (zpool_prop_feature(nvpair_name(nvp)))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Add a property pair (name, string-value) into a property nvlist.
*/
static int
add_prop_list(const char *propname, const char *propval, nvlist_t **props,
boolean_t poolprop)
{
zpool_prop_t prop = ZPOOL_PROP_INVAL;
nvlist_t *proplist;
const char *normnm;
const char *strval;
if (*props == NULL &&
nvlist_alloc(props, NV_UNIQUE_NAME, 0) != 0) {
(void) fprintf(stderr,
gettext("internal error: out of memory\n"));
return (1);
}
proplist = *props;
if (poolprop) {
const char *vname = zpool_prop_to_name(ZPOOL_PROP_VERSION);
const char *cname =
zpool_prop_to_name(ZPOOL_PROP_COMPATIBILITY);
if ((prop = zpool_name_to_prop(propname)) == ZPOOL_PROP_INVAL &&
(!zpool_prop_feature(propname) &&
!zpool_prop_vdev(propname))) {
(void) fprintf(stderr, gettext("property '%s' is "
"not a valid pool or vdev property\n"), propname);
return (2);
}
/*
* feature@ properties and version should not be specified
* at the same time.
*/
if ((prop == ZPOOL_PROP_INVAL && zpool_prop_feature(propname) &&
nvlist_exists(proplist, vname)) ||
(prop == ZPOOL_PROP_VERSION &&
prop_list_contains_feature(proplist))) {
(void) fprintf(stderr, gettext("'feature@' and "
"'version' properties cannot be specified "
"together\n"));
return (2);
}
/*
* if version is specified, only "legacy" compatibility
* may be requested
*/
if ((prop == ZPOOL_PROP_COMPATIBILITY &&
strcmp(propval, ZPOOL_COMPAT_LEGACY) != 0 &&
nvlist_exists(proplist, vname)) ||
(prop == ZPOOL_PROP_VERSION &&
nvlist_exists(proplist, cname) &&
strcmp(fnvlist_lookup_string(proplist, cname),
ZPOOL_COMPAT_LEGACY) != 0)) {
(void) fprintf(stderr, gettext("when 'version' is "
"specified, the 'compatibility' feature may only "
"be set to '" ZPOOL_COMPAT_LEGACY "'\n"));
return (2);
}
if (zpool_prop_feature(propname) || zpool_prop_vdev(propname))
normnm = propname;
else
normnm = zpool_prop_to_name(prop);
} else {
zfs_prop_t fsprop = zfs_name_to_prop(propname);
if (zfs_prop_valid_for_type(fsprop, ZFS_TYPE_FILESYSTEM,
B_FALSE)) {
normnm = zfs_prop_to_name(fsprop);
} else if (zfs_prop_user(propname) ||
zfs_prop_userquota(propname)) {
normnm = propname;
} else {
(void) fprintf(stderr, gettext("property '%s' is "
"not a valid filesystem property\n"), propname);
return (2);
}
}
if (nvlist_lookup_string(proplist, normnm, &strval) == 0 &&
prop != ZPOOL_PROP_CACHEFILE) {
(void) fprintf(stderr, gettext("property '%s' "
"specified multiple times\n"), propname);
return (2);
}
if (nvlist_add_string(proplist, normnm, propval) != 0) {
(void) fprintf(stderr, gettext("internal "
"error: out of memory\n"));
return (1);
}
return (0);
}
/*
* Set a default property pair (name, string-value) in a property nvlist
*/
static int
add_prop_list_default(const char *propname, const char *propval,
nvlist_t **props)
{
const char *pval;
if (nvlist_lookup_string(*props, propname, &pval) == 0)
return (0);
return (add_prop_list(propname, propval, props, B_TRUE));
}
/*
* zpool add [-fgLnP] [-o property=value] <pool> <vdev> ...
*
* -f Force addition of devices, even if they appear in use
* -g Display guid for individual vdev name.
* -L Follow links when resolving vdev path name.
* -n Do not add the devices, but display the resulting layout if
* they were to be added.
* -o Set property=value.
* -P Display full path for vdev name.
*
* Adds the given vdevs to 'pool'. As with create, the bulk of this work is
* handled by make_root_vdev(), which constructs the nvlist needed to pass to
* libzfs.
*/
int
zpool_do_add(int argc, char **argv)
{
boolean_t force = B_FALSE;
boolean_t dryrun = B_FALSE;
int name_flags = 0;
int c;
nvlist_t *nvroot;
char *poolname;
int ret;
zpool_handle_t *zhp;
nvlist_t *config;
nvlist_t *props = NULL;
char *propval;
/* check options */
while ((c = getopt(argc, argv, "fgLno:P")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
case 'g':
name_flags |= VDEV_NAME_GUID;
break;
case 'L':
name_flags |= VDEV_NAME_FOLLOW_LINKS;
break;
case 'n':
dryrun = B_TRUE;
break;
case 'o':
if ((propval = strchr(optarg, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for -o option\n"));
usage(B_FALSE);
}
*propval = '\0';
propval++;
if ((strcmp(optarg, ZPOOL_CONFIG_ASHIFT) != 0) ||
(add_prop_list(optarg, propval, &props, B_TRUE)))
usage(B_FALSE);
break;
case 'P':
name_flags |= VDEV_NAME_PATH;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing vdev specification\n"));
usage(B_FALSE);
}
poolname = argv[0];
argc--;
argv++;
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
if ((config = zpool_get_config(zhp, NULL)) == NULL) {
(void) fprintf(stderr, gettext("pool '%s' is unavailable\n"),
poolname);
zpool_close(zhp);
return (1);
}
/* unless manually specified use "ashift" pool property (if set) */
if (!nvlist_exists(props, ZPOOL_CONFIG_ASHIFT)) {
int intval;
zprop_source_t src;
char strval[ZPOOL_MAXPROPLEN];
intval = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &src);
if (src != ZPROP_SRC_DEFAULT) {
(void) sprintf(strval, "%" PRId32, intval);
verify(add_prop_list(ZPOOL_CONFIG_ASHIFT, strval,
&props, B_TRUE) == 0);
}
}
/* pass off to make_root_vdev for processing */
nvroot = make_root_vdev(zhp, props, force, !force, B_FALSE, dryrun,
argc, argv);
if (nvroot == NULL) {
zpool_close(zhp);
return (1);
}
if (dryrun) {
nvlist_t *poolnvroot;
nvlist_t **l2child, **sparechild;
uint_t l2children, sparechildren, c;
char *vname;
boolean_t hadcache = B_FALSE, hadspare = B_FALSE;
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&poolnvroot) == 0);
(void) printf(gettext("would update '%s' to the following "
"configuration:\n\n"), zpool_get_name(zhp));
/* print original main pool and new tree */
print_vdev_tree(zhp, poolname, poolnvroot, 0, "",
name_flags | VDEV_NAME_TYPE_ID);
print_vdev_tree(zhp, NULL, nvroot, 0, "", name_flags);
/* print other classes: 'dedup', 'special', and 'log' */
if (zfs_special_devs(poolnvroot, VDEV_ALLOC_BIAS_DEDUP)) {
print_vdev_tree(zhp, "dedup", poolnvroot, 0,
VDEV_ALLOC_BIAS_DEDUP, name_flags);
print_vdev_tree(zhp, NULL, nvroot, 0,
VDEV_ALLOC_BIAS_DEDUP, name_flags);
} else if (zfs_special_devs(nvroot, VDEV_ALLOC_BIAS_DEDUP)) {
print_vdev_tree(zhp, "dedup", nvroot, 0,
VDEV_ALLOC_BIAS_DEDUP, name_flags);
}
if (zfs_special_devs(poolnvroot, VDEV_ALLOC_BIAS_SPECIAL)) {
print_vdev_tree(zhp, "special", poolnvroot, 0,
VDEV_ALLOC_BIAS_SPECIAL, name_flags);
print_vdev_tree(zhp, NULL, nvroot, 0,
VDEV_ALLOC_BIAS_SPECIAL, name_flags);
} else if (zfs_special_devs(nvroot, VDEV_ALLOC_BIAS_SPECIAL)) {
print_vdev_tree(zhp, "special", nvroot, 0,
VDEV_ALLOC_BIAS_SPECIAL, name_flags);
}
if (num_logs(poolnvroot) > 0) {
print_vdev_tree(zhp, "logs", poolnvroot, 0,
VDEV_ALLOC_BIAS_LOG, name_flags);
print_vdev_tree(zhp, NULL, nvroot, 0,
VDEV_ALLOC_BIAS_LOG, name_flags);
} else if (num_logs(nvroot) > 0) {
print_vdev_tree(zhp, "logs", nvroot, 0,
VDEV_ALLOC_BIAS_LOG, name_flags);
}
/* Do the same for the caches */
if (nvlist_lookup_nvlist_array(poolnvroot, ZPOOL_CONFIG_L2CACHE,
&l2child, &l2children) == 0 && l2children) {
hadcache = B_TRUE;
(void) printf(gettext("\tcache\n"));
for (c = 0; c < l2children; c++) {
vname = zpool_vdev_name(g_zfs, NULL,
l2child[c], name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2child, &l2children) == 0 && l2children) {
if (!hadcache)
(void) printf(gettext("\tcache\n"));
for (c = 0; c < l2children; c++) {
vname = zpool_vdev_name(g_zfs, NULL,
l2child[c], name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
/* And finally the spares */
if (nvlist_lookup_nvlist_array(poolnvroot, ZPOOL_CONFIG_SPARES,
&sparechild, &sparechildren) == 0 && sparechildren > 0) {
hadspare = B_TRUE;
(void) printf(gettext("\tspares\n"));
for (c = 0; c < sparechildren; c++) {
vname = zpool_vdev_name(g_zfs, NULL,
sparechild[c], name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&sparechild, &sparechildren) == 0 && sparechildren > 0) {
if (!hadspare)
(void) printf(gettext("\tspares\n"));
for (c = 0; c < sparechildren; c++) {
vname = zpool_vdev_name(g_zfs, NULL,
sparechild[c], name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
ret = 0;
} else {
ret = (zpool_add(zhp, nvroot) != 0);
}
nvlist_free(props);
nvlist_free(nvroot);
zpool_close(zhp);
return (ret);
}
/*
* zpool remove [-npsw] <pool> <vdev> ...
*
* Removes the given vdev from the pool.
*/
int
zpool_do_remove(int argc, char **argv)
{
char *poolname;
int i, ret = 0;
zpool_handle_t *zhp = NULL;
boolean_t stop = B_FALSE;
int c;
boolean_t noop = B_FALSE;
boolean_t parsable = B_FALSE;
boolean_t wait = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "npsw")) != -1) {
switch (c) {
case 'n':
noop = B_TRUE;
break;
case 'p':
parsable = B_TRUE;
break;
case 's':
stop = B_TRUE;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
poolname = argv[0];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
if (stop && noop) {
zpool_close(zhp);
(void) fprintf(stderr, gettext("stop request ignored\n"));
return (0);
}
if (stop) {
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if (zpool_vdev_remove_cancel(zhp) != 0)
ret = 1;
if (wait) {
(void) fprintf(stderr, gettext("invalid option "
"combination: -w cannot be used with -s\n"));
usage(B_FALSE);
}
} else {
if (argc < 2) {
(void) fprintf(stderr, gettext("missing device\n"));
usage(B_FALSE);
}
for (i = 1; i < argc; i++) {
if (noop) {
uint64_t size;
if (zpool_vdev_indirect_size(zhp, argv[i],
&size) != 0) {
ret = 1;
break;
}
if (parsable) {
(void) printf("%s %llu\n",
argv[i], (unsigned long long)size);
} else {
char valstr[32];
zfs_nicenum(size, valstr,
sizeof (valstr));
(void) printf("Memory that will be "
"used after removing %s: %s\n",
argv[i], valstr);
}
} else {
if (zpool_vdev_remove(zhp, argv[i]) != 0)
ret = 1;
}
}
if (ret == 0 && wait)
ret = zpool_wait(zhp, ZPOOL_WAIT_REMOVE);
}
zpool_close(zhp);
return (ret);
}
/*
* Return 1 if a vdev is active (being used in a pool)
* Return 0 if a vdev is inactive (offlined or faulted, or not in active pool)
*
* This is useful for checking if a disk in an active pool is offlined or
* faulted.
*/
static int
vdev_is_active(char *vdev_path)
{
int fd;
fd = open(vdev_path, O_EXCL);
if (fd < 0) {
return (1); /* cant open O_EXCL - disk is active */
}
close(fd);
return (0); /* disk is inactive in the pool */
}
/*
* zpool labelclear [-f] <vdev>
*
* -f Force clearing the label for the vdevs which are members of
* the exported or foreign pools.
*
* Verifies that the vdev is not active and zeros out the label information
* on the device.
*/
int
zpool_do_labelclear(int argc, char **argv)
{
char vdev[MAXPATHLEN];
char *name = NULL;
int c, fd = -1, ret = 0;
nvlist_t *config;
pool_state_t state;
boolean_t inuse = B_FALSE;
boolean_t force = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "f")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
default:
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get vdev name */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing vdev name\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
(void) strlcpy(vdev, argv[0], sizeof (vdev));
/*
* If we cannot open an absolute path, we quit.
* Otherwise if the provided vdev name doesn't point to a file,
* try prepending expected disk paths and partition numbers.
*/
if ((fd = open(vdev, O_RDWR)) < 0) {
int error;
if (vdev[0] == '/') {
(void) fprintf(stderr, gettext("failed to open "
"%s: %s\n"), vdev, strerror(errno));
return (1);
}
error = zfs_resolve_shortname(argv[0], vdev, MAXPATHLEN);
if (error == 0 && zfs_dev_is_whole_disk(vdev)) {
if (zfs_append_partition(vdev, MAXPATHLEN) == -1)
error = ENOENT;
}
if (error || ((fd = open(vdev, O_RDWR)) < 0)) {
if (errno == ENOENT) {
(void) fprintf(stderr, gettext(
"failed to find device %s, try "
"specifying absolute path instead\n"),
argv[0]);
return (1);
}
(void) fprintf(stderr, gettext("failed to open %s:"
" %s\n"), vdev, strerror(errno));
return (1);
}
}
/*
* Flush all dirty pages for the block device. This should not be
* fatal when the device does not support BLKFLSBUF as would be the
* case for a file vdev.
*/
if ((zfs_dev_flush(fd) != 0) && (errno != ENOTTY))
(void) fprintf(stderr, gettext("failed to invalidate "
"cache for %s: %s\n"), vdev, strerror(errno));
if (zpool_read_label(fd, &config, NULL) != 0) {
(void) fprintf(stderr,
gettext("failed to read label from %s\n"), vdev);
ret = 1;
goto errout;
}
nvlist_free(config);
ret = zpool_in_use(g_zfs, fd, &state, &name, &inuse);
if (ret != 0) {
(void) fprintf(stderr,
gettext("failed to check state for %s\n"), vdev);
ret = 1;
goto errout;
}
if (!inuse)
goto wipe_label;
switch (state) {
default:
case POOL_STATE_ACTIVE:
case POOL_STATE_SPARE:
case POOL_STATE_L2CACHE:
/*
* We allow the user to call 'zpool offline -f'
* on an offlined disk in an active pool. We can check if
* the disk is online by calling vdev_is_active().
*/
if (force && !vdev_is_active(vdev))
break;
(void) fprintf(stderr, gettext(
"%s is a member (%s) of pool \"%s\""),
vdev, zpool_pool_state_to_name(state), name);
if (force) {
(void) fprintf(stderr, gettext(
". Offline the disk first to clear its label."));
}
printf("\n");
ret = 1;
goto errout;
case POOL_STATE_EXPORTED:
if (force)
break;
(void) fprintf(stderr, gettext(
"use '-f' to override the following error:\n"
"%s is a member of exported pool \"%s\"\n"),
vdev, name);
ret = 1;
goto errout;
case POOL_STATE_POTENTIALLY_ACTIVE:
if (force)
break;
(void) fprintf(stderr, gettext(
"use '-f' to override the following error:\n"
"%s is a member of potentially active pool \"%s\"\n"),
vdev, name);
ret = 1;
goto errout;
case POOL_STATE_DESTROYED:
/* inuse should never be set for a destroyed pool */
assert(0);
break;
}
wipe_label:
ret = zpool_clear_label(fd);
if (ret != 0) {
(void) fprintf(stderr,
gettext("failed to clear label for %s\n"), vdev);
}
errout:
free(name);
(void) close(fd);
return (ret);
}
/*
* zpool create [-fnd] [-o property=value] ...
* [-O file-system-property=value] ...
* [-R root] [-m mountpoint] <pool> <dev> ...
*
* -f Force creation, even if devices appear in use
* -n Do not create the pool, but display the resulting layout if it
* were to be created.
* -R Create a pool under an alternate root
* -m Set default mountpoint for the root dataset. By default it's
* '/<pool>'
* -o Set property=value.
* -o Set feature@feature=enabled|disabled.
* -d Don't automatically enable all supported pool features
* (individual features can be enabled with -o).
* -O Set fsproperty=value in the pool's root file system
*
* Creates the named pool according to the given vdev specification. The
* bulk of the vdev processing is done in make_root_vdev() in zpool_vdev.c.
* Once we get the nvlist back from make_root_vdev(), we either print out the
* contents (if '-n' was specified), or pass it to libzfs to do the creation.
*/
int
zpool_do_create(int argc, char **argv)
{
boolean_t force = B_FALSE;
boolean_t dryrun = B_FALSE;
boolean_t enable_pool_features = B_TRUE;
int c;
nvlist_t *nvroot = NULL;
char *poolname;
char *tname = NULL;
int ret = 1;
char *altroot = NULL;
char *compat = NULL;
char *mountpoint = NULL;
nvlist_t *fsprops = NULL;
nvlist_t *props = NULL;
char *propval;
/* check options */
while ((c = getopt(argc, argv, ":fndR:m:o:O:t:")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
case 'n':
dryrun = B_TRUE;
break;
case 'd':
enable_pool_features = B_FALSE;
break;
case 'R':
altroot = optarg;
if (add_prop_list(zpool_prop_to_name(
ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE))
goto errout;
if (add_prop_list_default(zpool_prop_to_name(
ZPOOL_PROP_CACHEFILE), "none", &props))
goto errout;
break;
case 'm':
/* Equivalent to -O mountpoint=optarg */
mountpoint = optarg;
break;
case 'o':
if ((propval = strchr(optarg, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for -o option\n"));
goto errout;
}
*propval = '\0';
propval++;
if (add_prop_list(optarg, propval, &props, B_TRUE))
goto errout;
/*
* If the user is creating a pool that doesn't support
* feature flags, don't enable any features.
*/
if (zpool_name_to_prop(optarg) == ZPOOL_PROP_VERSION) {
char *end;
u_longlong_t ver;
ver = strtoull(propval, &end, 10);
if (*end == '\0' &&
ver < SPA_VERSION_FEATURES) {
enable_pool_features = B_FALSE;
}
}
if (zpool_name_to_prop(optarg) == ZPOOL_PROP_ALTROOT)
altroot = propval;
if (zpool_name_to_prop(optarg) ==
ZPOOL_PROP_COMPATIBILITY)
compat = propval;
break;
case 'O':
if ((propval = strchr(optarg, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for -O option\n"));
goto errout;
}
*propval = '\0';
propval++;
/*
* Mountpoints are checked and then added later.
* Uniquely among properties, they can be specified
* more than once, to avoid conflict with -m.
*/
if (0 == strcmp(optarg,
zfs_prop_to_name(ZFS_PROP_MOUNTPOINT))) {
mountpoint = propval;
} else if (add_prop_list(optarg, propval, &fsprops,
B_FALSE)) {
goto errout;
}
break;
case 't':
/*
* Sanity check temporary pool name.
*/
if (strchr(optarg, '/') != NULL) {
(void) fprintf(stderr, gettext("cannot create "
"'%s': invalid character '/' in temporary "
"name\n"), optarg);
(void) fprintf(stderr, gettext("use 'zfs "
"create' to create a dataset\n"));
goto errout;
}
if (add_prop_list(zpool_prop_to_name(
ZPOOL_PROP_TNAME), optarg, &props, B_TRUE))
goto errout;
if (add_prop_list_default(zpool_prop_to_name(
ZPOOL_PROP_CACHEFILE), "none", &props))
goto errout;
tname = optarg;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
goto badusage;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
goto badusage;
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
goto badusage;
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing vdev specification\n"));
goto badusage;
}
poolname = argv[0];
/*
* As a special case, check for use of '/' in the name, and direct the
* user to use 'zfs create' instead.
*/
if (strchr(poolname, '/') != NULL) {
(void) fprintf(stderr, gettext("cannot create '%s': invalid "
"character '/' in pool name\n"), poolname);
(void) fprintf(stderr, gettext("use 'zfs create' to "
"create a dataset\n"));
goto errout;
}
/* pass off to make_root_vdev for bulk processing */
nvroot = make_root_vdev(NULL, props, force, !force, B_FALSE, dryrun,
argc - 1, argv + 1);
if (nvroot == NULL)
goto errout;
/* make_root_vdev() allows 0 toplevel children if there are spares */
if (!zfs_allocatable_devs(nvroot)) {
(void) fprintf(stderr, gettext("invalid vdev "
"specification: at least one toplevel vdev must be "
"specified\n"));
goto errout;
}
if (altroot != NULL && altroot[0] != '/') {
(void) fprintf(stderr, gettext("invalid alternate root '%s': "
"must be an absolute path\n"), altroot);
goto errout;
}
/*
* Check the validity of the mountpoint and direct the user to use the
* '-m' mountpoint option if it looks like its in use.
*/
if (mountpoint == NULL ||
(strcmp(mountpoint, ZFS_MOUNTPOINT_LEGACY) != 0 &&
strcmp(mountpoint, ZFS_MOUNTPOINT_NONE) != 0)) {
char buf[MAXPATHLEN];
DIR *dirp;
if (mountpoint && mountpoint[0] != '/') {
(void) fprintf(stderr, gettext("invalid mountpoint "
"'%s': must be an absolute path, 'legacy', or "
"'none'\n"), mountpoint);
goto errout;
}
if (mountpoint == NULL) {
if (altroot != NULL)
(void) snprintf(buf, sizeof (buf), "%s/%s",
altroot, poolname);
else
(void) snprintf(buf, sizeof (buf), "/%s",
poolname);
} else {
if (altroot != NULL)
(void) snprintf(buf, sizeof (buf), "%s%s",
altroot, mountpoint);
else
(void) snprintf(buf, sizeof (buf), "%s",
mountpoint);
}
if ((dirp = opendir(buf)) == NULL && errno != ENOENT) {
(void) fprintf(stderr, gettext("mountpoint '%s' : "
"%s\n"), buf, strerror(errno));
(void) fprintf(stderr, gettext("use '-m' "
"option to provide a different default\n"));
goto errout;
} else if (dirp) {
int count = 0;
while (count < 3 && readdir(dirp) != NULL)
count++;
(void) closedir(dirp);
if (count > 2) {
(void) fprintf(stderr, gettext("mountpoint "
"'%s' exists and is not empty\n"), buf);
(void) fprintf(stderr, gettext("use '-m' "
"option to provide a "
"different default\n"));
goto errout;
}
}
}
/*
* Now that the mountpoint's validity has been checked, ensure that
* the property is set appropriately prior to creating the pool.
*/
if (mountpoint != NULL) {
ret = add_prop_list(zfs_prop_to_name(ZFS_PROP_MOUNTPOINT),
mountpoint, &fsprops, B_FALSE);
if (ret != 0)
goto errout;
}
ret = 1;
if (dryrun) {
/*
* For a dry run invocation, print out a basic message and run
* through all the vdevs in the list and print out in an
* appropriate hierarchy.
*/
(void) printf(gettext("would create '%s' with the "
"following layout:\n\n"), poolname);
print_vdev_tree(NULL, poolname, nvroot, 0, "", 0);
print_vdev_tree(NULL, "dedup", nvroot, 0,
VDEV_ALLOC_BIAS_DEDUP, 0);
print_vdev_tree(NULL, "special", nvroot, 0,
VDEV_ALLOC_BIAS_SPECIAL, 0);
print_vdev_tree(NULL, "logs", nvroot, 0,
VDEV_ALLOC_BIAS_LOG, 0);
print_cache_list(nvroot, 0);
print_spare_list(nvroot, 0);
ret = 0;
} else {
/*
* Load in feature set.
* Note: if compatibility property not given, we'll have
* NULL, which means 'all features'.
*/
boolean_t requested_features[SPA_FEATURES];
if (zpool_do_load_compat(compat, requested_features) !=
ZPOOL_COMPATIBILITY_OK)
goto errout;
/*
* props contains list of features to enable.
* For each feature:
* - remove it if feature@name=disabled
* - leave it there if feature@name=enabled
* - add it if:
* - enable_pool_features (ie: no '-d' or '-o version')
* - it's supported by the kernel module
* - it's in the requested feature set
* - warn if it's enabled but not in compat
*/
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
char propname[MAXPATHLEN];
const char *propval;
zfeature_info_t *feat = &spa_feature_table[i];
(void) snprintf(propname, sizeof (propname),
"feature@%s", feat->fi_uname);
if (!nvlist_lookup_string(props, propname, &propval)) {
if (strcmp(propval,
ZFS_FEATURE_DISABLED) == 0) {
(void) nvlist_remove_all(props,
propname);
} else if (strcmp(propval,
ZFS_FEATURE_ENABLED) == 0 &&
!requested_features[i]) {
(void) fprintf(stderr, gettext(
"Warning: feature \"%s\" enabled "
"but is not in specified "
"'compatibility' feature set.\n"),
feat->fi_uname);
}
} else if (
enable_pool_features &&
feat->fi_zfs_mod_supported &&
requested_features[i]) {
ret = add_prop_list(propname,
ZFS_FEATURE_ENABLED, &props, B_TRUE);
if (ret != 0)
goto errout;
}
}
ret = 1;
if (zpool_create(g_zfs, poolname,
nvroot, props, fsprops) == 0) {
zfs_handle_t *pool = zfs_open(g_zfs,
tname ? tname : poolname, ZFS_TYPE_FILESYSTEM);
if (pool != NULL) {
if (zfs_mount(pool, NULL, 0) == 0) {
ret = zfs_share(pool, NULL);
zfs_commit_shares(NULL);
}
zfs_close(pool);
}
} else if (libzfs_errno(g_zfs) == EZFS_INVALIDNAME) {
(void) fprintf(stderr, gettext("pool name may have "
"been omitted\n"));
}
}
errout:
nvlist_free(nvroot);
nvlist_free(fsprops);
nvlist_free(props);
return (ret);
badusage:
nvlist_free(fsprops);
nvlist_free(props);
usage(B_FALSE);
return (2);
}
/*
* zpool destroy <pool>
*
* -f Forcefully unmount any datasets
*
* Destroy the given pool. Automatically unmounts any datasets in the pool.
*/
int
zpool_do_destroy(int argc, char **argv)
{
boolean_t force = B_FALSE;
int c;
char *pool;
zpool_handle_t *zhp;
int ret;
/* check options */
while ((c = getopt(argc, argv, "f")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* check arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
pool = argv[0];
if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) {
/*
* As a special case, check for use of '/' in the name, and
* direct the user to use 'zfs destroy' instead.
*/
if (strchr(pool, '/') != NULL)
(void) fprintf(stderr, gettext("use 'zfs destroy' to "
"destroy a dataset\n"));
return (1);
}
if (zpool_disable_datasets(zhp, force) != 0) {
(void) fprintf(stderr, gettext("could not destroy '%s': "
"could not unmount datasets\n"), zpool_get_name(zhp));
zpool_close(zhp);
return (1);
}
/* The history must be logged as part of the export */
log_history = B_FALSE;
ret = (zpool_destroy(zhp, history_str) != 0);
zpool_close(zhp);
return (ret);
}
typedef struct export_cbdata {
boolean_t force;
boolean_t hardforce;
} export_cbdata_t;
/*
* Export one pool
*/
static int
zpool_export_one(zpool_handle_t *zhp, void *data)
{
export_cbdata_t *cb = data;
if (zpool_disable_datasets(zhp, cb->force) != 0)
return (1);
/* The history must be logged as part of the export */
log_history = B_FALSE;
if (cb->hardforce) {
if (zpool_export_force(zhp, history_str) != 0)
return (1);
} else if (zpool_export(zhp, cb->force, history_str) != 0) {
return (1);
}
return (0);
}
/*
* zpool export [-f] <pool> ...
*
* -a Export all pools
* -f Forcefully unmount datasets
*
* Export the given pools. By default, the command will attempt to cleanly
* unmount any active datasets within the pool. If the '-f' flag is specified,
* then the datasets will be forcefully unmounted.
*/
int
zpool_do_export(int argc, char **argv)
{
export_cbdata_t cb;
boolean_t do_all = B_FALSE;
boolean_t force = B_FALSE;
boolean_t hardforce = B_FALSE;
int c, ret;
/* check options */
while ((c = getopt(argc, argv, "afF")) != -1) {
switch (c) {
case 'a':
do_all = B_TRUE;
break;
case 'f':
force = B_TRUE;
break;
case 'F':
hardforce = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
cb.force = force;
cb.hardforce = hardforce;
argc -= optind;
argv += optind;
if (do_all) {
if (argc != 0) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
return (for_each_pool(argc, argv, B_TRUE, NULL,
ZFS_TYPE_POOL, B_FALSE, zpool_export_one, &cb));
}
/* check arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool argument\n"));
usage(B_FALSE);
}
ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, zpool_export_one, &cb);
return (ret);
}
/*
* Given a vdev configuration, determine the maximum width needed for the device
* name column.
*/
static int
max_width(zpool_handle_t *zhp, nvlist_t *nv, int depth, int max,
int name_flags)
{
static const char *const subtypes[] =
{ZPOOL_CONFIG_SPARES, ZPOOL_CONFIG_L2CACHE, ZPOOL_CONFIG_CHILDREN};
char *name = zpool_vdev_name(g_zfs, zhp, nv, name_flags);
max = MAX(strlen(name) + depth, max);
free(name);
nvlist_t **child;
uint_t children;
for (size_t i = 0; i < ARRAY_SIZE(subtypes); ++i)
if (nvlist_lookup_nvlist_array(nv, subtypes[i],
&child, &children) == 0)
for (uint_t c = 0; c < children; ++c)
max = MAX(max_width(zhp, child[c], depth + 2,
max, name_flags), max);
return (max);
}
typedef struct spare_cbdata {
uint64_t cb_guid;
zpool_handle_t *cb_zhp;
} spare_cbdata_t;
static boolean_t
find_vdev(nvlist_t *nv, uint64_t search)
{
uint64_t guid;
nvlist_t **child;
uint_t c, children;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0 &&
search == guid)
return (B_TRUE);
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0) {
for (c = 0; c < children; c++)
if (find_vdev(child[c], search))
return (B_TRUE);
}
return (B_FALSE);
}
static int
find_spare(zpool_handle_t *zhp, void *data)
{
spare_cbdata_t *cbp = data;
nvlist_t *config, *nvroot;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
if (find_vdev(nvroot, cbp->cb_guid)) {
cbp->cb_zhp = zhp;
return (1);
}
zpool_close(zhp);
return (0);
}
typedef struct status_cbdata {
int cb_count;
int cb_name_flags;
int cb_namewidth;
boolean_t cb_allpools;
boolean_t cb_verbose;
boolean_t cb_literal;
boolean_t cb_explain;
boolean_t cb_first;
boolean_t cb_dedup_stats;
boolean_t cb_print_status;
boolean_t cb_print_slow_ios;
boolean_t cb_print_vdev_init;
boolean_t cb_print_vdev_trim;
vdev_cmd_data_list_t *vcdl;
} status_cbdata_t;
/* Return 1 if string is NULL, empty, or whitespace; return 0 otherwise. */
static boolean_t
is_blank_str(const char *str)
{
for (; str != NULL && *str != '\0'; ++str)
if (!isblank(*str))
return (B_FALSE);
return (B_TRUE);
}
/* Print command output lines for specific vdev in a specific pool */
static void
zpool_print_cmd(vdev_cmd_data_list_t *vcdl, const char *pool, const char *path)
{
vdev_cmd_data_t *data;
int i, j;
const char *val;
for (i = 0; i < vcdl->count; i++) {
if ((strcmp(vcdl->data[i].path, path) != 0) ||
(strcmp(vcdl->data[i].pool, pool) != 0)) {
/* Not the vdev we're looking for */
continue;
}
data = &vcdl->data[i];
/* Print out all the output values for this vdev */
for (j = 0; j < vcdl->uniq_cols_cnt; j++) {
val = NULL;
/* Does this vdev have values for this column? */
for (int k = 0; k < data->cols_cnt; k++) {
if (strcmp(data->cols[k],
vcdl->uniq_cols[j]) == 0) {
/* yes it does, record the value */
val = data->lines[k];
break;
}
}
/*
* Mark empty values with dashes to make output
* awk-able.
*/
if (val == NULL || is_blank_str(val))
val = "-";
printf("%*s", vcdl->uniq_cols_width[j], val);
if (j < vcdl->uniq_cols_cnt - 1)
fputs(" ", stdout);
}
/* Print out any values that aren't in a column at the end */
for (j = data->cols_cnt; j < data->lines_cnt; j++) {
/* Did we have any columns? If so print a spacer. */
if (vcdl->uniq_cols_cnt > 0)
fputs(" ", stdout);
val = data->lines[j];
fputs(val ?: "", stdout);
}
break;
}
}
/*
* Print vdev initialization status for leaves
*/
static void
print_status_initialize(vdev_stat_t *vs, boolean_t verbose)
{
if (verbose) {
if ((vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE ||
vs->vs_initialize_state == VDEV_INITIALIZE_SUSPENDED ||
vs->vs_initialize_state == VDEV_INITIALIZE_COMPLETE) &&
!vs->vs_scan_removing) {
char zbuf[1024];
char tbuf[256];
struct tm zaction_ts;
time_t t = vs->vs_initialize_action_time;
int initialize_pct = 100;
if (vs->vs_initialize_state !=
VDEV_INITIALIZE_COMPLETE) {
initialize_pct = (vs->vs_initialize_bytes_done *
100 / (vs->vs_initialize_bytes_est + 1));
}
(void) localtime_r(&t, &zaction_ts);
(void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts);
switch (vs->vs_initialize_state) {
case VDEV_INITIALIZE_SUSPENDED:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("suspended, started at"), tbuf);
break;
case VDEV_INITIALIZE_ACTIVE:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("started at"), tbuf);
break;
case VDEV_INITIALIZE_COMPLETE:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("completed at"), tbuf);
break;
}
(void) printf(gettext(" (%d%% initialized%s)"),
initialize_pct, zbuf);
} else {
(void) printf(gettext(" (uninitialized)"));
}
} else if (vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE) {
(void) printf(gettext(" (initializing)"));
}
}
/*
* Print vdev TRIM status for leaves
*/
static void
print_status_trim(vdev_stat_t *vs, boolean_t verbose)
{
if (verbose) {
if ((vs->vs_trim_state == VDEV_TRIM_ACTIVE ||
vs->vs_trim_state == VDEV_TRIM_SUSPENDED ||
vs->vs_trim_state == VDEV_TRIM_COMPLETE) &&
!vs->vs_scan_removing) {
char zbuf[1024];
char tbuf[256];
struct tm zaction_ts;
time_t t = vs->vs_trim_action_time;
int trim_pct = 100;
if (vs->vs_trim_state != VDEV_TRIM_COMPLETE) {
trim_pct = (vs->vs_trim_bytes_done *
100 / (vs->vs_trim_bytes_est + 1));
}
(void) localtime_r(&t, &zaction_ts);
(void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts);
switch (vs->vs_trim_state) {
case VDEV_TRIM_SUSPENDED:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("suspended, started at"), tbuf);
break;
case VDEV_TRIM_ACTIVE:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("started at"), tbuf);
break;
case VDEV_TRIM_COMPLETE:
(void) snprintf(zbuf, sizeof (zbuf), ", %s %s",
gettext("completed at"), tbuf);
break;
}
(void) printf(gettext(" (%d%% trimmed%s)"),
trim_pct, zbuf);
} else if (vs->vs_trim_notsup) {
(void) printf(gettext(" (trim unsupported)"));
} else {
(void) printf(gettext(" (untrimmed)"));
}
} else if (vs->vs_trim_state == VDEV_TRIM_ACTIVE) {
(void) printf(gettext(" (trimming)"));
}
}
/*
* Return the color associated with a health string. This includes returning
* NULL for no color change.
*/
static const char *
health_str_to_color(const char *health)
{
if (strcmp(health, gettext("FAULTED")) == 0 ||
strcmp(health, gettext("SUSPENDED")) == 0 ||
strcmp(health, gettext("UNAVAIL")) == 0) {
return (ANSI_RED);
}
if (strcmp(health, gettext("OFFLINE")) == 0 ||
strcmp(health, gettext("DEGRADED")) == 0 ||
strcmp(health, gettext("REMOVED")) == 0) {
return (ANSI_YELLOW);
}
return (NULL);
}
/*
* Print out configuration state as requested by status_callback.
*/
static void
print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name,
nvlist_t *nv, int depth, boolean_t isspare, vdev_rebuild_stat_t *vrs)
{
nvlist_t **child, *root;
uint_t c, i, vsc, children;
pool_scan_stat_t *ps = NULL;
vdev_stat_t *vs;
char rbuf[6], wbuf[6], cbuf[6];
char *vname;
uint64_t notpresent;
spare_cbdata_t spare_cb;
const char *state;
const char *type;
const char *path = NULL;
const char *rcolor = NULL, *wcolor = NULL, *ccolor = NULL;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &vsc) == 0);
verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0);
if (strcmp(type, VDEV_TYPE_INDIRECT) == 0)
return;
state = zpool_state_to_name(vs->vs_state, vs->vs_aux);
if (isspare) {
/*
* For hot spares, we use the terms 'INUSE' and 'AVAILABLE' for
* online drives.
*/
if (vs->vs_aux == VDEV_AUX_SPARED)
state = gettext("INUSE");
else if (vs->vs_state == VDEV_STATE_HEALTHY)
state = gettext("AVAIL");
}
printf_color(health_str_to_color(state),
"\t%*s%-*s %-8s", depth, "", cb->cb_namewidth - depth,
name, state);
if (!isspare) {
if (vs->vs_read_errors)
rcolor = ANSI_RED;
if (vs->vs_write_errors)
wcolor = ANSI_RED;
if (vs->vs_checksum_errors)
ccolor = ANSI_RED;
if (cb->cb_literal) {
fputc(' ', stdout);
printf_color(rcolor, "%5llu",
(u_longlong_t)vs->vs_read_errors);
fputc(' ', stdout);
printf_color(wcolor, "%5llu",
(u_longlong_t)vs->vs_write_errors);
fputc(' ', stdout);
printf_color(ccolor, "%5llu",
(u_longlong_t)vs->vs_checksum_errors);
} else {
zfs_nicenum(vs->vs_read_errors, rbuf, sizeof (rbuf));
zfs_nicenum(vs->vs_write_errors, wbuf, sizeof (wbuf));
zfs_nicenum(vs->vs_checksum_errors, cbuf,
sizeof (cbuf));
fputc(' ', stdout);
printf_color(rcolor, "%5s", rbuf);
fputc(' ', stdout);
printf_color(wcolor, "%5s", wbuf);
fputc(' ', stdout);
printf_color(ccolor, "%5s", cbuf);
}
if (cb->cb_print_slow_ios) {
if (children == 0) {
/* Only leafs vdevs have slow IOs */
zfs_nicenum(vs->vs_slow_ios, rbuf,
sizeof (rbuf));
} else {
snprintf(rbuf, sizeof (rbuf), "-");
}
if (cb->cb_literal)
printf(" %5llu", (u_longlong_t)vs->vs_slow_ios);
else
printf(" %5s", rbuf);
}
}
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT,
&notpresent) == 0) {
verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0);
(void) printf(" %s %s", gettext("was"), path);
} else if (vs->vs_aux != 0) {
(void) printf(" ");
color_start(ANSI_RED);
switch (vs->vs_aux) {
case VDEV_AUX_OPEN_FAILED:
(void) printf(gettext("cannot open"));
break;
case VDEV_AUX_BAD_GUID_SUM:
(void) printf(gettext("missing device"));
break;
case VDEV_AUX_NO_REPLICAS:
(void) printf(gettext("insufficient replicas"));
break;
case VDEV_AUX_VERSION_NEWER:
(void) printf(gettext("newer version"));
break;
case VDEV_AUX_UNSUP_FEAT:
(void) printf(gettext("unsupported feature(s)"));
break;
case VDEV_AUX_ASHIFT_TOO_BIG:
(void) printf(gettext("unsupported minimum blocksize"));
break;
case VDEV_AUX_SPARED:
verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID,
&spare_cb.cb_guid) == 0);
if (zpool_iter(g_zfs, find_spare, &spare_cb) == 1) {
if (strcmp(zpool_get_name(spare_cb.cb_zhp),
zpool_get_name(zhp)) == 0)
(void) printf(gettext("currently in "
"use"));
else
(void) printf(gettext("in use by "
"pool '%s'"),
zpool_get_name(spare_cb.cb_zhp));
zpool_close(spare_cb.cb_zhp);
} else {
(void) printf(gettext("currently in use"));
}
break;
case VDEV_AUX_ERR_EXCEEDED:
(void) printf(gettext("too many errors"));
break;
case VDEV_AUX_IO_FAILURE:
(void) printf(gettext("experienced I/O failures"));
break;
case VDEV_AUX_BAD_LOG:
(void) printf(gettext("bad intent log"));
break;
case VDEV_AUX_EXTERNAL:
(void) printf(gettext("external device fault"));
break;
case VDEV_AUX_SPLIT_POOL:
(void) printf(gettext("split into new pool"));
break;
case VDEV_AUX_ACTIVE:
(void) printf(gettext("currently in use"));
break;
case VDEV_AUX_CHILDREN_OFFLINE:
(void) printf(gettext("all children offline"));
break;
case VDEV_AUX_BAD_LABEL:
(void) printf(gettext("invalid label"));
break;
default:
(void) printf(gettext("corrupted data"));
break;
}
color_end();
} else if (children == 0 && !isspare &&
getenv("ZPOOL_STATUS_NON_NATIVE_ASHIFT_IGNORE") == NULL &&
VDEV_STAT_VALID(vs_physical_ashift, vsc) &&
vs->vs_configured_ashift < vs->vs_physical_ashift) {
(void) printf(
gettext(" block size: %dB configured, %dB native"),
1 << vs->vs_configured_ashift, 1 << vs->vs_physical_ashift);
}
if (vs->vs_scan_removing != 0) {
(void) printf(gettext(" (removing)"));
} else if (VDEV_STAT_VALID(vs_noalloc, vsc) && vs->vs_noalloc != 0) {
(void) printf(gettext(" (non-allocating)"));
}
/* The root vdev has the scrub/resilver stats */
root = fnvlist_lookup_nvlist(zpool_get_config(zhp, NULL),
ZPOOL_CONFIG_VDEV_TREE);
(void) nvlist_lookup_uint64_array(root, ZPOOL_CONFIG_SCAN_STATS,
(uint64_t **)&ps, &c);
/*
* If you force fault a drive that's resilvering, its scan stats can
* get frozen in time, giving the false impression that it's
* being resilvered. That's why we check the state to see if the vdev
* is healthy before reporting "resilvering" or "repairing".
*/
if (ps != NULL && ps->pss_state == DSS_SCANNING && children == 0 &&
vs->vs_state == VDEV_STATE_HEALTHY) {
if (vs->vs_scan_processed != 0) {
(void) printf(gettext(" (%s)"),
(ps->pss_func == POOL_SCAN_RESILVER) ?
"resilvering" : "repairing");
} else if (vs->vs_resilver_deferred) {
(void) printf(gettext(" (awaiting resilver)"));
}
}
/* The top-level vdevs have the rebuild stats */
if (vrs != NULL && vrs->vrs_state == VDEV_REBUILD_ACTIVE &&
children == 0 && vs->vs_state == VDEV_STATE_HEALTHY) {
if (vs->vs_rebuild_processed != 0) {
(void) printf(gettext(" (resilvering)"));
}
}
if (cb->vcdl != NULL) {
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0) {
printf(" ");
zpool_print_cmd(cb->vcdl, zpool_get_name(zhp), path);
}
}
/* Display vdev initialization and trim status for leaves. */
if (children == 0) {
print_status_initialize(vs, cb->cb_print_vdev_init);
print_status_trim(vs, cb->cb_print_vdev_trim);
}
(void) printf("\n");
for (c = 0; c < children; c++) {
uint64_t islog = B_FALSE, ishole = B_FALSE;
/* Don't print logs or holes here */
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&islog);
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE,
&ishole);
if (islog || ishole)
continue;
/* Only print normal classes here */
if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS))
continue;
/* Provide vdev_rebuild_stats to children if available */
if (vrs == NULL) {
(void) nvlist_lookup_uint64_array(nv,
ZPOOL_CONFIG_REBUILD_STATS,
(uint64_t **)&vrs, &i);
}
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
print_status_config(zhp, cb, vname, child[c], depth + 2,
isspare, vrs);
free(vname);
}
}
/*
* Print the configuration of an exported pool. Iterate over all vdevs in the
* pool, printing out the name and status for each one.
*/
static void
print_import_config(status_cbdata_t *cb, const char *name, nvlist_t *nv,
int depth)
{
nvlist_t **child;
uint_t c, children;
vdev_stat_t *vs;
const char *type;
char *vname;
verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0);
if (strcmp(type, VDEV_TYPE_MISSING) == 0 ||
strcmp(type, VDEV_TYPE_HOLE) == 0)
return;
verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &c) == 0);
(void) printf("\t%*s%-*s", depth, "", cb->cb_namewidth - depth, name);
(void) printf(" %s", zpool_state_to_name(vs->vs_state, vs->vs_aux));
if (vs->vs_aux != 0) {
(void) printf(" ");
switch (vs->vs_aux) {
case VDEV_AUX_OPEN_FAILED:
(void) printf(gettext("cannot open"));
break;
case VDEV_AUX_BAD_GUID_SUM:
(void) printf(gettext("missing device"));
break;
case VDEV_AUX_NO_REPLICAS:
(void) printf(gettext("insufficient replicas"));
break;
case VDEV_AUX_VERSION_NEWER:
(void) printf(gettext("newer version"));
break;
case VDEV_AUX_UNSUP_FEAT:
(void) printf(gettext("unsupported feature(s)"));
break;
case VDEV_AUX_ERR_EXCEEDED:
(void) printf(gettext("too many errors"));
break;
case VDEV_AUX_ACTIVE:
(void) printf(gettext("currently in use"));
break;
case VDEV_AUX_CHILDREN_OFFLINE:
(void) printf(gettext("all children offline"));
break;
case VDEV_AUX_BAD_LABEL:
(void) printf(gettext("invalid label"));
break;
default:
(void) printf(gettext("corrupted data"));
break;
}
}
(void) printf("\n");
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
return;
for (c = 0; c < children; c++) {
uint64_t is_log = B_FALSE;
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&is_log);
if (is_log)
continue;
if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS))
continue;
vname = zpool_vdev_name(g_zfs, NULL, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
print_import_config(cb, vname, child[c], depth + 2);
free(vname);
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE,
&child, &children) == 0) {
(void) printf(gettext("\tcache\n"));
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, NULL, child[c],
cb->cb_name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
&child, &children) == 0) {
(void) printf(gettext("\tspares\n"));
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, NULL, child[c],
cb->cb_name_flags);
(void) printf("\t %s\n", vname);
free(vname);
}
}
}
/*
* Print specialized class vdevs.
*
* These are recorded as top level vdevs in the main pool child array
* but with "is_log" set to 1 or an "alloc_bias" string. We use either
* print_status_config() or print_import_config() to print the top level
* class vdevs then any of their children (eg mirrored slogs) are printed
* recursively - which works because only the top level vdev is marked.
*/
static void
print_class_vdevs(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t *nv,
const char *class)
{
uint_t c, children;
nvlist_t **child;
boolean_t printed = B_FALSE;
assert(zhp != NULL || !cb->cb_verbose);
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child,
&children) != 0)
return;
for (c = 0; c < children; c++) {
uint64_t is_log = B_FALSE;
const char *bias = NULL;
const char *type = NULL;
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&is_log);
if (is_log) {
bias = (char *)VDEV_ALLOC_CLASS_LOGS;
} else {
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias);
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_TYPE, &type);
}
if (bias == NULL || strcmp(bias, class) != 0)
continue;
if (!is_log && strcmp(type, VDEV_TYPE_INDIRECT) == 0)
continue;
if (!printed) {
(void) printf("\t%s\t\n", gettext(class));
printed = B_TRUE;
}
char *name = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
if (cb->cb_print_status)
print_status_config(zhp, cb, name, child[c], 2,
B_FALSE, NULL);
else
print_import_config(cb, name, child[c], 2);
free(name);
}
}
/*
* Display the status for the given pool.
*/
static int
show_import(nvlist_t *config, boolean_t report_error)
{
uint64_t pool_state;
vdev_stat_t *vs;
const char *name;
uint64_t guid;
uint64_t hostid = 0;
const char *msgid;
const char *hostname = "unknown";
nvlist_t *nvroot, *nvinfo;
zpool_status_t reason;
zpool_errata_t errata;
const char *health;
uint_t vsc;
const char *comment;
status_cbdata_t cb = { 0 };
verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
&name) == 0);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
&guid) == 0);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
&pool_state) == 0);
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &vsc) == 0);
health = zpool_state_to_name(vs->vs_state, vs->vs_aux);
reason = zpool_import_status(config, &msgid, &errata);
/*
* If we're importing using a cachefile, then we won't report any
* errors unless we are in the scan phase of the import.
*/
if (reason != ZPOOL_STATUS_OK && !report_error)
return (reason);
(void) printf(gettext(" pool: %s\n"), name);
(void) printf(gettext(" id: %llu\n"), (u_longlong_t)guid);
(void) printf(gettext(" state: %s"), health);
if (pool_state == POOL_STATE_DESTROYED)
(void) printf(gettext(" (DESTROYED)"));
(void) printf("\n");
switch (reason) {
case ZPOOL_STATUS_MISSING_DEV_R:
case ZPOOL_STATUS_MISSING_DEV_NR:
case ZPOOL_STATUS_BAD_GUID_SUM:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"missing from the system.\n"));
break;
case ZPOOL_STATUS_CORRUPT_LABEL_R:
case ZPOOL_STATUS_CORRUPT_LABEL_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices contains"
" corrupted data.\n"));
break;
case ZPOOL_STATUS_CORRUPT_DATA:
(void) printf(
gettext(" status: The pool data is corrupted.\n"));
break;
case ZPOOL_STATUS_OFFLINE_DEV:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices "
"are offlined.\n"));
break;
case ZPOOL_STATUS_CORRUPT_POOL:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool metadata is "
"corrupted.\n"));
break;
case ZPOOL_STATUS_VERSION_OLDER:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is formatted using "
"a legacy on-disk version.\n"));
break;
case ZPOOL_STATUS_VERSION_NEWER:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is formatted using "
"an incompatible version.\n"));
break;
case ZPOOL_STATUS_FEAT_DISABLED:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Some supported "
"features are not enabled on the pool.\n\t"
"(Note that they may be intentionally disabled "
"if the\n\t'compatibility' property is set.)\n"));
break;
case ZPOOL_STATUS_COMPATIBILITY_ERR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Error reading or parsing "
"the file(s) indicated by the 'compatibility'\n"
"property.\n"));
break;
case ZPOOL_STATUS_INCOMPATIBLE_FEAT:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more features "
"are enabled on the pool despite not being\n"
"requested by the 'compatibility' property.\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_READ:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool uses the following "
"feature(s) not supported on this system:\n"));
color_start(ANSI_YELLOW);
zpool_print_unsup_feat(config);
color_end();
break;
case ZPOOL_STATUS_UNSUP_FEAT_WRITE:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool can only be "
"accessed in read-only mode on this system. It\n\tcannot be"
" accessed in read-write mode because it uses the "
"following\n\tfeature(s) not supported on this system:\n"));
color_start(ANSI_YELLOW);
zpool_print_unsup_feat(config);
color_end();
break;
case ZPOOL_STATUS_HOSTID_ACTIVE:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is currently "
"imported by another system.\n"));
break;
case ZPOOL_STATUS_HOSTID_REQUIRED:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool has the "
"multihost property on. It cannot\n\tbe safely imported "
"when the system hostid is not set.\n"));
break;
case ZPOOL_STATUS_HOSTID_MISMATCH:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool was last accessed "
"by another system.\n"));
break;
case ZPOOL_STATUS_FAULTED_DEV_R:
case ZPOOL_STATUS_FAULTED_DEV_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"faulted.\n"));
break;
case ZPOOL_STATUS_BAD_LOG:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("An intent log record cannot "
"be read.\n"));
break;
case ZPOOL_STATUS_RESILVERING:
case ZPOOL_STATUS_REBUILDING:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices were "
"being resilvered.\n"));
break;
case ZPOOL_STATUS_ERRATA:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Errata #%d detected.\n"),
errata);
break;
case ZPOOL_STATUS_NON_NATIVE_ASHIFT:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"configured to use a non-native block size.\n"
"\tExpect reduced performance.\n"));
break;
default:
/*
* No other status can be seen when importing pools.
*/
assert(reason == ZPOOL_STATUS_OK);
}
/*
* Print out an action according to the overall state of the pool.
*/
if (vs->vs_state == VDEV_STATE_HEALTHY) {
if (reason == ZPOOL_STATUS_VERSION_OLDER ||
reason == ZPOOL_STATUS_FEAT_DISABLED) {
(void) printf(gettext(" action: The pool can be "
"imported using its name or numeric identifier, "
"though\n\tsome features will not be available "
"without an explicit 'zpool upgrade'.\n"));
} else if (reason == ZPOOL_STATUS_COMPATIBILITY_ERR) {
(void) printf(gettext(" action: The pool can be "
"imported using its name or numeric\n\tidentifier, "
"though the file(s) indicated by its "
"'compatibility'\n\tproperty cannot be parsed at "
"this time.\n"));
} else if (reason == ZPOOL_STATUS_HOSTID_MISMATCH) {
(void) printf(gettext(" action: The pool can be "
"imported using its name or numeric "
"identifier and\n\tthe '-f' flag.\n"));
} else if (reason == ZPOOL_STATUS_ERRATA) {
switch (errata) {
case ZPOOL_ERRATA_NONE:
break;
case ZPOOL_ERRATA_ZOL_2094_SCRUB:
(void) printf(gettext(" action: The pool can "
"be imported using its name or numeric "
"identifier,\n\thowever there is a compat"
"ibility issue which should be corrected"
"\n\tby running 'zpool scrub'\n"));
break;
case ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY:
(void) printf(gettext(" action: The pool can"
"not be imported with this version of ZFS "
"due to\n\tan active asynchronous destroy. "
"Revert to an earlier version\n\tand "
"allow the destroy to complete before "
"updating.\n"));
break;
case ZPOOL_ERRATA_ZOL_6845_ENCRYPTION:
(void) printf(gettext(" action: Existing "
"encrypted datasets contain an on-disk "
"incompatibility, which\n\tneeds to be "
"corrected. Backup these datasets to new "
"encrypted datasets\n\tand destroy the "
"old ones.\n"));
break;
case ZPOOL_ERRATA_ZOL_8308_ENCRYPTION:
(void) printf(gettext(" action: Existing "
"encrypted snapshots and bookmarks contain "
"an on-disk\n\tincompatibility. This may "
"cause on-disk corruption if they are used"
"\n\twith 'zfs recv'. To correct the "
"issue, enable the bookmark_v2 feature.\n\t"
"No additional action is needed if there "
"are no encrypted snapshots or\n\t"
"bookmarks. If preserving the encrypted "
"snapshots and bookmarks is\n\trequired, "
"use a non-raw send to backup and restore "
"them. Alternately,\n\tthey may be removed"
" to resolve the incompatibility.\n"));
break;
default:
/*
* All errata must contain an action message.
*/
assert(0);
}
} else {
(void) printf(gettext(" action: The pool can be "
"imported using its name or numeric "
"identifier.\n"));
}
} else if (vs->vs_state == VDEV_STATE_DEGRADED) {
(void) printf(gettext(" action: The pool can be imported "
"despite missing or damaged devices. The\n\tfault "
"tolerance of the pool may be compromised if imported.\n"));
} else {
switch (reason) {
case ZPOOL_STATUS_VERSION_NEWER:
(void) printf(gettext(" action: The pool cannot be "
"imported. Access the pool on a system running "
"newer\n\tsoftware, or recreate the pool from "
"backup.\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_READ:
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("The pool cannot be "
"imported. Access the pool on a system that "
"supports\n\tthe required feature(s), or recreate "
"the pool from backup.\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_WRITE:
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("The pool cannot be "
"imported in read-write mode. Import the pool "
"with\n"
"\t\"-o readonly=on\", access the pool on a system "
"that supports the\n\trequired feature(s), or "
"recreate the pool from backup.\n"));
break;
case ZPOOL_STATUS_MISSING_DEV_R:
case ZPOOL_STATUS_MISSING_DEV_NR:
case ZPOOL_STATUS_BAD_GUID_SUM:
(void) printf(gettext(" action: The pool cannot be "
"imported. Attach the missing\n\tdevices and try "
"again.\n"));
break;
case ZPOOL_STATUS_HOSTID_ACTIVE:
VERIFY0(nvlist_lookup_nvlist(config,
ZPOOL_CONFIG_LOAD_INFO, &nvinfo));
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME))
hostname = fnvlist_lookup_string(nvinfo,
ZPOOL_CONFIG_MMP_HOSTNAME);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID))
hostid = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_HOSTID);
(void) printf(gettext(" action: The pool must be "
"exported from %s (hostid=%"PRIx64")\n\tbefore it "
"can be safely imported.\n"), hostname, hostid);
break;
case ZPOOL_STATUS_HOSTID_REQUIRED:
(void) printf(gettext(" action: Set a unique system "
"hostid with the zgenhostid(8) command.\n"));
break;
default:
(void) printf(gettext(" action: The pool cannot be "
"imported due to damaged devices or data.\n"));
}
}
/* Print the comment attached to the pool. */
if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMMENT, &comment) == 0)
(void) printf(gettext("comment: %s\n"), comment);
/*
* If the state is "closed" or "can't open", and the aux state
* is "corrupt data":
*/
if (((vs->vs_state == VDEV_STATE_CLOSED) ||
(vs->vs_state == VDEV_STATE_CANT_OPEN)) &&
(vs->vs_aux == VDEV_AUX_CORRUPT_DATA)) {
if (pool_state == POOL_STATE_DESTROYED)
(void) printf(gettext("\tThe pool was destroyed, "
"but can be imported using the '-Df' flags.\n"));
else if (pool_state != POOL_STATE_EXPORTED)
(void) printf(gettext("\tThe pool may be active on "
"another system, but can be imported using\n\t"
"the '-f' flag.\n"));
}
if (msgid != NULL) {
(void) printf(gettext(
" see: https://openzfs.github.io/openzfs-docs/msg/%s\n"),
msgid);
}
(void) printf(gettext(" config:\n\n"));
cb.cb_namewidth = max_width(NULL, nvroot, 0, strlen(name),
VDEV_NAME_TYPE_ID);
if (cb.cb_namewidth < 10)
cb.cb_namewidth = 10;
print_import_config(&cb, name, nvroot, 0);
print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_BIAS_DEDUP);
print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_BIAS_SPECIAL);
print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_CLASS_LOGS);
if (reason == ZPOOL_STATUS_BAD_GUID_SUM) {
(void) printf(gettext("\n\tAdditional devices are known to "
"be part of this pool, though their\n\texact "
"configuration cannot be determined.\n"));
}
return (0);
}
static boolean_t
zfs_force_import_required(nvlist_t *config)
{
uint64_t state;
uint64_t hostid = 0;
nvlist_t *nvinfo;
state = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE);
(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid);
if (state != POOL_STATE_EXPORTED && hostid != get_system_hostid())
return (B_TRUE);
nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_STATE)) {
mmp_state_t mmp_state = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_STATE);
if (mmp_state != MMP_STATE_INACTIVE)
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Perform the import for the given configuration. This passes the heavy
* lifting off to zpool_import_props(), and then mounts the datasets contained
* within the pool.
*/
static int
do_import(nvlist_t *config, const char *newname, const char *mntopts,
nvlist_t *props, int flags)
{
int ret = 0;
zpool_handle_t *zhp;
const char *name;
uint64_t version;
name = fnvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME);
version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION);
if (!SPA_VERSION_IS_SUPPORTED(version)) {
(void) fprintf(stderr, gettext("cannot import '%s': pool "
"is formatted using an unsupported ZFS version\n"), name);
return (1);
} else if (zfs_force_import_required(config) &&
!(flags & ZFS_IMPORT_ANY_HOST)) {
mmp_state_t mmp_state = MMP_STATE_INACTIVE;
nvlist_t *nvinfo;
nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_STATE))
mmp_state = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_STATE);
if (mmp_state == MMP_STATE_ACTIVE) {
const char *hostname = "<unknown>";
uint64_t hostid = 0;
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME))
hostname = fnvlist_lookup_string(nvinfo,
ZPOOL_CONFIG_MMP_HOSTNAME);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID))
hostid = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_HOSTID);
(void) fprintf(stderr, gettext("cannot import '%s': "
"pool is imported on %s (hostid: "
"0x%"PRIx64")\nExport the pool on the other "
"system, then run 'zpool import'.\n"),
name, hostname, hostid);
} else if (mmp_state == MMP_STATE_NO_HOSTID) {
(void) fprintf(stderr, gettext("Cannot import '%s': "
"pool has the multihost property on and the\n"
"system's hostid is not set. Set a unique hostid "
"with the zgenhostid(8) command.\n"), name);
} else {
const char *hostname = "<unknown>";
time_t timestamp = 0;
uint64_t hostid = 0;
if (nvlist_exists(config, ZPOOL_CONFIG_HOSTNAME))
hostname = fnvlist_lookup_string(config,
ZPOOL_CONFIG_HOSTNAME);
if (nvlist_exists(config, ZPOOL_CONFIG_TIMESTAMP))
timestamp = fnvlist_lookup_uint64(config,
ZPOOL_CONFIG_TIMESTAMP);
if (nvlist_exists(config, ZPOOL_CONFIG_HOSTID))
hostid = fnvlist_lookup_uint64(config,
ZPOOL_CONFIG_HOSTID);
(void) fprintf(stderr, gettext("cannot import '%s': "
"pool was previously in use from another system.\n"
"Last accessed by %s (hostid=%"PRIx64") at %s"
"The pool can be imported, use 'zpool import -f' "
"to import the pool.\n"), name, hostname,
hostid, ctime(&timestamp));
}
return (1);
}
if (zpool_import_props(g_zfs, config, newname, props, flags) != 0)
return (1);
if (newname != NULL)
name = newname;
if ((zhp = zpool_open_canfail(g_zfs, name)) == NULL)
return (1);
/*
* Loading keys is best effort. We don't want to return immediately
* if it fails but we do want to give the error to the caller.
*/
if (flags & ZFS_IMPORT_LOAD_KEYS &&
zfs_crypto_attempt_load_keys(g_zfs, name) != 0)
ret = 1;
if (zpool_get_state(zhp) != POOL_STATE_UNAVAIL &&
!(flags & ZFS_IMPORT_ONLY) &&
zpool_enable_datasets(zhp, mntopts, 0) != 0) {
zpool_close(zhp);
return (1);
}
zpool_close(zhp);
return (ret);
}
static int
import_pools(nvlist_t *pools, nvlist_t *props, char *mntopts, int flags,
char *orig_name, char *new_name,
boolean_t do_destroyed, boolean_t pool_specified, boolean_t do_all,
importargs_t *import)
{
nvlist_t *config = NULL;
nvlist_t *found_config = NULL;
uint64_t pool_state;
/*
* At this point we have a list of import candidate configs. Even if
* we were searching by pool name or guid, we still need to
* post-process the list to deal with pool state and possible
* duplicate names.
*/
int err = 0;
nvpair_t *elem = NULL;
boolean_t first = B_TRUE;
while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) {
verify(nvpair_value_nvlist(elem, &config) == 0);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
&pool_state) == 0);
if (!do_destroyed && pool_state == POOL_STATE_DESTROYED)
continue;
if (do_destroyed && pool_state != POOL_STATE_DESTROYED)
continue;
verify(nvlist_add_nvlist(config, ZPOOL_LOAD_POLICY,
import->policy) == 0);
if (!pool_specified) {
if (first)
first = B_FALSE;
else if (!do_all)
(void) fputc('\n', stdout);
if (do_all) {
err |= do_import(config, NULL, mntopts,
props, flags);
} else {
/*
* If we're importing from cachefile, then
* we don't want to report errors until we
* are in the scan phase of the import. If
* we get an error, then we return that error
* to invoke the scan phase.
*/
if (import->cachefile && !import->scan)
err = show_import(config, B_FALSE);
else
(void) show_import(config, B_TRUE);
}
} else if (import->poolname != NULL) {
const char *name;
/*
* We are searching for a pool based on name.
*/
verify(nvlist_lookup_string(config,
ZPOOL_CONFIG_POOL_NAME, &name) == 0);
if (strcmp(name, import->poolname) == 0) {
if (found_config != NULL) {
(void) fprintf(stderr, gettext(
"cannot import '%s': more than "
"one matching pool\n"),
import->poolname);
(void) fprintf(stderr, gettext(
"import by numeric ID instead\n"));
err = B_TRUE;
}
found_config = config;
}
} else {
uint64_t guid;
/*
* Search for a pool by guid.
*/
verify(nvlist_lookup_uint64(config,
ZPOOL_CONFIG_POOL_GUID, &guid) == 0);
if (guid == import->guid)
found_config = config;
}
}
/*
* If we were searching for a specific pool, verify that we found a
* pool, and then do the import.
*/
if (pool_specified && err == 0) {
if (found_config == NULL) {
(void) fprintf(stderr, gettext("cannot import '%s': "
"no such pool available\n"), orig_name);
err = B_TRUE;
} else {
err |= do_import(found_config, new_name,
mntopts, props, flags);
}
}
/*
* If we were just looking for pools, report an error if none were
* found.
*/
if (!pool_specified && first)
(void) fprintf(stderr,
gettext("no pools available to import\n"));
return (err);
}
typedef struct target_exists_args {
const char *poolname;
uint64_t poolguid;
} target_exists_args_t;
static int
name_or_guid_exists(zpool_handle_t *zhp, void *data)
{
target_exists_args_t *args = data;
nvlist_t *config = zpool_get_config(zhp, NULL);
int found = 0;
if (config == NULL)
return (0);
if (args->poolname != NULL) {
const char *pool_name;
verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
&pool_name) == 0);
if (strcmp(pool_name, args->poolname) == 0)
found = 1;
} else {
uint64_t pool_guid;
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
&pool_guid) == 0);
if (pool_guid == args->poolguid)
found = 1;
}
zpool_close(zhp);
return (found);
}
/*
* zpool checkpoint <pool>
* checkpoint --discard <pool>
*
* -d Discard the checkpoint from a checkpointed
* --discard pool.
*
* -w Wait for discarding a checkpoint to complete.
* --wait
*
* Checkpoints the specified pool, by taking a "snapshot" of its
* current state. A pool can only have one checkpoint at a time.
*/
int
zpool_do_checkpoint(int argc, char **argv)
{
boolean_t discard, wait;
char *pool;
zpool_handle_t *zhp;
int c, err;
struct option long_options[] = {
{"discard", no_argument, NULL, 'd'},
{"wait", no_argument, NULL, 'w'},
{0, 0, 0, 0}
};
discard = B_FALSE;
wait = B_FALSE;
while ((c = getopt_long(argc, argv, ":dw", long_options, NULL)) != -1) {
switch (c) {
case 'd':
discard = B_TRUE;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
if (wait && !discard) {
(void) fprintf(stderr, gettext("--wait only valid when "
"--discard also specified\n"));
usage(B_FALSE);
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
pool = argv[0];
if ((zhp = zpool_open(g_zfs, pool)) == NULL) {
/* As a special case, check for use of '/' in the name */
if (strchr(pool, '/') != NULL)
(void) fprintf(stderr, gettext("'zpool checkpoint' "
"doesn't work on datasets. To save the state "
"of a dataset from a specific point in time "
"please use 'zfs snapshot'\n"));
return (1);
}
if (discard) {
err = (zpool_discard_checkpoint(zhp) != 0);
if (err == 0 && wait)
err = zpool_wait(zhp, ZPOOL_WAIT_CKPT_DISCARD);
} else {
err = (zpool_checkpoint(zhp) != 0);
}
zpool_close(zhp);
return (err);
}
#define CHECKPOINT_OPT 1024
/*
* zpool import [-d dir] [-D]
* import [-o mntopts] [-o prop=value] ... [-R root] [-D] [-l]
* [-d dir | -c cachefile | -s] [-f] -a
* import [-o mntopts] [-o prop=value] ... [-R root] [-D] [-l]
* [-d dir | -c cachefile | -s] [-f] [-n] [-F] <pool | id>
* [newpool]
*
* -c Read pool information from a cachefile instead of searching
* devices. If importing from a cachefile config fails, then
* fallback to searching for devices only in the directories that
* exist in the cachefile.
*
* -d Scan in a specific directory, other than /dev/. More than
* one directory can be specified using multiple '-d' options.
*
* -D Scan for previously destroyed pools or import all or only
* specified destroyed pools.
*
* -R Temporarily import the pool, with all mountpoints relative to
* the given root. The pool will remain exported when the machine
* is rebooted.
*
* -V Import even in the presence of faulted vdevs. This is an
* intentionally undocumented option for testing purposes, and
* treats the pool configuration as complete, leaving any bad
* vdevs in the FAULTED state. In other words, it does verbatim
* import.
*
* -f Force import, even if it appears that the pool is active.
*
* -F Attempt rewind if necessary.
*
* -n See if rewind would work, but don't actually rewind.
*
* -N Import the pool but don't mount datasets.
*
* -T Specify a starting txg to use for import. This option is
* intentionally undocumented option for testing purposes.
*
* -a Import all pools found.
*
* -l Load encryption keys while importing.
*
* -o Set property=value and/or temporary mount options (without '=').
*
* -s Scan using the default search path, the libblkid cache will
* not be consulted.
*
* --rewind-to-checkpoint
* Import the pool and revert back to the checkpoint.
*
* The import command scans for pools to import, and import pools based on pool
* name and GUID. The pool can also be renamed as part of the import process.
*/
int
zpool_do_import(int argc, char **argv)
{
char **searchdirs = NULL;
char *env, *envdup = NULL;
int nsearch = 0;
int c;
int err = 0;
nvlist_t *pools = NULL;
boolean_t do_all = B_FALSE;
boolean_t do_destroyed = B_FALSE;
char *mntopts = NULL;
uint64_t searchguid = 0;
char *searchname = NULL;
char *propval;
nvlist_t *policy = NULL;
nvlist_t *props = NULL;
int flags = ZFS_IMPORT_NORMAL;
uint32_t rewind_policy = ZPOOL_NO_REWIND;
boolean_t dryrun = B_FALSE;
boolean_t do_rewind = B_FALSE;
boolean_t xtreme_rewind = B_FALSE;
boolean_t do_scan = B_FALSE;
boolean_t pool_exists = B_FALSE;
boolean_t pool_specified = B_FALSE;
uint64_t txg = -1ULL;
char *cachefile = NULL;
importargs_t idata = { 0 };
char *endptr;
struct option long_options[] = {
{"rewind-to-checkpoint", no_argument, NULL, CHECKPOINT_OPT},
{0, 0, 0, 0}
};
/* check options */
while ((c = getopt_long(argc, argv, ":aCc:d:DEfFlmnNo:R:stT:VX",
long_options, NULL)) != -1) {
switch (c) {
case 'a':
do_all = B_TRUE;
break;
case 'c':
cachefile = optarg;
break;
case 'd':
searchdirs = safe_realloc(searchdirs,
(nsearch + 1) * sizeof (char *));
searchdirs[nsearch++] = optarg;
break;
case 'D':
do_destroyed = B_TRUE;
break;
case 'f':
flags |= ZFS_IMPORT_ANY_HOST;
break;
case 'F':
do_rewind = B_TRUE;
break;
case 'l':
flags |= ZFS_IMPORT_LOAD_KEYS;
break;
case 'm':
flags |= ZFS_IMPORT_MISSING_LOG;
break;
case 'n':
dryrun = B_TRUE;
break;
case 'N':
flags |= ZFS_IMPORT_ONLY;
break;
case 'o':
if ((propval = strchr(optarg, '=')) != NULL) {
*propval = '\0';
propval++;
if (add_prop_list(optarg, propval,
&props, B_TRUE))
goto error;
} else {
mntopts = optarg;
}
break;
case 'R':
if (add_prop_list(zpool_prop_to_name(
ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE))
goto error;
if (add_prop_list_default(zpool_prop_to_name(
ZPOOL_PROP_CACHEFILE), "none", &props))
goto error;
break;
case 's':
do_scan = B_TRUE;
break;
case 't':
flags |= ZFS_IMPORT_TEMP_NAME;
if (add_prop_list_default(zpool_prop_to_name(
ZPOOL_PROP_CACHEFILE), "none", &props))
goto error;
break;
case 'T':
errno = 0;
txg = strtoull(optarg, &endptr, 0);
if (errno != 0 || *endptr != '\0') {
(void) fprintf(stderr,
gettext("invalid txg value\n"));
usage(B_FALSE);
}
rewind_policy = ZPOOL_DO_REWIND | ZPOOL_EXTREME_REWIND;
break;
case 'V':
flags |= ZFS_IMPORT_VERBATIM;
break;
case 'X':
xtreme_rewind = B_TRUE;
break;
case CHECKPOINT_OPT:
flags |= ZFS_IMPORT_CHECKPOINT;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (cachefile && nsearch != 0) {
(void) fprintf(stderr, gettext("-c is incompatible with -d\n"));
usage(B_FALSE);
}
if (cachefile && do_scan) {
(void) fprintf(stderr, gettext("-c is incompatible with -s\n"));
usage(B_FALSE);
}
if ((flags & ZFS_IMPORT_LOAD_KEYS) && (flags & ZFS_IMPORT_ONLY)) {
(void) fprintf(stderr, gettext("-l is incompatible with -N\n"));
usage(B_FALSE);
}
if ((flags & ZFS_IMPORT_LOAD_KEYS) && !do_all && argc == 0) {
(void) fprintf(stderr, gettext("-l is only meaningful during "
"an import\n"));
usage(B_FALSE);
}
if ((dryrun || xtreme_rewind) && !do_rewind) {
(void) fprintf(stderr,
gettext("-n or -X only meaningful with -F\n"));
usage(B_FALSE);
}
if (dryrun)
rewind_policy = ZPOOL_TRY_REWIND;
else if (do_rewind)
rewind_policy = ZPOOL_DO_REWIND;
if (xtreme_rewind)
rewind_policy |= ZPOOL_EXTREME_REWIND;
/* In the future, we can capture further policy and include it here */
if (nvlist_alloc(&policy, NV_UNIQUE_NAME, 0) != 0 ||
nvlist_add_uint64(policy, ZPOOL_LOAD_REQUEST_TXG, txg) != 0 ||
nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY,
rewind_policy) != 0)
goto error;
/* check argument count */
if (do_all) {
if (argc != 0) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
} else {
if (argc > 2) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
}
/*
* Check for the effective uid. We do this explicitly here because
* otherwise any attempt to discover pools will silently fail.
*/
if (argc == 0 && geteuid() != 0) {
(void) fprintf(stderr, gettext("cannot "
"discover pools: permission denied\n"));
free(searchdirs);
nvlist_free(props);
nvlist_free(policy);
return (1);
}
/*
* Depending on the arguments given, we do one of the following:
*
* <none> Iterate through all pools and display information about
* each one.
*
* -a Iterate through all pools and try to import each one.
*
* <id> Find the pool that corresponds to the given GUID/pool
* name and import that one.
*
* -D Above options applies only to destroyed pools.
*/
if (argc != 0) {
char *endptr;
errno = 0;
searchguid = strtoull(argv[0], &endptr, 10);
if (errno != 0 || *endptr != '\0') {
searchname = argv[0];
searchguid = 0;
}
pool_specified = B_TRUE;
/*
* User specified a name or guid. Ensure it's unique.
*/
target_exists_args_t search = {searchname, searchguid};
pool_exists = zpool_iter(g_zfs, name_or_guid_exists, &search);
}
/*
* Check the environment for the preferred search path.
*/
if ((searchdirs == NULL) && (env = getenv("ZPOOL_IMPORT_PATH"))) {
char *dir, *tmp = NULL;
envdup = strdup(env);
for (dir = strtok_r(envdup, ":", &tmp);
dir != NULL;
dir = strtok_r(NULL, ":", &tmp)) {
searchdirs = safe_realloc(searchdirs,
(nsearch + 1) * sizeof (char *));
searchdirs[nsearch++] = dir;
}
}
idata.path = searchdirs;
idata.paths = nsearch;
idata.poolname = searchname;
idata.guid = searchguid;
idata.cachefile = cachefile;
idata.scan = do_scan;
idata.policy = policy;
libpc_handle_t lpch = {
.lpc_lib_handle = g_zfs,
.lpc_ops = &libzfs_config_ops,
.lpc_printerr = B_TRUE
};
pools = zpool_search_import(&lpch, &idata);
if (pools != NULL && pool_exists &&
(argc == 1 || strcmp(argv[0], argv[1]) == 0)) {
(void) fprintf(stderr, gettext("cannot import '%s': "
"a pool with that name already exists\n"),
argv[0]);
(void) fprintf(stderr, gettext("use the form '%s "
"<pool | id> <newpool>' to give it a new name\n"),
"zpool import");
err = 1;
} else if (pools == NULL && pool_exists) {
(void) fprintf(stderr, gettext("cannot import '%s': "
"a pool with that name is already created/imported,\n"),
argv[0]);
(void) fprintf(stderr, gettext("and no additional pools "
"with that name were found\n"));
err = 1;
} else if (pools == NULL) {
if (argc != 0) {
(void) fprintf(stderr, gettext("cannot import '%s': "
"no such pool available\n"), argv[0]);
}
err = 1;
}
if (err == 1) {
free(searchdirs);
free(envdup);
nvlist_free(policy);
nvlist_free(pools);
nvlist_free(props);
return (1);
}
err = import_pools(pools, props, mntopts, flags,
argc >= 1 ? argv[0] : NULL,
argc >= 2 ? argv[1] : NULL,
do_destroyed, pool_specified, do_all, &idata);
/*
* If we're using the cachefile and we failed to import, then
* fallback to scanning the directory for pools that match
* those in the cachefile.
*/
if (err != 0 && cachefile != NULL) {
(void) printf(gettext("cachefile import failed, retrying\n"));
/*
* We use the scan flag to gather the directories that exist
* in the cachefile. If we need to fallback to searching for
* the pool config, we will only search devices in these
* directories.
*/
idata.scan = B_TRUE;
nvlist_free(pools);
pools = zpool_search_import(&lpch, &idata);
err = import_pools(pools, props, mntopts, flags,
argc >= 1 ? argv[0] : NULL,
argc >= 2 ? argv[1] : NULL,
do_destroyed, pool_specified, do_all, &idata);
}
error:
nvlist_free(props);
nvlist_free(pools);
nvlist_free(policy);
free(searchdirs);
free(envdup);
return (err ? 1 : 0);
}
/*
* zpool sync [-f] [pool] ...
*
* -f (undocumented) force uberblock (and config including zpool cache file)
* update.
*
* Sync the specified pool(s).
* Without arguments "zpool sync" will sync all pools.
* This command initiates TXG sync(s) and will return after the TXG(s) commit.
*
*/
static int
zpool_do_sync(int argc, char **argv)
{
int ret;
boolean_t force = B_FALSE;
/* check options */
while ((ret = getopt(argc, argv, "f")) != -1) {
switch (ret) {
case 'f':
force = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* if argc == 0 we will execute zpool_sync_one on all pools */
ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL,
B_FALSE, zpool_sync_one, &force);
return (ret);
}
typedef struct iostat_cbdata {
uint64_t cb_flags;
int cb_namewidth;
int cb_iteration;
boolean_t cb_verbose;
boolean_t cb_literal;
boolean_t cb_scripted;
zpool_list_t *cb_list;
vdev_cmd_data_list_t *vcdl;
vdev_cbdata_t cb_vdevs;
} iostat_cbdata_t;
/* iostat labels */
typedef struct name_and_columns {
const char *name; /* Column name */
unsigned int columns; /* Center name to this number of columns */
} name_and_columns_t;
#define IOSTAT_MAX_LABELS 15 /* Max number of labels on one line */
static const name_and_columns_t iostat_top_labels[][IOSTAT_MAX_LABELS] =
{
[IOS_DEFAULT] = {{"capacity", 2}, {"operations", 2}, {"bandwidth", 2},
{NULL}},
[IOS_LATENCY] = {{"total_wait", 2}, {"disk_wait", 2}, {"syncq_wait", 2},
{"asyncq_wait", 2}, {"scrub", 1}, {"trim", 1}, {"rebuild", 1},
{NULL}},
[IOS_QUEUES] = {{"syncq_read", 2}, {"syncq_write", 2},
{"asyncq_read", 2}, {"asyncq_write", 2}, {"scrubq_read", 2},
{"trimq_write", 2}, {"rebuildq_write", 2}, {NULL}},
[IOS_L_HISTO] = {{"total_wait", 2}, {"disk_wait", 2}, {"syncq_wait", 2},
{"asyncq_wait", 2}, {NULL}},
[IOS_RQ_HISTO] = {{"sync_read", 2}, {"sync_write", 2},
{"async_read", 2}, {"async_write", 2}, {"scrub", 2},
{"trim", 2}, {"rebuild", 2}, {NULL}},
};
/* Shorthand - if "columns" field not set, default to 1 column */
static const name_and_columns_t iostat_bottom_labels[][IOSTAT_MAX_LABELS] =
{
[IOS_DEFAULT] = {{"alloc"}, {"free"}, {"read"}, {"write"}, {"read"},
{"write"}, {NULL}},
[IOS_LATENCY] = {{"read"}, {"write"}, {"read"}, {"write"}, {"read"},
{"write"}, {"read"}, {"write"}, {"wait"}, {"wait"}, {"wait"},
{NULL}},
[IOS_QUEUES] = {{"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"},
{"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"},
{"pend"}, {"activ"}, {"pend"}, {"activ"}, {NULL}},
[IOS_L_HISTO] = {{"read"}, {"write"}, {"read"}, {"write"}, {"read"},
{"write"}, {"read"}, {"write"}, {"scrub"}, {"trim"}, {"rebuild"},
{NULL}},
[IOS_RQ_HISTO] = {{"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"},
{"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"},
{"ind"}, {"agg"}, {NULL}},
};
static const char *histo_to_title[] = {
[IOS_L_HISTO] = "latency",
[IOS_RQ_HISTO] = "req_size",
};
/*
* Return the number of labels in a null-terminated name_and_columns_t
* array.
*
*/
static unsigned int
label_array_len(const name_and_columns_t *labels)
{
int i = 0;
while (labels[i].name)
i++;
return (i);
}
/*
* Return the number of strings in a null-terminated string array.
* For example:
*
* const char foo[] = {"bar", "baz", NULL}
*
* returns 2
*/
static uint64_t
str_array_len(const char *array[])
{
uint64_t i = 0;
while (array[i])
i++;
return (i);
}
/*
* Return a default column width for default/latency/queue columns. This does
* not include histograms, which have their columns autosized.
*/
static unsigned int
default_column_width(iostat_cbdata_t *cb, enum iostat_type type)
{
unsigned long column_width = 5; /* Normal niceprint */
static unsigned long widths[] = {
/*
* Choose some sane default column sizes for printing the
* raw numbers.
*/
[IOS_DEFAULT] = 15, /* 1PB capacity */
[IOS_LATENCY] = 10, /* 1B ns = 10sec */
[IOS_QUEUES] = 6, /* 1M queue entries */
[IOS_L_HISTO] = 10, /* 1B ns = 10sec */
[IOS_RQ_HISTO] = 6, /* 1M queue entries */
};
if (cb->cb_literal)
column_width = widths[type];
return (column_width);
}
/*
* Print the column labels, i.e:
*
* capacity operations bandwidth
* alloc free read write read write ...
*
* If force_column_width is set, use it for the column width. If not set, use
* the default column width.
*/
static void
print_iostat_labels(iostat_cbdata_t *cb, unsigned int force_column_width,
const name_and_columns_t labels[][IOSTAT_MAX_LABELS])
{
int i, idx, s;
int text_start, rw_column_width, spaces_to_end;
uint64_t flags = cb->cb_flags;
uint64_t f;
unsigned int column_width = force_column_width;
/* For each bit set in flags */
for (f = flags; f; f &= ~(1ULL << idx)) {
idx = lowbit64(f) - 1;
if (!force_column_width)
column_width = default_column_width(cb, idx);
/* Print our top labels centered over "read write" label. */
for (i = 0; i < label_array_len(labels[idx]); i++) {
const char *name = labels[idx][i].name;
/*
* We treat labels[][].columns == 0 as shorthand
* for one column. It makes writing out the label
* tables more concise.
*/
unsigned int columns = MAX(1, labels[idx][i].columns);
unsigned int slen = strlen(name);
rw_column_width = (column_width * columns) +
(2 * (columns - 1));
text_start = (int)((rw_column_width) / columns -
slen / columns);
if (text_start < 0)
text_start = 0;
printf(" "); /* Two spaces between columns */
/* Space from beginning of column to label */
for (s = 0; s < text_start; s++)
printf(" ");
printf("%s", name);
/* Print space after label to end of column */
spaces_to_end = rw_column_width - text_start - slen;
if (spaces_to_end < 0)
spaces_to_end = 0;
for (s = 0; s < spaces_to_end; s++)
printf(" ");
}
}
}
/*
* print_cmd_columns - Print custom column titles from -c
*
* If the user specified the "zpool status|iostat -c" then print their custom
* column titles in the header. For example, print_cmd_columns() would print
* the " col1 col2" part of this:
*
* $ zpool iostat -vc 'echo col1=val1; echo col2=val2'
* ...
* capacity operations bandwidth
* pool alloc free read write read write col1 col2
* ---------- ----- ----- ----- ----- ----- ----- ---- ----
* mypool 269K 1008M 0 0 107 946
* mirror 269K 1008M 0 0 107 946
* sdb - - 0 0 102 473 val1 val2
* sdc - - 0 0 5 473 val1 val2
* ---------- ----- ----- ----- ----- ----- ----- ---- ----
*/
static void
print_cmd_columns(vdev_cmd_data_list_t *vcdl, int use_dashes)
{
int i, j;
vdev_cmd_data_t *data = &vcdl->data[0];
if (vcdl->count == 0 || data == NULL)
return;
/*
* Each vdev cmd should have the same column names unless the user did
* something weird with their cmd. Just take the column names from the
* first vdev and assume it works for all of them.
*/
for (i = 0; i < vcdl->uniq_cols_cnt; i++) {
printf(" ");
if (use_dashes) {
for (j = 0; j < vcdl->uniq_cols_width[i]; j++)
printf("-");
} else {
printf_color(ANSI_BOLD, "%*s", vcdl->uniq_cols_width[i],
vcdl->uniq_cols[i]);
}
}
}
/*
* Utility function to print out a line of dashes like:
*
* -------------------------------- ----- ----- ----- ----- -----
*
* ...or a dashed named-row line like:
*
* logs - - - - -
*
* @cb: iostat data
*
* @force_column_width If non-zero, use the value as the column width.
* Otherwise use the default column widths.
*
* @name: Print a dashed named-row line starting
* with @name. Otherwise, print a regular
* dashed line.
*/
static void
print_iostat_dashes(iostat_cbdata_t *cb, unsigned int force_column_width,
const char *name)
{
int i;
unsigned int namewidth;
uint64_t flags = cb->cb_flags;
uint64_t f;
int idx;
const name_and_columns_t *labels;
const char *title;
if (cb->cb_flags & IOS_ANYHISTO_M) {
title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)];
} else if (cb->cb_vdevs.cb_names_count) {
title = "vdev";
} else {
title = "pool";
}
namewidth = MAX(MAX(strlen(title), cb->cb_namewidth),
name ? strlen(name) : 0);
if (name) {
printf("%-*s", namewidth, name);
} else {
for (i = 0; i < namewidth; i++)
(void) printf("-");
}
/* For each bit in flags */
for (f = flags; f; f &= ~(1ULL << idx)) {
unsigned int column_width;
idx = lowbit64(f) - 1;
if (force_column_width)
column_width = force_column_width;
else
column_width = default_column_width(cb, idx);
labels = iostat_bottom_labels[idx];
for (i = 0; i < label_array_len(labels); i++) {
if (name)
printf(" %*s-", column_width - 1, " ");
else
printf(" %.*s", column_width,
"--------------------");
}
}
}
static void
print_iostat_separator_impl(iostat_cbdata_t *cb,
unsigned int force_column_width)
{
print_iostat_dashes(cb, force_column_width, NULL);
}
static void
print_iostat_separator(iostat_cbdata_t *cb)
{
print_iostat_separator_impl(cb, 0);
}
static void
print_iostat_header_impl(iostat_cbdata_t *cb, unsigned int force_column_width,
const char *histo_vdev_name)
{
unsigned int namewidth;
const char *title;
color_start(ANSI_BOLD);
if (cb->cb_flags & IOS_ANYHISTO_M) {
title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)];
} else if (cb->cb_vdevs.cb_names_count) {
title = "vdev";
} else {
title = "pool";
}
namewidth = MAX(MAX(strlen(title), cb->cb_namewidth),
histo_vdev_name ? strlen(histo_vdev_name) : 0);
if (histo_vdev_name)
printf("%-*s", namewidth, histo_vdev_name);
else
printf("%*s", namewidth, "");
print_iostat_labels(cb, force_column_width, iostat_top_labels);
printf("\n");
printf("%-*s", namewidth, title);
print_iostat_labels(cb, force_column_width, iostat_bottom_labels);
if (cb->vcdl != NULL)
print_cmd_columns(cb->vcdl, 0);
printf("\n");
print_iostat_separator_impl(cb, force_column_width);
if (cb->vcdl != NULL)
print_cmd_columns(cb->vcdl, 1);
color_end();
printf("\n");
}
static void
print_iostat_header(iostat_cbdata_t *cb)
{
print_iostat_header_impl(cb, 0, NULL);
}
/*
* Prints a size string (i.e. 120M) with the suffix ("M") colored
* by order of magnitude. Uses column_size to add padding.
*/
static void
-print_stat_color(char *statbuf, unsigned int column_size)
+print_stat_color(const char *statbuf, unsigned int column_size)
{
fputs(" ", stdout);
+ size_t len = strlen(statbuf);
+ while (len < column_size) {
+ fputc(' ', stdout);
+ column_size--;
+ }
if (*statbuf == '0') {
color_start(ANSI_GRAY);
fputc('0', stdout);
- column_size--;
} else {
for (; *statbuf; statbuf++) {
if (*statbuf == 'K') color_start(ANSI_GREEN);
else if (*statbuf == 'M') color_start(ANSI_YELLOW);
else if (*statbuf == 'G') color_start(ANSI_RED);
else if (*statbuf == 'T') color_start(ANSI_BOLD_BLUE);
else if (*statbuf == 'P') color_start(ANSI_MAGENTA);
else if (*statbuf == 'E') color_start(ANSI_CYAN);
fputc(*statbuf, stdout);
if (--column_size <= 0)
break;
}
}
color_end();
- for (; column_size > 0; column_size--)
- fputc(' ', stdout);
}
/*
* Display a single statistic.
*/
static void
print_one_stat(uint64_t value, enum zfs_nicenum_format format,
unsigned int column_size, boolean_t scripted)
{
char buf[64];
zfs_nicenum_format(value, buf, sizeof (buf), format);
if (scripted)
printf("\t%s", buf);
else
print_stat_color(buf, column_size);
}
/*
* Calculate the default vdev stats
*
* Subtract oldvs from newvs, apply a scaling factor, and save the resulting
* stats into calcvs.
*/
static void
calc_default_iostats(vdev_stat_t *oldvs, vdev_stat_t *newvs,
vdev_stat_t *calcvs)
{
int i;
memcpy(calcvs, newvs, sizeof (*calcvs));
for (i = 0; i < ARRAY_SIZE(calcvs->vs_ops); i++)
calcvs->vs_ops[i] = (newvs->vs_ops[i] - oldvs->vs_ops[i]);
for (i = 0; i < ARRAY_SIZE(calcvs->vs_bytes); i++)
calcvs->vs_bytes[i] = (newvs->vs_bytes[i] - oldvs->vs_bytes[i]);
}
/*
* Internal representation of the extended iostats data.
*
* The extended iostat stats are exported in nvlists as either uint64_t arrays
* or single uint64_t's. We make both look like arrays to make them easier
* to process. In order to make single uint64_t's look like arrays, we set
* __data to the stat data, and then set *data = &__data with count = 1. Then,
* we can just use *data and count.
*/
struct stat_array {
uint64_t *data;
uint_t count; /* Number of entries in data[] */
uint64_t __data; /* Only used when data is a single uint64_t */
};
static uint64_t
stat_histo_max(struct stat_array *nva, unsigned int len)
{
uint64_t max = 0;
int i;
for (i = 0; i < len; i++)
max = MAX(max, array64_max(nva[i].data, nva[i].count));
return (max);
}
/*
* Helper function to lookup a uint64_t array or uint64_t value and store its
* data as a stat_array. If the nvpair is a single uint64_t value, then we make
* it look like a one element array to make it easier to process.
*/
static int
nvpair64_to_stat_array(nvlist_t *nvl, const char *name,
struct stat_array *nva)
{
nvpair_t *tmp;
int ret;
verify(nvlist_lookup_nvpair(nvl, name, &tmp) == 0);
switch (nvpair_type(tmp)) {
case DATA_TYPE_UINT64_ARRAY:
ret = nvpair_value_uint64_array(tmp, &nva->data, &nva->count);
break;
case DATA_TYPE_UINT64:
ret = nvpair_value_uint64(tmp, &nva->__data);
nva->data = &nva->__data;
nva->count = 1;
break;
default:
/* Not a uint64_t */
ret = EINVAL;
break;
}
return (ret);
}
/*
* Given a list of nvlist names, look up the extended stats in newnv and oldnv,
* subtract them, and return the results in a newly allocated stat_array.
* You must free the returned array after you are done with it with
* free_calc_stats().
*
* Additionally, you can set "oldnv" to NULL if you simply want the newnv
* values.
*/
static struct stat_array *
calc_and_alloc_stats_ex(const char **names, unsigned int len, nvlist_t *oldnv,
nvlist_t *newnv)
{
nvlist_t *oldnvx = NULL, *newnvx;
struct stat_array *oldnva, *newnva, *calcnva;
int i, j;
unsigned int alloc_size = (sizeof (struct stat_array)) * len;
/* Extract our extended stats nvlist from the main list */
verify(nvlist_lookup_nvlist(newnv, ZPOOL_CONFIG_VDEV_STATS_EX,
&newnvx) == 0);
if (oldnv) {
verify(nvlist_lookup_nvlist(oldnv, ZPOOL_CONFIG_VDEV_STATS_EX,
&oldnvx) == 0);
}
newnva = safe_malloc(alloc_size);
oldnva = safe_malloc(alloc_size);
calcnva = safe_malloc(alloc_size);
for (j = 0; j < len; j++) {
verify(nvpair64_to_stat_array(newnvx, names[j],
&newnva[j]) == 0);
calcnva[j].count = newnva[j].count;
alloc_size = calcnva[j].count * sizeof (calcnva[j].data[0]);
calcnva[j].data = safe_malloc(alloc_size);
memcpy(calcnva[j].data, newnva[j].data, alloc_size);
if (oldnvx) {
verify(nvpair64_to_stat_array(oldnvx, names[j],
&oldnva[j]) == 0);
for (i = 0; i < oldnva[j].count; i++)
calcnva[j].data[i] -= oldnva[j].data[i];
}
}
free(newnva);
free(oldnva);
return (calcnva);
}
static void
free_calc_stats(struct stat_array *nva, unsigned int len)
{
int i;
for (i = 0; i < len; i++)
free(nva[i].data);
free(nva);
}
static void
print_iostat_histo(struct stat_array *nva, unsigned int len,
iostat_cbdata_t *cb, unsigned int column_width, unsigned int namewidth,
double scale)
{
int i, j;
char buf[6];
uint64_t val;
enum zfs_nicenum_format format;
unsigned int buckets;
unsigned int start_bucket;
if (cb->cb_literal)
format = ZFS_NICENUM_RAW;
else
format = ZFS_NICENUM_1024;
/* All these histos are the same size, so just use nva[0].count */
buckets = nva[0].count;
if (cb->cb_flags & IOS_RQ_HISTO_M) {
/* Start at 512 - req size should never be lower than this */
start_bucket = 9;
} else {
start_bucket = 0;
}
for (j = start_bucket; j < buckets; j++) {
/* Print histogram bucket label */
if (cb->cb_flags & IOS_L_HISTO_M) {
/* Ending range of this bucket */
val = (1UL << (j + 1)) - 1;
zfs_nicetime(val, buf, sizeof (buf));
} else {
/* Request size (starting range of bucket) */
val = (1UL << j);
zfs_nicenum(val, buf, sizeof (buf));
}
if (cb->cb_scripted)
printf("%llu", (u_longlong_t)val);
else
printf("%-*s", namewidth, buf);
/* Print the values on the line */
for (i = 0; i < len; i++) {
print_one_stat(nva[i].data[j] * scale, format,
column_width, cb->cb_scripted);
}
printf("\n");
}
}
static void
print_solid_separator(unsigned int length)
{
while (length--)
printf("-");
printf("\n");
}
static void
print_iostat_histos(iostat_cbdata_t *cb, nvlist_t *oldnv,
nvlist_t *newnv, double scale, const char *name)
{
unsigned int column_width;
unsigned int namewidth;
unsigned int entire_width;
enum iostat_type type;
struct stat_array *nva;
const char **names;
unsigned int names_len;
/* What type of histo are we? */
type = IOS_HISTO_IDX(cb->cb_flags);
/* Get NULL-terminated array of nvlist names for our histo */
names = vsx_type_to_nvlist[type];
names_len = str_array_len(names); /* num of names */
nva = calc_and_alloc_stats_ex(names, names_len, oldnv, newnv);
if (cb->cb_literal) {
column_width = MAX(5,
(unsigned int) log10(stat_histo_max(nva, names_len)) + 1);
} else {
column_width = 5;
}
namewidth = MAX(cb->cb_namewidth,
strlen(histo_to_title[IOS_HISTO_IDX(cb->cb_flags)]));
/*
* Calculate the entire line width of what we're printing. The
* +2 is for the two spaces between columns:
*/
/* read write */
/* ----- ----- */
/* |___| <---------- column_width */
/* */
/* |__________| <--- entire_width */
/* */
entire_width = namewidth + (column_width + 2) *
label_array_len(iostat_bottom_labels[type]);
if (cb->cb_scripted)
printf("%s\n", name);
else
print_iostat_header_impl(cb, column_width, name);
print_iostat_histo(nva, names_len, cb, column_width,
namewidth, scale);
free_calc_stats(nva, names_len);
if (!cb->cb_scripted)
print_solid_separator(entire_width);
}
/*
* Calculate the average latency of a power-of-two latency histogram
*/
static uint64_t
single_histo_average(uint64_t *histo, unsigned int buckets)
{
int i;
uint64_t count = 0, total = 0;
for (i = 0; i < buckets; i++) {
/*
* Our buckets are power-of-two latency ranges. Use the
* midpoint latency of each bucket to calculate the average.
* For example:
*
* Bucket Midpoint
* 8ns-15ns: 12ns
* 16ns-31ns: 24ns
* ...
*/
if (histo[i] != 0) {
total += histo[i] * (((1UL << i) + ((1UL << i)/2)));
count += histo[i];
}
}
/* Prevent divide by zero */
return (count == 0 ? 0 : total / count);
}
static void
print_iostat_queues(iostat_cbdata_t *cb, nvlist_t *newnv)
{
const char *names[] = {
ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE,
ZPOOL_CONFIG_VDEV_REBUILD_PEND_QUEUE,
ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE,
};
struct stat_array *nva;
unsigned int column_width = default_column_width(cb, IOS_QUEUES);
enum zfs_nicenum_format format;
nva = calc_and_alloc_stats_ex(names, ARRAY_SIZE(names), NULL, newnv);
if (cb->cb_literal)
format = ZFS_NICENUM_RAW;
else
format = ZFS_NICENUM_1024;
for (int i = 0; i < ARRAY_SIZE(names); i++) {
uint64_t val = nva[i].data[0];
print_one_stat(val, format, column_width, cb->cb_scripted);
}
free_calc_stats(nva, ARRAY_SIZE(names));
}
static void
print_iostat_latency(iostat_cbdata_t *cb, nvlist_t *oldnv,
nvlist_t *newnv)
{
int i;
uint64_t val;
const char *names[] = {
ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO,
ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO,
ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO,
ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO,
ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO,
};
struct stat_array *nva;
unsigned int column_width = default_column_width(cb, IOS_LATENCY);
enum zfs_nicenum_format format;
nva = calc_and_alloc_stats_ex(names, ARRAY_SIZE(names), oldnv, newnv);
if (cb->cb_literal)
format = ZFS_NICENUM_RAWTIME;
else
format = ZFS_NICENUM_TIME;
/* Print our avg latencies on the line */
for (i = 0; i < ARRAY_SIZE(names); i++) {
/* Compute average latency for a latency histo */
val = single_histo_average(nva[i].data, nva[i].count);
print_one_stat(val, format, column_width, cb->cb_scripted);
}
free_calc_stats(nva, ARRAY_SIZE(names));
}
/*
* Print default statistics (capacity/operations/bandwidth)
*/
static void
print_iostat_default(vdev_stat_t *vs, iostat_cbdata_t *cb, double scale)
{
unsigned int column_width = default_column_width(cb, IOS_DEFAULT);
enum zfs_nicenum_format format;
char na; /* char to print for "not applicable" values */
if (cb->cb_literal) {
format = ZFS_NICENUM_RAW;
na = '0';
} else {
format = ZFS_NICENUM_1024;
na = '-';
}
/* only toplevel vdevs have capacity stats */
if (vs->vs_space == 0) {
if (cb->cb_scripted)
printf("\t%c\t%c", na, na);
else
printf(" %*c %*c", column_width, na, column_width,
na);
} else {
print_one_stat(vs->vs_alloc, format, column_width,
cb->cb_scripted);
print_one_stat(vs->vs_space - vs->vs_alloc, format,
column_width, cb->cb_scripted);
}
print_one_stat((uint64_t)(vs->vs_ops[ZIO_TYPE_READ] * scale),
format, column_width, cb->cb_scripted);
print_one_stat((uint64_t)(vs->vs_ops[ZIO_TYPE_WRITE] * scale),
format, column_width, cb->cb_scripted);
print_one_stat((uint64_t)(vs->vs_bytes[ZIO_TYPE_READ] * scale),
format, column_width, cb->cb_scripted);
print_one_stat((uint64_t)(vs->vs_bytes[ZIO_TYPE_WRITE] * scale),
format, column_width, cb->cb_scripted);
}
static const char *const class_name[] = {
VDEV_ALLOC_BIAS_DEDUP,
VDEV_ALLOC_BIAS_SPECIAL,
VDEV_ALLOC_CLASS_LOGS
};
/*
* Print out all the statistics for the given vdev. This can either be the
* toplevel configuration, or called recursively. If 'name' is NULL, then this
* is a verbose output, and we don't want to display the toplevel pool stats.
*
* Returns the number of stat lines printed.
*/
static unsigned int
print_vdev_stats(zpool_handle_t *zhp, const char *name, nvlist_t *oldnv,
nvlist_t *newnv, iostat_cbdata_t *cb, int depth)
{
nvlist_t **oldchild, **newchild;
uint_t c, children, oldchildren;
vdev_stat_t *oldvs, *newvs, *calcvs;
vdev_stat_t zerovs = { 0 };
char *vname;
int i;
int ret = 0;
uint64_t tdelta;
double scale;
if (strcmp(name, VDEV_TYPE_INDIRECT) == 0)
return (ret);
calcvs = safe_malloc(sizeof (*calcvs));
if (oldnv != NULL) {
verify(nvlist_lookup_uint64_array(oldnv,
ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&oldvs, &c) == 0);
} else {
oldvs = &zerovs;
}
/* Do we only want to see a specific vdev? */
for (i = 0; i < cb->cb_vdevs.cb_names_count; i++) {
/* Yes we do. Is this the vdev? */
if (strcmp(name, cb->cb_vdevs.cb_names[i]) == 0) {
/*
* This is our vdev. Since it is the only vdev we
* will be displaying, make depth = 0 so that it
* doesn't get indented.
*/
depth = 0;
break;
}
}
if (cb->cb_vdevs.cb_names_count && (i == cb->cb_vdevs.cb_names_count)) {
/* Couldn't match the name */
goto children;
}
verify(nvlist_lookup_uint64_array(newnv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&newvs, &c) == 0);
/*
* Print the vdev name unless it's is a histogram. Histograms
* display the vdev name in the header itself.
*/
if (!(cb->cb_flags & IOS_ANYHISTO_M)) {
if (cb->cb_scripted) {
printf("%s", name);
} else {
if (strlen(name) + depth > cb->cb_namewidth)
(void) printf("%*s%s", depth, "", name);
else
(void) printf("%*s%s%*s", depth, "", name,
(int)(cb->cb_namewidth - strlen(name) -
depth), "");
}
}
/* Calculate our scaling factor */
tdelta = newvs->vs_timestamp - oldvs->vs_timestamp;
if ((oldvs->vs_timestamp == 0) && (cb->cb_flags & IOS_ANYHISTO_M)) {
/*
* If we specify printing histograms with no time interval, then
* print the histogram numbers over the entire lifetime of the
* vdev.
*/
scale = 1;
} else {
if (tdelta == 0)
scale = 1.0;
else
scale = (double)NANOSEC / tdelta;
}
if (cb->cb_flags & IOS_DEFAULT_M) {
calc_default_iostats(oldvs, newvs, calcvs);
print_iostat_default(calcvs, cb, scale);
}
if (cb->cb_flags & IOS_LATENCY_M)
print_iostat_latency(cb, oldnv, newnv);
if (cb->cb_flags & IOS_QUEUES_M)
print_iostat_queues(cb, newnv);
if (cb->cb_flags & IOS_ANYHISTO_M) {
printf("\n");
print_iostat_histos(cb, oldnv, newnv, scale, name);
}
if (cb->vcdl != NULL) {
const char *path;
if (nvlist_lookup_string(newnv, ZPOOL_CONFIG_PATH,
&path) == 0) {
printf(" ");
zpool_print_cmd(cb->vcdl, zpool_get_name(zhp), path);
}
}
if (!(cb->cb_flags & IOS_ANYHISTO_M))
printf("\n");
ret++;
children:
free(calcvs);
if (!cb->cb_verbose)
return (ret);
if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_CHILDREN,
&newchild, &children) != 0)
return (ret);
if (oldnv) {
if (nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_CHILDREN,
&oldchild, &oldchildren) != 0)
return (ret);
children = MIN(oldchildren, children);
}
/*
* print normal top-level devices
*/
for (c = 0; c < children; c++) {
uint64_t ishole = B_FALSE, islog = B_FALSE;
(void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_HOLE,
&ishole);
(void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_LOG,
&islog);
if (ishole || islog)
continue;
if (nvlist_exists(newchild[c], ZPOOL_CONFIG_ALLOCATION_BIAS))
continue;
vname = zpool_vdev_name(g_zfs, zhp, newchild[c],
cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID);
ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL,
newchild[c], cb, depth + 2);
free(vname);
}
/*
* print all other top-level devices
*/
for (uint_t n = 0; n < ARRAY_SIZE(class_name); n++) {
boolean_t printed = B_FALSE;
for (c = 0; c < children; c++) {
uint64_t islog = B_FALSE;
const char *bias = NULL;
const char *type = NULL;
(void) nvlist_lookup_uint64(newchild[c],
ZPOOL_CONFIG_IS_LOG, &islog);
if (islog) {
bias = VDEV_ALLOC_CLASS_LOGS;
} else {
(void) nvlist_lookup_string(newchild[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias);
(void) nvlist_lookup_string(newchild[c],
ZPOOL_CONFIG_TYPE, &type);
}
if (bias == NULL || strcmp(bias, class_name[n]) != 0)
continue;
if (!islog && strcmp(type, VDEV_TYPE_INDIRECT) == 0)
continue;
if (!printed) {
if ((!(cb->cb_flags & IOS_ANYHISTO_M)) &&
!cb->cb_scripted &&
!cb->cb_vdevs.cb_names) {
print_iostat_dashes(cb, 0,
class_name[n]);
}
printf("\n");
printed = B_TRUE;
}
vname = zpool_vdev_name(g_zfs, zhp, newchild[c],
cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID);
ret += print_vdev_stats(zhp, vname, oldnv ?
oldchild[c] : NULL, newchild[c], cb, depth + 2);
free(vname);
}
}
/*
* Include level 2 ARC devices in iostat output
*/
if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_L2CACHE,
&newchild, &children) != 0)
return (ret);
if (oldnv) {
if (nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_L2CACHE,
&oldchild, &oldchildren) != 0)
return (ret);
children = MIN(oldchildren, children);
}
if (children > 0) {
if ((!(cb->cb_flags & IOS_ANYHISTO_M)) && !cb->cb_scripted &&
!cb->cb_vdevs.cb_names) {
print_iostat_dashes(cb, 0, "cache");
}
printf("\n");
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, zhp, newchild[c],
cb->cb_vdevs.cb_name_flags);
ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c]
: NULL, newchild[c], cb, depth + 2);
free(vname);
}
}
return (ret);
}
static int
refresh_iostat(zpool_handle_t *zhp, void *data)
{
iostat_cbdata_t *cb = data;
boolean_t missing;
/*
* If the pool has disappeared, remove it from the list and continue.
*/
if (zpool_refresh_stats(zhp, &missing) != 0)
return (-1);
if (missing)
pool_list_remove(cb->cb_list, zhp);
return (0);
}
/*
* Callback to print out the iostats for the given pool.
*/
static int
print_iostat(zpool_handle_t *zhp, void *data)
{
iostat_cbdata_t *cb = data;
nvlist_t *oldconfig, *newconfig;
nvlist_t *oldnvroot, *newnvroot;
int ret;
newconfig = zpool_get_config(zhp, &oldconfig);
if (cb->cb_iteration == 1)
oldconfig = NULL;
verify(nvlist_lookup_nvlist(newconfig, ZPOOL_CONFIG_VDEV_TREE,
&newnvroot) == 0);
if (oldconfig == NULL)
oldnvroot = NULL;
else
verify(nvlist_lookup_nvlist(oldconfig, ZPOOL_CONFIG_VDEV_TREE,
&oldnvroot) == 0);
ret = print_vdev_stats(zhp, zpool_get_name(zhp), oldnvroot, newnvroot,
cb, 0);
if ((ret != 0) && !(cb->cb_flags & IOS_ANYHISTO_M) &&
!cb->cb_scripted && cb->cb_verbose &&
!cb->cb_vdevs.cb_names_count) {
print_iostat_separator(cb);
if (cb->vcdl != NULL) {
print_cmd_columns(cb->vcdl, 1);
}
printf("\n");
}
return (ret);
}
static int
get_columns(void)
{
struct winsize ws;
int columns = 80;
int error;
if (isatty(STDOUT_FILENO)) {
error = ioctl(STDOUT_FILENO, TIOCGWINSZ, &ws);
if (error == 0)
columns = ws.ws_col;
} else {
columns = 999;
}
return (columns);
}
/*
* Return the required length of the pool/vdev name column. The minimum
* allowed width and output formatting flags must be provided.
*/
static int
get_namewidth(zpool_handle_t *zhp, int min_width, int flags, boolean_t verbose)
{
nvlist_t *config, *nvroot;
int width = min_width;
if ((config = zpool_get_config(zhp, NULL)) != NULL) {
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
size_t poolname_len = strlen(zpool_get_name(zhp));
if (verbose == B_FALSE) {
width = MAX(poolname_len, min_width);
} else {
width = MAX(poolname_len,
max_width(zhp, nvroot, 0, min_width, flags));
}
}
return (width);
}
/*
* Parse the input string, get the 'interval' and 'count' value if there is one.
*/
static void
get_interval_count(int *argcp, char **argv, float *iv,
unsigned long *cnt)
{
float interval = 0;
unsigned long count = 0;
int argc = *argcp;
/*
* Determine if the last argument is an integer or a pool name
*/
if (argc > 0 && zfs_isnumber(argv[argc - 1])) {
char *end;
errno = 0;
interval = strtof(argv[argc - 1], &end);
if (*end == '\0' && errno == 0) {
if (interval == 0) {
(void) fprintf(stderr, gettext(
"interval cannot be zero\n"));
usage(B_FALSE);
}
/*
* Ignore the last parameter
*/
argc--;
} else {
/*
* If this is not a valid number, just plow on. The
* user will get a more informative error message later
* on.
*/
interval = 0;
}
}
/*
* If the last argument is also an integer, then we have both a count
* and an interval.
*/
if (argc > 0 && zfs_isnumber(argv[argc - 1])) {
char *end;
errno = 0;
count = interval;
interval = strtof(argv[argc - 1], &end);
if (*end == '\0' && errno == 0) {
if (interval == 0) {
(void) fprintf(stderr, gettext(
"interval cannot be zero\n"));
usage(B_FALSE);
}
/*
* Ignore the last parameter
*/
argc--;
} else {
interval = 0;
}
}
*iv = interval;
*cnt = count;
*argcp = argc;
}
static void
get_timestamp_arg(char c)
{
if (c == 'u')
timestamp_fmt = UDATE;
else if (c == 'd')
timestamp_fmt = DDATE;
else
usage(B_FALSE);
}
/*
* Return stat flags that are supported by all pools by both the module and
* zpool iostat. "*data" should be initialized to all 0xFFs before running.
* It will get ANDed down until only the flags that are supported on all pools
* remain.
*/
static int
get_stat_flags_cb(zpool_handle_t *zhp, void *data)
{
uint64_t *mask = data;
nvlist_t *config, *nvroot, *nvx;
uint64_t flags = 0;
int i, j;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
/* Default stats are always supported, but for completeness.. */
if (nvlist_exists(nvroot, ZPOOL_CONFIG_VDEV_STATS))
flags |= IOS_DEFAULT_M;
/* Get our extended stats nvlist from the main list */
if (nvlist_lookup_nvlist(nvroot, ZPOOL_CONFIG_VDEV_STATS_EX,
&nvx) != 0) {
/*
* No extended stats; they're probably running an older
* module. No big deal, we support that too.
*/
goto end;
}
/* For each extended stat, make sure all its nvpairs are supported */
for (j = 0; j < ARRAY_SIZE(vsx_type_to_nvlist); j++) {
if (!vsx_type_to_nvlist[j][0])
continue;
/* Start off by assuming the flag is supported, then check */
flags |= (1ULL << j);
for (i = 0; vsx_type_to_nvlist[j][i]; i++) {
if (!nvlist_exists(nvx, vsx_type_to_nvlist[j][i])) {
/* flag isn't supported */
flags = flags & ~(1ULL << j);
break;
}
}
}
end:
*mask = *mask & flags;
return (0);
}
/*
* Return a bitmask of stats that are supported on all pools by both the module
* and zpool iostat.
*/
static uint64_t
get_stat_flags(zpool_list_t *list)
{
uint64_t mask = -1;
/*
* get_stat_flags_cb() will lop off bits from "mask" until only the
* flags that are supported on all pools remain.
*/
pool_list_iter(list, B_FALSE, get_stat_flags_cb, &mask);
return (mask);
}
/*
* Return 1 if cb_data->cb_names[0] is this vdev's name, 0 otherwise.
*/
static int
is_vdev_cb(void *zhp_data, nvlist_t *nv, void *cb_data)
{
uint64_t guid;
vdev_cbdata_t *cb = cb_data;
zpool_handle_t *zhp = zhp_data;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (0);
return (guid == zpool_vdev_path_to_guid(zhp, cb->cb_names[0]));
}
/*
* Returns 1 if cb_data->cb_names[0] is a vdev name, 0 otherwise.
*/
static int
is_vdev(zpool_handle_t *zhp, void *cb_data)
{
return (for_each_vdev(zhp, is_vdev_cb, cb_data));
}
/*
* Check if vdevs are in a pool
*
* Return 1 if all argv[] strings are vdev names in pool "pool_name". Otherwise
* return 0. If pool_name is NULL, then search all pools.
*/
static int
are_vdevs_in_pool(int argc, char **argv, char *pool_name,
vdev_cbdata_t *cb)
{
char **tmp_name;
int ret = 0;
int i;
int pool_count = 0;
if ((argc == 0) || !*argv)
return (0);
if (pool_name)
pool_count = 1;
/* Temporarily hijack cb_names for a second... */
tmp_name = cb->cb_names;
/* Go though our list of prospective vdev names */
for (i = 0; i < argc; i++) {
cb->cb_names = argv + i;
/* Is this name a vdev in our pools? */
ret = for_each_pool(pool_count, &pool_name, B_TRUE, NULL,
ZFS_TYPE_POOL, B_FALSE, is_vdev, cb);
if (!ret) {
/* No match */
break;
}
}
cb->cb_names = tmp_name;
return (ret);
}
static int
is_pool_cb(zpool_handle_t *zhp, void *data)
{
char *name = data;
if (strcmp(name, zpool_get_name(zhp)) == 0)
return (1);
return (0);
}
/*
* Do we have a pool named *name? If so, return 1, otherwise 0.
*/
static int
is_pool(char *name)
{
return (for_each_pool(0, NULL, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE,
is_pool_cb, name));
}
/* Are all our argv[] strings pool names? If so return 1, 0 otherwise. */
static int
are_all_pools(int argc, char **argv)
{
if ((argc == 0) || !*argv)
return (0);
while (--argc >= 0)
if (!is_pool(argv[argc]))
return (0);
return (1);
}
/*
* Helper function to print out vdev/pool names we can't resolve. Used for an
* error message.
*/
static void
error_list_unresolved_vdevs(int argc, char **argv, char *pool_name,
vdev_cbdata_t *cb)
{
int i;
char *name;
char *str;
for (i = 0; i < argc; i++) {
name = argv[i];
if (is_pool(name))
str = gettext("pool");
else if (are_vdevs_in_pool(1, &name, pool_name, cb))
str = gettext("vdev in this pool");
else if (are_vdevs_in_pool(1, &name, NULL, cb))
str = gettext("vdev in another pool");
else
str = gettext("unknown");
fprintf(stderr, "\t%s (%s)\n", name, str);
}
}
/*
* Same as get_interval_count(), but with additional checks to not misinterpret
* guids as interval/count values. Assumes VDEV_NAME_GUID is set in
* cb.cb_vdevs.cb_name_flags.
*/
static void
get_interval_count_filter_guids(int *argc, char **argv, float *interval,
unsigned long *count, iostat_cbdata_t *cb)
{
char **tmpargv = argv;
int argc_for_interval = 0;
/* Is the last arg an interval value? Or a guid? */
if (*argc >= 1 && !are_vdevs_in_pool(1, &argv[*argc - 1], NULL,
&cb->cb_vdevs)) {
/*
* The last arg is not a guid, so it's probably an
* interval value.
*/
argc_for_interval++;
if (*argc >= 2 &&
!are_vdevs_in_pool(1, &argv[*argc - 2], NULL,
&cb->cb_vdevs)) {
/*
* The 2nd to last arg is not a guid, so it's probably
* an interval value.
*/
argc_for_interval++;
}
}
/* Point to our list of possible intervals */
tmpargv = &argv[*argc - argc_for_interval];
*argc = *argc - argc_for_interval;
get_interval_count(&argc_for_interval, tmpargv,
interval, count);
}
/*
* Floating point sleep(). Allows you to pass in a floating point value for
* seconds.
*/
static void
fsleep(float sec)
{
struct timespec req;
req.tv_sec = floor(sec);
req.tv_nsec = (sec - (float)req.tv_sec) * NANOSEC;
nanosleep(&req, NULL);
}
/*
* Terminal height, in rows. Returns -1 if stdout is not connected to a TTY or
* if we were unable to determine its size.
*/
static int
terminal_height(void)
{
struct winsize win;
if (isatty(STDOUT_FILENO) == 0)
return (-1);
if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &win) != -1 && win.ws_row > 0)
return (win.ws_row);
return (-1);
}
/*
* Run one of the zpool status/iostat -c scripts with the help (-h) option and
* print the result.
*
* name: Short name of the script ('iostat').
* path: Full path to the script ('/usr/local/etc/zfs/zpool.d/iostat');
*/
static void
print_zpool_script_help(char *name, char *path)
{
char *argv[] = {path, (char *)"-h", NULL};
char **lines = NULL;
int lines_cnt = 0;
int rc;
rc = libzfs_run_process_get_stdout_nopath(path, argv, NULL, &lines,
&lines_cnt);
if (rc != 0 || lines == NULL || lines_cnt <= 0) {
if (lines != NULL)
libzfs_free_str_array(lines, lines_cnt);
return;
}
for (int i = 0; i < lines_cnt; i++)
if (!is_blank_str(lines[i]))
printf(" %-14s %s\n", name, lines[i]);
libzfs_free_str_array(lines, lines_cnt);
}
/*
* Go though the zpool status/iostat -c scripts in the user's path, run their
* help option (-h), and print out the results.
*/
static void
print_zpool_dir_scripts(char *dirpath)
{
DIR *dir;
struct dirent *ent;
char fullpath[MAXPATHLEN];
struct stat dir_stat;
if ((dir = opendir(dirpath)) != NULL) {
/* print all the files and directories within directory */
while ((ent = readdir(dir)) != NULL) {
if (snprintf(fullpath, sizeof (fullpath), "%s/%s",
dirpath, ent->d_name) >= sizeof (fullpath)) {
(void) fprintf(stderr,
gettext("internal error: "
"ZPOOL_SCRIPTS_PATH too large.\n"));
exit(1);
}
/* Print the scripts */
if (stat(fullpath, &dir_stat) == 0)
if (dir_stat.st_mode & S_IXUSR &&
S_ISREG(dir_stat.st_mode))
print_zpool_script_help(ent->d_name,
fullpath);
}
closedir(dir);
}
}
/*
* Print out help text for all zpool status/iostat -c scripts.
*/
static void
print_zpool_script_list(const char *subcommand)
{
char *dir, *sp, *tmp;
printf(gettext("Available 'zpool %s -c' commands:\n"), subcommand);
sp = zpool_get_cmd_search_path();
if (sp == NULL)
return;
for (dir = strtok_r(sp, ":", &tmp);
dir != NULL;
dir = strtok_r(NULL, ":", &tmp))
print_zpool_dir_scripts(dir);
free(sp);
}
/*
* Set the minimum pool/vdev name column width. The width must be at least 10,
* but may be as large as the column width - 42 so it still fits on one line.
* NOTE: 42 is the width of the default capacity/operations/bandwidth output
*/
static int
get_namewidth_iostat(zpool_handle_t *zhp, void *data)
{
iostat_cbdata_t *cb = data;
int width, available_width;
/*
* get_namewidth() returns the maximum width of any name in that column
* for any pool/vdev/device line that will be output.
*/
width = get_namewidth(zhp, cb->cb_namewidth,
cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID, cb->cb_verbose);
/*
* The width we are calculating is the width of the header and also the
* padding width for names that are less than maximum width. The stats
* take up 42 characters, so the width available for names is:
*/
available_width = get_columns() - 42;
/*
* If the maximum width fits on a screen, then great! Make everything
* line up by justifying all lines to the same width. If that max
* width is larger than what's available, the name plus stats won't fit
* on one line, and justifying to that width would cause every line to
* wrap on the screen. We only want lines with long names to wrap.
* Limit the padding to what won't wrap.
*/
if (width > available_width)
width = available_width;
/*
* And regardless of whatever the screen width is (get_columns can
* return 0 if the width is not known or less than 42 for a narrow
* terminal) have the width be a minimum of 10.
*/
if (width < 10)
width = 10;
/* Save the calculated width */
cb->cb_namewidth = width;
return (0);
}
/*
* zpool iostat [[-c [script1,script2,...]] [-lq]|[-rw]] [-ghHLpPvy] [-n name]
* [-T d|u] [[ pool ...]|[pool vdev ...]|[vdev ...]]
* [interval [count]]
*
* -c CMD For each vdev, run command CMD
* -g Display guid for individual vdev name.
* -L Follow links when resolving vdev path name.
* -P Display full path for vdev name.
* -v Display statistics for individual vdevs
* -h Display help
* -p Display values in parsable (exact) format.
* -H Scripted mode. Don't display headers, and separate properties
* by a single tab.
* -l Display average latency
* -q Display queue depths
* -w Display latency histograms
* -r Display request size histogram
* -T Display a timestamp in date(1) or Unix format
* -n Only print headers once
*
* This command can be tricky because we want to be able to deal with pool
* creation/destruction as well as vdev configuration changes. The bulk of this
* processing is handled by the pool_list_* routines in zpool_iter.c. We rely
* on pool_list_update() to detect the addition of new pools. Configuration
* changes are all handled within libzfs.
*/
int
zpool_do_iostat(int argc, char **argv)
{
int c;
int ret;
int npools;
float interval = 0;
unsigned long count = 0;
int winheight = 24;
zpool_list_t *list;
boolean_t verbose = B_FALSE;
boolean_t latency = B_FALSE, l_histo = B_FALSE, rq_histo = B_FALSE;
boolean_t queues = B_FALSE, parsable = B_FALSE, scripted = B_FALSE;
boolean_t omit_since_boot = B_FALSE;
boolean_t guid = B_FALSE;
boolean_t follow_links = B_FALSE;
boolean_t full_name = B_FALSE;
boolean_t headers_once = B_FALSE;
iostat_cbdata_t cb = { 0 };
char *cmd = NULL;
/* Used for printing error message */
const char flag_to_arg[] = {[IOS_LATENCY] = 'l', [IOS_QUEUES] = 'q',
[IOS_L_HISTO] = 'w', [IOS_RQ_HISTO] = 'r'};
uint64_t unsupported_flags;
/* check options */
while ((c = getopt(argc, argv, "c:gLPT:vyhplqrwnH")) != -1) {
switch (c) {
case 'c':
if (cmd != NULL) {
fprintf(stderr,
gettext("Can't set -c flag twice\n"));
exit(1);
}
if (getenv("ZPOOL_SCRIPTS_ENABLED") != NULL &&
!libzfs_envvar_is_set("ZPOOL_SCRIPTS_ENABLED")) {
fprintf(stderr, gettext(
"Can't run -c, disabled by "
"ZPOOL_SCRIPTS_ENABLED.\n"));
exit(1);
}
if ((getuid() <= 0 || geteuid() <= 0) &&
!libzfs_envvar_is_set("ZPOOL_SCRIPTS_AS_ROOT")) {
fprintf(stderr, gettext(
"Can't run -c with root privileges "
"unless ZPOOL_SCRIPTS_AS_ROOT is set.\n"));
exit(1);
}
cmd = optarg;
verbose = B_TRUE;
break;
case 'g':
guid = B_TRUE;
break;
case 'L':
follow_links = B_TRUE;
break;
case 'P':
full_name = B_TRUE;
break;
case 'T':
get_timestamp_arg(*optarg);
break;
case 'v':
verbose = B_TRUE;
break;
case 'p':
parsable = B_TRUE;
break;
case 'l':
latency = B_TRUE;
break;
case 'q':
queues = B_TRUE;
break;
case 'H':
scripted = B_TRUE;
break;
case 'w':
l_histo = B_TRUE;
break;
case 'r':
rq_histo = B_TRUE;
break;
case 'y':
omit_since_boot = B_TRUE;
break;
case 'n':
headers_once = B_TRUE;
break;
case 'h':
usage(B_FALSE);
break;
case '?':
if (optopt == 'c') {
print_zpool_script_list("iostat");
exit(0);
} else {
fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
}
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
cb.cb_literal = parsable;
cb.cb_scripted = scripted;
if (guid)
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_GUID;
if (follow_links)
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS;
if (full_name)
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_PATH;
cb.cb_iteration = 0;
cb.cb_namewidth = 0;
cb.cb_verbose = verbose;
/* Get our interval and count values (if any) */
if (guid) {
get_interval_count_filter_guids(&argc, argv, &interval,
&count, &cb);
} else {
get_interval_count(&argc, argv, &interval, &count);
}
if (argc == 0) {
/* No args, so just print the defaults. */
} else if (are_all_pools(argc, argv)) {
/* All the args are pool names */
} else if (are_vdevs_in_pool(argc, argv, NULL, &cb.cb_vdevs)) {
/* All the args are vdevs */
cb.cb_vdevs.cb_names = argv;
cb.cb_vdevs.cb_names_count = argc;
argc = 0; /* No pools to process */
} else if (are_all_pools(1, argv)) {
/* The first arg is a pool name */
if (are_vdevs_in_pool(argc - 1, argv + 1, argv[0],
&cb.cb_vdevs)) {
/* ...and the rest are vdev names */
cb.cb_vdevs.cb_names = argv + 1;
cb.cb_vdevs.cb_names_count = argc - 1;
argc = 1; /* One pool to process */
} else {
fprintf(stderr, gettext("Expected either a list of "));
fprintf(stderr, gettext("pools, or list of vdevs in"));
fprintf(stderr, " \"%s\", ", argv[0]);
fprintf(stderr, gettext("but got:\n"));
error_list_unresolved_vdevs(argc - 1, argv + 1,
argv[0], &cb.cb_vdevs);
fprintf(stderr, "\n");
usage(B_FALSE);
return (1);
}
} else {
/*
* The args don't make sense. The first arg isn't a pool name,
* nor are all the args vdevs.
*/
fprintf(stderr, gettext("Unable to parse pools/vdevs list.\n"));
fprintf(stderr, "\n");
return (1);
}
if (cb.cb_vdevs.cb_names_count != 0) {
/*
* If user specified vdevs, it implies verbose.
*/
cb.cb_verbose = B_TRUE;
}
/*
* Construct the list of all interesting pools.
*/
ret = 0;
if ((list = pool_list_get(argc, argv, NULL, ZFS_TYPE_POOL, parsable,
&ret)) == NULL)
return (1);
if (pool_list_count(list) == 0 && argc != 0) {
pool_list_free(list);
return (1);
}
if (pool_list_count(list) == 0 && interval == 0) {
pool_list_free(list);
(void) fprintf(stderr, gettext("no pools available\n"));
return (1);
}
if ((l_histo || rq_histo) && (cmd != NULL || latency || queues)) {
pool_list_free(list);
(void) fprintf(stderr,
gettext("[-r|-w] isn't allowed with [-c|-l|-q]\n"));
usage(B_FALSE);
return (1);
}
if (l_histo && rq_histo) {
pool_list_free(list);
(void) fprintf(stderr,
gettext("Only one of [-r|-w] can be passed at a time\n"));
usage(B_FALSE);
return (1);
}
/*
* Enter the main iostat loop.
*/
cb.cb_list = list;
if (l_histo) {
/*
* Histograms tables look out of place when you try to display
* them with the other stats, so make a rule that you can only
* print histograms by themselves.
*/
cb.cb_flags = IOS_L_HISTO_M;
} else if (rq_histo) {
cb.cb_flags = IOS_RQ_HISTO_M;
} else {
cb.cb_flags = IOS_DEFAULT_M;
if (latency)
cb.cb_flags |= IOS_LATENCY_M;
if (queues)
cb.cb_flags |= IOS_QUEUES_M;
}
/*
* See if the module supports all the stats we want to display.
*/
unsupported_flags = cb.cb_flags & ~get_stat_flags(list);
if (unsupported_flags) {
uint64_t f;
int idx;
fprintf(stderr,
gettext("The loaded zfs module doesn't support:"));
/* for each bit set in unsupported_flags */
for (f = unsupported_flags; f; f &= ~(1ULL << idx)) {
idx = lowbit64(f) - 1;
fprintf(stderr, " -%c", flag_to_arg[idx]);
}
fprintf(stderr, ". Try running a newer module.\n");
pool_list_free(list);
return (1);
}
for (;;) {
if ((npools = pool_list_count(list)) == 0)
(void) fprintf(stderr, gettext("no pools available\n"));
else {
/*
* If this is the first iteration and -y was supplied
* we skip any printing.
*/
boolean_t skip = (omit_since_boot &&
cb.cb_iteration == 0);
/*
* Refresh all statistics. This is done as an
* explicit step before calculating the maximum name
* width, so that any * configuration changes are
* properly accounted for.
*/
(void) pool_list_iter(list, B_FALSE, refresh_iostat,
&cb);
/*
* Iterate over all pools to determine the maximum width
* for the pool / device name column across all pools.
*/
cb.cb_namewidth = 0;
(void) pool_list_iter(list, B_FALSE,
get_namewidth_iostat, &cb);
if (timestamp_fmt != NODATE)
print_timestamp(timestamp_fmt);
if (cmd != NULL && cb.cb_verbose &&
!(cb.cb_flags & IOS_ANYHISTO_M)) {
cb.vcdl = all_pools_for_each_vdev_run(argc,
argv, cmd, g_zfs, cb.cb_vdevs.cb_names,
cb.cb_vdevs.cb_names_count,
cb.cb_vdevs.cb_name_flags);
} else {
cb.vcdl = NULL;
}
/*
* Check terminal size so we can print headers
* even when terminal window has its height
* changed.
*/
winheight = terminal_height();
/*
* Are we connected to TTY? If not, headers_once
* should be true, to avoid breaking scripts.
*/
if (winheight < 0)
headers_once = B_TRUE;
/*
* If it's the first time and we're not skipping it,
* or either skip or verbose mode, print the header.
*
* The histogram code explicitly prints its header on
* every vdev, so skip this for histograms.
*/
if (((++cb.cb_iteration == 1 && !skip) ||
(skip != verbose) ||
(!headers_once &&
(cb.cb_iteration % winheight) == 0)) &&
(!(cb.cb_flags & IOS_ANYHISTO_M)) &&
!cb.cb_scripted)
print_iostat_header(&cb);
if (skip) {
(void) fsleep(interval);
continue;
}
pool_list_iter(list, B_FALSE, print_iostat, &cb);
/*
* If there's more than one pool, and we're not in
* verbose mode (which prints a separator for us),
* then print a separator.
*
* In addition, if we're printing specific vdevs then
* we also want an ending separator.
*/
if (((npools > 1 && !verbose &&
!(cb.cb_flags & IOS_ANYHISTO_M)) ||
(!(cb.cb_flags & IOS_ANYHISTO_M) &&
cb.cb_vdevs.cb_names_count)) &&
!cb.cb_scripted) {
print_iostat_separator(&cb);
if (cb.vcdl != NULL)
print_cmd_columns(cb.vcdl, 1);
printf("\n");
}
if (cb.vcdl != NULL)
free_vdev_cmd_data_list(cb.vcdl);
}
/*
* Flush the output so that redirection to a file isn't buffered
* indefinitely.
*/
(void) fflush(stdout);
if (interval == 0)
break;
if (count != 0 && --count == 0)
break;
(void) fsleep(interval);
}
pool_list_free(list);
return (ret);
}
typedef struct list_cbdata {
boolean_t cb_verbose;
int cb_name_flags;
int cb_namewidth;
boolean_t cb_scripted;
zprop_list_t *cb_proplist;
boolean_t cb_literal;
} list_cbdata_t;
/*
* Given a list of columns to display, output appropriate headers for each one.
*/
static void
print_header(list_cbdata_t *cb)
{
zprop_list_t *pl = cb->cb_proplist;
char headerbuf[ZPOOL_MAXPROPLEN];
const char *header;
boolean_t first = B_TRUE;
boolean_t right_justify;
size_t width = 0;
for (; pl != NULL; pl = pl->pl_next) {
width = pl->pl_width;
if (first && cb->cb_verbose) {
/*
* Reset the width to accommodate the verbose listing
* of devices.
*/
width = cb->cb_namewidth;
}
if (!first)
(void) fputs(" ", stdout);
else
first = B_FALSE;
right_justify = B_FALSE;
if (pl->pl_prop != ZPROP_USERPROP) {
header = zpool_prop_column_name(pl->pl_prop);
right_justify = zpool_prop_align_right(pl->pl_prop);
} else {
int i;
for (i = 0; pl->pl_user_prop[i] != '\0'; i++)
headerbuf[i] = toupper(pl->pl_user_prop[i]);
headerbuf[i] = '\0';
header = headerbuf;
}
if (pl->pl_next == NULL && !right_justify)
(void) fputs(header, stdout);
else if (right_justify)
(void) printf("%*s", (int)width, header);
else
(void) printf("%-*s", (int)width, header);
}
(void) fputc('\n', stdout);
}
/*
* Given a pool and a list of properties, print out all the properties according
* to the described layout. Used by zpool_do_list().
*/
static void
print_pool(zpool_handle_t *zhp, list_cbdata_t *cb)
{
zprop_list_t *pl = cb->cb_proplist;
boolean_t first = B_TRUE;
char property[ZPOOL_MAXPROPLEN];
const char *propstr;
boolean_t right_justify;
size_t width;
for (; pl != NULL; pl = pl->pl_next) {
width = pl->pl_width;
if (first && cb->cb_verbose) {
/*
* Reset the width to accommodate the verbose listing
* of devices.
*/
width = cb->cb_namewidth;
}
if (!first) {
if (cb->cb_scripted)
(void) fputc('\t', stdout);
else
(void) fputs(" ", stdout);
} else {
first = B_FALSE;
}
right_justify = B_FALSE;
if (pl->pl_prop != ZPROP_USERPROP) {
if (zpool_get_prop(zhp, pl->pl_prop, property,
sizeof (property), NULL, cb->cb_literal) != 0)
propstr = "-";
else
propstr = property;
right_justify = zpool_prop_align_right(pl->pl_prop);
} else if ((zpool_prop_feature(pl->pl_user_prop) ||
zpool_prop_unsupported(pl->pl_user_prop)) &&
zpool_prop_get_feature(zhp, pl->pl_user_prop, property,
sizeof (property)) == 0) {
propstr = property;
} else if (zfs_prop_user(pl->pl_user_prop) &&
zpool_get_userprop(zhp, pl->pl_user_prop, property,
sizeof (property), NULL) == 0) {
propstr = property;
} else {
propstr = "-";
}
/*
* If this is being called in scripted mode, or if this is the
* last column and it is left-justified, don't include a width
* format specifier.
*/
if (cb->cb_scripted || (pl->pl_next == NULL && !right_justify))
(void) fputs(propstr, stdout);
else if (right_justify)
(void) printf("%*s", (int)width, propstr);
else
(void) printf("%-*s", (int)width, propstr);
}
(void) fputc('\n', stdout);
}
static void
print_one_column(zpool_prop_t prop, uint64_t value, const char *str,
boolean_t scripted, boolean_t valid, enum zfs_nicenum_format format)
{
char propval[64];
boolean_t fixed;
size_t width = zprop_width(prop, &fixed, ZFS_TYPE_POOL);
switch (prop) {
case ZPOOL_PROP_SIZE:
case ZPOOL_PROP_EXPANDSZ:
case ZPOOL_PROP_CHECKPOINT:
case ZPOOL_PROP_DEDUPRATIO:
if (value == 0)
(void) strlcpy(propval, "-", sizeof (propval));
else
zfs_nicenum_format(value, propval, sizeof (propval),
format);
break;
case ZPOOL_PROP_FRAGMENTATION:
if (value == ZFS_FRAG_INVALID) {
(void) strlcpy(propval, "-", sizeof (propval));
} else if (format == ZFS_NICENUM_RAW) {
(void) snprintf(propval, sizeof (propval), "%llu",
(unsigned long long)value);
} else {
(void) snprintf(propval, sizeof (propval), "%llu%%",
(unsigned long long)value);
}
break;
case ZPOOL_PROP_CAPACITY:
/* capacity value is in parts-per-10,000 (aka permyriad) */
if (format == ZFS_NICENUM_RAW)
(void) snprintf(propval, sizeof (propval), "%llu",
(unsigned long long)value / 100);
else
(void) snprintf(propval, sizeof (propval),
value < 1000 ? "%1.2f%%" : value < 10000 ?
"%2.1f%%" : "%3.0f%%", value / 100.0);
break;
case ZPOOL_PROP_HEALTH:
width = 8;
(void) strlcpy(propval, str, sizeof (propval));
break;
default:
zfs_nicenum_format(value, propval, sizeof (propval), format);
}
if (!valid)
(void) strlcpy(propval, "-", sizeof (propval));
if (scripted)
(void) printf("\t%s", propval);
else
(void) printf(" %*s", (int)width, propval);
}
/*
* print static default line per vdev
* not compatible with '-o' <proplist> option
*/
static void
print_list_stats(zpool_handle_t *zhp, const char *name, nvlist_t *nv,
list_cbdata_t *cb, int depth, boolean_t isspare)
{
nvlist_t **child;
vdev_stat_t *vs;
uint_t c, children;
char *vname;
boolean_t scripted = cb->cb_scripted;
uint64_t islog = B_FALSE;
const char *dashes = "%-*s - - - - "
"- - - - -\n";
verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &c) == 0);
if (name != NULL) {
boolean_t toplevel = (vs->vs_space != 0);
uint64_t cap;
enum zfs_nicenum_format format;
const char *state;
if (cb->cb_literal)
format = ZFS_NICENUM_RAW;
else
format = ZFS_NICENUM_1024;
if (strcmp(name, VDEV_TYPE_INDIRECT) == 0)
return;
if (scripted)
(void) printf("\t%s", name);
else if (strlen(name) + depth > cb->cb_namewidth)
(void) printf("%*s%s", depth, "", name);
else
(void) printf("%*s%s%*s", depth, "", name,
(int)(cb->cb_namewidth - strlen(name) - depth), "");
/*
* Print the properties for the individual vdevs. Some
* properties are only applicable to toplevel vdevs. The
* 'toplevel' boolean value is passed to the print_one_column()
* to indicate that the value is valid.
*/
if (VDEV_STAT_VALID(vs_pspace, c) && vs->vs_pspace)
print_one_column(ZPOOL_PROP_SIZE, vs->vs_pspace, NULL,
scripted, B_TRUE, format);
else
print_one_column(ZPOOL_PROP_SIZE, vs->vs_space, NULL,
scripted, toplevel, format);
print_one_column(ZPOOL_PROP_ALLOCATED, vs->vs_alloc, NULL,
scripted, toplevel, format);
print_one_column(ZPOOL_PROP_FREE, vs->vs_space - vs->vs_alloc,
NULL, scripted, toplevel, format);
print_one_column(ZPOOL_PROP_CHECKPOINT,
vs->vs_checkpoint_space, NULL, scripted, toplevel, format);
print_one_column(ZPOOL_PROP_EXPANDSZ, vs->vs_esize, NULL,
scripted, B_TRUE, format);
print_one_column(ZPOOL_PROP_FRAGMENTATION,
vs->vs_fragmentation, NULL, scripted,
(vs->vs_fragmentation != ZFS_FRAG_INVALID && toplevel),
format);
cap = (vs->vs_space == 0) ? 0 :
(vs->vs_alloc * 10000 / vs->vs_space);
print_one_column(ZPOOL_PROP_CAPACITY, cap, NULL,
scripted, toplevel, format);
print_one_column(ZPOOL_PROP_DEDUPRATIO, 0, NULL,
scripted, toplevel, format);
state = zpool_state_to_name(vs->vs_state, vs->vs_aux);
if (isspare) {
if (vs->vs_aux == VDEV_AUX_SPARED)
state = "INUSE";
else if (vs->vs_state == VDEV_STATE_HEALTHY)
state = "AVAIL";
}
print_one_column(ZPOOL_PROP_HEALTH, 0, state, scripted,
B_TRUE, format);
(void) fputc('\n', stdout);
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
return;
/* list the normal vdevs first */
for (c = 0; c < children; c++) {
uint64_t ishole = B_FALSE;
if (nvlist_lookup_uint64(child[c],
ZPOOL_CONFIG_IS_HOLE, &ishole) == 0 && ishole)
continue;
if (nvlist_lookup_uint64(child[c],
ZPOOL_CONFIG_IS_LOG, &islog) == 0 && islog)
continue;
if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS))
continue;
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
print_list_stats(zhp, vname, child[c], cb, depth + 2, B_FALSE);
free(vname);
}
/* list the classes: 'logs', 'dedup', and 'special' */
for (uint_t n = 0; n < ARRAY_SIZE(class_name); n++) {
boolean_t printed = B_FALSE;
for (c = 0; c < children; c++) {
const char *bias = NULL;
const char *type = NULL;
if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&islog) == 0 && islog) {
bias = VDEV_ALLOC_CLASS_LOGS;
} else {
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias);
(void) nvlist_lookup_string(child[c],
ZPOOL_CONFIG_TYPE, &type);
}
if (bias == NULL || strcmp(bias, class_name[n]) != 0)
continue;
if (!islog && strcmp(type, VDEV_TYPE_INDIRECT) == 0)
continue;
if (!printed) {
/* LINTED E_SEC_PRINTF_VAR_FMT */
(void) printf(dashes, cb->cb_namewidth,
class_name[n]);
printed = B_TRUE;
}
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags | VDEV_NAME_TYPE_ID);
print_list_stats(zhp, vname, child[c], cb, depth + 2,
B_FALSE);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE,
&child, &children) == 0 && children > 0) {
/* LINTED E_SEC_PRINTF_VAR_FMT */
(void) printf(dashes, cb->cb_namewidth, "cache");
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags);
print_list_stats(zhp, vname, child[c], cb, depth + 2,
B_FALSE);
free(vname);
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child,
&children) == 0 && children > 0) {
/* LINTED E_SEC_PRINTF_VAR_FMT */
(void) printf(dashes, cb->cb_namewidth, "spare");
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(g_zfs, zhp, child[c],
cb->cb_name_flags);
print_list_stats(zhp, vname, child[c], cb, depth + 2,
B_TRUE);
free(vname);
}
}
}
/*
* Generic callback function to list a pool.
*/
static int
list_callback(zpool_handle_t *zhp, void *data)
{
list_cbdata_t *cbp = data;
print_pool(zhp, cbp);
if (cbp->cb_verbose) {
nvlist_t *config, *nvroot;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
print_list_stats(zhp, NULL, nvroot, cbp, 0, B_FALSE);
}
return (0);
}
/*
* Set the minimum pool/vdev name column width. The width must be at least 9,
* but may be as large as needed.
*/
static int
get_namewidth_list(zpool_handle_t *zhp, void *data)
{
list_cbdata_t *cb = data;
int width;
width = get_namewidth(zhp, cb->cb_namewidth,
cb->cb_name_flags | VDEV_NAME_TYPE_ID, cb->cb_verbose);
if (width < 9)
width = 9;
cb->cb_namewidth = width;
return (0);
}
/*
* zpool list [-gHLpP] [-o prop[,prop]*] [-T d|u] [pool] ... [interval [count]]
*
* -g Display guid for individual vdev name.
* -H Scripted mode. Don't display headers, and separate properties
* by a single tab.
* -L Follow links when resolving vdev path name.
* -o List of properties to display. Defaults to
* "name,size,allocated,free,expandsize,fragmentation,capacity,"
* "dedupratio,health,altroot"
* -p Display values in parsable (exact) format.
* -P Display full path for vdev name.
* -T Display a timestamp in date(1) or Unix format
*
* List all pools in the system, whether or not they're healthy. Output space
* statistics for each one, as well as health status summary.
*/
int
zpool_do_list(int argc, char **argv)
{
int c;
int ret = 0;
list_cbdata_t cb = { 0 };
static char default_props[] =
"name,size,allocated,free,checkpoint,expandsize,fragmentation,"
"capacity,dedupratio,health,altroot";
char *props = default_props;
float interval = 0;
unsigned long count = 0;
zpool_list_t *list;
boolean_t first = B_TRUE;
current_prop_type = ZFS_TYPE_POOL;
/* check options */
while ((c = getopt(argc, argv, ":gHLo:pPT:v")) != -1) {
switch (c) {
case 'g':
cb.cb_name_flags |= VDEV_NAME_GUID;
break;
case 'H':
cb.cb_scripted = B_TRUE;
break;
case 'L':
cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS;
break;
case 'o':
props = optarg;
break;
case 'P':
cb.cb_name_flags |= VDEV_NAME_PATH;
break;
case 'p':
cb.cb_literal = B_TRUE;
break;
case 'T':
get_timestamp_arg(*optarg);
break;
case 'v':
cb.cb_verbose = B_TRUE;
cb.cb_namewidth = 8; /* 8 until precalc is avail */
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
get_interval_count(&argc, argv, &interval, &count);
if (zprop_get_list(g_zfs, props, &cb.cb_proplist, ZFS_TYPE_POOL) != 0)
usage(B_FALSE);
for (;;) {
if ((list = pool_list_get(argc, argv, &cb.cb_proplist,
ZFS_TYPE_POOL, cb.cb_literal, &ret)) == NULL)
return (1);
if (pool_list_count(list) == 0)
break;
cb.cb_namewidth = 0;
(void) pool_list_iter(list, B_FALSE, get_namewidth_list, &cb);
if (timestamp_fmt != NODATE)
print_timestamp(timestamp_fmt);
if (!cb.cb_scripted && (first || cb.cb_verbose)) {
print_header(&cb);
first = B_FALSE;
}
ret = pool_list_iter(list, B_TRUE, list_callback, &cb);
if (interval == 0)
break;
if (count != 0 && --count == 0)
break;
pool_list_free(list);
(void) fsleep(interval);
}
if (argc == 0 && !cb.cb_scripted && pool_list_count(list) == 0) {
(void) printf(gettext("no pools available\n"));
ret = 0;
}
pool_list_free(list);
zprop_free_list(cb.cb_proplist);
return (ret);
}
static int
zpool_do_attach_or_replace(int argc, char **argv, int replacing)
{
boolean_t force = B_FALSE;
boolean_t rebuild = B_FALSE;
boolean_t wait = B_FALSE;
int c;
nvlist_t *nvroot;
char *poolname, *old_disk, *new_disk;
zpool_handle_t *zhp;
nvlist_t *props = NULL;
char *propval;
int ret;
/* check options */
while ((c = getopt(argc, argv, "fo:sw")) != -1) {
switch (c) {
case 'f':
force = B_TRUE;
break;
case 'o':
if ((propval = strchr(optarg, '=')) == NULL) {
(void) fprintf(stderr, gettext("missing "
"'=' for -o option\n"));
usage(B_FALSE);
}
*propval = '\0';
propval++;
if ((strcmp(optarg, ZPOOL_CONFIG_ASHIFT) != 0) ||
(add_prop_list(optarg, propval, &props, B_TRUE)))
usage(B_FALSE);
break;
case 's':
rebuild = B_TRUE;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
poolname = argv[0];
if (argc < 2) {
(void) fprintf(stderr,
gettext("missing <device> specification\n"));
usage(B_FALSE);
}
old_disk = argv[1];
if (argc < 3) {
if (!replacing) {
(void) fprintf(stderr,
gettext("missing <new_device> specification\n"));
usage(B_FALSE);
}
new_disk = old_disk;
argc -= 1;
argv += 1;
} else {
new_disk = argv[2];
argc -= 2;
argv += 2;
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if ((zhp = zpool_open(g_zfs, poolname)) == NULL) {
nvlist_free(props);
return (1);
}
if (zpool_get_config(zhp, NULL) == NULL) {
(void) fprintf(stderr, gettext("pool '%s' is unavailable\n"),
poolname);
zpool_close(zhp);
nvlist_free(props);
return (1);
}
/* unless manually specified use "ashift" pool property (if set) */
if (!nvlist_exists(props, ZPOOL_CONFIG_ASHIFT)) {
int intval;
zprop_source_t src;
char strval[ZPOOL_MAXPROPLEN];
intval = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &src);
if (src != ZPROP_SRC_DEFAULT) {
(void) sprintf(strval, "%" PRId32, intval);
verify(add_prop_list(ZPOOL_CONFIG_ASHIFT, strval,
&props, B_TRUE) == 0);
}
}
nvroot = make_root_vdev(zhp, props, force, B_FALSE, replacing, B_FALSE,
argc, argv);
if (nvroot == NULL) {
zpool_close(zhp);
nvlist_free(props);
return (1);
}
ret = zpool_vdev_attach(zhp, old_disk, new_disk, nvroot, replacing,
rebuild);
if (ret == 0 && wait)
ret = zpool_wait(zhp,
replacing ? ZPOOL_WAIT_REPLACE : ZPOOL_WAIT_RESILVER);
nvlist_free(props);
nvlist_free(nvroot);
zpool_close(zhp);
return (ret);
}
/*
* zpool replace [-fsw] [-o property=value] <pool> <device> <new_device>
*
* -f Force attach, even if <new_device> appears to be in use.
* -s Use sequential instead of healing reconstruction for resilver.
* -o Set property=value.
* -w Wait for replacing to complete before returning
*
* Replace <device> with <new_device>.
*/
int
zpool_do_replace(int argc, char **argv)
{
return (zpool_do_attach_or_replace(argc, argv, B_TRUE));
}
/*
* zpool attach [-fsw] [-o property=value] <pool> <device> <new_device>
*
* -f Force attach, even if <new_device> appears to be in use.
* -s Use sequential instead of healing reconstruction for resilver.
* -o Set property=value.
* -w Wait for resilvering to complete before returning
*
* Attach <new_device> to the mirror containing <device>. If <device> is not
* part of a mirror, then <device> will be transformed into a mirror of
* <device> and <new_device>. In either case, <new_device> will begin life
* with a DTL of [0, now], and will immediately begin to resilver itself.
*/
int
zpool_do_attach(int argc, char **argv)
{
return (zpool_do_attach_or_replace(argc, argv, B_FALSE));
}
/*
* zpool detach [-f] <pool> <device>
*
* -f Force detach of <device>, even if DTLs argue against it
* (not supported yet)
*
* Detach a device from a mirror. The operation will be refused if <device>
* is the last device in the mirror, or if the DTLs indicate that this device
* has the only valid copy of some data.
*/
int
zpool_do_detach(int argc, char **argv)
{
int c;
char *poolname, *path;
zpool_handle_t *zhp;
int ret;
/* check options */
while ((c = getopt(argc, argv, "")) != -1) {
switch (c) {
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr,
gettext("missing <device> specification\n"));
usage(B_FALSE);
}
poolname = argv[0];
path = argv[1];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
ret = zpool_vdev_detach(zhp, path);
zpool_close(zhp);
return (ret);
}
/*
* zpool split [-gLnP] [-o prop=val] ...
* [-o mntopt] ...
* [-R altroot] <pool> <newpool> [<device> ...]
*
* -g Display guid for individual vdev name.
* -L Follow links when resolving vdev path name.
* -n Do not split the pool, but display the resulting layout if
* it were to be split.
* -o Set property=value, or set mount options.
* -P Display full path for vdev name.
* -R Mount the split-off pool under an alternate root.
* -l Load encryption keys while importing.
*
* Splits the named pool and gives it the new pool name. Devices to be split
* off may be listed, provided that no more than one device is specified
* per top-level vdev mirror. The newly split pool is left in an exported
* state unless -R is specified.
*
* Restrictions: the top-level of the pool pool must only be made up of
* mirrors; all devices in the pool must be healthy; no device may be
* undergoing a resilvering operation.
*/
int
zpool_do_split(int argc, char **argv)
{
char *srcpool, *newpool, *propval;
char *mntopts = NULL;
splitflags_t flags;
int c, ret = 0;
boolean_t loadkeys = B_FALSE;
zpool_handle_t *zhp;
nvlist_t *config, *props = NULL;
flags.dryrun = B_FALSE;
flags.import = B_FALSE;
flags.name_flags = 0;
/* check options */
while ((c = getopt(argc, argv, ":gLR:lno:P")) != -1) {
switch (c) {
case 'g':
flags.name_flags |= VDEV_NAME_GUID;
break;
case 'L':
flags.name_flags |= VDEV_NAME_FOLLOW_LINKS;
break;
case 'R':
flags.import = B_TRUE;
if (add_prop_list(
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), optarg,
&props, B_TRUE) != 0) {
nvlist_free(props);
usage(B_FALSE);
}
break;
case 'l':
loadkeys = B_TRUE;
break;
case 'n':
flags.dryrun = B_TRUE;
break;
case 'o':
if ((propval = strchr(optarg, '=')) != NULL) {
*propval = '\0';
propval++;
if (add_prop_list(optarg, propval,
&props, B_TRUE) != 0) {
nvlist_free(props);
usage(B_FALSE);
}
} else {
mntopts = optarg;
}
break;
case 'P':
flags.name_flags |= VDEV_NAME_PATH;
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
break;
}
}
if (!flags.import && mntopts != NULL) {
(void) fprintf(stderr, gettext("setting mntopts is only "
"valid when importing the pool\n"));
usage(B_FALSE);
}
if (!flags.import && loadkeys) {
(void) fprintf(stderr, gettext("loading keys is only "
"valid when importing the pool\n"));
usage(B_FALSE);
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("Missing pool name\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("Missing new pool name\n"));
usage(B_FALSE);
}
srcpool = argv[0];
newpool = argv[1];
argc -= 2;
argv += 2;
if ((zhp = zpool_open(g_zfs, srcpool)) == NULL) {
nvlist_free(props);
return (1);
}
config = split_mirror_vdev(zhp, newpool, props, flags, argc, argv);
if (config == NULL) {
ret = 1;
} else {
if (flags.dryrun) {
(void) printf(gettext("would create '%s' with the "
"following layout:\n\n"), newpool);
print_vdev_tree(NULL, newpool, config, 0, "",
flags.name_flags);
print_vdev_tree(NULL, "dedup", config, 0,
VDEV_ALLOC_BIAS_DEDUP, 0);
print_vdev_tree(NULL, "special", config, 0,
VDEV_ALLOC_BIAS_SPECIAL, 0);
}
}
zpool_close(zhp);
if (ret != 0 || flags.dryrun || !flags.import) {
nvlist_free(config);
nvlist_free(props);
return (ret);
}
/*
* The split was successful. Now we need to open the new
* pool and import it.
*/
if ((zhp = zpool_open_canfail(g_zfs, newpool)) == NULL) {
nvlist_free(config);
nvlist_free(props);
return (1);
}
if (loadkeys) {
ret = zfs_crypto_attempt_load_keys(g_zfs, newpool);
if (ret != 0)
ret = 1;
}
if (zpool_get_state(zhp) != POOL_STATE_UNAVAIL &&
zpool_enable_datasets(zhp, mntopts, 0) != 0) {
ret = 1;
(void) fprintf(stderr, gettext("Split was successful, but "
"the datasets could not all be mounted\n"));
(void) fprintf(stderr, gettext("Try doing '%s' with a "
"different altroot\n"), "zpool import");
}
zpool_close(zhp);
nvlist_free(config);
nvlist_free(props);
return (ret);
}
/*
* zpool online <pool> <device> ...
*/
int
zpool_do_online(int argc, char **argv)
{
int c, i;
char *poolname;
zpool_handle_t *zhp;
int ret = 0;
vdev_state_t newstate;
int flags = 0;
/* check options */
while ((c = getopt(argc, argv, "e")) != -1) {
switch (c) {
case 'e':
flags |= ZFS_ONLINE_EXPAND;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing device name\n"));
usage(B_FALSE);
}
poolname = argv[0];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
for (i = 1; i < argc; i++) {
+ vdev_state_t oldstate;
+ boolean_t avail_spare, l2cache;
+ nvlist_t *tgt = zpool_find_vdev(zhp, argv[i], &avail_spare,
+ &l2cache, NULL);
+ if (tgt == NULL) {
+ ret = 1;
+ continue;
+ }
+ uint_t vsc;
+ oldstate = ((vdev_stat_t *)fnvlist_lookup_uint64_array(tgt,
+ ZPOOL_CONFIG_VDEV_STATS, &vsc))->vs_state;
if (zpool_vdev_online(zhp, argv[i], flags, &newstate) == 0) {
if (newstate != VDEV_STATE_HEALTHY) {
(void) printf(gettext("warning: device '%s' "
"onlined, but remains in faulted state\n"),
argv[i]);
if (newstate == VDEV_STATE_FAULTED)
(void) printf(gettext("use 'zpool "
"clear' to restore a faulted "
"device\n"));
else
(void) printf(gettext("use 'zpool "
"replace' to replace devices "
"that are no longer present\n"));
+ if ((flags & ZFS_ONLINE_EXPAND)) {
+ (void) printf(gettext("%s: failed "
+ "to expand usable space on "
+ "unhealthy device '%s'\n"),
+ (oldstate >= VDEV_STATE_DEGRADED ?
+ "error" : "warning"), argv[i]);
+ if (oldstate >= VDEV_STATE_DEGRADED) {
+ ret = 1;
+ break;
+ }
+ }
}
} else {
ret = 1;
}
}
zpool_close(zhp);
return (ret);
}
/*
* zpool offline [-ft] <pool> <device> ...
*
* -f Force the device into a faulted state.
*
* -t Only take the device off-line temporarily. The offline/faulted
* state will not be persistent across reboots.
*/
int
zpool_do_offline(int argc, char **argv)
{
int c, i;
char *poolname;
zpool_handle_t *zhp;
int ret = 0;
boolean_t istmp = B_FALSE;
boolean_t fault = B_FALSE;
/* check options */
while ((c = getopt(argc, argv, "ft")) != -1) {
switch (c) {
case 'f':
fault = B_TRUE;
break;
case 't':
istmp = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing device name\n"));
usage(B_FALSE);
}
poolname = argv[0];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
for (i = 1; i < argc; i++) {
if (fault) {
uint64_t guid = zpool_vdev_path_to_guid(zhp, argv[i]);
vdev_aux_t aux;
if (istmp == B_FALSE) {
/* Force the fault to persist across imports */
aux = VDEV_AUX_EXTERNAL_PERSIST;
} else {
aux = VDEV_AUX_EXTERNAL;
}
if (guid == 0 || zpool_vdev_fault(zhp, guid, aux) != 0)
ret = 1;
} else {
if (zpool_vdev_offline(zhp, argv[i], istmp) != 0)
ret = 1;
}
}
zpool_close(zhp);
return (ret);
}
/*
* zpool clear <pool> [device]
*
* Clear all errors associated with a pool or a particular device.
*/
int
zpool_do_clear(int argc, char **argv)
{
int c;
int ret = 0;
boolean_t dryrun = B_FALSE;
boolean_t do_rewind = B_FALSE;
boolean_t xtreme_rewind = B_FALSE;
uint32_t rewind_policy = ZPOOL_NO_REWIND;
nvlist_t *policy = NULL;
zpool_handle_t *zhp;
char *pool, *device;
/* check options */
while ((c = getopt(argc, argv, "FnX")) != -1) {
switch (c) {
case 'F':
do_rewind = B_TRUE;
break;
case 'n':
dryrun = B_TRUE;
break;
case 'X':
xtreme_rewind = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc > 2) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if ((dryrun || xtreme_rewind) && !do_rewind) {
(void) fprintf(stderr,
gettext("-n or -X only meaningful with -F\n"));
usage(B_FALSE);
}
if (dryrun)
rewind_policy = ZPOOL_TRY_REWIND;
else if (do_rewind)
rewind_policy = ZPOOL_DO_REWIND;
if (xtreme_rewind)
rewind_policy |= ZPOOL_EXTREME_REWIND;
/* In future, further rewind policy choices can be passed along here */
if (nvlist_alloc(&policy, NV_UNIQUE_NAME, 0) != 0 ||
nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY,
rewind_policy) != 0) {
return (1);
}
pool = argv[0];
device = argc == 2 ? argv[1] : NULL;
if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) {
nvlist_free(policy);
return (1);
}
if (zpool_clear(zhp, device, policy) != 0)
ret = 1;
zpool_close(zhp);
nvlist_free(policy);
return (ret);
}
/*
* zpool reguid <pool>
*/
int
zpool_do_reguid(int argc, char **argv)
{
int c;
char *poolname;
zpool_handle_t *zhp;
int ret = 0;
/* check options */
while ((c = getopt(argc, argv, "")) != -1) {
switch (c) {
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* get pool name and check number of arguments */
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
poolname = argv[0];
if ((zhp = zpool_open(g_zfs, poolname)) == NULL)
return (1);
ret = zpool_reguid(zhp);
zpool_close(zhp);
return (ret);
}
/*
* zpool reopen <pool>
*
* Reopen the pool so that the kernel can update the sizes of all vdevs.
*/
int
zpool_do_reopen(int argc, char **argv)
{
int c;
int ret = 0;
boolean_t scrub_restart = B_TRUE;
/* check options */
while ((c = getopt(argc, argv, "n")) != -1) {
switch (c) {
case 'n':
scrub_restart = B_FALSE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
/* if argc == 0 we will execute zpool_reopen_one on all pools */
ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, zpool_reopen_one, &scrub_restart);
return (ret);
}
typedef struct scrub_cbdata {
int cb_type;
pool_scrub_cmd_t cb_scrub_cmd;
} scrub_cbdata_t;
static boolean_t
zpool_has_checkpoint(zpool_handle_t *zhp)
{
nvlist_t *config, *nvroot;
config = zpool_get_config(zhp, NULL);
if (config != NULL) {
pool_checkpoint_stat_t *pcs = NULL;
uint_t c;
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c);
if (pcs == NULL || pcs->pcs_state == CS_NONE)
return (B_FALSE);
assert(pcs->pcs_state == CS_CHECKPOINT_EXISTS ||
pcs->pcs_state == CS_CHECKPOINT_DISCARDING);
return (B_TRUE);
}
return (B_FALSE);
}
static int
scrub_callback(zpool_handle_t *zhp, void *data)
{
scrub_cbdata_t *cb = data;
int err;
/*
* Ignore faulted pools.
*/
if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) {
(void) fprintf(stderr, gettext("cannot scan '%s': pool is "
"currently unavailable\n"), zpool_get_name(zhp));
return (1);
}
err = zpool_scan(zhp, cb->cb_type, cb->cb_scrub_cmd);
if (err == 0 && zpool_has_checkpoint(zhp) &&
cb->cb_type == POOL_SCAN_SCRUB) {
(void) printf(gettext("warning: will not scrub state that "
"belongs to the checkpoint of pool '%s'\n"),
zpool_get_name(zhp));
}
return (err != 0);
}
static int
wait_callback(zpool_handle_t *zhp, void *data)
{
zpool_wait_activity_t *act = data;
return (zpool_wait(zhp, *act));
}
/*
* zpool scrub [-s | -p] [-w] <pool> ...
*
* -s Stop. Stops any in-progress scrub.
* -p Pause. Pause in-progress scrub.
* -w Wait. Blocks until scrub has completed.
*/
int
zpool_do_scrub(int argc, char **argv)
{
int c;
scrub_cbdata_t cb;
boolean_t wait = B_FALSE;
int error;
cb.cb_type = POOL_SCAN_SCRUB;
cb.cb_scrub_cmd = POOL_SCRUB_NORMAL;
/* check options */
while ((c = getopt(argc, argv, "spw")) != -1) {
switch (c) {
case 's':
cb.cb_type = POOL_SCAN_NONE;
break;
case 'p':
cb.cb_scrub_cmd = POOL_SCRUB_PAUSE;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
if (cb.cb_type == POOL_SCAN_NONE &&
cb.cb_scrub_cmd == POOL_SCRUB_PAUSE) {
(void) fprintf(stderr, gettext("invalid option combination: "
"-s and -p are mutually exclusive\n"));
usage(B_FALSE);
}
if (wait && (cb.cb_type == POOL_SCAN_NONE ||
cb.cb_scrub_cmd == POOL_SCRUB_PAUSE)) {
(void) fprintf(stderr, gettext("invalid option combination: "
"-w cannot be used with -p or -s\n"));
usage(B_FALSE);
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
error = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, scrub_callback, &cb);
if (wait && !error) {
zpool_wait_activity_t act = ZPOOL_WAIT_SCRUB;
error = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, wait_callback, &act);
}
return (error);
}
/*
* zpool resilver <pool> ...
*
* Restarts any in-progress resilver
*/
int
zpool_do_resilver(int argc, char **argv)
{
int c;
scrub_cbdata_t cb;
cb.cb_type = POOL_SCAN_RESILVER;
cb.cb_scrub_cmd = POOL_SCRUB_NORMAL;
/* check options */
while ((c = getopt(argc, argv, "")) != -1) {
switch (c) {
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
}
return (for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, scrub_callback, &cb));
}
/*
* zpool trim [-d] [-r <rate>] [-c | -s] <pool> [<device> ...]
*
* -c Cancel. Ends any in-progress trim.
* -d Secure trim. Requires kernel and device support.
* -r <rate> Sets the TRIM rate in bytes (per second). Supports
* adding a multiplier suffix such as 'k' or 'm'.
* -s Suspend. TRIM can then be restarted with no flags.
* -w Wait. Blocks until trimming has completed.
*/
int
zpool_do_trim(int argc, char **argv)
{
struct option long_options[] = {
{"cancel", no_argument, NULL, 'c'},
{"secure", no_argument, NULL, 'd'},
{"rate", required_argument, NULL, 'r'},
{"suspend", no_argument, NULL, 's'},
{"wait", no_argument, NULL, 'w'},
{0, 0, 0, 0}
};
pool_trim_func_t cmd_type = POOL_TRIM_START;
uint64_t rate = 0;
boolean_t secure = B_FALSE;
boolean_t wait = B_FALSE;
int c;
while ((c = getopt_long(argc, argv, "cdr:sw", long_options, NULL))
!= -1) {
switch (c) {
case 'c':
if (cmd_type != POOL_TRIM_START &&
cmd_type != POOL_TRIM_CANCEL) {
(void) fprintf(stderr, gettext("-c cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_TRIM_CANCEL;
break;
case 'd':
if (cmd_type != POOL_TRIM_START) {
(void) fprintf(stderr, gettext("-d cannot be "
"combined with the -c or -s options\n"));
usage(B_FALSE);
}
secure = B_TRUE;
break;
case 'r':
if (cmd_type != POOL_TRIM_START) {
(void) fprintf(stderr, gettext("-r cannot be "
"combined with the -c or -s options\n"));
usage(B_FALSE);
}
if (zfs_nicestrtonum(g_zfs, optarg, &rate) == -1) {
(void) fprintf(stderr, "%s: %s\n",
gettext("invalid value for rate"),
libzfs_error_description(g_zfs));
usage(B_FALSE);
}
break;
case 's':
if (cmd_type != POOL_TRIM_START &&
cmd_type != POOL_TRIM_SUSPEND) {
(void) fprintf(stderr, gettext("-s cannot be "
"combined with other options\n"));
usage(B_FALSE);
}
cmd_type = POOL_TRIM_SUSPEND;
break;
case 'w':
wait = B_TRUE;
break;
case '?':
if (optopt != 0) {
(void) fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
} else {
(void) fprintf(stderr,
gettext("invalid option '%s'\n"),
argv[optind - 1]);
}
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing pool name argument\n"));
usage(B_FALSE);
return (-1);
}
if (wait && (cmd_type != POOL_TRIM_START)) {
(void) fprintf(stderr, gettext("-w cannot be used with -c or "
"-s\n"));
usage(B_FALSE);
}
char *poolname = argv[0];
zpool_handle_t *zhp = zpool_open(g_zfs, poolname);
if (zhp == NULL)
return (-1);
trimflags_t trim_flags = {
.secure = secure,
.rate = rate,
.wait = wait,
};
nvlist_t *vdevs = fnvlist_alloc();
if (argc == 1) {
/* no individual leaf vdevs specified, so add them all */
nvlist_t *config = zpool_get_config(zhp, NULL);
nvlist_t *nvroot = fnvlist_lookup_nvlist(config,
ZPOOL_CONFIG_VDEV_TREE);
zpool_collect_leaves(zhp, nvroot, vdevs);
trim_flags.fullpool = B_TRUE;
} else {
trim_flags.fullpool = B_FALSE;
for (int i = 1; i < argc; i++) {
fnvlist_add_boolean(vdevs, argv[i]);
}
}
int error = zpool_trim(zhp, cmd_type, vdevs, &trim_flags);
fnvlist_free(vdevs);
zpool_close(zhp);
return (error);
}
/*
* Converts a total number of seconds to a human readable string broken
* down in to days/hours/minutes/seconds.
*/
static void
secs_to_dhms(uint64_t total, char *buf)
{
uint64_t days = total / 60 / 60 / 24;
uint64_t hours = (total / 60 / 60) % 24;
uint64_t mins = (total / 60) % 60;
uint64_t secs = (total % 60);
if (days > 0) {
(void) sprintf(buf, "%llu days %02llu:%02llu:%02llu",
(u_longlong_t)days, (u_longlong_t)hours,
(u_longlong_t)mins, (u_longlong_t)secs);
} else {
(void) sprintf(buf, "%02llu:%02llu:%02llu",
(u_longlong_t)hours, (u_longlong_t)mins,
(u_longlong_t)secs);
}
}
/*
* Print out detailed scrub status.
*/
static void
print_scan_scrub_resilver_status(pool_scan_stat_t *ps)
{
time_t start, end, pause;
uint64_t pass_scanned, scanned, pass_issued, issued, total;
uint64_t elapsed, scan_rate, issue_rate;
double fraction_done;
char processed_buf[7], scanned_buf[7], issued_buf[7], total_buf[7];
char srate_buf[7], irate_buf[7], time_buf[32];
printf(" ");
printf_color(ANSI_BOLD, gettext("scan:"));
printf(" ");
/* If there's never been a scan, there's not much to say. */
if (ps == NULL || ps->pss_func == POOL_SCAN_NONE ||
ps->pss_func >= POOL_SCAN_FUNCS) {
(void) printf(gettext("none requested\n"));
return;
}
start = ps->pss_start_time;
end = ps->pss_end_time;
pause = ps->pss_pass_scrub_pause;
zfs_nicebytes(ps->pss_processed, processed_buf, sizeof (processed_buf));
int is_resilver = ps->pss_func == POOL_SCAN_RESILVER;
int is_scrub = ps->pss_func == POOL_SCAN_SCRUB;
assert(is_resilver || is_scrub);
/* Scan is finished or canceled. */
if (ps->pss_state == DSS_FINISHED) {
secs_to_dhms(end - start, time_buf);
if (is_scrub) {
(void) printf(gettext("scrub repaired %s "
"in %s with %llu errors on %s"), processed_buf,
time_buf, (u_longlong_t)ps->pss_errors,
ctime(&end));
} else if (is_resilver) {
(void) printf(gettext("resilvered %s "
"in %s with %llu errors on %s"), processed_buf,
time_buf, (u_longlong_t)ps->pss_errors,
ctime(&end));
}
return;
} else if (ps->pss_state == DSS_CANCELED) {
if (is_scrub) {
(void) printf(gettext("scrub canceled on %s"),
ctime(&end));
} else if (is_resilver) {
(void) printf(gettext("resilver canceled on %s"),
ctime(&end));
}
return;
}
assert(ps->pss_state == DSS_SCANNING);
/* Scan is in progress. Resilvers can't be paused. */
if (is_scrub) {
if (pause == 0) {
(void) printf(gettext("scrub in progress since %s"),
ctime(&start));
} else {
(void) printf(gettext("scrub paused since %s"),
ctime(&pause));
(void) printf(gettext("\tscrub started on %s"),
ctime(&start));
}
} else if (is_resilver) {
(void) printf(gettext("resilver in progress since %s"),
ctime(&start));
}
scanned = ps->pss_examined;
pass_scanned = ps->pss_pass_exam;
issued = ps->pss_issued;
pass_issued = ps->pss_pass_issued;
total = ps->pss_to_examine;
/* we are only done with a block once we have issued the IO for it */
fraction_done = (double)issued / total;
/* elapsed time for this pass, rounding up to 1 if it's 0 */
elapsed = time(NULL) - ps->pss_pass_start;
elapsed -= ps->pss_pass_scrub_spent_paused;
elapsed = (elapsed != 0) ? elapsed : 1;
scan_rate = pass_scanned / elapsed;
issue_rate = pass_issued / elapsed;
uint64_t total_secs_left = (issue_rate != 0 && total >= issued) ?
((total - issued) / issue_rate) : UINT64_MAX;
secs_to_dhms(total_secs_left, time_buf);
/* format all of the numbers we will be reporting */
zfs_nicebytes(scanned, scanned_buf, sizeof (scanned_buf));
zfs_nicebytes(issued, issued_buf, sizeof (issued_buf));
zfs_nicebytes(total, total_buf, sizeof (total_buf));
zfs_nicebytes(scan_rate, srate_buf, sizeof (srate_buf));
zfs_nicebytes(issue_rate, irate_buf, sizeof (irate_buf));
/* do not print estimated time if we have a paused scrub */
if (pause == 0) {
(void) printf(gettext("\t%s scanned at %s/s, "
"%s issued at %s/s, %s total\n"),
scanned_buf, srate_buf, issued_buf, irate_buf, total_buf);
} else {
(void) printf(gettext("\t%s scanned, %s issued, %s total\n"),
scanned_buf, issued_buf, total_buf);
}
if (is_resilver) {
(void) printf(gettext("\t%s resilvered, %.2f%% done"),
processed_buf, 100 * fraction_done);
} else if (is_scrub) {
(void) printf(gettext("\t%s repaired, %.2f%% done"),
processed_buf, 100 * fraction_done);
}
if (pause == 0) {
/*
* Only provide an estimate iff:
* 1) the time remaining is valid, and
* 2) the issue rate exceeds 10 MB/s, and
* 3) it's either:
* a) a resilver which has started repairs, or
* b) a scrub which has entered the issue phase.
*/
if (total_secs_left != UINT64_MAX &&
issue_rate >= 10 * 1024 * 1024 &&
((is_resilver && ps->pss_processed > 0) ||
(is_scrub && issued > 0))) {
(void) printf(gettext(", %s to go\n"), time_buf);
} else {
(void) printf(gettext(", no estimated "
"completion time\n"));
}
} else {
(void) printf(gettext("\n"));
}
}
static void
print_rebuild_status_impl(vdev_rebuild_stat_t *vrs, char *vdev_name)
{
if (vrs == NULL || vrs->vrs_state == VDEV_REBUILD_NONE)
return;
printf(" ");
printf_color(ANSI_BOLD, gettext("scan:"));
printf(" ");
uint64_t bytes_scanned = vrs->vrs_bytes_scanned;
uint64_t bytes_issued = vrs->vrs_bytes_issued;
uint64_t bytes_rebuilt = vrs->vrs_bytes_rebuilt;
uint64_t bytes_est = vrs->vrs_bytes_est;
uint64_t scan_rate = (vrs->vrs_pass_bytes_scanned /
(vrs->vrs_pass_time_ms + 1)) * 1000;
uint64_t issue_rate = (vrs->vrs_pass_bytes_issued /
(vrs->vrs_pass_time_ms + 1)) * 1000;
double scan_pct = MIN((double)bytes_scanned * 100 /
(bytes_est + 1), 100);
/* Format all of the numbers we will be reporting */
char bytes_scanned_buf[7], bytes_issued_buf[7];
char bytes_rebuilt_buf[7], bytes_est_buf[7];
char scan_rate_buf[7], issue_rate_buf[7], time_buf[32];
zfs_nicebytes(bytes_scanned, bytes_scanned_buf,
sizeof (bytes_scanned_buf));
zfs_nicebytes(bytes_issued, bytes_issued_buf,
sizeof (bytes_issued_buf));
zfs_nicebytes(bytes_rebuilt, bytes_rebuilt_buf,
sizeof (bytes_rebuilt_buf));
zfs_nicebytes(bytes_est, bytes_est_buf, sizeof (bytes_est_buf));
zfs_nicebytes(scan_rate, scan_rate_buf, sizeof (scan_rate_buf));
zfs_nicebytes(issue_rate, issue_rate_buf, sizeof (issue_rate_buf));
time_t start = vrs->vrs_start_time;
time_t end = vrs->vrs_end_time;
/* Rebuild is finished or canceled. */
if (vrs->vrs_state == VDEV_REBUILD_COMPLETE) {
secs_to_dhms(vrs->vrs_scan_time_ms / 1000, time_buf);
(void) printf(gettext("resilvered (%s) %s in %s "
"with %llu errors on %s"), vdev_name, bytes_rebuilt_buf,
time_buf, (u_longlong_t)vrs->vrs_errors, ctime(&end));
return;
} else if (vrs->vrs_state == VDEV_REBUILD_CANCELED) {
(void) printf(gettext("resilver (%s) canceled on %s"),
vdev_name, ctime(&end));
return;
} else if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) {
(void) printf(gettext("resilver (%s) in progress since %s"),
vdev_name, ctime(&start));
}
assert(vrs->vrs_state == VDEV_REBUILD_ACTIVE);
secs_to_dhms(MAX((int64_t)bytes_est - (int64_t)bytes_scanned, 0) /
MAX(scan_rate, 1), time_buf);
(void) printf(gettext("\t%s scanned at %s/s, %s issued %s/s, "
"%s total\n"), bytes_scanned_buf, scan_rate_buf,
bytes_issued_buf, issue_rate_buf, bytes_est_buf);
(void) printf(gettext("\t%s resilvered, %.2f%% done"),
bytes_rebuilt_buf, scan_pct);
if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) {
if (scan_rate >= 10 * 1024 * 1024) {
(void) printf(gettext(", %s to go\n"), time_buf);
} else {
(void) printf(gettext(", no estimated "
"completion time\n"));
}
} else {
(void) printf(gettext("\n"));
}
}
/*
* Print rebuild status for top-level vdevs.
*/
static void
print_rebuild_status(zpool_handle_t *zhp, nvlist_t *nvroot)
{
nvlist_t **child;
uint_t children;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (uint_t c = 0; c < children; c++) {
vdev_rebuild_stat_t *vrs;
uint_t i;
if (nvlist_lookup_uint64_array(child[c],
ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i) == 0) {
char *name = zpool_vdev_name(g_zfs, zhp,
child[c], VDEV_NAME_TYPE_ID);
print_rebuild_status_impl(vrs, name);
free(name);
}
}
}
/*
* As we don't scrub checkpointed blocks, we want to warn the user that we
* skipped scanning some blocks if a checkpoint exists or existed at any
* time during the scan. If a sequential instead of healing reconstruction
* was performed then the blocks were reconstructed. However, their checksums
* have not been verified so we still print the warning.
*/
static void
print_checkpoint_scan_warning(pool_scan_stat_t *ps, pool_checkpoint_stat_t *pcs)
{
if (ps == NULL || pcs == NULL)
return;
if (pcs->pcs_state == CS_NONE ||
pcs->pcs_state == CS_CHECKPOINT_DISCARDING)
return;
assert(pcs->pcs_state == CS_CHECKPOINT_EXISTS);
if (ps->pss_state == DSS_NONE)
return;
if ((ps->pss_state == DSS_FINISHED || ps->pss_state == DSS_CANCELED) &&
ps->pss_end_time < pcs->pcs_start_time)
return;
if (ps->pss_state == DSS_FINISHED || ps->pss_state == DSS_CANCELED) {
(void) printf(gettext(" scan warning: skipped blocks "
"that are only referenced by the checkpoint.\n"));
} else {
assert(ps->pss_state == DSS_SCANNING);
(void) printf(gettext(" scan warning: skipping blocks "
"that are only referenced by the checkpoint.\n"));
}
}
/*
* Returns B_TRUE if there is an active rebuild in progress. Otherwise,
* B_FALSE is returned and 'rebuild_end_time' is set to the end time for
* the last completed (or cancelled) rebuild.
*/
static boolean_t
check_rebuilding(nvlist_t *nvroot, uint64_t *rebuild_end_time)
{
nvlist_t **child;
uint_t children;
boolean_t rebuilding = B_FALSE;
uint64_t end_time = 0;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (uint_t c = 0; c < children; c++) {
vdev_rebuild_stat_t *vrs;
uint_t i;
if (nvlist_lookup_uint64_array(child[c],
ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i) == 0) {
if (vrs->vrs_end_time > end_time)
end_time = vrs->vrs_end_time;
if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) {
rebuilding = B_TRUE;
end_time = 0;
break;
}
}
}
if (rebuild_end_time != NULL)
*rebuild_end_time = end_time;
return (rebuilding);
}
/*
* Print the scan status.
*/
static void
print_scan_status(zpool_handle_t *zhp, nvlist_t *nvroot)
{
uint64_t rebuild_end_time = 0, resilver_end_time = 0;
boolean_t have_resilver = B_FALSE, have_scrub = B_FALSE;
boolean_t active_resilver = B_FALSE;
pool_checkpoint_stat_t *pcs = NULL;
pool_scan_stat_t *ps = NULL;
uint_t c;
if (nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS,
(uint64_t **)&ps, &c) == 0) {
if (ps->pss_func == POOL_SCAN_RESILVER) {
resilver_end_time = ps->pss_end_time;
active_resilver = (ps->pss_state == DSS_SCANNING);
}
have_resilver = (ps->pss_func == POOL_SCAN_RESILVER);
have_scrub = (ps->pss_func == POOL_SCAN_SCRUB);
}
boolean_t active_rebuild = check_rebuilding(nvroot, &rebuild_end_time);
boolean_t have_rebuild = (active_rebuild || (rebuild_end_time > 0));
/* Always print the scrub status when available. */
if (have_scrub)
print_scan_scrub_resilver_status(ps);
/*
* When there is an active resilver or rebuild print its status.
* Otherwise print the status of the last resilver or rebuild.
*/
if (active_resilver || (!active_rebuild && have_resilver &&
resilver_end_time && resilver_end_time > rebuild_end_time)) {
print_scan_scrub_resilver_status(ps);
} else if (active_rebuild || (!active_resilver && have_rebuild &&
rebuild_end_time && rebuild_end_time > resilver_end_time)) {
print_rebuild_status(zhp, nvroot);
}
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c);
print_checkpoint_scan_warning(ps, pcs);
}
/*
* Print out detailed removal status.
*/
static void
print_removal_status(zpool_handle_t *zhp, pool_removal_stat_t *prs)
{
char copied_buf[7], examined_buf[7], total_buf[7], rate_buf[7];
time_t start, end;
nvlist_t *config, *nvroot;
nvlist_t **child;
uint_t children;
char *vdev_name;
if (prs == NULL || prs->prs_state == DSS_NONE)
return;
/*
* Determine name of vdev.
*/
config = zpool_get_config(zhp, NULL);
nvroot = fnvlist_lookup_nvlist(config,
ZPOOL_CONFIG_VDEV_TREE);
verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0);
assert(prs->prs_removing_vdev < children);
vdev_name = zpool_vdev_name(g_zfs, zhp,
child[prs->prs_removing_vdev], B_TRUE);
printf_color(ANSI_BOLD, gettext("remove: "));
start = prs->prs_start_time;
end = prs->prs_end_time;
zfs_nicenum(prs->prs_copied, copied_buf, sizeof (copied_buf));
/*
* Removal is finished or canceled.
*/
if (prs->prs_state == DSS_FINISHED) {
uint64_t minutes_taken = (end - start) / 60;
(void) printf(gettext("Removal of vdev %llu copied %s "
"in %lluh%um, completed on %s"),
(longlong_t)prs->prs_removing_vdev,
copied_buf,
(u_longlong_t)(minutes_taken / 60),
(uint_t)(minutes_taken % 60),
ctime((time_t *)&end));
} else if (prs->prs_state == DSS_CANCELED) {
(void) printf(gettext("Removal of %s canceled on %s"),
vdev_name, ctime(&end));
} else {
uint64_t copied, total, elapsed, mins_left, hours_left;
double fraction_done;
uint_t rate;
assert(prs->prs_state == DSS_SCANNING);
/*
* Removal is in progress.
*/
(void) printf(gettext(
"Evacuation of %s in progress since %s"),
vdev_name, ctime(&start));
copied = prs->prs_copied > 0 ? prs->prs_copied : 1;
total = prs->prs_to_copy;
fraction_done = (double)copied / total;
/* elapsed time for this pass */
elapsed = time(NULL) - prs->prs_start_time;
elapsed = elapsed > 0 ? elapsed : 1;
rate = copied / elapsed;
rate = rate > 0 ? rate : 1;
mins_left = ((total - copied) / rate) / 60;
hours_left = mins_left / 60;
zfs_nicenum(copied, examined_buf, sizeof (examined_buf));
zfs_nicenum(total, total_buf, sizeof (total_buf));
zfs_nicenum(rate, rate_buf, sizeof (rate_buf));
/*
* do not print estimated time if hours_left is more than
* 30 days
*/
(void) printf(gettext(
"\t%s copied out of %s at %s/s, %.2f%% done"),
examined_buf, total_buf, rate_buf, 100 * fraction_done);
if (hours_left < (30 * 24)) {
(void) printf(gettext(", %lluh%um to go\n"),
(u_longlong_t)hours_left, (uint_t)(mins_left % 60));
} else {
(void) printf(gettext(
", (copy is slow, no estimated time)\n"));
}
}
free(vdev_name);
if (prs->prs_mapping_memory > 0) {
char mem_buf[7];
zfs_nicenum(prs->prs_mapping_memory, mem_buf, sizeof (mem_buf));
(void) printf(gettext(
"\t%s memory used for removed device mappings\n"),
mem_buf);
}
}
static void
print_checkpoint_status(pool_checkpoint_stat_t *pcs)
{
time_t start;
char space_buf[7];
if (pcs == NULL || pcs->pcs_state == CS_NONE)
return;
(void) printf(gettext("checkpoint: "));
start = pcs->pcs_start_time;
zfs_nicenum(pcs->pcs_space, space_buf, sizeof (space_buf));
if (pcs->pcs_state == CS_CHECKPOINT_EXISTS) {
char *date = ctime(&start);
/*
* ctime() adds a newline at the end of the generated
* string, thus the weird format specifier and the
* strlen() call used to chop it off from the output.
*/
(void) printf(gettext("created %.*s, consumes %s\n"),
(int)(strlen(date) - 1), date, space_buf);
return;
}
assert(pcs->pcs_state == CS_CHECKPOINT_DISCARDING);
(void) printf(gettext("discarding, %s remaining.\n"),
space_buf);
}
static void
print_error_log(zpool_handle_t *zhp)
{
nvlist_t *nverrlist = NULL;
nvpair_t *elem;
char *pathname;
size_t len = MAXPATHLEN * 2;
if (zpool_get_errlog(zhp, &nverrlist) != 0)
return;
(void) printf("errors: Permanent errors have been "
"detected in the following files:\n\n");
pathname = safe_malloc(len);
elem = NULL;
while ((elem = nvlist_next_nvpair(nverrlist, elem)) != NULL) {
nvlist_t *nv;
uint64_t dsobj, obj;
verify(nvpair_value_nvlist(elem, &nv) == 0);
verify(nvlist_lookup_uint64(nv, ZPOOL_ERR_DATASET,
&dsobj) == 0);
verify(nvlist_lookup_uint64(nv, ZPOOL_ERR_OBJECT,
&obj) == 0);
zpool_obj_to_path(zhp, dsobj, obj, pathname, len);
(void) printf("%7s %s\n", "", pathname);
}
free(pathname);
nvlist_free(nverrlist);
}
static void
print_spares(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t **spares,
uint_t nspares)
{
uint_t i;
char *name;
if (nspares == 0)
return;
(void) printf(gettext("\tspares\n"));
for (i = 0; i < nspares; i++) {
name = zpool_vdev_name(g_zfs, zhp, spares[i],
cb->cb_name_flags);
print_status_config(zhp, cb, name, spares[i], 2, B_TRUE, NULL);
free(name);
}
}
static void
print_l2cache(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t **l2cache,
uint_t nl2cache)
{
uint_t i;
char *name;
if (nl2cache == 0)
return;
(void) printf(gettext("\tcache\n"));
for (i = 0; i < nl2cache; i++) {
name = zpool_vdev_name(g_zfs, zhp, l2cache[i],
cb->cb_name_flags);
print_status_config(zhp, cb, name, l2cache[i], 2,
B_FALSE, NULL);
free(name);
}
}
static void
print_dedup_stats(nvlist_t *config)
{
ddt_histogram_t *ddh;
ddt_stat_t *dds;
ddt_object_t *ddo;
uint_t c;
char dspace[6], mspace[6];
/*
* If the pool was faulted then we may not have been able to
* obtain the config. Otherwise, if we have anything in the dedup
* table continue processing the stats.
*/
if (nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_OBJ_STATS,
(uint64_t **)&ddo, &c) != 0)
return;
(void) printf("\n");
(void) printf(gettext(" dedup: "));
if (ddo->ddo_count == 0) {
(void) printf(gettext("no DDT entries\n"));
return;
}
zfs_nicebytes(ddo->ddo_dspace, dspace, sizeof (dspace));
zfs_nicebytes(ddo->ddo_mspace, mspace, sizeof (mspace));
(void) printf("DDT entries %llu, size %s on disk, %s in core\n",
(u_longlong_t)ddo->ddo_count,
dspace,
mspace);
verify(nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_STATS,
(uint64_t **)&dds, &c) == 0);
verify(nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_HISTOGRAM,
(uint64_t **)&ddh, &c) == 0);
zpool_dump_ddt(dds, ddh);
}
/*
* Display a summary of pool status. Displays a summary such as:
*
* pool: tank
* status: DEGRADED
* reason: One or more devices ...
* see: https://openzfs.github.io/openzfs-docs/msg/ZFS-xxxx-01
* config:
* mirror DEGRADED
* c1t0d0 OK
* c2t0d0 UNAVAIL
*
* When given the '-v' option, we print out the complete config. If the '-e'
* option is specified, then we print out error rate information as well.
*/
static int
status_callback(zpool_handle_t *zhp, void *data)
{
status_cbdata_t *cbp = data;
nvlist_t *config, *nvroot;
const char *msgid;
zpool_status_t reason;
zpool_errata_t errata;
const char *health;
uint_t c;
vdev_stat_t *vs;
config = zpool_get_config(zhp, NULL);
reason = zpool_get_status(zhp, &msgid, &errata);
cbp->cb_count++;
/*
* If we were given 'zpool status -x', only report those pools with
* problems.
*/
if (cbp->cb_explain &&
(reason == ZPOOL_STATUS_OK ||
reason == ZPOOL_STATUS_VERSION_OLDER ||
reason == ZPOOL_STATUS_FEAT_DISABLED ||
reason == ZPOOL_STATUS_COMPATIBILITY_ERR ||
reason == ZPOOL_STATUS_INCOMPATIBLE_FEAT)) {
if (!cbp->cb_allpools) {
(void) printf(gettext("pool '%s' is healthy\n"),
zpool_get_name(zhp));
if (cbp->cb_first)
cbp->cb_first = B_FALSE;
}
return (0);
}
if (cbp->cb_first)
cbp->cb_first = B_FALSE;
else
(void) printf("\n");
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &c) == 0);
health = zpool_get_state_str(zhp);
printf(" ");
printf_color(ANSI_BOLD, gettext("pool:"));
printf(" %s\n", zpool_get_name(zhp));
fputc(' ', stdout);
printf_color(ANSI_BOLD, gettext("state: "));
printf_color(health_str_to_color(health), "%s", health);
fputc('\n', stdout);
switch (reason) {
case ZPOOL_STATUS_MISSING_DEV_R:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices could "
"not be opened. Sufficient replicas exist for\n\tthe pool "
"to continue functioning in a degraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Attach the missing device "
"and online it using 'zpool online'.\n"));
break;
case ZPOOL_STATUS_MISSING_DEV_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices could "
"not be opened. There are insufficient\n\treplicas for the"
" pool to continue functioning.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Attach the missing device "
"and online it using 'zpool online'.\n"));
break;
case ZPOOL_STATUS_CORRUPT_LABEL_R:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices could "
"not be used because the label is missing or\n\tinvalid. "
"Sufficient replicas exist for the pool to continue\n\t"
"functioning in a degraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Replace the device using "
"'zpool replace'.\n"));
break;
case ZPOOL_STATUS_CORRUPT_LABEL_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices could "
"not be used because the label is missing \n\tor invalid. "
"There are insufficient replicas for the pool to "
"continue\n\tfunctioning.\n"));
zpool_explain_recover(zpool_get_handle(zhp),
zpool_get_name(zhp), reason, config);
break;
case ZPOOL_STATUS_FAILING_DEV:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices has "
"experienced an unrecoverable error. An\n\tattempt was "
"made to correct the error. Applications are "
"unaffected.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Determine if the "
"device needs to be replaced, and clear the errors\n\tusing"
" 'zpool clear' or replace the device with 'zpool "
"replace'.\n"));
break;
case ZPOOL_STATUS_OFFLINE_DEV:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices has "
"been taken offline by the administrator.\n\tSufficient "
"replicas exist for the pool to continue functioning in "
"a\n\tdegraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Online the device "
"using 'zpool online' or replace the device with\n\t'zpool "
"replace'.\n"));
break;
case ZPOOL_STATUS_REMOVED_DEV:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices has "
"been removed by the administrator.\n\tSufficient "
"replicas exist for the pool to continue functioning in "
"a\n\tdegraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Online the device "
"using zpool online' or replace the device with\n\t'zpool "
"replace'.\n"));
break;
case ZPOOL_STATUS_RESILVERING:
case ZPOOL_STATUS_REBUILDING:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices is "
"currently being resilvered. The pool will\n\tcontinue "
"to function, possibly in a degraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Wait for the resilver to "
"complete.\n"));
break;
case ZPOOL_STATUS_REBUILD_SCRUB:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices have "
"been sequentially resilvered, scrubbing\n\tthe pool "
"is recommended.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Use 'zpool scrub' to "
"verify all data checksums.\n"));
break;
case ZPOOL_STATUS_CORRUPT_DATA:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices has "
"experienced an error resulting in data\n\tcorruption. "
"Applications may be affected.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Restore the file in question"
" if possible. Otherwise restore the\n\tentire pool from "
"backup.\n"));
break;
case ZPOOL_STATUS_CORRUPT_POOL:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool metadata is "
"corrupted and the pool cannot be opened.\n"));
zpool_explain_recover(zpool_get_handle(zhp),
zpool_get_name(zhp), reason, config);
break;
case ZPOOL_STATUS_VERSION_OLDER:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is formatted using "
"a legacy on-disk format. The pool can\n\tstill be used, "
"but some features are unavailable.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Upgrade the pool using "
"'zpool upgrade'. Once this is done, the\n\tpool will no "
"longer be accessible on software that does not support\n\t"
"feature flags.\n"));
break;
case ZPOOL_STATUS_VERSION_NEWER:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool has been upgraded "
"to a newer, incompatible on-disk version.\n\tThe pool "
"cannot be accessed on this system.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Access the pool from a "
"system running more recent software, or\n\trestore the "
"pool from backup.\n"));
break;
case ZPOOL_STATUS_FEAT_DISABLED:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Some supported and "
"requested features are not enabled on the pool.\n\t"
"The pool can still be used, but some features are "
"unavailable.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Enable all features using "
"'zpool upgrade'. Once this is done,\n\tthe pool may no "
"longer be accessible by software that does not support\n\t"
"the features. See zpool-features(7) for details.\n"));
break;
case ZPOOL_STATUS_COMPATIBILITY_ERR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("This pool has a "
"compatibility list specified, but it could not be\n\t"
"read/parsed at this time. The pool can still be used, "
"but this\n\tshould be investigated.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Check the value of the "
"'compatibility' property against the\n\t"
"appropriate file in " ZPOOL_SYSCONF_COMPAT_D " or "
ZPOOL_DATA_COMPAT_D ".\n"));
break;
case ZPOOL_STATUS_INCOMPATIBLE_FEAT:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more features "
"are enabled on the pool despite not being\n\t"
"requested by the 'compatibility' property.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Consider setting "
"'compatibility' to an appropriate value, or\n\t"
"adding needed features to the relevant file in\n\t"
ZPOOL_SYSCONF_COMPAT_D " or " ZPOOL_DATA_COMPAT_D ".\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_READ:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool cannot be accessed "
"on this system because it uses the\n\tfollowing feature(s)"
" not supported on this system:\n"));
zpool_print_unsup_feat(config);
(void) printf("\n");
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Access the pool from a "
"system that supports the required feature(s),\n\tor "
"restore the pool from backup.\n"));
break;
case ZPOOL_STATUS_UNSUP_FEAT_WRITE:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool can only be "
"accessed in read-only mode on this system. It\n\tcannot be"
" accessed in read-write mode because it uses the "
"following\n\tfeature(s) not supported on this system:\n"));
zpool_print_unsup_feat(config);
(void) printf("\n");
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("The pool cannot be accessed "
"in read-write mode. Import the pool with\n"
"\t\"-o readonly=on\", access the pool from a system that "
"supports the\n\trequired feature(s), or restore the "
"pool from backup.\n"));
break;
case ZPOOL_STATUS_FAULTED_DEV_R:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"faulted in response to persistent errors.\n\tSufficient "
"replicas exist for the pool to continue functioning "
"in a\n\tdegraded state.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Replace the faulted device, "
"or use 'zpool clear' to mark the device\n\trepaired.\n"));
break;
case ZPOOL_STATUS_FAULTED_DEV_NR:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"faulted in response to persistent errors. There are "
"insufficient replicas for the pool to\n\tcontinue "
"functioning.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Destroy and re-create the "
"pool from a backup source. Manually marking the device\n"
"\trepaired using 'zpool clear' may allow some data "
"to be recovered.\n"));
break;
case ZPOOL_STATUS_IO_FAILURE_MMP:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("The pool is suspended "
"because multihost writes failed or were delayed;\n\t"
"another system could import the pool undetected.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Make sure the pool's devices"
" are connected, then reboot your system and\n\timport the "
"pool.\n"));
break;
case ZPOOL_STATUS_IO_FAILURE_WAIT:
case ZPOOL_STATUS_IO_FAILURE_CONTINUE:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("One or more devices are "
"faulted in response to IO failures.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Make sure the affected "
"devices are connected, then run 'zpool clear'.\n"));
break;
case ZPOOL_STATUS_BAD_LOG:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("An intent log record "
"could not be read.\n"
"\tWaiting for administrator intervention to fix the "
"faulted pool.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Either restore the affected "
"device(s) and run 'zpool online',\n"
"\tor ignore the intent log records by running "
"'zpool clear'.\n"));
break;
case ZPOOL_STATUS_NON_NATIVE_ASHIFT:
(void) printf(gettext("status: One or more devices are "
"configured to use a non-native block size.\n"
"\tExpect reduced performance.\n"));
(void) printf(gettext("action: Replace affected devices with "
"devices that support the\n\tconfigured block size, or "
"migrate data to a properly configured\n\tpool.\n"));
break;
case ZPOOL_STATUS_HOSTID_MISMATCH:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Mismatch between pool hostid"
" and system hostid on imported pool.\n\tThis pool was "
"previously imported into a system with a different "
"hostid,\n\tand then was verbatim imported into this "
"system.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("Export this pool on all "
"systems on which it is imported.\n"
"\tThen import it to correct the mismatch.\n"));
break;
case ZPOOL_STATUS_ERRATA:
printf_color(ANSI_BOLD, gettext("status: "));
printf_color(ANSI_YELLOW, gettext("Errata #%d detected.\n"),
errata);
switch (errata) {
case ZPOOL_ERRATA_NONE:
break;
case ZPOOL_ERRATA_ZOL_2094_SCRUB:
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("To correct the issue"
" run 'zpool scrub'.\n"));
break;
case ZPOOL_ERRATA_ZOL_6845_ENCRYPTION:
(void) printf(gettext("\tExisting encrypted datasets "
"contain an on-disk incompatibility\n\twhich "
"needs to be corrected.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("To correct the issue"
" backup existing encrypted datasets to new\n\t"
"encrypted datasets and destroy the old ones. "
"'zfs mount -o ro' can\n\tbe used to temporarily "
"mount existing encrypted datasets readonly.\n"));
break;
case ZPOOL_ERRATA_ZOL_8308_ENCRYPTION:
(void) printf(gettext("\tExisting encrypted snapshots "
"and bookmarks contain an on-disk\n\tincompat"
"ibility. This may cause on-disk corruption if "
"they are used\n\twith 'zfs recv'.\n"));
printf_color(ANSI_BOLD, gettext("action: "));
printf_color(ANSI_YELLOW, gettext("To correct the"
"issue, enable the bookmark_v2 feature. No "
"additional\n\taction is needed if there are no "
"encrypted snapshots or bookmarks.\n\tIf preserving"
"the encrypted snapshots and bookmarks is required,"
" use\n\ta non-raw send to backup and restore them."
" Alternately, they may be\n\tremoved to resolve "
"the incompatibility.\n"));
break;
default:
/*
* All errata which allow the pool to be imported
* must contain an action message.
*/
assert(0);
}
break;
default:
/*
* The remaining errors can't actually be generated, yet.
*/
assert(reason == ZPOOL_STATUS_OK);
}
if (msgid != NULL) {
printf(" ");
printf_color(ANSI_BOLD, gettext("see:"));
printf(gettext(
" https://openzfs.github.io/openzfs-docs/msg/%s\n"),
msgid);
}
if (config != NULL) {
uint64_t nerr;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
pool_checkpoint_stat_t *pcs = NULL;
pool_removal_stat_t *prs = NULL;
print_scan_status(zhp, nvroot);
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c);
print_removal_status(zhp, prs);
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c);
print_checkpoint_status(pcs);
cbp->cb_namewidth = max_width(zhp, nvroot, 0, 0,
cbp->cb_name_flags | VDEV_NAME_TYPE_ID);
if (cbp->cb_namewidth < 10)
cbp->cb_namewidth = 10;
color_start(ANSI_BOLD);
(void) printf(gettext("config:\n\n"));
(void) printf(gettext("\t%-*s %-8s %5s %5s %5s"),
cbp->cb_namewidth, "NAME", "STATE", "READ", "WRITE",
"CKSUM");
color_end();
if (cbp->cb_print_slow_ios) {
printf_color(ANSI_BOLD, " %5s", gettext("SLOW"));
}
if (cbp->vcdl != NULL)
print_cmd_columns(cbp->vcdl, 0);
printf("\n");
print_status_config(zhp, cbp, zpool_get_name(zhp), nvroot, 0,
B_FALSE, NULL);
print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_BIAS_DEDUP);
print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_BIAS_SPECIAL);
print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_CLASS_LOGS);
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0)
print_l2cache(zhp, cbp, l2cache, nl2cache);
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0)
print_spares(zhp, cbp, spares, nspares);
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT,
&nerr) == 0) {
(void) printf("\n");
if (nerr == 0) {
(void) printf(gettext(
"errors: No known data errors\n"));
} else if (!cbp->cb_verbose) {
(void) printf(gettext("errors: %llu data "
"errors, use '-v' for a list\n"),
(u_longlong_t)nerr);
} else {
print_error_log(zhp);
}
}
if (cbp->cb_dedup_stats)
print_dedup_stats(config);
} else {
(void) printf(gettext("config: The configuration cannot be "
"determined.\n"));
}
return (0);
}
/*
* zpool status [-c [script1,script2,...]] [-igLpPstvx] [-T d|u] [pool] ...
* [interval [count]]
*
* -c CMD For each vdev, run command CMD
* -i Display vdev initialization status.
* -g Display guid for individual vdev name.
* -L Follow links when resolving vdev path name.
* -p Display values in parsable (exact) format.
* -P Display full path for vdev name.
* -s Display slow IOs column.
* -v Display complete error logs
* -x Display only pools with potential problems
* -D Display dedup status (undocumented)
* -t Display vdev TRIM status.
* -T Display a timestamp in date(1) or Unix format
*
* Describes the health status of all pools or some subset.
*/
int
zpool_do_status(int argc, char **argv)
{
int c;
int ret;
float interval = 0;
unsigned long count = 0;
status_cbdata_t cb = { 0 };
char *cmd = NULL;
/* check options */
while ((c = getopt(argc, argv, "c:igLpPsvxDtT:")) != -1) {
switch (c) {
case 'c':
if (cmd != NULL) {
fprintf(stderr,
gettext("Can't set -c flag twice\n"));
exit(1);
}
if (getenv("ZPOOL_SCRIPTS_ENABLED") != NULL &&
!libzfs_envvar_is_set("ZPOOL_SCRIPTS_ENABLED")) {
fprintf(stderr, gettext(
"Can't run -c, disabled by "
"ZPOOL_SCRIPTS_ENABLED.\n"));
exit(1);
}
if ((getuid() <= 0 || geteuid() <= 0) &&
!libzfs_envvar_is_set("ZPOOL_SCRIPTS_AS_ROOT")) {
fprintf(stderr, gettext(
"Can't run -c with root privileges "
"unless ZPOOL_SCRIPTS_AS_ROOT is set.\n"));
exit(1);
}
cmd = optarg;
break;
case 'i':
cb.cb_print_vdev_init = B_TRUE;
break;
case 'g':
cb.cb_name_flags |= VDEV_NAME_GUID;
break;
case 'L':
cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS;
break;
case 'p':
cb.cb_literal = B_TRUE;
break;
case 'P':
cb.cb_name_flags |= VDEV_NAME_PATH;
break;
case 's':
cb.cb_print_slow_ios = B_TRUE;
break;
case 'v':
cb.cb_verbose = B_TRUE;
break;
case 'x':
cb.cb_explain = B_TRUE;
break;
case 'D':
cb.cb_dedup_stats = B_TRUE;
break;
case 't':
cb.cb_print_vdev_trim = B_TRUE;
break;
case 'T':
get_timestamp_arg(*optarg);
break;
case '?':
if (optopt == 'c') {
print_zpool_script_list("status");
exit(0);
} else {
fprintf(stderr,
gettext("invalid option '%c'\n"), optopt);
}
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
get_interval_count(&argc, argv, &interval, &count);
if (argc == 0)
cb.cb_allpools = B_TRUE;
cb.cb_first = B_TRUE;
cb.cb_print_status = B_TRUE;
for (;;) {
if (timestamp_fmt != NODATE)
print_timestamp(timestamp_fmt);
if (cmd != NULL)
cb.vcdl = all_pools_for_each_vdev_run(argc, argv, cmd,
NULL, NULL, 0, 0);
ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
cb.cb_literal, status_callback, &cb);
if (cb.vcdl != NULL)
free_vdev_cmd_data_list(cb.vcdl);
if (argc == 0 && cb.cb_count == 0)
(void) fprintf(stderr, gettext("no pools available\n"));
else if (cb.cb_explain && cb.cb_first && cb.cb_allpools)
(void) printf(gettext("all pools are healthy\n"));
if (ret != 0)
return (ret);
if (interval == 0)
break;
if (count != 0 && --count == 0)
break;
(void) fsleep(interval);
}
return (0);
}
typedef struct upgrade_cbdata {
int cb_first;
int cb_argc;
uint64_t cb_version;
char **cb_argv;
} upgrade_cbdata_t;
static int
check_unsupp_fs(zfs_handle_t *zhp, void *unsupp_fs)
{
int zfs_version = (int)zfs_prop_get_int(zhp, ZFS_PROP_VERSION);
int *count = (int *)unsupp_fs;
if (zfs_version > ZPL_VERSION) {
(void) printf(gettext("%s (v%d) is not supported by this "
"implementation of ZFS.\n"),
zfs_get_name(zhp), zfs_version);
(*count)++;
}
- zfs_iter_filesystems(zhp, 0, check_unsupp_fs, unsupp_fs);
+ zfs_iter_filesystems_v2(zhp, 0, check_unsupp_fs, unsupp_fs);
zfs_close(zhp);
return (0);
}
static int
upgrade_version(zpool_handle_t *zhp, uint64_t version)
{
int ret;
nvlist_t *config;
uint64_t oldversion;
int unsupp_fs = 0;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&oldversion) == 0);
char compat[ZFS_MAXPROPLEN];
if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat,
ZFS_MAXPROPLEN, NULL, B_FALSE) != 0)
compat[0] = '\0';
assert(SPA_VERSION_IS_SUPPORTED(oldversion));
assert(oldversion < version);
ret = zfs_iter_root(zpool_get_handle(zhp), check_unsupp_fs, &unsupp_fs);
if (ret != 0)
return (ret);
if (unsupp_fs) {
(void) fprintf(stderr, gettext("Upgrade not performed due "
"to %d unsupported filesystems (max v%d).\n"),
unsupp_fs, (int)ZPL_VERSION);
return (1);
}
if (strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0) {
(void) fprintf(stderr, gettext("Upgrade not performed because "
"'compatibility' property set to '"
ZPOOL_COMPAT_LEGACY "'.\n"));
return (1);
}
ret = zpool_upgrade(zhp, version);
if (ret != 0)
return (ret);
if (version >= SPA_VERSION_FEATURES) {
(void) printf(gettext("Successfully upgraded "
"'%s' from version %llu to feature flags.\n"),
zpool_get_name(zhp), (u_longlong_t)oldversion);
} else {
(void) printf(gettext("Successfully upgraded "
"'%s' from version %llu to version %llu.\n"),
zpool_get_name(zhp), (u_longlong_t)oldversion,
(u_longlong_t)version);
}
return (0);
}
static int
upgrade_enable_all(zpool_handle_t *zhp, int *countp)
{
int i, ret, count;
boolean_t firstff = B_TRUE;
nvlist_t *enabled = zpool_get_features(zhp);
char compat[ZFS_MAXPROPLEN];
if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat,
ZFS_MAXPROPLEN, NULL, B_FALSE) != 0)
compat[0] = '\0';
boolean_t requested_features[SPA_FEATURES];
if (zpool_do_load_compat(compat, requested_features) !=
ZPOOL_COMPATIBILITY_OK)
return (-1);
count = 0;
for (i = 0; i < SPA_FEATURES; i++) {
const char *fname = spa_feature_table[i].fi_uname;
const char *fguid = spa_feature_table[i].fi_guid;
if (!spa_feature_table[i].fi_zfs_mod_supported)
continue;
if (!nvlist_exists(enabled, fguid) && requested_features[i]) {
char *propname;
verify(-1 != asprintf(&propname, "feature@%s", fname));
ret = zpool_set_prop(zhp, propname,
ZFS_FEATURE_ENABLED);
if (ret != 0) {
free(propname);
return (ret);
}
count++;
if (firstff) {
(void) printf(gettext("Enabled the "
"following features on '%s':\n"),
zpool_get_name(zhp));
firstff = B_FALSE;
}
(void) printf(gettext(" %s\n"), fname);
free(propname);
}
}
if (countp != NULL)
*countp = count;
return (0);
}
static int
upgrade_cb(zpool_handle_t *zhp, void *arg)
{
upgrade_cbdata_t *cbp = arg;
nvlist_t *config;
uint64_t version;
boolean_t modified_pool = B_FALSE;
int ret;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&version) == 0);
assert(SPA_VERSION_IS_SUPPORTED(version));
if (version < cbp->cb_version) {
cbp->cb_first = B_FALSE;
ret = upgrade_version(zhp, cbp->cb_version);
if (ret != 0)
return (ret);
modified_pool = B_TRUE;
/*
* If they did "zpool upgrade -a", then we could
* be doing ioctls to different pools. We need
* to log this history once to each pool, and bypass
* the normal history logging that happens in main().
*/
(void) zpool_log_history(g_zfs, history_str);
log_history = B_FALSE;
}
if (cbp->cb_version >= SPA_VERSION_FEATURES) {
int count;
ret = upgrade_enable_all(zhp, &count);
if (ret != 0)
return (ret);
if (count > 0) {
cbp->cb_first = B_FALSE;
modified_pool = B_TRUE;
}
}
if (modified_pool) {
(void) printf("\n");
(void) after_zpool_upgrade(zhp);
}
return (0);
}
static int
upgrade_list_older_cb(zpool_handle_t *zhp, void *arg)
{
upgrade_cbdata_t *cbp = arg;
nvlist_t *config;
uint64_t version;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&version) == 0);
assert(SPA_VERSION_IS_SUPPORTED(version));
if (version < SPA_VERSION_FEATURES) {
if (cbp->cb_first) {
(void) printf(gettext("The following pools are "
"formatted with legacy version numbers and can\n"
"be upgraded to use feature flags. After "
"being upgraded, these pools\nwill no "
"longer be accessible by software that does not "
"support feature\nflags.\n\n"
"Note that setting a pool's 'compatibility' "
"feature to '" ZPOOL_COMPAT_LEGACY "' will\n"
"inhibit upgrades.\n\n"));
(void) printf(gettext("VER POOL\n"));
(void) printf(gettext("--- ------------\n"));
cbp->cb_first = B_FALSE;
}
(void) printf("%2llu %s\n", (u_longlong_t)version,
zpool_get_name(zhp));
}
return (0);
}
static int
upgrade_list_disabled_cb(zpool_handle_t *zhp, void *arg)
{
upgrade_cbdata_t *cbp = arg;
nvlist_t *config;
uint64_t version;
config = zpool_get_config(zhp, NULL);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&version) == 0);
if (version >= SPA_VERSION_FEATURES) {
int i;
boolean_t poolfirst = B_TRUE;
nvlist_t *enabled = zpool_get_features(zhp);
for (i = 0; i < SPA_FEATURES; i++) {
const char *fguid = spa_feature_table[i].fi_guid;
const char *fname = spa_feature_table[i].fi_uname;
if (!spa_feature_table[i].fi_zfs_mod_supported)
continue;
if (!nvlist_exists(enabled, fguid)) {
if (cbp->cb_first) {
(void) printf(gettext("\nSome "
"supported features are not "
"enabled on the following pools. "
"Once a\nfeature is enabled the "
"pool may become incompatible with "
"software\nthat does not support "
"the feature. See "
"zpool-features(7) for "
"details.\n\n"
"Note that the pool "
"'compatibility' feature can be "
"used to inhibit\nfeature "
"upgrades.\n\n"));
(void) printf(gettext("POOL "
"FEATURE\n"));
(void) printf(gettext("------"
"---------\n"));
cbp->cb_first = B_FALSE;
}
if (poolfirst) {
(void) printf(gettext("%s\n"),
zpool_get_name(zhp));
poolfirst = B_FALSE;
}
(void) printf(gettext(" %s\n"), fname);
}
/*
* If they did "zpool upgrade -a", then we could
* be doing ioctls to different pools. We need
* to log this history once to each pool, and bypass
* the normal history logging that happens in main().
*/
(void) zpool_log_history(g_zfs, history_str);
log_history = B_FALSE;
}
}
return (0);
}
static int
upgrade_one(zpool_handle_t *zhp, void *data)
{
boolean_t modified_pool = B_FALSE;
upgrade_cbdata_t *cbp = data;
uint64_t cur_version;
int ret;
if (strcmp("log", zpool_get_name(zhp)) == 0) {
(void) fprintf(stderr, gettext("'log' is now a reserved word\n"
"Pool 'log' must be renamed using export and import"
" to upgrade.\n"));
return (1);
}
cur_version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL);
if (cur_version > cbp->cb_version) {
(void) printf(gettext("Pool '%s' is already formatted "
"using more current version '%llu'.\n\n"),
zpool_get_name(zhp), (u_longlong_t)cur_version);
return (0);
}
if (cbp->cb_version != SPA_VERSION && cur_version == cbp->cb_version) {
(void) printf(gettext("Pool '%s' is already formatted "
"using version %llu.\n\n"), zpool_get_name(zhp),
(u_longlong_t)cbp->cb_version);
return (0);
}
if (cur_version != cbp->cb_version) {
modified_pool = B_TRUE;
ret = upgrade_version(zhp, cbp->cb_version);
if (ret != 0)
return (ret);
}
if (cbp->cb_version >= SPA_VERSION_FEATURES) {
int count = 0;
ret = upgrade_enable_all(zhp, &count);
if (ret != 0)
return (ret);
if (count != 0) {
modified_pool = B_TRUE;
} else if (cur_version == SPA_VERSION) {
(void) printf(gettext("Pool '%s' already has all "
"supported and requested features enabled.\n"),
zpool_get_name(zhp));
}
}
if (modified_pool) {
(void) printf("\n");
(void) after_zpool_upgrade(zhp);
}
return (0);
}
/*
* zpool upgrade
* zpool upgrade -v
* zpool upgrade [-V version] <-a | pool ...>
*
* With no arguments, display downrev'd ZFS pool available for upgrade.
* Individual pools can be upgraded by specifying the pool, and '-a' will
* upgrade all pools.
*/
int
zpool_do_upgrade(int argc, char **argv)
{
int c;
upgrade_cbdata_t cb = { 0 };
int ret = 0;
boolean_t showversions = B_FALSE;
boolean_t upgradeall = B_FALSE;
char *end;
/* check options */
while ((c = getopt(argc, argv, ":avV:")) != -1) {
switch (c) {
case 'a':
upgradeall = B_TRUE;
break;
case 'v':
showversions = B_TRUE;
break;
case 'V':
cb.cb_version = strtoll(optarg, &end, 10);
if (*end != '\0' ||
!SPA_VERSION_IS_SUPPORTED(cb.cb_version)) {
(void) fprintf(stderr,
gettext("invalid version '%s'\n"), optarg);
usage(B_FALSE);
}
break;
case ':':
(void) fprintf(stderr, gettext("missing argument for "
"'%c' option\n"), optopt);
usage(B_FALSE);
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
cb.cb_argc = argc;
cb.cb_argv = argv;
argc -= optind;
argv += optind;
if (cb.cb_version == 0) {
cb.cb_version = SPA_VERSION;
} else if (!upgradeall && argc == 0) {
(void) fprintf(stderr, gettext("-V option is "
"incompatible with other arguments\n"));
usage(B_FALSE);
}
if (showversions) {
if (upgradeall || argc != 0) {
(void) fprintf(stderr, gettext("-v option is "
"incompatible with other arguments\n"));
usage(B_FALSE);
}
} else if (upgradeall) {
if (argc != 0) {
(void) fprintf(stderr, gettext("-a option should not "
"be used along with a pool name\n"));
usage(B_FALSE);
}
}
(void) printf("%s", gettext("This system supports ZFS pool feature "
"flags.\n\n"));
if (showversions) {
int i;
(void) printf(gettext("The following features are "
"supported:\n\n"));
(void) printf(gettext("FEAT DESCRIPTION\n"));
(void) printf("----------------------------------------------"
"---------------\n");
for (i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t *fi = &spa_feature_table[i];
if (!fi->fi_zfs_mod_supported)
continue;
const char *ro =
(fi->fi_flags & ZFEATURE_FLAG_READONLY_COMPAT) ?
" (read-only compatible)" : "";
(void) printf("%-37s%s\n", fi->fi_uname, ro);
(void) printf(" %s\n", fi->fi_desc);
}
(void) printf("\n");
(void) printf(gettext("The following legacy versions are also "
"supported:\n\n"));
(void) printf(gettext("VER DESCRIPTION\n"));
(void) printf("--- -----------------------------------------"
"---------------\n");
(void) printf(gettext(" 1 Initial ZFS version\n"));
(void) printf(gettext(" 2 Ditto blocks "
"(replicated metadata)\n"));
(void) printf(gettext(" 3 Hot spares and double parity "
"RAID-Z\n"));
(void) printf(gettext(" 4 zpool history\n"));
(void) printf(gettext(" 5 Compression using the gzip "
"algorithm\n"));
(void) printf(gettext(" 6 bootfs pool property\n"));
(void) printf(gettext(" 7 Separate intent log devices\n"));
(void) printf(gettext(" 8 Delegated administration\n"));
(void) printf(gettext(" 9 refquota and refreservation "
"properties\n"));
(void) printf(gettext(" 10 Cache devices\n"));
(void) printf(gettext(" 11 Improved scrub performance\n"));
(void) printf(gettext(" 12 Snapshot properties\n"));
(void) printf(gettext(" 13 snapused property\n"));
(void) printf(gettext(" 14 passthrough-x aclinherit\n"));
(void) printf(gettext(" 15 user/group space accounting\n"));
(void) printf(gettext(" 16 stmf property support\n"));
(void) printf(gettext(" 17 Triple-parity RAID-Z\n"));
(void) printf(gettext(" 18 Snapshot user holds\n"));
(void) printf(gettext(" 19 Log device removal\n"));
(void) printf(gettext(" 20 Compression using zle "
"(zero-length encoding)\n"));
(void) printf(gettext(" 21 Deduplication\n"));
(void) printf(gettext(" 22 Received properties\n"));
(void) printf(gettext(" 23 Slim ZIL\n"));
(void) printf(gettext(" 24 System attributes\n"));
(void) printf(gettext(" 25 Improved scrub stats\n"));
(void) printf(gettext(" 26 Improved snapshot deletion "
"performance\n"));
(void) printf(gettext(" 27 Improved snapshot creation "
"performance\n"));
(void) printf(gettext(" 28 Multiple vdev replacements\n"));
(void) printf(gettext("\nFor more information on a particular "
"version, including supported releases,\n"));
(void) printf(gettext("see the ZFS Administration Guide.\n\n"));
} else if (argc == 0 && upgradeall) {
cb.cb_first = B_TRUE;
ret = zpool_iter(g_zfs, upgrade_cb, &cb);
if (ret == 0 && cb.cb_first) {
if (cb.cb_version == SPA_VERSION) {
(void) printf(gettext("All pools are already "
"formatted using feature flags.\n\n"));
(void) printf(gettext("Every feature flags "
"pool already has all supported and "
"requested features enabled.\n"));
} else {
(void) printf(gettext("All pools are already "
"formatted with version %llu or higher.\n"),
(u_longlong_t)cb.cb_version);
}
}
} else if (argc == 0) {
cb.cb_first = B_TRUE;
ret = zpool_iter(g_zfs, upgrade_list_older_cb, &cb);
assert(ret == 0);
if (cb.cb_first) {
(void) printf(gettext("All pools are formatted "
"using feature flags.\n\n"));
} else {
(void) printf(gettext("\nUse 'zpool upgrade -v' "
"for a list of available legacy versions.\n"));
}
cb.cb_first = B_TRUE;
ret = zpool_iter(g_zfs, upgrade_list_disabled_cb, &cb);
assert(ret == 0);
if (cb.cb_first) {
(void) printf(gettext("Every feature flags pool has "
"all supported and requested features enabled.\n"));
} else {
(void) printf(gettext("\n"));
}
} else {
ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL,
B_FALSE, upgrade_one, &cb);
}
return (ret);
}
typedef struct hist_cbdata {
boolean_t first;
boolean_t longfmt;
boolean_t internal;
} hist_cbdata_t;
static void
print_history_records(nvlist_t *nvhis, hist_cbdata_t *cb)
{
nvlist_t **records;
uint_t numrecords;
int i;
verify(nvlist_lookup_nvlist_array(nvhis, ZPOOL_HIST_RECORD,
&records, &numrecords) == 0);
for (i = 0; i < numrecords; i++) {
nvlist_t *rec = records[i];
char tbuf[64] = "";
if (nvlist_exists(rec, ZPOOL_HIST_TIME)) {
time_t tsec;
struct tm t;
tsec = fnvlist_lookup_uint64(records[i],
ZPOOL_HIST_TIME);
(void) localtime_r(&tsec, &t);
(void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t);
}
if (nvlist_exists(rec, ZPOOL_HIST_ELAPSED_NS)) {
uint64_t elapsed_ns = fnvlist_lookup_int64(records[i],
ZPOOL_HIST_ELAPSED_NS);
(void) snprintf(tbuf + strlen(tbuf),
sizeof (tbuf) - strlen(tbuf),
" (%lldms)", (long long)elapsed_ns / 1000 / 1000);
}
if (nvlist_exists(rec, ZPOOL_HIST_CMD)) {
(void) printf("%s %s", tbuf,
fnvlist_lookup_string(rec, ZPOOL_HIST_CMD));
} else if (nvlist_exists(rec, ZPOOL_HIST_INT_EVENT)) {
int ievent =
fnvlist_lookup_uint64(rec, ZPOOL_HIST_INT_EVENT);
if (!cb->internal)
continue;
if (ievent >= ZFS_NUM_LEGACY_HISTORY_EVENTS) {
(void) printf("%s unrecognized record:\n",
tbuf);
dump_nvlist(rec, 4);
continue;
}
(void) printf("%s [internal %s txg:%lld] %s", tbuf,
zfs_history_event_names[ievent],
(longlong_t)fnvlist_lookup_uint64(
rec, ZPOOL_HIST_TXG),
fnvlist_lookup_string(rec, ZPOOL_HIST_INT_STR));
} else if (nvlist_exists(rec, ZPOOL_HIST_INT_NAME)) {
if (!cb->internal)
continue;
(void) printf("%s [txg:%lld] %s", tbuf,
(longlong_t)fnvlist_lookup_uint64(
rec, ZPOOL_HIST_TXG),
fnvlist_lookup_string(rec, ZPOOL_HIST_INT_NAME));
if (nvlist_exists(rec, ZPOOL_HIST_DSNAME)) {
(void) printf(" %s (%llu)",
fnvlist_lookup_string(rec,
ZPOOL_HIST_DSNAME),
(u_longlong_t)fnvlist_lookup_uint64(rec,
ZPOOL_HIST_DSID));
}
(void) printf(" %s", fnvlist_lookup_string(rec,
ZPOOL_HIST_INT_STR));
} else if (nvlist_exists(rec, ZPOOL_HIST_IOCTL)) {
if (!cb->internal)
continue;
(void) printf("%s ioctl %s\n", tbuf,
fnvlist_lookup_string(rec, ZPOOL_HIST_IOCTL));
if (nvlist_exists(rec, ZPOOL_HIST_INPUT_NVL)) {
(void) printf(" input:\n");
dump_nvlist(fnvlist_lookup_nvlist(rec,
ZPOOL_HIST_INPUT_NVL), 8);
}
if (nvlist_exists(rec, ZPOOL_HIST_OUTPUT_NVL)) {
(void) printf(" output:\n");
dump_nvlist(fnvlist_lookup_nvlist(rec,
ZPOOL_HIST_OUTPUT_NVL), 8);
}
if (nvlist_exists(rec, ZPOOL_HIST_OUTPUT_SIZE)) {
(void) printf(" output nvlist omitted; "
"original size: %lldKB\n",
(longlong_t)fnvlist_lookup_int64(rec,
ZPOOL_HIST_OUTPUT_SIZE) / 1024);
}
if (nvlist_exists(rec, ZPOOL_HIST_ERRNO)) {
(void) printf(" errno: %lld\n",
(longlong_t)fnvlist_lookup_int64(rec,
ZPOOL_HIST_ERRNO));
}
} else {
if (!cb->internal)
continue;
(void) printf("%s unrecognized record:\n", tbuf);
dump_nvlist(rec, 4);
}
if (!cb->longfmt) {
(void) printf("\n");
continue;
}
(void) printf(" [");
if (nvlist_exists(rec, ZPOOL_HIST_WHO)) {
uid_t who = fnvlist_lookup_uint64(rec, ZPOOL_HIST_WHO);
struct passwd *pwd = getpwuid(who);
(void) printf("user %d ", (int)who);
if (pwd != NULL)
(void) printf("(%s) ", pwd->pw_name);
}
if (nvlist_exists(rec, ZPOOL_HIST_HOST)) {
(void) printf("on %s",
fnvlist_lookup_string(rec, ZPOOL_HIST_HOST));
}
if (nvlist_exists(rec, ZPOOL_HIST_ZONE)) {
(void) printf(":%s",
fnvlist_lookup_string(rec, ZPOOL_HIST_ZONE));
}
(void) printf("]");
(void) printf("\n");
}
}
/*
* Print out the command history for a specific pool.
*/
static int
get_history_one(zpool_handle_t *zhp, void *data)
{
nvlist_t *nvhis;
int ret;
hist_cbdata_t *cb = (hist_cbdata_t *)data;
uint64_t off = 0;
boolean_t eof = B_FALSE;
cb->first = B_FALSE;
(void) printf(gettext("History for '%s':\n"), zpool_get_name(zhp));
while (!eof) {
if ((ret = zpool_get_history(zhp, &nvhis, &off, &eof)) != 0)
return (ret);
print_history_records(nvhis, cb);
nvlist_free(nvhis);
}
(void) printf("\n");
return (ret);
}
/*
* zpool history <pool>
*
* Displays the history of commands that modified pools.
*/
int
zpool_do_history(int argc, char **argv)
{
hist_cbdata_t cbdata = { 0 };
int ret;
int c;
cbdata.first = B_TRUE;
/* check options */
while ((c = getopt(argc, argv, "li")) != -1) {
switch (c) {
case 'l':
cbdata.longfmt = B_TRUE;
break;
case 'i':
cbdata.internal = B_TRUE;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL,
B_FALSE, get_history_one, &cbdata);
if (argc == 0 && cbdata.first == B_TRUE) {
(void) fprintf(stderr, gettext("no pools available\n"));
return (0);
}
return (ret);
}
typedef struct ev_opts {
int verbose;
int scripted;
int follow;
int clear;
char poolname[ZFS_MAX_DATASET_NAME_LEN];
} ev_opts_t;
static void
zpool_do_events_short(nvlist_t *nvl, ev_opts_t *opts)
{
char ctime_str[26], str[32];
const char *ptr;
int64_t *tv;
uint_t n;
verify(nvlist_lookup_int64_array(nvl, FM_EREPORT_TIME, &tv, &n) == 0);
memset(str, ' ', 32);
(void) ctime_r((const time_t *)&tv[0], ctime_str);
(void) memcpy(str, ctime_str+4, 6); /* 'Jun 30' */
(void) memcpy(str+7, ctime_str+20, 4); /* '1993' */
(void) memcpy(str+12, ctime_str+11, 8); /* '21:49:08' */
(void) sprintf(str+20, ".%09lld", (longlong_t)tv[1]); /* '.123456789' */
if (opts->scripted)
(void) printf(gettext("%s\t"), str);
else
(void) printf(gettext("%s "), str);
verify(nvlist_lookup_string(nvl, FM_CLASS, &ptr) == 0);
(void) printf(gettext("%s\n"), ptr);
}
static void
zpool_do_events_nvprint(nvlist_t *nvl, int depth)
{
nvpair_t *nvp;
for (nvp = nvlist_next_nvpair(nvl, NULL);
nvp != NULL; nvp = nvlist_next_nvpair(nvl, nvp)) {
data_type_t type = nvpair_type(nvp);
const char *name = nvpair_name(nvp);
boolean_t b;
uint8_t i8;
uint16_t i16;
uint32_t i32;
uint64_t i64;
const char *str;
nvlist_t *cnv;
printf(gettext("%*s%s = "), depth, "", name);
switch (type) {
case DATA_TYPE_BOOLEAN:
printf(gettext("%s"), "1");
break;
case DATA_TYPE_BOOLEAN_VALUE:
(void) nvpair_value_boolean_value(nvp, &b);
printf(gettext("%s"), b ? "1" : "0");
break;
case DATA_TYPE_BYTE:
(void) nvpair_value_byte(nvp, &i8);
printf(gettext("0x%x"), i8);
break;
case DATA_TYPE_INT8:
(void) nvpair_value_int8(nvp, (void *)&i8);
printf(gettext("0x%x"), i8);
break;
case DATA_TYPE_UINT8:
(void) nvpair_value_uint8(nvp, &i8);
printf(gettext("0x%x"), i8);
break;
case DATA_TYPE_INT16:
(void) nvpair_value_int16(nvp, (void *)&i16);
printf(gettext("0x%x"), i16);
break;
case DATA_TYPE_UINT16:
(void) nvpair_value_uint16(nvp, &i16);
printf(gettext("0x%x"), i16);
break;
case DATA_TYPE_INT32:
(void) nvpair_value_int32(nvp, (void *)&i32);
printf(gettext("0x%x"), i32);
break;
case DATA_TYPE_UINT32:
(void) nvpair_value_uint32(nvp, &i32);
printf(gettext("0x%x"), i32);
break;
case DATA_TYPE_INT64:
(void) nvpair_value_int64(nvp, (void *)&i64);
printf(gettext("0x%llx"), (u_longlong_t)i64);
break;
case DATA_TYPE_UINT64:
(void) nvpair_value_uint64(nvp, &i64);
/*
* translate vdev state values to readable
* strings to aide zpool events consumers
*/
if (strcmp(name,
FM_EREPORT_PAYLOAD_ZFS_VDEV_STATE) == 0 ||
strcmp(name,
FM_EREPORT_PAYLOAD_ZFS_VDEV_LASTSTATE) == 0) {
printf(gettext("\"%s\" (0x%llx)"),
zpool_state_to_name(i64, VDEV_AUX_NONE),
(u_longlong_t)i64);
} else {
printf(gettext("0x%llx"), (u_longlong_t)i64);
}
break;
case DATA_TYPE_HRTIME:
(void) nvpair_value_hrtime(nvp, (void *)&i64);
printf(gettext("0x%llx"), (u_longlong_t)i64);
break;
case DATA_TYPE_STRING:
(void) nvpair_value_string(nvp, &str);
printf(gettext("\"%s\""), str ? str : "<NULL>");
break;
case DATA_TYPE_NVLIST:
printf(gettext("(embedded nvlist)\n"));
(void) nvpair_value_nvlist(nvp, &cnv);
zpool_do_events_nvprint(cnv, depth + 8);
printf(gettext("%*s(end %s)"), depth, "", name);
break;
case DATA_TYPE_NVLIST_ARRAY: {
nvlist_t **val;
uint_t i, nelem;
(void) nvpair_value_nvlist_array(nvp, &val, &nelem);
printf(gettext("(%d embedded nvlists)\n"), nelem);
for (i = 0; i < nelem; i++) {
printf(gettext("%*s%s[%d] = %s\n"),
depth, "", name, i, "(embedded nvlist)");
zpool_do_events_nvprint(val[i], depth + 8);
printf(gettext("%*s(end %s[%i])\n"),
depth, "", name, i);
}
printf(gettext("%*s(end %s)\n"), depth, "", name);
}
break;
case DATA_TYPE_INT8_ARRAY: {
int8_t *val;
uint_t i, nelem;
(void) nvpair_value_int8_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_UINT8_ARRAY: {
uint8_t *val;
uint_t i, nelem;
(void) nvpair_value_uint8_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_INT16_ARRAY: {
int16_t *val;
uint_t i, nelem;
(void) nvpair_value_int16_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_UINT16_ARRAY: {
uint16_t *val;
uint_t i, nelem;
(void) nvpair_value_uint16_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_INT32_ARRAY: {
int32_t *val;
uint_t i, nelem;
(void) nvpair_value_int32_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_UINT32_ARRAY: {
uint32_t *val;
uint_t i, nelem;
(void) nvpair_value_uint32_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%x "), val[i]);
break;
}
case DATA_TYPE_INT64_ARRAY: {
int64_t *val;
uint_t i, nelem;
(void) nvpair_value_int64_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%llx "),
(u_longlong_t)val[i]);
break;
}
case DATA_TYPE_UINT64_ARRAY: {
uint64_t *val;
uint_t i, nelem;
(void) nvpair_value_uint64_array(nvp, &val, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("0x%llx "),
(u_longlong_t)val[i]);
break;
}
case DATA_TYPE_STRING_ARRAY: {
const char **str;
uint_t i, nelem;
(void) nvpair_value_string_array(nvp, &str, &nelem);
for (i = 0; i < nelem; i++)
printf(gettext("\"%s\" "),
str[i] ? str[i] : "<NULL>");
break;
}
case DATA_TYPE_BOOLEAN_ARRAY:
case DATA_TYPE_BYTE_ARRAY:
case DATA_TYPE_DOUBLE:
case DATA_TYPE_DONTCARE:
case DATA_TYPE_UNKNOWN:
printf(gettext("<unknown>"));
break;
}
printf(gettext("\n"));
}
}
static int
zpool_do_events_next(ev_opts_t *opts)
{
nvlist_t *nvl;
int zevent_fd, ret, dropped;
const char *pool;
zevent_fd = open(ZFS_DEV, O_RDWR);
VERIFY(zevent_fd >= 0);
if (!opts->scripted)
(void) printf(gettext("%-30s %s\n"), "TIME", "CLASS");
while (1) {
ret = zpool_events_next(g_zfs, &nvl, &dropped,
(opts->follow ? ZEVENT_NONE : ZEVENT_NONBLOCK), zevent_fd);
if (ret || nvl == NULL)
break;
if (dropped > 0)
(void) printf(gettext("dropped %d events\n"), dropped);
if (strlen(opts->poolname) > 0 &&
nvlist_lookup_string(nvl, FM_FMRI_ZFS_POOL, &pool) == 0 &&
strcmp(opts->poolname, pool) != 0)
continue;
zpool_do_events_short(nvl, opts);
if (opts->verbose) {
zpool_do_events_nvprint(nvl, 8);
printf(gettext("\n"));
}
(void) fflush(stdout);
nvlist_free(nvl);
}
VERIFY(0 == close(zevent_fd));
return (ret);
}
static int
zpool_do_events_clear(void)
{
int count, ret;
ret = zpool_events_clear(g_zfs, &count);
if (!ret)
(void) printf(gettext("cleared %d events\n"), count);
return (ret);
}
/*
* zpool events [-vHf [pool] | -c]
*
* Displays events logs by ZFS.
*/
int
zpool_do_events(int argc, char **argv)
{
ev_opts_t opts = { 0 };
int ret;
int c;
/* check options */
while ((c = getopt(argc, argv, "vHfc")) != -1) {
switch (c) {
case 'v':
opts.verbose = 1;
break;
case 'H':
opts.scripted = 1;
break;
case 'f':
opts.follow = 1;
break;
case 'c':
opts.clear = 1;
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
} else if (argc == 1) {
(void) strlcpy(opts.poolname, argv[0], sizeof (opts.poolname));
if (!zfs_name_valid(opts.poolname, ZFS_TYPE_POOL)) {
(void) fprintf(stderr,
gettext("invalid pool name '%s'\n"), opts.poolname);
usage(B_FALSE);
}
}
if ((argc == 1 || opts.verbose || opts.scripted || opts.follow) &&
opts.clear) {
(void) fprintf(stderr,
gettext("invalid options combined with -c\n"));
usage(B_FALSE);
}
if (opts.clear)
ret = zpool_do_events_clear();
else
ret = zpool_do_events_next(&opts);
return (ret);
}
static int
get_callback_vdev(zpool_handle_t *zhp, char *vdevname, void *data)
{
zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data;
char value[ZFS_MAXPROPLEN];
zprop_source_t srctype;
for (zprop_list_t *pl = cbp->cb_proplist; pl != NULL;
pl = pl->pl_next) {
char *prop_name;
/*
* If the first property is pool name, it is a special
* placeholder that we can skip. This will also skip
* over the name property when 'all' is specified.
*/
if (pl->pl_prop == ZPOOL_PROP_NAME &&
pl == cbp->cb_proplist)
continue;
if (pl->pl_prop == ZPROP_INVAL) {
prop_name = pl->pl_user_prop;
} else {
prop_name = (char *)vdev_prop_to_name(pl->pl_prop);
}
if (zpool_get_vdev_prop(zhp, vdevname, pl->pl_prop,
prop_name, value, sizeof (value), &srctype,
cbp->cb_literal) == 0) {
zprop_print_one_property(vdevname, cbp, prop_name,
value, srctype, NULL, NULL);
}
}
return (0);
}
static int
-get_callback_vdev_width_cb(void *zhp_data, nvlist_t *nv, void *data)
+get_callback_vdev_cb(void *zhp_data, nvlist_t *nv, void *data)
{
zpool_handle_t *zhp = zhp_data;
zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data;
- char *vdevname = zpool_vdev_name(g_zfs, zhp, nv,
- cbp->cb_vdevs.cb_name_flags);
+ char *vdevname;
+ const char *type;
int ret;
- /* Adjust the column widths for the vdev properties */
- ret = vdev_expand_proplist(zhp, vdevname, &cbp->cb_proplist);
-
- return (ret);
-}
+ /*
+ * zpool_vdev_name() transforms the root vdev name (i.e., root-0) to the
+ * pool name for display purposes, which is not desired. Fallback to
+ * zpool_vdev_name() when not dealing with the root vdev.
+ */
+ type = fnvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE);
+ if (zhp != NULL && strcmp(type, "root") == 0)
+ vdevname = strdup("root-0");
+ else
+ vdevname = zpool_vdev_name(g_zfs, zhp, nv,
+ cbp->cb_vdevs.cb_name_flags);
-static int
-get_callback_vdev_cb(void *zhp_data, nvlist_t *nv, void *data)
-{
- zpool_handle_t *zhp = zhp_data;
- zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data;
- char *vdevname = zpool_vdev_name(g_zfs, zhp, nv,
- cbp->cb_vdevs.cb_name_flags);
- int ret;
+ (void) vdev_expand_proplist(zhp, vdevname, &cbp->cb_proplist);
- /* Display the properties */
ret = get_callback_vdev(zhp, vdevname, data);
+ free(vdevname);
+
return (ret);
}
static int
get_callback(zpool_handle_t *zhp, void *data)
{
zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data;
char value[ZFS_MAXPROPLEN];
zprop_source_t srctype;
zprop_list_t *pl;
int vid;
if (cbp->cb_type == ZFS_TYPE_VDEV) {
if (strcmp(cbp->cb_vdevs.cb_names[0], "all-vdevs") == 0) {
- for_each_vdev(zhp, get_callback_vdev_width_cb, data);
for_each_vdev(zhp, get_callback_vdev_cb, data);
} else {
/* Adjust column widths for vdev properties */
for (vid = 0; vid < cbp->cb_vdevs.cb_names_count;
vid++) {
vdev_expand_proplist(zhp,
cbp->cb_vdevs.cb_names[vid],
&cbp->cb_proplist);
}
/* Display the properties */
for (vid = 0; vid < cbp->cb_vdevs.cb_names_count;
vid++) {
get_callback_vdev(zhp,
cbp->cb_vdevs.cb_names[vid], data);
}
}
} else {
assert(cbp->cb_type == ZFS_TYPE_POOL);
for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) {
/*
* Skip the special fake placeholder. This will also
* skip over the name property when 'all' is specified.
*/
if (pl->pl_prop == ZPOOL_PROP_NAME &&
pl == cbp->cb_proplist)
continue;
if (pl->pl_prop == ZPROP_INVAL &&
zfs_prop_user(pl->pl_user_prop)) {
srctype = ZPROP_SRC_LOCAL;
if (zpool_get_userprop(zhp, pl->pl_user_prop,
value, sizeof (value), &srctype) != 0)
continue;
zprop_print_one_property(zpool_get_name(zhp),
cbp, pl->pl_user_prop, value, srctype,
NULL, NULL);
} else if (pl->pl_prop == ZPROP_INVAL &&
(zpool_prop_feature(pl->pl_user_prop) ||
zpool_prop_unsupported(pl->pl_user_prop))) {
srctype = ZPROP_SRC_LOCAL;
if (zpool_prop_get_feature(zhp,
pl->pl_user_prop, value,
sizeof (value)) == 0) {
zprop_print_one_property(
zpool_get_name(zhp), cbp,
pl->pl_user_prop, value, srctype,
NULL, NULL);
}
} else {
if (zpool_get_prop(zhp, pl->pl_prop, value,
sizeof (value), &srctype,
cbp->cb_literal) != 0)
continue;
zprop_print_one_property(zpool_get_name(zhp),
cbp, zpool_prop_to_name(pl->pl_prop),
value, srctype, NULL, NULL);
}
}
}
return (0);
}
/*
* zpool get [-Hp] [-o "all" | field[,...]] <"all" | property[,...]> <pool> ...
*
* -H Scripted mode. Don't display headers, and separate properties
* by a single tab.
* -o List of columns to display. Defaults to
* "name,property,value,source".
* -p Display values in parsable (exact) format.
*
* Get properties of pools in the system. Output space statistics
* for each one as well as other attributes.
*/
int
zpool_do_get(int argc, char **argv)
{
zprop_get_cbdata_t cb = { 0 };
zprop_list_t fake_name = { 0 };
int ret;
int c, i;
char *propstr = NULL;
+ char *vdev = NULL;
cb.cb_first = B_TRUE;
/*
* Set up default columns and sources.
*/
cb.cb_sources = ZPROP_SRC_ALL;
cb.cb_columns[0] = GET_COL_NAME;
cb.cb_columns[1] = GET_COL_PROPERTY;
cb.cb_columns[2] = GET_COL_VALUE;
cb.cb_columns[3] = GET_COL_SOURCE;
cb.cb_type = ZFS_TYPE_POOL;
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_TYPE_ID;
current_prop_type = cb.cb_type;
/* check options */
while ((c = getopt(argc, argv, ":Hpo:")) != -1) {
switch (c) {
case 'p':
cb.cb_literal = B_TRUE;
break;
case 'H':
cb.cb_scripted = B_TRUE;
break;
case 'o':
memset(&cb.cb_columns, 0, sizeof (cb.cb_columns));
i = 0;
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const col_opts[] =
{ "name", "property", "value", "source",
"all" };
static const zfs_get_column_t col_cols[] =
{ GET_COL_NAME, GET_COL_PROPERTY, GET_COL_VALUE,
GET_COL_SOURCE };
if (i == ZFS_GET_NCOLS - 1) {
(void) fprintf(stderr, gettext("too "
"many fields given to -o "
"option\n"));
usage(B_FALSE);
}
for (c = 0; c < ARRAY_SIZE(col_opts); ++c)
if (strcmp(tok, col_opts[c]) == 0)
goto found;
(void) fprintf(stderr,
gettext("invalid column name '%s'\n"), tok);
usage(B_FALSE);
found:
if (c >= 4) {
if (i > 0) {
(void) fprintf(stderr,
gettext("\"all\" conflicts "
"with specific fields "
"given to -o option\n"));
usage(B_FALSE);
}
memcpy(cb.cb_columns, col_cols,
sizeof (col_cols));
i = ZFS_GET_NCOLS - 1;
} else
cb.cb_columns[i++] = col_cols[c];
}
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing property "
"argument\n"));
usage(B_FALSE);
}
/* Properties list is needed later by zprop_get_list() */
propstr = argv[0];
argc--;
argv++;
if (argc == 0) {
/* No args, so just print the defaults. */
} else if (are_all_pools(argc, argv)) {
/* All the args are pool names */
} else if (are_all_pools(1, argv)) {
/* The first arg is a pool name */
if ((argc == 2 && strcmp(argv[1], "all-vdevs") == 0) ||
+ (argc == 2 && strcmp(argv[1], "root") == 0) ||
are_vdevs_in_pool(argc - 1, argv + 1, argv[0],
&cb.cb_vdevs)) {
+
+ if (strcmp(argv[1], "root") == 0)
+ vdev = strdup("root-0");
+ else
+ vdev = strdup(argv[1]);
+
/* ... and the rest are vdev names */
- cb.cb_vdevs.cb_names = argv + 1;
+ cb.cb_vdevs.cb_names = &vdev;
cb.cb_vdevs.cb_names_count = argc - 1;
cb.cb_type = ZFS_TYPE_VDEV;
argc = 1; /* One pool to process */
} else {
fprintf(stderr, gettext("Expected a list of vdevs in"
" \"%s\", but got:\n"), argv[0]);
error_list_unresolved_vdevs(argc - 1, argv + 1,
argv[0], &cb.cb_vdevs);
fprintf(stderr, "\n");
usage(B_FALSE);
return (1);
}
} else {
/*
* The first arg isn't a pool name,
*/
fprintf(stderr, gettext("missing pool name.\n"));
fprintf(stderr, "\n");
usage(B_FALSE);
return (1);
}
if (zprop_get_list(g_zfs, propstr, &cb.cb_proplist,
cb.cb_type) != 0) {
/* Use correct list of valid properties (pool or vdev) */
current_prop_type = cb.cb_type;
usage(B_FALSE);
}
if (cb.cb_proplist != NULL) {
fake_name.pl_prop = ZPOOL_PROP_NAME;
fake_name.pl_width = strlen(gettext("NAME"));
fake_name.pl_next = cb.cb_proplist;
cb.cb_proplist = &fake_name;
}
ret = for_each_pool(argc, argv, B_TRUE, &cb.cb_proplist, cb.cb_type,
cb.cb_literal, get_callback, &cb);
if (cb.cb_proplist == &fake_name)
zprop_free_list(fake_name.pl_next);
else
zprop_free_list(cb.cb_proplist);
+ if (vdev != NULL)
+ free(vdev);
+
return (ret);
}
typedef struct set_cbdata {
char *cb_propname;
char *cb_value;
zfs_type_t cb_type;
vdev_cbdata_t cb_vdevs;
boolean_t cb_any_successful;
} set_cbdata_t;
static int
set_pool_callback(zpool_handle_t *zhp, set_cbdata_t *cb)
{
int error;
/* Check if we have out-of-bounds features */
if (strcmp(cb->cb_propname, ZPOOL_CONFIG_COMPATIBILITY) == 0) {
boolean_t features[SPA_FEATURES];
if (zpool_do_load_compat(cb->cb_value, features) !=
ZPOOL_COMPATIBILITY_OK)
return (-1);
nvlist_t *enabled = zpool_get_features(zhp);
spa_feature_t i;
for (i = 0; i < SPA_FEATURES; i++) {
const char *fguid = spa_feature_table[i].fi_guid;
if (nvlist_exists(enabled, fguid) && !features[i])
break;
}
if (i < SPA_FEATURES)
(void) fprintf(stderr, gettext("Warning: one or "
"more features already enabled on pool '%s'\n"
"are not present in this compatibility set.\n"),
zpool_get_name(zhp));
}
/* if we're setting a feature, check it's in compatibility set */
if (zpool_prop_feature(cb->cb_propname) &&
strcmp(cb->cb_value, ZFS_FEATURE_ENABLED) == 0) {
char *fname = strchr(cb->cb_propname, '@') + 1;
spa_feature_t f;
if (zfeature_lookup_name(fname, &f) == 0) {
char compat[ZFS_MAXPROPLEN];
if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY,
compat, ZFS_MAXPROPLEN, NULL, B_FALSE) != 0)
compat[0] = '\0';
boolean_t features[SPA_FEATURES];
if (zpool_do_load_compat(compat, features) !=
ZPOOL_COMPATIBILITY_OK) {
(void) fprintf(stderr, gettext("Error: "
"cannot enable feature '%s' on pool '%s'\n"
"because the pool's 'compatibility' "
"property cannot be parsed.\n"),
fname, zpool_get_name(zhp));
return (-1);
}
if (!features[f]) {
(void) fprintf(stderr, gettext("Error: "
"cannot enable feature '%s' on pool '%s'\n"
"as it is not specified in this pool's "
"current compatibility set.\n"
"Consider setting 'compatibility' to a "
"less restrictive set, or to 'off'.\n"),
fname, zpool_get_name(zhp));
return (-1);
}
}
}
error = zpool_set_prop(zhp, cb->cb_propname, cb->cb_value);
return (error);
}
static int
set_callback(zpool_handle_t *zhp, void *data)
{
int error;
set_cbdata_t *cb = (set_cbdata_t *)data;
if (cb->cb_type == ZFS_TYPE_VDEV) {
error = zpool_set_vdev_prop(zhp, *cb->cb_vdevs.cb_names,
cb->cb_propname, cb->cb_value);
} else {
assert(cb->cb_type == ZFS_TYPE_POOL);
error = set_pool_callback(zhp, cb);
}
cb->cb_any_successful = !error;
return (error);
}
int
zpool_do_set(int argc, char **argv)
{
set_cbdata_t cb = { 0 };
int error;
+ char *vdev = NULL;
current_prop_type = ZFS_TYPE_POOL;
if (argc > 1 && argv[1][0] == '-') {
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
argv[1][1]);
usage(B_FALSE);
}
if (argc < 2) {
(void) fprintf(stderr, gettext("missing property=value "
"argument\n"));
usage(B_FALSE);
}
if (argc < 3) {
(void) fprintf(stderr, gettext("missing pool name\n"));
usage(B_FALSE);
}
if (argc > 4) {
(void) fprintf(stderr, gettext("too many pool names\n"));
usage(B_FALSE);
}
cb.cb_propname = argv[1];
cb.cb_type = ZFS_TYPE_POOL;
cb.cb_vdevs.cb_name_flags |= VDEV_NAME_TYPE_ID;
cb.cb_value = strchr(cb.cb_propname, '=');
if (cb.cb_value == NULL) {
(void) fprintf(stderr, gettext("missing value in "
"property=value argument\n"));
usage(B_FALSE);
}
*(cb.cb_value) = '\0';
cb.cb_value++;
argc -= 2;
argv += 2;
/* argv[0] is pool name */
if (!is_pool(argv[0])) {
(void) fprintf(stderr,
gettext("cannot open '%s': is not a pool\n"), argv[0]);
return (EINVAL);
}
/* argv[1], when supplied, is vdev name */
if (argc == 2) {
- if (!are_vdevs_in_pool(1, argv + 1, argv[0], &cb.cb_vdevs)) {
+
+ if (strcmp(argv[1], "root") == 0)
+ vdev = strdup("root-0");
+ else
+ vdev = strdup(argv[1]);
+
+ if (!are_vdevs_in_pool(1, &vdev, argv[0], &cb.cb_vdevs)) {
(void) fprintf(stderr, gettext(
"cannot find '%s' in '%s': device not in pool\n"),
- argv[1], argv[0]);
+ vdev, argv[0]);
+ free(vdev);
return (EINVAL);
}
- cb.cb_vdevs.cb_names = argv + 1;
+ cb.cb_vdevs.cb_names = &vdev;
cb.cb_vdevs.cb_names_count = 1;
cb.cb_type = ZFS_TYPE_VDEV;
}
error = for_each_pool(1, argv, B_TRUE, NULL, ZFS_TYPE_POOL,
B_FALSE, set_callback, &cb);
+ if (vdev != NULL)
+ free(vdev);
+
return (error);
}
/* Add up the total number of bytes left to initialize/trim across all vdevs */
static uint64_t
vdev_activity_remaining(nvlist_t *nv, zpool_wait_activity_t activity)
{
uint64_t bytes_remaining;
nvlist_t **child;
uint_t c, children;
vdev_stat_t *vs;
assert(activity == ZPOOL_WAIT_INITIALIZE ||
activity == ZPOOL_WAIT_TRIM);
verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t **)&vs, &c) == 0);
if (activity == ZPOOL_WAIT_INITIALIZE &&
vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE)
bytes_remaining = vs->vs_initialize_bytes_est -
vs->vs_initialize_bytes_done;
else if (activity == ZPOOL_WAIT_TRIM &&
vs->vs_trim_state == VDEV_TRIM_ACTIVE)
bytes_remaining = vs->vs_trim_bytes_est -
vs->vs_trim_bytes_done;
else
bytes_remaining = 0;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (c = 0; c < children; c++)
bytes_remaining += vdev_activity_remaining(child[c], activity);
return (bytes_remaining);
}
/* Add up the total number of bytes left to rebuild across top-level vdevs */
static uint64_t
vdev_activity_top_remaining(nvlist_t *nv)
{
uint64_t bytes_remaining = 0;
nvlist_t **child;
uint_t children;
int error;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (uint_t c = 0; c < children; c++) {
vdev_rebuild_stat_t *vrs;
uint_t i;
error = nvlist_lookup_uint64_array(child[c],
ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i);
if (error == 0) {
if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) {
bytes_remaining += (vrs->vrs_bytes_est -
vrs->vrs_bytes_rebuilt);
}
}
}
return (bytes_remaining);
}
/* Whether any vdevs are 'spare' or 'replacing' vdevs */
static boolean_t
vdev_any_spare_replacing(nvlist_t *nv)
{
nvlist_t **child;
uint_t c, children;
const char *vdev_type;
(void) nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &vdev_type);
if (strcmp(vdev_type, VDEV_TYPE_REPLACING) == 0 ||
strcmp(vdev_type, VDEV_TYPE_SPARE) == 0 ||
strcmp(vdev_type, VDEV_TYPE_DRAID_SPARE) == 0) {
return (B_TRUE);
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
children = 0;
for (c = 0; c < children; c++) {
if (vdev_any_spare_replacing(child[c]))
return (B_TRUE);
}
return (B_FALSE);
}
typedef struct wait_data {
char *wd_poolname;
boolean_t wd_scripted;
boolean_t wd_exact;
boolean_t wd_headers_once;
boolean_t wd_should_exit;
/* Which activities to wait for */
boolean_t wd_enabled[ZPOOL_WAIT_NUM_ACTIVITIES];
float wd_interval;
pthread_cond_t wd_cv;
pthread_mutex_t wd_mutex;
} wait_data_t;
/*
* Print to stdout a single line, containing one column for each activity that
* we are waiting for specifying how many bytes of work are left for that
* activity.
*/
static void
print_wait_status_row(wait_data_t *wd, zpool_handle_t *zhp, int row)
{
nvlist_t *config, *nvroot;
uint_t c;
int i;
pool_checkpoint_stat_t *pcs = NULL;
pool_scan_stat_t *pss = NULL;
pool_removal_stat_t *prs = NULL;
const char *const headers[] = {"DISCARD", "FREE", "INITIALIZE",
"REPLACE", "REMOVE", "RESILVER", "SCRUB", "TRIM"};
int col_widths[ZPOOL_WAIT_NUM_ACTIVITIES];
/* Calculate the width of each column */
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
/*
* Make sure we have enough space in the col for pretty-printed
* numbers and for the column header, and then leave a couple
* spaces between cols for readability.
*/
col_widths[i] = MAX(strlen(headers[i]), 6) + 2;
}
/* Print header if appropriate */
int term_height = terminal_height();
boolean_t reprint_header = (!wd->wd_headers_once && term_height > 0 &&
row % (term_height-1) == 0);
if (!wd->wd_scripted && (row == 0 || reprint_header)) {
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
if (wd->wd_enabled[i])
(void) printf("%*s", col_widths[i], headers[i]);
}
(void) fputc('\n', stdout);
}
/* Bytes of work remaining in each activity */
int64_t bytes_rem[ZPOOL_WAIT_NUM_ACTIVITIES] = {0};
bytes_rem[ZPOOL_WAIT_FREE] =
zpool_get_prop_int(zhp, ZPOOL_PROP_FREEING, NULL);
config = zpool_get_config(zhp, NULL);
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c);
if (pcs != NULL && pcs->pcs_state == CS_CHECKPOINT_DISCARDING)
bytes_rem[ZPOOL_WAIT_CKPT_DISCARD] = pcs->pcs_space;
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c);
if (prs != NULL && prs->prs_state == DSS_SCANNING)
bytes_rem[ZPOOL_WAIT_REMOVE] = prs->prs_to_copy -
prs->prs_copied;
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&pss, &c);
if (pss != NULL && pss->pss_state == DSS_SCANNING &&
pss->pss_pass_scrub_pause == 0) {
int64_t rem = pss->pss_to_examine - pss->pss_issued;
if (pss->pss_func == POOL_SCAN_SCRUB)
bytes_rem[ZPOOL_WAIT_SCRUB] = rem;
else
bytes_rem[ZPOOL_WAIT_RESILVER] = rem;
} else if (check_rebuilding(nvroot, NULL)) {
bytes_rem[ZPOOL_WAIT_RESILVER] =
vdev_activity_top_remaining(nvroot);
}
bytes_rem[ZPOOL_WAIT_INITIALIZE] =
vdev_activity_remaining(nvroot, ZPOOL_WAIT_INITIALIZE);
bytes_rem[ZPOOL_WAIT_TRIM] =
vdev_activity_remaining(nvroot, ZPOOL_WAIT_TRIM);
/*
* A replace finishes after resilvering finishes, so the amount of work
* left for a replace is the same as for resilvering.
*
* It isn't quite correct to say that if we have any 'spare' or
* 'replacing' vdevs and a resilver is happening, then a replace is in
* progress, like we do here. When a hot spare is used, the faulted vdev
* is not removed after the hot spare is resilvered, so parent 'spare'
* vdev is not removed either. So we could have a 'spare' vdev, but be
* resilvering for a different reason. However, we use it as a heuristic
* because we don't have access to the DTLs, which could tell us whether
* or not we have really finished resilvering a hot spare.
*/
if (vdev_any_spare_replacing(nvroot))
bytes_rem[ZPOOL_WAIT_REPLACE] = bytes_rem[ZPOOL_WAIT_RESILVER];
if (timestamp_fmt != NODATE)
print_timestamp(timestamp_fmt);
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
char buf[64];
if (!wd->wd_enabled[i])
continue;
if (wd->wd_exact)
(void) snprintf(buf, sizeof (buf), "%" PRIi64,
bytes_rem[i]);
else
zfs_nicenum(bytes_rem[i], buf, sizeof (buf));
if (wd->wd_scripted)
(void) printf(i == 0 ? "%s" : "\t%s", buf);
else
(void) printf(" %*s", col_widths[i] - 1, buf);
}
(void) printf("\n");
(void) fflush(stdout);
}
static void *
wait_status_thread(void *arg)
{
wait_data_t *wd = (wait_data_t *)arg;
zpool_handle_t *zhp;
if ((zhp = zpool_open(g_zfs, wd->wd_poolname)) == NULL)
return (void *)(1);
for (int row = 0; ; row++) {
boolean_t missing;
struct timespec timeout;
int ret = 0;
(void) clock_gettime(CLOCK_REALTIME, &timeout);
if (zpool_refresh_stats(zhp, &missing) != 0 || missing ||
zpool_props_refresh(zhp) != 0) {
zpool_close(zhp);
return (void *)(uintptr_t)(missing ? 0 : 1);
}
print_wait_status_row(wd, zhp, row);
timeout.tv_sec += floor(wd->wd_interval);
long nanos = timeout.tv_nsec +
(wd->wd_interval - floor(wd->wd_interval)) * NANOSEC;
if (nanos >= NANOSEC) {
timeout.tv_sec++;
timeout.tv_nsec = nanos - NANOSEC;
} else {
timeout.tv_nsec = nanos;
}
pthread_mutex_lock(&wd->wd_mutex);
if (!wd->wd_should_exit)
ret = pthread_cond_timedwait(&wd->wd_cv, &wd->wd_mutex,
&timeout);
pthread_mutex_unlock(&wd->wd_mutex);
if (ret == 0) {
break; /* signaled by main thread */
} else if (ret != ETIMEDOUT) {
(void) fprintf(stderr, gettext("pthread_cond_timedwait "
"failed: %s\n"), strerror(ret));
zpool_close(zhp);
return (void *)(uintptr_t)(1);
}
}
zpool_close(zhp);
return (void *)(0);
}
int
zpool_do_wait(int argc, char **argv)
{
boolean_t verbose = B_FALSE;
int c, i;
unsigned long count;
pthread_t status_thr;
int error = 0;
zpool_handle_t *zhp;
wait_data_t wd;
wd.wd_scripted = B_FALSE;
wd.wd_exact = B_FALSE;
wd.wd_headers_once = B_FALSE;
wd.wd_should_exit = B_FALSE;
pthread_mutex_init(&wd.wd_mutex, NULL);
pthread_cond_init(&wd.wd_cv, NULL);
/* By default, wait for all types of activity. */
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++)
wd.wd_enabled[i] = B_TRUE;
while ((c = getopt(argc, argv, "HpT:t:")) != -1) {
switch (c) {
case 'H':
wd.wd_scripted = B_TRUE;
break;
case 'n':
wd.wd_headers_once = B_TRUE;
break;
case 'p':
wd.wd_exact = B_TRUE;
break;
case 'T':
get_timestamp_arg(*optarg);
break;
case 't':
/* Reset activities array */
memset(&wd.wd_enabled, 0, sizeof (wd.wd_enabled));
for (char *tok; (tok = strsep(&optarg, ",")); ) {
static const char *const col_opts[] = {
"discard", "free", "initialize", "replace",
"remove", "resilver", "scrub", "trim" };
for (i = 0; i < ARRAY_SIZE(col_opts); ++i)
if (strcmp(tok, col_opts[i]) == 0) {
wd.wd_enabled[i] = B_TRUE;
goto found;
}
(void) fprintf(stderr,
gettext("invalid activity '%s'\n"), tok);
usage(B_FALSE);
found:;
}
break;
case '?':
(void) fprintf(stderr, gettext("invalid option '%c'\n"),
optopt);
usage(B_FALSE);
}
}
argc -= optind;
argv += optind;
get_interval_count(&argc, argv, &wd.wd_interval, &count);
if (count != 0) {
/* This subcmd only accepts an interval, not a count */
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
if (wd.wd_interval != 0)
verbose = B_TRUE;
if (argc < 1) {
(void) fprintf(stderr, gettext("missing 'pool' argument\n"));
usage(B_FALSE);
}
if (argc > 1) {
(void) fprintf(stderr, gettext("too many arguments\n"));
usage(B_FALSE);
}
wd.wd_poolname = argv[0];
if ((zhp = zpool_open(g_zfs, wd.wd_poolname)) == NULL)
return (1);
if (verbose) {
/*
* We use a separate thread for printing status updates because
* the main thread will call lzc_wait(), which blocks as long
* as an activity is in progress, which can be a long time.
*/
if (pthread_create(&status_thr, NULL, wait_status_thread, &wd)
!= 0) {
(void) fprintf(stderr, gettext("failed to create status"
"thread: %s\n"), strerror(errno));
zpool_close(zhp);
return (1);
}
}
/*
* Loop over all activities that we are supposed to wait for until none
* of them are in progress. Note that this means we can end up waiting
* for more activities to complete than just those that were in progress
* when we began waiting; if an activity we are interested in begins
* while we are waiting for another activity, we will wait for both to
* complete before exiting.
*/
for (;;) {
boolean_t missing = B_FALSE;
boolean_t any_waited = B_FALSE;
for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
boolean_t waited;
if (!wd.wd_enabled[i])
continue;
error = zpool_wait_status(zhp, i, &missing, &waited);
if (error != 0 || missing)
break;
any_waited = (any_waited || waited);
}
if (error != 0 || missing || !any_waited)
break;
}
zpool_close(zhp);
if (verbose) {
uintptr_t status;
pthread_mutex_lock(&wd.wd_mutex);
wd.wd_should_exit = B_TRUE;
pthread_cond_signal(&wd.wd_cv);
pthread_mutex_unlock(&wd.wd_mutex);
(void) pthread_join(status_thr, (void *)&status);
if (status != 0)
error = status;
}
pthread_mutex_destroy(&wd.wd_mutex);
pthread_cond_destroy(&wd.wd_cv);
return (error);
}
static int
find_command_idx(const char *command, int *idx)
{
for (int i = 0; i < NCOMMAND; ++i) {
if (command_table[i].name == NULL)
continue;
if (strcmp(command, command_table[i].name) == 0) {
*idx = i;
return (0);
}
}
return (1);
}
/*
* Display version message
*/
static int
zpool_do_version(int argc, char **argv)
{
(void) argc, (void) argv;
return (zfs_version_print() != 0);
}
/*
* Do zpool_load_compat() and print error message on failure
*/
static zpool_compat_status_t
zpool_do_load_compat(const char *compat, boolean_t *list)
{
char report[1024];
zpool_compat_status_t ret;
ret = zpool_load_compat(compat, list, report, 1024);
switch (ret) {
case ZPOOL_COMPATIBILITY_OK:
break;
case ZPOOL_COMPATIBILITY_NOFILES:
case ZPOOL_COMPATIBILITY_BADFILE:
case ZPOOL_COMPATIBILITY_BADTOKEN:
(void) fprintf(stderr, "Error: %s\n", report);
break;
case ZPOOL_COMPATIBILITY_WARNTOKEN:
(void) fprintf(stderr, "Warning: %s\n", report);
ret = ZPOOL_COMPATIBILITY_OK;
break;
}
return (ret);
}
int
main(int argc, char **argv)
{
int ret = 0;
int i = 0;
char *cmdname;
char **newargv;
(void) setlocale(LC_ALL, "");
(void) setlocale(LC_NUMERIC, "C");
(void) textdomain(TEXT_DOMAIN);
srand(time(NULL));
opterr = 0;
/*
* Make sure the user has specified some command.
*/
if (argc < 2) {
(void) fprintf(stderr, gettext("missing command\n"));
usage(B_FALSE);
}
cmdname = argv[1];
/*
* Special case '-?'
*/
if ((strcmp(cmdname, "-?") == 0) || strcmp(cmdname, "--help") == 0)
usage(B_TRUE);
/*
* Special case '-V|--version'
*/
if ((strcmp(cmdname, "-V") == 0) || (strcmp(cmdname, "--version") == 0))
return (zpool_do_version(argc, argv));
if ((g_zfs = libzfs_init()) == NULL) {
(void) fprintf(stderr, "%s\n", libzfs_error_init(errno));
return (1);
}
libzfs_print_on_error(g_zfs, B_TRUE);
zfs_save_arguments(argc, argv, history_str, sizeof (history_str));
/*
* Many commands modify input strings for string parsing reasons.
* We create a copy to protect the original argv.
*/
newargv = safe_malloc((argc + 1) * sizeof (newargv[0]));
for (i = 0; i < argc; i++)
newargv[i] = strdup(argv[i]);
newargv[argc] = NULL;
/*
* Run the appropriate command.
*/
if (find_command_idx(cmdname, &i) == 0) {
current_command = &command_table[i];
ret = command_table[i].func(argc - 1, newargv + 1);
} else if (strchr(cmdname, '=')) {
verify(find_command_idx("set", &i) == 0);
current_command = &command_table[i];
ret = command_table[i].func(argc, newargv);
} else if (strcmp(cmdname, "freeze") == 0 && argc == 3) {
/*
* 'freeze' is a vile debugging abomination, so we treat
* it as such.
*/
zfs_cmd_t zc = {"\0"};
(void) strlcpy(zc.zc_name, argv[2], sizeof (zc.zc_name));
ret = zfs_ioctl(g_zfs, ZFS_IOC_POOL_FREEZE, &zc);
if (ret != 0) {
(void) fprintf(stderr,
gettext("failed to freeze pool: %d\n"), errno);
ret = 1;
}
log_history = 0;
} else {
(void) fprintf(stderr, gettext("unrecognized "
"command '%s'\n"), cmdname);
usage(B_FALSE);
ret = 1;
}
for (i = 0; i < argc; i++)
free(newargv[i]);
free(newargv);
if (ret == 0 && log_history)
(void) zpool_log_history(g_zfs, history_str);
libzfs_fini(g_zfs);
/*
* The 'ZFS_ABORT' environment variable causes us to dump core on exit
* for the purposes of running ::findleaks.
*/
if (getenv("ZFS_ABORT") != NULL) {
(void) printf("dumping core by request\n");
abort();
}
return (ret);
}
diff --git a/sys/contrib/openzfs/config/kernel-acl.m4 b/sys/contrib/openzfs/config/kernel-acl.m4
index 6e92da97d0fe..be08c3c60724 100644
--- a/sys/contrib/openzfs/config/kernel-acl.m4
+++ b/sys/contrib/openzfs/config/kernel-acl.m4
@@ -1,373 +1,395 @@
dnl #
dnl # Check if posix_acl_release can be used from a ZFS_META_LICENSED
dnl # module. The is_owner_or_cap macro was replaced by
dnl # inode_owner_or_capable
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_POSIX_ACL_RELEASE], [
ZFS_LINUX_TEST_SRC([posix_acl_release], [
#include <linux/cred.h>
#include <linux/fs.h>
#include <linux/posix_acl.h>
], [
struct posix_acl *tmp = posix_acl_alloc(1, 0);
posix_acl_release(tmp);
], [], [ZFS_META_LICENSE])
])
AC_DEFUN([ZFS_AC_KERNEL_POSIX_ACL_RELEASE], [
AC_MSG_CHECKING([whether posix_acl_release() is available])
ZFS_LINUX_TEST_RESULT([posix_acl_release], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_POSIX_ACL_RELEASE, 1,
[posix_acl_release() is available])
AC_MSG_CHECKING([whether posix_acl_release() is GPL-only])
ZFS_LINUX_TEST_RESULT([posix_acl_release_license], [
AC_MSG_RESULT(no)
],[
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_POSIX_ACL_RELEASE_GPL_ONLY, 1,
[posix_acl_release() is GPL-only])
])
],[
AC_MSG_RESULT(no)
])
])
dnl #
dnl # 3.14 API change,
dnl # set_cached_acl() and forget_cached_acl() changed from inline to
dnl # EXPORT_SYMBOL. In the former case, they may not be usable because of
dnl # posix_acl_release. In the latter case, we can always use them.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_SET_CACHED_ACL_USABLE], [
ZFS_LINUX_TEST_SRC([set_cached_acl], [
#include <linux/cred.h>
#include <linux/fs.h>
#include <linux/posix_acl.h>
], [
struct inode *ip = NULL;
struct posix_acl *acl = posix_acl_alloc(1, 0);
set_cached_acl(ip, ACL_TYPE_ACCESS, acl);
forget_cached_acl(ip, ACL_TYPE_ACCESS);
], [], [ZFS_META_LICENSE])
])
AC_DEFUN([ZFS_AC_KERNEL_SET_CACHED_ACL_USABLE], [
AC_MSG_CHECKING([whether set_cached_acl() is usable])
ZFS_LINUX_TEST_RESULT([set_cached_acl_license], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_SET_CACHED_ACL_USABLE, 1,
[set_cached_acl() is usable])
],[
AC_MSG_RESULT(no)
])
])
dnl #
dnl # 3.1 API change,
dnl # posix_acl_chmod() was added as the preferred interface.
dnl #
dnl # 3.14 API change,
dnl # posix_acl_chmod() was changed to __posix_acl_chmod()
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_POSIX_ACL_CHMOD], [
ZFS_LINUX_TEST_SRC([posix_acl_chmod], [
#include <linux/fs.h>
#include <linux/posix_acl.h>
],[
posix_acl_chmod(NULL, 0, 0)
])
ZFS_LINUX_TEST_SRC([__posix_acl_chmod], [
#include <linux/fs.h>
#include <linux/posix_acl.h>
],[
__posix_acl_chmod(NULL, 0, 0)
])
])
AC_DEFUN([ZFS_AC_KERNEL_POSIX_ACL_CHMOD], [
AC_MSG_CHECKING([whether __posix_acl_chmod exists])
ZFS_LINUX_TEST_RESULT([__posix_acl_chmod], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE___POSIX_ACL_CHMOD, 1,
[__posix_acl_chmod() exists])
],[
AC_MSG_RESULT(no)
AC_MSG_CHECKING([whether posix_acl_chmod exists])
ZFS_LINUX_TEST_RESULT([posix_acl_chmod], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_POSIX_ACL_CHMOD, 1,
[posix_acl_chmod() exists])
],[
ZFS_LINUX_TEST_ERROR([posix_acl_chmod()])
])
])
])
dnl #
dnl # 3.1 API change,
dnl # posix_acl_equiv_mode now wants an umode_t instead of a mode_t
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_POSIX_ACL_EQUIV_MODE_WANTS_UMODE_T], [
ZFS_LINUX_TEST_SRC([posix_acl_equiv_mode], [
#include <linux/fs.h>
#include <linux/posix_acl.h>
],[
umode_t tmp;
posix_acl_equiv_mode(NULL, &tmp);
])
])
AC_DEFUN([ZFS_AC_KERNEL_POSIX_ACL_EQUIV_MODE_WANTS_UMODE_T], [
AC_MSG_CHECKING([whether posix_acl_equiv_mode() wants umode_t])
ZFS_LINUX_TEST_RESULT([posix_acl_equiv_mode], [
AC_MSG_RESULT(yes)
],[
ZFS_LINUX_TEST_ERROR([posix_acl_equiv_mode()])
])
])
dnl #
dnl # 4.8 API change,
dnl # The function posix_acl_valid now must be passed a namespace.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_POSIX_ACL_VALID_WITH_NS], [
ZFS_LINUX_TEST_SRC([posix_acl_valid_with_ns], [
#include <linux/fs.h>
#include <linux/posix_acl.h>
],[
struct user_namespace *user_ns = NULL;
const struct posix_acl *acl = NULL;
int error;
error = posix_acl_valid(user_ns, acl);
])
])
AC_DEFUN([ZFS_AC_KERNEL_POSIX_ACL_VALID_WITH_NS], [
AC_MSG_CHECKING([whether posix_acl_valid() wants user namespace])
ZFS_LINUX_TEST_RESULT([posix_acl_valid_with_ns], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_POSIX_ACL_VALID_WITH_NS, 1,
[posix_acl_valid() wants user namespace])
],[
AC_MSG_RESULT(no)
])
])
dnl #
dnl # 3.1 API change,
dnl # Check if inode_operations contains the function get_acl
dnl #
dnl # 5.15 API change,
dnl # Added the bool rcu argument to get_acl for rcu path walk.
dnl #
dnl # 6.2 API change,
dnl # get_acl() was renamed to get_inode_acl()
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_GET_ACL], [
ZFS_LINUX_TEST_SRC([inode_operations_get_acl], [
#include <linux/fs.h>
struct posix_acl *get_acl_fn(struct inode *inode, int type)
{ return NULL; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.get_acl = get_acl_fn,
};
],[])
ZFS_LINUX_TEST_SRC([inode_operations_get_acl_rcu], [
#include <linux/fs.h>
struct posix_acl *get_acl_fn(struct inode *inode, int type,
bool rcu) { return NULL; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.get_acl = get_acl_fn,
};
],[])
ZFS_LINUX_TEST_SRC([inode_operations_get_inode_acl], [
#include <linux/fs.h>
struct posix_acl *get_inode_acl_fn(struct inode *inode, int type,
bool rcu) { return NULL; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.get_inode_acl = get_inode_acl_fn,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_INODE_OPERATIONS_GET_ACL], [
AC_MSG_CHECKING([whether iops->get_acl() exists])
ZFS_LINUX_TEST_RESULT([inode_operations_get_acl], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_GET_ACL, 1, [iops->get_acl() exists])
],[
ZFS_LINUX_TEST_RESULT([inode_operations_get_acl_rcu], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_GET_ACL_RCU, 1, [iops->get_acl() takes rcu])
],[
ZFS_LINUX_TEST_RESULT([inode_operations_get_inode_acl], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_GET_INODE_ACL, 1, [has iops->get_inode_acl()])
],[
ZFS_LINUX_TEST_ERROR([iops->get_acl() or iops->get_inode_acl()])
])
])
])
])
dnl #
dnl # 3.14 API change,
dnl # Check if inode_operations contains the function set_acl
dnl #
dnl # 5.12 API change,
dnl # set_acl() added a user_namespace* parameter first
dnl #
dnl # 6.2 API change,
dnl # set_acl() second paramter changed to a struct dentry *
dnl #
+dnl # 6.3 API change,
+dnl # set_acl() first parameter changed to struct mnt_idmap *
+dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_SET_ACL], [
+ ZFS_LINUX_TEST_SRC([inode_operations_set_acl_mnt_idmap_dentry], [
+ #include <linux/fs.h>
+
+ int set_acl_fn(struct mnt_idmap *idmap,
+ struct dentry *dent, struct posix_acl *acl,
+ int type) { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .set_acl = set_acl_fn,
+ };
+ ],[])
ZFS_LINUX_TEST_SRC([inode_operations_set_acl_userns_dentry], [
#include <linux/fs.h>
int set_acl_fn(struct user_namespace *userns,
struct dentry *dent, struct posix_acl *acl,
int type) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.set_acl = set_acl_fn,
};
],[])
ZFS_LINUX_TEST_SRC([inode_operations_set_acl_userns], [
#include <linux/fs.h>
int set_acl_fn(struct user_namespace *userns,
struct inode *inode, struct posix_acl *acl,
int type) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.set_acl = set_acl_fn,
};
],[])
ZFS_LINUX_TEST_SRC([inode_operations_set_acl], [
#include <linux/fs.h>
int set_acl_fn(struct inode *inode, struct posix_acl *acl,
int type) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.set_acl = set_acl_fn,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_INODE_OPERATIONS_SET_ACL], [
AC_MSG_CHECKING([whether iops->set_acl() exists])
ZFS_LINUX_TEST_RESULT([inode_operations_set_acl_userns], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists])
AC_DEFINE(HAVE_SET_ACL_USERNS, 1, [iops->set_acl() takes 4 args])
],[
- ZFS_LINUX_TEST_RESULT([inode_operations_set_acl_userns_dentry], [
+ ZFS_LINUX_TEST_RESULT([inode_operations_set_acl_mnt_idmap_dentry], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists])
- AC_DEFINE(HAVE_SET_ACL_USERNS_DENTRY_ARG2, 1,
- [iops->set_acl() takes 4 args, arg2 is struct dentry *])
+ AC_DEFINE(HAVE_SET_ACL_IDMAP_DENTRY, 1,
+ [iops->set_acl() takes 4 args, arg1 is struct mnt_idmap *])
],[
- ZFS_LINUX_TEST_RESULT([inode_operations_set_acl], [
+ ZFS_LINUX_TEST_RESULT([inode_operations_set_acl_userns_dentry], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists, takes 3 args])
+ AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists])
+ AC_DEFINE(HAVE_SET_ACL_USERNS_DENTRY_ARG2, 1,
+ [iops->set_acl() takes 4 args, arg2 is struct dentry *])
],[
- ZFS_LINUX_REQUIRE_API([i_op->set_acl()], [3.14])
+ ZFS_LINUX_TEST_RESULT([inode_operations_set_acl], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_SET_ACL, 1, [iops->set_acl() exists, takes 3 args])
+ ],[
+ ZFS_LINUX_REQUIRE_API([i_op->set_acl()], [3.14])
+ ])
])
])
])
])
dnl #
dnl # 4.7 API change,
dnl # The kernel get_acl will now check cache before calling i_op->get_acl and
dnl # do set_cached_acl after that, so i_op->get_acl don't need to do that
dnl # anymore.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_GET_ACL_HANDLE_CACHE], [
ZFS_LINUX_TEST_SRC([get_acl_handle_cache], [
#include <linux/fs.h>
],[
void *sentinel __attribute__ ((unused)) =
uncached_acl_sentinel(NULL);
])
])
AC_DEFUN([ZFS_AC_KERNEL_GET_ACL_HANDLE_CACHE], [
AC_MSG_CHECKING([whether uncached_acl_sentinel() exists])
ZFS_LINUX_TEST_RESULT([get_acl_handle_cache], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_KERNEL_GET_ACL_HANDLE_CACHE, 1,
[uncached_acl_sentinel() exists])
],[
AC_MSG_RESULT(no)
])
])
dnl #
dnl # 4.16 kernel: check if struct posix_acl acl.a_refcount is a refcount_t.
dnl # It's an atomic_t on older kernels.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_ACL_HAS_REFCOUNT], [
ZFS_LINUX_TEST_SRC([acl_refcount], [
#include <linux/backing-dev.h>
#include <linux/refcount.h>
#include <linux/posix_acl.h>
],[
struct posix_acl acl;
refcount_t *r __attribute__ ((unused)) = &acl.a_refcount;
])
])
AC_DEFUN([ZFS_AC_KERNEL_ACL_HAS_REFCOUNT], [
AC_MSG_CHECKING([whether posix_acl has refcount_t])
ZFS_LINUX_TEST_RESULT([acl_refcount], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_ACL_REFCOUNT, 1, [posix_acl has refcount_t])
],[
AC_MSG_RESULT(no)
])
])
AC_DEFUN([ZFS_AC_KERNEL_SRC_ACL], [
ZFS_AC_KERNEL_SRC_POSIX_ACL_RELEASE
ZFS_AC_KERNEL_SRC_SET_CACHED_ACL_USABLE
ZFS_AC_KERNEL_SRC_POSIX_ACL_CHMOD
ZFS_AC_KERNEL_SRC_POSIX_ACL_EQUIV_MODE_WANTS_UMODE_T
ZFS_AC_KERNEL_SRC_POSIX_ACL_VALID_WITH_NS
ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_GET_ACL
ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_SET_ACL
ZFS_AC_KERNEL_SRC_GET_ACL_HANDLE_CACHE
ZFS_AC_KERNEL_SRC_ACL_HAS_REFCOUNT
])
AC_DEFUN([ZFS_AC_KERNEL_ACL], [
ZFS_AC_KERNEL_POSIX_ACL_RELEASE
ZFS_AC_KERNEL_SET_CACHED_ACL_USABLE
ZFS_AC_KERNEL_POSIX_ACL_CHMOD
ZFS_AC_KERNEL_POSIX_ACL_EQUIV_MODE_WANTS_UMODE_T
ZFS_AC_KERNEL_POSIX_ACL_VALID_WITH_NS
ZFS_AC_KERNEL_INODE_OPERATIONS_GET_ACL
ZFS_AC_KERNEL_INODE_OPERATIONS_SET_ACL
ZFS_AC_KERNEL_GET_ACL_HANDLE_CACHE
ZFS_AC_KERNEL_ACL_HAS_REFCOUNT
])
diff --git a/sys/contrib/openzfs/config/kernel-generic_fillattr.m4 b/sys/contrib/openzfs/config/kernel-generic_fillattr.m4
index 0acd5d53103f..02dee4d4c000 100644
--- a/sys/contrib/openzfs/config/kernel-generic_fillattr.m4
+++ b/sys/contrib/openzfs/config/kernel-generic_fillattr.m4
@@ -1,28 +1,47 @@
dnl #
dnl # 5.12 API
dnl #
dnl # generic_fillattr in linux/fs.h now requires a struct user_namespace*
dnl # as the first arg, to support idmapped mounts.
dnl #
-AC_DEFUN([ZFS_AC_KERNEL_SRC_GENERIC_FILLATTR_USERNS], [
+dnl # 6.3 API
+dnl # generic_fillattr() now takes struct mnt_idmap* as the first argument
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_GENERIC_FILLATTR], [
ZFS_LINUX_TEST_SRC([generic_fillattr_userns], [
#include <linux/fs.h>
],[
struct user_namespace *userns = NULL;
struct inode *in = NULL;
struct kstat *k = NULL;
generic_fillattr(userns, in, k);
])
+
+ ZFS_LINUX_TEST_SRC([generic_fillattr_mnt_idmap], [
+ #include <linux/fs.h>
+ ],[
+ struct mnt_idmap *idmap = NULL;
+ struct inode *in = NULL;
+ struct kstat *k = NULL;
+ generic_fillattr(idmap, in, k);
+ ])
])
-AC_DEFUN([ZFS_AC_KERNEL_GENERIC_FILLATTR_USERNS], [
- AC_MSG_CHECKING([whether generic_fillattr requires struct user_namespace*])
- ZFS_LINUX_TEST_RESULT([generic_fillattr_userns], [
+AC_DEFUN([ZFS_AC_KERNEL_GENERIC_FILLATTR], [
+ AC_MSG_CHECKING([whether generic_fillattr requires struct mnt_idmap*])
+ ZFS_LINUX_TEST_RESULT([generic_fillattr_mnt_idmap], [
AC_MSG_RESULT([yes])
- AC_DEFINE(HAVE_GENERIC_FILLATTR_USERNS, 1,
- [generic_fillattr requires struct user_namespace*])
+ AC_DEFINE(HAVE_GENERIC_FILLATTR_IDMAP, 1,
+ [generic_fillattr requires struct mnt_idmap*])
],[
- AC_MSG_RESULT([no])
+ AC_MSG_CHECKING([whether generic_fillattr requires struct user_namespace*])
+ ZFS_LINUX_TEST_RESULT([generic_fillattr_userns], [
+ AC_MSG_RESULT([yes])
+ AC_DEFINE(HAVE_GENERIC_FILLATTR_USERNS, 1,
+ [generic_fillattr requires struct user_namespace*])
+ ],[
+ AC_MSG_RESULT([no])
+ ])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-inode-create.m4 b/sys/contrib/openzfs/config/kernel-inode-create.m4
index a6ea11fb61b2..9e9e43180976 100644
--- a/sys/contrib/openzfs/config/kernel-inode-create.m4
+++ b/sys/contrib/openzfs/config/kernel-inode-create.m4
@@ -1,53 +1,80 @@
AC_DEFUN([ZFS_AC_KERNEL_SRC_CREATE], [
+ dnl #
+ dnl # 6.3 API change
+ dnl # The first arg is changed to struct mnt_idmap *
+ dnl #
+ ZFS_LINUX_TEST_SRC([create_mnt_idmap], [
+ #include <linux/fs.h>
+ #include <linux/sched.h>
+
+ int inode_create(struct mnt_idmap *idmap,
+ struct inode *inode ,struct dentry *dentry,
+ umode_t umode, bool flag) { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .create = inode_create,
+ };
+ ],[])
+
dnl #
dnl # 5.12 API change that added the struct user_namespace* arg
dnl # to the front of this function type's arg list.
dnl #
ZFS_LINUX_TEST_SRC([create_userns], [
#include <linux/fs.h>
#include <linux/sched.h>
int inode_create(struct user_namespace *userns,
struct inode *inode ,struct dentry *dentry,
umode_t umode, bool flag) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.create = inode_create,
};
],[])
dnl #
dnl # 3.6 API change
dnl #
ZFS_LINUX_TEST_SRC([create_flags], [
#include <linux/fs.h>
#include <linux/sched.h>
int inode_create(struct inode *inode ,struct dentry *dentry,
umode_t umode, bool flag) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.create = inode_create,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_CREATE], [
- AC_MSG_CHECKING([whether iops->create() takes struct user_namespace*])
- ZFS_LINUX_TEST_RESULT([create_userns], [
+ AC_MSG_CHECKING([whether iops->create() takes struct mnt_idmap*])
+ ZFS_LINUX_TEST_RESULT([create_mnt_idmap], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_IOPS_CREATE_USERNS, 1,
- [iops->create() takes struct user_namespace*])
+ AC_DEFINE(HAVE_IOPS_CREATE_IDMAP, 1,
+ [iops->create() takes struct mnt_idmap*])
],[
AC_MSG_RESULT(no)
- AC_MSG_CHECKING([whether iops->create() passes flags])
- ZFS_LINUX_TEST_RESULT([create_flags], [
+ AC_MSG_CHECKING([whether iops->create() takes struct user_namespace*])
+ ZFS_LINUX_TEST_RESULT([create_userns], [
AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_IOPS_CREATE_USERNS, 1,
+ [iops->create() takes struct user_namespace*])
],[
- ZFS_LINUX_TEST_ERROR([iops->create()])
+ AC_MSG_RESULT(no)
+
+ AC_MSG_CHECKING([whether iops->create() passes flags])
+ ZFS_LINUX_TEST_RESULT([create_flags], [
+ AC_MSG_RESULT(yes)
+ ],[
+ ZFS_LINUX_TEST_ERROR([iops->create()])
+ ])
])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-inode-getattr.m4 b/sys/contrib/openzfs/config/kernel-inode-getattr.m4
index f62e82f5230a..c8bfb07862ab 100644
--- a/sys/contrib/openzfs/config/kernel-inode-getattr.m4
+++ b/sys/contrib/openzfs/config/kernel-inode-getattr.m4
@@ -1,92 +1,123 @@
AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_GETATTR], [
+ dnl #
+ dnl # Linux 6.3 API
+ dnl # The first arg of getattr I/O operations handler type
+ dnl # is changed to struct mnt_idmap*
+ dnl #
+ ZFS_LINUX_TEST_SRC([inode_operations_getattr_mnt_idmap], [
+ #include <linux/fs.h>
+
+ int test_getattr(
+ struct mnt_idmap *idmap,
+ const struct path *p, struct kstat *k,
+ u32 request_mask, unsigned int query_flags)
+ { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .getattr = test_getattr,
+ };
+ ],[])
+
dnl #
dnl # Linux 5.12 API
dnl # The getattr I/O operations handler type was extended to require
dnl # a struct user_namespace* as its first arg, to support idmapped
dnl # mounts.
dnl #
ZFS_LINUX_TEST_SRC([inode_operations_getattr_userns], [
#include <linux/fs.h>
int test_getattr(
struct user_namespace *userns,
const struct path *p, struct kstat *k,
u32 request_mask, unsigned int query_flags)
{ return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.getattr = test_getattr,
};
],[])
dnl #
dnl # Linux 4.11 API
dnl # See torvalds/linux@a528d35
dnl #
ZFS_LINUX_TEST_SRC([inode_operations_getattr_path], [
#include <linux/fs.h>
int test_getattr(
const struct path *p, struct kstat *k,
u32 request_mask, unsigned int query_flags)
{ return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.getattr = test_getattr,
};
],[])
ZFS_LINUX_TEST_SRC([inode_operations_getattr_vfsmount], [
#include <linux/fs.h>
int test_getattr(
struct vfsmount *mnt, struct dentry *d,
struct kstat *k)
{ return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.getattr = test_getattr,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_INODE_GETATTR], [
dnl #
- dnl # Kernel 5.12 test
+ dnl # Kernel 6.3 test
dnl #
- AC_MSG_CHECKING([whether iops->getattr() takes user_namespace])
- ZFS_LINUX_TEST_RESULT([inode_operations_getattr_userns], [
+ AC_MSG_CHECKING([whether iops->getattr() takes mnt_idmap])
+ ZFS_LINUX_TEST_RESULT([inode_operations_getattr_mnt_idmap], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_USERNS_IOPS_GETATTR, 1,
- [iops->getattr() takes struct user_namespace*])
+ AC_DEFINE(HAVE_IDMAP_IOPS_GETATTR, 1,
+ [iops->getattr() takes struct mnt_idmap*])
],[
AC_MSG_RESULT(no)
-
dnl #
- dnl # Kernel 4.11 test
+ dnl # Kernel 5.12 test
dnl #
- AC_MSG_CHECKING([whether iops->getattr() takes a path])
- ZFS_LINUX_TEST_RESULT([inode_operations_getattr_path], [
+ AC_MSG_CHECKING([whether iops->getattr() takes user_namespace])
+ ZFS_LINUX_TEST_RESULT([inode_operations_getattr_userns], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_PATH_IOPS_GETATTR, 1,
- [iops->getattr() takes a path])
+ AC_DEFINE(HAVE_USERNS_IOPS_GETATTR, 1,
+ [iops->getattr() takes struct user_namespace*])
],[
AC_MSG_RESULT(no)
dnl #
- dnl # Kernel < 4.11 test
+ dnl # Kernel 4.11 test
dnl #
- AC_MSG_CHECKING([whether iops->getattr() takes a vfsmount])
- ZFS_LINUX_TEST_RESULT([inode_operations_getattr_vfsmount], [
+ AC_MSG_CHECKING([whether iops->getattr() takes a path])
+ ZFS_LINUX_TEST_RESULT([inode_operations_getattr_path], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_VFSMOUNT_IOPS_GETATTR, 1,
- [iops->getattr() takes a vfsmount])
+ AC_DEFINE(HAVE_PATH_IOPS_GETATTR, 1,
+ [iops->getattr() takes a path])
],[
AC_MSG_RESULT(no)
+
+ dnl #
+ dnl # Kernel < 4.11 test
+ dnl #
+ AC_MSG_CHECKING([whether iops->getattr() takes a vfsmount])
+ ZFS_LINUX_TEST_RESULT([inode_operations_getattr_vfsmount], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_VFSMOUNT_IOPS_GETATTR, 1,
+ [iops->getattr() takes a vfsmount])
+ ],[
+ AC_MSG_RESULT(no)
+ ])
])
])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-inode-permission.m4 b/sys/contrib/openzfs/config/kernel-inode-permission.m4
index ba9ff5d43d4d..01d23635b0c9 100644
--- a/sys/contrib/openzfs/config/kernel-inode-permission.m4
+++ b/sys/contrib/openzfs/config/kernel-inode-permission.m4
@@ -1,29 +1,54 @@
AC_DEFUN([ZFS_AC_KERNEL_SRC_PERMISSION], [
+ dnl #
+ dnl # 6.3 API change
+ dnl # iops->permission() now takes struct mnt_idmap*
+ dnl # as its first arg
+ dnl #
+ ZFS_LINUX_TEST_SRC([permission_mnt_idmap], [
+ #include <linux/fs.h>
+ #include <linux/sched.h>
+
+ int inode_permission(struct mnt_idmap *idmap,
+ struct inode *inode, int mask) { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .permission = inode_permission,
+ };
+ ],[])
+
dnl #
dnl # 5.12 API change that added the struct user_namespace* arg
dnl # to the front of this function type's arg list.
dnl #
ZFS_LINUX_TEST_SRC([permission_userns], [
#include <linux/fs.h>
#include <linux/sched.h>
int inode_permission(struct user_namespace *userns,
struct inode *inode, int mask) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.permission = inode_permission,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_PERMISSION], [
- AC_MSG_CHECKING([whether iops->permission() takes struct user_namespace*])
- ZFS_LINUX_TEST_RESULT([permission_userns], [
+ AC_MSG_CHECKING([whether iops->permission() takes struct mnt_idmap*])
+ ZFS_LINUX_TEST_RESULT([permission_mnt_idmap], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_IOPS_PERMISSION_USERNS, 1,
- [iops->permission() takes struct user_namespace*])
+ AC_DEFINE(HAVE_IOPS_PERMISSION_IDMAP, 1,
+ [iops->permission() takes struct mnt_idmap*])
],[
- AC_MSG_RESULT(no)
+ AC_MSG_CHECKING([whether iops->permission() takes struct user_namespace*])
+ ZFS_LINUX_TEST_RESULT([permission_userns], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_IOPS_PERMISSION_USERNS, 1,
+ [iops->permission() takes struct user_namespace*])
+ ],[
+ AC_MSG_RESULT(no)
+ ])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-inode-setattr.m4 b/sys/contrib/openzfs/config/kernel-inode-setattr.m4
new file mode 100644
index 000000000000..45755b4eb273
--- /dev/null
+++ b/sys/contrib/openzfs/config/kernel-inode-setattr.m4
@@ -0,0 +1,87 @@
+AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_SETATTR], [
+ dnl #
+ dnl # Linux 6.3 API
+ dnl # The first arg of setattr I/O operations handler type
+ dnl # is changed to struct mnt_idmap*
+ dnl #
+ ZFS_LINUX_TEST_SRC([inode_operations_setattr_mnt_idmap], [
+ #include <linux/fs.h>
+
+ int test_setattr(
+ struct mnt_idmap *idmap,
+ struct dentry *de, struct iattr *ia)
+ { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .setattr = test_setattr,
+ };
+ ],[])
+
+ dnl #
+ dnl # Linux 5.12 API
+ dnl # The setattr I/O operations handler type was extended to require
+ dnl # a struct user_namespace* as its first arg, to support idmapped
+ dnl # mounts.
+ dnl #
+ ZFS_LINUX_TEST_SRC([inode_operations_setattr_userns], [
+ #include <linux/fs.h>
+
+ int test_setattr(
+ struct user_namespace *userns,
+ struct dentry *de, struct iattr *ia)
+ { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .setattr = test_setattr,
+ };
+ ],[])
+
+ ZFS_LINUX_TEST_SRC([inode_operations_setattr], [
+ #include <linux/fs.h>
+
+ int test_setattr(
+ struct dentry *de, struct iattr *ia)
+ { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .setattr = test_setattr,
+ };
+ ],[])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_INODE_SETATTR], [
+ dnl #
+ dnl # Kernel 6.3 test
+ dnl #
+ AC_MSG_CHECKING([whether iops->setattr() takes mnt_idmap])
+ ZFS_LINUX_TEST_RESULT([inode_operations_setattr_mnt_idmap], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_IDMAP_IOPS_SETATTR, 1,
+ [iops->setattr() takes struct mnt_idmap*])
+ ],[
+ AC_MSG_RESULT(no)
+ dnl #
+ dnl # Kernel 5.12 test
+ dnl #
+ AC_MSG_CHECKING([whether iops->setattr() takes user_namespace])
+ ZFS_LINUX_TEST_RESULT([inode_operations_setattr_userns], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_USERNS_IOPS_SETATTR, 1,
+ [iops->setattr() takes struct user_namespace*])
+ ],[
+ AC_MSG_RESULT(no)
+
+ AC_MSG_CHECKING([whether iops->setattr() exists])
+ ZFS_LINUX_TEST_RESULT([inode_operations_setattr], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_IOPS_SETATTR, 1,
+ [iops->setattr() exists])
+ ],[
+ AC_MSG_RESULT(no)
+ ])
+ ])
+ ])
+])
diff --git a/sys/contrib/openzfs/config/kernel-is_owner_or_cap.m4 b/sys/contrib/openzfs/config/kernel-is_owner_or_cap.m4
index a90cf3da641d..4e9c002b77f2 100644
--- a/sys/contrib/openzfs/config/kernel-is_owner_or_cap.m4
+++ b/sys/contrib/openzfs/config/kernel-is_owner_or_cap.m4
@@ -1,46 +1,63 @@
dnl #
dnl # 2.6.39 API change,
dnl # The is_owner_or_cap() macro was renamed to inode_owner_or_capable(),
dnl # This is used for permission checks in the xattr and file attribute call
dnl # paths.
dnl #
dnl # 5.12 API change,
dnl # inode_owner_or_capable() now takes struct user_namespace *
dnl # to support idmapped mounts
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OWNER_OR_CAPABLE], [
ZFS_LINUX_TEST_SRC([inode_owner_or_capable], [
#include <linux/fs.h>
],[
struct inode *ip = NULL;
(void) inode_owner_or_capable(ip);
])
- ZFS_LINUX_TEST_SRC([inode_owner_or_capable_idmapped], [
+ ZFS_LINUX_TEST_SRC([inode_owner_or_capable_userns], [
#include <linux/fs.h>
],[
struct inode *ip = NULL;
(void) inode_owner_or_capable(&init_user_ns, ip);
])
+
+ ZFS_LINUX_TEST_SRC([inode_owner_or_capable_mnt_idmap], [
+ #include <linux/fs.h>
+ #include <linux/mnt_idmapping.h>
+ ],[
+ struct inode *ip = NULL;
+ (void) inode_owner_or_capable(&nop_mnt_idmap, ip);
+ ])
])
AC_DEFUN([ZFS_AC_KERNEL_INODE_OWNER_OR_CAPABLE], [
AC_MSG_CHECKING([whether inode_owner_or_capable() exists])
ZFS_LINUX_TEST_RESULT([inode_owner_or_capable], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_INODE_OWNER_OR_CAPABLE, 1,
[inode_owner_or_capable() exists])
], [
AC_MSG_RESULT(no)
AC_MSG_CHECKING(
[whether inode_owner_or_capable() takes user_ns])
- ZFS_LINUX_TEST_RESULT([inode_owner_or_capable_idmapped], [
+ ZFS_LINUX_TEST_RESULT([inode_owner_or_capable_userns], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_INODE_OWNER_OR_CAPABLE_IDMAPPED, 1,
+ AC_DEFINE(HAVE_INODE_OWNER_OR_CAPABLE_USERNS, 1,
[inode_owner_or_capable() takes user_ns])
],[
- ZFS_LINUX_TEST_ERROR([capability])
+ AC_MSG_RESULT(no)
+ AC_MSG_CHECKING(
+ [whether inode_owner_or_capable() takes mnt_idmap])
+ ZFS_LINUX_TEST_RESULT([inode_owner_or_capable_mnt_idmap], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_INODE_OWNER_OR_CAPABLE_IDMAP, 1,
+ [inode_owner_or_capable() takes mnt_idmap])
+ ], [
+ ZFS_LINUX_TEST_ERROR([capability])
+ ])
])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-mkdir.m4 b/sys/contrib/openzfs/config/kernel-mkdir.m4
index 6667ed04fa4c..7407a791b846 100644
--- a/sys/contrib/openzfs/config/kernel-mkdir.m4
+++ b/sys/contrib/openzfs/config/kernel-mkdir.m4
@@ -1,67 +1,94 @@
dnl #
dnl # Supported mkdir() interfaces checked newest to oldest.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_MKDIR], [
+ dnl #
+ dnl # 6.3 API change
+ dnl # mkdir() takes struct mnt_idmap * as the first arg
+ dnl #
+ ZFS_LINUX_TEST_SRC([mkdir_mnt_idmap], [
+ #include <linux/fs.h>
+
+ int mkdir(struct mnt_idmap *idmap,
+ struct inode *inode, struct dentry *dentry,
+ umode_t umode) { return 0; }
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .mkdir = mkdir,
+ };
+ ],[])
+
dnl #
dnl # 5.12 API change
dnl # The struct user_namespace arg was added as the first argument to
dnl # mkdir()
dnl #
ZFS_LINUX_TEST_SRC([mkdir_user_namespace], [
#include <linux/fs.h>
int mkdir(struct user_namespace *userns,
struct inode *inode, struct dentry *dentry,
umode_t umode) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.mkdir = mkdir,
};
],[])
dnl #
dnl # 3.3 API change
dnl # The VFS .create, .mkdir and .mknod callbacks were updated to take a
dnl # umode_t type rather than an int. The expectation is that any backport
dnl # would also change all three prototypes. However, if it turns out that
dnl # some distribution doesn't backport the whole thing this could be
dnl # broken apart into three separate checks.
dnl #
ZFS_LINUX_TEST_SRC([inode_operations_mkdir], [
#include <linux/fs.h>
int mkdir(struct inode *inode, struct dentry *dentry,
umode_t umode) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.mkdir = mkdir,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_MKDIR], [
dnl #
- dnl # 5.12 API change
- dnl # The struct user_namespace arg was added as the first argument to
- dnl # mkdir() of the iops structure.
+ dnl # 6.3 API change
+ dnl # mkdir() takes struct mnt_idmap * as the first arg
dnl #
- AC_MSG_CHECKING([whether iops->mkdir() takes struct user_namespace*])
- ZFS_LINUX_TEST_RESULT([mkdir_user_namespace], [
+ AC_MSG_CHECKING([whether iops->mkdir() takes struct mnt_idmap*])
+ ZFS_LINUX_TEST_RESULT([mkdir_mnt_idmap], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_IOPS_MKDIR_USERNS, 1,
- [iops->mkdir() takes struct user_namespace*])
+ AC_DEFINE(HAVE_IOPS_MKDIR_IDMAP, 1,
+ [iops->mkdir() takes struct mnt_idmap*])
],[
- AC_MSG_RESULT(no)
-
- AC_MSG_CHECKING([whether iops->mkdir() takes umode_t])
- ZFS_LINUX_TEST_RESULT([inode_operations_mkdir], [
+ dnl #
+ dnl # 5.12 API change
+ dnl # The struct user_namespace arg was added as the first argument to
+ dnl # mkdir() of the iops structure.
+ dnl #
+ AC_MSG_CHECKING([whether iops->mkdir() takes struct user_namespace*])
+ ZFS_LINUX_TEST_RESULT([mkdir_user_namespace], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_MKDIR_UMODE_T, 1,
- [iops->mkdir() takes umode_t])
+ AC_DEFINE(HAVE_IOPS_MKDIR_USERNS, 1,
+ [iops->mkdir() takes struct user_namespace*])
],[
- ZFS_LINUX_TEST_ERROR([mkdir()])
+ AC_MSG_RESULT(no)
+
+ AC_MSG_CHECKING([whether iops->mkdir() takes umode_t])
+ ZFS_LINUX_TEST_RESULT([inode_operations_mkdir], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_MKDIR_UMODE_T, 1,
+ [iops->mkdir() takes umode_t])
+ ],[
+ ZFS_LINUX_TEST_ERROR([mkdir()])
+ ])
])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-mknod.m4 b/sys/contrib/openzfs/config/kernel-mknod.m4
index ffe45106003a..1494ec1ae4d4 100644
--- a/sys/contrib/openzfs/config/kernel-mknod.m4
+++ b/sys/contrib/openzfs/config/kernel-mknod.m4
@@ -1,30 +1,56 @@
AC_DEFUN([ZFS_AC_KERNEL_SRC_MKNOD], [
+ dnl #
+ dnl # 6.3 API change
+ dnl # The first arg is now struct mnt_idmap*
+ dnl #
+ ZFS_LINUX_TEST_SRC([mknod_mnt_idmap], [
+ #include <linux/fs.h>
+ #include <linux/sched.h>
+
+ int tmp_mknod(struct mnt_idmap *idmap,
+ struct inode *inode ,struct dentry *dentry,
+ umode_t u, dev_t d) { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .mknod = tmp_mknod,
+ };
+ ],[])
+
dnl #
dnl # 5.12 API change that added the struct user_namespace* arg
dnl # to the front of this function type's arg list.
dnl #
ZFS_LINUX_TEST_SRC([mknod_userns], [
#include <linux/fs.h>
#include <linux/sched.h>
int tmp_mknod(struct user_namespace *userns,
struct inode *inode ,struct dentry *dentry,
umode_t u, dev_t d) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.mknod = tmp_mknod,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_MKNOD], [
- AC_MSG_CHECKING([whether iops->mknod() takes struct user_namespace*])
- ZFS_LINUX_TEST_RESULT([mknod_userns], [
+ AC_MSG_CHECKING([whether iops->mknod() takes struct mnt_idmap*])
+ ZFS_LINUX_TEST_RESULT([mknod_mnt_idmap], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_IOPS_MKNOD_USERNS, 1,
- [iops->mknod() takes struct user_namespace*])
+ AC_DEFINE(HAVE_IOPS_MKNOD_IDMAP, 1,
+ [iops->mknod() takes struct mnt_idmap*])
],[
AC_MSG_RESULT(no)
+ AC_MSG_CHECKING([whether iops->mknod() takes struct user_namespace*])
+ ZFS_LINUX_TEST_RESULT([mknod_userns], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_IOPS_MKNOD_USERNS, 1,
+ [iops->mknod() takes struct user_namespace*])
+ ],[
+ AC_MSG_RESULT(no)
+ ])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-rename.m4 b/sys/contrib/openzfs/config/kernel-rename.m4
index a2b0800ab4d2..57c3eed78974 100644
--- a/sys/contrib/openzfs/config/kernel-rename.m4
+++ b/sys/contrib/openzfs/config/kernel-rename.m4
@@ -1,111 +1,133 @@
AC_DEFUN([ZFS_AC_KERNEL_SRC_RENAME], [
dnl #
dnl # 3.9 (to 4.9) API change,
dnl #
dnl # A new version of iops->rename() was added (rename2) that takes a flag
dnl # argument (to support renameat2). However this separate function was
dnl # merged back into iops->rename() in Linux 4.9.
dnl #
ZFS_LINUX_TEST_SRC([inode_operations_rename2], [
#include <linux/fs.h>
int rename2_fn(struct inode *sip, struct dentry *sdp,
struct inode *tip, struct dentry *tdp,
unsigned int flags) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.rename2 = rename2_fn,
};
],[])
dnl #
dnl # 4.9 API change,
dnl #
dnl # iops->rename2() merged into iops->rename(), and iops->rename() now
dnl # wants flags.
dnl #
ZFS_LINUX_TEST_SRC([inode_operations_rename_flags], [
#include <linux/fs.h>
int rename_fn(struct inode *sip, struct dentry *sdp,
struct inode *tip, struct dentry *tdp,
unsigned int flags) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.rename = rename_fn,
};
],[])
dnl #
dnl # EL7 compatibility
dnl #
dnl # EL7 has backported renameat2 support, but it's done by defining a
dnl # separate iops wrapper structure that takes the .renameat2 function.
dnl #
ZFS_LINUX_TEST_SRC([dir_inode_operations_wrapper_rename2], [
#include <linux/fs.h>
int rename2_fn(struct inode *sip, struct dentry *sdp,
struct inode *tip, struct dentry *tdp,
unsigned int flags) { return 0; }
static const struct inode_operations_wrapper
iops __attribute__ ((unused)) = {
.rename2 = rename2_fn,
};
],[])
dnl #
dnl # 5.12 API change,
dnl #
dnl # Linux 5.12 introduced passing struct user_namespace* as the first
dnl # argument of the rename() and other inode_operations members.
dnl #
ZFS_LINUX_TEST_SRC([inode_operations_rename_userns], [
#include <linux/fs.h>
int rename_fn(struct user_namespace *user_ns, struct inode *sip,
struct dentry *sdp, struct inode *tip, struct dentry *tdp,
unsigned int flags) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.rename = rename_fn,
};
],[])
+
+ dnl #
+ dnl # 6.3 API change - the first arg is now struct mnt_idmap*
+ dnl #
+ ZFS_LINUX_TEST_SRC([inode_operations_rename_mnt_idmap], [
+ #include <linux/fs.h>
+ int rename_fn(struct mnt_idmap *idmap, struct inode *sip,
+ struct dentry *sdp, struct inode *tip, struct dentry *tdp,
+ unsigned int flags) { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .rename = rename_fn,
+ };
+ ],[])
])
AC_DEFUN([ZFS_AC_KERNEL_RENAME], [
- AC_MSG_CHECKING([whether iops->rename() takes struct user_namespace*])
- ZFS_LINUX_TEST_RESULT([inode_operations_rename_userns], [
+ AC_MSG_CHECKING([whether iops->rename() takes struct mnt_idmap*])
+ ZFS_LINUX_TEST_RESULT([inode_operations_rename_mnt_idmap], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_IOPS_RENAME_USERNS, 1,
- [iops->rename() takes struct user_namespace*])
+ AC_DEFINE(HAVE_IOPS_RENAME_IDMAP, 1,
+ [iops->rename() takes struct mnt_idmap*])
],[
- AC_MSG_RESULT(no)
-
- AC_MSG_CHECKING([whether iops->rename2() exists])
- ZFS_LINUX_TEST_RESULT([inode_operations_rename2], [
+ AC_MSG_CHECKING([whether iops->rename() takes struct user_namespace*])
+ ZFS_LINUX_TEST_RESULT([inode_operations_rename_userns], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_RENAME2, 1, [iops->rename2() exists])
+ AC_DEFINE(HAVE_IOPS_RENAME_USERNS, 1,
+ [iops->rename() takes struct user_namespace*])
],[
AC_MSG_RESULT(no)
- AC_MSG_CHECKING([whether iops->rename() wants flags])
- ZFS_LINUX_TEST_RESULT([inode_operations_rename_flags], [
+ AC_MSG_CHECKING([whether iops->rename2() exists])
+ ZFS_LINUX_TEST_RESULT([inode_operations_rename2], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_RENAME_WANTS_FLAGS, 1,
- [iops->rename() wants flags])
+ AC_DEFINE(HAVE_RENAME2, 1, [iops->rename2() exists])
],[
AC_MSG_RESULT(no)
- AC_MSG_CHECKING([whether struct inode_operations_wrapper takes .rename2()])
- ZFS_LINUX_TEST_RESULT([dir_inode_operations_wrapper_rename2], [
+ AC_MSG_CHECKING([whether iops->rename() wants flags])
+ ZFS_LINUX_TEST_RESULT([inode_operations_rename_flags], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_RENAME2_OPERATIONS_WRAPPER, 1,
- [struct inode_operations_wrapper takes .rename2()])
+ AC_DEFINE(HAVE_RENAME_WANTS_FLAGS, 1,
+ [iops->rename() wants flags])
],[
AC_MSG_RESULT(no)
+
+ AC_MSG_CHECKING([whether struct inode_operations_wrapper takes .rename2()])
+ ZFS_LINUX_TEST_RESULT([dir_inode_operations_wrapper_rename2], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_RENAME2_OPERATIONS_WRAPPER, 1,
+ [struct inode_operations_wrapper takes .rename2()])
+ ],[
+ AC_MSG_RESULT(no)
+ ])
])
])
])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-setattr-prepare.m4 b/sys/contrib/openzfs/config/kernel-setattr-prepare.m4
index 24245aa53448..e02d6263e9c9 100644
--- a/sys/contrib/openzfs/config/kernel-setattr-prepare.m4
+++ b/sys/contrib/openzfs/config/kernel-setattr-prepare.m4
@@ -1,52 +1,74 @@
AC_DEFUN([ZFS_AC_KERNEL_SRC_SETATTR_PREPARE], [
dnl #
dnl # 4.9 API change
dnl # The inode_change_ok() function has been renamed setattr_prepare()
dnl # and updated to take a dentry rather than an inode.
dnl #
ZFS_LINUX_TEST_SRC([setattr_prepare], [
#include <linux/fs.h>
], [
struct dentry *dentry = NULL;
struct iattr *attr = NULL;
int error __attribute__ ((unused)) =
setattr_prepare(dentry, attr);
])
dnl #
dnl # 5.12 API change
dnl # The setattr_prepare() function has been changed to accept a new argument
dnl # for struct user_namespace*
dnl #
ZFS_LINUX_TEST_SRC([setattr_prepare_userns], [
#include <linux/fs.h>
], [
struct dentry *dentry = NULL;
struct iattr *attr = NULL;
struct user_namespace *userns = NULL;
int error __attribute__ ((unused)) =
setattr_prepare(userns, dentry, attr);
])
+
+ dnl #
+ dnl # 6.3 API change
+ dnl # The first arg of setattr_prepare() is changed to struct mnt_idmap*
+ dnl #
+ ZFS_LINUX_TEST_SRC([setattr_prepare_mnt_idmap], [
+ #include <linux/fs.h>
+ ], [
+ struct dentry *dentry = NULL;
+ struct iattr *attr = NULL;
+ struct mnt_idmap *idmap = NULL;
+ int error __attribute__ ((unused)) =
+ setattr_prepare(idmap, dentry, attr);
+ ])
])
AC_DEFUN([ZFS_AC_KERNEL_SETATTR_PREPARE], [
- AC_MSG_CHECKING([whether setattr_prepare() is available and accepts struct user_namespace*])
- ZFS_LINUX_TEST_RESULT_SYMBOL([setattr_prepare_userns],
+ AC_MSG_CHECKING([whether setattr_prepare() is available and accepts struct mnt_idmap*])
+ ZFS_LINUX_TEST_RESULT_SYMBOL([setattr_prepare_mnt_idmap],
[setattr_prepare], [fs/attr.c], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_SETATTR_PREPARE_USERNS, 1,
- [setattr_prepare() accepts user_namespace])
+ AC_DEFINE(HAVE_SETATTR_PREPARE_IDMAP, 1,
+ [setattr_prepare() accepts mnt_idmap])
], [
- AC_MSG_RESULT(no)
-
- AC_MSG_CHECKING([whether setattr_prepare() is available, doesn't accept user_namespace])
- ZFS_LINUX_TEST_RESULT_SYMBOL([setattr_prepare],
- [setattr_prepare], [fs/attr.c], [
+ AC_MSG_CHECKING([whether setattr_prepare() is available and accepts struct user_namespace*])
+ ZFS_LINUX_TEST_RESULT_SYMBOL([setattr_prepare_userns],
+ [setattr_prepare], [fs/attr.c], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_SETATTR_PREPARE_NO_USERNS, 1,
- [setattr_prepare() is available, doesn't accept user_namespace])
+ AC_DEFINE(HAVE_SETATTR_PREPARE_USERNS, 1,
+ [setattr_prepare() accepts user_namespace])
], [
AC_MSG_RESULT(no)
+
+ AC_MSG_CHECKING([whether setattr_prepare() is available, doesn't accept user_namespace])
+ ZFS_LINUX_TEST_RESULT_SYMBOL([setattr_prepare],
+ [setattr_prepare], [fs/attr.c], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_SETATTR_PREPARE_NO_USERNS, 1,
+ [setattr_prepare() is available, doesn't accept user_namespace])
+ ], [
+ AC_MSG_RESULT(no)
+ ])
])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-symlink.m4 b/sys/contrib/openzfs/config/kernel-symlink.m4
index d90366d04b72..a0333ed66a7c 100644
--- a/sys/contrib/openzfs/config/kernel-symlink.m4
+++ b/sys/contrib/openzfs/config/kernel-symlink.m4
@@ -1,30 +1,53 @@
AC_DEFUN([ZFS_AC_KERNEL_SRC_SYMLINK], [
+ dnl #
+ dnl # 6.3 API change that changed the first arg
+ dnl # to struct mnt_idmap*
+ dnl #
+ ZFS_LINUX_TEST_SRC([symlink_mnt_idmap], [
+ #include <linux/fs.h>
+ #include <linux/sched.h>
+ int tmp_symlink(struct mnt_idmap *idmap,
+ struct inode *inode ,struct dentry *dentry,
+ const char *path) { return 0; }
+
+ static const struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .symlink = tmp_symlink,
+ };
+ ],[])
dnl #
dnl # 5.12 API change that added the struct user_namespace* arg
dnl # to the front of this function type's arg list.
dnl #
ZFS_LINUX_TEST_SRC([symlink_userns], [
#include <linux/fs.h>
#include <linux/sched.h>
int tmp_symlink(struct user_namespace *userns,
struct inode *inode ,struct dentry *dentry,
const char *path) { return 0; }
static const struct inode_operations
iops __attribute__ ((unused)) = {
.symlink = tmp_symlink,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_SYMLINK], [
- AC_MSG_CHECKING([whether iops->symlink() takes struct user_namespace*])
- ZFS_LINUX_TEST_RESULT([symlink_userns], [
+ AC_MSG_CHECKING([whether iops->symlink() takes struct mnt_idmap*])
+ ZFS_LINUX_TEST_RESULT([symlink_mnt_idmap], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_IOPS_SYMLINK_USERNS, 1,
- [iops->symlink() takes struct user_namespace*])
+ AC_DEFINE(HAVE_IOPS_SYMLINK_IDMAP, 1,
+ [iops->symlink() takes struct mnt_idmap*])
],[
- AC_MSG_RESULT(no)
+ AC_MSG_CHECKING([whether iops->symlink() takes struct user_namespace*])
+ ZFS_LINUX_TEST_RESULT([symlink_userns], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_IOPS_SYMLINK_USERNS, 1,
+ [iops->symlink() takes struct user_namespace*])
+ ],[
+ AC_MSG_RESULT(no)
+ ])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-tmpfile.m4 b/sys/contrib/openzfs/config/kernel-tmpfile.m4
index 0e1deb3612f3..cc18b8f65a88 100644
--- a/sys/contrib/openzfs/config/kernel-tmpfile.m4
+++ b/sys/contrib/openzfs/config/kernel-tmpfile.m4
@@ -1,67 +1,86 @@
dnl #
dnl # 3.11 API change
dnl # Add support for i_op->tmpfile
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_TMPFILE], [
dnl #
+ dnl # 6.3 API change
+ dnl # The first arg is now struct mnt_idmap *
+ dnl #
+ ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_mnt_idmap], [
+ #include <linux/fs.h>
+ int tmpfile(struct mnt_idmap *idmap,
+ struct inode *inode, struct file *file,
+ umode_t mode) { return 0; }
+ static struct inode_operations
+ iops __attribute__ ((unused)) = {
+ .tmpfile = tmpfile,
+ };
+ ],[])
dnl # 6.1 API change
dnl # use struct file instead of struct dentry
dnl #
ZFS_LINUX_TEST_SRC([inode_operations_tmpfile], [
#include <linux/fs.h>
int tmpfile(struct user_namespace *userns,
struct inode *inode, struct file *file,
umode_t mode) { return 0; }
static struct inode_operations
iops __attribute__ ((unused)) = {
.tmpfile = tmpfile,
};
],[])
dnl #
dnl # 5.11 API change
dnl # add support for userns parameter to tmpfile
dnl #
ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_dentry_userns], [
#include <linux/fs.h>
int tmpfile(struct user_namespace *userns,
struct inode *inode, struct dentry *dentry,
umode_t mode) { return 0; }
static struct inode_operations
iops __attribute__ ((unused)) = {
.tmpfile = tmpfile,
};
],[])
ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_dentry], [
#include <linux/fs.h>
int tmpfile(struct inode *inode, struct dentry *dentry,
umode_t mode) { return 0; }
static struct inode_operations
iops __attribute__ ((unused)) = {
.tmpfile = tmpfile,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_TMPFILE], [
AC_MSG_CHECKING([whether i_op->tmpfile() exists])
- ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile], [
+ ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile_mnt_idmap], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_TMPFILE, 1, [i_op->tmpfile() exists])
- AC_DEFINE(HAVE_TMPFILE_USERNS, 1, [i_op->tmpfile() has userns])
- ],[
- ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile_dentry_userns], [
+ AC_DEFINE(HAVE_TMPFILE_IDMAP, 1, [i_op->tmpfile() has mnt_idmap])
+ ], [
+ ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_TMPFILE, 1, [i_op->tmpfile() exists])
AC_DEFINE(HAVE_TMPFILE_USERNS, 1, [i_op->tmpfile() has userns])
- AC_DEFINE(HAVE_TMPFILE_DENTRY, 1, [i_op->tmpfile() uses old dentry signature])
],[
- ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile_dentry], [
+ ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile_dentry_userns], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_TMPFILE, 1, [i_op->tmpfile() exists])
+ AC_DEFINE(HAVE_TMPFILE_USERNS, 1, [i_op->tmpfile() has userns])
AC_DEFINE(HAVE_TMPFILE_DENTRY, 1, [i_op->tmpfile() uses old dentry signature])
],[
- ZFS_LINUX_REQUIRE_API([i_op->tmpfile()], [3.11])
+ ZFS_LINUX_TEST_RESULT([inode_operations_tmpfile_dentry], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_TMPFILE, 1, [i_op->tmpfile() exists])
+ AC_DEFINE(HAVE_TMPFILE_DENTRY, 1, [i_op->tmpfile() uses old dentry signature])
+ ],[
+ ZFS_LINUX_REQUIRE_API([i_op->tmpfile()], [3.11])
+ ])
])
])
])
])
diff --git a/sys/contrib/openzfs/config/kernel-writepage_t.m4 b/sys/contrib/openzfs/config/kernel-writepage_t.m4
new file mode 100644
index 000000000000..3a0cffd98570
--- /dev/null
+++ b/sys/contrib/openzfs/config/kernel-writepage_t.m4
@@ -0,0 +1,26 @@
+AC_DEFUN([ZFS_AC_KERNEL_SRC_WRITEPAGE_T], [
+ dnl #
+ dnl # 6.3 API change
+ dnl # The writepage_t function type now has its first argument as
+ dnl # struct folio* instead of struct page*
+ dnl #
+ ZFS_LINUX_TEST_SRC([writepage_t_folio], [
+ #include <linux/writeback.h>
+ int putpage(struct folio *folio,
+ struct writeback_control *wbc, void *data)
+ { return 0; }
+ writepage_t func = putpage;
+ ],[])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_WRITEPAGE_T], [
+ AC_MSG_CHECKING([whether int (*writepage_t)() takes struct folio*])
+ ZFS_LINUX_TEST_RESULT([writepage_t_folio], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_WRITEPAGE_T_FOLIO, 1,
+ [int (*writepage_t)() takes struct folio*])
+ ],[
+ AC_MSG_RESULT(no)
+ ])
+])
+
diff --git a/sys/contrib/openzfs/config/kernel-xattr-handler.m4 b/sys/contrib/openzfs/config/kernel-xattr-handler.m4
index b6cbfa155007..6b8a08dbcc80 100644
--- a/sys/contrib/openzfs/config/kernel-xattr-handler.m4
+++ b/sys/contrib/openzfs/config/kernel-xattr-handler.m4
@@ -1,452 +1,477 @@
dnl #
dnl # 2.6.35 API change,
dnl # The 'struct xattr_handler' was constified in the generic
dnl # super_block structure.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_CONST_XATTR_HANDLER], [
ZFS_LINUX_TEST_SRC([const_xattr_handler], [
#include <linux/fs.h>
#include <linux/xattr.h>
const struct xattr_handler xattr_test_handler = {
.prefix = "test",
.get = NULL,
.set = NULL,
};
const struct xattr_handler *xattr_handlers[] = {
&xattr_test_handler,
};
const struct super_block sb __attribute__ ((unused)) = {
.s_xattr = xattr_handlers,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_CONST_XATTR_HANDLER], [
AC_MSG_CHECKING([whether super_block uses const struct xattr_handler])
ZFS_LINUX_TEST_RESULT([const_xattr_handler], [
AC_MSG_RESULT([yes])
],[
ZFS_LINUX_TEST_ERROR([const xattr_handler])
])
])
dnl #
dnl # 4.5 API change,
dnl # struct xattr_handler added new member "name".
dnl # xattr_handler which matches to whole name rather than prefix should use
dnl # "name" instead of "prefix", e.g. "system.posix_acl_access"
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_NAME], [
ZFS_LINUX_TEST_SRC([xattr_handler_name], [
#include <linux/xattr.h>
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.name = XATTR_NAME_POSIX_ACL_ACCESS,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_XATTR_HANDLER_NAME], [
AC_MSG_CHECKING([whether xattr_handler has name])
ZFS_LINUX_TEST_RESULT([xattr_handler_name], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_XATTR_HANDLER_NAME, 1,
[xattr_handler has name])
],[
AC_MSG_RESULT(no)
])
])
dnl #
dnl # Supported xattr handler get() interfaces checked newest to oldest.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_GET], [
ZFS_LINUX_TEST_SRC([xattr_handler_get_dentry_inode], [
#include <linux/xattr.h>
int get(const struct xattr_handler *handler,
struct dentry *dentry, struct inode *inode,
const char *name, void *buffer, size_t size) { return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.get = get,
};
],[])
ZFS_LINUX_TEST_SRC([xattr_handler_get_xattr_handler], [
#include <linux/xattr.h>
int get(const struct xattr_handler *handler,
struct dentry *dentry, const char *name,
void *buffer, size_t size) { return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.get = get,
};
],[])
ZFS_LINUX_TEST_SRC([xattr_handler_get_dentry], [
#include <linux/xattr.h>
int get(struct dentry *dentry, const char *name,
void *buffer, size_t size, int handler_flags)
{ return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.get = get,
};
],[])
ZFS_LINUX_TEST_SRC([xattr_handler_get_dentry_inode_flags], [
#include <linux/xattr.h>
int get(const struct xattr_handler *handler,
struct dentry *dentry, struct inode *inode,
const char *name, void *buffer,
size_t size, int flags) { return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.get = get,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_XATTR_HANDLER_GET], [
dnl #
dnl # 4.7 API change,
dnl # The xattr_handler->get() callback was changed to take both
dnl # dentry and inode.
dnl #
AC_MSG_CHECKING([whether xattr_handler->get() wants dentry and inode])
ZFS_LINUX_TEST_RESULT([xattr_handler_get_dentry_inode], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_XATTR_GET_DENTRY_INODE, 1,
[xattr_handler->get() wants both dentry and inode])
],[
dnl #
dnl # 4.4 API change,
dnl # The xattr_handler->get() callback was changed to take a
dnl # attr_handler, and handler_flags argument was removed and
dnl # should be accessed by handler->flags.
dnl #
AC_MSG_RESULT(no)
AC_MSG_CHECKING(
[whether xattr_handler->get() wants xattr_handler])
ZFS_LINUX_TEST_RESULT([xattr_handler_get_xattr_handler], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_XATTR_GET_HANDLER, 1,
[xattr_handler->get() wants xattr_handler])
],[
dnl #
dnl # 2.6.33 API change,
dnl # The xattr_handler->get() callback was changed
dnl # to take a dentry instead of an inode, and a
dnl # handler_flags argument was added.
dnl #
AC_MSG_RESULT(no)
AC_MSG_CHECKING(
[whether xattr_handler->get() wants dentry])
ZFS_LINUX_TEST_RESULT([xattr_handler_get_dentry], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_XATTR_GET_DENTRY, 1,
[xattr_handler->get() wants dentry])
],[
dnl #
dnl # Android API change,
dnl # The xattr_handler->get() callback was
dnl # changed to take dentry, inode and flags.
dnl #
AC_MSG_RESULT(no)
AC_MSG_CHECKING(
[whether xattr_handler->get() wants dentry and inode and flags])
ZFS_LINUX_TEST_RESULT([xattr_handler_get_dentry_inode_flags], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_XATTR_GET_DENTRY_INODE_FLAGS, 1,
[xattr_handler->get() wants dentry and inode and flags])
],[
ZFS_LINUX_TEST_ERROR([xattr get()])
])
])
])
])
])
dnl #
dnl # Supported xattr handler set() interfaces checked newest to oldest.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_SET], [
+ ZFS_LINUX_TEST_SRC([xattr_handler_set_mnt_idmap], [
+ #include <linux/xattr.h>
+
+ int set(const struct xattr_handler *handler,
+ struct mnt_idmap *idmap,
+ struct dentry *dentry, struct inode *inode,
+ const char *name, const void *buffer,
+ size_t size, int flags)
+ { return 0; }
+ static const struct xattr_handler
+ xops __attribute__ ((unused)) = {
+ .set = set,
+ };
+ ],[])
+
ZFS_LINUX_TEST_SRC([xattr_handler_set_userns], [
#include <linux/xattr.h>
int set(const struct xattr_handler *handler,
struct user_namespace *mnt_userns,
struct dentry *dentry, struct inode *inode,
const char *name, const void *buffer,
size_t size, int flags)
{ return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.set = set,
};
],[])
ZFS_LINUX_TEST_SRC([xattr_handler_set_dentry_inode], [
#include <linux/xattr.h>
int set(const struct xattr_handler *handler,
struct dentry *dentry, struct inode *inode,
const char *name, const void *buffer,
size_t size, int flags)
{ return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.set = set,
};
],[])
ZFS_LINUX_TEST_SRC([xattr_handler_set_xattr_handler], [
#include <linux/xattr.h>
int set(const struct xattr_handler *handler,
struct dentry *dentry, const char *name,
const void *buffer, size_t size, int flags)
{ return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.set = set,
};
],[])
ZFS_LINUX_TEST_SRC([xattr_handler_set_dentry], [
#include <linux/xattr.h>
int set(struct dentry *dentry, const char *name,
const void *buffer, size_t size, int flags,
int handler_flags) { return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.set = set,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_XATTR_HANDLER_SET], [
dnl #
dnl # 5.12 API change,
dnl # The xattr_handler->set() callback was changed to 8 arguments, and
dnl # struct user_namespace* was inserted as arg #2
dnl #
- AC_MSG_CHECKING([whether xattr_handler->set() wants dentry, inode, and user_namespace])
- ZFS_LINUX_TEST_RESULT([xattr_handler_set_userns], [
+ dnl # 6.3 API change,
+ dnl # The xattr_handler->set() callback 2nd arg is now struct mnt_idmap *
+ dnl #
+ AC_MSG_CHECKING([whether xattr_handler->set() wants dentry, inode, and mnt_idmap])
+ ZFS_LINUX_TEST_RESULT([xattr_handler_set_mnt_idmap], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_XATTR_SET_USERNS, 1,
- [xattr_handler->set() takes user_namespace])
- ],[
- dnl #
- dnl # 4.7 API change,
- dnl # The xattr_handler->set() callback was changed to take both
- dnl # dentry and inode.
- dnl #
- AC_MSG_RESULT(no)
- AC_MSG_CHECKING([whether xattr_handler->set() wants dentry and inode])
- ZFS_LINUX_TEST_RESULT([xattr_handler_set_dentry_inode], [
+ AC_DEFINE(HAVE_XATTR_SET_IDMAP, 1,
+ [xattr_handler->set() takes mnt_idmap])
+ ], [
+ AC_MSG_CHECKING([whether xattr_handler->set() wants dentry, inode, and user_namespace])
+ ZFS_LINUX_TEST_RESULT([xattr_handler_set_userns], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_XATTR_SET_DENTRY_INODE, 1,
- [xattr_handler->set() wants both dentry and inode])
+ AC_DEFINE(HAVE_XATTR_SET_USERNS, 1,
+ [xattr_handler->set() takes user_namespace])
],[
dnl #
- dnl # 4.4 API change,
- dnl # The xattr_handler->set() callback was changed to take a
- dnl # xattr_handler, and handler_flags argument was removed and
- dnl # should be accessed by handler->flags.
+ dnl # 4.7 API change,
+ dnl # The xattr_handler->set() callback was changed to take both
+ dnl # dentry and inode.
dnl #
AC_MSG_RESULT(no)
- AC_MSG_CHECKING(
- [whether xattr_handler->set() wants xattr_handler])
- ZFS_LINUX_TEST_RESULT([xattr_handler_set_xattr_handler], [
+ AC_MSG_CHECKING([whether xattr_handler->set() wants dentry and inode])
+ ZFS_LINUX_TEST_RESULT([xattr_handler_set_dentry_inode], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_XATTR_SET_HANDLER, 1,
- [xattr_handler->set() wants xattr_handler])
+ AC_DEFINE(HAVE_XATTR_SET_DENTRY_INODE, 1,
+ [xattr_handler->set() wants both dentry and inode])
],[
dnl #
- dnl # 2.6.33 API change,
- dnl # The xattr_handler->set() callback was changed
- dnl # to take a dentry instead of an inode, and a
- dnl # handler_flags argument was added.
+ dnl # 4.4 API change,
+ dnl # The xattr_handler->set() callback was changed to take a
+ dnl # xattr_handler, and handler_flags argument was removed and
+ dnl # should be accessed by handler->flags.
dnl #
AC_MSG_RESULT(no)
AC_MSG_CHECKING(
- [whether xattr_handler->set() wants dentry])
- ZFS_LINUX_TEST_RESULT([xattr_handler_set_dentry], [
+ [whether xattr_handler->set() wants xattr_handler])
+ ZFS_LINUX_TEST_RESULT([xattr_handler_set_xattr_handler], [
AC_MSG_RESULT(yes)
- AC_DEFINE(HAVE_XATTR_SET_DENTRY, 1,
- [xattr_handler->set() wants dentry])
+ AC_DEFINE(HAVE_XATTR_SET_HANDLER, 1,
+ [xattr_handler->set() wants xattr_handler])
],[
- ZFS_LINUX_TEST_ERROR([xattr set()])
+ dnl #
+ dnl # 2.6.33 API change,
+ dnl # The xattr_handler->set() callback was changed
+ dnl # to take a dentry instead of an inode, and a
+ dnl # handler_flags argument was added.
+ dnl #
+ AC_MSG_RESULT(no)
+ AC_MSG_CHECKING(
+ [whether xattr_handler->set() wants dentry])
+ ZFS_LINUX_TEST_RESULT([xattr_handler_set_dentry], [
+ AC_MSG_RESULT(yes)
+ AC_DEFINE(HAVE_XATTR_SET_DENTRY, 1,
+ [xattr_handler->set() wants dentry])
+ ],[
+ ZFS_LINUX_TEST_ERROR([xattr set()])
+ ])
])
])
])
])
])
dnl #
dnl # Supported xattr handler list() interfaces checked newest to oldest.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_LIST], [
ZFS_LINUX_TEST_SRC([xattr_handler_list_simple], [
#include <linux/xattr.h>
bool list(struct dentry *dentry) { return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.list = list,
};
],[])
ZFS_LINUX_TEST_SRC([xattr_handler_list_xattr_handler], [
#include <linux/xattr.h>
size_t list(const struct xattr_handler *handler,
struct dentry *dentry, char *list, size_t list_size,
const char *name, size_t name_len) { return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.list = list,
};
],[])
ZFS_LINUX_TEST_SRC([xattr_handler_list_dentry], [
#include <linux/xattr.h>
size_t list(struct dentry *dentry,
char *list, size_t list_size,
const char *name, size_t name_len,
int handler_flags) { return 0; }
static const struct xattr_handler
xops __attribute__ ((unused)) = {
.list = list,
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_XATTR_HANDLER_LIST], [
dnl # 4.5 API change,
dnl # The xattr_handler->list() callback was changed to take only a
dnl # dentry and it only needs to return if it's accessible.
AC_MSG_CHECKING([whether xattr_handler->list() wants simple])
ZFS_LINUX_TEST_RESULT([xattr_handler_list_simple], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_XATTR_LIST_SIMPLE, 1,
[xattr_handler->list() wants simple])
],[
dnl #
dnl # 4.4 API change,
dnl # The xattr_handler->list() callback was changed to take a
dnl # xattr_handler, and handler_flags argument was removed
dnl # and should be accessed by handler->flags.
dnl #
AC_MSG_RESULT(no)
AC_MSG_CHECKING(
[whether xattr_handler->list() wants xattr_handler])
ZFS_LINUX_TEST_RESULT([xattr_handler_list_xattr_handler], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_XATTR_LIST_HANDLER, 1,
[xattr_handler->list() wants xattr_handler])
],[
dnl #
dnl # 2.6.33 API change,
dnl # The xattr_handler->list() callback was changed
dnl # to take a dentry instead of an inode, and a
dnl # handler_flags argument was added.
dnl #
AC_MSG_RESULT(no)
AC_MSG_CHECKING(
[whether xattr_handler->list() wants dentry])
ZFS_LINUX_TEST_RESULT([xattr_handler_list_dentry], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_XATTR_LIST_DENTRY, 1,
[xattr_handler->list() wants dentry])
],[
ZFS_LINUX_TEST_ERROR([xattr list()])
])
])
])
])
dnl #
dnl # 3.7 API change,
dnl # The posix_acl_{from,to}_xattr functions gained a new
dnl # parameter: user_ns
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_POSIX_ACL_FROM_XATTR_USERNS], [
ZFS_LINUX_TEST_SRC([posix_acl_from_xattr_userns], [
#include <linux/cred.h>
#include <linux/fs.h>
#include <linux/posix_acl_xattr.h>
],[
posix_acl_from_xattr(&init_user_ns, NULL, 0);
])
])
AC_DEFUN([ZFS_AC_KERNEL_POSIX_ACL_FROM_XATTR_USERNS], [
AC_MSG_CHECKING([whether posix_acl_from_xattr() needs user_ns])
ZFS_LINUX_TEST_RESULT([posix_acl_from_xattr_userns], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_POSIX_ACL_FROM_XATTR_USERNS, 1,
[posix_acl_from_xattr() needs user_ns])
],[
ZFS_LINUX_TEST_ERROR([posix_acl_from_xattr()])
])
])
dnl #
dnl # 4.9 API change,
dnl # iops->{set,get,remove}xattr and generic_{set,get,remove}xattr are
dnl # removed. xattr operations will directly go through sb->s_xattr.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_SRC_GENERIC_SETXATTR], [
ZFS_LINUX_TEST_SRC([have_generic_setxattr], [
#include <linux/fs.h>
#include <linux/xattr.h>
static const struct inode_operations
iops __attribute__ ((unused)) = {
.setxattr = generic_setxattr
};
],[])
])
AC_DEFUN([ZFS_AC_KERNEL_GENERIC_SETXATTR], [
AC_MSG_CHECKING([whether generic_setxattr() exists])
ZFS_LINUX_TEST_RESULT([have_generic_setxattr], [
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_GENERIC_SETXATTR, 1,
[generic_setxattr() exists])
],[
AC_MSG_RESULT(no)
])
])
AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR], [
ZFS_AC_KERNEL_SRC_CONST_XATTR_HANDLER
ZFS_AC_KERNEL_SRC_XATTR_HANDLER_NAME
ZFS_AC_KERNEL_SRC_XATTR_HANDLER_GET
ZFS_AC_KERNEL_SRC_XATTR_HANDLER_SET
ZFS_AC_KERNEL_SRC_XATTR_HANDLER_LIST
ZFS_AC_KERNEL_SRC_POSIX_ACL_FROM_XATTR_USERNS
ZFS_AC_KERNEL_SRC_GENERIC_SETXATTR
])
AC_DEFUN([ZFS_AC_KERNEL_XATTR], [
ZFS_AC_KERNEL_CONST_XATTR_HANDLER
ZFS_AC_KERNEL_XATTR_HANDLER_NAME
ZFS_AC_KERNEL_XATTR_HANDLER_GET
ZFS_AC_KERNEL_XATTR_HANDLER_SET
ZFS_AC_KERNEL_XATTR_HANDLER_LIST
ZFS_AC_KERNEL_POSIX_ACL_FROM_XATTR_USERNS
ZFS_AC_KERNEL_GENERIC_SETXATTR
])
diff --git a/sys/contrib/openzfs/config/kernel.m4 b/sys/contrib/openzfs/config/kernel.m4
index 4c7569841f33..439ffdf5a898 100644
--- a/sys/contrib/openzfs/config/kernel.m4
+++ b/sys/contrib/openzfs/config/kernel.m4
@@ -1,1006 +1,1010 @@
dnl #
dnl # Default ZFS kernel configuration
dnl #
AC_DEFUN([ZFS_AC_CONFIG_KERNEL], [
AM_COND_IF([BUILD_LINUX], [
dnl # Setup the kernel build environment.
ZFS_AC_KERNEL
ZFS_AC_QAT
dnl # Sanity checks for module building and CONFIG_* defines
ZFS_AC_KERNEL_CONFIG_DEFINED
ZFS_AC_MODULE_SYMVERS
dnl # Sequential ZFS_LINUX_TRY_COMPILE tests
ZFS_AC_KERNEL_FPU_HEADER
ZFS_AC_KERNEL_OBJTOOL_HEADER
ZFS_AC_KERNEL_WAIT_QUEUE_ENTRY_T
ZFS_AC_KERNEL_MISC_MINOR
ZFS_AC_KERNEL_DECLARE_EVENT_CLASS
dnl # Parallel ZFS_LINUX_TEST_SRC / ZFS_LINUX_TEST_RESULT tests
ZFS_AC_KERNEL_TEST_SRC
ZFS_AC_KERNEL_TEST_RESULT
AS_IF([test "$LINUX_OBJ" != "$LINUX"], [
KERNEL_MAKE="$KERNEL_MAKE O=$LINUX_OBJ"
])
AC_SUBST(KERNEL_MAKE)
])
])
dnl #
dnl # Generate and compile all of the kernel API test cases to determine
dnl # which interfaces are available. By invoking the kernel build system
dnl # only once the compilation can be done in parallel significantly
dnl # speeding up the process.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_TEST_SRC], [
ZFS_AC_KERNEL_SRC_OBJTOOL
ZFS_AC_KERNEL_SRC_GLOBAL_PAGE_STATE
ZFS_AC_KERNEL_SRC_ACCESS_OK_TYPE
ZFS_AC_KERNEL_SRC_PDE_DATA
ZFS_AC_KERNEL_SRC_FALLOCATE
ZFS_AC_KERNEL_SRC_FADVISE
ZFS_AC_KERNEL_SRC_GENERIC_FADVISE
ZFS_AC_KERNEL_SRC_2ARGS_ZLIB_DEFLATE_WORKSPACESIZE
ZFS_AC_KERNEL_SRC_RWSEM
ZFS_AC_KERNEL_SRC_SCHED
ZFS_AC_KERNEL_SRC_USLEEP_RANGE
ZFS_AC_KERNEL_SRC_KMEM_CACHE
ZFS_AC_KERNEL_SRC_KVMALLOC
ZFS_AC_KERNEL_SRC_VMALLOC_PAGE_KERNEL
ZFS_AC_KERNEL_SRC_WAIT
ZFS_AC_KERNEL_SRC_INODE_TIMES
ZFS_AC_KERNEL_SRC_INODE_LOCK
ZFS_AC_KERNEL_SRC_GROUP_INFO_GID
ZFS_AC_KERNEL_SRC_RW
ZFS_AC_KERNEL_SRC_TIMER_SETUP
ZFS_AC_KERNEL_SRC_SUPER_USER_NS
ZFS_AC_KERNEL_SRC_PROC_OPERATIONS
ZFS_AC_KERNEL_SRC_BLOCK_DEVICE_OPERATIONS
ZFS_AC_KERNEL_SRC_BIO
ZFS_AC_KERNEL_SRC_BLKDEV
ZFS_AC_KERNEL_SRC_BLK_QUEUE
ZFS_AC_KERNEL_SRC_GENHD_FLAGS
ZFS_AC_KERNEL_SRC_REVALIDATE_DISK
ZFS_AC_KERNEL_SRC_GET_DISK_RO
ZFS_AC_KERNEL_SRC_GENERIC_READLINK_GLOBAL
ZFS_AC_KERNEL_SRC_DISCARD_GRANULARITY
ZFS_AC_KERNEL_SRC_INODE_OWNER_OR_CAPABLE
ZFS_AC_KERNEL_SRC_XATTR
ZFS_AC_KERNEL_SRC_ACL
+ ZFS_AC_KERNEL_SRC_INODE_SETATTR
ZFS_AC_KERNEL_SRC_INODE_GETATTR
ZFS_AC_KERNEL_SRC_INODE_SET_FLAGS
ZFS_AC_KERNEL_SRC_INODE_SET_IVERSION
ZFS_AC_KERNEL_SRC_SHOW_OPTIONS
ZFS_AC_KERNEL_SRC_FILE_INODE
ZFS_AC_KERNEL_SRC_FILE_DENTRY
ZFS_AC_KERNEL_SRC_FSYNC
ZFS_AC_KERNEL_SRC_AIO_FSYNC
ZFS_AC_KERNEL_SRC_EVICT_INODE
ZFS_AC_KERNEL_SRC_DIRTY_INODE
ZFS_AC_KERNEL_SRC_SHRINKER
ZFS_AC_KERNEL_SRC_MKDIR
ZFS_AC_KERNEL_SRC_LOOKUP_FLAGS
ZFS_AC_KERNEL_SRC_CREATE
ZFS_AC_KERNEL_SRC_PERMISSION
ZFS_AC_KERNEL_SRC_GET_LINK
ZFS_AC_KERNEL_SRC_PUT_LINK
ZFS_AC_KERNEL_SRC_TMPFILE
ZFS_AC_KERNEL_SRC_AUTOMOUNT
ZFS_AC_KERNEL_SRC_ENCODE_FH_WITH_INODE
ZFS_AC_KERNEL_SRC_COMMIT_METADATA
ZFS_AC_KERNEL_SRC_CLEAR_INODE
ZFS_AC_KERNEL_SRC_SETATTR_PREPARE
ZFS_AC_KERNEL_SRC_INSERT_INODE_LOCKED
ZFS_AC_KERNEL_SRC_DENTRY
ZFS_AC_KERNEL_SRC_DENTRY_ALIAS_D_U
ZFS_AC_KERNEL_SRC_TRUNCATE_SETSIZE
ZFS_AC_KERNEL_SRC_SECURITY_INODE
ZFS_AC_KERNEL_SRC_FST_MOUNT
ZFS_AC_KERNEL_SRC_BDI
ZFS_AC_KERNEL_SRC_SET_NLINK
ZFS_AC_KERNEL_SRC_SGET
ZFS_AC_KERNEL_SRC_LSEEK_EXECUTE
ZFS_AC_KERNEL_SRC_VFS_FILEMAP_DIRTY_FOLIO
ZFS_AC_KERNEL_SRC_VFS_READ_FOLIO
ZFS_AC_KERNEL_SRC_VFS_GETATTR
ZFS_AC_KERNEL_SRC_VFS_FSYNC_2ARGS
ZFS_AC_KERNEL_SRC_VFS_ITERATE
ZFS_AC_KERNEL_SRC_VFS_DIRECT_IO
ZFS_AC_KERNEL_SRC_VFS_READPAGES
ZFS_AC_KERNEL_SRC_VFS_SET_PAGE_DIRTY_NOBUFFERS
ZFS_AC_KERNEL_SRC_VFS_RW_ITERATE
ZFS_AC_KERNEL_SRC_VFS_GENERIC_WRITE_CHECKS
ZFS_AC_KERNEL_SRC_VFS_IOV_ITER
ZFS_AC_KERNEL_SRC_KMAP_ATOMIC_ARGS
ZFS_AC_KERNEL_SRC_FOLLOW_DOWN_ONE
ZFS_AC_KERNEL_SRC_MAKE_REQUEST_FN
ZFS_AC_KERNEL_SRC_GENERIC_IO_ACCT
ZFS_AC_KERNEL_SRC_FPU
ZFS_AC_KERNEL_SRC_FMODE_T
ZFS_AC_KERNEL_SRC_KUIDGID_T
ZFS_AC_KERNEL_SRC_KUID_HELPERS
ZFS_AC_KERNEL_SRC_RENAME
ZFS_AC_KERNEL_SRC_CURRENT_TIME
ZFS_AC_KERNEL_SRC_USERNS_CAPABILITIES
ZFS_AC_KERNEL_SRC_IN_COMPAT_SYSCALL
ZFS_AC_KERNEL_SRC_KTIME
ZFS_AC_KERNEL_SRC_TOTALRAM_PAGES_FUNC
ZFS_AC_KERNEL_SRC_TOTALHIGH_PAGES
ZFS_AC_KERNEL_SRC_KSTRTOUL
ZFS_AC_KERNEL_SRC_PERCPU
ZFS_AC_KERNEL_SRC_CPU_HOTPLUG
- ZFS_AC_KERNEL_SRC_GENERIC_FILLATTR_USERNS
+ ZFS_AC_KERNEL_SRC_GENERIC_FILLATTR
ZFS_AC_KERNEL_SRC_MKNOD
ZFS_AC_KERNEL_SRC_SYMLINK
ZFS_AC_KERNEL_SRC_BIO_MAX_SEGS
ZFS_AC_KERNEL_SRC_SIGNAL_STOP
ZFS_AC_KERNEL_SRC_SIGINFO
ZFS_AC_KERNEL_SRC_SYSFS
ZFS_AC_KERNEL_SRC_SET_SPECIAL_STATE
ZFS_AC_KERNEL_SRC_STANDALONE_LINUX_STDARG
ZFS_AC_KERNEL_SRC_PAGEMAP_FOLIO_WAIT_BIT
ZFS_AC_KERNEL_SRC_ADD_DISK
ZFS_AC_KERNEL_SRC_KTHREAD
ZFS_AC_KERNEL_SRC_ZERO_PAGE
ZFS_AC_KERNEL_SRC___COPY_FROM_USER_INATOMIC
ZFS_AC_KERNEL_SRC_USER_NS_COMMON_INUM
ZFS_AC_KERNEL_SRC_IDMAP_MNT_API
ZFS_AC_KERNEL_SRC_IATTR_VFSID
ZFS_AC_KERNEL_SRC_FILEMAP
+ ZFS_AC_KERNEL_SRC_WRITEPAGE_T
case "$host_cpu" in
powerpc*)
ZFS_AC_KERNEL_SRC_CPU_HAS_FEATURE
ZFS_AC_KERNEL_SRC_FLUSH_DCACHE_PAGE
;;
esac
AC_MSG_CHECKING([for available kernel interfaces])
ZFS_LINUX_TEST_COMPILE_ALL([kabi])
AC_MSG_RESULT([done])
])
dnl #
dnl # Check results of kernel interface tests.
dnl #
AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [
ZFS_AC_KERNEL_ACCESS_OK_TYPE
ZFS_AC_KERNEL_GLOBAL_PAGE_STATE
ZFS_AC_KERNEL_OBJTOOL
ZFS_AC_KERNEL_PDE_DATA
ZFS_AC_KERNEL_FALLOCATE
ZFS_AC_KERNEL_FADVISE
ZFS_AC_KERNEL_GENERIC_FADVISE
ZFS_AC_KERNEL_2ARGS_ZLIB_DEFLATE_WORKSPACESIZE
ZFS_AC_KERNEL_RWSEM
ZFS_AC_KERNEL_SCHED
ZFS_AC_KERNEL_USLEEP_RANGE
ZFS_AC_KERNEL_KMEM_CACHE
ZFS_AC_KERNEL_KVMALLOC
ZFS_AC_KERNEL_VMALLOC_PAGE_KERNEL
ZFS_AC_KERNEL_WAIT
ZFS_AC_KERNEL_INODE_TIMES
ZFS_AC_KERNEL_INODE_LOCK
ZFS_AC_KERNEL_GROUP_INFO_GID
ZFS_AC_KERNEL_RW
ZFS_AC_KERNEL_TIMER_SETUP
ZFS_AC_KERNEL_SUPER_USER_NS
ZFS_AC_KERNEL_PROC_OPERATIONS
ZFS_AC_KERNEL_BLOCK_DEVICE_OPERATIONS
ZFS_AC_KERNEL_BIO
ZFS_AC_KERNEL_BLKDEV
ZFS_AC_KERNEL_BLK_QUEUE
ZFS_AC_KERNEL_GENHD_FLAGS
ZFS_AC_KERNEL_REVALIDATE_DISK
ZFS_AC_KERNEL_GET_DISK_RO
ZFS_AC_KERNEL_GENERIC_READLINK_GLOBAL
ZFS_AC_KERNEL_DISCARD_GRANULARITY
ZFS_AC_KERNEL_INODE_OWNER_OR_CAPABLE
ZFS_AC_KERNEL_XATTR
ZFS_AC_KERNEL_ACL
+ ZFS_AC_KERNEL_INODE_SETATTR
ZFS_AC_KERNEL_INODE_GETATTR
ZFS_AC_KERNEL_INODE_SET_FLAGS
ZFS_AC_KERNEL_INODE_SET_IVERSION
ZFS_AC_KERNEL_SHOW_OPTIONS
ZFS_AC_KERNEL_FILE_INODE
ZFS_AC_KERNEL_FILE_DENTRY
ZFS_AC_KERNEL_FSYNC
ZFS_AC_KERNEL_AIO_FSYNC
ZFS_AC_KERNEL_EVICT_INODE
ZFS_AC_KERNEL_DIRTY_INODE
ZFS_AC_KERNEL_SHRINKER
ZFS_AC_KERNEL_MKDIR
ZFS_AC_KERNEL_LOOKUP_FLAGS
ZFS_AC_KERNEL_CREATE
ZFS_AC_KERNEL_PERMISSION
ZFS_AC_KERNEL_GET_LINK
ZFS_AC_KERNEL_PUT_LINK
ZFS_AC_KERNEL_TMPFILE
ZFS_AC_KERNEL_AUTOMOUNT
ZFS_AC_KERNEL_ENCODE_FH_WITH_INODE
ZFS_AC_KERNEL_COMMIT_METADATA
ZFS_AC_KERNEL_CLEAR_INODE
ZFS_AC_KERNEL_SETATTR_PREPARE
ZFS_AC_KERNEL_INSERT_INODE_LOCKED
ZFS_AC_KERNEL_DENTRY
ZFS_AC_KERNEL_DENTRY_ALIAS_D_U
ZFS_AC_KERNEL_TRUNCATE_SETSIZE
ZFS_AC_KERNEL_SECURITY_INODE
ZFS_AC_KERNEL_FST_MOUNT
ZFS_AC_KERNEL_BDI
ZFS_AC_KERNEL_SET_NLINK
ZFS_AC_KERNEL_SGET
ZFS_AC_KERNEL_LSEEK_EXECUTE
ZFS_AC_KERNEL_VFS_FILEMAP_DIRTY_FOLIO
ZFS_AC_KERNEL_VFS_READ_FOLIO
ZFS_AC_KERNEL_VFS_GETATTR
ZFS_AC_KERNEL_VFS_FSYNC_2ARGS
ZFS_AC_KERNEL_VFS_ITERATE
ZFS_AC_KERNEL_VFS_DIRECT_IO
ZFS_AC_KERNEL_VFS_READPAGES
ZFS_AC_KERNEL_VFS_SET_PAGE_DIRTY_NOBUFFERS
ZFS_AC_KERNEL_VFS_RW_ITERATE
ZFS_AC_KERNEL_VFS_GENERIC_WRITE_CHECKS
ZFS_AC_KERNEL_VFS_IOV_ITER
ZFS_AC_KERNEL_KMAP_ATOMIC_ARGS
ZFS_AC_KERNEL_FOLLOW_DOWN_ONE
ZFS_AC_KERNEL_MAKE_REQUEST_FN
ZFS_AC_KERNEL_GENERIC_IO_ACCT
ZFS_AC_KERNEL_FPU
ZFS_AC_KERNEL_FMODE_T
ZFS_AC_KERNEL_KUIDGID_T
ZFS_AC_KERNEL_KUID_HELPERS
ZFS_AC_KERNEL_RENAME
ZFS_AC_KERNEL_CURRENT_TIME
ZFS_AC_KERNEL_USERNS_CAPABILITIES
ZFS_AC_KERNEL_IN_COMPAT_SYSCALL
ZFS_AC_KERNEL_KTIME
ZFS_AC_KERNEL_TOTALRAM_PAGES_FUNC
ZFS_AC_KERNEL_TOTALHIGH_PAGES
ZFS_AC_KERNEL_KSTRTOUL
ZFS_AC_KERNEL_PERCPU
ZFS_AC_KERNEL_CPU_HOTPLUG
- ZFS_AC_KERNEL_GENERIC_FILLATTR_USERNS
+ ZFS_AC_KERNEL_GENERIC_FILLATTR
ZFS_AC_KERNEL_MKNOD
ZFS_AC_KERNEL_SYMLINK
ZFS_AC_KERNEL_BIO_MAX_SEGS
ZFS_AC_KERNEL_SIGNAL_STOP
ZFS_AC_KERNEL_SIGINFO
ZFS_AC_KERNEL_SYSFS
ZFS_AC_KERNEL_SET_SPECIAL_STATE
ZFS_AC_KERNEL_STANDALONE_LINUX_STDARG
ZFS_AC_KERNEL_PAGEMAP_FOLIO_WAIT_BIT
ZFS_AC_KERNEL_ADD_DISK
ZFS_AC_KERNEL_KTHREAD
ZFS_AC_KERNEL_ZERO_PAGE
ZFS_AC_KERNEL___COPY_FROM_USER_INATOMIC
ZFS_AC_KERNEL_USER_NS_COMMON_INUM
ZFS_AC_KERNEL_IDMAP_MNT_API
ZFS_AC_KERNEL_IATTR_VFSID
ZFS_AC_KERNEL_FILEMAP
+ ZFS_AC_KERNEL_WRITEPAGE_T
case "$host_cpu" in
powerpc*)
ZFS_AC_KERNEL_CPU_HAS_FEATURE
ZFS_AC_KERNEL_FLUSH_DCACHE_PAGE
;;
esac
])
dnl #
dnl # Detect name used for Module.symvers file in kernel
dnl #
AC_DEFUN([ZFS_AC_MODULE_SYMVERS], [
modpost=$LINUX/scripts/Makefile.modpost
AC_MSG_CHECKING([kernel file name for module symbols])
AS_IF([test "x$enable_linux_builtin" != xyes -a -f "$modpost"], [
AS_IF([grep -q Modules.symvers $modpost], [
LINUX_SYMBOLS=Modules.symvers
], [
LINUX_SYMBOLS=Module.symvers
])
AS_IF([test ! -f "$LINUX_OBJ/$LINUX_SYMBOLS"], [
AC_MSG_ERROR([
*** Please make sure the kernel devel package for your distribution
*** is installed. If you are building with a custom kernel, make sure
*** the kernel is configured, built, and the '--with-linux=PATH'
*** configure option refers to the location of the kernel source.
])
])
], [
LINUX_SYMBOLS=NONE
])
AC_MSG_RESULT($LINUX_SYMBOLS)
AC_SUBST(LINUX_SYMBOLS)
])
dnl #
dnl # Detect the kernel to be built against
dnl #
dnl # Most modern Linux distributions have separate locations for bare
dnl # source (source) and prebuilt (build) files. Additionally, there are
dnl # `source` and `build` symlinks in `/lib/modules/$(KERNEL_VERSION)`
dnl # pointing to them. The directory search order is now:
dnl #
dnl # - `configure` command line values if both `--with-linux` and
dnl # `--with-linux-obj` were defined
dnl #
dnl # - If only `--with-linux` was defined, `--with-linux-obj` is assumed
dnl # to have the same value as `--with-linux`
dnl #
dnl # - If neither `--with-linux` nor `--with-linux-obj` were defined
dnl # autodetection is used:
dnl #
dnl # - `/lib/modules/$(uname -r)/{source,build}` respectively, if exist.
dnl #
dnl # - If only `/lib/modules/$(uname -r)/build` exists, it is assumed
dnl # to be both source and build directory.
dnl #
dnl # - The first directory in `/lib/modules` with the highest version
dnl # number according to `sort -V` which contains both `source` and
dnl # `build` symlinks/directories. If module directory contains only
dnl # `build` component, it is assumed to be both source and build
dnl # directory.
dnl #
dnl # - Last resort: the first directory matching `/usr/src/kernels/*`
dnl # and `/usr/src/linux-*` with the highest version number according
dnl # to `sort -V` is assumed to be both source and build directory.
dnl #
AC_DEFUN([ZFS_AC_KERNEL], [
AC_ARG_WITH([linux],
AS_HELP_STRING([--with-linux=PATH],
[Path to kernel source]),
[kernelsrc="$withval"])
AC_ARG_WITH(linux-obj,
AS_HELP_STRING([--with-linux-obj=PATH],
[Path to kernel build objects]),
[kernelbuild="$withval"])
AC_MSG_CHECKING([kernel source and build directories])
AS_IF([test -n "$kernelsrc" && test -z "$kernelbuild"], [
kernelbuild="$kernelsrc"
], [test -z "$kernelsrc"], [
AS_IF([test -e "/lib/modules/$(uname -r)/source" && \
test -e "/lib/modules/$(uname -r)/build"], [
src="/lib/modules/$(uname -r)/source"
build="/lib/modules/$(uname -r)/build"
], [test -e "/lib/modules/$(uname -r)/build"], [
build="/lib/modules/$(uname -r)/build"
src="$build"
], [
src=
for d in $(ls -1d /lib/modules/* 2>/dev/null | sort -Vr); do
if test -e "$d/source" && test -e "$d/build"; then
src="$d/source"
build="$d/build"
break
fi
if test -e "$d/build"; then
src="$d/build"
build="$d/build"
break
fi
done
# the least reliable method
if test -z "$src"; then
src=$(ls -1d /usr/src/kernels/* /usr/src/linux-* \
2>/dev/null | grep -v obj | sort -Vr | head -1)
build="$src"
fi
])
AS_IF([test -n "$src" && test -e "$src"], [
kernelsrc=$(readlink -e "$src")
], [
kernelsrc="[Not found]"
])
AS_IF([test -n "$build" && test -e "$build"], [
kernelbuild=$(readlink -e "$build")
], [
kernelbuild="[Not found]"
])
], [
AS_IF([test "$kernelsrc" = "NONE"], [
kernsrcver=NONE
])
withlinux=yes
])
AC_MSG_RESULT([done])
AC_MSG_CHECKING([kernel source directory])
AC_MSG_RESULT([$kernelsrc])
AC_MSG_CHECKING([kernel build directory])
AC_MSG_RESULT([$kernelbuild])
AS_IF([test ! -d "$kernelsrc" || test ! -d "$kernelbuild"], [
AC_MSG_ERROR([
*** Please make sure the kernel devel package for your distribution
*** is installed and then try again. If that fails, you can specify the
*** location of the kernel source and build with the '--with-linux=PATH' and
*** '--with-linux-obj=PATH' options respectively.])
])
AC_MSG_CHECKING([kernel source version])
utsrelease1=$kernelbuild/include/linux/version.h
utsrelease2=$kernelbuild/include/linux/utsrelease.h
utsrelease3=$kernelbuild/include/generated/utsrelease.h
AS_IF([test -r $utsrelease1 && grep -qF UTS_RELEASE $utsrelease1], [
utsrelease=$utsrelease1
], [test -r $utsrelease2 && grep -qF UTS_RELEASE $utsrelease2], [
utsrelease=$utsrelease2
], [test -r $utsrelease3 && grep -qF UTS_RELEASE $utsrelease3], [
utsrelease=$utsrelease3
])
AS_IF([test -n "$utsrelease"], [
kernsrcver=$($AWK '/UTS_RELEASE/ { gsub(/"/, "", $[3]); print $[3] }' $utsrelease)
AS_IF([test -z "$kernsrcver"], [
AC_MSG_RESULT([Not found])
AC_MSG_ERROR([
*** Cannot determine kernel version.
])
])
], [
AC_MSG_RESULT([Not found])
if test "x$enable_linux_builtin" != xyes; then
AC_MSG_ERROR([
*** Cannot find UTS_RELEASE definition.
])
else
AC_MSG_ERROR([
*** Cannot find UTS_RELEASE definition.
*** Please run 'make prepare' inside the kernel source tree.])
fi
])
AC_MSG_RESULT([$kernsrcver])
AS_VERSION_COMPARE([$kernsrcver], [$ZFS_META_KVER_MIN], [
AC_MSG_ERROR([
*** Cannot build against kernel version $kernsrcver.
*** The minimum supported kernel version is $ZFS_META_KVER_MIN.
])
])
LINUX=${kernelsrc}
LINUX_OBJ=${kernelbuild}
LINUX_VERSION=${kernsrcver}
AC_SUBST(LINUX)
AC_SUBST(LINUX_OBJ)
AC_SUBST(LINUX_VERSION)
])
dnl #
dnl # Detect the QAT module to be built against, QAT provides hardware
dnl # acceleration for data compression:
dnl #
dnl # https://01.org/intel-quickassist-technology
dnl #
dnl # 1) Download and install QAT driver from the above link
dnl # 2) Start QAT driver in your system:
dnl # service qat_service start
dnl # 3) Enable QAT in ZFS, e.g.:
dnl # ./configure --with-qat=<qat-driver-path>/QAT1.6
dnl # make
dnl # 4) Set GZIP compression in ZFS dataset:
dnl # zfs set compression = gzip <dataset>
dnl #
dnl # Then the data written to this ZFS pool is compressed by QAT accelerator
dnl # automatically, and de-compressed by QAT when read from the pool.
dnl #
dnl # 1) Get QAT hardware statistics with:
dnl # cat /proc/icp_dh895xcc_dev/qat
dnl # 2) To disable QAT:
dnl # insmod zfs.ko zfs_qat_disable=1
dnl #
AC_DEFUN([ZFS_AC_QAT], [
AC_ARG_WITH([qat],
AS_HELP_STRING([--with-qat=PATH],
[Path to qat source]),
AS_IF([test "$withval" = "yes"],
AC_MSG_ERROR([--with-qat=PATH requires a PATH]),
[qatsrc="$withval"]))
AC_ARG_WITH([qat-obj],
AS_HELP_STRING([--with-qat-obj=PATH],
[Path to qat build objects]),
[qatbuild="$withval"])
AS_IF([test ! -z "${qatsrc}"], [
AC_MSG_CHECKING([qat source directory])
AC_MSG_RESULT([$qatsrc])
QAT_SRC="${qatsrc}/quickassist"
AS_IF([ test ! -e "$QAT_SRC/include/cpa.h"], [
AC_MSG_ERROR([
*** Please make sure the qat driver package is installed
*** and specify the location of the qat source with the
*** '--with-qat=PATH' option then try again. Failed to
*** find cpa.h in:
${QAT_SRC}/include])
])
])
AS_IF([test ! -z "${qatsrc}"], [
AC_MSG_CHECKING([qat build directory])
AS_IF([test -z "$qatbuild"], [
qatbuild="${qatsrc}/build"
])
AC_MSG_RESULT([$qatbuild])
QAT_OBJ=${qatbuild}
AS_IF([ ! test -e "$QAT_OBJ/icp_qa_al.ko" && ! test -e "$QAT_OBJ/qat_api.ko"], [
AC_MSG_ERROR([
*** Please make sure the qat driver is installed then try again.
*** Failed to find icp_qa_al.ko or qat_api.ko in:
$QAT_OBJ])
])
AC_SUBST(QAT_SRC)
AC_SUBST(QAT_OBJ)
AC_DEFINE(HAVE_QAT, 1,
[qat is enabled and existed])
])
dnl #
dnl # Detect the name used for the QAT Module.symvers file.
dnl #
AS_IF([test ! -z "${qatsrc}"], [
AC_MSG_CHECKING([qat file for module symbols])
QAT_SYMBOLS=$QAT_SRC/lookaside/access_layer/src/Module.symvers
AS_IF([test -r $QAT_SYMBOLS], [
AC_MSG_RESULT([$QAT_SYMBOLS])
AC_SUBST(QAT_SYMBOLS)
],[
AC_MSG_ERROR([
*** Please make sure the qat driver is installed then try again.
*** Failed to find Module.symvers in:
$QAT_SYMBOLS
])
])
])
])
dnl #
dnl # ZFS_LINUX_CONFTEST_H
dnl #
AC_DEFUN([ZFS_LINUX_CONFTEST_H], [
test -d build/$2 || mkdir -p build/$2
cat - <<_ACEOF >build/$2/$2.h
$1
_ACEOF
])
dnl #
dnl # ZFS_LINUX_CONFTEST_C
dnl #
AC_DEFUN([ZFS_LINUX_CONFTEST_C], [
test -d build/$2 || mkdir -p build/$2
cat confdefs.h - <<_ACEOF >build/$2/$2.c
$1
_ACEOF
])
dnl #
dnl # ZFS_LINUX_CONFTEST_MAKEFILE
dnl #
dnl # $1 - test case name
dnl # $2 - add to top-level Makefile
dnl # $3 - additional build flags
dnl #
AC_DEFUN([ZFS_LINUX_CONFTEST_MAKEFILE], [
test -d build || mkdir -p build
test -d build/$1 || mkdir -p build/$1
file=build/$1/Makefile
dnl # Example command line to manually build source.
cat - <<_ACEOF >$file
# Example command line to manually build source
# make modules -C $LINUX_OBJ $ARCH_UM M=$PWD/build/$1
ccflags-y := -Werror $FRAME_LARGER_THAN
_ACEOF
dnl # Additional custom CFLAGS as requested.
m4_ifval($3, [echo "ccflags-y += $3" >>$file], [])
dnl # Test case source
echo "obj-m := $1.o" >>$file
AS_IF([test "x$2" = "xyes"], [echo "obj-m += $1/" >>build/Makefile], [])
])
dnl #
dnl # ZFS_LINUX_TEST_PROGRAM(C)([PROLOGUE], [BODY])
dnl #
m4_define([ZFS_LINUX_TEST_PROGRAM], [
#include <linux/module.h>
$1
int
main (void)
{
$2
;
return 0;
}
MODULE_DESCRIPTION("conftest");
MODULE_AUTHOR(ZFS_META_AUTHOR);
MODULE_VERSION(ZFS_META_VERSION "-" ZFS_META_RELEASE);
MODULE_LICENSE($3);
])
dnl #
dnl # ZFS_LINUX_TEST_REMOVE
dnl #
dnl # Removes the specified test source and results.
dnl #
AC_DEFUN([ZFS_LINUX_TEST_REMOVE], [
test -d build/$1 && rm -Rf build/$1
test -f build/Makefile && sed '/$1/d' build/Makefile
])
dnl #
dnl # ZFS_LINUX_COMPILE
dnl #
dnl # $1 - build dir
dnl # $2 - test command
dnl # $3 - pass command
dnl # $4 - fail command
dnl # $5 - set KBUILD_MODPOST_NOFINAL='yes'
dnl # $6 - set KBUILD_MODPOST_WARN='yes'
dnl #
dnl # Used internally by ZFS_LINUX_TEST_{COMPILE,MODPOST}
dnl #
AC_DEFUN([ZFS_LINUX_COMPILE], [
AC_ARG_VAR([KERNEL_CC], [C compiler for
building kernel modules])
AC_ARG_VAR([KERNEL_LD], [Linker for
building kernel modules])
AC_ARG_VAR([KERNEL_LLVM], [Binary option to
build kernel modules with LLVM/CLANG toolchain])
AC_TRY_COMMAND([
KBUILD_MODPOST_NOFINAL="$5" KBUILD_MODPOST_WARN="$6"
make modules -k -j$TEST_JOBS ${KERNEL_CC:+CC=$KERNEL_CC}
${KERNEL_LD:+LD=$KERNEL_LD} ${KERNEL_LLVM:+LLVM=$KERNEL_LLVM}
CONFIG_MODULES=y CFLAGS_MODULE=-DCONFIG_MODULES
-C $LINUX_OBJ $ARCH_UM M=$PWD/$1 >$1/build.log 2>&1])
AS_IF([AC_TRY_COMMAND([$2])], [$3], [$4])
])
dnl #
dnl # ZFS_LINUX_TEST_COMPILE
dnl #
dnl # Perform a full compile excluding the final modpost phase.
dnl #
AC_DEFUN([ZFS_LINUX_TEST_COMPILE], [
ZFS_LINUX_COMPILE([$2], [test -f $2/build.log], [
mv $2/Makefile $2/Makefile.compile.$1
mv $2/build.log $2/build.log.$1
],[
AC_MSG_ERROR([
*** Unable to compile test source to determine kernel interfaces.])
], [yes], [])
])
dnl #
dnl # ZFS_LINUX_TEST_MODPOST
dnl #
dnl # Perform a full compile including the modpost phase. This may
dnl # be an incremental build if the objects have already been built.
dnl #
AC_DEFUN([ZFS_LINUX_TEST_MODPOST], [
ZFS_LINUX_COMPILE([$2], [test -f $2/build.log], [
mv $2/Makefile $2/Makefile.modpost.$1
cat $2/build.log >>build/build.log.$1
],[
AC_MSG_ERROR([
*** Unable to modpost test source to determine kernel interfaces.])
], [], [yes])
])
dnl #
dnl # Perform the compilation of the test cases in two phases.
dnl #
dnl # Phase 1) attempt to build the object files for all of the tests
dnl # defined by the ZFS_LINUX_TEST_SRC macro. But do not
dnl # perform the final modpost stage.
dnl #
dnl # Phase 2) disable all tests which failed the initial compilation,
dnl # then invoke the final modpost step for the remaining tests.
dnl #
dnl # This allows us efficiently build the test cases in parallel while
dnl # remaining resilient to build failures which are expected when
dnl # detecting the available kernel interfaces.
dnl #
dnl # The maximum allowed parallelism can be controlled by setting the
dnl # TEST_JOBS environment variable. Otherwise, it default to $(nproc).
dnl #
AC_DEFUN([ZFS_LINUX_TEST_COMPILE_ALL], [
dnl # Phase 1 - Compilation only, final linking is skipped.
ZFS_LINUX_TEST_COMPILE([$1], [build])
dnl #
dnl # Phase 2 - When building external modules disable test cases
dnl # which failed to compile and invoke modpost to verify the
dnl # final linking.
dnl #
dnl # Test names suffixed with '_license' call modpost independently
dnl # to ensure that a single incompatibility does not result in the
dnl # modpost phase exiting early. This check is not performed on
dnl # every symbol since the majority are compatible and doing so
dnl # would significantly slow down this phase.
dnl #
dnl # When configuring for builtin (--enable-linux-builtin)
dnl # fake the linking step artificially create the expected .ko
dnl # files for tests which did compile. This is required for
dnl # kernels which do not have loadable module support or have
dnl # not yet been built.
dnl #
AS_IF([test "x$enable_linux_builtin" = "xno"], [
for dir in $(awk '/^obj-m/ { print [$]3 }' \
build/Makefile.compile.$1); do
name=${dir%/}
AS_IF([test -f build/$name/$name.o], [
AS_IF([test "${name##*_}" = "license"], [
ZFS_LINUX_TEST_MODPOST([$1],
[build/$name])
echo "obj-n += $dir" >>build/Makefile
], [
echo "obj-m += $dir" >>build/Makefile
])
], [
echo "obj-n += $dir" >>build/Makefile
])
done
ZFS_LINUX_TEST_MODPOST([$1], [build])
], [
for dir in $(awk '/^obj-m/ { print [$]3 }' \
build/Makefile.compile.$1); do
name=${dir%/}
AS_IF([test -f build/$name/$name.o], [
touch build/$name/$name.ko
])
done
])
])
dnl #
dnl # ZFS_LINUX_TEST_SRC
dnl #
dnl # $1 - name
dnl # $2 - global
dnl # $3 - source
dnl # $4 - extra cflags
dnl # $5 - check license-compatibility
dnl #
dnl # Check if the test source is buildable at all and then if it is
dnl # license compatible.
dnl #
dnl # N.B because all of the test cases are compiled in parallel they
dnl # must never depend on the results of previous tests. Each test
dnl # needs to be entirely independent.
dnl #
AC_DEFUN([ZFS_LINUX_TEST_SRC], [
ZFS_LINUX_CONFTEST_C([ZFS_LINUX_TEST_PROGRAM([[$2]], [[$3]],
[["Dual BSD/GPL"]])], [$1])
ZFS_LINUX_CONFTEST_MAKEFILE([$1], [yes], [$4])
AS_IF([ test -n "$5" ], [
ZFS_LINUX_CONFTEST_C([ZFS_LINUX_TEST_PROGRAM(
[[$2]], [[$3]], [[$5]])], [$1_license])
ZFS_LINUX_CONFTEST_MAKEFILE([$1_license], [yes], [$4])
])
])
dnl #
dnl # ZFS_LINUX_TEST_RESULT
dnl #
dnl # $1 - name of a test source (ZFS_LINUX_TEST_SRC)
dnl # $2 - run on success (valid .ko generated)
dnl # $3 - run on failure (unable to compile)
dnl #
AC_DEFUN([ZFS_LINUX_TEST_RESULT], [
AS_IF([test -d build/$1], [
AS_IF([test -f build/$1/$1.ko], [$2], [$3])
], [
AC_MSG_ERROR([
*** No matching source for the "$1" test, check that
*** both the test source and result macros refer to the same name.
])
])
])
dnl #
dnl # ZFS_LINUX_TEST_ERROR
dnl #
dnl # Generic error message which can be used when none of the expected
dnl # kernel interfaces were detected.
dnl #
AC_DEFUN([ZFS_LINUX_TEST_ERROR], [
AC_MSG_ERROR([
*** None of the expected "$1" interfaces were detected.
*** This may be because your kernel version is newer than what is
*** supported, or you are using a patched custom kernel with
*** incompatible modifications.
***
*** ZFS Version: $ZFS_META_ALIAS
*** Compatible Kernels: $ZFS_META_KVER_MIN - $ZFS_META_KVER_MAX
])
])
dnl #
dnl # ZFS_LINUX_TEST_RESULT_SYMBOL
dnl #
dnl # Like ZFS_LINUX_TEST_RESULT except ZFS_CHECK_SYMBOL_EXPORT is called to
dnl # verify symbol exports, unless --enable-linux-builtin was provided to
dnl # configure.
dnl #
AC_DEFUN([ZFS_LINUX_TEST_RESULT_SYMBOL], [
AS_IF([ ! test -f build/$1/$1.ko], [
$5
], [
AS_IF([test "x$enable_linux_builtin" != "xyes"], [
ZFS_CHECK_SYMBOL_EXPORT([$2], [$3], [$4], [$5])
], [
$4
])
])
])
dnl #
dnl # ZFS_LINUX_COMPILE_IFELSE
dnl #
AC_DEFUN([ZFS_LINUX_COMPILE_IFELSE], [
ZFS_LINUX_TEST_REMOVE([conftest])
m4_ifvaln([$1], [ZFS_LINUX_CONFTEST_C([$1], [conftest])])
m4_ifvaln([$5], [ZFS_LINUX_CONFTEST_H([$5], [conftest])],
[ZFS_LINUX_CONFTEST_H([], [conftest])])
ZFS_LINUX_CONFTEST_MAKEFILE([conftest], [no],
[m4_ifvaln([$5], [-I$PWD/build/conftest], [])])
ZFS_LINUX_COMPILE([build/conftest], [$2], [$3], [$4], [], [])
])
dnl #
dnl # ZFS_LINUX_TRY_COMPILE
dnl #
dnl # $1 - global
dnl # $2 - source
dnl # $3 - run on success (valid .ko generated)
dnl # $4 - run on failure (unable to compile)
dnl #
dnl # When configuring as builtin (--enable-linux-builtin) for kernels
dnl # without loadable module support (CONFIG_MODULES=n) only the object
dnl # file is created. See ZFS_LINUX_TEST_COMPILE_ALL for details.
dnl #
AC_DEFUN([ZFS_LINUX_TRY_COMPILE], [
AS_IF([test "x$enable_linux_builtin" = "xyes"], [
ZFS_LINUX_COMPILE_IFELSE(
[ZFS_LINUX_TEST_PROGRAM([[$1]], [[$2]],
[[ZFS_META_LICENSE]])],
[test -f build/conftest/conftest.o], [$3], [$4])
], [
ZFS_LINUX_COMPILE_IFELSE(
[ZFS_LINUX_TEST_PROGRAM([[$1]], [[$2]],
[[ZFS_META_LICENSE]])],
[test -f build/conftest/conftest.ko], [$3], [$4])
])
])
dnl #
dnl # ZFS_CHECK_SYMBOL_EXPORT
dnl #
dnl # Check if a symbol is exported on not by consulting the symbols
dnl # file, or optionally the source code.
dnl #
AC_DEFUN([ZFS_CHECK_SYMBOL_EXPORT], [
grep -q -E '[[[:space:]]]$1[[[:space:]]]' \
$LINUX_OBJ/$LINUX_SYMBOLS 2>/dev/null
rc=$?
if test $rc -ne 0; then
export=0
for file in $2; do
grep -q -E "EXPORT_SYMBOL.*($1)" \
"$LINUX/$file" 2>/dev/null
rc=$?
if test $rc -eq 0; then
export=1
break;
fi
done
if test $export -eq 0; then :
$4
else :
$3
fi
else :
$3
fi
])
dnl #
dnl # ZFS_LINUX_TRY_COMPILE_SYMBOL
dnl #
dnl # Like ZFS_LINUX_TRY_COMPILER except ZFS_CHECK_SYMBOL_EXPORT is called
dnl # to verify symbol exports, unless --enable-linux-builtin was provided
dnl # to configure.
dnl #
AC_DEFUN([ZFS_LINUX_TRY_COMPILE_SYMBOL], [
ZFS_LINUX_TRY_COMPILE([$1], [$2], [rc=0], [rc=1])
if test $rc -ne 0; then :
$6
else
if test "x$enable_linux_builtin" != xyes; then
ZFS_CHECK_SYMBOL_EXPORT([$3], [$4], [rc=0], [rc=1])
fi
if test $rc -ne 0; then :
$6
else :
$5
fi
fi
])
dnl #
dnl # ZFS_LINUX_TRY_COMPILE_HEADER
dnl # like ZFS_LINUX_TRY_COMPILE, except the contents conftest.h are
dnl # provided via the fifth parameter
dnl #
AC_DEFUN([ZFS_LINUX_TRY_COMPILE_HEADER], [
AS_IF([test "x$enable_linux_builtin" = "xyes"], [
ZFS_LINUX_COMPILE_IFELSE(
[ZFS_LINUX_TEST_PROGRAM([[$1]], [[$2]],
[[ZFS_META_LICENSE]])],
[test -f build/conftest/conftest.o], [$3], [$4], [$5])
], [
ZFS_LINUX_COMPILE_IFELSE(
[ZFS_LINUX_TEST_PROGRAM([[$1]], [[$2]],
[[ZFS_META_LICENSE]])],
[test -f build/conftest/conftest.ko], [$3], [$4], [$5])
])
])
dnl #
dnl # AS_VERSION_COMPARE_LE
dnl # like AS_VERSION_COMPARE_LE, but runs $3 if (and only if) $1 <= $2
dnl # AS_VERSION_COMPARE_LE (version-1, version-2, [action-if-less-or-equal], [action-if-greater])
dnl #
AC_DEFUN([AS_VERSION_COMPARE_LE], [
AS_VERSION_COMPARE([$1], [$2], [$3], [$3], [$4])
])
dnl #
dnl # ZFS_LINUX_REQUIRE_API
dnl # like ZFS_LINUX_TEST_ERROR, except only fails if the kernel is
dnl # at least some specified version.
dnl #
AC_DEFUN([ZFS_LINUX_REQUIRE_API], [
AS_VERSION_COMPARE_LE([$2], [$kernsrcver], [
AC_MSG_ERROR([
*** None of the expected "$1" interfaces were detected. This
*** interface is expected for kernels version "$2" and above.
*** This may be because your kernel version is newer than what is
*** supported, or you are using a patched custom kernel with
*** incompatible modifications. Newer kernels may have incompatible
*** APIs.
***
*** ZFS Version: $ZFS_META_ALIAS
*** Compatible Kernels: $ZFS_META_KVER_MIN - $ZFS_META_KVER_MAX
])
], [
AC_MSG_RESULT(no)
])
])
diff --git a/sys/contrib/openzfs/contrib/debian/control.modules.in b/sys/contrib/openzfs/contrib/debian/control.modules.in
index 70a165266d16..34eb7fafba7c 100644
--- a/sys/contrib/openzfs/contrib/debian/control.modules.in
+++ b/sys/contrib/openzfs/contrib/debian/control.modules.in
@@ -1,33 +1,33 @@
Source: openzfs-linux
Section: contrib/kernel
Priority: optional
Maintainer: ZFS on Linux specific mailing list <zfs-discuss@list.zfsonlinux.org>
Build-Depends: debhelper-compat (= 10),
dkms (>> 2.1.1.2-5),
libtool,
- linux-headers-_KVERS_
+ linux-headers-_KVERS_ | raspberrypi-kernel-headers
Standards-Version: 4.3.0
Homepage: http://www.openzfs.org/
Vcs-Git: https://github.com/openzfs/zfs.git
Vcs-Browser: https://github.com/openzfs/zfs
Package: openzfs-zfs-modules-_KVERS_
Architecture: _ARCH_
Provides: openzfs-zfs-modules
-Depends: linux-image-_KVERS_
+Depends: linux-image-_KVERS_ | raspberrypi-kernel
Recommends: openzfsutils
Replaces: zfs-modules-_KVERS_
Conflicts: zfs-modules-_KVERS_
Description: OpenZFS filesystem kernel modules for Linux (kernel _KVERS_)
An advanced integrated volume manager and filesystem that is designed for
performance and data integrity. Snapshots, clones, checksums, deduplication,
compression, and RAID redundancy are built-in features.
.
This package contains the compiled kernel module for _KVERS_
.
Includes the SPA, DMU, ZVOL, and ZPL components of OpenZFS.
.
If you have compiled your own kernel, you will most likely need to build
your own zfs-modules. The zfs-source package has been
provided for use with the Debian kernel-package utility to produce a
version of zfs-module for your kernel.
diff --git a/sys/contrib/openzfs/contrib/dracut/90zfs/module-setup.sh.in b/sys/contrib/openzfs/contrib/dracut/90zfs/module-setup.sh.in
index 78c74e7423bb..e55cb60e1612 100755
--- a/sys/contrib/openzfs/contrib/dracut/90zfs/module-setup.sh.in
+++ b/sys/contrib/openzfs/contrib/dracut/90zfs/module-setup.sh.in
@@ -1,128 +1,131 @@
#!/usr/bin/env bash
# shellcheck disable=SC2154
check() {
# We depend on udev-rules being loaded
[[ "${1}" = "-d" ]] && return 0
# Verify the zfs tool chain
for tool in "zgenhostid" "zpool" "zfs" "mount.zfs"; do
command -v "${tool}" >/dev/null || return 1
done
}
depends() {
echo udev-rules
}
installkernel() {
instmods -c zfs
}
install() {
inst_rules 90-zfs.rules 69-vdev.rules 60-zvol.rules
inst_multiple \
zgenhostid \
zfs \
zpool \
mount.zfs \
hostid \
grep \
awk \
tr \
cut \
head ||
{ dfatal "Failed to install essential binaries"; exit 1; }
# Adapted from https://github.com/zbm-dev/zfsbootmenu
if ! ldd "$(command -v zpool)" | grep -qF 'libgcc_s.so'; then
# On systems with gcc-config (Gentoo, Funtoo, etc.), use it to find libgcc_s
if command -v gcc-config >/dev/null; then
inst_simple "/usr/lib/gcc/$(s=$(gcc-config -c); echo "${s%-*}/${s##*-}")/libgcc_s.so.1" ||
{ dfatal "Unable to install libgcc_s.so"; exit 1; }
# Otherwise, use dracut's library installation function to find the right one
elif ! inst_libdir_file "libgcc_s.so*"; then
# If all else fails, just try looking for some gcc arch directory
inst_simple /usr/lib/gcc/*/*/libgcc_s.so* ||
{ dfatal "Unable to install libgcc_s.so"; exit 1; }
fi
fi
inst_hook cmdline 95 "${moddir}/parse-zfs.sh"
if [[ -n "${systemdutildir}" ]]; then
inst_script "${moddir}/zfs-generator.sh" "${systemdutildir}/system-generators/dracut-zfs-generator"
fi
inst_hook pre-mount 90 "${moddir}/zfs-load-key.sh"
inst_hook mount 98 "${moddir}/mount-zfs.sh"
inst_hook cleanup 99 "${moddir}/zfs-needshutdown.sh"
inst_hook shutdown 20 "${moddir}/export-zfs.sh"
inst_script "${moddir}/zfs-lib.sh" "/lib/dracut-zfs-lib.sh"
# -H ensures they are marked host-only
# -o ensures there is no error upon absence of these files
inst_multiple -o -H \
"@sysconfdir@/zfs/zpool.cache" \
"@sysconfdir@/zfs/vdev_id.conf"
# Synchronize initramfs and system hostid
if ! inst_simple -H @sysconfdir@/hostid; then
if HOSTID="$(hostid 2>/dev/null)" && [[ "${HOSTID}" != "00000000" ]]; then
zgenhostid -o "${initdir}@sysconfdir@/hostid" "${HOSTID}"
mark_hostonly @sysconfdir@/hostid
fi
fi
if dracut_module_included "systemd"; then
inst_simple "${systemdsystemunitdir}/zfs-import.target"
systemctl -q --root "${initdir}" add-wants initrd.target zfs-import.target
inst_simple "${moddir}/zfs-env-bootfs.service" "${systemdsystemunitdir}/zfs-env-bootfs.service"
systemctl -q --root "${initdir}" add-wants zfs-import.target zfs-env-bootfs.service
+ inst_simple "${moddir}/zfs-nonroot-necessities.service" "${systemdsystemunitdir}/zfs-nonroot-necessities.service"
+ systemctl -q --root "${initdir}" add-requires initrd-root-fs.target zfs-nonroot-necessities.service
+
# Add user-provided unit overrides:
# - /etc/systemd/system/${_service}
# - /etc/systemd/system/${_service}.d/overrides.conf
# -H ensures they are marked host-only
# -o ensures there is no error upon absence of these files
inst_multiple -o -H \
"${systemdsystemconfdir}/zfs-import.target" \
"${systemdsystemconfdir}/zfs-import.target.d/"*.conf
for _service in \
"zfs-import-scan.service" \
"zfs-import-cache.service"; do
inst_simple "${systemdsystemunitdir}/${_service}"
systemctl -q --root "${initdir}" add-wants zfs-import.target "${_service}"
# Add user-provided unit overrides:
# - /etc/systemd/system/${_service}
# - /etc/systemd/system/${_service}.d/overrides.conf
# -H ensures they are marked host-only
# -o ensures there is no error upon absence of these files
inst_multiple -o -H \
"${systemdsystemconfdir}/${_service}" \
"${systemdsystemconfdir}/${_service}.d/"*.conf
done
for _service in \
"zfs-snapshot-bootfs.service" \
"zfs-rollback-bootfs.service"; do
inst_simple "${moddir}/${_service}" "${systemdsystemunitdir}/${_service}"
systemctl -q --root "${initdir}" add-wants initrd.target "${_service}"
# Add user-provided unit overrides:
# - /etc/systemd/system/${_service}
# - /etc/systemd/system/${_service}.d/overrides.conf
# -H ensures they are marked host-only
# -o ensures there is no error upon absence of these files
inst_multiple -o -H \
"${systemdsystemconfdir}/${_service}" \
"${systemdsystemconfdir}/${_service}.d/"*.conf
done
inst_simple "${moddir}/import-opts-generator.sh" "${systemdutildir}/system-environment-generators/zfs-import-opts.sh"
fi
}
diff --git a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-env-bootfs.service.in b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-env-bootfs.service.in
index 34c88037cac2..7ebab4c1a58d 100644
--- a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-env-bootfs.service.in
+++ b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-env-bootfs.service.in
@@ -1,14 +1,23 @@
[Unit]
-Description=Set BOOTFS environment for dracut
-Documentation=man:zpool(8)
+Description=Set BOOTFS and BOOTFSFLAGS environment variables for dracut
DefaultDependencies=no
After=zfs-import-cache.service
After=zfs-import-scan.service
Before=zfs-import.target
[Service]
Type=oneshot
-ExecStart=/bin/sh -c "exec systemctl set-environment BOOTFS=$(@sbindir@/zpool list -H -o bootfs | grep -m1 -vFx -)"
+ExecStart=/bin/sh -c ' \
+ . /lib/dracut-zfs-lib.sh; \
+ decode_root_args || exit 0; \
+ [ "$root" = "zfs:AUTO" ] && root="$(@sbindir@/zpool list -H -o bootfs | grep -m1 -vFx -)"; \
+ rootflags="$(getarg rootflags=)"; \
+ case ",$rootflags," in \
+ *,zfsutil,*) ;; \
+ ,,) rootflags=zfsutil ;; \
+ *) rootflags="zfsutil,$rootflags" ;; \
+ esac; \
+ exec systemctl set-environment BOOTFS="$root" BOOTFSFLAGS="$rootflags"'
[Install]
WantedBy=zfs-import.target
diff --git a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-generator.sh.in b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-generator.sh.in
index 56f7ca9785ba..4e1eb7490e0d 100755
--- a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-generator.sh.in
+++ b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-generator.sh.in
@@ -1,99 +1,42 @@
#!/bin/sh
# shellcheck disable=SC2016,SC1004,SC2154
grep -wq debug /proc/cmdline && debug=1
[ -n "$debug" ] && echo "zfs-generator: starting" >> /dev/kmsg
GENERATOR_DIR="$1"
[ -n "$GENERATOR_DIR" ] || {
echo "zfs-generator: no generator directory specified, exiting" >> /dev/kmsg
exit 1
}
# shellcheck source=zfs-lib.sh.in
. /lib/dracut-zfs-lib.sh
decode_root_args || exit 0
-[ -z "${rootflags}" ] && rootflags=$(getarg rootflags=)
-case ",${rootflags}," in
- *,zfsutil,*) ;;
- ,,) rootflags=zfsutil ;;
- *) rootflags="zfsutil,${rootflags}" ;;
-esac
-
[ -n "$debug" ] && echo "zfs-generator: writing extension for sysroot.mount to $GENERATOR_DIR/sysroot.mount.d/zfs-enhancement.conf" >> /dev/kmsg
-mkdir -p "$GENERATOR_DIR"/sysroot.mount.d "$GENERATOR_DIR"/initrd-root-fs.target.requires "$GENERATOR_DIR"/dracut-pre-mount.service.d
+mkdir -p "$GENERATOR_DIR"/sysroot.mount.d "$GENERATOR_DIR"/dracut-pre-mount.service.d
+
{
echo "[Unit]"
echo "Before=initrd-root-fs.target"
echo "After=zfs-import.target"
echo
echo "[Mount]"
- if [ "${root}" = "zfs:AUTO" ]; then
- echo "PassEnvironment=BOOTFS"
- echo 'What=${BOOTFS}'
- else
- echo "What=${root}"
- fi
+ echo "PassEnvironment=BOOTFS BOOTFSFLAGS"
+ echo 'What=${BOOTFS}'
echo "Type=zfs"
- echo "Options=${rootflags}"
+ echo 'Options=${BOOTFSFLAGS}'
} > "$GENERATOR_DIR"/sysroot.mount.d/zfs-enhancement.conf
ln -fs ../sysroot.mount "$GENERATOR_DIR"/initrd-root-fs.target.requires/sysroot.mount
-
-if [ "${root}" = "zfs:AUTO" ]; then
- {
- echo "[Unit]"
- echo "Before=initrd-root-fs.target"
- echo "After=sysroot.mount"
- echo "DefaultDependencies=no"
- echo
- echo "[Service]"
- echo "Type=oneshot"
- echo "PassEnvironment=BOOTFS"
- echo "ExecStart=/bin/sh -c '" ' \
- . /lib/dracut-zfs-lib.sh; \
- _zfs_nonroot_necessities_cb() { \
- zfs mount | grep -m1 -q "^$1 " && return 0; \
- echo "Mounting $1 on /sysroot$2"; \
- mount -o zfsutil -t zfs "$1" "/sysroot$2"; \
- }; \
- for_relevant_root_children "${BOOTFS}" _zfs_nonroot_necessities_cb;' \
- "'"
- } > "$GENERATOR_DIR"/zfs-nonroot-necessities.service
- ln -fs ../zfs-nonroot-necessities.service "$GENERATOR_DIR"/initrd-root-fs.target.requires/zfs-nonroot-necessities.service
-else
- # We can solve this statically at generation time, so do!
- _zfs_generator_cb() {
- dset="${1}"
- mpnt="${2}"
- unit="$(systemd-escape --suffix=mount -p "/sysroot${mpnt}")"
-
- {
- echo "[Unit]"
- echo "Before=initrd-root-fs.target"
- echo "After=sysroot.mount"
- echo
- echo "[Mount]"
- echo "Where=/sysroot${mpnt}"
- echo "What=${dset}"
- echo "Type=zfs"
- echo "Options=zfsutil"
- } > "$GENERATOR_DIR/${unit}"
- ln -fs ../"${unit}" "$GENERATOR_DIR"/initrd-root-fs.target.requires/"${unit}"
- }
-
- for_relevant_root_children "${root}" _zfs_generator_cb
-fi
-
-
{
echo "[Unit]"
echo "After=zfs-import.target"
} > "$GENERATOR_DIR"/dracut-pre-mount.service.d/zfs-enhancement.conf
[ -n "$debug" ] && echo "zfs-generator: finished" >> /dev/kmsg
exit 0
diff --git a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-lib.sh.in b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-lib.sh.in
index 3a43e514d6f9..7139e2e6fe4b 100755
--- a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-lib.sh.in
+++ b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-lib.sh.in
@@ -1,120 +1,120 @@
#!/bin/sh
# shellcheck disable=SC2034
command -v getarg >/dev/null || . /lib/dracut-lib.sh || . /usr/lib/dracut/modules.d/99base/dracut-lib.sh
TAB=" "
ZPOOL_IMPORT_OPTS=
if getargbool 0 zfs_force -y zfs.force -y zfsforce; then
warn "ZFS: Will force-import pools if necessary."
ZPOOL_IMPORT_OPTS=-f
fi
_mount_dataset_cb() {
# shellcheck disable=SC2154
mount -o zfsutil -t zfs "${1}" "${NEWROOT}${2}"
}
# mount_dataset DATASET
# mounts the given zfs dataset.
mount_dataset() {
dataset="${1}"
mountpoint="$(zfs get -H -o value mountpoint "${dataset}")"
ret=0
# We need zfsutil for non-legacy mounts and not for legacy mounts.
if [ "${mountpoint}" = "legacy" ] ; then
mount -t zfs "${dataset}" "${NEWROOT}" || ret=$?
else
mount -o zfsutil -t zfs "${dataset}" "${NEWROOT}" || ret=$?
if [ "$ret" = "0" ]; then
for_relevant_root_children "${dataset}" _mount_dataset_cb || ret=$?
fi
fi
return "${ret}"
}
# for_relevant_root_children DATASET EXEC
# Runs "EXEC dataset mountpoint" for all children of DATASET that are needed for system bringup
-# Used by zfs-generator.sh and friends, too!
+# Used by zfs-nonroot-necessities.service and friends, too!
for_relevant_root_children() {
dataset="${1}"
exec="${2}"
zfs list -t filesystem -Ho name,mountpoint,canmount -r "${dataset}" |
(
_ret=0
while IFS="${TAB}" read -r dataset mountpoint canmount; do
[ "$canmount" != "on" ] && continue
case "$mountpoint" in
/etc|/bin|/lib|/lib??|/libx32|/usr)
# If these aren't mounted we may not be able to get to the real init at all, or pollute the dataset holding the rootfs
"${exec}" "${dataset}" "${mountpoint}" || _ret=$?
;;
*)
# Up to the real init to remount everything else it might need
;;
esac
done
exit "${_ret}"
)
}
# Parse root=, rootfstype=, return them decoded and normalised to zfs:AUTO for auto, plain dset for explicit
#
# True if ZFS-on-root, false if we shouldn't
#
# Supported values:
# root=
# root=zfs
# root=zfs:
# root=zfs:AUTO
#
# root=ZFS=data/set
# root=zfs:data/set
# root=zfs:ZFS=data/set (as a side-effect; allowed but undocumented)
#
# rootfstype=zfs AND root=data/set <=> root=data/set
# rootfstype=zfs AND root= <=> root=zfs:AUTO
#
# '+'es in explicit dataset decoded to ' 's.
decode_root_args() {
if [ -n "$rootfstype" ]; then
[ "$rootfstype" = zfs ]
return
fi
xroot=$(getarg root=)
rootfstype=$(getarg rootfstype=)
# shellcheck disable=SC2249
case "$xroot" in
""|zfs|zfs:|zfs:AUTO)
root=zfs:AUTO
rootfstype=zfs
return 0
;;
ZFS=*|zfs:*)
root="${xroot#zfs:}"
root="${root#ZFS=}"
root=$(echo "$root" | tr '+' ' ')
rootfstype=zfs
return 0
;;
esac
if [ "$rootfstype" = "zfs" ]; then
case "$xroot" in
"") root=zfs:AUTO ;;
*) root=$(echo "$xroot" | tr '+' ' ') ;;
esac
return 0
fi
return 1
}
diff --git a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-nonroot-necessities.service.in b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-nonroot-necessities.service.in
new file mode 100644
index 000000000000..8f420c737c72
--- /dev/null
+++ b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-nonroot-necessities.service.in
@@ -0,0 +1,20 @@
+[Unit]
+Before=initrd-root-fs.target
+After=sysroot.mount
+DefaultDependencies=no
+ConditionEnvironment=BOOTFS
+
+[Service]
+Type=oneshot
+PassEnvironment=BOOTFS
+ExecStart=/bin/sh -c ' \
+ . /lib/dracut-zfs-lib.sh; \
+ _zfs_nonroot_necessities_cb() { \
+ @sbindir@/zfs mount | grep -m1 -q "^$1 " && return 0; \
+ echo "Mounting $1 on /sysroot$2"; \
+ mount -o zfsutil -t zfs "$1" "/sysroot$2"; \
+ }; \
+ for_relevant_root_children "${BOOTFS}" _zfs_nonroot_necessities_cb'
+
+[Install]
+RequiredBy=initrd-root-fs.target
diff --git a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-rollback-bootfs.service.in b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-rollback-bootfs.service.in
index a29cf3a3dd81..68fdcb1f323e 100644
--- a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-rollback-bootfs.service.in
+++ b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-rollback-bootfs.service.in
@@ -1,12 +1,13 @@
[Unit]
Description=Rollback bootfs just before it is mounted
Requisite=zfs-import.target
After=zfs-import.target dracut-pre-mount.service zfs-snapshot-bootfs.service
Before=dracut-mount.service
DefaultDependencies=no
ConditionKernelCommandLine=bootfs.rollback
+ConditionEnvironment=BOOTFS
[Service]
Type=oneshot
-ExecStart=/bin/sh -c '. /lib/dracut-zfs-lib.sh; decode_root_args || exit; [ "$root" = "zfs:AUTO" ] && root="$BOOTFS"; SNAPNAME="$(getarg bootfs.rollback)"; exec @sbindir@/zfs rollback -Rf "$root@${SNAPNAME:-%v}"'
+ExecStart=/bin/sh -c '. /lib/dracut-lib.sh; SNAPNAME="$(getarg bootfs.rollback)"; exec @sbindir@/zfs rollback -Rf "$BOOTFS@${SNAPNAME:-%v}"'
RemainAfterExit=yes
diff --git a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-snapshot-bootfs.service.in b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-snapshot-bootfs.service.in
index 9e73d1a78724..a675b5b2ea98 100644
--- a/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-snapshot-bootfs.service.in
+++ b/sys/contrib/openzfs/contrib/dracut/90zfs/zfs-snapshot-bootfs.service.in
@@ -1,12 +1,13 @@
[Unit]
Description=Snapshot bootfs just before it is mounted
Requisite=zfs-import.target
After=zfs-import.target dracut-pre-mount.service
Before=dracut-mount.service
DefaultDependencies=no
ConditionKernelCommandLine=bootfs.snapshot
+ConditionEnvironment=BOOTFS
[Service]
Type=oneshot
-ExecStart=-/bin/sh -c '. /lib/dracut-zfs-lib.sh; decode_root_args || exit; [ "$root" = "zfs:AUTO" ] && root="$BOOTFS"; SNAPNAME="$(getarg bootfs.snapshot)"; exec @sbindir@/zfs snapshot "$root@${SNAPNAME:-%v}"'
+ExecStart=-/bin/sh -c '. /lib/dracut-lib.sh; SNAPNAME="$(getarg bootfs.snapshot)"; exec @sbindir@/zfs snapshot "$BOOTFS@${SNAPNAME:-%v}"'
RemainAfterExit=yes
diff --git a/sys/contrib/openzfs/contrib/dracut/Makefile.am b/sys/contrib/openzfs/contrib/dracut/Makefile.am
index 73ca52b66316..b432ab76a6d8 100644
--- a/sys/contrib/openzfs/contrib/dracut/Makefile.am
+++ b/sys/contrib/openzfs/contrib/dracut/Makefile.am
@@ -1,27 +1,28 @@
pkgdracut_02dir = $(dracutdir)/modules.d/02zfsexpandknowledge
pkgdracut_02_SCRIPTS = \
%D%/02zfsexpandknowledge/module-setup.sh
pkgdracut_90dir = $(dracutdir)/modules.d/90zfs
pkgdracut_90_SCRIPTS = \
%D%/90zfs/export-zfs.sh \
%D%/90zfs/import-opts-generator.sh \
%D%/90zfs/module-setup.sh \
%D%/90zfs/mount-zfs.sh \
%D%/90zfs/parse-zfs.sh \
%D%/90zfs/zfs-generator.sh \
%D%/90zfs/zfs-lib.sh \
%D%/90zfs/zfs-load-key.sh \
%D%/90zfs/zfs-needshutdown.sh
pkgdracut_90_DATA = \
%D%/90zfs/zfs-env-bootfs.service \
+ %D%/90zfs/zfs-nonroot-necessities.service \
%D%/90zfs/zfs-rollback-bootfs.service \
%D%/90zfs/zfs-snapshot-bootfs.service
SUBSTFILES += $(pkgdracut_02_SCRIPTS) $(pkgdracut_90_SCRIPTS) $(pkgdracut_90_DATA)
SHELLCHECKSCRIPTS += $(pkgdracut_02_SCRIPTS) $(pkgdracut_90_SCRIPTS)
# Provided by /bin/sleep, and, again, every implementation of that supports this
$(call SHELLCHECK_OPTS,$(pkgdracut_90_SCRIPTS)): CHECKBASHISMS_IGNORE = -e 'sleep only takes one integer' -e 'sleep 0.'
dist_noinst_DATA += %D%/README.md
diff --git a/sys/contrib/openzfs/contrib/initramfs/README.md b/sys/contrib/openzfs/contrib/initramfs/README.md
index 34e9bab3c756..68647fa9fc3d 100644
--- a/sys/contrib/openzfs/contrib/initramfs/README.md
+++ b/sys/contrib/openzfs/contrib/initramfs/README.md
@@ -1,84 +1,107 @@
## Description
These scripts are intended to be used with `initramfs-tools`, which is a
similar software product to `dracut` (which is used in Red Hat based
distributions), and is mainly used by Debian GNU/Linux and derivatives.
These scripts share some common functionality with the SysV init scripts,
primarily the `/etc/zfs/zfs-functions` script.
## Configuration
### Root pool/filesystem
Different distributions have their own standard on what to specify on the
kernel command line to boot off a ZFS filesystem.
This script supports the following kernel command line argument combinations
(in this order - first match wins):
* `rpool=<pool>`
* `bootfs=<pool>/<dataset>`
* `rpool=<pool> bootfs=<pool>/<dataset>`
* `-B zfs-bootfs=<pool>/<fs>`
* `root=<pool>/<dataset>`
* `root=ZFS=<pool>/<dataset>`
* `root=zfs:AUTO`
* `root=zfs:<pool>/<dataset>`
* `rpool=rpool`
If a pool is specified, it will be used. Otherwise, in `AUTO` mode, all pools
will be searched. Pools may be excluded from the search by listing them in
`ZFS_POOL_EXCEPTIONS` in `/etc/default/zfs`.
Pools will be imported as follows:
* Try `/dev/disk/by-vdev` if it exists; see `/etc/zfs/vdev_id.conf`.
* Try `/dev/disk/by-id` and any other `/dev/disk/by-*` directories.
* Try `/dev`.
* Use the cache file if nothing else worked.
This order may be modified by setting `ZPOOL_IMPORT_PATH` in
`/etc/default/zfs`.
If a dataset is specified, it will be used as the root filesystem. Otherwise,
this script will attempt to find a root filesystem automatically (in the
specified pool or all pools, as described above).
Filesystems below the root filesystem will be automatically mounted with no
additional configuration necessary. For example, if the root filesystem is
`rpool/ROOT/rootfs`, `rpool/root/rootfs/var`, `rpool/root/rootfs/usr`, etc.
will be mounted (if they exist).
### Snapshots
The `<dataset>` can be a snapshot. In this case, the snapshot will be cloned
and the clone used as the root filesystem. Note:
* If the snapshot does not exist, the base dataset (the part before `@`) is
used as the boot filesystem instead.
* If the resulting clone dataset already exists, it is destroyed.
* The clone is created with `mountpoint=none` and `canmount=noauto`. The root
filesystem is mounted manually by the initramfs script.
* If no snapshot is specified on the `root=` kernel command line, but
there is an `@`, the user will be prompted to choose a snapshot to use.
### Extra options
The following kernel command line arguments are supported:
* `zfsdebug=(on,yes,1)`: Show extra debugging information
* `zfsforce=(on,yes,1)`: Force import the pool
* `rollback=(on,yes,1)`: Rollback to (instead of clone) the snapshot
### Unlocking a ZFS encrypted root over SSH
To use this feature:
1. Install the `dropbear-initramfs` package. You may wish to uninstall the
`cryptsetup-initramfs` package to avoid warnings.
2. Add your SSH key(s) to `/etc/dropbear-initramfs/authorized_keys`. Note
- that Dropbear does not support ed25519 keys before version 2020.79;
+ that Dropbear does not support ed25519 keys before version 2020.79;
in that case, use RSA (2048-bit or more) instead.
3. Rebuild the initramfs with your keys: `update-initramfs -u`
4. During the system boot, login via SSH and run: `zfsunlock`
+
+### Unlocking a ZFS encrypted root via alternate means
+
+If present, a shell program at `/etc/zfs/initramfs-tools-load-key`
+and files matching `/etc/zfs/initramfs-tools-load-key.d/*`
+will be copied to the initramfs during generation
+and sourced to load the key, if required.
+
+The `$ENCRYPTIONROOT` to load the key for and `$KEYLOCATION` variables are set,
+and all initramfs-tools functions are available;
+use unquoted `$ZPOOL` and `$ZFS` to run `zpool` and `zfs`.
+
+A successful return (and loaded key) stops the search.
+A failure return is non-fatal,
+and loading keys proceeds as normal if no hook succeeds.
+
+A trivial example of a key-loading drop-in that uses the BLAKE2 checksum
+of the file at the `keylocation` as the key follows.
+
+```sh
+key="$(b2sum "${KEYLOCATION#file://}")" || return
+printf '%s\n' "${key%% *}" | $ZFS load-key -L prompt "$ENCRYPTIONROOT"
+```
diff --git a/sys/contrib/openzfs/contrib/initramfs/hooks/zfs.in b/sys/contrib/openzfs/contrib/initramfs/hooks/zfs.in
index 28dd252eea52..6cd7e6f1ea3b 100755
--- a/sys/contrib/openzfs/contrib/initramfs/hooks/zfs.in
+++ b/sys/contrib/openzfs/contrib/initramfs/hooks/zfs.in
@@ -1,56 +1,59 @@
#!/bin/sh
#
# Add OpenZFS filesystem capabilities to an initrd, usually for a native ZFS root.
#
if [ "$1" = "prereqs" ]; then
echo "udev"
exit
fi
. /usr/share/initramfs-tools/hook-functions
for req in "@sbindir@/zpool" "@sbindir@/zfs" "@mounthelperdir@/mount.zfs"; do
copy_exec "$req" || {
echo "$req not available!" >&2
exit 2
}
done
copy_exec "@udevdir@/vdev_id"
copy_exec "@udevdir@/zvol_id"
if command -v systemd-ask-password > /dev/null; then
copy_exec "$(command -v systemd-ask-password)"
fi
# We use pthreads, but i-t from buster doesn't automatically
# copy this indirect dependency: this can be removed when buster finally dies.
find /lib/ -type f -name "libgcc_s.so.[1-9]" | while read -r libgcc; do
copy_exec "$libgcc"
done
# shellcheck disable=SC2050
if [ @LIBFETCH_DYNAMIC@ -gt 0 ]; then
find /lib/ -name "@LIBFETCH_SONAME@" | while read -r libfetch; do
copy_exec "$libfetch"
done
fi
copy_file config "/etc/hostid"
copy_file cache "@sysconfdir@/zfs/zpool.cache"
copy_file config "@initconfdir@/zfs"
copy_file config "@sysconfdir@/zfs/zfs-functions"
copy_file config "@sysconfdir@/zfs/vdev_id.conf"
+for f in "@sysconfdir@/zfs/initramfs-tools-load-key" "@sysconfdir@/zfs/initramfs-tools-load-key.d/"*; do
+ copy_file config "$f"
+done
copy_file rule "@udevruledir@/60-zvol.rules"
copy_file rule "@udevruledir@/69-vdev.rules"
manual_add_modules zfs
if [ -f "/etc/hostname" ]; then
copy_file config "/etc/hostname"
else
hostname="$(mktemp -t hostname.XXXXXXXXXX)"
hostname > "$hostname"
copy_file config "$hostname" "/etc/hostname"
rm -f "$hostname"
fi
diff --git a/sys/contrib/openzfs/contrib/initramfs/scripts/zfs b/sys/contrib/openzfs/contrib/initramfs/scripts/zfs
index 23aa95efc8f4..7f977a30f75b 100644
--- a/sys/contrib/openzfs/contrib/initramfs/scripts/zfs
+++ b/sys/contrib/openzfs/contrib/initramfs/scripts/zfs
@@ -1,1012 +1,1022 @@
# ZFS boot stub for initramfs-tools.
#
# In the initramfs environment, the /init script sources this stub to
# override the default functions in the /scripts/local script.
#
# Enable this by passing boot=zfs on the kernel command line.
#
# $quiet, $root, $rpool, $bootfs come from the cmdline:
# shellcheck disable=SC2154
# Source the common functions
. /etc/zfs/zfs-functions
# Start interactive shell.
# Use debian's panic() if defined, because it allows to prevent shell access
# by setting panic in cmdline (e.g. panic=0 or panic=15).
# See "4.5 Disable root prompt on the initramfs" of Securing Debian Manual:
# https://www.debian.org/doc/manuals/securing-debian-howto/ch4.en.html
shell() {
if command -v panic > /dev/null 2>&1; then
panic
else
/bin/sh
fi
}
# This runs any scripts that should run before we start importing
# pools and mounting any filesystems.
pre_mountroot()
{
if command -v run_scripts > /dev/null 2>&1
then
if [ -f "/scripts/local-top" ] || [ -d "/scripts/local-top" ]
then
[ "$quiet" != "y" ] && \
zfs_log_begin_msg "Running /scripts/local-top"
run_scripts /scripts/local-top
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
if [ -f "/scripts/local-premount" ] || [ -d "/scripts/local-premount" ]
then
[ "$quiet" != "y" ] && \
zfs_log_begin_msg "Running /scripts/local-premount"
run_scripts /scripts/local-premount
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
fi
}
# If plymouth is available, hide the splash image.
disable_plymouth()
{
if [ -x /bin/plymouth ] && /bin/plymouth --ping
then
/bin/plymouth hide-splash >/dev/null 2>&1
fi
}
# Get a ZFS filesystem property value.
get_fs_value()
{
fs="$1"
value=$2
"${ZFS}" get -H -ovalue "$value" "$fs" 2> /dev/null
}
# Find the 'bootfs' property on pool $1.
# If the property does not contain '/', then ignore this
# pool by exporting it again.
find_rootfs()
{
pool="$1"
# If 'POOL_IMPORTED' isn't set, no pool imported and therefore
# we won't be able to find a root fs.
[ -z "${POOL_IMPORTED}" ] && return 1
# If it's already specified, just keep it mounted and exit
# User (kernel command line) must be correct.
if [ -n "${ZFS_BOOTFS}" ] && [ "${ZFS_BOOTFS}" != "zfs:AUTO" ]; then
return 0
fi
# Not set, try to find it in the 'bootfs' property of the pool.
# NOTE: zpool does not support 'get -H -ovalue bootfs'...
ZFS_BOOTFS=$("${ZPOOL}" list -H -obootfs "$pool")
# Make sure it's not '-' and that it starts with /.
if [ "${ZFS_BOOTFS}" != "-" ] && \
get_fs_value "${ZFS_BOOTFS}" mountpoint | grep -q '^/$'
then
# Keep it mounted
POOL_IMPORTED=1
return 0
fi
# Not boot fs here, export it and later try again..
"${ZPOOL}" export "$pool"
POOL_IMPORTED=
ZFS_BOOTFS=
return 1
}
# Support function to get a list of all pools, separated with ';'
find_pools()
{
pools=$("$@" 2> /dev/null | \
sed -Ee '/pool:|^[a-zA-Z0-9]/!d' -e 's@.*: @@' | \
tr '\n' ';')
echo "${pools%%;}" # Return without the last ';'.
}
# Get a list of all available pools
get_pools()
{
if [ -n "${ZFS_POOL_IMPORT}" ]; then
echo "$ZFS_POOL_IMPORT"
return 0
fi
# Get the base list of available pools.
available_pools=$(find_pools "$ZPOOL" import)
# Just in case - seen it happen (that a pool isn't visible/found
# with a simple "zpool import" but only when using the "-d"
# option or setting ZPOOL_IMPORT_PATH).
if [ -d "/dev/disk/by-id" ]
then
npools=$(find_pools "$ZPOOL" import -d /dev/disk/by-id)
if [ -n "$npools" ]
then
# Because we have found extra pool(s) here, which wasn't
# found 'normally', we need to force USE_DISK_BY_ID to
# make sure we're able to actually import it/them later.
USE_DISK_BY_ID='yes'
if [ -n "$available_pools" ]
then
# Filter out duplicates (pools found with the simple
# "zpool import" but which is also found with the
# "zpool import -d ...").
npools=$(echo "$npools" | sed "s,$available_pools,,")
# Add the list to the existing list of
# available pools
available_pools="$available_pools;$npools"
else
available_pools="$npools"
fi
fi
fi
# Filter out any exceptions...
if [ -n "$ZFS_POOL_EXCEPTIONS" ]
then
found=""
apools=""
OLD_IFS="$IFS" ; IFS=";"
for pool in $available_pools
do
for exception in $ZFS_POOL_EXCEPTIONS
do
[ "$pool" = "$exception" ] && continue 2
found="$pool"
done
if [ -n "$found" ]
then
if [ -n "$apools" ]
then
apools="$apools;$pool"
else
apools="$pool"
fi
fi
done
IFS="$OLD_IFS"
available_pools="$apools"
fi
# Return list of available pools.
echo "$available_pools"
}
# Import given pool $1
import_pool()
{
pool="$1"
# Verify that the pool isn't already imported
# Make as sure as we can to not require '-f' to import.
"${ZPOOL}" get -H -o value name,guid 2>/dev/null | grep -Fxq "$pool" && return 0
# For backwards compatibility, make sure that ZPOOL_IMPORT_PATH is set
# to something we can use later with the real import(s). We want to
# make sure we find all by* dirs, BUT by-vdev should be first (if it
# exists).
if [ -n "$USE_DISK_BY_ID" ] && [ -z "$ZPOOL_IMPORT_PATH" ]
then
dirs="$(for dir in /dev/disk/by-*
do
# Ignore by-vdev here - we want it first!
echo "$dir" | grep -q /by-vdev && continue
[ ! -d "$dir" ] && continue
printf "%s" "$dir:"
done | sed 's,:$,,g')"
if [ -d "/dev/disk/by-vdev" ]
then
# Add by-vdev at the beginning.
ZPOOL_IMPORT_PATH="/dev/disk/by-vdev:"
fi
# ... and /dev at the very end, just for good measure.
ZPOOL_IMPORT_PATH="$ZPOOL_IMPORT_PATH$dirs:/dev"
fi
# Needs to be exported for "zpool" to catch it.
[ -n "$ZPOOL_IMPORT_PATH" ] && export ZPOOL_IMPORT_PATH
[ "$quiet" != "y" ] && zfs_log_begin_msg \
"Importing pool '${pool}' using defaults"
ZFS_CMD="${ZPOOL} import -N ${ZPOOL_FORCE} ${ZPOOL_IMPORT_OPTS}"
ZFS_STDERR="$($ZFS_CMD "$pool" 2>&1)"
ZFS_ERROR="$?"
if [ "${ZFS_ERROR}" != 0 ]
then
[ "$quiet" != "y" ] && zfs_log_failure_msg "${ZFS_ERROR}"
if [ -f "${ZPOOL_CACHE}" ]
then
[ "$quiet" != "y" ] && zfs_log_begin_msg \
"Importing pool '${pool}' using cachefile."
ZFS_CMD="${ZPOOL} import -c ${ZPOOL_CACHE} -N ${ZPOOL_FORCE} ${ZPOOL_IMPORT_OPTS}"
ZFS_STDERR="$($ZFS_CMD "$pool" 2>&1)"
ZFS_ERROR="$?"
fi
if [ "${ZFS_ERROR}" != 0 ]
then
[ "$quiet" != "y" ] && zfs_log_failure_msg "${ZFS_ERROR}"
disable_plymouth
echo ""
echo "Command: ${ZFS_CMD} '$pool'"
echo "Message: $ZFS_STDERR"
echo "Error: $ZFS_ERROR"
echo ""
echo "Failed to import pool '$pool'."
echo "Manually import the pool and exit."
shell
fi
fi
[ "$quiet" != "y" ] && zfs_log_end_msg
POOL_IMPORTED=1
return 0
}
# Load ZFS modules
# Loading a module in a initrd require a slightly different approach,
# with more logging etc.
load_module_initrd()
{
ZFS_INITRD_PRE_MOUNTROOT_SLEEP=${ROOTDELAY:-0}
if [ "$ZFS_INITRD_PRE_MOUNTROOT_SLEEP" -gt 0 ]; then
[ "$quiet" != "y" ] && zfs_log_begin_msg "Delaying for up to '${ZFS_INITRD_PRE_MOUNTROOT_SLEEP}' seconds."
fi
START=$(/bin/date -u +%s)
END=$((START+ZFS_INITRD_PRE_MOUNTROOT_SLEEP))
while true; do
# Wait for all of the /dev/{hd,sd}[a-z] device nodes to appear.
if command -v wait_for_udev > /dev/null 2>&1 ; then
wait_for_udev 10
elif command -v wait_for_dev > /dev/null 2>&1 ; then
wait_for_dev
fi
#
# zpool import refuse to import without a valid
# /proc/self/mounts
#
[ ! -f /proc/self/mounts ] && mount proc /proc
# Load the module
if load_module "zfs"; then
ret=0
break
else
ret=1
fi
[ "$(/bin/date -u +%s)" -gt "$END" ] && break
sleep 1
done
if [ "$ZFS_INITRD_PRE_MOUNTROOT_SLEEP" -gt 0 ]; then
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
[ "$ret" -ne 0 ] && return 1
if [ "$ZFS_INITRD_POST_MODPROBE_SLEEP" -gt 0 ] 2>/dev/null
then
if [ "$quiet" != "y" ]; then
zfs_log_begin_msg "Sleeping for" \
"$ZFS_INITRD_POST_MODPROBE_SLEEP seconds..."
fi
sleep "$ZFS_INITRD_POST_MODPROBE_SLEEP"
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
return 0
}
# Mount a given filesystem
mount_fs()
{
fs="$1"
# Check that the filesystem exists
"${ZFS}" list -oname -tfilesystem -H "${fs}" > /dev/null 2>&1 || return 1
# Skip filesystems with canmount=off. The root fs should not have
# canmount=off, but ignore it for backwards compatibility just in case.
if [ "$fs" != "${ZFS_BOOTFS}" ]
then
canmount=$(get_fs_value "$fs" canmount)
[ "$canmount" = "off" ] && return 0
fi
# Need the _original_ datasets mountpoint!
mountpoint=$(get_fs_value "$fs" mountpoint)
ZFS_CMD="mount.zfs -o zfsutil"
if [ "$mountpoint" = "legacy" ] || [ "$mountpoint" = "none" ]; then
# Can't use the mountpoint property. Might be one of our
# clones. Check the 'org.zol:mountpoint' property set in
# clone_snap() if that's usable.
mountpoint1=$(get_fs_value "$fs" org.zol:mountpoint)
if [ "$mountpoint1" = "legacy" ] ||
[ "$mountpoint1" = "none" ] ||
[ "$mountpoint1" = "-" ]
then
if [ "$fs" != "${ZFS_BOOTFS}" ]; then
# We don't have a proper mountpoint and this
# isn't the root fs.
return 0
fi
# Don't use mount.zfs -o zfsutils for legacy mountpoint
if [ "$mountpoint" = "legacy" ]; then
ZFS_CMD="mount.zfs"
fi
# Last hail-mary: Hope 'rootmnt' is set!
mountpoint=""
else
mountpoint="$mountpoint1"
fi
fi
# Possibly decrypt a filesystem using native encryption.
decrypt_fs "$fs"
[ "$quiet" != "y" ] && \
zfs_log_begin_msg "Mounting '${fs}' on '${rootmnt}/${mountpoint}'"
[ -n "${ZFS_DEBUG}" ] && \
zfs_log_begin_msg "CMD: '$ZFS_CMD ${fs} ${rootmnt}/${mountpoint}'"
ZFS_STDERR=$(${ZFS_CMD} "${fs}" "${rootmnt}/${mountpoint}" 2>&1)
ZFS_ERROR=$?
if [ "${ZFS_ERROR}" != 0 ]
then
[ "$quiet" != "y" ] && zfs_log_failure_msg "${ZFS_ERROR}"
disable_plymouth
echo ""
echo "Command: ${ZFS_CMD} ${fs} ${rootmnt}/${mountpoint}"
echo "Message: $ZFS_STDERR"
echo "Error: $ZFS_ERROR"
echo ""
echo "Failed to mount ${fs} on ${rootmnt}/${mountpoint}."
echo "Manually mount the filesystem and exit."
shell
else
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
return 0
}
# Unlock a ZFS native encrypted filesystem.
decrypt_fs()
{
fs="$1"
# If pool encryption is active and the zfs command understands '-o encryption'
if [ "$(zpool list -H -o feature@encryption "${fs%%/*}")" = 'active' ]; then
# Determine dataset that holds key for root dataset
ENCRYPTIONROOT="$(get_fs_value "${fs}" encryptionroot)"
KEYLOCATION="$(get_fs_value "${ENCRYPTIONROOT}" keylocation)"
echo "${ENCRYPTIONROOT}" > /run/zfs_fs_name
# If root dataset is encrypted...
if ! [ "${ENCRYPTIONROOT}" = "-" ]; then
KEYSTATUS="$(get_fs_value "${ENCRYPTIONROOT}" keystatus)"
# Continue only if the key needs to be loaded
[ "$KEYSTATUS" = "unavailable" ] || return 0
+ # Try extensions first
+ for f in "/etc/zfs/initramfs-tools-load-key" "/etc/zfs/initramfs-tools-load-key.d/"*; do
+ [ -r "$f" ] || continue
+ (. "$f") && {
+ # Successful return and actually-loaded key: we're done
+ KEYSTATUS="$(get_fs_value "${ENCRYPTIONROOT}" keystatus)"
+ [ "$KEYSTATUS" = "unavailable" ] || return 0
+ }
+ done
+
# Do not prompt if key is stored noninteractively,
if ! [ "${KEYLOCATION}" = "prompt" ]; then
$ZFS load-key "${ENCRYPTIONROOT}"
# Prompt with plymouth, if active
elif /bin/plymouth --ping 2>/dev/null; then
echo "plymouth" > /run/zfs_console_askpwd_cmd
for _ in 1 2 3; do
plymouth ask-for-password --prompt "Encrypted ZFS password for ${ENCRYPTIONROOT}" | \
$ZFS load-key "${ENCRYPTIONROOT}" && break
done
# Prompt with systemd, if active
elif [ -e /run/systemd/system ]; then
echo "systemd-ask-password" > /run/zfs_console_askpwd_cmd
for _ in 1 2 3; do
systemd-ask-password --no-tty "Encrypted ZFS password for ${ENCRYPTIONROOT}" | \
$ZFS load-key "${ENCRYPTIONROOT}" && break
done
# Prompt with ZFS tty, otherwise
else
# Temporarily setting "printk" to "7" allows the prompt to appear even when the "quiet" kernel option has been used
echo "load-key" > /run/zfs_console_askpwd_cmd
read -r storeprintk _ < /proc/sys/kernel/printk
echo 7 > /proc/sys/kernel/printk
$ZFS load-key "${ENCRYPTIONROOT}"
echo "$storeprintk" > /proc/sys/kernel/printk
fi
fi
fi
return 0
}
# Destroy a given filesystem.
destroy_fs()
{
fs="$1"
[ "$quiet" != "y" ] && \
zfs_log_begin_msg "Destroying '$fs'"
ZFS_CMD="${ZFS} destroy $fs"
ZFS_STDERR="$(${ZFS_CMD} 2>&1)"
ZFS_ERROR="$?"
if [ "${ZFS_ERROR}" != 0 ]
then
[ "$quiet" != "y" ] && zfs_log_failure_msg "${ZFS_ERROR}"
disable_plymouth
echo ""
echo "Command: $ZFS_CMD"
echo "Message: $ZFS_STDERR"
echo "Error: $ZFS_ERROR"
echo ""
echo "Failed to destroy '$fs'. Please make sure that '$fs' is not available."
echo "Hint: Try: zfs destroy -Rfn $fs"
echo "If this dryrun looks good, then remove the 'n' from '-Rfn' and try again."
shell
else
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
return 0
}
# Clone snapshot $1 to destination filesystem $2
# Set 'canmount=noauto' and 'mountpoint=none' so that we get to keep
# manual control over it's mounting (i.e., make sure it's not automatically
# mounted with a 'zfs mount -a' in the init/systemd scripts).
clone_snap()
{
snap="$1"
destfs="$2"
mountpoint="$3"
[ "$quiet" != "y" ] && zfs_log_begin_msg "Cloning '$snap' to '$destfs'"
# Clone the snapshot into a dataset we can boot from
# + We don't want this filesystem to be automatically mounted, we
# want control over this here and nowhere else.
# + We don't need any mountpoint set for the same reason.
# We use the 'org.zol:mountpoint' property to remember the mountpoint.
ZFS_CMD="${ZFS} clone -o canmount=noauto -o mountpoint=none"
ZFS_CMD="${ZFS_CMD} -o org.zol:mountpoint=${mountpoint}"
ZFS_CMD="${ZFS_CMD} $snap $destfs"
ZFS_STDERR="$(${ZFS_CMD} 2>&1)"
ZFS_ERROR="$?"
if [ "${ZFS_ERROR}" != 0 ]
then
[ "$quiet" != "y" ] && zfs_log_failure_msg "${ZFS_ERROR}"
disable_plymouth
echo ""
echo "Command: $ZFS_CMD"
echo "Message: $ZFS_STDERR"
echo "Error: $ZFS_ERROR"
echo ""
echo "Failed to clone snapshot."
echo "Make sure that any problems are corrected and then make sure"
echo "that the dataset '$destfs' exists and is bootable."
shell
else
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
return 0
}
# Rollback a given snapshot.
rollback_snap()
{
snap="$1"
[ "$quiet" != "y" ] && zfs_log_begin_msg "Rollback $snap"
ZFS_CMD="${ZFS} rollback -Rf $snap"
ZFS_STDERR="$(${ZFS_CMD} 2>&1)"
ZFS_ERROR="$?"
if [ "${ZFS_ERROR}" != 0 ]
then
[ "$quiet" != "y" ] && zfs_log_failure_msg "${ZFS_ERROR}"
disable_plymouth
echo ""
echo "Command: $ZFS_CMD"
echo "Message: $ZFS_STDERR"
echo "Error: $ZFS_ERROR"
echo ""
echo "Failed to rollback snapshot."
shell
else
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
return 0
}
# Get a list of snapshots, give them as a numbered list
# to the user to choose from.
ask_user_snap()
{
fs="$1"
# We need to temporarily disable debugging. Set 'debug' so we
# remember to enabled it again.
if [ -n "${ZFS_DEBUG}" ]; then
unset ZFS_DEBUG
set +x
debug=1
fi
# Because we need the resulting snapshot, which is sent on
# stdout to the caller, we use stderr for our questions.
echo "What snapshot do you want to boot from?" > /dev/stderr
# shellcheck disable=SC2046
IFS="
" set -- $("${ZFS}" list -H -oname -tsnapshot -r "${fs}")
i=1
for snap in "$@"; do
echo " $i: $snap"
i=$((i + 1))
done > /dev/stderr
# expr instead of test here because [ a -lt 0 ] errors out,
# but expr falls back to lexicographical, which works out right
snapnr=0
while expr "$snapnr" "<" 1 > /dev/null ||
expr "$snapnr" ">" "$#" > /dev/null
do
printf "%s" "Snap nr [1-$#]? " > /dev/stderr
read -r snapnr
done
# Re-enable debugging.
if [ -n "${debug}" ]; then
ZFS_DEBUG=1
set -x
fi
eval echo '$'"$snapnr"
}
setup_snapshot_booting()
{
snap="$1"
retval=0
# Make sure that the snapshot specified actually exists.
if [ -z "$(get_fs_value "${snap}" type)" ]
then
# Snapshot does not exist (...@<null> ?)
# ask the user for a snapshot to use.
snap="$(ask_user_snap "${snap%%@*}")"
fi
# Separate the full snapshot ('$snap') into it's filesystem and
# snapshot names. Would have been nice with a split() function..
rootfs="${snap%%@*}"
snapname="${snap##*@}"
ZFS_BOOTFS="${rootfs}_${snapname}"
if ! grep -qiE '(^|[^\\](\\\\)* )(rollback)=(on|yes|1)( |$)' /proc/cmdline
then
# If the destination dataset for the clone
# already exists, destroy it. Recursively
if [ -n "$(get_fs_value "${rootfs}_${snapname}" type)" ]
then
filesystems=$("${ZFS}" list -oname -tfilesystem -H \
-r -Sname "${ZFS_BOOTFS}")
for fs in $filesystems; do
destroy_fs "${fs}"
done
fi
fi
# Get all snapshots, recursively (might need to clone /usr, /var etc
# as well).
for s in $("${ZFS}" list -H -oname -tsnapshot -r "${rootfs}" | \
grep "${snapname}")
do
if grep -qiE '(^|[^\\](\\\\)* )(rollback)=(on|yes|1)( |$)' /proc/cmdline
then
# Rollback snapshot
rollback_snap "$s" || retval=$((retval + 1))
ZFS_BOOTFS="${rootfs}"
else
# Setup a destination filesystem name.
# Ex: Called with 'rpool/ROOT/debian@snap2'
# rpool/ROOT/debian@snap2 => rpool/ROOT/debian_snap2
# rpool/ROOT/debian/boot@snap2 => rpool/ROOT/debian_snap2/boot
# rpool/ROOT/debian/usr@snap2 => rpool/ROOT/debian_snap2/usr
# rpool/ROOT/debian/var@snap2 => rpool/ROOT/debian_snap2/var
subfs="${s##"$rootfs"}"
subfs="${subfs%%@"$snapname"}"
destfs="${rootfs}_${snapname}" # base fs.
[ -n "$subfs" ] && destfs="${destfs}$subfs" # + sub fs.
# Get the mountpoint of the filesystem, to be used
# with clone_snap(). If legacy or none, then use
# the sub fs value.
mountpoint=$(get_fs_value "${s%%@*}" mountpoint)
if [ "$mountpoint" = "legacy" ] || \
[ "$mountpoint" = "none" ]
then
if [ -n "${subfs}" ]; then
mountpoint="${subfs}"
else
mountpoint="/"
fi
fi
# Clone the snapshot into its own
# filesystem
clone_snap "$s" "${destfs}" "${mountpoint}" || \
retval=$((retval + 1))
fi
done
# If we haven't return yet, we have a problem...
return "${retval}"
}
# ================================================================
# This is the main function.
mountroot()
{
# ----------------------------------------------------------------
# I N I T I A L S E T U P
# ------------
# Run the pre-mount scripts from /scripts/local-top.
pre_mountroot
# ------------
# Source the default setup variables.
[ -r '/etc/default/zfs' ] && . /etc/default/zfs
# ------------
# Support debug option
if grep -qiE '(^|[^\\](\\\\)* )(zfs_debug|zfs\.debug|zfsdebug)=(on|yes|1)( |$)' /proc/cmdline
then
ZFS_DEBUG=1
mkdir /var/log
#exec 2> /var/log/boot.debug
set -x
fi
# ------------
# Load ZFS module etc.
if ! load_module_initrd; then
disable_plymouth
echo ""
echo "Failed to load ZFS modules."
echo "Manually load the modules and exit."
shell
fi
# ------------
# Look for the cache file (if any).
[ -f "${ZPOOL_CACHE}" ] || unset ZPOOL_CACHE
[ -s "${ZPOOL_CACHE}" ] || unset ZPOOL_CACHE
# ------------
# Compatibility: 'ROOT' is for Debian GNU/Linux (etc),
# 'root' is for Redhat/Fedora (etc),
# 'REAL_ROOT' is for Gentoo
if [ -z "$ROOT" ]
then
[ -n "$root" ] && ROOT=${root}
[ -n "$REAL_ROOT" ] && ROOT=${REAL_ROOT}
fi
# ------------
# Where to mount the root fs in the initrd - set outside this script
# Compatibility: 'rootmnt' is for Debian GNU/Linux (etc),
# 'NEWROOT' is for RedHat/Fedora (etc),
# 'NEW_ROOT' is for Gentoo
if [ -z "$rootmnt" ]
then
[ -n "$NEWROOT" ] && rootmnt=${NEWROOT}
[ -n "$NEW_ROOT" ] && rootmnt=${NEW_ROOT}
fi
# ------------
# No longer set in the defaults file, but it could have been set in
# get_pools() in some circumstances. If it's something, but not 'yes',
# it's no good to us.
[ -n "$USE_DISK_BY_ID" ] && [ "$USE_DISK_BY_ID" != 'yes' ] && \
unset USE_DISK_BY_ID
# ----------------------------------------------------------------
# P A R S E C O M M A N D L I N E O P T I O N S
# This part is the really ugly part - there's so many options and permutations
# 'out there', and if we should make this the 'primary' source for ZFS initrd
# scripting, we need/should support them all.
#
# Supports the following kernel command line argument combinations
# (in this order - first match win):
#
# rpool=<pool> (tries to finds bootfs automatically)
# bootfs=<pool>/<dataset> (uses this for rpool - first part)
# rpool=<pool> bootfs=<pool>/<dataset>
# -B zfs-bootfs=<pool>/<fs> (uses this for rpool - first part)
# rpool=rpool (default if none of the above is used)
# root=<pool>/<dataset> (uses this for rpool - first part)
# root=ZFS=<pool>/<dataset> (uses this for rpool - first part, without 'ZFS=')
# root=zfs:AUTO (tries to detect both pool and rootfs)
# root=zfs:<pool>/<dataset> (uses this for rpool - first part, without 'zfs:')
#
# Option <dataset> could also be <snapshot>
# Option <pool> could also be <guid>
# ------------
# Support force option
# In addition, setting one of zfs_force, zfs.force or zfsforce to
# 'yes', 'on' or '1' will make sure we force import the pool.
# This should (almost) never be needed, but it's here for
# completeness.
ZPOOL_FORCE=""
if grep -qiE '(^|[^\\](\\\\)* )(zfs_force|zfs\.force|zfsforce)=(on|yes|1)( |$)' /proc/cmdline
then
ZPOOL_FORCE="-f"
fi
# ------------
# Look for 'rpool' and 'bootfs' parameter
[ -n "$rpool" ] && ZFS_RPOOL="${rpool#rpool=}"
[ -n "$bootfs" ] && ZFS_BOOTFS="${bootfs#bootfs=}"
# ------------
# If we have 'ROOT' (see above), but not 'ZFS_BOOTFS', then use
# 'ROOT'
[ -n "$ROOT" ] && [ -z "${ZFS_BOOTFS}" ] && ZFS_BOOTFS="$ROOT"
# ------------
# Check for the `-B zfs-bootfs=%s/%u,...` kind of parameter.
# NOTE: Only use the pool name and dataset. The rest is not
# supported by OpenZFS (whatever it's for).
if [ -z "$ZFS_RPOOL" ]
then
# The ${zfs-bootfs} variable is set at the kernel command
# line, usually by GRUB, but it cannot be referenced here
# directly because bourne variable names cannot contain a
# hyphen.
#
# Reassign the variable by dumping the environment and
# stripping the zfs-bootfs= prefix. Let the shell handle
# quoting through the eval command:
# shellcheck disable=SC2046
eval ZFS_RPOOL=$(set | sed -n -e 's,^zfs-bootfs=,,p')
fi
# ------------
# No root fs or pool specified - do auto detect.
if [ -z "$ZFS_RPOOL" ] && [ -z "${ZFS_BOOTFS}" ]
then
# Do auto detect. Do this by 'cheating' - set 'root=zfs:AUTO'
# which will be caught later
ROOT='zfs:AUTO'
fi
# ----------------------------------------------------------------
# F I N D A N D I M P O R T C O R R E C T P O O L
# ------------
if [ "$ROOT" = "zfs:AUTO" ]
then
# Try to detect both pool and root fs.
# If we got here, that means we don't have a hint so as to
# the root dataset, but with root=zfs:AUTO on cmdline,
# this says "zfs:AUTO" here and interferes with checks later
ZFS_BOOTFS=
[ "$quiet" != "y" ] && \
zfs_log_begin_msg "Attempting to import additional pools."
# Get a list of pools available for import
if [ -n "$ZFS_RPOOL" ]
then
# We've specified a pool - check only that
POOLS=$ZFS_RPOOL
else
POOLS=$(get_pools)
fi
OLD_IFS="$IFS" ; IFS=";"
for pool in $POOLS
do
[ -z "$pool" ] && continue
IFS="$OLD_IFS" import_pool "$pool"
IFS="$OLD_IFS" find_rootfs "$pool" && break
done
IFS="$OLD_IFS"
[ "$quiet" != "y" ] && zfs_log_end_msg "$ZFS_ERROR"
else
# No auto - use value from the command line option.
# Strip 'zfs:' and 'ZFS='.
ZFS_BOOTFS="${ROOT#*[:=]}"
# Strip everything after the first slash.
ZFS_RPOOL="${ZFS_BOOTFS%%/*}"
fi
# Import the pool (if not already done so in the AUTO check above).
if [ -n "$ZFS_RPOOL" ] && [ -z "${POOL_IMPORTED}" ]
then
[ "$quiet" != "y" ] && \
zfs_log_begin_msg "Importing ZFS root pool '$ZFS_RPOOL'"
import_pool "${ZFS_RPOOL}"
find_rootfs "${ZFS_RPOOL}"
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
if [ -z "${POOL_IMPORTED}" ]
then
# No pool imported, this is serious!
disable_plymouth
echo ""
echo "Command: $ZFS_CMD"
echo "Message: $ZFS_STDERR"
echo "Error: $ZFS_ERROR"
echo ""
echo "No pool imported. Manually import the root pool"
echo "at the command prompt and then exit."
echo "Hint: Try: zpool import -N ${ZFS_RPOOL}"
shell
fi
# In case the pool was specified as guid, resolve guid to name
pool="$("${ZPOOL}" get -H -o name,value name,guid | \
awk -v pool="${ZFS_RPOOL}" '$2 == pool { print $1 }')"
if [ -n "$pool" ]; then
# If $ZFS_BOOTFS contains guid, replace the guid portion with $pool
ZFS_BOOTFS=$(echo "$ZFS_BOOTFS" | \
sed -e "s/$("${ZPOOL}" get -H -o value guid "$pool")/$pool/g")
ZFS_RPOOL="${pool}"
fi
# ----------------------------------------------------------------
# P R E P A R E R O O T F I L E S Y S T E M
if [ -n "${ZFS_BOOTFS}" ]
then
# Booting from a snapshot?
# Will overwrite the ZFS_BOOTFS variable like so:
# rpool/ROOT/debian@snap2 => rpool/ROOT/debian_snap2
echo "${ZFS_BOOTFS}" | grep -q '@' && \
setup_snapshot_booting "${ZFS_BOOTFS}"
fi
if [ -z "${ZFS_BOOTFS}" ]
then
# Still nothing! Let the user sort this out.
disable_plymouth
echo ""
echo "Error: Unknown root filesystem - no 'bootfs' pool property and"
echo " not specified on the kernel command line."
echo ""
echo "Manually mount the root filesystem on $rootmnt and then exit."
echo "Hint: Try: mount.zfs -o zfsutil ${ZFS_RPOOL-rpool}/ROOT/system $rootmnt"
shell
fi
# ----------------------------------------------------------------
# M O U N T F I L E S Y S T E M S
# * Ideally, the root filesystem would be mounted like this:
#
# zpool import -R "$rootmnt" -N "$ZFS_RPOOL"
# zfs mount -o mountpoint=/ "${ZFS_BOOTFS}"
#
# but the MOUNTPOINT prefix is preserved on descendent filesystem
# after the pivot into the regular root, which later breaks things
# like `zfs mount -a` and the /proc/self/mounts refresh.
#
# * Mount additional filesystems required
# Such as /usr, /var, /usr/local etc.
# NOTE: Mounted in the order specified in the
# ZFS_INITRD_ADDITIONAL_DATASETS variable so take care!
# Go through the complete list (recursively) of all filesystems below
# the real root dataset
filesystems="$("${ZFS}" list -oname -tfilesystem -H -r "${ZFS_BOOTFS}")"
OLD_IFS="$IFS" ; IFS="
"
for fs in $filesystems; do
IFS="$OLD_IFS" mount_fs "$fs"
done
IFS="$OLD_IFS"
for fs in $ZFS_INITRD_ADDITIONAL_DATASETS; do
mount_fs "$fs"
done
touch /run/zfs_unlock_complete
if [ -e /run/zfs_unlock_complete_notify ]; then
read -r < /run/zfs_unlock_complete_notify
fi
# ------------
# Debugging information
if [ -n "${ZFS_DEBUG}" ]
then
#exec 2>&1-
echo "DEBUG: imported pools:"
"${ZPOOL}" list -H
echo
echo "DEBUG: mounted ZFS filesystems:"
mount | grep zfs
echo
echo "=> waiting for ENTER before continuing because of 'zfsdebug=1'. "
printf "%s" " 'c' for shell, 'r' for reboot, 'ENTER' to continue. "
read -r b
[ "$b" = "c" ] && /bin/sh
[ "$b" = "r" ] && reboot -f
set +x
fi
# ------------
# Run local bottom script
if command -v run_scripts > /dev/null 2>&1
then
if [ -f "/scripts/local-bottom" ] || [ -d "/scripts/local-bottom" ]
then
[ "$quiet" != "y" ] && \
zfs_log_begin_msg "Running /scripts/local-bottom"
run_scripts /scripts/local-bottom
[ "$quiet" != "y" ] && zfs_log_end_msg
fi
fi
}
diff --git a/sys/contrib/openzfs/contrib/pam_zfs_key/pam_zfs_key.c b/sys/contrib/openzfs/contrib/pam_zfs_key/pam_zfs_key.c
index 99cdb8d7733f..27c7d63781c5 100644
--- a/sys/contrib/openzfs/contrib/pam_zfs_key/pam_zfs_key.c
+++ b/sys/contrib/openzfs/contrib/pam_zfs_key/pam_zfs_key.c
@@ -1,853 +1,888 @@
/*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of the <organization> nor the
* names of its contributors may be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL <COPYRIGHT HOLDER> BE LIABLE FOR ANY
* DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Copyright (c) 2020, Felix Dörre
* All rights reserved.
*/
#include <sys/dsl_crypt.h>
#include <sys/byteorder.h>
#include <libzfs.h>
#include <syslog.h>
#include <sys/zio_crypt.h>
#include <openssl/evp.h>
#define PAM_SM_AUTH
#define PAM_SM_PASSWORD
#define PAM_SM_SESSION
#include <security/pam_modules.h>
#if defined(__linux__)
#include <security/pam_ext.h>
#define MAP_FLAGS MAP_PRIVATE | MAP_ANONYMOUS
#elif defined(__FreeBSD__)
#include <security/pam_appl.h>
static void
pam_syslog(pam_handle_t *pamh, int loglevel, const char *fmt, ...)
{
(void) pamh;
va_list args;
va_start(args, fmt);
vsyslog(loglevel, fmt, args);
va_end(args);
}
#define MAP_FLAGS MAP_PRIVATE | MAP_ANON | MAP_NOCORE
#endif
#include <string.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/stat.h>
#include <sys/file.h>
#include <sys/wait.h>
#include <pwd.h>
#include <sys/mman.h>
static const char PASSWORD_VAR_NAME[] = "pam_zfs_key_authtok";
static libzfs_handle_t *g_zfs;
static void destroy_pw(pam_handle_t *pamh, void *data, int errcode);
typedef int (*mlock_func_t) (const void *, size_t);
typedef struct {
size_t len;
char *value;
} pw_password_t;
/*
* Try to mlock(2) or munlock(2) addr while handling EAGAIN by retrying ten
* times and sleeping 10 milliseconds in between for a total of 0.1
* seconds. lock_func must point to either mlock(2) or munlock(2).
*/
static int
try_lock(mlock_func_t lock_func, const void *addr, size_t len)
{
int err;
int retries = 10;
useconds_t sleep_dur = 10 * 1000;
if ((err = (*lock_func)(addr, len)) != EAGAIN) {
return (err);
}
for (int i = retries; i > 0; --i) {
(void) usleep(sleep_dur);
if ((err = (*lock_func)(addr, len)) != EAGAIN) {
break;
}
}
return (err);
}
static pw_password_t *
alloc_pw_size(size_t len)
{
pw_password_t *pw = malloc(sizeof (pw_password_t));
if (!pw) {
return (NULL);
}
pw->len = len;
/*
* We use mmap(2) rather than malloc(3) since later on we mlock(2) the
* memory region. Since mlock(2) and munlock(2) operate on whole memory
* pages we should allocate a whole page here as mmap(2) does. Further
* this ensures that the addresses passed to mlock(2) an munlock(2) are
* on a page boundary as suggested by FreeBSD and required by some
* other implementations. Finally we avoid inadvertently munlocking
* memory mlocked by an concurrently running instance of us.
*/
pw->value = mmap(NULL, pw->len, PROT_READ | PROT_WRITE, MAP_FLAGS,
-1, 0);
if (pw->value == MAP_FAILED) {
free(pw);
return (NULL);
}
if (try_lock(mlock, pw->value, pw->len) != 0) {
(void) munmap(pw->value, pw->len);
free(pw);
return (NULL);
}
return (pw);
}
static pw_password_t *
alloc_pw_string(const char *source)
{
size_t len = strlen(source) + 1;
pw_password_t *pw = alloc_pw_size(len);
if (!pw) {
return (NULL);
}
memcpy(pw->value, source, pw->len);
return (pw);
}
static void
pw_free(pw_password_t *pw)
{
memset(pw->value, 0, pw->len);
if (try_lock(munlock, pw->value, pw->len) == 0) {
(void) munmap(pw->value, pw->len);
}
free(pw);
}
static pw_password_t *
pw_fetch(pam_handle_t *pamh)
{
const char *token;
if (pam_get_authtok(pamh, PAM_AUTHTOK, &token, NULL) != PAM_SUCCESS) {
pam_syslog(pamh, LOG_ERR,
"couldn't get password from PAM stack");
return (NULL);
}
if (!token) {
pam_syslog(pamh, LOG_ERR,
"token from PAM stack is null");
return (NULL);
}
return (alloc_pw_string(token));
}
static const pw_password_t *
pw_fetch_lazy(pam_handle_t *pamh)
{
pw_password_t *pw = pw_fetch(pamh);
if (pw == NULL) {
return (NULL);
}
int ret = pam_set_data(pamh, PASSWORD_VAR_NAME, pw, destroy_pw);
if (ret != PAM_SUCCESS) {
pw_free(pw);
pam_syslog(pamh, LOG_ERR, "pam_set_data failed");
return (NULL);
}
return (pw);
}
static const pw_password_t *
pw_get(pam_handle_t *pamh)
{
const pw_password_t *authtok = NULL;
int ret = pam_get_data(pamh, PASSWORD_VAR_NAME,
(const void**)(&authtok));
if (ret == PAM_SUCCESS)
return (authtok);
if (ret == PAM_NO_MODULE_DATA)
return (pw_fetch_lazy(pamh));
pam_syslog(pamh, LOG_ERR, "password not available");
return (NULL);
}
static int
pw_clear(pam_handle_t *pamh)
{
int ret = pam_set_data(pamh, PASSWORD_VAR_NAME, NULL, NULL);
if (ret != PAM_SUCCESS) {
pam_syslog(pamh, LOG_ERR, "clearing password failed");
return (-1);
}
return (0);
}
static void
destroy_pw(pam_handle_t *pamh, void *data, int errcode)
{
(void) pamh, (void) errcode;
if (data != NULL) {
pw_free((pw_password_t *)data);
}
}
static int
pam_zfs_init(pam_handle_t *pamh)
{
int error = 0;
if ((g_zfs = libzfs_init()) == NULL) {
error = errno;
pam_syslog(pamh, LOG_ERR, "Zfs initialization error: %s",
libzfs_error_init(error));
}
return (error);
}
static void
pam_zfs_free(void)
{
libzfs_fini(g_zfs);
}
static pw_password_t *
prepare_passphrase(pam_handle_t *pamh, zfs_handle_t *ds,
const char *passphrase, nvlist_t *nvlist)
{
pw_password_t *key = alloc_pw_size(WRAPPING_KEY_LEN);
if (!key) {
return (NULL);
}
uint64_t salt;
uint64_t iters;
if (nvlist != NULL) {
int fd = open("/dev/urandom", O_RDONLY);
if (fd < 0) {
pw_free(key);
return (NULL);
}
int bytes_read = 0;
char *buf = (char *)&salt;
size_t bytes = sizeof (uint64_t);
while (bytes_read < bytes) {
ssize_t len = read(fd, buf + bytes_read, bytes
- bytes_read);
if (len < 0) {
close(fd);
pw_free(key);
return (NULL);
}
bytes_read += len;
}
close(fd);
if (nvlist_add_uint64(nvlist,
zfs_prop_to_name(ZFS_PROP_PBKDF2_SALT), salt)) {
pam_syslog(pamh, LOG_ERR,
"failed to add salt to nvlist");
pw_free(key);
return (NULL);
}
iters = DEFAULT_PBKDF2_ITERATIONS;
if (nvlist_add_uint64(nvlist, zfs_prop_to_name(
ZFS_PROP_PBKDF2_ITERS), iters)) {
pam_syslog(pamh, LOG_ERR,
"failed to add iters to nvlist");
pw_free(key);
return (NULL);
}
} else {
salt = zfs_prop_get_int(ds, ZFS_PROP_PBKDF2_SALT);
iters = zfs_prop_get_int(ds, ZFS_PROP_PBKDF2_ITERS);
}
salt = LE_64(salt);
if (!PKCS5_PBKDF2_HMAC_SHA1((char *)passphrase,
strlen(passphrase), (uint8_t *)&salt,
sizeof (uint64_t), iters, WRAPPING_KEY_LEN,
(uint8_t *)key->value)) {
pam_syslog(pamh, LOG_ERR, "pbkdf failed");
pw_free(key);
return (NULL);
}
return (key);
}
static int
is_key_loaded(pam_handle_t *pamh, const char *ds_name)
{
zfs_handle_t *ds = zfs_open(g_zfs, ds_name, ZFS_TYPE_FILESYSTEM);
if (ds == NULL) {
pam_syslog(pamh, LOG_ERR, "dataset %s not found", ds_name);
return (-1);
}
int keystatus = zfs_prop_get_int(ds, ZFS_PROP_KEYSTATUS);
zfs_close(ds);
return (keystatus != ZFS_KEYSTATUS_UNAVAILABLE);
}
static int
change_key(pam_handle_t *pamh, const char *ds_name,
const char *passphrase)
{
zfs_handle_t *ds = zfs_open(g_zfs, ds_name, ZFS_TYPE_FILESYSTEM);
if (ds == NULL) {
pam_syslog(pamh, LOG_ERR, "dataset %s not found", ds_name);
return (-1);
}
nvlist_t *nvlist = fnvlist_alloc();
pw_password_t *key = prepare_passphrase(pamh, ds, passphrase, nvlist);
if (key == NULL) {
nvlist_free(nvlist);
zfs_close(ds);
return (-1);
}
if (nvlist_add_string(nvlist,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION),
"prompt")) {
pam_syslog(pamh, LOG_ERR, "nvlist_add failed for keylocation");
pw_free(key);
nvlist_free(nvlist);
zfs_close(ds);
return (-1);
}
if (nvlist_add_uint64(nvlist,
zfs_prop_to_name(ZFS_PROP_KEYFORMAT),
ZFS_KEYFORMAT_PASSPHRASE)) {
pam_syslog(pamh, LOG_ERR, "nvlist_add failed for keyformat");
pw_free(key);
nvlist_free(nvlist);
zfs_close(ds);
return (-1);
}
int ret = lzc_change_key(ds_name, DCP_CMD_NEW_KEY, nvlist,
(uint8_t *)key->value, WRAPPING_KEY_LEN);
pw_free(key);
if (ret) {
pam_syslog(pamh, LOG_ERR, "change_key failed: %d", ret);
nvlist_free(nvlist);
zfs_close(ds);
return (-1);
}
nvlist_free(nvlist);
zfs_close(ds);
return (0);
}
static int
decrypt_mount(pam_handle_t *pamh, const char *ds_name,
- const char *passphrase)
+ const char *passphrase, boolean_t noop)
{
zfs_handle_t *ds = zfs_open(g_zfs, ds_name, ZFS_TYPE_FILESYSTEM);
if (ds == NULL) {
pam_syslog(pamh, LOG_ERR, "dataset %s not found", ds_name);
return (-1);
}
pw_password_t *key = prepare_passphrase(pamh, ds, passphrase, NULL);
if (key == NULL) {
zfs_close(ds);
return (-1);
}
- int ret = lzc_load_key(ds_name, B_FALSE, (uint8_t *)key->value,
+ int ret = lzc_load_key(ds_name, noop, (uint8_t *)key->value,
WRAPPING_KEY_LEN);
pw_free(key);
if (ret) {
pam_syslog(pamh, LOG_ERR, "load_key failed: %d", ret);
zfs_close(ds);
return (-1);
}
+ if (noop) {
+ goto out;
+ }
ret = zfs_mount(ds, NULL, 0);
if (ret) {
pam_syslog(pamh, LOG_ERR, "mount failed: %d", ret);
zfs_close(ds);
return (-1);
}
+out:
zfs_close(ds);
return (0);
}
static int
unmount_unload(pam_handle_t *pamh, const char *ds_name)
{
zfs_handle_t *ds = zfs_open(g_zfs, ds_name, ZFS_TYPE_FILESYSTEM);
if (ds == NULL) {
pam_syslog(pamh, LOG_ERR, "dataset %s not found", ds_name);
return (-1);
}
int ret = zfs_unmount(ds, NULL, 0);
if (ret) {
pam_syslog(pamh, LOG_ERR, "zfs_unmount failed with: %d", ret);
zfs_close(ds);
return (-1);
}
ret = lzc_unload_key(ds_name);
if (ret) {
pam_syslog(pamh, LOG_ERR, "unload_key failed with: %d", ret);
zfs_close(ds);
return (-1);
}
zfs_close(ds);
return (0);
}
typedef struct {
char *homes_prefix;
char *runstatedir;
char *homedir;
char *dsname;
uid_t uid;
const char *username;
int unmount_and_unload;
} zfs_key_config_t;
static int
zfs_key_config_load(pam_handle_t *pamh, zfs_key_config_t *config,
int argc, const char **argv)
{
config->homes_prefix = strdup("rpool/home");
if (config->homes_prefix == NULL) {
pam_syslog(pamh, LOG_ERR, "strdup failure");
- return (-1);
+ return (PAM_SERVICE_ERR);
}
config->runstatedir = strdup(RUNSTATEDIR "/pam_zfs_key");
if (config->runstatedir == NULL) {
pam_syslog(pamh, LOG_ERR, "strdup failure");
free(config->homes_prefix);
- return (-1);
+ return (PAM_SERVICE_ERR);
}
const char *name;
if (pam_get_user(pamh, &name, NULL) != PAM_SUCCESS) {
pam_syslog(pamh, LOG_ERR,
"couldn't get username from PAM stack");
free(config->runstatedir);
free(config->homes_prefix);
- return (-1);
+ return (PAM_SERVICE_ERR);
}
struct passwd *entry = getpwnam(name);
if (!entry) {
free(config->runstatedir);
free(config->homes_prefix);
- return (-1);
+ return (PAM_USER_UNKNOWN);
}
config->uid = entry->pw_uid;
config->username = name;
config->unmount_and_unload = 1;
config->dsname = NULL;
config->homedir = NULL;
for (int c = 0; c < argc; c++) {
if (strncmp(argv[c], "homes=", 6) == 0) {
free(config->homes_prefix);
config->homes_prefix = strdup(argv[c] + 6);
} else if (strncmp(argv[c], "runstatedir=", 12) == 0) {
free(config->runstatedir);
config->runstatedir = strdup(argv[c] + 12);
} else if (strcmp(argv[c], "nounmount") == 0) {
config->unmount_and_unload = 0;
} else if (strcmp(argv[c], "prop_mountpoint") == 0) {
if (config->homedir == NULL)
config->homedir = strdup(entry->pw_dir);
}
}
- return (0);
+ return (PAM_SUCCESS);
}
static void
zfs_key_config_free(zfs_key_config_t *config)
{
free(config->homes_prefix);
free(config->runstatedir);
free(config->homedir);
free(config->dsname);
}
static int
find_dsname_by_prop_value(zfs_handle_t *zhp, void *data)
{
zfs_type_t type = zfs_get_type(zhp);
zfs_key_config_t *target = data;
char mountpoint[ZFS_MAXPROPLEN];
/* Skip any datasets whose type does not match */
if ((type & ZFS_TYPE_FILESYSTEM) == 0) {
zfs_close(zhp);
return (0);
}
/* Skip any datasets whose mountpoint does not match */
(void) zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, mountpoint,
sizeof (mountpoint), NULL, NULL, 0, B_FALSE);
if (strcmp(target->homedir, mountpoint) != 0) {
zfs_close(zhp);
return (0);
}
target->dsname = strdup(zfs_get_name(zhp));
zfs_close(zhp);
return (1);
}
static char *
zfs_key_config_get_dataset(zfs_key_config_t *config)
{
if (config->homedir != NULL &&
config->homes_prefix != NULL) {
zfs_handle_t *zhp = zfs_open(g_zfs, config->homes_prefix,
ZFS_TYPE_FILESYSTEM);
if (zhp == NULL) {
pam_syslog(NULL, LOG_ERR, "dataset %s not found",
config->homes_prefix);
return (NULL);
}
- (void) zfs_iter_filesystems(zhp, 0, find_dsname_by_prop_value,
- config);
+ (void) zfs_iter_filesystems_v2(zhp, 0,
+ find_dsname_by_prop_value, config);
zfs_close(zhp);
char *dsname = config->dsname;
config->dsname = NULL;
return (dsname);
}
if (config->homes_prefix == NULL) {
return (NULL);
}
size_t len = ZFS_MAX_DATASET_NAME_LEN;
size_t total_len = strlen(config->homes_prefix) + 1
+ strlen(config->username);
if (total_len > len) {
return (NULL);
}
char *ret = malloc(len + 1);
if (!ret) {
return (NULL);
}
ret[0] = 0;
(void) snprintf(ret, len + 1, "%s/%s", config->homes_prefix,
config->username);
return (ret);
}
static int
zfs_key_config_modify_session_counter(pam_handle_t *pamh,
zfs_key_config_t *config, int delta)
{
const char *runtime_path = config->runstatedir;
if (mkdir(runtime_path, S_IRWXU) != 0 && errno != EEXIST) {
pam_syslog(pamh, LOG_ERR, "Can't create runtime path: %d",
errno);
return (-1);
}
if (chown(runtime_path, 0, 0) != 0) {
pam_syslog(pamh, LOG_ERR, "Can't chown runtime path: %d",
errno);
return (-1);
}
if (chmod(runtime_path, S_IRWXU) != 0) {
pam_syslog(pamh, LOG_ERR, "Can't chmod runtime path: %d",
errno);
return (-1);
}
size_t runtime_path_len = strlen(runtime_path);
size_t counter_path_len = runtime_path_len + 1 + 10;
char *counter_path = malloc(counter_path_len + 1);
if (!counter_path) {
return (-1);
}
counter_path[0] = 0;
strcat(counter_path, runtime_path);
snprintf(counter_path + runtime_path_len, counter_path_len, "/%d",
config->uid);
const int fd = open(counter_path,
O_RDWR | O_CLOEXEC | O_CREAT | O_NOFOLLOW,
S_IRUSR | S_IWUSR);
free(counter_path);
if (fd < 0) {
pam_syslog(pamh, LOG_ERR, "Can't open counter file: %d", errno);
return (-1);
}
if (flock(fd, LOCK_EX) != 0) {
pam_syslog(pamh, LOG_ERR, "Can't lock counter file: %d", errno);
close(fd);
return (-1);
}
char counter[20];
char *pos = counter;
int remaining = sizeof (counter) - 1;
int ret;
counter[sizeof (counter) - 1] = 0;
while (remaining > 0 && (ret = read(fd, pos, remaining)) > 0) {
remaining -= ret;
pos += ret;
}
*pos = 0;
long int counter_value = strtol(counter, NULL, 10);
counter_value += delta;
if (counter_value < 0) {
counter_value = 0;
}
lseek(fd, 0, SEEK_SET);
if (ftruncate(fd, 0) != 0) {
pam_syslog(pamh, LOG_ERR, "Can't truncate counter file: %d",
errno);
close(fd);
return (-1);
}
snprintf(counter, sizeof (counter), "%ld", counter_value);
remaining = strlen(counter);
pos = counter;
while (remaining > 0 && (ret = write(fd, pos, remaining)) > 0) {
remaining -= ret;
pos += ret;
}
close(fd);
return (counter_value);
}
__attribute__((visibility("default")))
PAM_EXTERN int
pam_sm_authenticate(pam_handle_t *pamh, int flags,
int argc, const char **argv)
{
- (void) flags, (void) argc, (void) argv;
+ (void) flags;
- if (pw_fetch_lazy(pamh) == NULL) {
- return (PAM_AUTH_ERR);
+ if (geteuid() != 0) {
+ pam_syslog(pamh, LOG_ERR,
+ "Cannot zfs_mount when not being root.");
+ return (PAM_SERVICE_ERR);
+ }
+ zfs_key_config_t config;
+ int config_err = zfs_key_config_load(pamh, &config, argc, argv);
+ if (config_err != PAM_SUCCESS) {
+ return (config_err);
}
+ const pw_password_t *token = pw_fetch_lazy(pamh);
+ if (token == NULL) {
+ zfs_key_config_free(&config);
+ return (PAM_AUTH_ERR);
+ }
+ if (pam_zfs_init(pamh) != 0) {
+ zfs_key_config_free(&config);
+ return (PAM_SERVICE_ERR);
+ }
+ char *dataset = zfs_key_config_get_dataset(&config);
+ if (!dataset) {
+ pam_zfs_free();
+ zfs_key_config_free(&config);
+ return (PAM_SERVICE_ERR);
+ }
+ if (decrypt_mount(pamh, dataset, token->value, B_TRUE) == -1) {
+ free(dataset);
+ pam_zfs_free();
+ zfs_key_config_free(&config);
+ return (PAM_AUTH_ERR);
+ }
+ free(dataset);
+ pam_zfs_free();
+ zfs_key_config_free(&config);
return (PAM_SUCCESS);
}
__attribute__((visibility("default")))
PAM_EXTERN int
pam_sm_setcred(pam_handle_t *pamh, int flags,
int argc, const char **argv)
{
(void) pamh, (void) flags, (void) argc, (void) argv;
return (PAM_SUCCESS);
}
__attribute__((visibility("default")))
PAM_EXTERN int
pam_sm_chauthtok(pam_handle_t *pamh, int flags,
int argc, const char **argv)
{
if (geteuid() != 0) {
pam_syslog(pamh, LOG_ERR,
"Cannot zfs_mount when not being root.");
return (PAM_PERM_DENIED);
}
zfs_key_config_t config;
- if (zfs_key_config_load(pamh, &config, argc, argv) == -1) {
+ if (zfs_key_config_load(pamh, &config, argc, argv) != PAM_SUCCESS) {
return (PAM_SERVICE_ERR);
}
if (config.uid < 1000) {
zfs_key_config_free(&config);
return (PAM_SUCCESS);
}
{
if (pam_zfs_init(pamh) != 0) {
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
char *dataset = zfs_key_config_get_dataset(&config);
if (!dataset) {
pam_zfs_free();
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
int key_loaded = is_key_loaded(pamh, dataset);
if (key_loaded == -1) {
free(dataset);
pam_zfs_free();
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
free(dataset);
pam_zfs_free();
if (! key_loaded) {
pam_syslog(pamh, LOG_ERR,
"key not loaded, returning try_again");
zfs_key_config_free(&config);
return (PAM_PERM_DENIED);
}
}
if ((flags & PAM_UPDATE_AUTHTOK) != 0) {
const pw_password_t *token = pw_get(pamh);
if (token == NULL) {
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
if (pam_zfs_init(pamh) != 0) {
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
char *dataset = zfs_key_config_get_dataset(&config);
if (!dataset) {
pam_zfs_free();
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
if (change_key(pamh, dataset, token->value) == -1) {
free(dataset);
pam_zfs_free();
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
free(dataset);
pam_zfs_free();
zfs_key_config_free(&config);
if (pw_clear(pamh) == -1) {
return (PAM_SERVICE_ERR);
}
} else {
zfs_key_config_free(&config);
}
return (PAM_SUCCESS);
}
PAM_EXTERN int
pam_sm_open_session(pam_handle_t *pamh, int flags,
int argc, const char **argv)
{
(void) flags;
if (geteuid() != 0) {
pam_syslog(pamh, LOG_ERR,
"Cannot zfs_mount when not being root.");
return (PAM_SUCCESS);
}
zfs_key_config_t config;
- if (zfs_key_config_load(pamh, &config, argc, argv) != 0) {
+ if (zfs_key_config_load(pamh, &config, argc, argv) != PAM_SUCCESS) {
return (PAM_SESSION_ERR);
}
if (config.uid < 1000) {
zfs_key_config_free(&config);
return (PAM_SUCCESS);
}
int counter = zfs_key_config_modify_session_counter(pamh, &config, 1);
if (counter != 1) {
zfs_key_config_free(&config);
return (PAM_SUCCESS);
}
const pw_password_t *token = pw_get(pamh);
if (token == NULL) {
zfs_key_config_free(&config);
return (PAM_SESSION_ERR);
}
if (pam_zfs_init(pamh) != 0) {
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
char *dataset = zfs_key_config_get_dataset(&config);
if (!dataset) {
pam_zfs_free();
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
- if (decrypt_mount(pamh, dataset, token->value) == -1) {
+ if (decrypt_mount(pamh, dataset, token->value, B_FALSE) == -1) {
free(dataset);
pam_zfs_free();
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
free(dataset);
pam_zfs_free();
zfs_key_config_free(&config);
if (pw_clear(pamh) == -1) {
return (PAM_SERVICE_ERR);
}
return (PAM_SUCCESS);
}
__attribute__((visibility("default")))
PAM_EXTERN int
pam_sm_close_session(pam_handle_t *pamh, int flags,
int argc, const char **argv)
{
(void) flags;
if (geteuid() != 0) {
pam_syslog(pamh, LOG_ERR,
"Cannot zfs_mount when not being root.");
return (PAM_SUCCESS);
}
zfs_key_config_t config;
- if (zfs_key_config_load(pamh, &config, argc, argv) != 0) {
+ if (zfs_key_config_load(pamh, &config, argc, argv) != PAM_SUCCESS) {
return (PAM_SESSION_ERR);
}
if (config.uid < 1000) {
zfs_key_config_free(&config);
return (PAM_SUCCESS);
}
int counter = zfs_key_config_modify_session_counter(pamh, &config, -1);
if (counter != 0) {
zfs_key_config_free(&config);
return (PAM_SUCCESS);
}
if (config.unmount_and_unload) {
if (pam_zfs_init(pamh) != 0) {
zfs_key_config_free(&config);
return (PAM_SERVICE_ERR);
}
char *dataset = zfs_key_config_get_dataset(&config);
if (!dataset) {
pam_zfs_free();
zfs_key_config_free(&config);
return (PAM_SESSION_ERR);
}
if (unmount_unload(pamh, dataset) == -1) {
free(dataset);
pam_zfs_free();
zfs_key_config_free(&config);
return (PAM_SESSION_ERR);
}
free(dataset);
pam_zfs_free();
}
zfs_key_config_free(&config);
return (PAM_SUCCESS);
}
diff --git a/sys/contrib/openzfs/include/libzfs.h b/sys/contrib/openzfs/include/libzfs.h
index a265f2a19878..87d1ed738f2b 100644
--- a/sys/contrib/openzfs/include/libzfs.h
+++ b/sys/contrib/openzfs/include/libzfs.h
@@ -1,1016 +1,1028 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2022 by Delphix. All rights reserved.
* Copyright Joyent, Inc.
* Copyright (c) 2013 Steven Hartland. All rights reserved.
* Copyright (c) 2016, Intel Corporation.
* Copyright 2016 Nexenta Systems, Inc.
* Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
* Copyright (c) 2019 Datto Inc.
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
*/
#ifndef _LIBZFS_H
#define _LIBZFS_H extern __attribute__((visibility("default")))
#include <assert.h>
#include <libshare.h>
#include <libnvpair.h>
#include <sys/mnttab.h>
#include <sys/param.h>
#include <sys/types.h>
#include <sys/fs/zfs.h>
#include <sys/avl.h>
#include <libzfs_core.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* Miscellaneous ZFS constants
*/
#define ZFS_MAXPROPLEN MAXPATHLEN
#define ZPOOL_MAXPROPLEN MAXPATHLEN
/*
* libzfs errors
*/
typedef enum zfs_error {
EZFS_SUCCESS = 0, /* no error -- success */
EZFS_NOMEM = 2000, /* out of memory */
EZFS_BADPROP, /* invalid property value */
EZFS_PROPREADONLY, /* cannot set readonly property */
EZFS_PROPTYPE, /* property does not apply to dataset type */
EZFS_PROPNONINHERIT, /* property is not inheritable */
EZFS_PROPSPACE, /* bad quota or reservation */
EZFS_BADTYPE, /* dataset is not of appropriate type */
EZFS_BUSY, /* pool or dataset is busy */
EZFS_EXISTS, /* pool or dataset already exists */
EZFS_NOENT, /* no such pool or dataset */
EZFS_BADSTREAM, /* bad backup stream */
EZFS_DSREADONLY, /* dataset is readonly */
EZFS_VOLTOOBIG, /* volume is too large for 32-bit system */
EZFS_INVALIDNAME, /* invalid dataset name */
EZFS_BADRESTORE, /* unable to restore to destination */
EZFS_BADBACKUP, /* backup failed */
EZFS_BADTARGET, /* bad attach/detach/replace target */
EZFS_NODEVICE, /* no such device in pool */
EZFS_BADDEV, /* invalid device to add */
EZFS_NOREPLICAS, /* no valid replicas */
EZFS_RESILVERING, /* resilvering (healing reconstruction) */
EZFS_BADVERSION, /* unsupported version */
EZFS_POOLUNAVAIL, /* pool is currently unavailable */
EZFS_DEVOVERFLOW, /* too many devices in one vdev */
EZFS_BADPATH, /* must be an absolute path */
EZFS_CROSSTARGET, /* rename or clone across pool or dataset */
EZFS_ZONED, /* used improperly in local zone */
EZFS_MOUNTFAILED, /* failed to mount dataset */
EZFS_UMOUNTFAILED, /* failed to unmount dataset */
EZFS_UNSHARENFSFAILED, /* failed to unshare over nfs */
EZFS_SHARENFSFAILED, /* failed to share over nfs */
EZFS_PERM, /* permission denied */
EZFS_NOSPC, /* out of space */
EZFS_FAULT, /* bad address */
EZFS_IO, /* I/O error */
EZFS_INTR, /* signal received */
EZFS_ISSPARE, /* device is a hot spare */
EZFS_INVALCONFIG, /* invalid vdev configuration */
EZFS_RECURSIVE, /* recursive dependency */
EZFS_NOHISTORY, /* no history object */
EZFS_POOLPROPS, /* couldn't retrieve pool props */
EZFS_POOL_NOTSUP, /* ops not supported for this type of pool */
EZFS_POOL_INVALARG, /* invalid argument for this pool operation */
EZFS_NAMETOOLONG, /* dataset name is too long */
EZFS_OPENFAILED, /* open of device failed */
EZFS_NOCAP, /* couldn't get capacity */
EZFS_LABELFAILED, /* write of label failed */
EZFS_BADWHO, /* invalid permission who */
EZFS_BADPERM, /* invalid permission */
EZFS_BADPERMSET, /* invalid permission set name */
EZFS_NODELEGATION, /* delegated administration is disabled */
EZFS_UNSHARESMBFAILED, /* failed to unshare over smb */
EZFS_SHARESMBFAILED, /* failed to share over smb */
EZFS_BADCACHE, /* bad cache file */
EZFS_ISL2CACHE, /* device is for the level 2 ARC */
EZFS_VDEVNOTSUP, /* unsupported vdev type */
EZFS_NOTSUP, /* ops not supported on this dataset */
EZFS_ACTIVE_SPARE, /* pool has active shared spare devices */
EZFS_UNPLAYED_LOGS, /* log device has unplayed logs */
EZFS_REFTAG_RELE, /* snapshot release: tag not found */
EZFS_REFTAG_HOLD, /* snapshot hold: tag already exists */
EZFS_TAGTOOLONG, /* snapshot hold/rele: tag too long */
EZFS_PIPEFAILED, /* pipe create failed */
EZFS_THREADCREATEFAILED, /* thread create failed */
EZFS_POSTSPLIT_ONLINE, /* onlining a disk after splitting it */
EZFS_SCRUBBING, /* currently scrubbing */
EZFS_NO_SCRUB, /* no active scrub */
EZFS_DIFF, /* general failure of zfs diff */
EZFS_DIFFDATA, /* bad zfs diff data */
EZFS_POOLREADONLY, /* pool is in read-only mode */
EZFS_SCRUB_PAUSED, /* scrub currently paused */
EZFS_ACTIVE_POOL, /* pool is imported on a different system */
EZFS_CRYPTOFAILED, /* failed to setup encryption */
EZFS_NO_PENDING, /* cannot cancel, no operation is pending */
EZFS_CHECKPOINT_EXISTS, /* checkpoint exists */
EZFS_DISCARDING_CHECKPOINT, /* currently discarding a checkpoint */
EZFS_NO_CHECKPOINT, /* pool has no checkpoint */
EZFS_DEVRM_IN_PROGRESS, /* a device is currently being removed */
EZFS_VDEV_TOO_BIG, /* a device is too big to be used */
EZFS_IOC_NOTSUPPORTED, /* operation not supported by zfs module */
EZFS_TOOMANY, /* argument list too long */
EZFS_INITIALIZING, /* currently initializing */
EZFS_NO_INITIALIZE, /* no active initialize */
EZFS_WRONG_PARENT, /* invalid parent dataset (e.g ZVOL) */
EZFS_TRIMMING, /* currently trimming */
EZFS_NO_TRIM, /* no active trim */
EZFS_TRIM_NOTSUP, /* device does not support trim */
EZFS_NO_RESILVER_DEFER, /* pool doesn't support resilver_defer */
EZFS_EXPORT_IN_PROGRESS, /* currently exporting the pool */
EZFS_REBUILDING, /* resilvering (sequential reconstrution) */
EZFS_VDEV_NOTSUP, /* ops not supported for this type of vdev */
EZFS_NOT_USER_NAMESPACE, /* a file is not a user namespace */
EZFS_CKSUM, /* insufficient replicas */
EZFS_RESUME_EXISTS, /* Resume on existing dataset without force */
EZFS_UNKNOWN
} zfs_error_t;
/*
* The following data structures are all part
* of the zfs_allow_t data structure which is
* used for printing 'allow' permissions.
* It is a linked list of zfs_allow_t's which
* then contain avl tree's for user/group/sets/...
* and each one of the entries in those trees have
* avl tree's for the permissions they belong to and
* whether they are local,descendent or local+descendent
* permissions. The AVL trees are used primarily for
* sorting purposes, but also so that we can quickly find
* a given user and or permission.
*/
typedef struct zfs_perm_node {
avl_node_t z_node;
char z_pname[MAXPATHLEN];
} zfs_perm_node_t;
typedef struct zfs_allow_node {
avl_node_t z_node;
char z_key[MAXPATHLEN]; /* name, such as joe */
avl_tree_t z_localdescend; /* local+descendent perms */
avl_tree_t z_local; /* local permissions */
avl_tree_t z_descend; /* descendent permissions */
} zfs_allow_node_t;
typedef struct zfs_allow {
struct zfs_allow *z_next;
char z_setpoint[MAXPATHLEN];
avl_tree_t z_sets;
avl_tree_t z_crperms;
avl_tree_t z_user;
avl_tree_t z_group;
avl_tree_t z_everyone;
} zfs_allow_t;
/*
* Basic handle types
*/
typedef struct zfs_handle zfs_handle_t;
typedef struct zpool_handle zpool_handle_t;
typedef struct libzfs_handle libzfs_handle_t;
_LIBZFS_H int zpool_wait(zpool_handle_t *, zpool_wait_activity_t);
_LIBZFS_H int zpool_wait_status(zpool_handle_t *, zpool_wait_activity_t,
boolean_t *, boolean_t *);
/*
* Library initialization
*/
_LIBZFS_H libzfs_handle_t *libzfs_init(void);
_LIBZFS_H void libzfs_fini(libzfs_handle_t *);
_LIBZFS_H libzfs_handle_t *zpool_get_handle(zpool_handle_t *);
_LIBZFS_H libzfs_handle_t *zfs_get_handle(zfs_handle_t *);
_LIBZFS_H void libzfs_print_on_error(libzfs_handle_t *, boolean_t);
_LIBZFS_H void zfs_save_arguments(int argc, char **, char *, int);
_LIBZFS_H int zpool_log_history(libzfs_handle_t *, const char *);
_LIBZFS_H int libzfs_errno(libzfs_handle_t *);
_LIBZFS_H const char *libzfs_error_init(int);
_LIBZFS_H const char *libzfs_error_action(libzfs_handle_t *);
_LIBZFS_H const char *libzfs_error_description(libzfs_handle_t *);
_LIBZFS_H int zfs_standard_error(libzfs_handle_t *, int, const char *);
_LIBZFS_H void libzfs_mnttab_init(libzfs_handle_t *);
_LIBZFS_H void libzfs_mnttab_fini(libzfs_handle_t *);
_LIBZFS_H void libzfs_mnttab_cache(libzfs_handle_t *, boolean_t);
_LIBZFS_H int libzfs_mnttab_find(libzfs_handle_t *, const char *,
struct mnttab *);
_LIBZFS_H void libzfs_mnttab_add(libzfs_handle_t *, const char *,
const char *, const char *);
_LIBZFS_H void libzfs_mnttab_remove(libzfs_handle_t *, const char *);
/*
* Basic handle functions
*/
_LIBZFS_H zpool_handle_t *zpool_open(libzfs_handle_t *, const char *);
_LIBZFS_H zpool_handle_t *zpool_open_canfail(libzfs_handle_t *, const char *);
_LIBZFS_H void zpool_close(zpool_handle_t *);
_LIBZFS_H const char *zpool_get_name(zpool_handle_t *);
_LIBZFS_H int zpool_get_state(zpool_handle_t *);
_LIBZFS_H const char *zpool_state_to_name(vdev_state_t, vdev_aux_t);
_LIBZFS_H const char *zpool_pool_state_to_name(pool_state_t);
_LIBZFS_H void zpool_free_handles(libzfs_handle_t *);
/*
* Iterate over all active pools in the system.
*/
typedef int (*zpool_iter_f)(zpool_handle_t *, void *);
_LIBZFS_H int zpool_iter(libzfs_handle_t *, zpool_iter_f, void *);
_LIBZFS_H boolean_t zpool_skip_pool(const char *);
/*
* Functions to create and destroy pools
*/
_LIBZFS_H int zpool_create(libzfs_handle_t *, const char *, nvlist_t *,
nvlist_t *, nvlist_t *);
_LIBZFS_H int zpool_destroy(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_add(zpool_handle_t *, nvlist_t *);
typedef struct splitflags {
/* do not split, but return the config that would be split off */
unsigned int dryrun : 1;
/* after splitting, import the pool */
unsigned int import : 1;
int name_flags;
} splitflags_t;
typedef struct trimflags {
/* requested vdevs are for the entire pool */
boolean_t fullpool;
/* request a secure trim, requires support from device */
boolean_t secure;
/* after starting trim, block until trim completes */
boolean_t wait;
/* trim at the requested rate in bytes/second */
uint64_t rate;
} trimflags_t;
/*
* Functions to manipulate pool and vdev state
*/
_LIBZFS_H int zpool_scan(zpool_handle_t *, pool_scan_func_t, pool_scrub_cmd_t);
_LIBZFS_H int zpool_initialize(zpool_handle_t *, pool_initialize_func_t,
nvlist_t *);
_LIBZFS_H int zpool_initialize_wait(zpool_handle_t *, pool_initialize_func_t,
nvlist_t *);
_LIBZFS_H int zpool_trim(zpool_handle_t *, pool_trim_func_t, nvlist_t *,
trimflags_t *);
_LIBZFS_H int zpool_clear(zpool_handle_t *, const char *, nvlist_t *);
_LIBZFS_H int zpool_reguid(zpool_handle_t *);
_LIBZFS_H int zpool_reopen_one(zpool_handle_t *, void *);
_LIBZFS_H int zpool_sync_one(zpool_handle_t *, void *);
_LIBZFS_H int zpool_vdev_online(zpool_handle_t *, const char *, int,
vdev_state_t *);
_LIBZFS_H int zpool_vdev_offline(zpool_handle_t *, const char *, boolean_t);
_LIBZFS_H int zpool_vdev_attach(zpool_handle_t *, const char *,
const char *, nvlist_t *, int, boolean_t);
_LIBZFS_H int zpool_vdev_detach(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_vdev_remove(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_vdev_remove_cancel(zpool_handle_t *);
_LIBZFS_H int zpool_vdev_indirect_size(zpool_handle_t *, const char *,
uint64_t *);
_LIBZFS_H int zpool_vdev_split(zpool_handle_t *, char *, nvlist_t **,
nvlist_t *, splitflags_t);
_LIBZFS_H int zpool_vdev_remove_wanted(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_vdev_fault(zpool_handle_t *, uint64_t, vdev_aux_t);
_LIBZFS_H int zpool_vdev_degrade(zpool_handle_t *, uint64_t, vdev_aux_t);
_LIBZFS_H int zpool_vdev_clear(zpool_handle_t *, uint64_t);
_LIBZFS_H nvlist_t *zpool_find_vdev(zpool_handle_t *, const char *, boolean_t *,
boolean_t *, boolean_t *);
_LIBZFS_H nvlist_t *zpool_find_vdev_by_physpath(zpool_handle_t *, const char *,
boolean_t *, boolean_t *, boolean_t *);
_LIBZFS_H int zpool_label_disk(libzfs_handle_t *, zpool_handle_t *,
const char *);
_LIBZFS_H uint64_t zpool_vdev_path_to_guid(zpool_handle_t *zhp,
const char *path);
_LIBZFS_H const char *zpool_get_state_str(zpool_handle_t *);
/*
* Functions to manage pool properties
*/
_LIBZFS_H int zpool_set_prop(zpool_handle_t *, const char *, const char *);
_LIBZFS_H int zpool_get_prop(zpool_handle_t *, zpool_prop_t, char *,
size_t proplen, zprop_source_t *, boolean_t literal);
_LIBZFS_H int zpool_get_userprop(zpool_handle_t *, const char *, char *,
size_t proplen, zprop_source_t *);
_LIBZFS_H uint64_t zpool_get_prop_int(zpool_handle_t *, zpool_prop_t,
zprop_source_t *);
_LIBZFS_H int zpool_props_refresh(zpool_handle_t *);
_LIBZFS_H const char *zpool_prop_to_name(zpool_prop_t);
_LIBZFS_H const char *zpool_prop_values(zpool_prop_t);
/*
* Functions to manage vdev properties
*/
_LIBZFS_H int zpool_get_vdev_prop_value(nvlist_t *, vdev_prop_t, char *, char *,
size_t, zprop_source_t *, boolean_t);
_LIBZFS_H int zpool_get_vdev_prop(zpool_handle_t *, const char *, vdev_prop_t,
char *, char *, size_t, zprop_source_t *, boolean_t);
_LIBZFS_H int zpool_get_all_vdev_props(zpool_handle_t *, const char *,
nvlist_t **);
_LIBZFS_H int zpool_set_vdev_prop(zpool_handle_t *, const char *, const char *,
const char *);
_LIBZFS_H const char *vdev_prop_to_name(vdev_prop_t);
_LIBZFS_H const char *vdev_prop_values(vdev_prop_t);
_LIBZFS_H boolean_t vdev_prop_user(const char *name);
_LIBZFS_H const char *vdev_prop_column_name(vdev_prop_t);
_LIBZFS_H boolean_t vdev_prop_align_right(vdev_prop_t);
/*
* Pool health statistics.
*/
typedef enum {
/*
* The following correspond to faults as defined in the (fault.fs.zfs.*)
* event namespace. Each is associated with a corresponding message ID.
* This must be kept in sync with the zfs_msgid_table in
* lib/libzfs/libzfs_status.c.
*/
ZPOOL_STATUS_CORRUPT_CACHE, /* corrupt /kernel/drv/zpool.cache */
ZPOOL_STATUS_MISSING_DEV_R, /* missing device with replicas */
ZPOOL_STATUS_MISSING_DEV_NR, /* missing device with no replicas */
ZPOOL_STATUS_CORRUPT_LABEL_R, /* bad device label with replicas */
ZPOOL_STATUS_CORRUPT_LABEL_NR, /* bad device label with no replicas */
ZPOOL_STATUS_BAD_GUID_SUM, /* sum of device guids didn't match */
ZPOOL_STATUS_CORRUPT_POOL, /* pool metadata is corrupted */
ZPOOL_STATUS_CORRUPT_DATA, /* data errors in user (meta)data */
ZPOOL_STATUS_FAILING_DEV, /* device experiencing errors */
ZPOOL_STATUS_VERSION_NEWER, /* newer on-disk version */
ZPOOL_STATUS_HOSTID_MISMATCH, /* last accessed by another system */
ZPOOL_STATUS_HOSTID_ACTIVE, /* currently active on another system */
ZPOOL_STATUS_HOSTID_REQUIRED, /* multihost=on and hostid=0 */
ZPOOL_STATUS_IO_FAILURE_WAIT, /* failed I/O, failmode 'wait' */
ZPOOL_STATUS_IO_FAILURE_CONTINUE, /* failed I/O, failmode 'continue' */
ZPOOL_STATUS_IO_FAILURE_MMP, /* failed MMP, failmode not 'panic' */
ZPOOL_STATUS_BAD_LOG, /* cannot read log chain(s) */
ZPOOL_STATUS_ERRATA, /* informational errata available */
/*
* If the pool has unsupported features but can still be opened in
* read-only mode, its status is ZPOOL_STATUS_UNSUP_FEAT_WRITE. If the
* pool has unsupported features but cannot be opened at all, its
* status is ZPOOL_STATUS_UNSUP_FEAT_READ.
*/
ZPOOL_STATUS_UNSUP_FEAT_READ, /* unsupported features for read */
ZPOOL_STATUS_UNSUP_FEAT_WRITE, /* unsupported features for write */
/*
* These faults have no corresponding message ID. At the time we are
* checking the status, the original reason for the FMA fault (I/O or
* checksum errors) has been lost.
*/
ZPOOL_STATUS_FAULTED_DEV_R, /* faulted device with replicas */
ZPOOL_STATUS_FAULTED_DEV_NR, /* faulted device with no replicas */
/*
* The following are not faults per se, but still an error possibly
* requiring administrative attention. There is no corresponding
* message ID.
*/
ZPOOL_STATUS_VERSION_OLDER, /* older legacy on-disk version */
ZPOOL_STATUS_FEAT_DISABLED, /* supported features are disabled */
ZPOOL_STATUS_RESILVERING, /* device being resilvered */
ZPOOL_STATUS_OFFLINE_DEV, /* device offline */
ZPOOL_STATUS_REMOVED_DEV, /* removed device */
ZPOOL_STATUS_REBUILDING, /* device being rebuilt */
ZPOOL_STATUS_REBUILD_SCRUB, /* recommend scrubbing the pool */
ZPOOL_STATUS_NON_NATIVE_ASHIFT, /* (e.g. 512e dev with ashift of 9) */
ZPOOL_STATUS_COMPATIBILITY_ERR, /* bad 'compatibility' property */
ZPOOL_STATUS_INCOMPATIBLE_FEAT, /* feature set outside compatibility */
/*
* Finally, the following indicates a healthy pool.
*/
ZPOOL_STATUS_OK
} zpool_status_t;
_LIBZFS_H zpool_status_t zpool_get_status(zpool_handle_t *, const char **,
zpool_errata_t *);
_LIBZFS_H zpool_status_t zpool_import_status(nvlist_t *, const char **,
zpool_errata_t *);
/*
* Statistics and configuration functions.
*/
_LIBZFS_H nvlist_t *zpool_get_config(zpool_handle_t *, nvlist_t **);
_LIBZFS_H nvlist_t *zpool_get_features(zpool_handle_t *);
_LIBZFS_H int zpool_refresh_stats(zpool_handle_t *, boolean_t *);
_LIBZFS_H int zpool_get_errlog(zpool_handle_t *, nvlist_t **);
/*
* Import and export functions
*/
_LIBZFS_H int zpool_export(zpool_handle_t *, boolean_t, const char *);
_LIBZFS_H int zpool_export_force(zpool_handle_t *, const char *);
_LIBZFS_H int zpool_import(libzfs_handle_t *, nvlist_t *, const char *,
char *altroot);
_LIBZFS_H int zpool_import_props(libzfs_handle_t *, nvlist_t *, const char *,
nvlist_t *, int);
_LIBZFS_H void zpool_print_unsup_feat(nvlist_t *config);
/*
* Miscellaneous pool functions
*/
struct zfs_cmd;
_LIBZFS_H const char *const zfs_history_event_names[];
typedef enum {
VDEV_NAME_PATH = 1 << 0,
VDEV_NAME_GUID = 1 << 1,
VDEV_NAME_FOLLOW_LINKS = 1 << 2,
VDEV_NAME_TYPE_ID = 1 << 3,
} vdev_name_t;
_LIBZFS_H char *zpool_vdev_name(libzfs_handle_t *, zpool_handle_t *, nvlist_t *,
int name_flags);
_LIBZFS_H int zpool_upgrade(zpool_handle_t *, uint64_t);
_LIBZFS_H int zpool_get_history(zpool_handle_t *, nvlist_t **, uint64_t *,
boolean_t *);
_LIBZFS_H int zpool_events_next(libzfs_handle_t *, nvlist_t **, int *, unsigned,
int);
_LIBZFS_H int zpool_events_clear(libzfs_handle_t *, int *);
_LIBZFS_H int zpool_events_seek(libzfs_handle_t *, uint64_t, int);
_LIBZFS_H void zpool_obj_to_path_ds(zpool_handle_t *, uint64_t, uint64_t,
char *, size_t);
_LIBZFS_H void zpool_obj_to_path(zpool_handle_t *, uint64_t, uint64_t, char *,
size_t);
_LIBZFS_H int zfs_ioctl(libzfs_handle_t *, int, struct zfs_cmd *);
_LIBZFS_H void zpool_explain_recover(libzfs_handle_t *, const char *, int,
nvlist_t *);
_LIBZFS_H int zpool_checkpoint(zpool_handle_t *);
_LIBZFS_H int zpool_discard_checkpoint(zpool_handle_t *);
_LIBZFS_H boolean_t zpool_is_draid_spare(const char *);
/*
* Basic handle manipulations. These functions do not create or destroy the
* underlying datasets, only the references to them.
*/
_LIBZFS_H zfs_handle_t *zfs_open(libzfs_handle_t *, const char *, int);
_LIBZFS_H zfs_handle_t *zfs_handle_dup(zfs_handle_t *);
_LIBZFS_H void zfs_close(zfs_handle_t *);
_LIBZFS_H zfs_type_t zfs_get_type(const zfs_handle_t *);
_LIBZFS_H zfs_type_t zfs_get_underlying_type(const zfs_handle_t *);
_LIBZFS_H const char *zfs_get_name(const zfs_handle_t *);
_LIBZFS_H zpool_handle_t *zfs_get_pool_handle(const zfs_handle_t *);
_LIBZFS_H const char *zfs_get_pool_name(const zfs_handle_t *);
/*
* Property management functions. Some functions are shared with the kernel,
* and are found in sys/fs/zfs.h.
*/
/*
* zfs dataset property management
*/
_LIBZFS_H const char *zfs_prop_default_string(zfs_prop_t);
_LIBZFS_H uint64_t zfs_prop_default_numeric(zfs_prop_t);
_LIBZFS_H const char *zfs_prop_column_name(zfs_prop_t);
_LIBZFS_H boolean_t zfs_prop_align_right(zfs_prop_t);
_LIBZFS_H nvlist_t *zfs_valid_proplist(libzfs_handle_t *, zfs_type_t,
nvlist_t *, uint64_t, zfs_handle_t *, zpool_handle_t *, boolean_t,
const char *);
_LIBZFS_H const char *zfs_prop_to_name(zfs_prop_t);
_LIBZFS_H int zfs_prop_set(zfs_handle_t *, const char *, const char *);
_LIBZFS_H int zfs_prop_set_list(zfs_handle_t *, nvlist_t *);
_LIBZFS_H int zfs_prop_get(zfs_handle_t *, zfs_prop_t, char *, size_t,
zprop_source_t *, char *, size_t, boolean_t);
_LIBZFS_H int zfs_prop_get_recvd(zfs_handle_t *, const char *, char *, size_t,
boolean_t);
_LIBZFS_H int zfs_prop_get_numeric(zfs_handle_t *, zfs_prop_t, uint64_t *,
zprop_source_t *, char *, size_t);
_LIBZFS_H int zfs_prop_get_userquota_int(zfs_handle_t *zhp,
const char *propname, uint64_t *propvalue);
_LIBZFS_H int zfs_prop_get_userquota(zfs_handle_t *zhp, const char *propname,
char *propbuf, int proplen, boolean_t literal);
_LIBZFS_H int zfs_prop_get_written_int(zfs_handle_t *zhp, const char *propname,
uint64_t *propvalue);
_LIBZFS_H int zfs_prop_get_written(zfs_handle_t *zhp, const char *propname,
char *propbuf, int proplen, boolean_t literal);
_LIBZFS_H int zfs_prop_get_feature(zfs_handle_t *zhp, const char *propname,
char *buf, size_t len);
_LIBZFS_H uint64_t getprop_uint64(zfs_handle_t *, zfs_prop_t, const char **);
_LIBZFS_H uint64_t zfs_prop_get_int(zfs_handle_t *, zfs_prop_t);
_LIBZFS_H int zfs_prop_inherit(zfs_handle_t *, const char *, boolean_t);
_LIBZFS_H const char *zfs_prop_values(zfs_prop_t);
_LIBZFS_H int zfs_prop_is_string(zfs_prop_t prop);
_LIBZFS_H nvlist_t *zfs_get_all_props(zfs_handle_t *);
_LIBZFS_H nvlist_t *zfs_get_user_props(zfs_handle_t *);
_LIBZFS_H nvlist_t *zfs_get_recvd_props(zfs_handle_t *);
_LIBZFS_H nvlist_t *zfs_get_clones_nvl(zfs_handle_t *);
_LIBZFS_H int zfs_wait_status(zfs_handle_t *, zfs_wait_activity_t,
boolean_t *, boolean_t *);
/*
* zfs encryption management
*/
_LIBZFS_H int zfs_crypto_get_encryption_root(zfs_handle_t *, boolean_t *,
char *);
_LIBZFS_H int zfs_crypto_create(libzfs_handle_t *, char *, nvlist_t *,
nvlist_t *, boolean_t stdin_available, uint8_t **, uint_t *);
_LIBZFS_H int zfs_crypto_clone_check(libzfs_handle_t *, zfs_handle_t *, char *,
nvlist_t *);
_LIBZFS_H int zfs_crypto_attempt_load_keys(libzfs_handle_t *, const char *);
_LIBZFS_H int zfs_crypto_load_key(zfs_handle_t *, boolean_t, const char *);
_LIBZFS_H int zfs_crypto_unload_key(zfs_handle_t *);
_LIBZFS_H int zfs_crypto_rewrap(zfs_handle_t *, nvlist_t *, boolean_t);
typedef struct zprop_list {
int pl_prop;
char *pl_user_prop;
struct zprop_list *pl_next;
boolean_t pl_all;
size_t pl_width;
size_t pl_recvd_width;
boolean_t pl_fixed;
} zprop_list_t;
_LIBZFS_H int zfs_expand_proplist(zfs_handle_t *, zprop_list_t **, boolean_t,
boolean_t);
_LIBZFS_H void zfs_prune_proplist(zfs_handle_t *, uint8_t *);
_LIBZFS_H int vdev_expand_proplist(zpool_handle_t *, const char *,
zprop_list_t **);
#define ZFS_MOUNTPOINT_NONE "none"
#define ZFS_MOUNTPOINT_LEGACY "legacy"
#define ZFS_FEATURE_DISABLED "disabled"
#define ZFS_FEATURE_ENABLED "enabled"
#define ZFS_FEATURE_ACTIVE "active"
#define ZFS_UNSUPPORTED_INACTIVE "inactive"
#define ZFS_UNSUPPORTED_READONLY "readonly"
/*
* zpool property management
*/
_LIBZFS_H int zpool_expand_proplist(zpool_handle_t *, zprop_list_t **,
zfs_type_t, boolean_t);
_LIBZFS_H int zpool_prop_get_feature(zpool_handle_t *, const char *, char *,
size_t);
_LIBZFS_H const char *zpool_prop_default_string(zpool_prop_t);
_LIBZFS_H uint64_t zpool_prop_default_numeric(zpool_prop_t);
_LIBZFS_H const char *zpool_prop_column_name(zpool_prop_t);
_LIBZFS_H boolean_t zpool_prop_align_right(zpool_prop_t);
/*
* Functions shared by zfs and zpool property management.
*/
_LIBZFS_H int zprop_iter(zprop_func func, void *cb, boolean_t show_all,
boolean_t ordered, zfs_type_t type);
_LIBZFS_H int zprop_get_list(libzfs_handle_t *, char *, zprop_list_t **,
zfs_type_t);
_LIBZFS_H void zprop_free_list(zprop_list_t *);
#define ZFS_GET_NCOLS 5
typedef enum {
GET_COL_NONE,
GET_COL_NAME,
GET_COL_PROPERTY,
GET_COL_VALUE,
GET_COL_RECVD,
GET_COL_SOURCE
} zfs_get_column_t;
/*
* Functions for printing zfs or zpool properties
*/
typedef struct vdev_cbdata {
int cb_name_flags;
char **cb_names;
unsigned int cb_names_count;
} vdev_cbdata_t;
typedef struct zprop_get_cbdata {
int cb_sources;
zfs_get_column_t cb_columns[ZFS_GET_NCOLS];
int cb_colwidths[ZFS_GET_NCOLS + 1];
boolean_t cb_scripted;
boolean_t cb_literal;
boolean_t cb_first;
zprop_list_t *cb_proplist;
zfs_type_t cb_type;
vdev_cbdata_t cb_vdevs;
} zprop_get_cbdata_t;
_LIBZFS_H void zprop_print_one_property(const char *, zprop_get_cbdata_t *,
const char *, const char *, zprop_source_t, const char *,
const char *);
/*
* Iterator functions.
*/
#define ZFS_ITER_RECURSE (1 << 0)
#define ZFS_ITER_ARGS_CAN_BE_PATHS (1 << 1)
#define ZFS_ITER_PROP_LISTSNAPS (1 << 2)
#define ZFS_ITER_DEPTH_LIMIT (1 << 3)
#define ZFS_ITER_RECVD_PROPS (1 << 4)
#define ZFS_ITER_LITERAL_PROPS (1 << 5)
#define ZFS_ITER_SIMPLE (1 << 6)
typedef int (*zfs_iter_f)(zfs_handle_t *, void *);
_LIBZFS_H int zfs_iter_root(libzfs_handle_t *, zfs_iter_f, void *);
-_LIBZFS_H int zfs_iter_children(zfs_handle_t *, int, zfs_iter_f, void *);
-_LIBZFS_H int zfs_iter_dependents(zfs_handle_t *, int, boolean_t, zfs_iter_f,
+_LIBZFS_H int zfs_iter_children(zfs_handle_t *, zfs_iter_f, void *);
+_LIBZFS_H int zfs_iter_dependents(zfs_handle_t *, boolean_t, zfs_iter_f,
void *);
-_LIBZFS_H int zfs_iter_filesystems(zfs_handle_t *, int, zfs_iter_f, void *);
-_LIBZFS_H int zfs_iter_snapshots(zfs_handle_t *, int, zfs_iter_f, void *,
+_LIBZFS_H int zfs_iter_filesystems(zfs_handle_t *, zfs_iter_f, void *);
+_LIBZFS_H int zfs_iter_snapshots(zfs_handle_t *, boolean_t, zfs_iter_f, void *,
uint64_t, uint64_t);
-_LIBZFS_H int zfs_iter_snapshots_sorted(zfs_handle_t *, int, zfs_iter_f, void *,
+_LIBZFS_H int zfs_iter_snapshots_sorted(zfs_handle_t *, zfs_iter_f, void *,
uint64_t, uint64_t);
-_LIBZFS_H int zfs_iter_snapspec(zfs_handle_t *, int, const char *, zfs_iter_f,
+_LIBZFS_H int zfs_iter_snapspec(zfs_handle_t *, const char *, zfs_iter_f,
void *);
-_LIBZFS_H int zfs_iter_bookmarks(zfs_handle_t *, int, zfs_iter_f, void *);
+_LIBZFS_H int zfs_iter_bookmarks(zfs_handle_t *, zfs_iter_f, void *);
+
+_LIBZFS_H int zfs_iter_children_v2(zfs_handle_t *, int, zfs_iter_f, void *);
+_LIBZFS_H int zfs_iter_dependents_v2(zfs_handle_t *, int, boolean_t, zfs_iter_f,
+ void *);
+_LIBZFS_H int zfs_iter_filesystems_v2(zfs_handle_t *, int, zfs_iter_f, void *);
+_LIBZFS_H int zfs_iter_snapshots_v2(zfs_handle_t *, int, zfs_iter_f, void *,
+ uint64_t, uint64_t);
+_LIBZFS_H int zfs_iter_snapshots_sorted_v2(zfs_handle_t *, int, zfs_iter_f,
+ void *, uint64_t, uint64_t);
+_LIBZFS_H int zfs_iter_snapspec_v2(zfs_handle_t *, int, const char *,
+ zfs_iter_f, void *);
+_LIBZFS_H int zfs_iter_bookmarks_v2(zfs_handle_t *, int, zfs_iter_f, void *);
_LIBZFS_H int zfs_iter_mounted(zfs_handle_t *, zfs_iter_f, void *);
typedef struct get_all_cb {
zfs_handle_t **cb_handles;
size_t cb_alloc;
size_t cb_used;
} get_all_cb_t;
_LIBZFS_H void zfs_foreach_mountpoint(libzfs_handle_t *, zfs_handle_t **,
size_t, zfs_iter_f, void *, boolean_t);
_LIBZFS_H void libzfs_add_handle(get_all_cb_t *, zfs_handle_t *);
/*
* Functions to create and destroy datasets.
*/
_LIBZFS_H int zfs_create(libzfs_handle_t *, const char *, zfs_type_t,
nvlist_t *);
_LIBZFS_H int zfs_create_ancestors(libzfs_handle_t *, const char *);
_LIBZFS_H int zfs_destroy(zfs_handle_t *, boolean_t);
_LIBZFS_H int zfs_destroy_snaps(zfs_handle_t *, char *, boolean_t);
_LIBZFS_H int zfs_destroy_snaps_nvl(libzfs_handle_t *, nvlist_t *, boolean_t);
_LIBZFS_H int zfs_destroy_snaps_nvl_os(libzfs_handle_t *, nvlist_t *);
_LIBZFS_H int zfs_clone(zfs_handle_t *, const char *, nvlist_t *);
_LIBZFS_H int zfs_snapshot(libzfs_handle_t *, const char *, boolean_t,
nvlist_t *);
_LIBZFS_H int zfs_snapshot_nvl(libzfs_handle_t *hdl, nvlist_t *snaps,
nvlist_t *props);
_LIBZFS_H int zfs_rollback(zfs_handle_t *, zfs_handle_t *, boolean_t);
typedef struct renameflags {
/* recursive rename */
unsigned int recursive : 1;
/* don't unmount file systems */
unsigned int nounmount : 1;
/* force unmount file systems */
unsigned int forceunmount : 1;
} renameflags_t;
_LIBZFS_H int zfs_rename(zfs_handle_t *, const char *, renameflags_t);
typedef struct sendflags {
/* Amount of extra information to print. */
int verbosity;
/* recursive send (ie, -R) */
boolean_t replicate;
/* for recursive send, skip sending missing snapshots */
boolean_t skipmissing;
/* for incrementals, do all intermediate snapshots */
boolean_t doall;
/* if dataset is a clone, do incremental from its origin */
boolean_t fromorigin;
/* field no longer used, maintained for backwards compatibility */
boolean_t pad;
/* send properties (ie, -p) */
boolean_t props;
/* do not send (no-op, ie. -n) */
boolean_t dryrun;
/* parsable verbose output (ie. -P) */
boolean_t parsable;
/* show progress (ie. -v) */
boolean_t progress;
/* show progress as process title (ie. -V) */
boolean_t progressastitle;
/* large blocks (>128K) are permitted */
boolean_t largeblock;
/* WRITE_EMBEDDED records of type DATA are permitted */
boolean_t embed_data;
/* compressed WRITE records are permitted */
boolean_t compress;
/* raw encrypted records are permitted */
boolean_t raw;
/* only send received properties (ie. -b) */
boolean_t backup;
/* include snapshot holds in send stream */
boolean_t holds;
/* stream represents a partially received dataset */
boolean_t saved;
} sendflags_t;
typedef boolean_t (snapfilter_cb_t)(zfs_handle_t *, void *);
_LIBZFS_H int zfs_send(zfs_handle_t *, const char *, const char *,
sendflags_t *, int, snapfilter_cb_t, void *, nvlist_t **);
_LIBZFS_H int zfs_send_one(zfs_handle_t *, const char *, int, sendflags_t *,
const char *);
_LIBZFS_H int zfs_send_progress(zfs_handle_t *, int, uint64_t *, uint64_t *);
_LIBZFS_H int zfs_send_resume(libzfs_handle_t *, sendflags_t *, int outfd,
const char *);
_LIBZFS_H int zfs_send_saved(zfs_handle_t *, sendflags_t *, int, const char *);
_LIBZFS_H nvlist_t *zfs_send_resume_token_to_nvlist(libzfs_handle_t *hdl,
const char *token);
_LIBZFS_H int zfs_promote(zfs_handle_t *);
_LIBZFS_H int zfs_hold(zfs_handle_t *, const char *, const char *,
boolean_t, int);
_LIBZFS_H int zfs_hold_nvl(zfs_handle_t *, int, nvlist_t *);
_LIBZFS_H int zfs_release(zfs_handle_t *, const char *, const char *,
boolean_t);
_LIBZFS_H int zfs_get_holds(zfs_handle_t *, nvlist_t **);
_LIBZFS_H uint64_t zvol_volsize_to_reservation(zpool_handle_t *, uint64_t,
nvlist_t *);
typedef int (*zfs_userspace_cb_t)(void *arg, const char *domain,
uid_t rid, uint64_t space);
_LIBZFS_H int zfs_userspace(zfs_handle_t *, zfs_userquota_prop_t,
zfs_userspace_cb_t, void *);
_LIBZFS_H int zfs_get_fsacl(zfs_handle_t *, nvlist_t **);
_LIBZFS_H int zfs_set_fsacl(zfs_handle_t *, boolean_t, nvlist_t *);
typedef struct recvflags {
/* print informational messages (ie, -v was specified) */
boolean_t verbose;
/* the destination is a prefix, not the exact fs (ie, -d) */
boolean_t isprefix;
/*
* Only the tail of the sent snapshot path is appended to the
* destination to determine the received snapshot name (ie, -e).
*/
boolean_t istail;
/* do not actually do the recv, just check if it would work (ie, -n) */
boolean_t dryrun;
/* rollback/destroy filesystems as necessary (eg, -F) */
boolean_t force;
/* set "canmount=off" on all modified filesystems */
boolean_t canmountoff;
/*
* Mark the file systems as "resumable" and do not destroy them if the
* receive is interrupted
*/
boolean_t resumable;
/* byteswap flag is used internally; callers need not specify */
boolean_t byteswap;
/* do not mount file systems as they are extracted (private) */
boolean_t nomount;
/* Was holds flag set in the compound header? */
boolean_t holds;
/* skip receive of snapshot holds */
boolean_t skipholds;
/* mount the filesystem unless nomount is specified */
boolean_t domount;
/* force unmount while recv snapshot (private) */
boolean_t forceunmount;
/* use this recv to check (and heal if needed) an existing snapshot */
boolean_t heal;
} recvflags_t;
_LIBZFS_H int zfs_receive(libzfs_handle_t *, const char *, nvlist_t *,
recvflags_t *, int, avl_tree_t *);
typedef enum diff_flags {
ZFS_DIFF_PARSEABLE = 1 << 0,
ZFS_DIFF_TIMESTAMP = 1 << 1,
ZFS_DIFF_CLASSIFY = 1 << 2,
ZFS_DIFF_NO_MANGLE = 1 << 3
} diff_flags_t;
_LIBZFS_H int zfs_show_diffs(zfs_handle_t *, int, const char *, const char *,
int);
/*
* Miscellaneous functions.
*/
_LIBZFS_H const char *zfs_type_to_name(zfs_type_t);
_LIBZFS_H void zfs_refresh_properties(zfs_handle_t *);
_LIBZFS_H int zfs_name_valid(const char *, zfs_type_t);
_LIBZFS_H zfs_handle_t *zfs_path_to_zhandle(libzfs_handle_t *, const char *,
zfs_type_t);
_LIBZFS_H int zfs_parent_name(zfs_handle_t *, char *, size_t);
_LIBZFS_H boolean_t zfs_dataset_exists(libzfs_handle_t *, const char *,
zfs_type_t);
_LIBZFS_H int zfs_spa_version(zfs_handle_t *, int *);
_LIBZFS_H boolean_t zfs_bookmark_exists(const char *path);
/*
* Mount support functions.
*/
_LIBZFS_H boolean_t is_mounted(libzfs_handle_t *, const char *special, char **);
_LIBZFS_H boolean_t zfs_is_mounted(zfs_handle_t *, char **);
_LIBZFS_H int zfs_mount(zfs_handle_t *, const char *, int);
_LIBZFS_H int zfs_mount_at(zfs_handle_t *, const char *, int, const char *);
_LIBZFS_H int zfs_unmount(zfs_handle_t *, const char *, int);
_LIBZFS_H int zfs_unmountall(zfs_handle_t *, int);
_LIBZFS_H int zfs_mount_delegation_check(void);
#if defined(__linux__) || defined(__APPLE__)
_LIBZFS_H int zfs_parse_mount_options(const char *mntopts,
unsigned long *mntflags, unsigned long *zfsflags, int sloppy, char *badopt,
char *mtabopt);
_LIBZFS_H void zfs_adjust_mount_options(zfs_handle_t *zhp, const char *mntpoint,
char *mntopts, char *mtabopt);
#endif
/*
* Share support functions.
*
* enum sa_protocol * lists are terminated with SA_NO_PROTOCOL,
* NULL means "all/any known to this libzfs".
*/
#define SA_NO_PROTOCOL -1
_LIBZFS_H boolean_t zfs_is_shared(zfs_handle_t *zhp, char **where,
const enum sa_protocol *proto);
_LIBZFS_H int zfs_share(zfs_handle_t *zhp, const enum sa_protocol *proto);
_LIBZFS_H int zfs_unshare(zfs_handle_t *zhp, const char *mountpoint,
const enum sa_protocol *proto);
_LIBZFS_H int zfs_unshareall(zfs_handle_t *zhp,
const enum sa_protocol *proto);
_LIBZFS_H void zfs_commit_shares(const enum sa_protocol *proto);
_LIBZFS_H void zfs_truncate_shares(const enum sa_protocol *proto);
_LIBZFS_H int zfs_nicestrtonum(libzfs_handle_t *, const char *, uint64_t *);
/*
* Utility functions to run an external process.
*/
#define STDOUT_VERBOSE 0x01
#define STDERR_VERBOSE 0x02
#define NO_DEFAULT_PATH 0x04 /* Don't use $PATH to lookup the command */
_LIBZFS_H int libzfs_run_process(const char *, char **, int);
_LIBZFS_H int libzfs_run_process_get_stdout(const char *, char *[], char *[],
char **[], int *);
_LIBZFS_H int libzfs_run_process_get_stdout_nopath(const char *, char *[],
char *[], char **[], int *);
_LIBZFS_H void libzfs_free_str_array(char **, int);
_LIBZFS_H boolean_t libzfs_envvar_is_set(const char *);
/*
* Utility functions for zfs version
*/
_LIBZFS_H const char *zfs_version_userland(void);
_LIBZFS_H char *zfs_version_kernel(void);
_LIBZFS_H int zfs_version_print(void);
/*
* Given a device or file, determine if it is part of a pool.
*/
_LIBZFS_H int zpool_in_use(libzfs_handle_t *, int, pool_state_t *, char **,
boolean_t *);
/*
* Label manipulation.
*/
_LIBZFS_H int zpool_clear_label(int);
_LIBZFS_H int zpool_set_bootenv(zpool_handle_t *, const nvlist_t *);
_LIBZFS_H int zpool_get_bootenv(zpool_handle_t *, nvlist_t **);
/*
* Management interfaces for SMB ACL files
*/
_LIBZFS_H int zfs_smb_acl_add(libzfs_handle_t *, char *, char *, char *);
_LIBZFS_H int zfs_smb_acl_remove(libzfs_handle_t *, char *, char *, char *);
_LIBZFS_H int zfs_smb_acl_purge(libzfs_handle_t *, char *, char *);
_LIBZFS_H int zfs_smb_acl_rename(libzfs_handle_t *, char *, char *, char *,
char *);
/*
* Enable and disable datasets within a pool by mounting/unmounting and
* sharing/unsharing them.
*/
_LIBZFS_H int zpool_enable_datasets(zpool_handle_t *, const char *, int);
_LIBZFS_H int zpool_disable_datasets(zpool_handle_t *, boolean_t);
_LIBZFS_H void zpool_disable_datasets_os(zpool_handle_t *, boolean_t);
_LIBZFS_H void zpool_disable_volume_os(const char *);
/*
* Parse a features file for -o compatibility
*/
typedef enum {
ZPOOL_COMPATIBILITY_OK,
ZPOOL_COMPATIBILITY_WARNTOKEN,
ZPOOL_COMPATIBILITY_BADTOKEN,
ZPOOL_COMPATIBILITY_BADFILE,
ZPOOL_COMPATIBILITY_NOFILES
} zpool_compat_status_t;
_LIBZFS_H zpool_compat_status_t zpool_load_compat(const char *,
boolean_t *, char *, size_t);
#ifdef __FreeBSD__
/*
* Attach/detach the given filesystem to/from the given jail.
*/
_LIBZFS_H int zfs_jail(zfs_handle_t *zhp, int jailid, int attach);
/*
* Set loader options for next boot.
*/
_LIBZFS_H int zpool_nextboot(libzfs_handle_t *, uint64_t, uint64_t,
const char *);
#endif /* __FreeBSD__ */
#ifdef __linux__
/*
* Add or delete the given filesystem to/from the given user namespace.
*/
_LIBZFS_H int zfs_userns(zfs_handle_t *zhp, const char *nspath, int attach);
#endif
#ifdef __cplusplus
}
#endif
#endif /* _LIBZFS_H */
diff --git a/sys/contrib/openzfs/include/libzutil.h b/sys/contrib/openzfs/include/libzutil.h
index 465e463f0c1f..237ff976ba62 100644
--- a/sys/contrib/openzfs/include/libzutil.h
+++ b/sys/contrib/openzfs/include/libzutil.h
@@ -1,214 +1,215 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2018 by Delphix. All rights reserved.
*/
#ifndef _LIBZUTIL_H
#define _LIBZUTIL_H extern __attribute__((visibility("default")))
#include <sys/nvpair.h>
#include <sys/fs/zfs.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* Default wait time for a device name to be created.
*/
#define DISK_LABEL_WAIT (30 * 1000) /* 30 seconds */
/*
* Pool Config Operations
*
* These are specific to the library libzfs or libzpool instance.
*/
typedef nvlist_t *refresh_config_func_t(void *, nvlist_t *);
typedef int pool_active_func_t(void *, const char *, uint64_t, boolean_t *);
typedef const struct pool_config_ops {
refresh_config_func_t *pco_refresh_config;
pool_active_func_t *pco_pool_active;
} pool_config_ops_t;
/*
* An instance of pool_config_ops_t is expected in the caller's binary.
*/
_LIBZUTIL_H pool_config_ops_t libzfs_config_ops;
_LIBZUTIL_H pool_config_ops_t libzpool_config_ops;
typedef enum lpc_error {
LPC_SUCCESS = 0, /* no error -- success */
LPC_BADCACHE = 2000, /* out of memory */
LPC_BADPATH, /* must be an absolute path */
LPC_NOMEM, /* out of memory */
LPC_EACCESS, /* some devices require root privileges */
LPC_UNKNOWN
} lpc_error_t;
typedef struct importargs {
char **path; /* a list of paths to search */
int paths; /* number of paths to search */
const char *poolname; /* name of a pool to find */
uint64_t guid; /* guid of a pool to find */
const char *cachefile; /* cachefile to use for import */
boolean_t can_be_active; /* can the pool be active? */
boolean_t scan; /* prefer scanning to libblkid cache */
nvlist_t *policy; /* load policy (max txg, rewind, etc.) */
} importargs_t;
typedef struct libpc_handle {
int lpc_error;
boolean_t lpc_printerr;
boolean_t lpc_open_access_error;
boolean_t lpc_desc_active;
char lpc_desc[1024];
pool_config_ops_t *lpc_ops;
void *lpc_lib_handle;
} libpc_handle_t;
_LIBZUTIL_H const char *libpc_error_description(libpc_handle_t *);
_LIBZUTIL_H nvlist_t *zpool_search_import(libpc_handle_t *, importargs_t *);
_LIBZUTIL_H int zpool_find_config(libpc_handle_t *, const char *, nvlist_t **,
importargs_t *);
_LIBZUTIL_H const char * const * zpool_default_search_paths(size_t *count);
_LIBZUTIL_H int zpool_read_label(int, nvlist_t **, int *);
_LIBZUTIL_H int zpool_label_disk_wait(const char *, int);
struct udev_device;
_LIBZUTIL_H int zfs_device_get_devid(struct udev_device *, char *, size_t);
_LIBZUTIL_H int zfs_device_get_physical(struct udev_device *, char *, size_t);
_LIBZUTIL_H void update_vdev_config_dev_strs(nvlist_t *);
/*
* Default device paths
*/
#define DISK_ROOT "/dev"
#define UDISK_ROOT "/dev/disk"
#define ZVOL_ROOT "/dev/zvol"
_LIBZUTIL_H int zfs_append_partition(char *path, size_t max_len);
_LIBZUTIL_H int zfs_resolve_shortname(const char *name, char *path,
size_t pathlen);
_LIBZUTIL_H char *zfs_strip_partition(const char *);
_LIBZUTIL_H const char *zfs_strip_path(const char *);
_LIBZUTIL_H int zfs_strcmp_pathname(const char *, const char *, int);
_LIBZUTIL_H boolean_t zfs_dev_is_dm(const char *);
_LIBZUTIL_H boolean_t zfs_dev_is_whole_disk(const char *);
_LIBZUTIL_H int zfs_dev_flush(int);
_LIBZUTIL_H char *zfs_get_underlying_path(const char *);
_LIBZUTIL_H char *zfs_get_enclosure_sysfs_path(const char *);
_LIBZUTIL_H boolean_t is_mpath_whole_disk(const char *);
_LIBZUTIL_H boolean_t zfs_isnumber(const char *);
/*
* Formats for iostat numbers. Examples: "12K", "30ms", "4B", "2321234", "-".
*
* ZFS_NICENUM_1024: Print kilo, mega, tera, peta, exa..
* ZFS_NICENUM_BYTES: Print single bytes ("13B"), kilo, mega, tera...
* ZFS_NICENUM_TIME: Print nanosecs, microsecs, millisecs, seconds...
* ZFS_NICENUM_RAW: Print the raw number without any formatting
* ZFS_NICENUM_RAWTIME: Same as RAW, but print dashes ('-') for zero.
*/
enum zfs_nicenum_format {
ZFS_NICENUM_1024 = 0,
ZFS_NICENUM_BYTES = 1,
ZFS_NICENUM_TIME = 2,
ZFS_NICENUM_RAW = 3,
ZFS_NICENUM_RAWTIME = 4
};
/*
* Convert a number to a human-readable form.
*/
_LIBZUTIL_H void zfs_nicebytes(uint64_t, char *, size_t);
_LIBZUTIL_H void zfs_nicenum(uint64_t, char *, size_t);
_LIBZUTIL_H void zfs_nicenum_format(uint64_t, char *, size_t,
enum zfs_nicenum_format);
_LIBZUTIL_H void zfs_nicetime(uint64_t, char *, size_t);
_LIBZUTIL_H void zfs_niceraw(uint64_t, char *, size_t);
#define nicenum(num, buf, size) zfs_nicenum(num, buf, size)
_LIBZUTIL_H void zpool_dump_ddt(const ddt_stat_t *, const ddt_histogram_t *);
_LIBZUTIL_H int zpool_history_unpack(char *, uint64_t, uint64_t *, nvlist_t ***,
uint_t *);
struct zfs_cmd;
/*
* List of colors to use
*/
#define ANSI_BLACK "\033[0;30m"
#define ANSI_RED "\033[0;31m"
#define ANSI_GREEN "\033[0;32m"
#define ANSI_YELLOW "\033[0;33m"
#define ANSI_BLUE "\033[0;34m"
#define ANSI_BOLD_BLUE "\033[1;34m" /* light blue */
#define ANSI_MAGENTA "\033[0;35m"
#define ANSI_CYAN "\033[0;36m"
#define ANSI_GRAY "\033[0;37m"
#define ANSI_RESET "\033[0m"
#define ANSI_BOLD "\033[1m"
+_LIBZUTIL_H int use_color(void);
_LIBZUTIL_H void color_start(const char *color);
_LIBZUTIL_H void color_end(void);
_LIBZUTIL_H int printf_color(const char *color, const char *format, ...);
_LIBZUTIL_H const char *zfs_basename(const char *path);
_LIBZUTIL_H ssize_t zfs_dirnamelen(const char *path);
#ifdef __linux__
extern char **environ;
_LIBZUTIL_H void zfs_setproctitle_init(int argc, char *argv[], char *envp[]);
_LIBZUTIL_H void zfs_setproctitle(const char *fmt, ...);
#else
#define zfs_setproctitle(fmt, ...) setproctitle(fmt, ##__VA_ARGS__)
#define zfs_setproctitle_init(x, y, z) ((void)0)
#endif
/*
* These functions are used by the ZFS libraries and cmd/zpool code, but are
* not exported in the ABI.
*/
typedef int (*pool_vdev_iter_f)(void *, nvlist_t *, void *);
int for_each_vdev_cb(void *zhp, nvlist_t *nv, pool_vdev_iter_f func,
void *data);
int for_each_vdev_in_nvlist(nvlist_t *nvroot, pool_vdev_iter_f func,
void *data);
void update_vdevs_config_dev_sysfs_path(nvlist_t *config);
#ifdef __cplusplus
}
#endif
#endif /* _LIBZUTIL_H */
diff --git a/sys/contrib/openzfs/include/os/freebsd/spl/sys/simd_aarch64.h b/sys/contrib/openzfs/include/os/freebsd/spl/sys/simd_aarch64.h
index 4fe211983c33..a17d409a4c5a 100644
--- a/sys/contrib/openzfs/include/os/freebsd/spl/sys/simd_aarch64.h
+++ b/sys/contrib/openzfs/include/os/freebsd/spl/sys/simd_aarch64.h
@@ -1,105 +1,105 @@
/*
* Copyright (C) 2022 Tino Reichardt <milky-zfs@mcmilk.de>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
/*
* SIMD support:
*
* Following functions should be called to determine whether CPU feature
* is supported. All functions are usable in kernel and user space.
* If a SIMD algorithm is using more than one instruction set
* all relevant feature test functions should be called.
*
* Supported features:
* zfs_neon_available()
* zfs_sha256_available()
* zfs_sha512_available()
*/
#ifndef _FREEBSD_SIMD_AARCH64_H
#define _FREEBSD_SIMD_AARCH64_H
#include <sys/types.h>
#include <sys/ucontext.h>
#include <machine/elf.h>
#include <machine/fpu.h>
#include <machine/md_var.h>
#include <machine/pcb.h>
#ifdef _STANDALONE
#define kfpu_allowed() 0
#define kfpu_begin() do {} while (0)
#define kfpu_end() do {} while (0)
#else
#define kfpu_allowed() 1
#define kfpu_begin() do { \
if (__predict_false(!is_fpu_kern_thread(0))) \
fpu_kern_enter(curthread, NULL, FPU_KERN_NOCTX); \
-} while(0)
+} while (0)
#define kfpu_end() do { \
- if (__predict_false(curthread->td_pcb->pcb_fpflags & PCB_FP_NOSAVE)) \
+ if (__predict_false(curthread->td_pcb->pcb_fpflags & PCB_FP_NOSAVE)) \
fpu_kern_leave(curthread, NULL); \
-} while(0)
+} while (0)
#endif
#define kfpu_initialize(tsk) do {} while (0)
#define kfpu_init() (0)
#define kfpu_fini() do {} while (0)
/*
* Check if NEON is available
*/
static inline boolean_t
zfs_neon_available(void)
{
return (elf_hwcap & HWCAP_FP);
}
/*
* Check if SHA256 is available
*/
static inline boolean_t
zfs_sha256_available(void)
{
return (elf_hwcap & HWCAP_SHA2);
}
/*
* Check if SHA512 is available
*/
static inline boolean_t
zfs_sha512_available(void)
{
return (elf_hwcap & HWCAP_SHA512);
}
#endif /* _FREEBSD_SIMD_AARCH64_H */
diff --git a/sys/contrib/openzfs/include/os/freebsd/spl/sys/types.h b/sys/contrib/openzfs/include/os/freebsd/spl/sys/types.h
index 558843dcaa74..ebc93f4f4485 100644
--- a/sys/contrib/openzfs/include/os/freebsd/spl/sys/types.h
+++ b/sys/contrib/openzfs/include/os/freebsd/spl/sys/types.h
@@ -1,111 +1,111 @@
/*
* Copyright (c) 2007 Pawel Jakub Dawidek <pjd@FreeBSD.org>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
#ifndef _SPL_SYS_TYPES_H_
#define _SPL_SYS_TYPES_H_
#pragma once
/*
* This is a bag of dirty hacks to keep things compiling.
*/
#include_next <sys/types.h>
#ifdef __ILP32__
typedef __uint64_t u_longlong_t;
typedef __int64_t longlong_t;
#else
typedef unsigned long long u_longlong_t;
typedef long long longlong_t;
#endif
#include <sys/stdint.h>
#define _CLOCK_T_DECLARED
#include <sys/types32.h>
#include <sys/_stdarg.h>
#include <linux/types.h>
#define MAXNAMELEN 256
typedef void zfs_kernel_param_t;
typedef struct timespec timestruc_t;
typedef struct timespec timespec_t;
typedef struct timespec inode_timespec_t;
/* BEGIN CSTYLED */
typedef u_int uint_t;
typedef u_char uchar_t;
typedef u_short ushort_t;
typedef u_long ulong_t;
/* END CSTYLED */
typedef int minor_t;
#ifndef _OFF64_T_DECLARED
#define _OFF64_T_DECLARED
typedef off_t off64_t;
#endif
typedef id_t taskid_t;
typedef id_t projid_t;
typedef id_t poolid_t;
typedef uint_t zoneid_t;
typedef id_t ctid_t;
typedef mode_t o_mode_t;
typedef uint64_t pgcnt_t;
typedef short index_t;
typedef off_t offset_t;
#ifndef _PTRDIFF_T_DECLARED
typedef __ptrdiff_t ptrdiff_t; /* pointer difference */
#define _PTRDIFF_T_DECLARED
#endif
typedef int64_t rlim64_t;
typedef int major_t;
#ifdef NEED_SOLARIS_BOOLEAN
#if defined(__XOPEN_OR_POSIX)
typedef enum { _B_FALSE, _B_TRUE } boolean_t;
#else
typedef enum { B_FALSE, B_TRUE } boolean_t;
#endif /* defined(__XOPEN_OR_POSIX) */
#else
#define B_FALSE 0
#define B_TRUE 1
#endif
typedef u_longlong_t u_offset_t;
typedef u_longlong_t len_t;
typedef longlong_t diskaddr_t;
-typedef void zuserns_t;
+typedef void zidmap_t;
#include <sys/debug.h>
#endif /* !_OPENSOLARIS_SYS_TYPES_H_ */
diff --git a/sys/contrib/openzfs/include/os/freebsd/zfs/sys/zfs_vnops_os.h b/sys/contrib/openzfs/include/os/freebsd/zfs/sys/zfs_vnops_os.h
index 839ee629a5ab..eddcab575b91 100644
--- a/sys/contrib/openzfs/include/os/freebsd/zfs/sys/zfs_vnops_os.h
+++ b/sys/contrib/openzfs/include/os/freebsd/zfs/sys/zfs_vnops_os.h
@@ -1,60 +1,60 @@
/*
* Copyright (c) 2020 iXsystems, Inc.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
#ifndef _SYS_FS_ZFS_VNOPS_OS_H
#define _SYS_FS_ZFS_VNOPS_OS_H
int dmu_write_pages(objset_t *os, uint64_t object, uint64_t offset,
uint64_t size, struct vm_page **ppa, dmu_tx_t *tx);
int dmu_read_pages(objset_t *os, uint64_t object, vm_page_t *ma, int count,
int *rbehind, int *rahead, int last_size);
extern int zfs_remove(znode_t *dzp, const char *name, cred_t *cr, int flags);
extern int zfs_mkdir(znode_t *dzp, const char *dirname, vattr_t *vap,
- znode_t **zpp, cred_t *cr, int flags, vsecattr_t *vsecp, zuserns_t *mnt_ns);
+ znode_t **zpp, cred_t *cr, int flags, vsecattr_t *vsecp, zidmap_t *mnt_ns);
extern int zfs_rmdir(znode_t *dzp, const char *name, znode_t *cwd,
cred_t *cr, int flags);
extern int zfs_setattr(znode_t *zp, vattr_t *vap, int flag, cred_t *cr,
- zuserns_t *mnt_ns);
+ zidmap_t *mnt_ns);
extern int zfs_rename(znode_t *sdzp, const char *snm, znode_t *tdzp,
const char *tnm, cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap,
- zuserns_t *mnt_ns);
+ zidmap_t *mnt_ns);
extern int zfs_symlink(znode_t *dzp, const char *name, vattr_t *vap,
- const char *link, znode_t **zpp, cred_t *cr, int flags, zuserns_t *mnt_ns);
+ const char *link, znode_t **zpp, cred_t *cr, int flags, zidmap_t *mnt_ns);
extern int zfs_link(znode_t *tdzp, znode_t *sp,
const char *name, cred_t *cr, int flags);
extern int zfs_space(znode_t *zp, int cmd, struct flock *bfp, int flag,
offset_t offset, cred_t *cr);
extern int zfs_create(znode_t *dzp, const char *name, vattr_t *vap, int excl,
int mode, znode_t **zpp, cred_t *cr, int flag, vsecattr_t *vsecp,
- zuserns_t *mnt_ns);
+ zidmap_t *mnt_ns);
extern int zfs_setsecattr(znode_t *zp, vsecattr_t *vsecp, int flag,
cred_t *cr);
extern int zfs_write_simple(znode_t *zp, const void *data, size_t len,
loff_t pos, size_t *resid);
#endif
diff --git a/sys/contrib/openzfs/include/os/linux/kernel/linux/blkdev_compat.h b/sys/contrib/openzfs/include/os/linux/kernel/linux/blkdev_compat.h
index c7405ffab8ba..c5c6385be6ff 100644
--- a/sys/contrib/openzfs/include/os/linux/kernel/linux/blkdev_compat.h
+++ b/sys/contrib/openzfs/include/os/linux/kernel/linux/blkdev_compat.h
@@ -1,764 +1,764 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2011 Lawrence Livermore National Security, LLC.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
* Written by Brian Behlendorf <behlendorf1@llnl.gov>.
* LLNL-CODE-403049.
*/
#ifndef _ZFS_BLKDEV_H
#define _ZFS_BLKDEV_H
#include <linux/blkdev.h>
#include <linux/backing-dev.h>
#include <linux/hdreg.h>
#include <linux/major.h>
#include <linux/msdos_fs.h> /* for SECTOR_* */
#include <linux/bio.h>
#ifdef HAVE_BLK_MQ
#include <linux/blk-mq.h>
#endif
#ifndef HAVE_BLK_QUEUE_FLAG_SET
static inline void
blk_queue_flag_set(unsigned int flag, struct request_queue *q)
{
queue_flag_set(flag, q);
}
#endif
#ifndef HAVE_BLK_QUEUE_FLAG_CLEAR
static inline void
blk_queue_flag_clear(unsigned int flag, struct request_queue *q)
{
queue_flag_clear(flag, q);
}
#endif
/*
* 4.7 API,
* The blk_queue_write_cache() interface has replaced blk_queue_flush()
* interface. However, the new interface is GPL-only thus we implement
* our own trivial wrapper when the GPL-only version is detected.
*
* 2.6.36 - 4.6 API,
* The blk_queue_flush() interface has replaced blk_queue_ordered()
* interface. However, while the old interface was available to all the
* new one is GPL-only. Thus if the GPL-only version is detected we
* implement our own trivial helper.
*/
static inline void
blk_queue_set_write_cache(struct request_queue *q, bool wc, bool fua)
{
#if defined(HAVE_BLK_QUEUE_WRITE_CACHE_GPL_ONLY)
if (wc)
blk_queue_flag_set(QUEUE_FLAG_WC, q);
else
blk_queue_flag_clear(QUEUE_FLAG_WC, q);
if (fua)
blk_queue_flag_set(QUEUE_FLAG_FUA, q);
else
blk_queue_flag_clear(QUEUE_FLAG_FUA, q);
#elif defined(HAVE_BLK_QUEUE_WRITE_CACHE)
blk_queue_write_cache(q, wc, fua);
#elif defined(HAVE_BLK_QUEUE_FLUSH_GPL_ONLY)
if (wc)
q->flush_flags |= REQ_FLUSH;
if (fua)
q->flush_flags |= REQ_FUA;
#elif defined(HAVE_BLK_QUEUE_FLUSH)
blk_queue_flush(q, (wc ? REQ_FLUSH : 0) | (fua ? REQ_FUA : 0));
#else
#error "Unsupported kernel"
#endif
}
static inline void
blk_queue_set_read_ahead(struct request_queue *q, unsigned long ra_pages)
{
#if !defined(HAVE_BLK_QUEUE_UPDATE_READAHEAD) && \
!defined(HAVE_DISK_UPDATE_READAHEAD)
#ifdef HAVE_BLK_QUEUE_BDI_DYNAMIC
q->backing_dev_info->ra_pages = ra_pages;
#else
q->backing_dev_info.ra_pages = ra_pages;
#endif
#endif
}
#ifdef HAVE_BIO_BVEC_ITER
#define BIO_BI_SECTOR(bio) (bio)->bi_iter.bi_sector
#define BIO_BI_SIZE(bio) (bio)->bi_iter.bi_size
#define BIO_BI_IDX(bio) (bio)->bi_iter.bi_idx
#define BIO_BI_SKIP(bio) (bio)->bi_iter.bi_bvec_done
#define bio_for_each_segment4(bv, bvp, b, i) \
bio_for_each_segment((bv), (b), (i))
typedef struct bvec_iter bvec_iterator_t;
#else
#define BIO_BI_SECTOR(bio) (bio)->bi_sector
#define BIO_BI_SIZE(bio) (bio)->bi_size
#define BIO_BI_IDX(bio) (bio)->bi_idx
#define BIO_BI_SKIP(bio) (0)
#define bio_for_each_segment4(bv, bvp, b, i) \
bio_for_each_segment((bvp), (b), (i))
typedef int bvec_iterator_t;
#endif
static inline void
bio_set_flags_failfast(struct block_device *bdev, int *flags, bool dev,
bool transport, bool driver)
{
#ifdef CONFIG_BUG
/*
* Disable FAILFAST for loopback devices because of the
* following incorrect BUG_ON() in loop_make_request().
* This support is also disabled for md devices because the
* test suite layers md devices on top of loopback devices.
* This may be removed when the loopback driver is fixed.
*
* BUG_ON(!lo || (rw != READ && rw != WRITE));
*/
if ((MAJOR(bdev->bd_dev) == LOOP_MAJOR) ||
(MAJOR(bdev->bd_dev) == MD_MAJOR))
return;
#ifdef BLOCK_EXT_MAJOR
if (MAJOR(bdev->bd_dev) == BLOCK_EXT_MAJOR)
return;
#endif /* BLOCK_EXT_MAJOR */
#endif /* CONFIG_BUG */
if (dev)
*flags |= REQ_FAILFAST_DEV;
if (transport)
*flags |= REQ_FAILFAST_TRANSPORT;
if (driver)
*flags |= REQ_FAILFAST_DRIVER;
}
/*
* Maximum disk label length, it may be undefined for some kernels.
*/
#if !defined(DISK_NAME_LEN)
#define DISK_NAME_LEN 32
#endif /* DISK_NAME_LEN */
#ifdef HAVE_BIO_BI_STATUS
static inline int
bi_status_to_errno(blk_status_t status)
{
switch (status) {
case BLK_STS_OK:
return (0);
case BLK_STS_NOTSUPP:
return (EOPNOTSUPP);
case BLK_STS_TIMEOUT:
return (ETIMEDOUT);
case BLK_STS_NOSPC:
return (ENOSPC);
case BLK_STS_TRANSPORT:
return (ENOLINK);
case BLK_STS_TARGET:
return (EREMOTEIO);
case BLK_STS_NEXUS:
return (EBADE);
case BLK_STS_MEDIUM:
return (ENODATA);
case BLK_STS_PROTECTION:
return (EILSEQ);
case BLK_STS_RESOURCE:
return (ENOMEM);
case BLK_STS_AGAIN:
return (EAGAIN);
case BLK_STS_IOERR:
return (EIO);
default:
return (EIO);
}
}
static inline blk_status_t
errno_to_bi_status(int error)
{
switch (error) {
case 0:
return (BLK_STS_OK);
case EOPNOTSUPP:
return (BLK_STS_NOTSUPP);
case ETIMEDOUT:
return (BLK_STS_TIMEOUT);
case ENOSPC:
return (BLK_STS_NOSPC);
case ENOLINK:
return (BLK_STS_TRANSPORT);
case EREMOTEIO:
return (BLK_STS_TARGET);
case EBADE:
return (BLK_STS_NEXUS);
case ENODATA:
return (BLK_STS_MEDIUM);
case EILSEQ:
return (BLK_STS_PROTECTION);
case ENOMEM:
return (BLK_STS_RESOURCE);
case EAGAIN:
return (BLK_STS_AGAIN);
case EIO:
return (BLK_STS_IOERR);
default:
return (BLK_STS_IOERR);
}
}
#endif /* HAVE_BIO_BI_STATUS */
/*
* 4.3 API change
* The bio_endio() prototype changed slightly. These are helper
* macro's to ensure the prototype and invocation are handled.
*/
#ifdef HAVE_1ARG_BIO_END_IO_T
#ifdef HAVE_BIO_BI_STATUS
#define BIO_END_IO_ERROR(bio) bi_status_to_errno(bio->bi_status)
#define BIO_END_IO_PROTO(fn, x, z) static void fn(struct bio *x)
#define BIO_END_IO(bio, error) bio_set_bi_status(bio, error)
static inline void
bio_set_bi_status(struct bio *bio, int error)
{
ASSERT3S(error, <=, 0);
bio->bi_status = errno_to_bi_status(-error);
bio_endio(bio);
}
#else
#define BIO_END_IO_ERROR(bio) (-(bio->bi_error))
#define BIO_END_IO_PROTO(fn, x, z) static void fn(struct bio *x)
#define BIO_END_IO(bio, error) bio_set_bi_error(bio, error)
static inline void
bio_set_bi_error(struct bio *bio, int error)
{
ASSERT3S(error, <=, 0);
bio->bi_error = error;
bio_endio(bio);
}
#endif /* HAVE_BIO_BI_STATUS */
#else
#define BIO_END_IO_PROTO(fn, x, z) static void fn(struct bio *x, int z)
#define BIO_END_IO(bio, error) bio_endio(bio, error);
#endif /* HAVE_1ARG_BIO_END_IO_T */
/*
* 5.15 MACRO,
* GD_DEAD
*
* 2.6.36 - 5.14 MACRO,
* GENHD_FL_UP
*
* Check the disk status and return B_TRUE if alive
* otherwise B_FALSE
*/
static inline boolean_t
zfs_check_disk_status(struct block_device *bdev)
{
#if defined(GENHD_FL_UP)
return (!!(bdev->bd_disk->flags & GENHD_FL_UP));
#elif defined(GD_DEAD)
return (!test_bit(GD_DEAD, &bdev->bd_disk->state));
#else
/*
* This is encountered if neither GENHD_FL_UP nor GD_DEAD is available in
* the kernel - likely due to an MACRO change that needs to be chased down.
*/
#error "Unsupported kernel: no usable disk status check"
#endif
}
/*
* 4.1 API,
* 3.10.0 CentOS 7.x API,
* blkdev_reread_part()
*
* For older kernels trigger a re-reading of the partition table by calling
* check_disk_change() which calls flush_disk() to invalidate the device.
*
* For newer kernels (as of 5.10), bdev_check_media_change is used, in favor of
* check_disk_change(), with the modification that invalidation is no longer
* forced.
*/
#ifdef HAVE_CHECK_DISK_CHANGE
#define zfs_check_media_change(bdev) check_disk_change(bdev)
#ifdef HAVE_BLKDEV_REREAD_PART
#define vdev_bdev_reread_part(bdev) blkdev_reread_part(bdev)
#else
#define vdev_bdev_reread_part(bdev) check_disk_change(bdev)
#endif /* HAVE_BLKDEV_REREAD_PART */
#else
#ifdef HAVE_BDEV_CHECK_MEDIA_CHANGE
static inline int
zfs_check_media_change(struct block_device *bdev)
{
#ifdef HAVE_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK
struct gendisk *gd = bdev->bd_disk;
const struct block_device_operations *bdo = gd->fops;
#endif
if (!bdev_check_media_change(bdev))
return (0);
#ifdef HAVE_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK
/*
* Force revalidation, to mimic the old behavior of
* check_disk_change()
*/
if (bdo->revalidate_disk)
bdo->revalidate_disk(gd);
#endif
return (0);
}
#define vdev_bdev_reread_part(bdev) zfs_check_media_change(bdev)
#else
/*
* This is encountered if check_disk_change() and bdev_check_media_change()
* are not available in the kernel - likely due to an API change that needs
* to be chased down.
*/
#error "Unsupported kernel: no usable disk change check"
#endif /* HAVE_BDEV_CHECK_MEDIA_CHANGE */
#endif /* HAVE_CHECK_DISK_CHANGE */
/*
* 2.6.27 API change
* The function was exported for use, prior to this it existed but the
* symbol was not exported.
*
* 4.4.0-6.21 API change for Ubuntu
* lookup_bdev() gained a second argument, FMODE_*, to check inode permissions.
*
* 5.11 API change
* Changed to take a dev_t argument which is set on success and return a
* non-zero error code on failure.
*/
static inline int
vdev_lookup_bdev(const char *path, dev_t *dev)
{
#if defined(HAVE_DEVT_LOOKUP_BDEV)
return (lookup_bdev(path, dev));
#elif defined(HAVE_1ARG_LOOKUP_BDEV)
struct block_device *bdev = lookup_bdev(path);
if (IS_ERR(bdev))
return (PTR_ERR(bdev));
*dev = bdev->bd_dev;
bdput(bdev);
return (0);
#elif defined(HAVE_MODE_LOOKUP_BDEV)
struct block_device *bdev = lookup_bdev(path, FMODE_READ);
if (IS_ERR(bdev))
return (PTR_ERR(bdev));
*dev = bdev->bd_dev;
bdput(bdev);
return (0);
#else
#error "Unsupported kernel"
#endif
}
/*
* Kernels without bio_set_op_attrs use bi_rw for the bio flags.
*/
#if !defined(HAVE_BIO_SET_OP_ATTRS)
static inline void
bio_set_op_attrs(struct bio *bio, unsigned rw, unsigned flags)
{
#if defined(HAVE_BIO_BI_OPF)
bio->bi_opf = rw | flags;
#else
bio->bi_rw |= rw | flags;
#endif /* HAVE_BIO_BI_OPF */
}
#endif
/*
* bio_set_flush - Set the appropriate flags in a bio to guarantee
* data are on non-volatile media on completion.
*
* 2.6.37 - 4.8 API,
* Introduce WRITE_FLUSH, WRITE_FUA, and WRITE_FLUSH_FUA flags as a
* replacement for WRITE_BARRIER to allow expressing richer semantics
* to the block layer. It's up to the block layer to implement the
* semantics correctly. Use the WRITE_FLUSH_FUA flag combination.
*
* 4.8 - 4.9 API,
* REQ_FLUSH was renamed to REQ_PREFLUSH. For consistency with previous
* OpenZFS releases, prefer the WRITE_FLUSH_FUA flag set if it's available.
*
* 4.10 API,
* The read/write flags and their modifiers, including WRITE_FLUSH,
* WRITE_FUA and WRITE_FLUSH_FUA were removed from fs.h in
* torvalds/linux@70fd7614 and replaced by direct flag modification
* of the REQ_ flags in bio->bi_opf. Use REQ_PREFLUSH.
*/
static inline void
bio_set_flush(struct bio *bio)
{
#if defined(HAVE_REQ_PREFLUSH) /* >= 4.10 */
- bio_set_op_attrs(bio, 0, REQ_PREFLUSH);
+ bio_set_op_attrs(bio, 0, REQ_PREFLUSH | REQ_OP_WRITE);
#elif defined(WRITE_FLUSH_FUA) /* >= 2.6.37 and <= 4.9 */
bio_set_op_attrs(bio, 0, WRITE_FLUSH_FUA);
#else
#error "Allowing the build will cause bio_set_flush requests to be ignored."
#endif
}
/*
* 4.8 API,
* REQ_OP_FLUSH
*
* 4.8-rc0 - 4.8-rc1,
* REQ_PREFLUSH
*
* 2.6.36 - 4.7 API,
* REQ_FLUSH
*
* in all cases but may have a performance impact for some kernels. It
* has the advantage of minimizing kernel specific changes in the zvol code.
*
*/
static inline boolean_t
bio_is_flush(struct bio *bio)
{
#if defined(HAVE_REQ_OP_FLUSH) && defined(HAVE_BIO_BI_OPF)
return ((bio_op(bio) == REQ_OP_FLUSH) || (bio->bi_opf & REQ_PREFLUSH));
#elif defined(HAVE_REQ_PREFLUSH) && defined(HAVE_BIO_BI_OPF)
return (bio->bi_opf & REQ_PREFLUSH);
#elif defined(HAVE_REQ_PREFLUSH) && !defined(HAVE_BIO_BI_OPF)
return (bio->bi_rw & REQ_PREFLUSH);
#elif defined(HAVE_REQ_FLUSH)
return (bio->bi_rw & REQ_FLUSH);
#else
#error "Unsupported kernel"
#endif
}
/*
* 4.8 API,
* REQ_FUA flag moved to bio->bi_opf
*
* 2.6.x - 4.7 API,
* REQ_FUA
*/
static inline boolean_t
bio_is_fua(struct bio *bio)
{
#if defined(HAVE_BIO_BI_OPF)
return (bio->bi_opf & REQ_FUA);
#elif defined(REQ_FUA)
return (bio->bi_rw & REQ_FUA);
#else
#error "Allowing the build will cause fua requests to be ignored."
#endif
}
/*
* 4.8 API,
* REQ_OP_DISCARD
*
* 2.6.36 - 4.7 API,
* REQ_DISCARD
*
* In all cases the normal I/O path is used for discards. The only
* difference is how the kernel tags individual I/Os as discards.
*/
static inline boolean_t
bio_is_discard(struct bio *bio)
{
#if defined(HAVE_REQ_OP_DISCARD)
return (bio_op(bio) == REQ_OP_DISCARD);
#elif defined(HAVE_REQ_DISCARD)
return (bio->bi_rw & REQ_DISCARD);
#else
#error "Unsupported kernel"
#endif
}
/*
* 4.8 API,
* REQ_OP_SECURE_ERASE
*
* 2.6.36 - 4.7 API,
* REQ_SECURE
*/
static inline boolean_t
bio_is_secure_erase(struct bio *bio)
{
#if defined(HAVE_REQ_OP_SECURE_ERASE)
return (bio_op(bio) == REQ_OP_SECURE_ERASE);
#elif defined(REQ_SECURE)
return (bio->bi_rw & REQ_SECURE);
#else
return (0);
#endif
}
/*
* 2.6.33 API change
* Discard granularity and alignment restrictions may now be set. For
* older kernels which do not support this it is safe to skip it.
*/
static inline void
blk_queue_discard_granularity(struct request_queue *q, unsigned int dg)
{
q->limits.discard_granularity = dg;
}
/*
* 5.19 API,
* bdev_max_discard_sectors()
*
* 2.6.32 API,
* blk_queue_discard()
*/
static inline boolean_t
bdev_discard_supported(struct block_device *bdev)
{
#if defined(HAVE_BDEV_MAX_DISCARD_SECTORS)
return (!!bdev_max_discard_sectors(bdev));
#elif defined(HAVE_BLK_QUEUE_DISCARD)
return (!!blk_queue_discard(bdev_get_queue(bdev)));
#else
#error "Unsupported kernel"
#endif
}
/*
* 5.19 API,
* bdev_max_secure_erase_sectors()
*
* 4.8 API,
* blk_queue_secure_erase()
*
* 2.6.36 - 4.7 API,
* blk_queue_secdiscard()
*/
static inline boolean_t
bdev_secure_discard_supported(struct block_device *bdev)
{
#if defined(HAVE_BDEV_MAX_SECURE_ERASE_SECTORS)
return (!!bdev_max_secure_erase_sectors(bdev));
#elif defined(HAVE_BLK_QUEUE_SECURE_ERASE)
return (!!blk_queue_secure_erase(bdev_get_queue(bdev)));
#elif defined(HAVE_BLK_QUEUE_SECDISCARD)
return (!!blk_queue_secdiscard(bdev_get_queue(bdev)));
#else
#error "Unsupported kernel"
#endif
}
/*
* A common holder for vdev_bdev_open() is used to relax the exclusive open
* semantics slightly. Internal vdev disk callers may pass VDEV_HOLDER to
* allow them to open the device multiple times. Other kernel callers and
* user space processes which don't pass this value will get EBUSY. This is
* currently required for the correct operation of hot spares.
*/
#define VDEV_HOLDER ((void *)0x2401de7)
static inline unsigned long
blk_generic_start_io_acct(struct request_queue *q __attribute__((unused)),
struct gendisk *disk __attribute__((unused)),
int rw __attribute__((unused)), struct bio *bio)
{
#if defined(HAVE_BDEV_IO_ACCT_63)
return (bdev_start_io_acct(bio->bi_bdev, bio_op(bio),
jiffies));
#elif defined(HAVE_BDEV_IO_ACCT_OLD)
return (bdev_start_io_acct(bio->bi_bdev, bio_sectors(bio),
bio_op(bio), jiffies));
#elif defined(HAVE_DISK_IO_ACCT)
return (disk_start_io_acct(disk, bio_sectors(bio), bio_op(bio)));
#elif defined(HAVE_BIO_IO_ACCT)
return (bio_start_io_acct(bio));
#elif defined(HAVE_GENERIC_IO_ACCT_3ARG)
unsigned long start_time = jiffies;
generic_start_io_acct(rw, bio_sectors(bio), &disk->part0);
return (start_time);
#elif defined(HAVE_GENERIC_IO_ACCT_4ARG)
unsigned long start_time = jiffies;
generic_start_io_acct(q, rw, bio_sectors(bio), &disk->part0);
return (start_time);
#else
/* Unsupported */
return (0);
#endif
}
static inline void
blk_generic_end_io_acct(struct request_queue *q __attribute__((unused)),
struct gendisk *disk __attribute__((unused)),
int rw __attribute__((unused)), struct bio *bio, unsigned long start_time)
{
#if defined(HAVE_BDEV_IO_ACCT_63)
bdev_end_io_acct(bio->bi_bdev, bio_op(bio), bio_sectors(bio),
start_time);
#elif defined(HAVE_BDEV_IO_ACCT_OLD)
bdev_end_io_acct(bio->bi_bdev, bio_op(bio), start_time);
#elif defined(HAVE_DISK_IO_ACCT)
disk_end_io_acct(disk, bio_op(bio), start_time);
#elif defined(HAVE_BIO_IO_ACCT)
bio_end_io_acct(bio, start_time);
#elif defined(HAVE_GENERIC_IO_ACCT_3ARG)
generic_end_io_acct(rw, &disk->part0, start_time);
#elif defined(HAVE_GENERIC_IO_ACCT_4ARG)
generic_end_io_acct(q, rw, &disk->part0, start_time);
#endif
}
#ifndef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS
static inline struct request_queue *
blk_generic_alloc_queue(make_request_fn make_request, int node_id)
{
#if defined(HAVE_BLK_ALLOC_QUEUE_REQUEST_FN)
return (blk_alloc_queue(make_request, node_id));
#elif defined(HAVE_BLK_ALLOC_QUEUE_REQUEST_FN_RH)
return (blk_alloc_queue_rh(make_request, node_id));
#else
struct request_queue *q = blk_alloc_queue(GFP_KERNEL);
if (q != NULL)
blk_queue_make_request(q, make_request);
return (q);
#endif
}
#endif /* !HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS */
/*
* All the io_*() helper functions below can operate on a bio, or a rq, but
* not both. The older submit_bio() codepath will pass a bio, and the
* newer blk-mq codepath will pass a rq.
*/
static inline int
io_data_dir(struct bio *bio, struct request *rq)
{
#ifdef HAVE_BLK_MQ
if (rq != NULL) {
if (op_is_write(req_op(rq))) {
return (WRITE);
} else {
return (READ);
}
}
#else
ASSERT3P(rq, ==, NULL);
#endif
return (bio_data_dir(bio));
}
static inline int
io_is_flush(struct bio *bio, struct request *rq)
{
#ifdef HAVE_BLK_MQ
if (rq != NULL)
return (req_op(rq) == REQ_OP_FLUSH);
#else
ASSERT3P(rq, ==, NULL);
#endif
return (bio_is_flush(bio));
}
static inline int
io_is_discard(struct bio *bio, struct request *rq)
{
#ifdef HAVE_BLK_MQ
if (rq != NULL)
return (req_op(rq) == REQ_OP_DISCARD);
#else
ASSERT3P(rq, ==, NULL);
#endif
return (bio_is_discard(bio));
}
static inline int
io_is_secure_erase(struct bio *bio, struct request *rq)
{
#ifdef HAVE_BLK_MQ
if (rq != NULL)
return (req_op(rq) == REQ_OP_SECURE_ERASE);
#else
ASSERT3P(rq, ==, NULL);
#endif
return (bio_is_secure_erase(bio));
}
static inline int
io_is_fua(struct bio *bio, struct request *rq)
{
#ifdef HAVE_BLK_MQ
if (rq != NULL)
return (rq->cmd_flags & REQ_FUA);
#else
ASSERT3P(rq, ==, NULL);
#endif
return (bio_is_fua(bio));
}
static inline uint64_t
io_offset(struct bio *bio, struct request *rq)
{
#ifdef HAVE_BLK_MQ
if (rq != NULL)
return (blk_rq_pos(rq) << 9);
#else
ASSERT3P(rq, ==, NULL);
#endif
return (BIO_BI_SECTOR(bio) << 9);
}
static inline uint64_t
io_size(struct bio *bio, struct request *rq)
{
#ifdef HAVE_BLK_MQ
if (rq != NULL)
return (blk_rq_bytes(rq));
#else
ASSERT3P(rq, ==, NULL);
#endif
return (BIO_BI_SIZE(bio));
}
static inline int
io_has_data(struct bio *bio, struct request *rq)
{
#ifdef HAVE_BLK_MQ
if (rq != NULL)
return (bio_has_data(rq->bio));
#else
ASSERT3P(rq, ==, NULL);
#endif
return (bio_has_data(bio));
}
#endif /* _ZFS_BLKDEV_H */
diff --git a/sys/contrib/openzfs/include/os/linux/kernel/linux/vfs_compat.h b/sys/contrib/openzfs/include/os/linux/kernel/linux/vfs_compat.h
index fd0b9e8e1068..e156ed41c28c 100644
--- a/sys/contrib/openzfs/include/os/linux/kernel/linux/vfs_compat.h
+++ b/sys/contrib/openzfs/include/os/linux/kernel/linux/vfs_compat.h
@@ -1,458 +1,475 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2011 Lawrence Livermore National Security, LLC.
* Copyright (C) 2015 Jörg Thalheim.
*/
#ifndef _ZFS_VFS_H
#define _ZFS_VFS_H
#include <sys/taskq.h>
#include <sys/cred.h>
#include <linux/backing-dev.h>
#include <linux/compat.h>
/*
* 2.6.34 - 3.19, bdi_setup_and_register() takes 3 arguments.
* 4.0 - 4.11, bdi_setup_and_register() takes 2 arguments.
* 4.12 - x.y, super_setup_bdi_name() new interface.
*/
#if defined(HAVE_SUPER_SETUP_BDI_NAME)
extern atomic_long_t zfs_bdi_seq;
static inline int
zpl_bdi_setup(struct super_block *sb, char *name)
{
return super_setup_bdi_name(sb, "%.28s-%ld", name,
atomic_long_inc_return(&zfs_bdi_seq));
}
static inline void
zpl_bdi_destroy(struct super_block *sb)
{
}
#elif defined(HAVE_2ARGS_BDI_SETUP_AND_REGISTER)
static inline int
zpl_bdi_setup(struct super_block *sb, char *name)
{
struct backing_dev_info *bdi;
int error;
bdi = kmem_zalloc(sizeof (struct backing_dev_info), KM_SLEEP);
error = bdi_setup_and_register(bdi, name);
if (error) {
kmem_free(bdi, sizeof (struct backing_dev_info));
return (error);
}
sb->s_bdi = bdi;
return (0);
}
static inline void
zpl_bdi_destroy(struct super_block *sb)
{
struct backing_dev_info *bdi = sb->s_bdi;
bdi_destroy(bdi);
kmem_free(bdi, sizeof (struct backing_dev_info));
sb->s_bdi = NULL;
}
#elif defined(HAVE_3ARGS_BDI_SETUP_AND_REGISTER)
static inline int
zpl_bdi_setup(struct super_block *sb, char *name)
{
struct backing_dev_info *bdi;
int error;
bdi = kmem_zalloc(sizeof (struct backing_dev_info), KM_SLEEP);
error = bdi_setup_and_register(bdi, name, BDI_CAP_MAP_COPY);
if (error) {
kmem_free(sb->s_bdi, sizeof (struct backing_dev_info));
return (error);
}
sb->s_bdi = bdi;
return (0);
}
static inline void
zpl_bdi_destroy(struct super_block *sb)
{
struct backing_dev_info *bdi = sb->s_bdi;
bdi_destroy(bdi);
kmem_free(bdi, sizeof (struct backing_dev_info));
sb->s_bdi = NULL;
}
#else
#error "Unsupported kernel"
#endif
/*
* 4.14 adds SB_* flag definitions, define them to MS_* equivalents
* if not set.
*/
#ifndef SB_RDONLY
#define SB_RDONLY MS_RDONLY
#endif
#ifndef SB_SILENT
#define SB_SILENT MS_SILENT
#endif
#ifndef SB_ACTIVE
#define SB_ACTIVE MS_ACTIVE
#endif
#ifndef SB_POSIXACL
#define SB_POSIXACL MS_POSIXACL
#endif
#ifndef SB_MANDLOCK
#define SB_MANDLOCK MS_MANDLOCK
#endif
#ifndef SB_NOATIME
#define SB_NOATIME MS_NOATIME
#endif
/*
* 3.5 API change,
* The clear_inode() function replaces end_writeback() and introduces an
* ordering change regarding when the inode_sync_wait() occurs. See the
* configure check in config/kernel-clear-inode.m4 for full details.
*/
#if defined(HAVE_EVICT_INODE) && !defined(HAVE_CLEAR_INODE)
#define clear_inode(ip) end_writeback(ip)
#endif /* HAVE_EVICT_INODE && !HAVE_CLEAR_INODE */
#if defined(SEEK_HOLE) && defined(SEEK_DATA) && !defined(HAVE_LSEEK_EXECUTE)
static inline loff_t
lseek_execute(
struct file *filp,
struct inode *inode,
loff_t offset,
loff_t maxsize)
{
if (offset < 0 && !(filp->f_mode & FMODE_UNSIGNED_OFFSET))
return (-EINVAL);
if (offset > maxsize)
return (-EINVAL);
if (offset != filp->f_pos) {
spin_lock(&filp->f_lock);
filp->f_pos = offset;
filp->f_version = 0;
spin_unlock(&filp->f_lock);
}
return (offset);
}
#endif /* SEEK_HOLE && SEEK_DATA && !HAVE_LSEEK_EXECUTE */
#if defined(CONFIG_FS_POSIX_ACL)
/*
* These functions safely approximates the behavior of posix_acl_release()
* which cannot be used because it calls the GPL-only symbol kfree_rcu().
* The in-kernel version, which can access the RCU, frees the ACLs after
* the grace period expires. Because we're unsure how long that grace
* period may be this implementation conservatively delays for 60 seconds.
* This is several orders of magnitude larger than expected grace period.
* At 60 seconds the kernel will also begin issuing RCU stall warnings.
*/
#include <linux/posix_acl.h>
#if defined(HAVE_POSIX_ACL_RELEASE) && !defined(HAVE_POSIX_ACL_RELEASE_GPL_ONLY)
#define zpl_posix_acl_release(arg) posix_acl_release(arg)
#else
void zpl_posix_acl_release_impl(struct posix_acl *);
static inline void
zpl_posix_acl_release(struct posix_acl *acl)
{
if ((acl == NULL) || (acl == ACL_NOT_CACHED))
return;
#ifdef HAVE_ACL_REFCOUNT
if (refcount_dec_and_test(&acl->a_refcount))
zpl_posix_acl_release_impl(acl);
#else
if (atomic_dec_and_test(&acl->a_refcount))
zpl_posix_acl_release_impl(acl);
#endif
}
#endif /* HAVE_POSIX_ACL_RELEASE */
#ifdef HAVE_SET_CACHED_ACL_USABLE
#define zpl_set_cached_acl(ip, ty, n) set_cached_acl(ip, ty, n)
#define zpl_forget_cached_acl(ip, ty) forget_cached_acl(ip, ty)
#else
static inline void
zpl_set_cached_acl(struct inode *ip, int type, struct posix_acl *newer)
{
struct posix_acl *older = NULL;
spin_lock(&ip->i_lock);
if ((newer != ACL_NOT_CACHED) && (newer != NULL))
posix_acl_dup(newer);
switch (type) {
case ACL_TYPE_ACCESS:
older = ip->i_acl;
rcu_assign_pointer(ip->i_acl, newer);
break;
case ACL_TYPE_DEFAULT:
older = ip->i_default_acl;
rcu_assign_pointer(ip->i_default_acl, newer);
break;
}
spin_unlock(&ip->i_lock);
zpl_posix_acl_release(older);
}
static inline void
zpl_forget_cached_acl(struct inode *ip, int type)
{
zpl_set_cached_acl(ip, type, (struct posix_acl *)ACL_NOT_CACHED);
}
#endif /* HAVE_SET_CACHED_ACL_USABLE */
/*
* 3.1 API change,
* posix_acl_chmod() was added as the preferred interface.
*
* 3.14 API change,
* posix_acl_chmod() was changed to __posix_acl_chmod()
*/
#ifndef HAVE___POSIX_ACL_CHMOD
#ifdef HAVE_POSIX_ACL_CHMOD
#define __posix_acl_chmod(acl, gfp, mode) posix_acl_chmod(acl, gfp, mode)
#define __posix_acl_create(acl, gfp, mode) posix_acl_create(acl, gfp, mode)
#else
#error "Unsupported kernel"
#endif /* HAVE_POSIX_ACL_CHMOD */
#endif /* HAVE___POSIX_ACL_CHMOD */
/*
* 4.8 API change,
* posix_acl_valid() now must be passed a namespace, the namespace from
* from super block associated with the given inode is used for this purpose.
*/
#ifdef HAVE_POSIX_ACL_VALID_WITH_NS
#define zpl_posix_acl_valid(ip, acl) posix_acl_valid(ip->i_sb->s_user_ns, acl)
#else
#define zpl_posix_acl_valid(ip, acl) posix_acl_valid(acl)
#endif
#endif /* CONFIG_FS_POSIX_ACL */
/*
* 3.19 API change
* struct access f->f_dentry->d_inode was replaced by accessor function
* file_inode(f)
*/
#ifndef HAVE_FILE_INODE
static inline struct inode *file_inode(const struct file *f)
{
return (f->f_dentry->d_inode);
}
#endif /* HAVE_FILE_INODE */
/*
* 4.1 API change
* struct access file->f_path.dentry was replaced by accessor function
* file_dentry(f)
*/
#ifndef HAVE_FILE_DENTRY
static inline struct dentry *file_dentry(const struct file *f)
{
return (f->f_path.dentry);
}
#endif /* HAVE_FILE_DENTRY */
static inline uid_t zfs_uid_read_impl(struct inode *ip)
{
return (from_kuid(kcred->user_ns, ip->i_uid));
}
static inline uid_t zfs_uid_read(struct inode *ip)
{
return (zfs_uid_read_impl(ip));
}
static inline gid_t zfs_gid_read_impl(struct inode *ip)
{
return (from_kgid(kcred->user_ns, ip->i_gid));
}
static inline gid_t zfs_gid_read(struct inode *ip)
{
return (zfs_gid_read_impl(ip));
}
static inline void zfs_uid_write(struct inode *ip, uid_t uid)
{
ip->i_uid = make_kuid(kcred->user_ns, uid);
}
static inline void zfs_gid_write(struct inode *ip, gid_t gid)
{
ip->i_gid = make_kgid(kcred->user_ns, gid);
}
/*
* 3.15 API change
*/
#ifndef RENAME_NOREPLACE
#define RENAME_NOREPLACE (1 << 0) /* Don't overwrite target */
#endif
#ifndef RENAME_EXCHANGE
#define RENAME_EXCHANGE (1 << 1) /* Exchange source and dest */
#endif
#ifndef RENAME_WHITEOUT
#define RENAME_WHITEOUT (1 << 2) /* Whiteout source */
#endif
/*
* 4.9 API change
*/
#if !(defined(HAVE_SETATTR_PREPARE_NO_USERNS) || \
- defined(HAVE_SETATTR_PREPARE_USERNS))
+ defined(HAVE_SETATTR_PREPARE_USERNS) || \
+ defined(HAVE_SETATTR_PREPARE_IDMAP))
static inline int
setattr_prepare(struct dentry *dentry, struct iattr *ia)
{
return (inode_change_ok(dentry->d_inode, ia));
}
#endif
/*
* 4.11 API change
* These macros are defined by kernel 4.11. We define them so that the same
* code builds under kernels < 4.11 and >= 4.11. The macros are set to 0 so
* that it will create obvious failures if they are accidentally used when built
* against a kernel >= 4.11.
*/
#ifndef STATX_BASIC_STATS
#define STATX_BASIC_STATS 0
#endif
#ifndef AT_STATX_SYNC_AS_STAT
#define AT_STATX_SYNC_AS_STAT 0
#endif
/*
* 4.11 API change
* 4.11 takes struct path *, < 4.11 takes vfsmount *
*/
#ifdef HAVE_VFSMOUNT_IOPS_GETATTR
#define ZPL_GETATTR_WRAPPER(func) \
static int \
func(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) \
{ \
struct path path = { .mnt = mnt, .dentry = dentry }; \
return func##_impl(&path, stat, STATX_BASIC_STATS, \
AT_STATX_SYNC_AS_STAT); \
}
#elif defined(HAVE_PATH_IOPS_GETATTR)
#define ZPL_GETATTR_WRAPPER(func) \
static int \
func(const struct path *path, struct kstat *stat, u32 request_mask, \
unsigned int query_flags) \
{ \
return (func##_impl(path, stat, request_mask, query_flags)); \
}
#elif defined(HAVE_USERNS_IOPS_GETATTR)
#define ZPL_GETATTR_WRAPPER(func) \
static int \
func(struct user_namespace *user_ns, const struct path *path, \
struct kstat *stat, u32 request_mask, unsigned int query_flags) \
{ \
return (func##_impl(user_ns, path, stat, request_mask, \
query_flags)); \
}
+#elif defined(HAVE_IDMAP_IOPS_GETATTR)
+#define ZPL_GETATTR_WRAPPER(func) \
+static int \
+func(struct mnt_idmap *user_ns, const struct path *path, \
+ struct kstat *stat, u32 request_mask, unsigned int query_flags) \
+{ \
+ return (func##_impl(user_ns, path, stat, request_mask, \
+ query_flags)); \
+}
#else
#error
#endif
/*
* 4.9 API change
* Preferred interface to get the current FS time.
*/
#if !defined(HAVE_CURRENT_TIME)
static inline struct timespec
current_time(struct inode *ip)
{
return (timespec_trunc(current_kernel_time(), ip->i_sb->s_time_gran));
}
#endif
/*
* 4.16 API change
* Added iversion interface for managing inode version field.
*/
#ifdef HAVE_INODE_SET_IVERSION
#include <linux/iversion.h>
#else
static inline void
inode_set_iversion(struct inode *ip, u64 val)
{
ip->i_version = val;
}
#endif
/*
* Returns true when called in the context of a 32-bit system call.
*/
static inline int
zpl_is_32bit_api(void)
{
#ifdef CONFIG_COMPAT
#ifdef HAVE_IN_COMPAT_SYSCALL
return (in_compat_syscall());
#else
return (is_compat_task());
#endif
#else
return (BITS_PER_LONG == 32);
#endif
}
/*
* 5.12 API change
* To support id-mapped mounts, generic_fillattr() was modified to
* accept a new struct user_namespace* as its first arg.
+ *
+ * 6.3 API change
+ * generic_fillattr() first arg is changed to struct mnt_idmap *
+ *
*/
-#ifdef HAVE_GENERIC_FILLATTR_USERNS
+#ifdef HAVE_GENERIC_FILLATTR_IDMAP
+#define zpl_generic_fillattr(idmap, ip, sp) \
+ generic_fillattr(idmap, ip, sp)
+#elif defined(HAVE_GENERIC_FILLATTR_USERNS)
#define zpl_generic_fillattr(user_ns, ip, sp) \
generic_fillattr(user_ns, ip, sp)
#else
#define zpl_generic_fillattr(user_ns, ip, sp) generic_fillattr(ip, sp)
#endif
#endif /* _ZFS_VFS_H */
diff --git a/sys/contrib/openzfs/include/os/linux/kernel/linux/xattr_compat.h b/sys/contrib/openzfs/include/os/linux/kernel/linux/xattr_compat.h
index ff80fbb06413..bcc7289ad857 100644
--- a/sys/contrib/openzfs/include/os/linux/kernel/linux/xattr_compat.h
+++ b/sys/contrib/openzfs/include/os/linux/kernel/linux/xattr_compat.h
@@ -1,215 +1,230 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2011 Lawrence Livermore National Security, LLC.
*/
#ifndef _ZFS_XATTR_H
#define _ZFS_XATTR_H
#include <linux/posix_acl_xattr.h>
/*
* 2.6.35 API change,
* The const keyword was added to the 'struct xattr_handler' in the
* generic Linux super_block structure. To handle this we define an
* appropriate xattr_handler_t typedef which can be used. This was
* the preferred solution because it keeps the code clean and readable.
*/
typedef const struct xattr_handler xattr_handler_t;
/*
* 4.5 API change,
*/
#if defined(HAVE_XATTR_LIST_SIMPLE)
#define ZPL_XATTR_LIST_WRAPPER(fn) \
static bool \
fn(struct dentry *dentry) \
{ \
return (!!__ ## fn(dentry->d_inode, NULL, 0, NULL, 0)); \
}
/*
* 4.4 API change,
*/
#elif defined(HAVE_XATTR_LIST_DENTRY)
#define ZPL_XATTR_LIST_WRAPPER(fn) \
static size_t \
fn(struct dentry *dentry, char *list, size_t list_size, \
const char *name, size_t name_len, int type) \
{ \
return (__ ## fn(dentry->d_inode, \
list, list_size, name, name_len)); \
}
/*
* 2.6.33 API change,
*/
#elif defined(HAVE_XATTR_LIST_HANDLER)
#define ZPL_XATTR_LIST_WRAPPER(fn) \
static size_t \
fn(const struct xattr_handler *handler, struct dentry *dentry, \
char *list, size_t list_size, const char *name, size_t name_len) \
{ \
return (__ ## fn(dentry->d_inode, \
list, list_size, name, name_len)); \
}
#else
#error "Unsupported kernel"
#endif
/*
* 4.7 API change,
* The xattr_handler->get() callback was changed to take a both dentry and
* inode, because the dentry might not be attached to an inode yet.
*/
#if defined(HAVE_XATTR_GET_DENTRY_INODE)
#define ZPL_XATTR_GET_WRAPPER(fn) \
static int \
fn(const struct xattr_handler *handler, struct dentry *dentry, \
struct inode *inode, const char *name, void *buffer, size_t size) \
{ \
return (__ ## fn(inode, name, buffer, size)); \
}
/*
* 4.4 API change,
* The xattr_handler->get() callback was changed to take a xattr_handler,
* and handler_flags argument was removed and should be accessed by
* handler->flags.
*/
#elif defined(HAVE_XATTR_GET_HANDLER)
#define ZPL_XATTR_GET_WRAPPER(fn) \
static int \
fn(const struct xattr_handler *handler, struct dentry *dentry, \
const char *name, void *buffer, size_t size) \
{ \
return (__ ## fn(dentry->d_inode, name, buffer, size)); \
}
/*
* 2.6.33 API change,
* The xattr_handler->get() callback was changed to take a dentry
* instead of an inode, and a handler_flags argument was added.
*/
#elif defined(HAVE_XATTR_GET_DENTRY)
#define ZPL_XATTR_GET_WRAPPER(fn) \
static int \
fn(struct dentry *dentry, const char *name, void *buffer, size_t size, \
int unused_handler_flags) \
{ \
return (__ ## fn(dentry->d_inode, name, buffer, size)); \
}
/*
* Android API change,
* The xattr_handler->get() callback was changed to take a dentry and inode
* and flags, because the dentry might not be attached to an inode yet.
*/
#elif defined(HAVE_XATTR_GET_DENTRY_INODE_FLAGS)
#define ZPL_XATTR_GET_WRAPPER(fn) \
static int \
fn(const struct xattr_handler *handler, struct dentry *dentry, \
struct inode *inode, const char *name, void *buffer, \
size_t size, int flags) \
{ \
return (__ ## fn(inode, name, buffer, size)); \
}
#else
#error "Unsupported kernel"
#endif
+/*
+ * 6.3 API change,
+ * The xattr_handler->set() callback was changed to take the
+ * struct mnt_idmap* as the first arg, to support idmapped
+ * mounts.
+ */
+#if defined(HAVE_XATTR_SET_IDMAP)
+#define ZPL_XATTR_SET_WRAPPER(fn) \
+static int \
+fn(const struct xattr_handler *handler, struct mnt_idmap *user_ns, \
+ struct dentry *dentry, struct inode *inode, const char *name, \
+ const void *buffer, size_t size, int flags) \
+{ \
+ return (__ ## fn(user_ns, inode, name, buffer, size, flags)); \
+}
/*
* 5.12 API change,
* The xattr_handler->set() callback was changed to take the
* struct user_namespace* as the first arg, to support idmapped
* mounts.
*/
-#if defined(HAVE_XATTR_SET_USERNS)
+#elif defined(HAVE_XATTR_SET_USERNS)
#define ZPL_XATTR_SET_WRAPPER(fn) \
static int \
fn(const struct xattr_handler *handler, struct user_namespace *user_ns, \
struct dentry *dentry, struct inode *inode, const char *name, \
const void *buffer, size_t size, int flags) \
{ \
return (__ ## fn(user_ns, inode, name, buffer, size, flags)); \
}
/*
* 4.7 API change,
* The xattr_handler->set() callback was changed to take a both dentry and
* inode, because the dentry might not be attached to an inode yet.
*/
#elif defined(HAVE_XATTR_SET_DENTRY_INODE)
#define ZPL_XATTR_SET_WRAPPER(fn) \
static int \
fn(const struct xattr_handler *handler, struct dentry *dentry, \
struct inode *inode, const char *name, const void *buffer, \
size_t size, int flags) \
{ \
return (__ ## fn(kcred->user_ns, inode, name, buffer, size, flags));\
}
/*
* 4.4 API change,
* The xattr_handler->set() callback was changed to take a xattr_handler,
* and handler_flags argument was removed and should be accessed by
* handler->flags.
*/
#elif defined(HAVE_XATTR_SET_HANDLER)
#define ZPL_XATTR_SET_WRAPPER(fn) \
static int \
fn(const struct xattr_handler *handler, struct dentry *dentry, \
const char *name, const void *buffer, size_t size, int flags) \
{ \
return (__ ## fn(kcred->user_ns, dentry->d_inode, name, \
buffer, size, flags)); \
}
/*
* 2.6.33 API change,
* The xattr_handler->set() callback was changed to take a dentry
* instead of an inode, and a handler_flags argument was added.
*/
#elif defined(HAVE_XATTR_SET_DENTRY)
#define ZPL_XATTR_SET_WRAPPER(fn) \
static int \
fn(struct dentry *dentry, const char *name, const void *buffer, \
size_t size, int flags, int unused_handler_flags) \
{ \
return (__ ## fn(kcred->user_ns, dentry->d_inode, name, \
buffer, size, flags)); \
}
#else
#error "Unsupported kernel"
#endif
/*
* Linux 3.7 API change. posix_acl_{from,to}_xattr gained the user_ns
* parameter. All callers are expected to pass the &init_user_ns which
* is available through the init credential (kcred).
*/
static inline struct posix_acl *
zpl_acl_from_xattr(const void *value, int size)
{
return (posix_acl_from_xattr(kcred->user_ns, value, size));
}
static inline int
zpl_acl_to_xattr(struct posix_acl *acl, void *value, int size)
{
return (posix_acl_to_xattr(kcred->user_ns, acl, value, size));
}
#endif /* _ZFS_XATTR_H */
diff --git a/sys/contrib/openzfs/include/os/linux/spl/sys/cred.h b/sys/contrib/openzfs/include/os/linux/spl/sys/cred.h
index 75ad400d312d..7fd5f644863f 100644
--- a/sys/contrib/openzfs/include/os/linux/spl/sys/cred.h
+++ b/sys/contrib/openzfs/include/os/linux/spl/sys/cred.h
@@ -1,135 +1,141 @@
/*
* Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
* Copyright (C) 2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
* Written by Brian Behlendorf <behlendorf1@llnl.gov>.
* UCRL-CODE-235197
*
* This file is part of the SPL, Solaris Porting Layer.
*
* The SPL is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation; either version 2 of the License, or (at your
* option) any later version.
*
* The SPL is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* for more details.
*
* You should have received a copy of the GNU General Public License along
* with the SPL. If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef _SPL_CRED_H
#define _SPL_CRED_H
#include <linux/module.h>
#include <linux/cred.h>
#include <linux/sched.h>
#include <sys/types.h>
#include <sys/vfs.h>
typedef struct cred cred_t;
extern struct task_struct init_task;
#define kcred ((cred_t *)(init_task.cred))
#define CRED() ((cred_t *)current_cred())
/* Linux 4.9 API change, GROUP_AT was removed */
#ifndef GROUP_AT
#define GROUP_AT(gi, i) ((gi)->gid[i])
#endif
#define KUID_TO_SUID(x) (__kuid_val(x))
#define KGID_TO_SGID(x) (__kgid_val(x))
#define SUID_TO_KUID(x) (KUIDT_INIT(x))
#define SGID_TO_KGID(x) (KGIDT_INIT(x))
#define KGIDP_TO_SGIDP(x) (&(x)->val)
+extern zidmap_t *zfs_get_init_idmap(void);
+
/* Check if the user ns is the initial one */
static inline boolean_t
zfs_is_init_userns(struct user_namespace *user_ns)
{
#if defined(CONFIG_USER_NS)
return (user_ns == kcred->user_ns);
#else
return (B_FALSE);
#endif
}
static inline struct user_namespace *zfs_i_user_ns(struct inode *inode)
{
#ifdef HAVE_SUPER_USER_NS
return (inode->i_sb->s_user_ns);
#else
return (kcred->user_ns);
#endif
}
static inline boolean_t zfs_no_idmapping(struct user_namespace *mnt_userns,
struct user_namespace *fs_userns)
{
return (zfs_is_init_userns(mnt_userns) || mnt_userns == fs_userns);
}
-static inline uid_t zfs_uid_to_vfsuid(struct user_namespace *mnt_userns,
+static inline uid_t zfs_uid_to_vfsuid(zidmap_t *mnt_userns,
struct user_namespace *fs_userns, uid_t uid)
{
- if (zfs_no_idmapping(mnt_userns, fs_userns))
+ struct user_namespace *owner = idmap_owner(mnt_userns);
+ if (zfs_no_idmapping(owner, fs_userns))
return (uid);
if (!zfs_is_init_userns(fs_userns))
uid = from_kuid(fs_userns, KUIDT_INIT(uid));
if (uid == (uid_t)-1)
return (uid);
- return (__kuid_val(make_kuid(mnt_userns, uid)));
+ return (__kuid_val(make_kuid(owner, uid)));
}
-static inline gid_t zfs_gid_to_vfsgid(struct user_namespace *mnt_userns,
+static inline gid_t zfs_gid_to_vfsgid(zidmap_t *mnt_userns,
struct user_namespace *fs_userns, gid_t gid)
{
- if (zfs_no_idmapping(mnt_userns, fs_userns))
+ struct user_namespace *owner = idmap_owner(mnt_userns);
+ if (zfs_no_idmapping(owner, fs_userns))
return (gid);
if (!zfs_is_init_userns(fs_userns))
gid = from_kgid(fs_userns, KGIDT_INIT(gid));
if (gid == (gid_t)-1)
return (gid);
- return (__kgid_val(make_kgid(mnt_userns, gid)));
+ return (__kgid_val(make_kgid(owner, gid)));
}
-static inline uid_t zfs_vfsuid_to_uid(struct user_namespace *mnt_userns,
+static inline uid_t zfs_vfsuid_to_uid(zidmap_t *mnt_userns,
struct user_namespace *fs_userns, uid_t uid)
{
- if (zfs_no_idmapping(mnt_userns, fs_userns))
+ struct user_namespace *owner = idmap_owner(mnt_userns);
+ if (zfs_no_idmapping(owner, fs_userns))
return (uid);
- uid = from_kuid(mnt_userns, KUIDT_INIT(uid));
+ uid = from_kuid(owner, KUIDT_INIT(uid));
if (uid == (uid_t)-1)
return (uid);
if (zfs_is_init_userns(fs_userns))
return (uid);
return (__kuid_val(make_kuid(fs_userns, uid)));
}
-static inline gid_t zfs_vfsgid_to_gid(struct user_namespace *mnt_userns,
+static inline gid_t zfs_vfsgid_to_gid(zidmap_t *mnt_userns,
struct user_namespace *fs_userns, gid_t gid)
{
- if (zfs_no_idmapping(mnt_userns, fs_userns))
+ struct user_namespace *owner = idmap_owner(mnt_userns);
+ if (zfs_no_idmapping(owner, fs_userns))
return (gid);
- gid = from_kgid(mnt_userns, KGIDT_INIT(gid));
+ gid = from_kgid(owner, KGIDT_INIT(gid));
if (gid == (gid_t)-1)
return (gid);
if (zfs_is_init_userns(fs_userns))
return (gid);
return (__kgid_val(make_kgid(fs_userns, gid)));
}
extern void crhold(cred_t *cr);
extern void crfree(cred_t *cr);
extern uid_t crgetuid(const cred_t *cr);
extern uid_t crgetruid(const cred_t *cr);
extern gid_t crgetgid(const cred_t *cr);
extern int crgetngroups(const cred_t *cr);
extern gid_t *crgetgroups(const cred_t *cr);
extern int groupmember(gid_t gid, const cred_t *cr);
#endif /* _SPL_CRED_H */
diff --git a/sys/contrib/openzfs/include/os/linux/spl/sys/isa_defs.h b/sys/contrib/openzfs/include/os/linux/spl/sys/isa_defs.h
index 5801ec92bc2b..7c95c94c1cf1 100644
--- a/sys/contrib/openzfs/include/os/linux/spl/sys/isa_defs.h
+++ b/sys/contrib/openzfs/include/os/linux/spl/sys/isa_defs.h
@@ -1,232 +1,248 @@
/*
* Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
* Copyright (C) 2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
* Written by Brian Behlendorf <behlendorf1@llnl.gov>.
* UCRL-CODE-235197
*
* This file is part of the SPL, Solaris Porting Layer.
*
* The SPL is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation; either version 2 of the License, or (at your
* option) any later version.
*
* The SPL is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* for more details.
*
* You should have received a copy of the GNU General Public License along
* with the SPL. If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef _SPL_ISA_DEFS_H
#define _SPL_ISA_DEFS_H
/* x86_64 arch specific defines */
#if defined(__x86_64) || defined(__x86_64__)
#if !defined(__x86_64)
#define __x86_64
#endif
#if !defined(__amd64)
#define __amd64
#endif
#if !defined(__x86)
#define __x86
#endif
#if defined(_ILP32)
/* x32-specific defines; careful to *not* define _LP64 here */
#else
#if !defined(_LP64)
#define _LP64
#endif
#endif
/* i386 arch specific defines */
#elif defined(__i386) || defined(__i386__)
#if !defined(__i386)
#define __i386
#endif
#if !defined(__x86)
#define __x86
#endif
#if !defined(_ILP32)
#define _ILP32
#endif
/* powerpc (ppc64) arch specific defines */
#elif defined(__powerpc) || defined(__powerpc__) || defined(__powerpc64__)
#if !defined(__powerpc)
#define __powerpc
#endif
#if !defined(__powerpc__)
#define __powerpc__
#endif
#if defined(__powerpc64__)
#if !defined(_LP64)
#define _LP64
#endif
#else
#if !defined(_ILP32)
#define _ILP32
#endif
#endif
/* arm arch specific defines */
#elif defined(__arm) || defined(__arm__)
#if !defined(__arm)
#define __arm
#endif
#if !defined(__arm__)
#define __arm__
#endif
#if !defined(_ILP32)
#define _ILP32
#endif
#if defined(__ARMEL__)
#define _ZFS_LITTLE_ENDIAN
#else
#define _ZFS_BIG_ENDIAN
#endif
/* aarch64 arch specific defines */
#elif defined(__aarch64__)
#if !defined(_LP64)
#define _LP64
#endif
#if defined(__AARCH64EL__)
#define _ZFS_LITTLE_ENDIAN
#else
#define _ZFS_BIG_ENDIAN
#endif
/* sparc arch specific defines */
#elif defined(__sparc) || defined(__sparc__)
#if !defined(__sparc)
#define __sparc
#endif
#if !defined(__sparc__)
#define __sparc__
#endif
#if defined(__arch64__)
#if !defined(_LP64)
#define _LP64
#endif
#else
#if !defined(_ILP32)
#define _ILP32
#endif
#endif
#define _ZFS_BIG_ENDIAN
#define _SUNOS_VTOC_16
/* s390 arch specific defines */
#elif defined(__s390__)
#if defined(__s390x__)
#if !defined(_LP64)
#define _LP64
#endif
#else
#if !defined(_ILP32)
#define _ILP32
#endif
#endif
#define _ZFS_BIG_ENDIAN
/* MIPS arch specific defines */
#elif defined(__mips__)
#if defined(__MIPSEB__)
#define _ZFS_BIG_ENDIAN
#elif defined(__MIPSEL__)
#define _ZFS_LITTLE_ENDIAN
#else
#error MIPS no endian specified
#endif
#ifndef _LP64
#define _ILP32
#endif
#define _SUNOS_VTOC_16
/*
* RISC-V arch specific defines
* only RV64G (including atomic) LP64 is supported yet
*/
#elif defined(__riscv) && defined(__riscv_xlen) && __riscv_xlen == 64 && \
defined(__riscv_atomic) && __riscv_atomic
#if !defined(_LP64)
#define _LP64 1
#endif
#ifndef __riscv__
#define __riscv__
#endif
#ifndef __rv64g__
#define __rv64g__
#endif
#define _ZFS_LITTLE_ENDIAN
#define _SUNOS_VTOC_16
+/*
+ * LoongArch arch specific defines
+ * only LoongArch64 is supported yet
+ */
+#elif defined(__loongarch__) && defined(__loongarch_lp64)
+
+#if !defined(_LP64)
+#define _LP64
+#endif
+
+#define _ZFS_LITTLE_ENDIAN
+#define _SUNOS_VTOC_16
+
+/* not all LoongArch cores support unaligned accesses in hardware */
+#define _ALIGNMENT_REQUIRED 1
+
#else
/*
* Currently supported:
- * x86_64, x32, i386, arm, powerpc, s390, sparc, mips, and RV64G
+ * x86_64, x32, i386, arm, powerpc, s390, sparc, mips, RV64G, and LoongArch64
*/
#error "Unsupported ISA type"
#endif
#if defined(_ILP32) && defined(_LP64)
#error "Both _ILP32 and _LP64 are defined"
#endif
#if !defined(_ILP32) && !defined(_LP64)
#error "Neither _ILP32 or _LP64 are defined"
#endif
#include <sys/byteorder.h>
/*
* CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS will be defined by the Linux
* kernel for architectures which support efficient unaligned access.
*/
#if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)
#define HAVE_EFFICIENT_UNALIGNED_ACCESS
#endif
#if defined(_ZFS_LITTLE_ENDIAN) && defined(_ZFS_BIG_ENDIAN)
#error "Both _ZFS_LITTLE_ENDIAN and _ZFS_BIG_ENDIAN are defined"
#endif
#if !defined(_ZFS_LITTLE_ENDIAN) && !defined(_ZFS_BIG_ENDIAN)
#error "Neither _ZFS_LITTLE_ENDIAN or _ZFS_BIG_ENDIAN are defined"
#endif
#endif /* _SPL_ISA_DEFS_H */
diff --git a/sys/contrib/openzfs/include/os/linux/spl/sys/types.h b/sys/contrib/openzfs/include/os/linux/spl/sys/types.h
index cae1bbddf105..a7666187ec23 100644
--- a/sys/contrib/openzfs/include/os/linux/spl/sys/types.h
+++ b/sys/contrib/openzfs/include/os/linux/spl/sys/types.h
@@ -1,60 +1,73 @@
/*
* Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
* Copyright (C) 2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
* Written by Brian Behlendorf <behlendorf1@llnl.gov>.
* UCRL-CODE-235197
*
* This file is part of the SPL, Solaris Porting Layer.
*
* The SPL is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation; either version 2 of the License, or (at your
* option) any later version.
*
* The SPL is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* for more details.
*
* You should have received a copy of the GNU General Public License along
* with the SPL. If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef _SPL_TYPES_H
#define _SPL_TYPES_H
#include <linux/types.h>
typedef enum {
B_FALSE = 0,
B_TRUE = 1
} boolean_t;
typedef unsigned char uchar_t;
typedef unsigned short ushort_t;
typedef unsigned int uint_t;
typedef unsigned long ulong_t;
typedef unsigned long long u_longlong_t;
typedef long long longlong_t;
typedef unsigned long intptr_t;
typedef unsigned long long rlim64_t;
typedef struct task_struct kthread_t;
typedef struct task_struct proc_t;
typedef int id_t;
typedef short pri_t;
typedef short index_t;
typedef longlong_t offset_t;
typedef u_longlong_t u_offset_t;
typedef ulong_t pgcnt_t;
typedef int major_t;
typedef int minor_t;
struct user_namespace;
-typedef struct user_namespace zuserns_t;
+#ifdef HAVE_IOPS_CREATE_IDMAP
+#include <linux/refcount.h>
+struct mnt_idmap {
+ struct user_namespace *owner;
+ refcount_t count;
+};
+typedef struct mnt_idmap zidmap_t;
+#define idmap_owner(p) (((struct mnt_idmap *)p)->owner)
+#else
+typedef struct user_namespace zidmap_t;
+#define idmap_owner(p) ((struct user_namespace *)p)
+#endif
+
+extern zidmap_t *zfs_init_idmap;
#endif /* _SPL_TYPES_H */
diff --git a/sys/contrib/openzfs/include/os/linux/zfs/sys/policy.h b/sys/contrib/openzfs/include/os/linux/zfs/sys/policy.h
index b182da95b8e0..0c265db78591 100644
--- a/sys/contrib/openzfs/include/os/linux/zfs/sys/policy.h
+++ b/sys/contrib/openzfs/include/os/linux/zfs/sys/policy.h
@@ -1,64 +1,64 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2015, Joyent, Inc. All rights reserved.
* Copyright (c) 2016, Lawrence Livermore National Security, LLC.
*/
#ifndef _SYS_POLICY_H
#define _SYS_POLICY_H
#ifdef _KERNEL
#include <sys/cred.h>
#include <sys/types.h>
#include <sys/xvattr.h>
#include <sys/zpl.h>
struct znode;
int secpolicy_nfs(const cred_t *);
int secpolicy_sys_config(const cred_t *, boolean_t);
int secpolicy_vnode_access2(const cred_t *, struct inode *,
uid_t, mode_t, mode_t);
int secpolicy_vnode_any_access(const cred_t *, struct inode *, uid_t);
int secpolicy_vnode_chown(const cred_t *, uid_t);
int secpolicy_vnode_create_gid(const cred_t *);
int secpolicy_vnode_remove(const cred_t *);
int secpolicy_vnode_setdac(const cred_t *, uid_t);
int secpolicy_vnode_setid_retain(struct znode *, const cred_t *, boolean_t);
-int secpolicy_vnode_setids_setgids(const cred_t *, gid_t, zuserns_t *,
- zuserns_t *);
+int secpolicy_vnode_setids_setgids(const cred_t *, gid_t, zidmap_t *,
+ struct user_namespace *);
int secpolicy_zinject(const cred_t *);
int secpolicy_zfs(const cred_t *);
int secpolicy_zfs_proc(const cred_t *, proc_t *);
void secpolicy_setid_clear(vattr_t *, cred_t *);
int secpolicy_setid_setsticky_clear(struct inode *, vattr_t *,
- const vattr_t *, cred_t *, zuserns_t *, zuserns_t *);
+ const vattr_t *, cred_t *, zidmap_t *, struct user_namespace *);
int secpolicy_xvattr(xvattr_t *, uid_t, cred_t *, mode_t);
int secpolicy_vnode_setattr(cred_t *, struct inode *, struct vattr *,
const struct vattr *, int, int (void *, int, cred_t *), void *);
int secpolicy_basic_link(const cred_t *);
#endif /* _KERNEL */
#endif /* _SYS_POLICY_H */
diff --git a/sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h b/sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h
index 6dd18c5974b9..7bddd9d1f469 100644
--- a/sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h
+++ b/sys/contrib/openzfs/include/os/linux/zfs/sys/trace_zil.h
@@ -1,225 +1,233 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
#if defined(_KERNEL)
#if defined(HAVE_DECLARE_EVENT_CLASS)
#undef TRACE_SYSTEM
#define TRACE_SYSTEM zfs
#undef TRACE_SYSTEM_VAR
#define TRACE_SYSTEM_VAR zfs_zil
#if !defined(_TRACE_ZIL_H) || defined(TRACE_HEADER_MULTI_READ)
#define _TRACE_ZIL_H
#include <linux/tracepoint.h>
#include <sys/types.h>
#define ZILOG_TP_STRUCT_ENTRY \
__field(uint64_t, zl_lr_seq) \
__field(uint64_t, zl_commit_lr_seq) \
__field(uint64_t, zl_destroy_txg) \
__field(uint64_t, zl_replaying_seq) \
__field(uint32_t, zl_suspend) \
__field(uint8_t, zl_suspending) \
__field(uint8_t, zl_keep_first) \
__field(uint8_t, zl_replay) \
__field(uint8_t, zl_stop_sync) \
__field(uint8_t, zl_logbias) \
__field(uint8_t, zl_sync) \
__field(int, zl_parse_error) \
__field(uint64_t, zl_parse_blk_seq) \
__field(uint64_t, zl_parse_lr_seq) \
__field(uint64_t, zl_parse_blk_count) \
__field(uint64_t, zl_parse_lr_count) \
__field(uint64_t, zl_cur_used) \
__field(clock_t, zl_replay_time) \
__field(uint64_t, zl_replay_blks)
#define ZILOG_TP_FAST_ASSIGN \
__entry->zl_lr_seq = zilog->zl_lr_seq; \
__entry->zl_commit_lr_seq = zilog->zl_commit_lr_seq; \
__entry->zl_destroy_txg = zilog->zl_destroy_txg; \
__entry->zl_replaying_seq = zilog->zl_replaying_seq; \
__entry->zl_suspend = zilog->zl_suspend; \
__entry->zl_suspending = zilog->zl_suspending; \
__entry->zl_keep_first = zilog->zl_keep_first; \
__entry->zl_replay = zilog->zl_replay; \
__entry->zl_stop_sync = zilog->zl_stop_sync; \
__entry->zl_logbias = zilog->zl_logbias; \
__entry->zl_sync = zilog->zl_sync; \
__entry->zl_parse_error = zilog->zl_parse_error; \
__entry->zl_parse_blk_seq = zilog->zl_parse_blk_seq; \
__entry->zl_parse_lr_seq = zilog->zl_parse_lr_seq; \
__entry->zl_parse_blk_count = zilog->zl_parse_blk_count;\
__entry->zl_parse_lr_count = zilog->zl_parse_lr_count; \
__entry->zl_cur_used = zilog->zl_cur_used; \
__entry->zl_replay_time = zilog->zl_replay_time; \
__entry->zl_replay_blks = zilog->zl_replay_blks;
#define ZILOG_TP_PRINTK_FMT \
"zl { lr_seq %llu commit_lr_seq %llu destroy_txg %llu " \
"replaying_seq %llu suspend %u suspending %u keep_first %u " \
"replay %u stop_sync %u logbias %u sync %u " \
"parse_error %u parse_blk_seq %llu parse_lr_seq %llu " \
"parse_blk_count %llu parse_lr_count %llu " \
"cur_used %llu replay_time %lu replay_blks %llu }"
#define ZILOG_TP_PRINTK_ARGS \
__entry->zl_lr_seq, __entry->zl_commit_lr_seq, \
__entry->zl_destroy_txg, __entry->zl_replaying_seq, \
__entry->zl_suspend, __entry->zl_suspending, \
__entry->zl_keep_first, __entry->zl_replay, \
__entry->zl_stop_sync, __entry->zl_logbias, __entry->zl_sync, \
__entry->zl_parse_error, __entry->zl_parse_blk_seq, \
__entry->zl_parse_lr_seq, __entry->zl_parse_blk_count, \
__entry->zl_parse_lr_count, __entry->zl_cur_used, \
__entry->zl_replay_time, __entry->zl_replay_blks
#define ITX_TP_STRUCT_ENTRY \
__field(itx_wr_state_t, itx_wr_state) \
__field(uint8_t, itx_sync) \
__field(zil_callback_t, itx_callback) \
__field(void *, itx_callback_data) \
__field(uint64_t, itx_oid) \
\
__field(uint64_t, lrc_txtype) \
__field(uint64_t, lrc_reclen) \
__field(uint64_t, lrc_txg) \
__field(uint64_t, lrc_seq)
#define ITX_TP_FAST_ASSIGN \
__entry->itx_wr_state = itx->itx_wr_state; \
__entry->itx_sync = itx->itx_sync; \
__entry->itx_callback = itx->itx_callback; \
__entry->itx_callback_data = itx->itx_callback_data; \
__entry->itx_oid = itx->itx_oid; \
\
__entry->lrc_txtype = itx->itx_lr.lrc_txtype; \
__entry->lrc_reclen = itx->itx_lr.lrc_reclen; \
__entry->lrc_txg = itx->itx_lr.lrc_txg; \
__entry->lrc_seq = itx->itx_lr.lrc_seq;
#define ITX_TP_PRINTK_FMT \
"itx { wr_state %u sync %u callback %p callback_data %p oid %llu" \
" { txtype %llu reclen %llu txg %llu seq %llu } }"
#define ITX_TP_PRINTK_ARGS \
__entry->itx_wr_state, __entry->itx_sync, __entry->itx_callback,\
__entry->itx_callback_data, __entry->itx_oid, \
__entry->lrc_txtype, __entry->lrc_reclen, __entry->lrc_txg, \
__entry->lrc_seq
#define ZCW_TP_STRUCT_ENTRY \
__field(lwb_t *, zcw_lwb) \
__field(boolean_t, zcw_done) \
__field(int, zcw_zio_error) \
#define ZCW_TP_FAST_ASSIGN \
__entry->zcw_lwb = zcw->zcw_lwb; \
__entry->zcw_done = zcw->zcw_done; \
__entry->zcw_zio_error = zcw->zcw_zio_error;
#define ZCW_TP_PRINTK_FMT \
"zcw { lwb %p done %u error %u }"
#define ZCW_TP_PRINTK_ARGS \
__entry->zcw_lwb, __entry->zcw_done, __entry->zcw_zio_error
/*
* Generic support for two argument tracepoints of the form:
*
* DTRACE_PROBE2(...,
* zilog_t *, ...,
* itx_t *, ...);
*/
+
+#if defined(__clang__)
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Wordered-compare-function-pointers"
+#endif
/* BEGIN CSTYLED */
DECLARE_EVENT_CLASS(zfs_zil_process_itx_class,
TP_PROTO(zilog_t *zilog, itx_t *itx),
TP_ARGS(zilog, itx),
TP_STRUCT__entry(
ZILOG_TP_STRUCT_ENTRY
ITX_TP_STRUCT_ENTRY
),
TP_fast_assign(
ZILOG_TP_FAST_ASSIGN
ITX_TP_FAST_ASSIGN
),
TP_printk(
ZILOG_TP_PRINTK_FMT " " ITX_TP_PRINTK_FMT,
ZILOG_TP_PRINTK_ARGS, ITX_TP_PRINTK_ARGS)
);
/* END CSTYLED */
+#if defined(__clang__)
+#pragma clang diagnostic pop
+#endif
#define DEFINE_ZIL_PROCESS_ITX_EVENT(name) \
DEFINE_EVENT(zfs_zil_process_itx_class, name, \
TP_PROTO(zilog_t *zilog, itx_t *itx), \
TP_ARGS(zilog, itx))
DEFINE_ZIL_PROCESS_ITX_EVENT(zfs_zil__process__commit__itx);
DEFINE_ZIL_PROCESS_ITX_EVENT(zfs_zil__process__normal__itx);
/*
* Generic support for two argument tracepoints of the form:
*
* DTRACE_PROBE2(...,
* zilog_t *, ...,
* zil_commit_waiter_t *, ...);
*/
/* BEGIN CSTYLED */
DECLARE_EVENT_CLASS(zfs_zil_commit_io_error_class,
TP_PROTO(zilog_t *zilog, zil_commit_waiter_t *zcw),
TP_ARGS(zilog, zcw),
TP_STRUCT__entry(
ZILOG_TP_STRUCT_ENTRY
ZCW_TP_STRUCT_ENTRY
),
TP_fast_assign(
ZILOG_TP_FAST_ASSIGN
ZCW_TP_FAST_ASSIGN
),
TP_printk(
ZILOG_TP_PRINTK_FMT " " ZCW_TP_PRINTK_FMT,
ZILOG_TP_PRINTK_ARGS, ZCW_TP_PRINTK_ARGS)
);
#define DEFINE_ZIL_COMMIT_IO_ERROR_EVENT(name) \
DEFINE_EVENT(zfs_zil_commit_io_error_class, name, \
TP_PROTO(zilog_t *zilog, zil_commit_waiter_t *zcw), \
TP_ARGS(zilog, zcw))
DEFINE_ZIL_COMMIT_IO_ERROR_EVENT(zfs_zil__commit__io__error);
#endif /* _TRACE_ZIL_H */
#undef TRACE_INCLUDE_PATH
#undef TRACE_INCLUDE_FILE
#define TRACE_INCLUDE_PATH sys
#define TRACE_INCLUDE_FILE trace_zil
#include <trace/define_trace.h>
#else
DEFINE_DTRACE_PROBE2(zil__process__commit__itx);
DEFINE_DTRACE_PROBE2(zil__process__normal__itx);
DEFINE_DTRACE_PROBE2(zil__commit__io__error);
#endif /* HAVE_DECLARE_EVENT_CLASS */
#endif /* _KERNEL */
diff --git a/sys/contrib/openzfs/include/os/linux/zfs/sys/zfs_vnops_os.h b/sys/contrib/openzfs/include/os/linux/zfs/sys/zfs_vnops_os.h
index 1caec0ef4f1d..7a1db7deeec8 100644
--- a/sys/contrib/openzfs/include/os/linux/zfs/sys/zfs_vnops_os.h
+++ b/sys/contrib/openzfs/include/os/linux/zfs/sys/zfs_vnops_os.h
@@ -1,87 +1,86 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
*/
#ifndef _SYS_FS_ZFS_VNOPS_OS_H
#define _SYS_FS_ZFS_VNOPS_OS_H
#include <sys/vnode.h>
#include <sys/xvattr.h>
#include <sys/uio.h>
#include <sys/cred.h>
#include <sys/fcntl.h>
#include <sys/pathname.h>
#include <sys/zpl.h>
#include <sys/zfs_file.h>
#ifdef __cplusplus
extern "C" {
#endif
extern int zfs_open(struct inode *ip, int mode, int flag, cred_t *cr);
extern int zfs_close(struct inode *ip, int flag, cred_t *cr);
extern int zfs_write_simple(znode_t *zp, const void *data, size_t len,
loff_t pos, size_t *resid);
extern int zfs_lookup(znode_t *dzp, char *nm, znode_t **zpp, int flags,
cred_t *cr, int *direntflags, pathname_t *realpnp);
extern int zfs_create(znode_t *dzp, char *name, vattr_t *vap, int excl,
int mode, znode_t **zpp, cred_t *cr, int flag, vsecattr_t *vsecp,
- zuserns_t *mnt_ns);
+ zidmap_t *mnt_ns);
extern int zfs_tmpfile(struct inode *dip, vattr_t *vapzfs, int excl,
int mode, struct inode **ipp, cred_t *cr, int flag, vsecattr_t *vsecp,
- zuserns_t *mnt_ns);
+ zidmap_t *mnt_ns);
extern int zfs_remove(znode_t *dzp, char *name, cred_t *cr, int flags);
extern int zfs_mkdir(znode_t *dzp, char *dirname, vattr_t *vap,
- znode_t **zpp, cred_t *cr, int flags, vsecattr_t *vsecp, zuserns_t *mnt_ns);
+ znode_t **zpp, cred_t *cr, int flags, vsecattr_t *vsecp, zidmap_t *mnt_ns);
extern int zfs_rmdir(znode_t *dzp, char *name, znode_t *cwd,
cred_t *cr, int flags);
extern int zfs_readdir(struct inode *ip, zpl_dir_context_t *ctx, cred_t *cr);
-extern int zfs_getattr_fast(struct user_namespace *, struct inode *ip,
- struct kstat *sp);
+extern int zfs_getattr_fast(zidmap_t *, struct inode *ip, struct kstat *sp);
extern int zfs_setattr(znode_t *zp, vattr_t *vap, int flag, cred_t *cr,
- zuserns_t *mnt_ns);
+ zidmap_t *mnt_ns);
extern int zfs_rename(znode_t *sdzp, char *snm, znode_t *tdzp,
char *tnm, cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap,
- zuserns_t *mnt_ns);
+ zidmap_t *mnt_ns);
extern int zfs_symlink(znode_t *dzp, char *name, vattr_t *vap,
- char *link, znode_t **zpp, cred_t *cr, int flags, zuserns_t *mnt_ns);
+ char *link, znode_t **zpp, cred_t *cr, int flags, zidmap_t *mnt_ns);
extern int zfs_readlink(struct inode *ip, zfs_uio_t *uio, cred_t *cr);
extern int zfs_link(znode_t *tdzp, znode_t *szp,
char *name, cred_t *cr, int flags);
extern void zfs_inactive(struct inode *ip);
extern int zfs_space(znode_t *zp, int cmd, flock64_t *bfp, int flag,
offset_t offset, cred_t *cr);
extern int zfs_fid(struct inode *ip, fid_t *fidp);
extern int zfs_getpage(struct inode *ip, struct page *pp);
extern int zfs_putpage(struct inode *ip, struct page *pp,
struct writeback_control *wbc, boolean_t for_sync);
extern int zfs_dirty_inode(struct inode *ip, int flags);
extern int zfs_map(struct inode *ip, offset_t off, caddr_t *addrp,
size_t len, unsigned long vm_flags);
extern void zfs_zrele_async(znode_t *zp);
#ifdef __cplusplus
}
#endif
#endif /* _SYS_FS_ZFS_VNOPS_H */
diff --git a/sys/contrib/openzfs/include/os/linux/zfs/sys/zpl.h b/sys/contrib/openzfs/include/os/linux/zfs/sys/zpl.h
index ac1f01a86c41..2b302e9dab07 100644
--- a/sys/contrib/openzfs/include/os/linux/zfs/sys/zpl.h
+++ b/sys/contrib/openzfs/include/os/linux/zfs/sys/zpl.h
@@ -1,208 +1,213 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2011, Lawrence Livermore National Security, LLC.
*/
#ifndef _SYS_ZPL_H
#define _SYS_ZPL_H
#include <sys/mntent.h>
#include <sys/vfs.h>
#include <linux/aio.h>
#include <linux/dcache_compat.h>
#include <linux/exportfs.h>
#include <linux/falloc.h>
#include <linux/parser.h>
#include <linux/task_io_accounting_ops.h>
#include <linux/vfs_compat.h>
#include <linux/writeback.h>
#include <linux/xattr_compat.h>
/* zpl_inode.c */
extern void zpl_vap_init(vattr_t *vap, struct inode *dir,
- umode_t mode, cred_t *cr, zuserns_t *mnt_ns);
+ umode_t mode, cred_t *cr, zidmap_t *mnt_ns);
extern const struct inode_operations zpl_inode_operations;
#ifdef HAVE_RENAME2_OPERATIONS_WRAPPER
extern const struct inode_operations_wrapper zpl_dir_inode_operations;
#else
extern const struct inode_operations zpl_dir_inode_operations;
#endif
extern const struct inode_operations zpl_symlink_inode_operations;
extern const struct inode_operations zpl_special_inode_operations;
/* zpl_file.c */
extern const struct address_space_operations zpl_address_space_operations;
extern const struct file_operations zpl_file_operations;
extern const struct file_operations zpl_dir_file_operations;
/* zpl_super.c */
extern void zpl_prune_sb(int64_t nr_to_scan, void *arg);
extern const struct super_operations zpl_super_operations;
extern const struct export_operations zpl_export_operations;
extern struct file_system_type zpl_fs_type;
/* zpl_xattr.c */
extern ssize_t zpl_xattr_list(struct dentry *dentry, char *buf, size_t size);
extern int zpl_xattr_security_init(struct inode *ip, struct inode *dip,
const struct qstr *qstr);
#if defined(CONFIG_FS_POSIX_ACL)
#if defined(HAVE_SET_ACL)
-#if defined(HAVE_SET_ACL_USERNS)
+#if defined(HAVE_SET_ACL_IDMAP_DENTRY)
+extern int zpl_set_acl(struct mnt_idmap *idmap, struct dentry *dentry,
+ struct posix_acl *acl, int type);
+#elif defined(HAVE_SET_ACL_USERNS)
extern int zpl_set_acl(struct user_namespace *userns, struct inode *ip,
struct posix_acl *acl, int type);
#elif defined(HAVE_SET_ACL_USERNS_DENTRY_ARG2)
extern int zpl_set_acl(struct user_namespace *userns, struct dentry *dentry,
struct posix_acl *acl, int type);
#else
extern int zpl_set_acl(struct inode *ip, struct posix_acl *acl, int type);
#endif /* HAVE_SET_ACL_USERNS */
#endif /* HAVE_SET_ACL */
#if defined(HAVE_GET_ACL_RCU) || defined(HAVE_GET_INODE_ACL)
extern struct posix_acl *zpl_get_acl(struct inode *ip, int type, bool rcu);
#elif defined(HAVE_GET_ACL)
extern struct posix_acl *zpl_get_acl(struct inode *ip, int type);
#endif
extern int zpl_init_acl(struct inode *ip, struct inode *dir);
extern int zpl_chmod_acl(struct inode *ip);
#else
static inline int
zpl_init_acl(struct inode *ip, struct inode *dir)
{
return (0);
}
static inline int
zpl_chmod_acl(struct inode *ip)
{
return (0);
}
#endif /* CONFIG_FS_POSIX_ACL */
extern xattr_handler_t *zpl_xattr_handlers[];
/* zpl_ctldir.c */
extern const struct file_operations zpl_fops_root;
extern const struct inode_operations zpl_ops_root;
extern const struct file_operations zpl_fops_snapdir;
extern const struct inode_operations zpl_ops_snapdir;
extern const struct file_operations zpl_fops_shares;
extern const struct inode_operations zpl_ops_shares;
#if defined(HAVE_VFS_ITERATE) || defined(HAVE_VFS_ITERATE_SHARED)
#define ZPL_DIR_CONTEXT_INIT(_dirent, _actor, _pos) { \
.actor = _actor, \
.pos = _pos, \
}
typedef struct dir_context zpl_dir_context_t;
#define zpl_dir_emit dir_emit
#define zpl_dir_emit_dot dir_emit_dot
#define zpl_dir_emit_dotdot dir_emit_dotdot
#define zpl_dir_emit_dots dir_emit_dots
#else
typedef struct zpl_dir_context {
void *dirent;
const filldir_t actor;
loff_t pos;
} zpl_dir_context_t;
#define ZPL_DIR_CONTEXT_INIT(_dirent, _actor, _pos) { \
.dirent = _dirent, \
.actor = _actor, \
.pos = _pos, \
}
static inline bool
zpl_dir_emit(zpl_dir_context_t *ctx, const char *name, int namelen,
uint64_t ino, unsigned type)
{
return (!ctx->actor(ctx->dirent, name, namelen, ctx->pos, ino, type));
}
static inline bool
zpl_dir_emit_dot(struct file *file, zpl_dir_context_t *ctx)
{
return (ctx->actor(ctx->dirent, ".", 1, ctx->pos,
file_inode(file)->i_ino, DT_DIR) == 0);
}
static inline bool
zpl_dir_emit_dotdot(struct file *file, zpl_dir_context_t *ctx)
{
return (ctx->actor(ctx->dirent, "..", 2, ctx->pos,
parent_ino(file_dentry(file)), DT_DIR) == 0);
}
static inline bool
zpl_dir_emit_dots(struct file *file, zpl_dir_context_t *ctx)
{
if (ctx->pos == 0) {
if (!zpl_dir_emit_dot(file, ctx))
return (false);
ctx->pos = 1;
}
if (ctx->pos == 1) {
if (!zpl_dir_emit_dotdot(file, ctx))
return (false);
ctx->pos = 2;
}
return (true);
}
#endif /* HAVE_VFS_ITERATE */
#if defined(HAVE_INODE_TIMESTAMP_TRUNCATE)
#define zpl_inode_timestamp_truncate(ts, ip) timestamp_truncate(ts, ip)
#elif defined(HAVE_INODE_TIMESPEC64_TIMES)
#define zpl_inode_timestamp_truncate(ts, ip) \
timespec64_trunc(ts, (ip)->i_sb->s_time_gran)
#else
#define zpl_inode_timestamp_truncate(ts, ip) \
timespec_trunc(ts, (ip)->i_sb->s_time_gran)
#endif
#if defined(HAVE_INODE_OWNER_OR_CAPABLE)
#define zpl_inode_owner_or_capable(ns, ip) inode_owner_or_capable(ip)
-#elif defined(HAVE_INODE_OWNER_OR_CAPABLE_IDMAPPED)
+#elif defined(HAVE_INODE_OWNER_OR_CAPABLE_USERNS)
#define zpl_inode_owner_or_capable(ns, ip) inode_owner_or_capable(ns, ip)
+#elif defined(HAVE_INODE_OWNER_OR_CAPABLE_IDMAP)
+#define zpl_inode_owner_or_capable(idmap, ip) inode_owner_or_capable(idmap, ip)
#else
#error "Unsupported kernel"
#endif
-#ifdef HAVE_SETATTR_PREPARE_USERNS
+#if defined(HAVE_SETATTR_PREPARE_USERNS) || defined(HAVE_SETATTR_PREPARE_IDMAP)
#define zpl_setattr_prepare(ns, dentry, ia) setattr_prepare(ns, dentry, ia)
#else
/*
* Use kernel-provided version, or our own from
* linux/vfs_compat.h
*/
#define zpl_setattr_prepare(ns, dentry, ia) setattr_prepare(dentry, ia)
#endif
#endif /* _SYS_ZPL_H */
diff --git a/sys/contrib/openzfs/include/sys/dmu_objset.h b/sys/contrib/openzfs/include/sys/dmu_objset.h
index d22c682875d8..9f6e0fdd601b 100644
--- a/sys/contrib/openzfs/include/sys/dmu_objset.h
+++ b/sys/contrib/openzfs/include/sys/dmu_objset.h
@@ -1,269 +1,276 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
*/
/* Portions Copyright 2010 Robert Milkowski */
#ifndef _SYS_DMU_OBJSET_H
#define _SYS_DMU_OBJSET_H
#include <sys/spa.h>
#include <sys/arc.h>
#include <sys/txg.h>
#include <sys/zfs_context.h>
#include <sys/dnode.h>
#include <sys/zio.h>
#include <sys/zil.h>
#include <sys/sa.h>
#include <sys/zfs_ioctl.h>
#ifdef __cplusplus
extern "C" {
#endif
extern krwlock_t os_lock;
struct dsl_pool;
struct dsl_dataset;
struct dmu_tx;
#define OBJSET_PHYS_SIZE_V1 1024
#define OBJSET_PHYS_SIZE_V2 2048
#define OBJSET_PHYS_SIZE_V3 4096
#define OBJSET_BUF_HAS_USERUSED(buf) \
(arc_buf_size(buf) >= OBJSET_PHYS_SIZE_V2)
#define OBJSET_BUF_HAS_PROJECTUSED(buf) \
(arc_buf_size(buf) >= OBJSET_PHYS_SIZE_V3)
#define OBJSET_FLAG_USERACCOUNTING_COMPLETE (1ULL << 0)
#define OBJSET_FLAG_USEROBJACCOUNTING_COMPLETE (1ULL << 1)
#define OBJSET_FLAG_PROJECTQUOTA_COMPLETE (1ULL << 2)
/*
* This mask defines the set of flags which are "portable", meaning
* that they can be preserved when doing a raw encrypted zfs send.
* Flags included in this mask will be protected by os_portable_mac
* when the block of dnodes is encrypted. No portable flags currently
* exist.
*/
#define OBJSET_CRYPT_PORTABLE_FLAGS_MASK (0)
+#if defined(__clang__)
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Wgnu-variable-sized-type-not-at-end"
+#endif
typedef struct objset_phys {
dnode_phys_t os_meta_dnode;
zil_header_t os_zil_header;
uint64_t os_type;
uint64_t os_flags;
uint8_t os_portable_mac[ZIO_OBJSET_MAC_LEN];
uint8_t os_local_mac[ZIO_OBJSET_MAC_LEN];
char os_pad0[OBJSET_PHYS_SIZE_V2 - sizeof (dnode_phys_t)*3 -
sizeof (zil_header_t) - sizeof (uint64_t)*2 -
2*ZIO_OBJSET_MAC_LEN];
dnode_phys_t os_userused_dnode;
dnode_phys_t os_groupused_dnode;
dnode_phys_t os_projectused_dnode;
char os_pad1[OBJSET_PHYS_SIZE_V3 - OBJSET_PHYS_SIZE_V2 -
sizeof (dnode_phys_t)];
} objset_phys_t;
+#if defined(__clang__)
+#pragma clang diagnostic pop
+#endif
typedef int (*dmu_objset_upgrade_cb_t)(objset_t *);
#define OBJSET_PROP_UNINITIALIZED ((uint64_t)-1)
struct objset {
/* Immutable: */
struct dsl_dataset *os_dsl_dataset;
spa_t *os_spa;
arc_buf_t *os_phys_buf;
objset_phys_t *os_phys;
boolean_t os_encrypted;
/*
* The following "special" dnodes have no parent, are exempt
* from dnode_move(), and are not recorded in os_dnodes, but they
* root their descendents in this objset using handles anyway, so
* that all access to dnodes from dbufs consistently uses handles.
*/
dnode_handle_t os_meta_dnode;
dnode_handle_t os_userused_dnode;
dnode_handle_t os_groupused_dnode;
dnode_handle_t os_projectused_dnode;
zilog_t *os_zil;
list_node_t os_evicting_node;
/* can change, under dsl_dir's locks: */
uint64_t os_dnodesize; /* default dnode size for new objects */
enum zio_checksum os_checksum;
enum zio_compress os_compress;
uint8_t os_complevel;
uint8_t os_copies;
enum zio_checksum os_dedup_checksum;
boolean_t os_dedup_verify;
zfs_logbias_op_t os_logbias;
zfs_cache_type_t os_primary_cache;
zfs_cache_type_t os_secondary_cache;
zfs_sync_type_t os_sync;
zfs_redundant_metadata_type_t os_redundant_metadata;
uint64_t os_recordsize;
/*
* The next four values are used as a cache of whatever's on disk, and
* are initialized the first time these properties are queried. Before
* being initialized with their real values, their values are
* OBJSET_PROP_UNINITIALIZED.
*/
uint64_t os_version;
uint64_t os_normalization;
uint64_t os_utf8only;
uint64_t os_casesensitivity;
/*
* The largest zpl file block allowed in special class.
* cached here instead of zfsvfs for easier access.
*/
int os_zpl_special_smallblock;
/*
* Pointer is constant; the blkptr it points to is protected by
* os_dsl_dataset->ds_bp_rwlock
*/
blkptr_t *os_rootbp;
/* no lock needed: */
struct dmu_tx *os_synctx; /* XXX sketchy */
zil_header_t os_zil_header;
multilist_t os_synced_dnodes;
uint64_t os_flags;
uint64_t os_freed_dnodes;
boolean_t os_rescan_dnodes;
boolean_t os_raw_receive;
/* os_phys_buf should be written raw next txg */
boolean_t os_next_write_raw[TXG_SIZE];
/* Protected by os_obj_lock */
kmutex_t os_obj_lock;
uint64_t os_obj_next_chunk;
/* Per-CPU next object to allocate, protected by atomic ops. */
uint64_t *os_obj_next_percpu;
int os_obj_next_percpu_len;
/* Protected by os_lock */
kmutex_t os_lock;
multilist_t os_dirty_dnodes[TXG_SIZE];
list_t os_dnodes;
list_t os_downgraded_dbufs;
/* Protects changes to DMU_{USER,GROUP,PROJECT}USED_OBJECT */
kmutex_t os_userused_lock;
/* stuff we store for the user */
kmutex_t os_user_ptr_lock;
void *os_user_ptr;
sa_os_t *os_sa;
/* kernel thread to upgrade this dataset */
kmutex_t os_upgrade_lock;
taskqid_t os_upgrade_id;
dmu_objset_upgrade_cb_t os_upgrade_cb;
boolean_t os_upgrade_exit;
int os_upgrade_status;
};
#define DMU_META_OBJSET 0
#define DMU_META_DNODE_OBJECT 0
#define DMU_OBJECT_IS_SPECIAL(obj) ((int64_t)(obj) <= 0)
#define DMU_META_DNODE(os) ((os)->os_meta_dnode.dnh_dnode)
#define DMU_USERUSED_DNODE(os) ((os)->os_userused_dnode.dnh_dnode)
#define DMU_GROUPUSED_DNODE(os) ((os)->os_groupused_dnode.dnh_dnode)
#define DMU_PROJECTUSED_DNODE(os) ((os)->os_projectused_dnode.dnh_dnode)
/* called from zpl */
int dmu_objset_hold(const char *name, const void *tag, objset_t **osp);
int dmu_objset_hold_flags(const char *name, boolean_t decrypt, const void *tag,
objset_t **osp);
int dmu_objset_own(const char *name, dmu_objset_type_t type,
boolean_t readonly, boolean_t decrypt, const void *tag, objset_t **osp);
int dmu_objset_own_obj(struct dsl_pool *dp, uint64_t obj,
dmu_objset_type_t type, boolean_t readonly, boolean_t decrypt,
const void *tag, objset_t **osp);
void dmu_objset_refresh_ownership(struct dsl_dataset *ds,
struct dsl_dataset **newds, boolean_t decrypt, const void *tag);
void dmu_objset_rele(objset_t *os, const void *tag);
void dmu_objset_rele_flags(objset_t *os, boolean_t decrypt, const void *tag);
void dmu_objset_disown(objset_t *os, boolean_t decrypt, const void *tag);
int dmu_objset_from_ds(struct dsl_dataset *ds, objset_t **osp);
void dmu_objset_stats(objset_t *os, nvlist_t *nv);
void dmu_objset_fast_stat(objset_t *os, dmu_objset_stats_t *stat);
void dmu_objset_space(objset_t *os, uint64_t *refdbytesp, uint64_t *availbytesp,
uint64_t *usedobjsp, uint64_t *availobjsp);
uint64_t dmu_objset_fsid_guid(objset_t *os);
int dmu_objset_find_dp(struct dsl_pool *dp, uint64_t ddobj,
int func(struct dsl_pool *, struct dsl_dataset *, void *),
void *arg, int flags);
void dmu_objset_evict_dbufs(objset_t *os);
inode_timespec_t dmu_objset_snap_cmtime(objset_t *os);
/* called from dsl */
void dmu_objset_sync(objset_t *os, zio_t *zio, dmu_tx_t *tx);
boolean_t dmu_objset_is_dirty(objset_t *os, uint64_t txg);
objset_t *dmu_objset_create_impl_dnstats(spa_t *spa, struct dsl_dataset *ds,
blkptr_t *bp, dmu_objset_type_t type, int levels, int blksz, int ibs,
dmu_tx_t *tx);
objset_t *dmu_objset_create_impl(spa_t *spa, struct dsl_dataset *ds,
blkptr_t *bp, dmu_objset_type_t type, dmu_tx_t *tx);
int dmu_objset_open_impl(spa_t *spa, struct dsl_dataset *ds, blkptr_t *bp,
objset_t **osp);
void dmu_objset_evict(objset_t *os);
void dmu_objset_sync_done(objset_t *os, dmu_tx_t *tx);
void dmu_objset_userquota_get_ids(dnode_t *dn, boolean_t before, dmu_tx_t *tx);
boolean_t dmu_objset_userused_enabled(objset_t *os);
void dmu_objset_userspace_upgrade(objset_t *os);
boolean_t dmu_objset_userspace_present(objset_t *os);
boolean_t dmu_objset_userobjused_enabled(objset_t *os);
boolean_t dmu_objset_userobjspace_upgradable(objset_t *os);
boolean_t dmu_objset_userobjspace_present(objset_t *os);
boolean_t dmu_objset_incompatible_encryption_version(objset_t *os);
boolean_t dmu_objset_projectquota_enabled(objset_t *os);
boolean_t dmu_objset_projectquota_present(objset_t *os);
boolean_t dmu_objset_projectquota_upgradable(objset_t *os);
void dmu_objset_id_quota_upgrade(objset_t *os);
int dmu_get_file_info(objset_t *os, dmu_object_type_t bonustype,
const void *data, zfs_file_info_t *zfi);
int dmu_fsname(const char *snapname, char *buf);
void dmu_objset_evict_done(objset_t *os);
void dmu_objset_willuse_space(objset_t *os, int64_t space, dmu_tx_t *tx);
void dmu_objset_init(void);
void dmu_objset_fini(void);
#ifdef __cplusplus
}
#endif
#endif /* _SYS_DMU_OBJSET_H */
diff --git a/sys/contrib/openzfs/include/sys/dnode.h b/sys/contrib/openzfs/include/sys/dnode.h
index 2d741ea36bd0..dbe7350d4da7 100644
--- a/sys/contrib/openzfs/include/sys/dnode.h
+++ b/sys/contrib/openzfs/include/sys/dnode.h
@@ -1,657 +1,665 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
*/
#ifndef _SYS_DNODE_H
#define _SYS_DNODE_H
#include <sys/zfs_context.h>
#include <sys/avl.h>
#include <sys/spa.h>
#include <sys/txg.h>
#include <sys/zio.h>
#include <sys/zfs_refcount.h>
#include <sys/dmu_zfetch.h>
#include <sys/zrlock.h>
#include <sys/multilist.h>
#include <sys/wmsum.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* dnode_hold() flags.
*/
#define DNODE_MUST_BE_ALLOCATED 1
#define DNODE_MUST_BE_FREE 2
#define DNODE_DRY_RUN 4
/*
* dnode_next_offset() flags.
*/
#define DNODE_FIND_HOLE 1
#define DNODE_FIND_BACKWARDS 2
#define DNODE_FIND_HAVELOCK 4
/*
* Fixed constants.
*/
#define DNODE_SHIFT 9 /* 512 bytes */
#define DN_MIN_INDBLKSHIFT 12 /* 4k */
/*
* If we ever increase this value beyond 20, we need to revisit all logic that
* does x << level * ebps to handle overflow. With a 1M indirect block size,
* 4 levels of indirect blocks would not be able to guarantee addressing an
* entire object, so 5 levels will be used, but 5 * (20 - 7) = 65.
*/
#define DN_MAX_INDBLKSHIFT 17 /* 128k */
#define DNODE_BLOCK_SHIFT 14 /* 16k */
#define DNODE_CORE_SIZE 64 /* 64 bytes for dnode sans blkptrs */
#define DN_MAX_OBJECT_SHIFT 48 /* 256 trillion (zfs_fid_t limit) */
#define DN_MAX_OFFSET_SHIFT 64 /* 2^64 bytes in a dnode */
/*
* dnode id flags
*
* Note: a file will never ever have its ids moved from bonus->spill
*/
#define DN_ID_CHKED_BONUS 0x1
#define DN_ID_CHKED_SPILL 0x2
#define DN_ID_OLD_EXIST 0x4
#define DN_ID_NEW_EXIST 0x8
/*
* Derived constants.
*/
#define DNODE_MIN_SIZE (1 << DNODE_SHIFT)
#define DNODE_MAX_SIZE (1 << DNODE_BLOCK_SHIFT)
#define DNODE_BLOCK_SIZE (1 << DNODE_BLOCK_SHIFT)
#define DNODE_MIN_SLOTS (DNODE_MIN_SIZE >> DNODE_SHIFT)
#define DNODE_MAX_SLOTS (DNODE_MAX_SIZE >> DNODE_SHIFT)
#define DN_BONUS_SIZE(dnsize) ((dnsize) - DNODE_CORE_SIZE - \
(1 << SPA_BLKPTRSHIFT))
#define DN_SLOTS_TO_BONUSLEN(slots) DN_BONUS_SIZE((slots) << DNODE_SHIFT)
#define DN_OLD_MAX_BONUSLEN (DN_BONUS_SIZE(DNODE_MIN_SIZE))
#define DN_MAX_NBLKPTR ((DNODE_MIN_SIZE - DNODE_CORE_SIZE) >> SPA_BLKPTRSHIFT)
#define DN_MAX_OBJECT (1ULL << DN_MAX_OBJECT_SHIFT)
#define DN_ZERO_BONUSLEN (DN_BONUS_SIZE(DNODE_MAX_SIZE) + 1)
#define DN_KILL_SPILLBLK (1)
#define DN_SLOT_UNINIT ((void *)NULL) /* Uninitialized */
#define DN_SLOT_FREE ((void *)1UL) /* Free slot */
#define DN_SLOT_ALLOCATED ((void *)2UL) /* Allocated slot */
#define DN_SLOT_INTERIOR ((void *)3UL) /* Interior allocated slot */
#define DN_SLOT_IS_PTR(dn) ((void *)dn > DN_SLOT_INTERIOR)
#define DN_SLOT_IS_VALID(dn) ((void *)dn != NULL)
#define DNODES_PER_BLOCK_SHIFT (DNODE_BLOCK_SHIFT - DNODE_SHIFT)
#define DNODES_PER_BLOCK (1ULL << DNODES_PER_BLOCK_SHIFT)
/*
* This is inaccurate if the indblkshift of the particular object is not the
* max. But it's only used by userland to calculate the zvol reservation.
*/
#define DNODES_PER_LEVEL_SHIFT (DN_MAX_INDBLKSHIFT - SPA_BLKPTRSHIFT)
#define DNODES_PER_LEVEL (1ULL << DNODES_PER_LEVEL_SHIFT)
#define DN_MAX_LEVELS (DIV_ROUND_UP(DN_MAX_OFFSET_SHIFT - SPA_MINBLOCKSHIFT, \
DN_MIN_INDBLKSHIFT - SPA_BLKPTRSHIFT) + 1)
-#define DN_BONUS(dnp) ((void*)((dnp)->dn_bonus + \
+/*
+ * Use the flexible array instead of the fixed length one dn_bonus
+ * to address memcpy/memmove fortify error
+ */
+#define DN_BONUS(dnp) ((void*)((dnp)->dn_bonus_flexible + \
(((dnp)->dn_nblkptr - 1) * sizeof (blkptr_t))))
#define DN_MAX_BONUS_LEN(dnp) \
((dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) ? \
(uint8_t *)DN_SPILL_BLKPTR(dnp) - (uint8_t *)DN_BONUS(dnp) : \
(uint8_t *)(dnp + (dnp->dn_extra_slots + 1)) - (uint8_t *)DN_BONUS(dnp))
#define DN_USED_BYTES(dnp) (((dnp)->dn_flags & DNODE_FLAG_USED_BYTES) ? \
(dnp)->dn_used : (dnp)->dn_used << SPA_MINBLOCKSHIFT)
#define EPB(blkshift, typeshift) (1 << (blkshift - typeshift))
struct dmu_buf_impl;
struct objset;
struct zio;
enum dnode_dirtycontext {
DN_UNDIRTIED,
DN_DIRTY_OPEN,
DN_DIRTY_SYNC
};
/* Is dn_used in bytes? if not, it's in multiples of SPA_MINBLOCKSIZE */
#define DNODE_FLAG_USED_BYTES (1 << 0)
#define DNODE_FLAG_USERUSED_ACCOUNTED (1 << 1)
/* Does dnode have a SA spill blkptr in bonus? */
#define DNODE_FLAG_SPILL_BLKPTR (1 << 2)
/* User/Group/Project dnode accounting */
#define DNODE_FLAG_USEROBJUSED_ACCOUNTED (1 << 3)
/*
* This mask defines the set of flags which are "portable", meaning
* that they can be preserved when doing a raw encrypted zfs send.
* Flags included in this mask will be protected by AAD when the block
* of dnodes is encrypted.
*/
#define DNODE_CRYPT_PORTABLE_FLAGS_MASK (DNODE_FLAG_SPILL_BLKPTR)
/*
* VARIABLE-LENGTH (LARGE) DNODES
*
* The motivation for variable-length dnodes is to eliminate the overhead
* associated with using spill blocks. Spill blocks are used to store
* system attribute data (i.e. file metadata) that does not fit in the
* dnode's bonus buffer. By allowing a larger bonus buffer area the use of
* a spill block can be avoided. Spill blocks potentially incur an
* additional read I/O for every dnode in a dnode block. As a worst case
* example, reading 32 dnodes from a 16k dnode block and all of the spill
* blocks could issue 33 separate reads. Now suppose those dnodes have size
* 1024 and therefore don't need spill blocks. Then the worst case number
* of blocks read is reduced from 33 to two--one per dnode block.
*
* ZFS-on-Linux systems that make heavy use of extended attributes benefit
* from this feature. In particular, ZFS-on-Linux supports the xattr=sa
* dataset property which allows file extended attribute data to be stored
* in the dnode bonus buffer as an alternative to the traditional
* directory-based format. Workloads such as SELinux and the Lustre
* distributed filesystem often store enough xattr data to force spill
* blocks when xattr=sa is in effect. Large dnodes may therefore provide a
* performance benefit to such systems. Other use cases that benefit from
* this feature include files with large ACLs and symbolic links with long
* target names.
*
* The size of a dnode may be a multiple of 512 bytes up to the size of a
* dnode block (currently 16384 bytes). The dn_extra_slots field of the
* on-disk dnode_phys_t structure describes the size of the physical dnode
* on disk. The field represents how many "extra" dnode_phys_t slots a
* dnode consumes in its dnode block. This convention results in a value of
* 0 for 512 byte dnodes which preserves on-disk format compatibility with
* older software which doesn't support large dnodes.
*
* Similarly, the in-memory dnode_t structure has a dn_num_slots field
* to represent the total number of dnode_phys_t slots consumed on disk.
* Thus dn->dn_num_slots is 1 greater than the corresponding
* dnp->dn_extra_slots. This difference in convention was adopted
* because, unlike on-disk structures, backward compatibility is not a
* concern for in-memory objects, so we used a more natural way to
* represent size for a dnode_t.
*
* The default size for newly created dnodes is determined by the value of
* the "dnodesize" dataset property. By default the property is set to
* "legacy" which is compatible with older software. Setting the property
* to "auto" will allow the filesystem to choose the most suitable dnode
* size. Currently this just sets the default dnode size to 1k, but future
* code improvements could dynamically choose a size based on observed
* workload patterns. Dnodes of varying sizes can coexist within the same
* dataset and even within the same dnode block.
*/
typedef struct dnode_phys {
uint8_t dn_type; /* dmu_object_type_t */
uint8_t dn_indblkshift; /* ln2(indirect block size) */
uint8_t dn_nlevels; /* 1=dn_blkptr->data blocks */
uint8_t dn_nblkptr; /* length of dn_blkptr */
uint8_t dn_bonustype; /* type of data in bonus buffer */
uint8_t dn_checksum; /* ZIO_CHECKSUM type */
uint8_t dn_compress; /* ZIO_COMPRESS type */
uint8_t dn_flags; /* DNODE_FLAG_* */
uint16_t dn_datablkszsec; /* data block size in 512b sectors */
uint16_t dn_bonuslen; /* length of dn_bonus */
uint8_t dn_extra_slots; /* # of subsequent slots consumed */
uint8_t dn_pad2[3];
/* accounting is protected by dn_dirty_mtx */
uint64_t dn_maxblkid; /* largest allocated block ID */
uint64_t dn_used; /* bytes (or sectors) of disk space */
/*
* Both dn_pad2 and dn_pad3 are protected by the block's MAC. This
* allows us to protect any fields that might be added here in the
* future. In either case, developers will want to check
* zio_crypt_init_uios_dnode() and zio_crypt_do_dnode_hmac_updates()
* to ensure the new field is being protected and updated properly.
*/
uint64_t dn_pad3[4];
/*
* The tail region is 448 bytes for a 512 byte dnode, and
* correspondingly larger for larger dnode sizes. The spill
* block pointer, when present, is always at the end of the tail
* region. There are three ways this space may be used, using
* a 512 byte dnode for this diagram:
*
* 0 64 128 192 256 320 384 448 (offset)
* +---------------+---------------+---------------+-------+
* | dn_blkptr[0] | dn_blkptr[1] | dn_blkptr[2] | / |
* +---------------+---------------+---------------+-------+
* | dn_blkptr[0] | dn_bonus[0..319] |
* +---------------+-----------------------+---------------+
* | dn_blkptr[0] | dn_bonus[0..191] | dn_spill |
* +---------------+-----------------------+---------------+
*/
union {
blkptr_t dn_blkptr[1+DN_OLD_MAX_BONUSLEN/sizeof (blkptr_t)];
struct {
blkptr_t __dn_ignore1;
uint8_t dn_bonus[DN_OLD_MAX_BONUSLEN];
};
struct {
blkptr_t __dn_ignore2;
uint8_t __dn_ignore3[DN_OLD_MAX_BONUSLEN -
sizeof (blkptr_t)];
blkptr_t dn_spill;
};
+ struct {
+ blkptr_t __dn_ignore4;
+ uint8_t dn_bonus_flexible[];
+ };
};
} dnode_phys_t;
#define DN_SPILL_BLKPTR(dnp) ((blkptr_t *)((char *)(dnp) + \
(((dnp)->dn_extra_slots + 1) << DNODE_SHIFT) - (1 << SPA_BLKPTRSHIFT)))
struct dnode {
/*
* Protects the structure of the dnode, including the number of levels
* of indirection (dn_nlevels), dn_maxblkid, and dn_next_*
*/
krwlock_t dn_struct_rwlock;
/* Our link on dn_objset->os_dnodes list; protected by os_lock. */
list_node_t dn_link;
/* immutable: */
struct objset *dn_objset;
uint64_t dn_object;
struct dmu_buf_impl *dn_dbuf;
struct dnode_handle *dn_handle;
dnode_phys_t *dn_phys; /* pointer into dn->dn_dbuf->db.db_data */
/*
* Copies of stuff in dn_phys. They're valid in the open
* context (eg. even before the dnode is first synced).
* Where necessary, these are protected by dn_struct_rwlock.
*/
dmu_object_type_t dn_type; /* object type */
uint16_t dn_bonuslen; /* bonus length */
uint8_t dn_bonustype; /* bonus type */
uint8_t dn_nblkptr; /* number of blkptrs (immutable) */
uint8_t dn_checksum; /* ZIO_CHECKSUM type */
uint8_t dn_compress; /* ZIO_COMPRESS type */
uint8_t dn_nlevels;
uint8_t dn_indblkshift;
uint8_t dn_datablkshift; /* zero if blksz not power of 2! */
uint8_t dn_moved; /* Has this dnode been moved? */
uint16_t dn_datablkszsec; /* in 512b sectors */
uint32_t dn_datablksz; /* in bytes */
uint64_t dn_maxblkid;
uint8_t dn_next_type[TXG_SIZE];
uint8_t dn_num_slots; /* metadnode slots consumed on disk */
uint8_t dn_next_nblkptr[TXG_SIZE];
uint8_t dn_next_nlevels[TXG_SIZE];
uint8_t dn_next_indblkshift[TXG_SIZE];
uint8_t dn_next_bonustype[TXG_SIZE];
uint8_t dn_rm_spillblk[TXG_SIZE]; /* for removing spill blk */
uint16_t dn_next_bonuslen[TXG_SIZE];
uint32_t dn_next_blksz[TXG_SIZE]; /* next block size in bytes */
uint64_t dn_next_maxblkid[TXG_SIZE]; /* next maxblkid in bytes */
/* protected by dn_dbufs_mtx; declared here to fill 32-bit hole */
uint32_t dn_dbufs_count; /* count of dn_dbufs */
/* protected by os_lock: */
multilist_node_t dn_dirty_link[TXG_SIZE]; /* next on dataset's dirty */
/* protected by dn_mtx: */
kmutex_t dn_mtx;
list_t dn_dirty_records[TXG_SIZE];
struct range_tree *dn_free_ranges[TXG_SIZE];
uint64_t dn_allocated_txg;
uint64_t dn_free_txg;
uint64_t dn_assigned_txg;
uint64_t dn_dirty_txg; /* txg dnode was last dirtied */
kcondvar_t dn_notxholds;
kcondvar_t dn_nodnholds;
enum dnode_dirtycontext dn_dirtyctx;
const void *dn_dirtyctx_firstset; /* dbg: contents meaningless */
/* protected by own devices */
zfs_refcount_t dn_tx_holds;
zfs_refcount_t dn_holds;
kmutex_t dn_dbufs_mtx;
/*
* Descendent dbufs, ordered by dbuf_compare. Note that dn_dbufs
* can contain multiple dbufs of the same (level, blkid) when a
* dbuf is marked DB_EVICTING without being removed from
* dn_dbufs. To maintain the avl invariant that there cannot be
* duplicate entries, we order the dbufs by an arbitrary value -
* their address in memory. This means that dn_dbufs cannot be used to
* directly look up a dbuf. Instead, callers must use avl_walk, have
* a reference to the dbuf, or look up a non-existent node with
* db_state = DB_SEARCH (see dbuf_free_range for an example).
*/
avl_tree_t dn_dbufs;
/* protected by dn_struct_rwlock */
struct dmu_buf_impl *dn_bonus; /* bonus buffer dbuf */
boolean_t dn_have_spill; /* have spill or are spilling */
/* parent IO for current sync write */
zio_t *dn_zio;
/* used in syncing context */
uint64_t dn_oldused; /* old phys used bytes */
uint64_t dn_oldflags; /* old phys dn_flags */
uint64_t dn_olduid, dn_oldgid, dn_oldprojid;
uint64_t dn_newuid, dn_newgid, dn_newprojid;
int dn_id_flags;
/* holds prefetch structure */
struct zfetch dn_zfetch;
};
/*
* Since AVL already has embedded element counter, use dn_dbufs_count
* only for dbufs not counted there (bonus buffers) and just add them.
*/
#define DN_DBUFS_COUNT(dn) ((dn)->dn_dbufs_count + \
avl_numnodes(&(dn)->dn_dbufs))
/*
* We use this (otherwise unused) bit to indicate if the value of
* dn_next_maxblkid[txgoff] is valid to use in dnode_sync().
*/
#define DMU_NEXT_MAXBLKID_SET (1ULL << 63)
/*
* Adds a level of indirection between the dbuf and the dnode to avoid
* iterating descendent dbufs in dnode_move(). Handles are not allocated
* individually, but as an array of child dnodes in dnode_hold_impl().
*/
typedef struct dnode_handle {
/* Protects dnh_dnode from modification by dnode_move(). */
zrlock_t dnh_zrlock;
dnode_t *dnh_dnode;
} dnode_handle_t;
typedef struct dnode_children {
dmu_buf_user_t dnc_dbu; /* User evict data */
size_t dnc_count; /* number of children */
dnode_handle_t dnc_children[]; /* sized dynamically */
} dnode_children_t;
typedef struct free_range {
avl_node_t fr_node;
uint64_t fr_blkid;
uint64_t fr_nblks;
} free_range_t;
void dnode_special_open(struct objset *dd, dnode_phys_t *dnp,
uint64_t object, dnode_handle_t *dnh);
void dnode_special_close(dnode_handle_t *dnh);
void dnode_setbonuslen(dnode_t *dn, int newsize, dmu_tx_t *tx);
void dnode_setbonus_type(dnode_t *dn, dmu_object_type_t, dmu_tx_t *tx);
void dnode_rm_spill(dnode_t *dn, dmu_tx_t *tx);
int dnode_hold(struct objset *dd, uint64_t object,
const void *ref, dnode_t **dnp);
int dnode_hold_impl(struct objset *dd, uint64_t object, int flag, int dn_slots,
const void *ref, dnode_t **dnp);
boolean_t dnode_add_ref(dnode_t *dn, const void *ref);
void dnode_rele(dnode_t *dn, const void *ref);
void dnode_rele_and_unlock(dnode_t *dn, const void *tag, boolean_t evicting);
int dnode_try_claim(objset_t *os, uint64_t object, int slots);
boolean_t dnode_is_dirty(dnode_t *dn);
void dnode_setdirty(dnode_t *dn, dmu_tx_t *tx);
void dnode_set_dirtyctx(dnode_t *dn, dmu_tx_t *tx, const void *tag);
void dnode_sync(dnode_t *dn, dmu_tx_t *tx);
void dnode_allocate(dnode_t *dn, dmu_object_type_t ot, int blocksize, int ibs,
dmu_object_type_t bonustype, int bonuslen, int dn_slots, dmu_tx_t *tx);
void dnode_reallocate(dnode_t *dn, dmu_object_type_t ot, int blocksize,
dmu_object_type_t bonustype, int bonuslen, int dn_slots,
boolean_t keep_spill, dmu_tx_t *tx);
void dnode_free(dnode_t *dn, dmu_tx_t *tx);
void dnode_byteswap(dnode_phys_t *dnp);
void dnode_buf_byteswap(void *buf, size_t size);
void dnode_verify(dnode_t *dn);
int dnode_set_nlevels(dnode_t *dn, int nlevels, dmu_tx_t *tx);
int dnode_set_blksz(dnode_t *dn, uint64_t size, int ibs, dmu_tx_t *tx);
void dnode_free_range(dnode_t *dn, uint64_t off, uint64_t len, dmu_tx_t *tx);
void dnode_diduse_space(dnode_t *dn, int64_t space);
void dnode_new_blkid(dnode_t *dn, uint64_t blkid, dmu_tx_t *tx,
boolean_t have_read, boolean_t force);
uint64_t dnode_block_freed(dnode_t *dn, uint64_t blkid);
void dnode_init(void);
void dnode_fini(void);
int dnode_next_offset(dnode_t *dn, int flags, uint64_t *off,
int minlvl, uint64_t blkfill, uint64_t txg);
void dnode_evict_dbufs(dnode_t *dn);
void dnode_evict_bonus(dnode_t *dn);
void dnode_free_interior_slots(dnode_t *dn);
#define DNODE_IS_DIRTY(_dn) \
((_dn)->dn_dirty_txg >= spa_syncing_txg((_dn)->dn_objset->os_spa))
#define DNODE_LEVEL_IS_CACHEABLE(_dn, _level) \
((_dn)->dn_objset->os_primary_cache == ZFS_CACHE_ALL || \
(((_level) > 0 || DMU_OT_IS_METADATA((_dn)->dn_type)) && \
(_dn)->dn_objset->os_primary_cache == ZFS_CACHE_METADATA))
/*
* Used for dnodestats kstat.
*/
typedef struct dnode_stats {
/*
* Number of failed attempts to hold a meta dnode dbuf.
*/
kstat_named_t dnode_hold_dbuf_hold;
/*
* Number of failed attempts to read a meta dnode dbuf.
*/
kstat_named_t dnode_hold_dbuf_read;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_ALLOCATED) was able
* to hold the requested object number which was allocated. This is
* the common case when looking up any allocated object number.
*/
kstat_named_t dnode_hold_alloc_hits;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_ALLOCATED) was not
* able to hold the request object number because it was not allocated.
*/
kstat_named_t dnode_hold_alloc_misses;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_ALLOCATED) was not
* able to hold the request object number because the object number
* refers to an interior large dnode slot.
*/
kstat_named_t dnode_hold_alloc_interior;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_ALLOCATED) needed
* to retry acquiring slot zrl locks due to contention.
*/
kstat_named_t dnode_hold_alloc_lock_retry;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_ALLOCATED) did not
* need to create the dnode because another thread did so after
* dropping the read lock but before acquiring the write lock.
*/
kstat_named_t dnode_hold_alloc_lock_misses;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_ALLOCATED) found
* a free dnode instantiated by dnode_create() but not yet allocated
* by dnode_allocate().
*/
kstat_named_t dnode_hold_alloc_type_none;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_FREE) was able
* to hold the requested range of free dnode slots.
*/
kstat_named_t dnode_hold_free_hits;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_FREE) was not
* able to hold the requested range of free dnode slots because
* at least one slot was allocated.
*/
kstat_named_t dnode_hold_free_misses;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_FREE) was not
* able to hold the requested range of free dnode slots because
* after acquiring the zrl lock at least one slot was allocated.
*/
kstat_named_t dnode_hold_free_lock_misses;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_FREE) needed
* to retry acquiring slot zrl locks due to contention.
*/
kstat_named_t dnode_hold_free_lock_retry;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_FREE) requested
* a range of dnode slots which were held by another thread.
*/
kstat_named_t dnode_hold_free_refcount;
/*
* Number of times dnode_hold(..., DNODE_MUST_BE_FREE) requested
* a range of dnode slots which would overflow the dnode_phys_t.
*/
kstat_named_t dnode_hold_free_overflow;
/*
* Number of times dnode_free_interior_slots() needed to retry
* acquiring a slot zrl lock due to contention.
*/
kstat_named_t dnode_free_interior_lock_retry;
/*
* Number of new dnodes allocated by dnode_allocate().
*/
kstat_named_t dnode_allocate;
/*
* Number of dnodes re-allocated by dnode_reallocate().
*/
kstat_named_t dnode_reallocate;
/*
* Number of meta dnode dbufs evicted.
*/
kstat_named_t dnode_buf_evict;
/*
* Number of times dmu_object_alloc*() reached the end of the existing
* object ID chunk and advanced to a new one.
*/
kstat_named_t dnode_alloc_next_chunk;
/*
* Number of times multiple threads attempted to allocate a dnode
* from the same block of free dnodes.
*/
kstat_named_t dnode_alloc_race;
/*
* Number of times dmu_object_alloc*() was forced to advance to the
* next meta dnode dbuf due to an error from dmu_object_next().
*/
kstat_named_t dnode_alloc_next_block;
/*
* Statistics for tracking dnodes which have been moved.
*/
kstat_named_t dnode_move_invalid;
kstat_named_t dnode_move_recheck1;
kstat_named_t dnode_move_recheck2;
kstat_named_t dnode_move_special;
kstat_named_t dnode_move_handle;
kstat_named_t dnode_move_rwlock;
kstat_named_t dnode_move_active;
} dnode_stats_t;
typedef struct dnode_sums {
wmsum_t dnode_hold_dbuf_hold;
wmsum_t dnode_hold_dbuf_read;
wmsum_t dnode_hold_alloc_hits;
wmsum_t dnode_hold_alloc_misses;
wmsum_t dnode_hold_alloc_interior;
wmsum_t dnode_hold_alloc_lock_retry;
wmsum_t dnode_hold_alloc_lock_misses;
wmsum_t dnode_hold_alloc_type_none;
wmsum_t dnode_hold_free_hits;
wmsum_t dnode_hold_free_misses;
wmsum_t dnode_hold_free_lock_misses;
wmsum_t dnode_hold_free_lock_retry;
wmsum_t dnode_hold_free_refcount;
wmsum_t dnode_hold_free_overflow;
wmsum_t dnode_free_interior_lock_retry;
wmsum_t dnode_allocate;
wmsum_t dnode_reallocate;
wmsum_t dnode_buf_evict;
wmsum_t dnode_alloc_next_chunk;
wmsum_t dnode_alloc_race;
wmsum_t dnode_alloc_next_block;
wmsum_t dnode_move_invalid;
wmsum_t dnode_move_recheck1;
wmsum_t dnode_move_recheck2;
wmsum_t dnode_move_special;
wmsum_t dnode_move_handle;
wmsum_t dnode_move_rwlock;
wmsum_t dnode_move_active;
} dnode_sums_t;
extern dnode_stats_t dnode_stats;
extern dnode_sums_t dnode_sums;
#define DNODE_STAT_INCR(stat, val) \
wmsum_add(&dnode_sums.stat, (val))
#define DNODE_STAT_BUMP(stat) \
DNODE_STAT_INCR(stat, 1);
#ifdef ZFS_DEBUG
#define dprintf_dnode(dn, fmt, ...) do { \
if (zfs_flags & ZFS_DEBUG_DPRINTF) { \
char __db_buf[32]; \
uint64_t __db_obj = (dn)->dn_object; \
if (__db_obj == DMU_META_DNODE_OBJECT) \
(void) strlcpy(__db_buf, "mdn", sizeof (__db_buf)); \
else \
(void) snprintf(__db_buf, sizeof (__db_buf), "%lld", \
(u_longlong_t)__db_obj);\
dprintf_ds((dn)->dn_objset->os_dsl_dataset, "obj=%s " fmt, \
__db_buf, __VA_ARGS__); \
} \
} while (0)
#define DNODE_VERIFY(dn) dnode_verify(dn)
#define FREE_VERIFY(db, start, end, tx) free_verify(db, start, end, tx)
#else
#define dprintf_dnode(db, fmt, ...)
#define DNODE_VERIFY(dn) ((void) sizeof ((uintptr_t)(dn)))
#define FREE_VERIFY(db, start, end, tx)
#endif
#ifdef __cplusplus
}
#endif
#endif /* _SYS_DNODE_H */
diff --git a/sys/contrib/openzfs/include/sys/fs/zfs.h b/sys/contrib/openzfs/include/sys/fs/zfs.h
index 25babd4ea8cf..0734ff12280e 100644
--- a/sys/contrib/openzfs/include/sys/fs/zfs.h
+++ b/sys/contrib/openzfs/include/sys/fs/zfs.h
@@ -1,1840 +1,1841 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2013, 2017 Joyent, Inc. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019 Datto Inc.
* Portions Copyright 2010 Robert Milkowski
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
* Copyright (c) 2022 Hewlett Packard Enterprise Development LP.
*/
#ifndef _SYS_FS_ZFS_H
#define _SYS_FS_ZFS_H extern __attribute__((visibility("default")))
#include <sys/time.h>
#include <sys/zio_priority.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* Types and constants shared between userland and the kernel.
*/
/*
* Each dataset can be one of the following types. These constants can be
* combined into masks that can be passed to various functions.
*/
typedef enum {
ZFS_TYPE_INVALID = 0,
ZFS_TYPE_FILESYSTEM = (1 << 0),
ZFS_TYPE_SNAPSHOT = (1 << 1),
ZFS_TYPE_VOLUME = (1 << 2),
ZFS_TYPE_POOL = (1 << 3),
ZFS_TYPE_BOOKMARK = (1 << 4),
ZFS_TYPE_VDEV = (1 << 5),
} zfs_type_t;
/*
* NB: lzc_dataset_type should be updated whenever a new objset type is added,
* if it represents a real type of a dataset that can be created from userland.
*/
typedef enum dmu_objset_type {
DMU_OST_NONE,
DMU_OST_META,
DMU_OST_ZFS,
DMU_OST_ZVOL,
DMU_OST_OTHER, /* For testing only! */
DMU_OST_ANY, /* Be careful! */
DMU_OST_NUMTYPES
} dmu_objset_type_t;
#define ZFS_TYPE_DATASET \
(ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME | ZFS_TYPE_SNAPSHOT)
/*
* All of these include the terminating NUL byte.
*/
#define ZAP_MAXNAMELEN 256
#define ZAP_MAXVALUELEN (1024 * 8)
#define ZAP_OLDMAXVALUELEN 1024
#define ZFS_MAX_DATASET_NAME_LEN 256
/*
* Dataset properties are identified by these constants and must be added to
* the end of this list to ensure that external consumers are not affected
* by the change. If you make any changes to this list, be sure to update
* the property table in module/zcommon/zfs_prop.c.
*/
typedef enum {
ZPROP_CONT = -2,
ZPROP_INVAL = -1,
ZPROP_USERPROP = ZPROP_INVAL,
ZFS_PROP_TYPE = 0,
ZFS_PROP_CREATION,
ZFS_PROP_USED,
ZFS_PROP_AVAILABLE,
ZFS_PROP_REFERENCED,
ZFS_PROP_COMPRESSRATIO,
ZFS_PROP_MOUNTED,
ZFS_PROP_ORIGIN,
ZFS_PROP_QUOTA,
ZFS_PROP_RESERVATION,
ZFS_PROP_VOLSIZE,
ZFS_PROP_VOLBLOCKSIZE,
ZFS_PROP_RECORDSIZE,
ZFS_PROP_MOUNTPOINT,
ZFS_PROP_SHARENFS,
ZFS_PROP_CHECKSUM,
ZFS_PROP_COMPRESSION,
ZFS_PROP_ATIME,
ZFS_PROP_DEVICES,
ZFS_PROP_EXEC,
ZFS_PROP_SETUID,
ZFS_PROP_READONLY,
ZFS_PROP_ZONED,
ZFS_PROP_SNAPDIR,
ZFS_PROP_ACLMODE,
ZFS_PROP_ACLINHERIT,
ZFS_PROP_CREATETXG,
ZFS_PROP_NAME, /* not exposed to the user */
ZFS_PROP_CANMOUNT,
ZFS_PROP_ISCSIOPTIONS, /* not exposed to the user */
ZFS_PROP_XATTR,
ZFS_PROP_NUMCLONES, /* not exposed to the user */
ZFS_PROP_COPIES,
ZFS_PROP_VERSION,
ZFS_PROP_UTF8ONLY,
ZFS_PROP_NORMALIZE,
ZFS_PROP_CASE,
ZFS_PROP_VSCAN,
ZFS_PROP_NBMAND,
ZFS_PROP_SHARESMB,
ZFS_PROP_REFQUOTA,
ZFS_PROP_REFRESERVATION,
ZFS_PROP_GUID,
ZFS_PROP_PRIMARYCACHE,
ZFS_PROP_SECONDARYCACHE,
ZFS_PROP_USEDSNAP,
ZFS_PROP_USEDDS,
ZFS_PROP_USEDCHILD,
ZFS_PROP_USEDREFRESERV,
ZFS_PROP_USERACCOUNTING, /* not exposed to the user */
ZFS_PROP_STMF_SHAREINFO, /* not exposed to the user */
ZFS_PROP_DEFER_DESTROY,
ZFS_PROP_USERREFS,
ZFS_PROP_LOGBIAS,
ZFS_PROP_UNIQUE, /* not exposed to the user */
ZFS_PROP_OBJSETID,
ZFS_PROP_DEDUP,
ZFS_PROP_MLSLABEL,
ZFS_PROP_SYNC,
ZFS_PROP_DNODESIZE,
ZFS_PROP_REFRATIO,
ZFS_PROP_WRITTEN,
ZFS_PROP_CLONES,
ZFS_PROP_LOGICALUSED,
ZFS_PROP_LOGICALREFERENCED,
ZFS_PROP_INCONSISTENT, /* not exposed to the user */
ZFS_PROP_VOLMODE,
ZFS_PROP_FILESYSTEM_LIMIT,
ZFS_PROP_SNAPSHOT_LIMIT,
ZFS_PROP_FILESYSTEM_COUNT,
ZFS_PROP_SNAPSHOT_COUNT,
ZFS_PROP_SNAPDEV,
ZFS_PROP_ACLTYPE,
ZFS_PROP_SELINUX_CONTEXT,
ZFS_PROP_SELINUX_FSCONTEXT,
ZFS_PROP_SELINUX_DEFCONTEXT,
ZFS_PROP_SELINUX_ROOTCONTEXT,
ZFS_PROP_RELATIME,
ZFS_PROP_REDUNDANT_METADATA,
ZFS_PROP_OVERLAY,
ZFS_PROP_PREV_SNAP,
ZFS_PROP_RECEIVE_RESUME_TOKEN,
ZFS_PROP_ENCRYPTION,
ZFS_PROP_KEYLOCATION,
ZFS_PROP_KEYFORMAT,
ZFS_PROP_PBKDF2_SALT,
ZFS_PROP_PBKDF2_ITERS,
ZFS_PROP_ENCRYPTION_ROOT,
ZFS_PROP_KEY_GUID,
ZFS_PROP_KEYSTATUS,
ZFS_PROP_REMAPTXG, /* obsolete - no longer used */
ZFS_PROP_SPECIAL_SMALL_BLOCKS,
ZFS_PROP_IVSET_GUID, /* not exposed to the user */
ZFS_PROP_REDACTED,
ZFS_PROP_REDACT_SNAPS,
ZFS_PROP_SNAPSHOTS_CHANGED,
ZFS_NUM_PROPS
} zfs_prop_t;
typedef enum {
ZFS_PROP_USERUSED,
ZFS_PROP_USERQUOTA,
ZFS_PROP_GROUPUSED,
ZFS_PROP_GROUPQUOTA,
ZFS_PROP_USEROBJUSED,
ZFS_PROP_USEROBJQUOTA,
ZFS_PROP_GROUPOBJUSED,
ZFS_PROP_GROUPOBJQUOTA,
ZFS_PROP_PROJECTUSED,
ZFS_PROP_PROJECTQUOTA,
ZFS_PROP_PROJECTOBJUSED,
ZFS_PROP_PROJECTOBJQUOTA,
ZFS_NUM_USERQUOTA_PROPS
} zfs_userquota_prop_t;
_SYS_FS_ZFS_H const char *const zfs_userquota_prop_prefixes[
ZFS_NUM_USERQUOTA_PROPS];
/*
* Pool properties are identified by these constants and must be added to the
* end of this list to ensure that external consumers are not affected
* by the change. Properties must be registered in zfs_prop_init().
*/
typedef enum {
ZPOOL_PROP_INVAL = -1,
ZPOOL_PROP_NAME,
ZPOOL_PROP_SIZE,
ZPOOL_PROP_CAPACITY,
ZPOOL_PROP_ALTROOT,
ZPOOL_PROP_HEALTH,
ZPOOL_PROP_GUID,
ZPOOL_PROP_VERSION,
ZPOOL_PROP_BOOTFS,
ZPOOL_PROP_DELEGATION,
ZPOOL_PROP_AUTOREPLACE,
ZPOOL_PROP_CACHEFILE,
ZPOOL_PROP_FAILUREMODE,
ZPOOL_PROP_LISTSNAPS,
ZPOOL_PROP_AUTOEXPAND,
ZPOOL_PROP_DEDUPDITTO,
ZPOOL_PROP_DEDUPRATIO,
ZPOOL_PROP_FREE,
ZPOOL_PROP_ALLOCATED,
ZPOOL_PROP_READONLY,
ZPOOL_PROP_ASHIFT,
ZPOOL_PROP_COMMENT,
ZPOOL_PROP_EXPANDSZ,
ZPOOL_PROP_FREEING,
ZPOOL_PROP_FRAGMENTATION,
ZPOOL_PROP_LEAKED,
ZPOOL_PROP_MAXBLOCKSIZE,
ZPOOL_PROP_TNAME,
ZPOOL_PROP_MAXDNODESIZE,
ZPOOL_PROP_MULTIHOST,
ZPOOL_PROP_CHECKPOINT,
ZPOOL_PROP_LOAD_GUID,
ZPOOL_PROP_AUTOTRIM,
ZPOOL_PROP_COMPATIBILITY,
ZPOOL_PROP_BCLONEUSED,
ZPOOL_PROP_BCLONESAVED,
ZPOOL_PROP_BCLONERATIO,
ZPOOL_NUM_PROPS
} zpool_prop_t;
/* Small enough to not hog a whole line of printout in zpool(8). */
#define ZPROP_MAX_COMMENT 32
#define ZPROP_BOOLEAN_NA 2
#define ZPROP_VALUE "value"
#define ZPROP_SOURCE "source"
typedef enum {
ZPROP_SRC_NONE = 0x1,
ZPROP_SRC_DEFAULT = 0x2,
ZPROP_SRC_TEMPORARY = 0x4,
ZPROP_SRC_LOCAL = 0x8,
ZPROP_SRC_INHERITED = 0x10,
ZPROP_SRC_RECEIVED = 0x20
} zprop_source_t;
#define ZPROP_SRC_ALL 0x3f
#define ZPROP_SOURCE_VAL_RECVD "$recvd"
#define ZPROP_N_MORE_ERRORS "N_MORE_ERRORS"
/*
* Dataset flag implemented as a special entry in the props zap object
* indicating that the dataset has received properties on or after
* SPA_VERSION_RECVD_PROPS. The first such receive blows away local properties
* just as it did in earlier versions, and thereafter, local properties are
* preserved.
*/
#define ZPROP_HAS_RECVD "$hasrecvd"
typedef enum {
ZPROP_ERR_NOCLEAR = 0x1, /* failure to clear existing props */
ZPROP_ERR_NORESTORE = 0x2 /* failure to restore props on error */
} zprop_errflags_t;
typedef int (*zprop_func)(int, void *);
/*
* Properties to be set on the root file system of a new pool
* are stuffed into their own nvlist, which is then included in
* the properties nvlist with the pool properties.
*/
#define ZPOOL_ROOTFS_PROPS "root-props-nvl"
/*
* Length of 'written@' and 'written#'
*/
#define ZFS_WRITTEN_PROP_PREFIX_LEN 8
/*
* VDEV properties are identified by these constants and must be added to the
* end of this list to ensure that external consumers are not affected
* by the change. If you make any changes to this list, be sure to update
* the property table in usr/src/common/zfs/zpool_prop.c.
*/
typedef enum {
VDEV_PROP_INVAL = -1,
VDEV_PROP_USERPROP = VDEV_PROP_INVAL,
VDEV_PROP_NAME,
VDEV_PROP_CAPACITY,
VDEV_PROP_STATE,
VDEV_PROP_GUID,
VDEV_PROP_ASIZE,
VDEV_PROP_PSIZE,
VDEV_PROP_ASHIFT,
VDEV_PROP_SIZE,
VDEV_PROP_FREE,
VDEV_PROP_ALLOCATED,
VDEV_PROP_COMMENT,
VDEV_PROP_EXPANDSZ,
VDEV_PROP_FRAGMENTATION,
VDEV_PROP_BOOTSIZE,
VDEV_PROP_PARITY,
VDEV_PROP_PATH,
VDEV_PROP_DEVID,
VDEV_PROP_PHYS_PATH,
VDEV_PROP_ENC_PATH,
VDEV_PROP_FRU,
VDEV_PROP_PARENT,
VDEV_PROP_CHILDREN,
VDEV_PROP_NUMCHILDREN,
VDEV_PROP_READ_ERRORS,
VDEV_PROP_WRITE_ERRORS,
VDEV_PROP_CHECKSUM_ERRORS,
VDEV_PROP_INITIALIZE_ERRORS,
VDEV_PROP_OPS_NULL,
VDEV_PROP_OPS_READ,
VDEV_PROP_OPS_WRITE,
VDEV_PROP_OPS_FREE,
VDEV_PROP_OPS_CLAIM,
VDEV_PROP_OPS_TRIM,
VDEV_PROP_BYTES_NULL,
VDEV_PROP_BYTES_READ,
VDEV_PROP_BYTES_WRITE,
VDEV_PROP_BYTES_FREE,
VDEV_PROP_BYTES_CLAIM,
VDEV_PROP_BYTES_TRIM,
VDEV_PROP_REMOVING,
VDEV_PROP_ALLOCATING,
VDEV_PROP_FAILFAST,
VDEV_PROP_CHECKSUM_N,
VDEV_PROP_CHECKSUM_T,
VDEV_PROP_IO_N,
VDEV_PROP_IO_T,
VDEV_NUM_PROPS
} vdev_prop_t;
/*
* Dataset property functions shared between libzfs and kernel.
*/
_SYS_FS_ZFS_H const char *zfs_prop_default_string(zfs_prop_t);
_SYS_FS_ZFS_H uint64_t zfs_prop_default_numeric(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_readonly(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_visible(zfs_prop_t prop);
_SYS_FS_ZFS_H boolean_t zfs_prop_inheritable(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_setonce(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_encryption_key_param(zfs_prop_t);
_SYS_FS_ZFS_H boolean_t zfs_prop_valid_keylocation(const char *, boolean_t);
_SYS_FS_ZFS_H const char *zfs_prop_to_name(zfs_prop_t);
_SYS_FS_ZFS_H zfs_prop_t zfs_name_to_prop(const char *);
_SYS_FS_ZFS_H boolean_t zfs_prop_user(const char *);
_SYS_FS_ZFS_H boolean_t zfs_prop_userquota(const char *);
_SYS_FS_ZFS_H boolean_t zfs_prop_written(const char *);
_SYS_FS_ZFS_H int zfs_prop_index_to_string(zfs_prop_t, uint64_t, const char **);
_SYS_FS_ZFS_H int zfs_prop_string_to_index(zfs_prop_t, const char *,
uint64_t *);
_SYS_FS_ZFS_H uint64_t zfs_prop_random_value(zfs_prop_t, uint64_t seed);
_SYS_FS_ZFS_H boolean_t zfs_prop_valid_for_type(int, zfs_type_t, boolean_t);
/*
* Pool property functions shared between libzfs and kernel.
*/
_SYS_FS_ZFS_H zpool_prop_t zpool_name_to_prop(const char *);
_SYS_FS_ZFS_H const char *zpool_prop_to_name(zpool_prop_t);
_SYS_FS_ZFS_H const char *zpool_prop_default_string(zpool_prop_t);
_SYS_FS_ZFS_H uint64_t zpool_prop_default_numeric(zpool_prop_t);
_SYS_FS_ZFS_H boolean_t zpool_prop_readonly(zpool_prop_t);
_SYS_FS_ZFS_H boolean_t zpool_prop_setonce(zpool_prop_t);
_SYS_FS_ZFS_H boolean_t zpool_prop_feature(const char *);
_SYS_FS_ZFS_H boolean_t zpool_prop_unsupported(const char *);
_SYS_FS_ZFS_H int zpool_prop_index_to_string(zpool_prop_t, uint64_t,
const char **);
_SYS_FS_ZFS_H int zpool_prop_string_to_index(zpool_prop_t, const char *,
uint64_t *);
_SYS_FS_ZFS_H uint64_t zpool_prop_random_value(zpool_prop_t, uint64_t seed);
/*
* VDEV property functions shared between libzfs and kernel.
*/
_SYS_FS_ZFS_H vdev_prop_t vdev_name_to_prop(const char *);
_SYS_FS_ZFS_H boolean_t vdev_prop_user(const char *name);
_SYS_FS_ZFS_H const char *vdev_prop_to_name(vdev_prop_t);
_SYS_FS_ZFS_H const char *vdev_prop_default_string(vdev_prop_t);
_SYS_FS_ZFS_H uint64_t vdev_prop_default_numeric(vdev_prop_t);
_SYS_FS_ZFS_H boolean_t vdev_prop_readonly(vdev_prop_t prop);
_SYS_FS_ZFS_H int vdev_prop_index_to_string(vdev_prop_t, uint64_t,
const char **);
_SYS_FS_ZFS_H int vdev_prop_string_to_index(vdev_prop_t, const char *,
uint64_t *);
_SYS_FS_ZFS_H boolean_t zpool_prop_vdev(const char *name);
_SYS_FS_ZFS_H uint64_t vdev_prop_random_value(vdev_prop_t prop, uint64_t seed);
/*
* Definitions for the Delegation.
*/
typedef enum {
ZFS_DELEG_WHO_UNKNOWN = 0,
ZFS_DELEG_USER = 'u',
ZFS_DELEG_USER_SETS = 'U',
ZFS_DELEG_GROUP = 'g',
ZFS_DELEG_GROUP_SETS = 'G',
ZFS_DELEG_EVERYONE = 'e',
ZFS_DELEG_EVERYONE_SETS = 'E',
ZFS_DELEG_CREATE = 'c',
ZFS_DELEG_CREATE_SETS = 'C',
ZFS_DELEG_NAMED_SET = 's',
ZFS_DELEG_NAMED_SET_SETS = 'S'
} zfs_deleg_who_type_t;
typedef enum {
ZFS_DELEG_NONE = 0,
ZFS_DELEG_PERM_LOCAL = 1,
ZFS_DELEG_PERM_DESCENDENT = 2,
ZFS_DELEG_PERM_LOCALDESCENDENT = 3,
ZFS_DELEG_PERM_CREATE = 4
} zfs_deleg_inherit_t;
#define ZFS_DELEG_PERM_UID "uid"
#define ZFS_DELEG_PERM_GID "gid"
#define ZFS_DELEG_PERM_GROUPS "groups"
#define ZFS_MLSLABEL_DEFAULT "none"
#define ZFS_SMB_ACL_SRC "src"
#define ZFS_SMB_ACL_TARGET "target"
typedef enum {
ZFS_CANMOUNT_OFF = 0,
ZFS_CANMOUNT_ON = 1,
ZFS_CANMOUNT_NOAUTO = 2
} zfs_canmount_type_t;
typedef enum {
ZFS_LOGBIAS_LATENCY = 0,
ZFS_LOGBIAS_THROUGHPUT = 1
} zfs_logbias_op_t;
typedef enum zfs_share_op {
ZFS_SHARE_NFS = 0,
ZFS_UNSHARE_NFS = 1,
ZFS_SHARE_SMB = 2,
ZFS_UNSHARE_SMB = 3
} zfs_share_op_t;
typedef enum zfs_smb_acl_op {
ZFS_SMB_ACL_ADD,
ZFS_SMB_ACL_REMOVE,
ZFS_SMB_ACL_RENAME,
ZFS_SMB_ACL_PURGE
} zfs_smb_acl_op_t;
typedef enum zfs_cache_type {
ZFS_CACHE_NONE = 0,
ZFS_CACHE_METADATA = 1,
ZFS_CACHE_ALL = 2
} zfs_cache_type_t;
typedef enum {
ZFS_SYNC_STANDARD = 0,
ZFS_SYNC_ALWAYS = 1,
ZFS_SYNC_DISABLED = 2
} zfs_sync_type_t;
typedef enum {
ZFS_XATTR_OFF = 0,
ZFS_XATTR_DIR = 1,
ZFS_XATTR_SA = 2
} zfs_xattr_type_t;
typedef enum {
ZFS_DNSIZE_LEGACY = 0,
ZFS_DNSIZE_AUTO = 1,
ZFS_DNSIZE_1K = 1024,
ZFS_DNSIZE_2K = 2048,
ZFS_DNSIZE_4K = 4096,
ZFS_DNSIZE_8K = 8192,
ZFS_DNSIZE_16K = 16384
} zfs_dnsize_type_t;
typedef enum {
ZFS_REDUNDANT_METADATA_ALL,
ZFS_REDUNDANT_METADATA_MOST,
ZFS_REDUNDANT_METADATA_SOME,
ZFS_REDUNDANT_METADATA_NONE
} zfs_redundant_metadata_type_t;
typedef enum {
ZFS_VOLMODE_DEFAULT = 0,
ZFS_VOLMODE_GEOM = 1,
ZFS_VOLMODE_DEV = 2,
ZFS_VOLMODE_NONE = 3
} zfs_volmode_t;
typedef enum zfs_keystatus {
ZFS_KEYSTATUS_NONE = 0,
ZFS_KEYSTATUS_UNAVAILABLE,
ZFS_KEYSTATUS_AVAILABLE,
} zfs_keystatus_t;
typedef enum zfs_keyformat {
ZFS_KEYFORMAT_NONE = 0,
ZFS_KEYFORMAT_RAW,
ZFS_KEYFORMAT_HEX,
ZFS_KEYFORMAT_PASSPHRASE,
ZFS_KEYFORMAT_FORMATS
} zfs_keyformat_t;
typedef enum zfs_key_location {
ZFS_KEYLOCATION_NONE = 0,
ZFS_KEYLOCATION_PROMPT,
ZFS_KEYLOCATION_URI,
ZFS_KEYLOCATION_LOCATIONS
} zfs_keylocation_t;
#define DEFAULT_PBKDF2_ITERATIONS 350000
#define MIN_PBKDF2_ITERATIONS 100000
/*
* On-disk version number.
*/
#define SPA_VERSION_1 1ULL
#define SPA_VERSION_2 2ULL
#define SPA_VERSION_3 3ULL
#define SPA_VERSION_4 4ULL
#define SPA_VERSION_5 5ULL
#define SPA_VERSION_6 6ULL
#define SPA_VERSION_7 7ULL
#define SPA_VERSION_8 8ULL
#define SPA_VERSION_9 9ULL
#define SPA_VERSION_10 10ULL
#define SPA_VERSION_11 11ULL
#define SPA_VERSION_12 12ULL
#define SPA_VERSION_13 13ULL
#define SPA_VERSION_14 14ULL
#define SPA_VERSION_15 15ULL
#define SPA_VERSION_16 16ULL
#define SPA_VERSION_17 17ULL
#define SPA_VERSION_18 18ULL
#define SPA_VERSION_19 19ULL
#define SPA_VERSION_20 20ULL
#define SPA_VERSION_21 21ULL
#define SPA_VERSION_22 22ULL
#define SPA_VERSION_23 23ULL
#define SPA_VERSION_24 24ULL
#define SPA_VERSION_25 25ULL
#define SPA_VERSION_26 26ULL
#define SPA_VERSION_27 27ULL
#define SPA_VERSION_28 28ULL
#define SPA_VERSION_5000 5000ULL
/*
* The incrementing pool version number has been replaced by pool feature
* flags. For more details, see zfeature.c.
*/
#define SPA_VERSION SPA_VERSION_5000
#define SPA_VERSION_STRING "5000"
/*
* Symbolic names for the changes that caused a SPA_VERSION switch.
* Used in the code when checking for presence or absence of a feature.
* Feel free to define multiple symbolic names for each version if there
* were multiple changes to on-disk structures during that version.
*
* NOTE: When checking the current SPA_VERSION in your code, be sure
* to use spa_version() since it reports the version of the
* last synced uberblock. Checking the in-flight version can
* be dangerous in some cases.
*/
#define SPA_VERSION_INITIAL SPA_VERSION_1
#define SPA_VERSION_DITTO_BLOCKS SPA_VERSION_2
#define SPA_VERSION_SPARES SPA_VERSION_3
#define SPA_VERSION_RAIDZ2 SPA_VERSION_3
#define SPA_VERSION_BPOBJ_ACCOUNT SPA_VERSION_3
#define SPA_VERSION_RAIDZ_DEFLATE SPA_VERSION_3
#define SPA_VERSION_DNODE_BYTES SPA_VERSION_3
#define SPA_VERSION_ZPOOL_HISTORY SPA_VERSION_4
#define SPA_VERSION_GZIP_COMPRESSION SPA_VERSION_5
#define SPA_VERSION_BOOTFS SPA_VERSION_6
#define SPA_VERSION_SLOGS SPA_VERSION_7
#define SPA_VERSION_DELEGATED_PERMS SPA_VERSION_8
#define SPA_VERSION_FUID SPA_VERSION_9
#define SPA_VERSION_REFRESERVATION SPA_VERSION_9
#define SPA_VERSION_REFQUOTA SPA_VERSION_9
#define SPA_VERSION_UNIQUE_ACCURATE SPA_VERSION_9
#define SPA_VERSION_L2CACHE SPA_VERSION_10
#define SPA_VERSION_NEXT_CLONES SPA_VERSION_11
#define SPA_VERSION_ORIGIN SPA_VERSION_11
#define SPA_VERSION_DSL_SCRUB SPA_VERSION_11
#define SPA_VERSION_SNAP_PROPS SPA_VERSION_12
#define SPA_VERSION_USED_BREAKDOWN SPA_VERSION_13
#define SPA_VERSION_PASSTHROUGH_X SPA_VERSION_14
#define SPA_VERSION_USERSPACE SPA_VERSION_15
#define SPA_VERSION_STMF_PROP SPA_VERSION_16
#define SPA_VERSION_RAIDZ3 SPA_VERSION_17
#define SPA_VERSION_USERREFS SPA_VERSION_18
#define SPA_VERSION_HOLES SPA_VERSION_19
#define SPA_VERSION_ZLE_COMPRESSION SPA_VERSION_20
#define SPA_VERSION_DEDUP SPA_VERSION_21
#define SPA_VERSION_RECVD_PROPS SPA_VERSION_22
#define SPA_VERSION_SLIM_ZIL SPA_VERSION_23
#define SPA_VERSION_SA SPA_VERSION_24
#define SPA_VERSION_SCAN SPA_VERSION_25
#define SPA_VERSION_DIR_CLONES SPA_VERSION_26
#define SPA_VERSION_DEADLISTS SPA_VERSION_26
#define SPA_VERSION_FAST_SNAP SPA_VERSION_27
#define SPA_VERSION_MULTI_REPLACE SPA_VERSION_28
#define SPA_VERSION_BEFORE_FEATURES SPA_VERSION_28
#define SPA_VERSION_FEATURES SPA_VERSION_5000
#define SPA_VERSION_IS_SUPPORTED(v) \
(((v) >= SPA_VERSION_INITIAL && (v) <= SPA_VERSION_BEFORE_FEATURES) || \
((v) >= SPA_VERSION_FEATURES && (v) <= SPA_VERSION))
/*
* ZPL version - rev'd whenever an incompatible on-disk format change
* occurs. This is independent of SPA/DMU/ZAP versioning. You must
* also update the version_table[] and help message in zfs_prop.c.
*/
#define ZPL_VERSION_1 1ULL
#define ZPL_VERSION_2 2ULL
#define ZPL_VERSION_3 3ULL
#define ZPL_VERSION_4 4ULL
#define ZPL_VERSION_5 5ULL
#define ZPL_VERSION ZPL_VERSION_5
#define ZPL_VERSION_STRING "5"
#define ZPL_VERSION_INITIAL ZPL_VERSION_1
#define ZPL_VERSION_DIRENT_TYPE ZPL_VERSION_2
#define ZPL_VERSION_FUID ZPL_VERSION_3
#define ZPL_VERSION_NORMALIZATION ZPL_VERSION_3
#define ZPL_VERSION_SYSATTR ZPL_VERSION_3
#define ZPL_VERSION_USERSPACE ZPL_VERSION_4
#define ZPL_VERSION_SA ZPL_VERSION_5
/* Persistent L2ARC version */
#define L2ARC_PERSISTENT_VERSION_1 1ULL
#define L2ARC_PERSISTENT_VERSION L2ARC_PERSISTENT_VERSION_1
#define L2ARC_PERSISTENT_VERSION_STRING "1"
/* Rewind policy information */
#define ZPOOL_NO_REWIND 1 /* No policy - default behavior */
#define ZPOOL_NEVER_REWIND 2 /* Do not search for best txg or rewind */
#define ZPOOL_TRY_REWIND 4 /* Search for best txg, but do not rewind */
#define ZPOOL_DO_REWIND 8 /* Rewind to best txg w/in deferred frees */
#define ZPOOL_EXTREME_REWIND 16 /* Allow extreme measures to find best txg */
#define ZPOOL_REWIND_MASK 28 /* All the possible rewind bits */
#define ZPOOL_REWIND_POLICIES 31 /* All the possible policy bits */
typedef struct zpool_load_policy {
uint32_t zlp_rewind; /* rewind policy requested */
uint64_t zlp_maxmeta; /* max acceptable meta-data errors */
uint64_t zlp_maxdata; /* max acceptable data errors */
uint64_t zlp_txg; /* specific txg to load */
} zpool_load_policy_t;
/*
* The following are configuration names used in the nvlist describing a pool's
* configuration. New on-disk names should be prefixed with "<reversed-DNS>:"
* (e.g. "org.openzfs:") to avoid conflicting names being developed
* independently.
*/
#define ZPOOL_CONFIG_VERSION "version"
#define ZPOOL_CONFIG_POOL_NAME "name"
#define ZPOOL_CONFIG_POOL_STATE "state"
#define ZPOOL_CONFIG_POOL_TXG "txg"
#define ZPOOL_CONFIG_POOL_GUID "pool_guid"
#define ZPOOL_CONFIG_CREATE_TXG "create_txg"
#define ZPOOL_CONFIG_TOP_GUID "top_guid"
#define ZPOOL_CONFIG_VDEV_TREE "vdev_tree"
#define ZPOOL_CONFIG_TYPE "type"
#define ZPOOL_CONFIG_CHILDREN "children"
#define ZPOOL_CONFIG_ID "id"
#define ZPOOL_CONFIG_GUID "guid"
#define ZPOOL_CONFIG_INDIRECT_OBJECT "com.delphix:indirect_object"
#define ZPOOL_CONFIG_INDIRECT_BIRTHS "com.delphix:indirect_births"
#define ZPOOL_CONFIG_PREV_INDIRECT_VDEV "com.delphix:prev_indirect_vdev"
#define ZPOOL_CONFIG_PATH "path"
#define ZPOOL_CONFIG_DEVID "devid"
#define ZPOOL_CONFIG_SPARE_ID "spareid"
#define ZPOOL_CONFIG_METASLAB_ARRAY "metaslab_array"
#define ZPOOL_CONFIG_METASLAB_SHIFT "metaslab_shift"
#define ZPOOL_CONFIG_ASHIFT "ashift"
#define ZPOOL_CONFIG_ASIZE "asize"
#define ZPOOL_CONFIG_DTL "DTL"
#define ZPOOL_CONFIG_SCAN_STATS "scan_stats" /* not stored on disk */
#define ZPOOL_CONFIG_REMOVAL_STATS "removal_stats" /* not stored on disk */
#define ZPOOL_CONFIG_CHECKPOINT_STATS "checkpoint_stats" /* not on disk */
#define ZPOOL_CONFIG_VDEV_STATS "vdev_stats" /* not stored on disk */
#define ZPOOL_CONFIG_INDIRECT_SIZE "indirect_size" /* not stored on disk */
/* container nvlist of extended stats */
#define ZPOOL_CONFIG_VDEV_STATS_EX "vdev_stats_ex"
/* Active queue read/write stats */
#define ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE "vdev_sync_r_active_queue"
#define ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE "vdev_sync_w_active_queue"
#define ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE "vdev_async_r_active_queue"
#define ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE "vdev_async_w_active_queue"
#define ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE "vdev_async_scrub_active_queue"
#define ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE "vdev_async_trim_active_queue"
#define ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE "vdev_rebuild_active_queue"
/* Queue sizes */
#define ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE "vdev_sync_r_pend_queue"
#define ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE "vdev_sync_w_pend_queue"
#define ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE "vdev_async_r_pend_queue"
#define ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE "vdev_async_w_pend_queue"
#define ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE "vdev_async_scrub_pend_queue"
#define ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE "vdev_async_trim_pend_queue"
#define ZPOOL_CONFIG_VDEV_REBUILD_PEND_QUEUE "vdev_rebuild_pend_queue"
/* Latency read/write histogram stats */
#define ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO "vdev_tot_r_lat_histo"
#define ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO "vdev_tot_w_lat_histo"
#define ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO "vdev_disk_r_lat_histo"
#define ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO "vdev_disk_w_lat_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO "vdev_sync_r_lat_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO "vdev_sync_w_lat_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO "vdev_async_r_lat_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO "vdev_async_w_lat_histo"
#define ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO "vdev_scrub_histo"
#define ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO "vdev_trim_histo"
#define ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO "vdev_rebuild_histo"
/* Request size histograms */
#define ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO "vdev_sync_ind_r_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO "vdev_sync_ind_w_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO "vdev_async_ind_r_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO "vdev_async_ind_w_histo"
#define ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO "vdev_ind_scrub_histo"
#define ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO "vdev_ind_trim_histo"
#define ZPOOL_CONFIG_VDEV_IND_REBUILD_HISTO "vdev_ind_rebuild_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO "vdev_sync_agg_r_histo"
#define ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO "vdev_sync_agg_w_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO "vdev_async_agg_r_histo"
#define ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO "vdev_async_agg_w_histo"
#define ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO "vdev_agg_scrub_histo"
#define ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO "vdev_agg_trim_histo"
#define ZPOOL_CONFIG_VDEV_AGG_REBUILD_HISTO "vdev_agg_rebuild_histo"
/* Number of slow IOs */
#define ZPOOL_CONFIG_VDEV_SLOW_IOS "vdev_slow_ios"
/* vdev enclosure sysfs path */
#define ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH "vdev_enc_sysfs_path"
#define ZPOOL_CONFIG_WHOLE_DISK "whole_disk"
#define ZPOOL_CONFIG_ERRCOUNT "error_count"
#define ZPOOL_CONFIG_NOT_PRESENT "not_present"
#define ZPOOL_CONFIG_SPARES "spares"
#define ZPOOL_CONFIG_IS_SPARE "is_spare"
#define ZPOOL_CONFIG_NPARITY "nparity"
#define ZPOOL_CONFIG_HOSTID "hostid"
#define ZPOOL_CONFIG_HOSTNAME "hostname"
#define ZPOOL_CONFIG_LOADED_TIME "initial_load_time"
#define ZPOOL_CONFIG_UNSPARE "unspare"
#define ZPOOL_CONFIG_PHYS_PATH "phys_path"
#define ZPOOL_CONFIG_IS_LOG "is_log"
#define ZPOOL_CONFIG_L2CACHE "l2cache"
#define ZPOOL_CONFIG_HOLE_ARRAY "hole_array"
#define ZPOOL_CONFIG_VDEV_CHILDREN "vdev_children"
#define ZPOOL_CONFIG_IS_HOLE "is_hole"
#define ZPOOL_CONFIG_DDT_HISTOGRAM "ddt_histogram"
#define ZPOOL_CONFIG_DDT_OBJ_STATS "ddt_object_stats"
#define ZPOOL_CONFIG_DDT_STATS "ddt_stats"
#define ZPOOL_CONFIG_SPLIT "splitcfg"
#define ZPOOL_CONFIG_ORIG_GUID "orig_guid"
#define ZPOOL_CONFIG_SPLIT_GUID "split_guid"
#define ZPOOL_CONFIG_SPLIT_LIST "guid_list"
#define ZPOOL_CONFIG_NONALLOCATING "non_allocating"
#define ZPOOL_CONFIG_REMOVING "removing"
#define ZPOOL_CONFIG_RESILVER_TXG "resilver_txg"
#define ZPOOL_CONFIG_REBUILD_TXG "rebuild_txg"
#define ZPOOL_CONFIG_COMMENT "comment"
#define ZPOOL_CONFIG_SUSPENDED "suspended" /* not stored on disk */
#define ZPOOL_CONFIG_SUSPENDED_REASON "suspended_reason" /* not stored */
#define ZPOOL_CONFIG_TIMESTAMP "timestamp" /* not stored on disk */
#define ZPOOL_CONFIG_BOOTFS "bootfs" /* not stored on disk */
#define ZPOOL_CONFIG_MISSING_DEVICES "missing_vdevs" /* not stored on disk */
#define ZPOOL_CONFIG_LOAD_INFO "load_info" /* not stored on disk */
#define ZPOOL_CONFIG_REWIND_INFO "rewind_info" /* not stored on disk */
#define ZPOOL_CONFIG_UNSUP_FEAT "unsup_feat" /* not stored on disk */
#define ZPOOL_CONFIG_ENABLED_FEAT "enabled_feat" /* not stored on disk */
#define ZPOOL_CONFIG_CAN_RDONLY "can_rdonly" /* not stored on disk */
#define ZPOOL_CONFIG_FEATURES_FOR_READ "features_for_read"
#define ZPOOL_CONFIG_FEATURE_STATS "feature_stats" /* not stored on disk */
#define ZPOOL_CONFIG_ERRATA "errata" /* not stored on disk */
+#define ZPOOL_CONFIG_VDEV_ROOT_ZAP "com.klarasystems:vdev_zap_root"
#define ZPOOL_CONFIG_VDEV_TOP_ZAP "com.delphix:vdev_zap_top"
#define ZPOOL_CONFIG_VDEV_LEAF_ZAP "com.delphix:vdev_zap_leaf"
#define ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS "com.delphix:has_per_vdev_zaps"
#define ZPOOL_CONFIG_RESILVER_DEFER "com.datto:resilver_defer"
#define ZPOOL_CONFIG_CACHEFILE "cachefile" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_STATE "mmp_state" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_TXG "mmp_txg" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_SEQ "mmp_seq" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_HOSTNAME "mmp_hostname" /* not stored on disk */
#define ZPOOL_CONFIG_MMP_HOSTID "mmp_hostid" /* not stored on disk */
#define ZPOOL_CONFIG_ALLOCATION_BIAS "alloc_bias" /* not stored on disk */
#define ZPOOL_CONFIG_EXPANSION_TIME "expansion_time" /* not stored */
#define ZPOOL_CONFIG_REBUILD_STATS "org.openzfs:rebuild_stats"
#define ZPOOL_CONFIG_COMPATIBILITY "compatibility"
/*
* The persistent vdev state is stored as separate values rather than a single
* 'vdev_state' entry. This is because a device can be in multiple states, such
* as offline and degraded.
*/
#define ZPOOL_CONFIG_OFFLINE "offline"
#define ZPOOL_CONFIG_FAULTED "faulted"
#define ZPOOL_CONFIG_DEGRADED "degraded"
#define ZPOOL_CONFIG_REMOVED "removed"
#define ZPOOL_CONFIG_FRU "fru"
#define ZPOOL_CONFIG_AUX_STATE "aux_state"
/* Pool load policy parameters */
#define ZPOOL_LOAD_POLICY "load-policy"
#define ZPOOL_LOAD_REWIND_POLICY "load-rewind-policy"
#define ZPOOL_LOAD_REQUEST_TXG "load-request-txg"
#define ZPOOL_LOAD_META_THRESH "load-meta-thresh"
#define ZPOOL_LOAD_DATA_THRESH "load-data-thresh"
/* Rewind data discovered */
#define ZPOOL_CONFIG_LOAD_TIME "rewind_txg_ts"
#define ZPOOL_CONFIG_LOAD_META_ERRORS "verify_meta_errors"
#define ZPOOL_CONFIG_LOAD_DATA_ERRORS "verify_data_errors"
#define ZPOOL_CONFIG_REWIND_TIME "seconds_of_rewind"
/* dRAID configuration */
#define ZPOOL_CONFIG_DRAID_NDATA "draid_ndata"
#define ZPOOL_CONFIG_DRAID_NSPARES "draid_nspares"
#define ZPOOL_CONFIG_DRAID_NGROUPS "draid_ngroups"
#define VDEV_TYPE_ROOT "root"
#define VDEV_TYPE_MIRROR "mirror"
#define VDEV_TYPE_REPLACING "replacing"
#define VDEV_TYPE_RAIDZ "raidz"
#define VDEV_TYPE_DRAID "draid"
#define VDEV_TYPE_DRAID_SPARE "dspare"
#define VDEV_TYPE_DISK "disk"
#define VDEV_TYPE_FILE "file"
#define VDEV_TYPE_MISSING "missing"
#define VDEV_TYPE_HOLE "hole"
#define VDEV_TYPE_SPARE "spare"
#define VDEV_TYPE_LOG "log"
#define VDEV_TYPE_L2CACHE "l2cache"
#define VDEV_TYPE_INDIRECT "indirect"
#define VDEV_RAIDZ_MAXPARITY 3
#define VDEV_DRAID_MAXPARITY 3
#define VDEV_DRAID_MIN_CHILDREN 2
#define VDEV_DRAID_MAX_CHILDREN UINT8_MAX
/* VDEV_TOP_ZAP_* are used in top-level vdev ZAP objects. */
#define VDEV_TOP_ZAP_INDIRECT_OBSOLETE_SM \
"com.delphix:indirect_obsolete_sm"
#define VDEV_TOP_ZAP_OBSOLETE_COUNTS_ARE_PRECISE \
"com.delphix:obsolete_counts_are_precise"
#define VDEV_TOP_ZAP_POOL_CHECKPOINT_SM \
"com.delphix:pool_checkpoint_sm"
#define VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS \
"com.delphix:ms_unflushed_phys_txgs"
#define VDEV_TOP_ZAP_VDEV_REBUILD_PHYS \
"org.openzfs:vdev_rebuild"
#define VDEV_TOP_ZAP_ALLOCATION_BIAS \
"org.zfsonlinux:allocation_bias"
/* vdev metaslab allocation bias */
#define VDEV_ALLOC_BIAS_LOG "log"
#define VDEV_ALLOC_BIAS_SPECIAL "special"
#define VDEV_ALLOC_BIAS_DEDUP "dedup"
/* vdev initialize state */
#define VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET \
"com.delphix:next_offset_to_initialize"
#define VDEV_LEAF_ZAP_INITIALIZE_STATE \
"com.delphix:vdev_initialize_state"
#define VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME \
"com.delphix:vdev_initialize_action_time"
/* vdev TRIM state */
#define VDEV_LEAF_ZAP_TRIM_LAST_OFFSET \
"org.zfsonlinux:next_offset_to_trim"
#define VDEV_LEAF_ZAP_TRIM_STATE \
"org.zfsonlinux:vdev_trim_state"
#define VDEV_LEAF_ZAP_TRIM_ACTION_TIME \
"org.zfsonlinux:vdev_trim_action_time"
#define VDEV_LEAF_ZAP_TRIM_RATE \
"org.zfsonlinux:vdev_trim_rate"
#define VDEV_LEAF_ZAP_TRIM_PARTIAL \
"org.zfsonlinux:vdev_trim_partial"
#define VDEV_LEAF_ZAP_TRIM_SECURE \
"org.zfsonlinux:vdev_trim_secure"
/*
* This is needed in userland to report the minimum necessary device size.
*/
#define SPA_MINDEVSIZE (64ULL << 20)
/*
* Set if the fragmentation has not yet been calculated. This can happen
* because the space maps have not been upgraded or the histogram feature
* is not enabled.
*/
#define ZFS_FRAG_INVALID UINT64_MAX
/*
* The location of the pool configuration repository, shared between kernel and
* userland.
*/
#define ZPOOL_CACHE_BOOT "/boot/zfs/zpool.cache"
#define ZPOOL_CACHE "/etc/zfs/zpool.cache"
/*
* Settings for zpool compatibility features files
*/
#define ZPOOL_SYSCONF_COMPAT_D SYSCONFDIR "/zfs/compatibility.d"
#define ZPOOL_DATA_COMPAT_D PKGDATADIR "/compatibility.d"
#define ZPOOL_COMPAT_MAXSIZE 16384
/*
* Hard-wired compatibility settings
*/
#define ZPOOL_COMPAT_LEGACY "legacy"
#define ZPOOL_COMPAT_OFF "off"
/*
* vdev states are ordered from least to most healthy.
* A vdev that's CANT_OPEN or below is considered unusable.
*/
typedef enum vdev_state {
VDEV_STATE_UNKNOWN = 0, /* Uninitialized vdev */
VDEV_STATE_CLOSED, /* Not currently open */
VDEV_STATE_OFFLINE, /* Not allowed to open */
VDEV_STATE_REMOVED, /* Explicitly removed from system */
VDEV_STATE_CANT_OPEN, /* Tried to open, but failed */
VDEV_STATE_FAULTED, /* External request to fault device */
VDEV_STATE_DEGRADED, /* Replicated vdev with unhealthy kids */
VDEV_STATE_HEALTHY /* Presumed good */
} vdev_state_t;
#define VDEV_STATE_ONLINE VDEV_STATE_HEALTHY
/*
* vdev aux states. When a vdev is in the CANT_OPEN state, the aux field
* of the vdev stats structure uses these constants to distinguish why.
*/
typedef enum vdev_aux {
VDEV_AUX_NONE, /* no error */
VDEV_AUX_OPEN_FAILED, /* ldi_open_*() or vn_open() failed */
VDEV_AUX_CORRUPT_DATA, /* bad label or disk contents */
VDEV_AUX_NO_REPLICAS, /* insufficient number of replicas */
VDEV_AUX_BAD_GUID_SUM, /* vdev guid sum doesn't match */
VDEV_AUX_TOO_SMALL, /* vdev size is too small */
VDEV_AUX_BAD_LABEL, /* the label is OK but invalid */
VDEV_AUX_VERSION_NEWER, /* on-disk version is too new */
VDEV_AUX_VERSION_OLDER, /* on-disk version is too old */
VDEV_AUX_UNSUP_FEAT, /* unsupported features */
VDEV_AUX_SPARED, /* hot spare used in another pool */
VDEV_AUX_ERR_EXCEEDED, /* too many errors */
VDEV_AUX_IO_FAILURE, /* experienced I/O failure */
VDEV_AUX_BAD_LOG, /* cannot read log chain(s) */
VDEV_AUX_EXTERNAL, /* external diagnosis or forced fault */
VDEV_AUX_SPLIT_POOL, /* vdev was split off into another pool */
VDEV_AUX_BAD_ASHIFT, /* vdev ashift is invalid */
VDEV_AUX_EXTERNAL_PERSIST, /* persistent forced fault */
VDEV_AUX_ACTIVE, /* vdev active on a different host */
VDEV_AUX_CHILDREN_OFFLINE, /* all children are offline */
VDEV_AUX_ASHIFT_TOO_BIG, /* vdev's min block size is too large */
} vdev_aux_t;
/*
* pool state. The following states are written to disk as part of the normal
* SPA lifecycle: ACTIVE, EXPORTED, DESTROYED, SPARE, L2CACHE. The remaining
* states are software abstractions used at various levels to communicate
* pool state.
*/
typedef enum pool_state {
POOL_STATE_ACTIVE = 0, /* In active use */
POOL_STATE_EXPORTED, /* Explicitly exported */
POOL_STATE_DESTROYED, /* Explicitly destroyed */
POOL_STATE_SPARE, /* Reserved for hot spare use */
POOL_STATE_L2CACHE, /* Level 2 ARC device */
POOL_STATE_UNINITIALIZED, /* Internal spa_t state */
POOL_STATE_UNAVAIL, /* Internal libzfs state */
POOL_STATE_POTENTIALLY_ACTIVE /* Internal libzfs state */
} pool_state_t;
/*
* mmp state. The following states provide additional detail describing
* why a pool couldn't be safely imported.
*/
typedef enum mmp_state {
MMP_STATE_ACTIVE = 0, /* In active use */
MMP_STATE_INACTIVE, /* Inactive and safe to import */
MMP_STATE_NO_HOSTID /* System hostid is not set */
} mmp_state_t;
/*
* Scan Functions.
*/
typedef enum pool_scan_func {
POOL_SCAN_NONE,
POOL_SCAN_SCRUB,
POOL_SCAN_RESILVER,
POOL_SCAN_FUNCS
} pool_scan_func_t;
/*
* Used to control scrub pause and resume.
*/
typedef enum pool_scrub_cmd {
POOL_SCRUB_NORMAL = 0,
POOL_SCRUB_PAUSE,
POOL_SCRUB_FLAGS_END
} pool_scrub_cmd_t;
typedef enum {
CS_NONE,
CS_CHECKPOINT_EXISTS,
CS_CHECKPOINT_DISCARDING,
CS_NUM_STATES
} checkpoint_state_t;
typedef struct pool_checkpoint_stat {
uint64_t pcs_state; /* checkpoint_state_t */
uint64_t pcs_start_time; /* time checkpoint/discard started */
uint64_t pcs_space; /* checkpointed space */
} pool_checkpoint_stat_t;
/*
* ZIO types. Needed to interpret vdev statistics below.
*/
typedef enum zio_type {
ZIO_TYPE_NULL = 0,
ZIO_TYPE_READ,
ZIO_TYPE_WRITE,
ZIO_TYPE_FREE,
ZIO_TYPE_CLAIM,
ZIO_TYPE_IOCTL,
ZIO_TYPE_TRIM,
ZIO_TYPES
} zio_type_t;
/*
* Pool statistics. Note: all fields should be 64-bit because this
* is passed between kernel and userland as an nvlist uint64 array.
*/
typedef struct pool_scan_stat {
/* values stored on disk */
uint64_t pss_func; /* pool_scan_func_t */
uint64_t pss_state; /* dsl_scan_state_t */
uint64_t pss_start_time; /* scan start time */
uint64_t pss_end_time; /* scan end time */
uint64_t pss_to_examine; /* total bytes to scan */
uint64_t pss_examined; /* total bytes located by scanner */
uint64_t pss_to_process; /* total bytes to process */
uint64_t pss_processed; /* total processed bytes */
uint64_t pss_errors; /* scan errors */
/* values not stored on disk */
uint64_t pss_pass_exam; /* examined bytes per scan pass */
uint64_t pss_pass_start; /* start time of a scan pass */
uint64_t pss_pass_scrub_pause; /* pause time of a scrub pass */
/* cumulative time scrub spent paused, needed for rate calculation */
uint64_t pss_pass_scrub_spent_paused;
uint64_t pss_pass_issued; /* issued bytes per scan pass */
uint64_t pss_issued; /* total bytes checked by scanner */
} pool_scan_stat_t;
typedef struct pool_removal_stat {
uint64_t prs_state; /* dsl_scan_state_t */
uint64_t prs_removing_vdev;
uint64_t prs_start_time;
uint64_t prs_end_time;
uint64_t prs_to_copy; /* bytes that need to be copied */
uint64_t prs_copied; /* bytes copied so far */
/*
* bytes of memory used for indirect mappings.
* This includes all removed vdevs.
*/
uint64_t prs_mapping_memory;
} pool_removal_stat_t;
typedef enum dsl_scan_state {
DSS_NONE,
DSS_SCANNING,
DSS_FINISHED,
DSS_CANCELED,
DSS_NUM_STATES
} dsl_scan_state_t;
typedef struct vdev_rebuild_stat {
uint64_t vrs_state; /* vdev_rebuild_state_t */
uint64_t vrs_start_time; /* time_t */
uint64_t vrs_end_time; /* time_t */
uint64_t vrs_scan_time_ms; /* total run time (millisecs) */
uint64_t vrs_bytes_scanned; /* allocated bytes scanned */
uint64_t vrs_bytes_issued; /* read bytes issued */
uint64_t vrs_bytes_rebuilt; /* rebuilt bytes */
uint64_t vrs_bytes_est; /* total bytes to scan */
uint64_t vrs_errors; /* scanning errors */
uint64_t vrs_pass_time_ms; /* pass run time (millisecs) */
uint64_t vrs_pass_bytes_scanned; /* bytes scanned since start/resume */
uint64_t vrs_pass_bytes_issued; /* bytes rebuilt since start/resume */
} vdev_rebuild_stat_t;
/*
* Errata described by https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-ER.
* The ordering of this enum must be maintained to ensure the errata identifiers
* map to the correct documentation. New errata may only be appended to the
* list and must contain corresponding documentation at the above link.
*/
typedef enum zpool_errata {
ZPOOL_ERRATA_NONE,
ZPOOL_ERRATA_ZOL_2094_SCRUB,
ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY,
ZPOOL_ERRATA_ZOL_6845_ENCRYPTION,
ZPOOL_ERRATA_ZOL_8308_ENCRYPTION,
} zpool_errata_t;
/*
* Vdev statistics. Note: all fields should be 64-bit because this
* is passed between kernel and user land as an nvlist uint64 array.
*
* The vs_ops[] and vs_bytes[] arrays must always be an array size of 6 in
* order to keep subsequent members at their known fixed offsets. When
* adding a new field it must be added to the end the structure.
*/
#define VS_ZIO_TYPES 6
typedef struct vdev_stat {
hrtime_t vs_timestamp; /* time since vdev load */
uint64_t vs_state; /* vdev state */
uint64_t vs_aux; /* see vdev_aux_t */
uint64_t vs_alloc; /* space allocated */
uint64_t vs_space; /* total capacity */
uint64_t vs_dspace; /* deflated capacity */
uint64_t vs_rsize; /* replaceable dev size */
uint64_t vs_esize; /* expandable dev size */
uint64_t vs_ops[VS_ZIO_TYPES]; /* operation count */
uint64_t vs_bytes[VS_ZIO_TYPES]; /* bytes read/written */
uint64_t vs_read_errors; /* read errors */
uint64_t vs_write_errors; /* write errors */
uint64_t vs_checksum_errors; /* checksum errors */
uint64_t vs_initialize_errors; /* initializing errors */
uint64_t vs_self_healed; /* self-healed bytes */
uint64_t vs_scan_removing; /* removing? */
uint64_t vs_scan_processed; /* scan processed bytes */
uint64_t vs_fragmentation; /* device fragmentation */
uint64_t vs_initialize_bytes_done; /* bytes initialized */
uint64_t vs_initialize_bytes_est; /* total bytes to initialize */
uint64_t vs_initialize_state; /* vdev_initializing_state_t */
uint64_t vs_initialize_action_time; /* time_t */
uint64_t vs_checkpoint_space; /* checkpoint-consumed space */
uint64_t vs_resilver_deferred; /* resilver deferred */
uint64_t vs_slow_ios; /* slow IOs */
uint64_t vs_trim_errors; /* trimming errors */
uint64_t vs_trim_notsup; /* supported by device */
uint64_t vs_trim_bytes_done; /* bytes trimmed */
uint64_t vs_trim_bytes_est; /* total bytes to trim */
uint64_t vs_trim_state; /* vdev_trim_state_t */
uint64_t vs_trim_action_time; /* time_t */
uint64_t vs_rebuild_processed; /* bytes rebuilt */
uint64_t vs_configured_ashift; /* TLV vdev_ashift */
uint64_t vs_logical_ashift; /* vdev_logical_ashift */
uint64_t vs_physical_ashift; /* vdev_physical_ashift */
uint64_t vs_noalloc; /* allocations halted? */
uint64_t vs_pspace; /* physical capacity */
} vdev_stat_t;
#define VDEV_STAT_VALID(field, uint64_t_field_count) \
((uint64_t_field_count * sizeof (uint64_t)) >= \
(offsetof(vdev_stat_t, field) + sizeof (((vdev_stat_t *)NULL)->field)))
/*
* Extended stats
*
* These are stats which aren't included in the original iostat output. For
* convenience, they are grouped together in vdev_stat_ex, although each stat
* is individually exported as an nvlist.
*/
typedef struct vdev_stat_ex {
/* Number of ZIOs issued to disk and waiting to finish */
uint64_t vsx_active_queue[ZIO_PRIORITY_NUM_QUEUEABLE];
/* Number of ZIOs pending to be issued to disk */
uint64_t vsx_pend_queue[ZIO_PRIORITY_NUM_QUEUEABLE];
/*
* Below are the histograms for various latencies. Buckets are in
* units of nanoseconds.
*/
/*
* 2^37 nanoseconds = 134s. Timeouts will probably start kicking in
* before this.
*/
#define VDEV_L_HISTO_BUCKETS 37 /* Latency histo buckets */
#define VDEV_RQ_HISTO_BUCKETS 25 /* Request size histo buckets */
/* Amount of time in ZIO queue (ns) */
uint64_t vsx_queue_histo[ZIO_PRIORITY_NUM_QUEUEABLE]
[VDEV_L_HISTO_BUCKETS];
/* Total ZIO latency (ns). Includes queuing and disk access time */
uint64_t vsx_total_histo[ZIO_TYPES][VDEV_L_HISTO_BUCKETS];
/* Amount of time to read/write the disk (ns) */
uint64_t vsx_disk_histo[ZIO_TYPES][VDEV_L_HISTO_BUCKETS];
/* "lookup the bucket for a value" histogram macros */
#define HISTO(val, buckets) (val != 0 ? MIN(highbit64(val) - 1, \
buckets - 1) : 0)
#define L_HISTO(a) HISTO(a, VDEV_L_HISTO_BUCKETS)
#define RQ_HISTO(a) HISTO(a, VDEV_RQ_HISTO_BUCKETS)
/* Physical IO histogram */
uint64_t vsx_ind_histo[ZIO_PRIORITY_NUM_QUEUEABLE]
[VDEV_RQ_HISTO_BUCKETS];
/* Delegated (aggregated) physical IO histogram */
uint64_t vsx_agg_histo[ZIO_PRIORITY_NUM_QUEUEABLE]
[VDEV_RQ_HISTO_BUCKETS];
} vdev_stat_ex_t;
/*
* Initialize functions.
*/
typedef enum pool_initialize_func {
POOL_INITIALIZE_START,
POOL_INITIALIZE_CANCEL,
POOL_INITIALIZE_SUSPEND,
POOL_INITIALIZE_FUNCS
} pool_initialize_func_t;
/*
* TRIM functions.
*/
typedef enum pool_trim_func {
POOL_TRIM_START,
POOL_TRIM_CANCEL,
POOL_TRIM_SUSPEND,
POOL_TRIM_FUNCS
} pool_trim_func_t;
/*
* DDT statistics. Note: all fields should be 64-bit because this
* is passed between kernel and userland as an nvlist uint64 array.
*/
typedef struct ddt_object {
uint64_t ddo_count; /* number of elements in ddt */
uint64_t ddo_dspace; /* size of ddt on disk */
uint64_t ddo_mspace; /* size of ddt in-core */
} ddt_object_t;
typedef struct ddt_stat {
uint64_t dds_blocks; /* blocks */
uint64_t dds_lsize; /* logical size */
uint64_t dds_psize; /* physical size */
uint64_t dds_dsize; /* deflated allocated size */
uint64_t dds_ref_blocks; /* referenced blocks */
uint64_t dds_ref_lsize; /* referenced lsize * refcnt */
uint64_t dds_ref_psize; /* referenced psize * refcnt */
uint64_t dds_ref_dsize; /* referenced dsize * refcnt */
} ddt_stat_t;
typedef struct ddt_histogram {
ddt_stat_t ddh_stat[64]; /* power-of-two histogram buckets */
} ddt_histogram_t;
#define ZVOL_DRIVER "zvol"
#define ZFS_DRIVER "zfs"
#define ZFS_DEV "/dev/zfs"
#define ZFS_DEVDIR "/dev"
#define ZFS_SUPER_MAGIC 0x2fc12fc1
/* general zvol path */
#define ZVOL_DIR "/dev/zvol/"
#define ZVOL_MAJOR 230
#define ZVOL_MINOR_BITS 4
#define ZVOL_MINOR_MASK ((1U << ZVOL_MINOR_BITS) - 1)
#define ZVOL_MINORS (1 << 4)
#define ZVOL_DEV_NAME "zd"
#define ZVOL_PROP_NAME "name"
#define ZVOL_DEFAULT_BLOCKSIZE 16384
typedef enum {
VDEV_INITIALIZE_NONE,
VDEV_INITIALIZE_ACTIVE,
VDEV_INITIALIZE_CANCELED,
VDEV_INITIALIZE_SUSPENDED,
VDEV_INITIALIZE_COMPLETE
} vdev_initializing_state_t;
typedef enum {
VDEV_TRIM_NONE,
VDEV_TRIM_ACTIVE,
VDEV_TRIM_CANCELED,
VDEV_TRIM_SUSPENDED,
VDEV_TRIM_COMPLETE,
} vdev_trim_state_t;
typedef enum {
VDEV_REBUILD_NONE,
VDEV_REBUILD_ACTIVE,
VDEV_REBUILD_CANCELED,
VDEV_REBUILD_COMPLETE,
} vdev_rebuild_state_t;
/*
* nvlist name constants. Facilitate restricting snapshot iteration range for
* the "list next snapshot" ioctl
*/
#define SNAP_ITER_MIN_TXG "snap_iter_min_txg"
#define SNAP_ITER_MAX_TXG "snap_iter_max_txg"
/*
* /dev/zfs ioctl numbers.
*
* These numbers cannot change over time. New ioctl numbers must be appended.
*/
typedef enum zfs_ioc {
/*
* Core features - 81/128 numbers reserved.
*/
#ifdef __FreeBSD__
ZFS_IOC_FIRST = 0,
#else
ZFS_IOC_FIRST = ('Z' << 8),
#endif
ZFS_IOC = ZFS_IOC_FIRST,
ZFS_IOC_POOL_CREATE = ZFS_IOC_FIRST, /* 0x5a00 */
ZFS_IOC_POOL_DESTROY, /* 0x5a01 */
ZFS_IOC_POOL_IMPORT, /* 0x5a02 */
ZFS_IOC_POOL_EXPORT, /* 0x5a03 */
ZFS_IOC_POOL_CONFIGS, /* 0x5a04 */
ZFS_IOC_POOL_STATS, /* 0x5a05 */
ZFS_IOC_POOL_TRYIMPORT, /* 0x5a06 */
ZFS_IOC_POOL_SCAN, /* 0x5a07 */
ZFS_IOC_POOL_FREEZE, /* 0x5a08 */
ZFS_IOC_POOL_UPGRADE, /* 0x5a09 */
ZFS_IOC_POOL_GET_HISTORY, /* 0x5a0a */
ZFS_IOC_VDEV_ADD, /* 0x5a0b */
ZFS_IOC_VDEV_REMOVE, /* 0x5a0c */
ZFS_IOC_VDEV_SET_STATE, /* 0x5a0d */
ZFS_IOC_VDEV_ATTACH, /* 0x5a0e */
ZFS_IOC_VDEV_DETACH, /* 0x5a0f */
ZFS_IOC_VDEV_SETPATH, /* 0x5a10 */
ZFS_IOC_VDEV_SETFRU, /* 0x5a11 */
ZFS_IOC_OBJSET_STATS, /* 0x5a12 */
ZFS_IOC_OBJSET_ZPLPROPS, /* 0x5a13 */
ZFS_IOC_DATASET_LIST_NEXT, /* 0x5a14 */
ZFS_IOC_SNAPSHOT_LIST_NEXT, /* 0x5a15 */
ZFS_IOC_SET_PROP, /* 0x5a16 */
ZFS_IOC_CREATE, /* 0x5a17 */
ZFS_IOC_DESTROY, /* 0x5a18 */
ZFS_IOC_ROLLBACK, /* 0x5a19 */
ZFS_IOC_RENAME, /* 0x5a1a */
ZFS_IOC_RECV, /* 0x5a1b */
ZFS_IOC_SEND, /* 0x5a1c */
ZFS_IOC_INJECT_FAULT, /* 0x5a1d */
ZFS_IOC_CLEAR_FAULT, /* 0x5a1e */
ZFS_IOC_INJECT_LIST_NEXT, /* 0x5a1f */
ZFS_IOC_ERROR_LOG, /* 0x5a20 */
ZFS_IOC_CLEAR, /* 0x5a21 */
ZFS_IOC_PROMOTE, /* 0x5a22 */
ZFS_IOC_SNAPSHOT, /* 0x5a23 */
ZFS_IOC_DSOBJ_TO_DSNAME, /* 0x5a24 */
ZFS_IOC_OBJ_TO_PATH, /* 0x5a25 */
ZFS_IOC_POOL_SET_PROPS, /* 0x5a26 */
ZFS_IOC_POOL_GET_PROPS, /* 0x5a27 */
ZFS_IOC_SET_FSACL, /* 0x5a28 */
ZFS_IOC_GET_FSACL, /* 0x5a29 */
ZFS_IOC_SHARE, /* 0x5a2a */
ZFS_IOC_INHERIT_PROP, /* 0x5a2b */
ZFS_IOC_SMB_ACL, /* 0x5a2c */
ZFS_IOC_USERSPACE_ONE, /* 0x5a2d */
ZFS_IOC_USERSPACE_MANY, /* 0x5a2e */
ZFS_IOC_USERSPACE_UPGRADE, /* 0x5a2f */
ZFS_IOC_HOLD, /* 0x5a30 */
ZFS_IOC_RELEASE, /* 0x5a31 */
ZFS_IOC_GET_HOLDS, /* 0x5a32 */
ZFS_IOC_OBJSET_RECVD_PROPS, /* 0x5a33 */
ZFS_IOC_VDEV_SPLIT, /* 0x5a34 */
ZFS_IOC_NEXT_OBJ, /* 0x5a35 */
ZFS_IOC_DIFF, /* 0x5a36 */
ZFS_IOC_TMP_SNAPSHOT, /* 0x5a37 */
ZFS_IOC_OBJ_TO_STATS, /* 0x5a38 */
ZFS_IOC_SPACE_WRITTEN, /* 0x5a39 */
ZFS_IOC_SPACE_SNAPS, /* 0x5a3a */
ZFS_IOC_DESTROY_SNAPS, /* 0x5a3b */
ZFS_IOC_POOL_REGUID, /* 0x5a3c */
ZFS_IOC_POOL_REOPEN, /* 0x5a3d */
ZFS_IOC_SEND_PROGRESS, /* 0x5a3e */
ZFS_IOC_LOG_HISTORY, /* 0x5a3f */
ZFS_IOC_SEND_NEW, /* 0x5a40 */
ZFS_IOC_SEND_SPACE, /* 0x5a41 */
ZFS_IOC_CLONE, /* 0x5a42 */
ZFS_IOC_BOOKMARK, /* 0x5a43 */
ZFS_IOC_GET_BOOKMARKS, /* 0x5a44 */
ZFS_IOC_DESTROY_BOOKMARKS, /* 0x5a45 */
ZFS_IOC_RECV_NEW, /* 0x5a46 */
ZFS_IOC_POOL_SYNC, /* 0x5a47 */
ZFS_IOC_CHANNEL_PROGRAM, /* 0x5a48 */
ZFS_IOC_LOAD_KEY, /* 0x5a49 */
ZFS_IOC_UNLOAD_KEY, /* 0x5a4a */
ZFS_IOC_CHANGE_KEY, /* 0x5a4b */
ZFS_IOC_REMAP, /* 0x5a4c */
ZFS_IOC_POOL_CHECKPOINT, /* 0x5a4d */
ZFS_IOC_POOL_DISCARD_CHECKPOINT, /* 0x5a4e */
ZFS_IOC_POOL_INITIALIZE, /* 0x5a4f */
ZFS_IOC_POOL_TRIM, /* 0x5a50 */
ZFS_IOC_REDACT, /* 0x5a51 */
ZFS_IOC_GET_BOOKMARK_PROPS, /* 0x5a52 */
ZFS_IOC_WAIT, /* 0x5a53 */
ZFS_IOC_WAIT_FS, /* 0x5a54 */
ZFS_IOC_VDEV_GET_PROPS, /* 0x5a55 */
ZFS_IOC_VDEV_SET_PROPS, /* 0x5a56 */
/*
* Per-platform (Optional) - 8/128 numbers reserved.
*/
ZFS_IOC_PLATFORM = ZFS_IOC_FIRST + 0x80,
ZFS_IOC_EVENTS_NEXT, /* 0x81 (Linux) */
ZFS_IOC_EVENTS_CLEAR, /* 0x82 (Linux) */
ZFS_IOC_EVENTS_SEEK, /* 0x83 (Linux) */
ZFS_IOC_NEXTBOOT, /* 0x84 (FreeBSD) */
ZFS_IOC_JAIL, /* 0x85 (FreeBSD) */
ZFS_IOC_USERNS_ATTACH = ZFS_IOC_JAIL, /* 0x85 (Linux) */
ZFS_IOC_UNJAIL, /* 0x86 (FreeBSD) */
ZFS_IOC_USERNS_DETACH = ZFS_IOC_UNJAIL, /* 0x86 (Linux) */
ZFS_IOC_SET_BOOTENV, /* 0x87 */
ZFS_IOC_GET_BOOTENV, /* 0x88 */
ZFS_IOC_LAST
} zfs_ioc_t;
/*
* zvol ioctl to get dataset name
*/
#define BLKZNAME _IOR(0x12, 125, char[ZFS_MAX_DATASET_NAME_LEN])
#ifdef __linux__
/*
* IOCTLs to update and retrieve additional file level attributes on
* Linux.
*/
#define ZFS_IOC_GETDOSFLAGS _IOR(0x83, 1, uint64_t)
#define ZFS_IOC_SETDOSFLAGS _IOW(0x83, 2, uint64_t)
/*
* Additional file level attributes, that are stored
* in the upper half of z_pflags
*/
#define ZFS_READONLY 0x0000000100000000ull
#define ZFS_HIDDEN 0x0000000200000000ull
#define ZFS_SYSTEM 0x0000000400000000ull
#define ZFS_ARCHIVE 0x0000000800000000ull
#define ZFS_IMMUTABLE 0x0000001000000000ull
#define ZFS_NOUNLINK 0x0000002000000000ull
#define ZFS_APPENDONLY 0x0000004000000000ull
#define ZFS_NODUMP 0x0000008000000000ull
#define ZFS_OPAQUE 0x0000010000000000ull
#define ZFS_AV_QUARANTINED 0x0000020000000000ull
#define ZFS_AV_MODIFIED 0x0000040000000000ull
#define ZFS_REPARSE 0x0000080000000000ull
#define ZFS_OFFLINE 0x0000100000000000ull
#define ZFS_SPARSE 0x0000200000000000ull
#define ZFS_DOS_FL_USER_VISIBLE (ZFS_IMMUTABLE | ZFS_APPENDONLY | \
ZFS_NOUNLINK | ZFS_ARCHIVE | ZFS_NODUMP | ZFS_SYSTEM | \
ZFS_HIDDEN | ZFS_READONLY | ZFS_REPARSE | ZFS_OFFLINE | \
ZFS_SPARSE)
#endif
/*
* ZFS-specific error codes used for returning descriptive errors
* to the userland through zfs ioctls.
*
* The enum implicitly includes all the error codes from errno.h.
* New code should use and extend this enum for errors that are
* not described precisely by generic errno codes.
*
* These numbers should not change over time. New entries should be appended.
*
* (Keep in sync with contrib/pyzfs/libzfs_core/_constants.py)
*/
typedef enum {
ZFS_ERR_CHECKPOINT_EXISTS = 1024,
ZFS_ERR_DISCARDING_CHECKPOINT,
ZFS_ERR_NO_CHECKPOINT,
ZFS_ERR_DEVRM_IN_PROGRESS,
ZFS_ERR_VDEV_TOO_BIG,
ZFS_ERR_IOC_CMD_UNAVAIL,
ZFS_ERR_IOC_ARG_UNAVAIL,
ZFS_ERR_IOC_ARG_REQUIRED,
ZFS_ERR_IOC_ARG_BADTYPE,
ZFS_ERR_WRONG_PARENT,
ZFS_ERR_FROM_IVSET_GUID_MISSING,
ZFS_ERR_FROM_IVSET_GUID_MISMATCH,
ZFS_ERR_SPILL_BLOCK_FLAG_MISSING,
ZFS_ERR_UNKNOWN_SEND_STREAM_FEATURE,
ZFS_ERR_EXPORT_IN_PROGRESS,
ZFS_ERR_BOOKMARK_SOURCE_NOT_ANCESTOR,
ZFS_ERR_STREAM_TRUNCATED,
ZFS_ERR_STREAM_LARGE_BLOCK_MISMATCH,
ZFS_ERR_RESILVER_IN_PROGRESS,
ZFS_ERR_REBUILD_IN_PROGRESS,
ZFS_ERR_BADPROP,
ZFS_ERR_VDEV_NOTSUP,
ZFS_ERR_NOT_USER_NAMESPACE,
ZFS_ERR_RESUME_EXISTS,
ZFS_ERR_CRYPTO_NOTSUP,
} zfs_errno_t;
/*
* Internal SPA load state. Used by FMA diagnosis engine.
*/
typedef enum {
SPA_LOAD_NONE, /* no load in progress */
SPA_LOAD_OPEN, /* normal open */
SPA_LOAD_IMPORT, /* import in progress */
SPA_LOAD_TRYIMPORT, /* tryimport in progress */
SPA_LOAD_RECOVER, /* recovery requested */
SPA_LOAD_ERROR, /* load failed */
SPA_LOAD_CREATE /* creation in progress */
} spa_load_state_t;
typedef enum {
ZPOOL_WAIT_CKPT_DISCARD,
ZPOOL_WAIT_FREE,
ZPOOL_WAIT_INITIALIZE,
ZPOOL_WAIT_REPLACE,
ZPOOL_WAIT_REMOVE,
ZPOOL_WAIT_RESILVER,
ZPOOL_WAIT_SCRUB,
ZPOOL_WAIT_TRIM,
ZPOOL_WAIT_NUM_ACTIVITIES
} zpool_wait_activity_t;
typedef enum {
ZFS_WAIT_DELETEQ,
ZFS_WAIT_NUM_ACTIVITIES
} zfs_wait_activity_t;
/*
* Bookmark name values.
*/
#define ZPOOL_ERR_LIST "error list"
#define ZPOOL_ERR_DATASET "dataset"
#define ZPOOL_ERR_OBJECT "object"
#define HIS_MAX_RECORD_LEN (MAXPATHLEN + MAXPATHLEN + 1)
/*
* The following are names used in the nvlist describing
* the pool's history log.
*/
#define ZPOOL_HIST_RECORD "history record"
#define ZPOOL_HIST_TIME "history time"
#define ZPOOL_HIST_CMD "history command"
#define ZPOOL_HIST_WHO "history who"
#define ZPOOL_HIST_ZONE "history zone"
#define ZPOOL_HIST_HOST "history hostname"
#define ZPOOL_HIST_TXG "history txg"
#define ZPOOL_HIST_INT_EVENT "history internal event"
#define ZPOOL_HIST_INT_STR "history internal str"
#define ZPOOL_HIST_INT_NAME "internal_name"
#define ZPOOL_HIST_IOCTL "ioctl"
#define ZPOOL_HIST_INPUT_NVL "in_nvl"
#define ZPOOL_HIST_OUTPUT_NVL "out_nvl"
#define ZPOOL_HIST_OUTPUT_SIZE "out_size"
#define ZPOOL_HIST_DSNAME "dsname"
#define ZPOOL_HIST_DSID "dsid"
#define ZPOOL_HIST_ERRNO "errno"
#define ZPOOL_HIST_ELAPSED_NS "elapsed_ns"
/*
* Special nvlist name that will not have its args recorded in the pool's
* history log.
*/
#define ZPOOL_HIDDEN_ARGS "hidden_args"
/*
* The following are names used when invoking ZFS_IOC_POOL_INITIALIZE.
*/
#define ZPOOL_INITIALIZE_COMMAND "initialize_command"
#define ZPOOL_INITIALIZE_VDEVS "initialize_vdevs"
/*
* The following are names used when invoking ZFS_IOC_POOL_TRIM.
*/
#define ZPOOL_TRIM_COMMAND "trim_command"
#define ZPOOL_TRIM_VDEVS "trim_vdevs"
#define ZPOOL_TRIM_RATE "trim_rate"
#define ZPOOL_TRIM_SECURE "trim_secure"
/*
* The following are names used when invoking ZFS_IOC_POOL_WAIT.
*/
#define ZPOOL_WAIT_ACTIVITY "wait_activity"
#define ZPOOL_WAIT_TAG "wait_tag"
#define ZPOOL_WAIT_WAITED "wait_waited"
/*
* The following are names used when invoking ZFS_IOC_VDEV_GET_PROP.
*/
#define ZPOOL_VDEV_PROPS_GET_VDEV "vdevprops_get_vdev"
#define ZPOOL_VDEV_PROPS_GET_PROPS "vdevprops_get_props"
/*
* The following are names used when invoking ZFS_IOC_VDEV_SET_PROP.
*/
#define ZPOOL_VDEV_PROPS_SET_VDEV "vdevprops_set_vdev"
#define ZPOOL_VDEV_PROPS_SET_PROPS "vdevprops_set_props"
/*
* The following are names used when invoking ZFS_IOC_WAIT_FS.
*/
#define ZFS_WAIT_ACTIVITY "wait_activity"
#define ZFS_WAIT_WAITED "wait_waited"
/*
* Flags for ZFS_IOC_VDEV_SET_STATE
*/
#define ZFS_ONLINE_CHECKREMOVE 0x1
#define ZFS_ONLINE_UNSPARE 0x2
#define ZFS_ONLINE_FORCEFAULT 0x4
#define ZFS_ONLINE_EXPAND 0x8
#define ZFS_ONLINE_SPARE 0x10
#define ZFS_OFFLINE_TEMPORARY 0x1
/*
* Flags for ZFS_IOC_POOL_IMPORT
*/
#define ZFS_IMPORT_NORMAL 0x0
#define ZFS_IMPORT_VERBATIM 0x1
#define ZFS_IMPORT_ANY_HOST 0x2
#define ZFS_IMPORT_MISSING_LOG 0x4
#define ZFS_IMPORT_ONLY 0x8
#define ZFS_IMPORT_TEMP_NAME 0x10
#define ZFS_IMPORT_SKIP_MMP 0x20
#define ZFS_IMPORT_LOAD_KEYS 0x40
#define ZFS_IMPORT_CHECKPOINT 0x80
/*
* Channel program argument/return nvlist keys and defaults.
*/
#define ZCP_ARG_PROGRAM "program"
#define ZCP_ARG_ARGLIST "arg"
#define ZCP_ARG_SYNC "sync"
#define ZCP_ARG_INSTRLIMIT "instrlimit"
#define ZCP_ARG_MEMLIMIT "memlimit"
#define ZCP_ARG_CLIARGV "argv"
#define ZCP_RET_ERROR "error"
#define ZCP_RET_RETURN "return"
#define ZCP_DEFAULT_INSTRLIMIT (10 * 1000 * 1000)
#define ZCP_MAX_INSTRLIMIT (10 * ZCP_DEFAULT_INSTRLIMIT)
#define ZCP_DEFAULT_MEMLIMIT (10 * 1024 * 1024)
#define ZCP_MAX_MEMLIMIT (10 * ZCP_DEFAULT_MEMLIMIT)
/*
* Sysevent payload members. ZFS will generate the following sysevents with the
* given payloads:
*
* ESC_ZFS_RESILVER_START
* ESC_ZFS_RESILVER_FINISH
*
* ZFS_EV_POOL_NAME DATA_TYPE_STRING
* ZFS_EV_POOL_GUID DATA_TYPE_UINT64
* ZFS_EV_RESILVER_TYPE DATA_TYPE_STRING
*
* ESC_ZFS_POOL_DESTROY
* ESC_ZFS_POOL_REGUID
*
* ZFS_EV_POOL_NAME DATA_TYPE_STRING
* ZFS_EV_POOL_GUID DATA_TYPE_UINT64
*
* ESC_ZFS_VDEV_REMOVE
* ESC_ZFS_VDEV_CLEAR
* ESC_ZFS_VDEV_CHECK
*
* ZFS_EV_POOL_NAME DATA_TYPE_STRING
* ZFS_EV_POOL_GUID DATA_TYPE_UINT64
* ZFS_EV_VDEV_PATH DATA_TYPE_STRING (optional)
* ZFS_EV_VDEV_GUID DATA_TYPE_UINT64
*
* ESC_ZFS_HISTORY_EVENT
*
* ZFS_EV_POOL_NAME DATA_TYPE_STRING
* ZFS_EV_POOL_GUID DATA_TYPE_UINT64
* ZFS_EV_HIST_TIME DATA_TYPE_UINT64 (optional)
* ZFS_EV_HIST_CMD DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_WHO DATA_TYPE_UINT64 (optional)
* ZFS_EV_HIST_ZONE DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_HOST DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_TXG DATA_TYPE_UINT64 (optional)
* ZFS_EV_HIST_INT_EVENT DATA_TYPE_UINT64 (optional)
* ZFS_EV_HIST_INT_STR DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_INT_NAME DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_IOCTL DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_DSNAME DATA_TYPE_STRING (optional)
* ZFS_EV_HIST_DSID DATA_TYPE_UINT64 (optional)
*
* The ZFS_EV_HIST_* members will correspond to the ZPOOL_HIST_* members in the
* history log nvlist. The keynames will be free of any spaces or other
* characters that could be potentially unexpected to consumers of the
* sysevents.
*/
#define ZFS_EV_POOL_NAME "pool_name"
#define ZFS_EV_POOL_GUID "pool_guid"
#define ZFS_EV_VDEV_PATH "vdev_path"
#define ZFS_EV_VDEV_GUID "vdev_guid"
#define ZFS_EV_HIST_TIME "history_time"
#define ZFS_EV_HIST_CMD "history_command"
#define ZFS_EV_HIST_WHO "history_who"
#define ZFS_EV_HIST_ZONE "history_zone"
#define ZFS_EV_HIST_HOST "history_hostname"
#define ZFS_EV_HIST_TXG "history_txg"
#define ZFS_EV_HIST_INT_EVENT "history_internal_event"
#define ZFS_EV_HIST_INT_STR "history_internal_str"
#define ZFS_EV_HIST_INT_NAME "history_internal_name"
#define ZFS_EV_HIST_IOCTL "history_ioctl"
#define ZFS_EV_HIST_DSNAME "history_dsname"
#define ZFS_EV_HIST_DSID "history_dsid"
#define ZFS_EV_RESILVER_TYPE "resilver_type"
/*
* We currently support block sizes from 512 bytes to 16MB.
* The benefits of larger blocks, and thus larger IO, need to be weighed
* against the cost of COWing a giant block to modify one byte, and the
* large latency of reading or writing a large block.
*
* The recordsize property can not be set larger than zfs_max_recordsize
* (default 16MB on 64-bit and 1MB on 32-bit). See the comment near
* zfs_max_recordsize in dsl_dataset.c for details.
*
* Note that although the LSIZE field of the blkptr_t can store sizes up
* to 32MB, the dnode's dn_datablkszsec can only store sizes up to
* 32MB - 512 bytes. Therefore, we limit SPA_MAXBLOCKSIZE to 16MB.
*/
#define SPA_MINBLOCKSHIFT 9
#define SPA_OLD_MAXBLOCKSHIFT 17
#define SPA_MAXBLOCKSHIFT 24
#define SPA_MINBLOCKSIZE (1ULL << SPA_MINBLOCKSHIFT)
#define SPA_OLD_MAXBLOCKSIZE (1ULL << SPA_OLD_MAXBLOCKSHIFT)
#define SPA_MAXBLOCKSIZE (1ULL << SPA_MAXBLOCKSHIFT)
/* supported encryption algorithms */
enum zio_encrypt {
ZIO_CRYPT_INHERIT = 0,
ZIO_CRYPT_ON,
ZIO_CRYPT_OFF,
ZIO_CRYPT_AES_128_CCM,
ZIO_CRYPT_AES_192_CCM,
ZIO_CRYPT_AES_256_CCM,
ZIO_CRYPT_AES_128_GCM,
ZIO_CRYPT_AES_192_GCM,
ZIO_CRYPT_AES_256_GCM,
ZIO_CRYPT_FUNCTIONS
};
#define ZIO_CRYPT_ON_VALUE ZIO_CRYPT_AES_256_GCM
#define ZIO_CRYPT_DEFAULT ZIO_CRYPT_OFF
/*
* xattr namespace prefixes. These are forbidden in xattr names.
*
* For cross-platform compatibility, xattrs in the user namespace should not be
* prefixed with the namespace name, but for backwards compatibility with older
* ZFS on Linux versions we do prefix the namespace.
*/
#define ZFS_XA_NS_FREEBSD_PREFIX "freebsd:"
#define ZFS_XA_NS_FREEBSD_PREFIX_LEN strlen("freebsd:")
#define ZFS_XA_NS_LINUX_SECURITY_PREFIX "security."
#define ZFS_XA_NS_LINUX_SECURITY_PREFIX_LEN strlen("security.")
#define ZFS_XA_NS_LINUX_SYSTEM_PREFIX "system."
#define ZFS_XA_NS_LINUX_SYSTEM_PREFIX_LEN strlen("system.")
#define ZFS_XA_NS_LINUX_TRUSTED_PREFIX "trusted."
#define ZFS_XA_NS_LINUX_TRUSTED_PREFIX_LEN strlen("trusted.")
#define ZFS_XA_NS_LINUX_USER_PREFIX "user."
#define ZFS_XA_NS_LINUX_USER_PREFIX_LEN strlen("user.")
#define ZFS_XA_NS_PREFIX_MATCH(ns, name) \
(strncmp(name, ZFS_XA_NS_##ns##_PREFIX, \
ZFS_XA_NS_##ns##_PREFIX_LEN) == 0)
#define ZFS_XA_NS_PREFIX_FORBIDDEN(name) \
(ZFS_XA_NS_PREFIX_MATCH(FREEBSD, name) || \
ZFS_XA_NS_PREFIX_MATCH(LINUX_SECURITY, name) || \
ZFS_XA_NS_PREFIX_MATCH(LINUX_SYSTEM, name) || \
ZFS_XA_NS_PREFIX_MATCH(LINUX_TRUSTED, name) || \
ZFS_XA_NS_PREFIX_MATCH(LINUX_USER, name))
#ifdef __cplusplus
}
#endif
#endif /* _SYS_FS_ZFS_H */
diff --git a/sys/contrib/openzfs/include/sys/spa.h b/sys/contrib/openzfs/include/sys/spa.h
index 79c46aa07709..b96a9ef1d42f 100644
--- a/sys/contrib/openzfs/include/sys/spa.h
+++ b/sys/contrib/openzfs/include/sys/spa.h
@@ -1,1235 +1,1238 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2021 by Delphix. All rights reserved.
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
* Copyright 2013 Saso Kiselkov. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2017 Joyent, Inc.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, Allan Jude
* Copyright (c) 2019, Klara Inc.
* Copyright (c) 2019, Datto Inc.
*/
#ifndef _SYS_SPA_H
#define _SYS_SPA_H
#include <sys/avl.h>
#include <sys/zfs_context.h>
#include <sys/kstat.h>
#include <sys/nvpair.h>
#include <sys/sysmacros.h>
#include <sys/types.h>
#include <sys/fs/zfs.h>
#include <sys/spa_checksum.h>
#include <sys/dmu.h>
#include <sys/space_map.h>
#include <sys/bitops.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* Forward references that lots of things need.
*/
typedef struct spa spa_t;
typedef struct vdev vdev_t;
typedef struct metaslab metaslab_t;
typedef struct metaslab_group metaslab_group_t;
typedef struct metaslab_class metaslab_class_t;
typedef struct zio zio_t;
typedef struct zilog zilog_t;
typedef struct spa_aux_vdev spa_aux_vdev_t;
typedef struct ddt ddt_t;
typedef struct ddt_entry ddt_entry_t;
typedef struct zbookmark_phys zbookmark_phys_t;
typedef struct zbookmark_err_phys zbookmark_err_phys_t;
struct bpobj;
struct bplist;
struct dsl_pool;
struct dsl_dataset;
struct dsl_crypto_params;
/*
* Alignment Shift (ashift) is an immutable, internal top-level vdev property
* which can only be set at vdev creation time. Physical writes are always done
* according to it, which makes 2^ashift the smallest possible IO on a vdev.
*
* We currently allow values ranging from 512 bytes (2^9 = 512) to 64 KiB
* (2^16 = 65,536).
*/
#define ASHIFT_MIN 9
#define ASHIFT_MAX 16
/*
* Size of block to hold the configuration data (a packed nvlist)
*/
#define SPA_CONFIG_BLOCKSIZE (1ULL << 14)
/*
* The DVA size encodings for LSIZE and PSIZE support blocks up to 32MB.
* The ASIZE encoding should be at least 64 times larger (6 more bits)
* to support up to 4-way RAID-Z mirror mode with worst-case gang block
* overhead, three DVAs per bp, plus one more bit in case we do anything
* else that expands the ASIZE.
*/
#define SPA_LSIZEBITS 16 /* LSIZE up to 32M (2^16 * 512) */
#define SPA_PSIZEBITS 16 /* PSIZE up to 32M (2^16 * 512) */
#define SPA_ASIZEBITS 24 /* ASIZE up to 64 times larger */
#define SPA_COMPRESSBITS 7
#define SPA_VDEVBITS 24
#define SPA_COMPRESSMASK ((1U << SPA_COMPRESSBITS) - 1)
/*
* All SPA data is represented by 128-bit data virtual addresses (DVAs).
* The members of the dva_t should be considered opaque outside the SPA.
*/
typedef struct dva {
uint64_t dva_word[2];
} dva_t;
/*
* Some checksums/hashes need a 256-bit initialization salt. This salt is kept
* secret and is suitable for use in MAC algorithms as the key.
*/
typedef struct zio_cksum_salt {
uint8_t zcs_bytes[32];
} zio_cksum_salt_t;
/*
* Each block is described by its DVAs, time of birth, checksum, etc.
* The word-by-word, bit-by-bit layout of the blkptr is as follows:
*
* 64 56 48 40 32 24 16 8 0
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 0 | pad | vdev1 | GRID | ASIZE |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 1 |G| offset1 |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 2 | pad | vdev2 | GRID | ASIZE |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 3 |G| offset2 |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 4 | pad | vdev3 | GRID | ASIZE |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 5 |G| offset3 |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 6 |BDX|lvl| type | cksum |E| comp| PSIZE | LSIZE |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 7 | padding |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 8 | padding |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 9 | physical birth txg |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* a | logical birth txg |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* b | fill count |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* c | checksum[0] |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* d | checksum[1] |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* e | checksum[2] |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* f | checksum[3] |
* +-------+-------+-------+-------+-------+-------+-------+-------+
*
* Legend:
*
* vdev virtual device ID
* offset offset into virtual device
* LSIZE logical size
* PSIZE physical size (after compression)
* ASIZE allocated size (including RAID-Z parity and gang block headers)
* GRID RAID-Z layout information (reserved for future use)
* cksum checksum function
* comp compression function
* G gang block indicator
* B byteorder (endianness)
* D dedup
* X encryption
* E blkptr_t contains embedded data (see below)
* lvl level of indirection
* type DMU object type
* phys birth txg when dva[0] was written; zero if same as logical birth txg
* note that typically all the dva's would be written in this
* txg, but they could be different if they were moved by
* device removal.
* log. birth transaction group in which the block was logically born
* fill count number of non-zero blocks under this bp
* checksum[4] 256-bit checksum of the data this bp describes
*/
/*
* The blkptr_t's of encrypted blocks also need to store the encryption
* parameters so that the block can be decrypted. This layout is as follows:
*
* 64 56 48 40 32 24 16 8 0
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 0 | vdev1 | GRID | ASIZE |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 1 |G| offset1 |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 2 | vdev2 | GRID | ASIZE |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 3 |G| offset2 |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 4 | salt |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 5 | IV1 |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 6 |BDX|lvl| type | cksum |E| comp| PSIZE | LSIZE |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 7 | padding |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 8 | padding |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 9 | physical birth txg |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* a | logical birth txg |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* b | IV2 | fill count |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* c | checksum[0] |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* d | checksum[1] |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* e | MAC[0] |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* f | MAC[1] |
* +-------+-------+-------+-------+-------+-------+-------+-------+
*
* Legend:
*
* salt Salt for generating encryption keys
* IV1 First 64 bits of encryption IV
* X Block requires encryption handling (set to 1)
* E blkptr_t contains embedded data (set to 0, see below)
* fill count number of non-zero blocks under this bp (truncated to 32 bits)
* IV2 Last 32 bits of encryption IV
* checksum[2] 128-bit checksum of the data this bp describes
* MAC[2] 128-bit message authentication code for this data
*
* The X bit being set indicates that this block is one of 3 types. If this is
* a level 0 block with an encrypted object type, the block is encrypted
* (see BP_IS_ENCRYPTED()). If this is a level 0 block with an unencrypted
* object type, this block is authenticated with an HMAC (see
* BP_IS_AUTHENTICATED()). Otherwise (if level > 0), this bp will use the MAC
* words to store a checksum-of-MACs from the level below (see
* BP_HAS_INDIRECT_MAC_CKSUM()). For convenience in the code, BP_IS_PROTECTED()
* refers to both encrypted and authenticated blocks and BP_USES_CRYPT()
* refers to any of these 3 kinds of blocks.
*
* The additional encryption parameters are the salt, IV, and MAC which are
* explained in greater detail in the block comment at the top of zio_crypt.c.
* The MAC occupies half of the checksum space since it serves a very similar
* purpose: to prevent data corruption on disk. The only functional difference
* is that the checksum is used to detect on-disk corruption whether or not the
* encryption key is loaded and the MAC provides additional protection against
* malicious disk tampering. We use the 3rd DVA to store the salt and first
* 64 bits of the IV. As a result encrypted blocks can only have 2 copies
* maximum instead of the normal 3. The last 32 bits of the IV are stored in
* the upper bits of what is usually the fill count. Note that only blocks at
* level 0 or -2 are ever encrypted, which allows us to guarantee that these
* 32 bits are not trampled over by other code (see zio_crypt.c for details).
* The salt and IV are not used for authenticated bps or bps with an indirect
* MAC checksum, so these blocks can utilize all 3 DVAs and the full 64 bits
* for the fill count.
*/
/*
* "Embedded" blkptr_t's don't actually point to a block, instead they
* have a data payload embedded in the blkptr_t itself. See the comment
* in blkptr.c for more details.
*
* The blkptr_t is laid out as follows:
*
* 64 56 48 40 32 24 16 8 0
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 0 | payload |
* 1 | payload |
* 2 | payload |
* 3 | payload |
* 4 | payload |
* 5 | payload |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 6 |BDX|lvl| type | etype |E| comp| PSIZE| LSIZE |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* 7 | payload |
* 8 | payload |
* 9 | payload |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* a | logical birth txg |
* +-------+-------+-------+-------+-------+-------+-------+-------+
* b | payload |
* c | payload |
* d | payload |
* e | payload |
* f | payload |
* +-------+-------+-------+-------+-------+-------+-------+-------+
*
* Legend:
*
* payload contains the embedded data
* B (byteorder) byteorder (endianness)
* D (dedup) padding (set to zero)
* X encryption (set to zero)
* E (embedded) set to one
* lvl indirection level
* type DMU object type
* etype how to interpret embedded data (BP_EMBEDDED_TYPE_*)
* comp compression function of payload
* PSIZE size of payload after compression, in bytes
* LSIZE logical size of payload, in bytes
* note that 25 bits is enough to store the largest
* "normal" BP's LSIZE (2^16 * 2^9) in bytes
* log. birth transaction group in which the block was logically born
*
* Note that LSIZE and PSIZE are stored in bytes, whereas for non-embedded
* bp's they are stored in units of SPA_MINBLOCKSHIFT.
* Generally, the generic BP_GET_*() macros can be used on embedded BP's.
* The B, D, X, lvl, type, and comp fields are stored the same as with normal
* BP's so the BP_SET_* macros can be used with them. etype, PSIZE, LSIZE must
* be set with the BPE_SET_* macros. BP_SET_EMBEDDED() should be called before
* other macros, as they assert that they are only used on BP's of the correct
* "embedded-ness". Encrypted blkptr_t's cannot be embedded because they use
* the payload space for encryption parameters (see the comment above on
* how encryption parameters are stored).
*/
#define BPE_GET_ETYPE(bp) \
(ASSERT(BP_IS_EMBEDDED(bp)), \
BF64_GET((bp)->blk_prop, 40, 8))
#define BPE_SET_ETYPE(bp, t) do { \
ASSERT(BP_IS_EMBEDDED(bp)); \
BF64_SET((bp)->blk_prop, 40, 8, t); \
} while (0)
#define BPE_GET_LSIZE(bp) \
(ASSERT(BP_IS_EMBEDDED(bp)), \
BF64_GET_SB((bp)->blk_prop, 0, 25, 0, 1))
#define BPE_SET_LSIZE(bp, x) do { \
ASSERT(BP_IS_EMBEDDED(bp)); \
BF64_SET_SB((bp)->blk_prop, 0, 25, 0, 1, x); \
} while (0)
#define BPE_GET_PSIZE(bp) \
(ASSERT(BP_IS_EMBEDDED(bp)), \
BF64_GET_SB((bp)->blk_prop, 25, 7, 0, 1))
#define BPE_SET_PSIZE(bp, x) do { \
ASSERT(BP_IS_EMBEDDED(bp)); \
BF64_SET_SB((bp)->blk_prop, 25, 7, 0, 1, x); \
} while (0)
typedef enum bp_embedded_type {
BP_EMBEDDED_TYPE_DATA,
BP_EMBEDDED_TYPE_RESERVED, /* Reserved for Delphix byteswap feature. */
BP_EMBEDDED_TYPE_REDACTED,
NUM_BP_EMBEDDED_TYPES
} bp_embedded_type_t;
#define BPE_NUM_WORDS 14
#define BPE_PAYLOAD_SIZE (BPE_NUM_WORDS * sizeof (uint64_t))
#define BPE_IS_PAYLOADWORD(bp, wp) \
((wp) != &(bp)->blk_prop && (wp) != &(bp)->blk_birth)
#define SPA_BLKPTRSHIFT 7 /* blkptr_t is 128 bytes */
#define SPA_DVAS_PER_BP 3 /* Number of DVAs in a bp */
#define SPA_SYNC_MIN_VDEVS 3 /* min vdevs to update during sync */
/*
* A block is a hole when it has either 1) never been written to, or
* 2) is zero-filled. In both cases, ZFS can return all zeroes for all reads
* without physically allocating disk space. Holes are represented in the
* blkptr_t structure by zeroed blk_dva. Correct checking for holes is
* done through the BP_IS_HOLE macro. For holes, the logical size, level,
* DMU object type, and birth times are all also stored for holes that
* were written to at some point (i.e. were punched after having been filled).
*/
typedef struct blkptr {
dva_t blk_dva[SPA_DVAS_PER_BP]; /* Data Virtual Addresses */
uint64_t blk_prop; /* size, compression, type, etc */
uint64_t blk_pad[2]; /* Extra space for the future */
uint64_t blk_phys_birth; /* txg when block was allocated */
uint64_t blk_birth; /* transaction group at birth */
uint64_t blk_fill; /* fill count */
zio_cksum_t blk_cksum; /* 256-bit checksum */
} blkptr_t;
/*
* Macros to get and set fields in a bp or DVA.
*/
/*
* Note, for gang blocks, DVA_GET_ASIZE() is the total space allocated for
* this gang DVA including its children BP's. The space allocated at this
* DVA's vdev/offset is vdev_gang_header_asize(vdev).
*/
#define DVA_GET_ASIZE(dva) \
BF64_GET_SB((dva)->dva_word[0], 0, SPA_ASIZEBITS, SPA_MINBLOCKSHIFT, 0)
#define DVA_SET_ASIZE(dva, x) \
BF64_SET_SB((dva)->dva_word[0], 0, SPA_ASIZEBITS, \
SPA_MINBLOCKSHIFT, 0, x)
#define DVA_GET_GRID(dva) BF64_GET((dva)->dva_word[0], 24, 8)
#define DVA_SET_GRID(dva, x) BF64_SET((dva)->dva_word[0], 24, 8, x)
#define DVA_GET_VDEV(dva) BF64_GET((dva)->dva_word[0], 32, SPA_VDEVBITS)
#define DVA_SET_VDEV(dva, x) \
BF64_SET((dva)->dva_word[0], 32, SPA_VDEVBITS, x)
#define DVA_GET_OFFSET(dva) \
BF64_GET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0)
#define DVA_SET_OFFSET(dva, x) \
BF64_SET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0, x)
#define DVA_GET_GANG(dva) BF64_GET((dva)->dva_word[1], 63, 1)
#define DVA_SET_GANG(dva, x) BF64_SET((dva)->dva_word[1], 63, 1, x)
#define BP_GET_LSIZE(bp) \
(BP_IS_EMBEDDED(bp) ? \
(BPE_GET_ETYPE(bp) == BP_EMBEDDED_TYPE_DATA ? BPE_GET_LSIZE(bp) : 0): \
BF64_GET_SB((bp)->blk_prop, 0, SPA_LSIZEBITS, SPA_MINBLOCKSHIFT, 1))
#define BP_SET_LSIZE(bp, x) do { \
ASSERT(!BP_IS_EMBEDDED(bp)); \
BF64_SET_SB((bp)->blk_prop, \
0, SPA_LSIZEBITS, SPA_MINBLOCKSHIFT, 1, x); \
} while (0)
#define BP_GET_PSIZE(bp) \
(BP_IS_EMBEDDED(bp) ? 0 : \
BF64_GET_SB((bp)->blk_prop, 16, SPA_PSIZEBITS, SPA_MINBLOCKSHIFT, 1))
#define BP_SET_PSIZE(bp, x) do { \
ASSERT(!BP_IS_EMBEDDED(bp)); \
BF64_SET_SB((bp)->blk_prop, \
16, SPA_PSIZEBITS, SPA_MINBLOCKSHIFT, 1, x); \
} while (0)
#define BP_GET_COMPRESS(bp) \
BF64_GET((bp)->blk_prop, 32, SPA_COMPRESSBITS)
#define BP_SET_COMPRESS(bp, x) \
BF64_SET((bp)->blk_prop, 32, SPA_COMPRESSBITS, x)
#define BP_IS_EMBEDDED(bp) BF64_GET((bp)->blk_prop, 39, 1)
#define BP_SET_EMBEDDED(bp, x) BF64_SET((bp)->blk_prop, 39, 1, x)
#define BP_GET_CHECKSUM(bp) \
(BP_IS_EMBEDDED(bp) ? ZIO_CHECKSUM_OFF : \
BF64_GET((bp)->blk_prop, 40, 8))
#define BP_SET_CHECKSUM(bp, x) do { \
ASSERT(!BP_IS_EMBEDDED(bp)); \
BF64_SET((bp)->blk_prop, 40, 8, x); \
} while (0)
#define BP_GET_TYPE(bp) BF64_GET((bp)->blk_prop, 48, 8)
#define BP_SET_TYPE(bp, x) BF64_SET((bp)->blk_prop, 48, 8, x)
#define BP_GET_LEVEL(bp) BF64_GET((bp)->blk_prop, 56, 5)
#define BP_SET_LEVEL(bp, x) BF64_SET((bp)->blk_prop, 56, 5, x)
/* encrypted, authenticated, and MAC cksum bps use the same bit */
#define BP_USES_CRYPT(bp) BF64_GET((bp)->blk_prop, 61, 1)
#define BP_SET_CRYPT(bp, x) BF64_SET((bp)->blk_prop, 61, 1, x)
#define BP_IS_ENCRYPTED(bp) \
(BP_USES_CRYPT(bp) && \
BP_GET_LEVEL(bp) <= 0 && \
DMU_OT_IS_ENCRYPTED(BP_GET_TYPE(bp)))
#define BP_IS_AUTHENTICATED(bp) \
(BP_USES_CRYPT(bp) && \
BP_GET_LEVEL(bp) <= 0 && \
!DMU_OT_IS_ENCRYPTED(BP_GET_TYPE(bp)))
#define BP_HAS_INDIRECT_MAC_CKSUM(bp) \
(BP_USES_CRYPT(bp) && BP_GET_LEVEL(bp) > 0)
#define BP_IS_PROTECTED(bp) \
(BP_IS_ENCRYPTED(bp) || BP_IS_AUTHENTICATED(bp))
#define BP_GET_DEDUP(bp) BF64_GET((bp)->blk_prop, 62, 1)
#define BP_SET_DEDUP(bp, x) BF64_SET((bp)->blk_prop, 62, 1, x)
#define BP_GET_BYTEORDER(bp) BF64_GET((bp)->blk_prop, 63, 1)
#define BP_SET_BYTEORDER(bp, x) BF64_SET((bp)->blk_prop, 63, 1, x)
#define BP_GET_FREE(bp) BF64_GET((bp)->blk_fill, 0, 1)
#define BP_SET_FREE(bp, x) BF64_SET((bp)->blk_fill, 0, 1, x)
#define BP_PHYSICAL_BIRTH(bp) \
(BP_IS_EMBEDDED(bp) ? 0 : \
(bp)->blk_phys_birth ? (bp)->blk_phys_birth : (bp)->blk_birth)
#define BP_SET_BIRTH(bp, logical, physical) \
{ \
ASSERT(!BP_IS_EMBEDDED(bp)); \
(bp)->blk_birth = (logical); \
(bp)->blk_phys_birth = ((logical) == (physical) ? 0 : (physical)); \
}
#define BP_GET_FILL(bp) \
((BP_IS_ENCRYPTED(bp)) ? BF64_GET((bp)->blk_fill, 0, 32) : \
((BP_IS_EMBEDDED(bp)) ? 1 : (bp)->blk_fill))
#define BP_SET_FILL(bp, fill) \
{ \
if (BP_IS_ENCRYPTED(bp)) \
BF64_SET((bp)->blk_fill, 0, 32, fill); \
else \
(bp)->blk_fill = fill; \
}
#define BP_GET_IV2(bp) \
(ASSERT(BP_IS_ENCRYPTED(bp)), \
BF64_GET((bp)->blk_fill, 32, 32))
#define BP_SET_IV2(bp, iv2) \
{ \
ASSERT(BP_IS_ENCRYPTED(bp)); \
BF64_SET((bp)->blk_fill, 32, 32, iv2); \
}
#define BP_IS_METADATA(bp) \
(BP_GET_LEVEL(bp) > 0 || DMU_OT_IS_METADATA(BP_GET_TYPE(bp)))
#define BP_GET_ASIZE(bp) \
(BP_IS_EMBEDDED(bp) ? 0 : \
DVA_GET_ASIZE(&(bp)->blk_dva[0]) + \
DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \
(DVA_GET_ASIZE(&(bp)->blk_dva[2]) * !BP_IS_ENCRYPTED(bp)))
#define BP_GET_UCSIZE(bp) \
(BP_IS_METADATA(bp) ? BP_GET_PSIZE(bp) : BP_GET_LSIZE(bp))
#define BP_GET_NDVAS(bp) \
(BP_IS_EMBEDDED(bp) ? 0 : \
!!DVA_GET_ASIZE(&(bp)->blk_dva[0]) + \
!!DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \
(!!DVA_GET_ASIZE(&(bp)->blk_dva[2]) * !BP_IS_ENCRYPTED(bp)))
#define BP_COUNT_GANG(bp) \
(BP_IS_EMBEDDED(bp) ? 0 : \
(DVA_GET_GANG(&(bp)->blk_dva[0]) + \
DVA_GET_GANG(&(bp)->blk_dva[1]) + \
(DVA_GET_GANG(&(bp)->blk_dva[2]) * !BP_IS_ENCRYPTED(bp))))
#define DVA_EQUAL(dva1, dva2) \
((dva1)->dva_word[1] == (dva2)->dva_word[1] && \
(dva1)->dva_word[0] == (dva2)->dva_word[0])
#define BP_EQUAL(bp1, bp2) \
(BP_PHYSICAL_BIRTH(bp1) == BP_PHYSICAL_BIRTH(bp2) && \
(bp1)->blk_birth == (bp2)->blk_birth && \
DVA_EQUAL(&(bp1)->blk_dva[0], &(bp2)->blk_dva[0]) && \
DVA_EQUAL(&(bp1)->blk_dva[1], &(bp2)->blk_dva[1]) && \
DVA_EQUAL(&(bp1)->blk_dva[2], &(bp2)->blk_dva[2]))
#define DVA_IS_VALID(dva) (DVA_GET_ASIZE(dva) != 0)
#define BP_IDENTITY(bp) (ASSERT(!BP_IS_EMBEDDED(bp)), &(bp)->blk_dva[0])
#define BP_IS_GANG(bp) \
(BP_IS_EMBEDDED(bp) ? B_FALSE : DVA_GET_GANG(BP_IDENTITY(bp)))
#define DVA_IS_EMPTY(dva) ((dva)->dva_word[0] == 0ULL && \
(dva)->dva_word[1] == 0ULL)
#define BP_IS_HOLE(bp) \
(!BP_IS_EMBEDDED(bp) && DVA_IS_EMPTY(BP_IDENTITY(bp)))
#define BP_SET_REDACTED(bp) \
{ \
BP_SET_EMBEDDED(bp, B_TRUE); \
BPE_SET_ETYPE(bp, BP_EMBEDDED_TYPE_REDACTED); \
}
#define BP_IS_REDACTED(bp) \
(BP_IS_EMBEDDED(bp) && BPE_GET_ETYPE(bp) == BP_EMBEDDED_TYPE_REDACTED)
/* BP_IS_RAIDZ(bp) assumes no block compression */
#define BP_IS_RAIDZ(bp) (DVA_GET_ASIZE(&(bp)->blk_dva[0]) > \
BP_GET_PSIZE(bp))
#define BP_ZERO(bp) \
{ \
(bp)->blk_dva[0].dva_word[0] = 0; \
(bp)->blk_dva[0].dva_word[1] = 0; \
(bp)->blk_dva[1].dva_word[0] = 0; \
(bp)->blk_dva[1].dva_word[1] = 0; \
(bp)->blk_dva[2].dva_word[0] = 0; \
(bp)->blk_dva[2].dva_word[1] = 0; \
(bp)->blk_prop = 0; \
(bp)->blk_pad[0] = 0; \
(bp)->blk_pad[1] = 0; \
(bp)->blk_phys_birth = 0; \
(bp)->blk_birth = 0; \
(bp)->blk_fill = 0; \
ZIO_SET_CHECKSUM(&(bp)->blk_cksum, 0, 0, 0, 0); \
}
#ifdef _ZFS_BIG_ENDIAN
#define ZFS_HOST_BYTEORDER (0ULL)
#else
#define ZFS_HOST_BYTEORDER (1ULL)
#endif
#define BP_SHOULD_BYTESWAP(bp) (BP_GET_BYTEORDER(bp) != ZFS_HOST_BYTEORDER)
#define BP_SPRINTF_LEN 400
/*
* This macro allows code sharing between zfs, libzpool, and mdb.
* 'func' is either kmem_scnprintf() or mdb_snprintf().
* 'ws' (whitespace) can be ' ' for single-line format, '\n' for multi-line.
*/
#define SNPRINTF_BLKPTR(func, ws, buf, size, bp, type, checksum, compress) \
{ \
static const char *const copyname[] = \
{ "zero", "single", "double", "triple" }; \
int len = 0; \
int copies = 0; \
const char *crypt_type; \
if (bp != NULL) { \
if (BP_IS_ENCRYPTED(bp)) { \
crypt_type = "encrypted"; \
/* LINTED E_SUSPICIOUS_COMPARISON */ \
} else if (BP_IS_AUTHENTICATED(bp)) { \
crypt_type = "authenticated"; \
} else if (BP_HAS_INDIRECT_MAC_CKSUM(bp)) { \
crypt_type = "indirect-MAC"; \
} else { \
crypt_type = "unencrypted"; \
} \
} \
if (bp == NULL) { \
len += func(buf + len, size - len, "<NULL>"); \
} else if (BP_IS_HOLE(bp)) { \
len += func(buf + len, size - len, \
"HOLE [L%llu %s] " \
"size=%llxL birth=%lluL", \
(u_longlong_t)BP_GET_LEVEL(bp), \
type, \
(u_longlong_t)BP_GET_LSIZE(bp), \
(u_longlong_t)bp->blk_birth); \
} else if (BP_IS_EMBEDDED(bp)) { \
len = func(buf + len, size - len, \
"EMBEDDED [L%llu %s] et=%u %s " \
"size=%llxL/%llxP birth=%lluL", \
(u_longlong_t)BP_GET_LEVEL(bp), \
type, \
(int)BPE_GET_ETYPE(bp), \
compress, \
(u_longlong_t)BPE_GET_LSIZE(bp), \
(u_longlong_t)BPE_GET_PSIZE(bp), \
(u_longlong_t)bp->blk_birth); \
} else if (BP_IS_REDACTED(bp)) { \
len += func(buf + len, size - len, \
"REDACTED [L%llu %s] size=%llxL birth=%lluL", \
(u_longlong_t)BP_GET_LEVEL(bp), \
type, \
(u_longlong_t)BP_GET_LSIZE(bp), \
(u_longlong_t)bp->blk_birth); \
} else { \
for (int d = 0; d < BP_GET_NDVAS(bp); d++) { \
const dva_t *dva = &bp->blk_dva[d]; \
if (DVA_IS_VALID(dva)) \
copies++; \
len += func(buf + len, size - len, \
"DVA[%d]=<%llu:%llx:%llx>%c", d, \
(u_longlong_t)DVA_GET_VDEV(dva), \
(u_longlong_t)DVA_GET_OFFSET(dva), \
(u_longlong_t)DVA_GET_ASIZE(dva), \
ws); \
} \
ASSERT3S(copies, >, 0); \
if (BP_IS_ENCRYPTED(bp)) { \
len += func(buf + len, size - len, \
"salt=%llx iv=%llx:%llx%c", \
(u_longlong_t)bp->blk_dva[2].dva_word[0], \
(u_longlong_t)bp->blk_dva[2].dva_word[1], \
(u_longlong_t)BP_GET_IV2(bp), \
ws); \
} \
if (BP_IS_GANG(bp) && \
DVA_GET_ASIZE(&bp->blk_dva[2]) <= \
DVA_GET_ASIZE(&bp->blk_dva[1]) / 2) \
copies--; \
len += func(buf + len, size - len, \
"[L%llu %s] %s %s %s %s %s %s %s%c" \
"size=%llxL/%llxP birth=%lluL/%lluP fill=%llu%c" \
"cksum=%016llx:%016llx:%016llx:%016llx", \
(u_longlong_t)BP_GET_LEVEL(bp), \
type, \
checksum, \
compress, \
crypt_type, \
BP_GET_BYTEORDER(bp) == 0 ? "BE" : "LE", \
BP_IS_GANG(bp) ? "gang" : "contiguous", \
BP_GET_DEDUP(bp) ? "dedup" : "unique", \
copyname[copies], \
ws, \
(u_longlong_t)BP_GET_LSIZE(bp), \
(u_longlong_t)BP_GET_PSIZE(bp), \
(u_longlong_t)bp->blk_birth, \
(u_longlong_t)BP_PHYSICAL_BIRTH(bp), \
(u_longlong_t)BP_GET_FILL(bp), \
ws, \
(u_longlong_t)bp->blk_cksum.zc_word[0], \
(u_longlong_t)bp->blk_cksum.zc_word[1], \
(u_longlong_t)bp->blk_cksum.zc_word[2], \
(u_longlong_t)bp->blk_cksum.zc_word[3]); \
} \
ASSERT(len < size); \
}
#define BP_GET_BUFC_TYPE(bp) \
(BP_IS_METADATA(bp) ? ARC_BUFC_METADATA : ARC_BUFC_DATA)
typedef enum spa_import_type {
SPA_IMPORT_EXISTING,
SPA_IMPORT_ASSEMBLE
} spa_import_type_t;
typedef enum spa_mode {
SPA_MODE_UNINIT = 0,
SPA_MODE_READ = 1,
SPA_MODE_WRITE = 2,
} spa_mode_t;
/*
* Send TRIM commands in-line during normal pool operation while deleting.
* OFF: no
* ON: yes
* NB: IN_FREEBSD_BASE is defined within the FreeBSD sources.
*/
typedef enum {
SPA_AUTOTRIM_OFF = 0, /* default */
SPA_AUTOTRIM_ON,
#ifdef IN_FREEBSD_BASE
SPA_AUTOTRIM_DEFAULT = SPA_AUTOTRIM_ON,
#else
SPA_AUTOTRIM_DEFAULT = SPA_AUTOTRIM_OFF,
#endif
} spa_autotrim_t;
/*
* Reason TRIM command was issued, used internally for accounting purposes.
*/
typedef enum trim_type {
TRIM_TYPE_MANUAL = 0,
TRIM_TYPE_AUTO = 1,
TRIM_TYPE_SIMPLE = 2
} trim_type_t;
/* state manipulation functions */
extern int spa_open(const char *pool, spa_t **, const void *tag);
extern int spa_open_rewind(const char *pool, spa_t **, const void *tag,
nvlist_t *policy, nvlist_t **config);
extern int spa_get_stats(const char *pool, nvlist_t **config, char *altroot,
size_t buflen);
extern int spa_create(const char *pool, nvlist_t *nvroot, nvlist_t *props,
nvlist_t *zplprops, struct dsl_crypto_params *dcp);
extern int spa_import(char *pool, nvlist_t *config, nvlist_t *props,
uint64_t flags);
extern nvlist_t *spa_tryimport(nvlist_t *tryconfig);
extern int spa_destroy(const char *pool);
extern int spa_checkpoint(const char *pool);
extern int spa_checkpoint_discard(const char *pool);
extern int spa_export(const char *pool, nvlist_t **oldconfig, boolean_t force,
boolean_t hardforce);
extern int spa_reset(const char *pool);
extern void spa_async_request(spa_t *spa, int flag);
extern void spa_async_unrequest(spa_t *spa, int flag);
extern void spa_async_suspend(spa_t *spa);
extern void spa_async_resume(spa_t *spa);
extern int spa_async_tasks(spa_t *spa);
extern spa_t *spa_inject_addref(char *pool);
extern void spa_inject_delref(spa_t *spa);
extern void spa_scan_stat_init(spa_t *spa);
extern int spa_scan_get_stats(spa_t *spa, pool_scan_stat_t *ps);
extern int bpobj_enqueue_alloc_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx);
extern int bpobj_enqueue_free_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx);
#define SPA_ASYNC_CONFIG_UPDATE 0x01
#define SPA_ASYNC_REMOVE 0x02
#define SPA_ASYNC_PROBE 0x04
#define SPA_ASYNC_RESILVER_DONE 0x08
#define SPA_ASYNC_RESILVER 0x10
#define SPA_ASYNC_AUTOEXPAND 0x20
#define SPA_ASYNC_REMOVE_DONE 0x40
#define SPA_ASYNC_REMOVE_STOP 0x80
#define SPA_ASYNC_INITIALIZE_RESTART 0x100
#define SPA_ASYNC_TRIM_RESTART 0x200
#define SPA_ASYNC_AUTOTRIM_RESTART 0x400
#define SPA_ASYNC_L2CACHE_REBUILD 0x800
#define SPA_ASYNC_L2CACHE_TRIM 0x1000
#define SPA_ASYNC_REBUILD_DONE 0x2000
+#define SPA_ASYNC_DETACH_SPARE 0x4000
/* device manipulation */
extern int spa_vdev_add(spa_t *spa, nvlist_t *nvroot);
extern int spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot,
int replacing, int rebuild);
extern int spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid,
int replace_done);
extern int spa_vdev_alloc(spa_t *spa, uint64_t guid);
extern int spa_vdev_noalloc(spa_t *spa, uint64_t guid);
extern boolean_t spa_vdev_remove_active(spa_t *spa);
extern int spa_vdev_initialize(spa_t *spa, nvlist_t *nv, uint64_t cmd_type,
nvlist_t *vdev_errlist);
extern int spa_vdev_trim(spa_t *spa, nvlist_t *nv, uint64_t cmd_type,
uint64_t rate, boolean_t partial, boolean_t secure, nvlist_t *vdev_errlist);
extern int spa_vdev_setpath(spa_t *spa, uint64_t guid, const char *newpath);
extern int spa_vdev_setfru(spa_t *spa, uint64_t guid, const char *newfru);
extern int spa_vdev_split_mirror(spa_t *spa, const char *newname,
nvlist_t *config, nvlist_t *props, boolean_t exp);
/* spare state (which is global across all pools) */
extern void spa_spare_add(vdev_t *vd);
extern void spa_spare_remove(vdev_t *vd);
extern boolean_t spa_spare_exists(uint64_t guid, uint64_t *pool, int *refcnt);
extern void spa_spare_activate(vdev_t *vd);
/* L2ARC state (which is global across all pools) */
extern void spa_l2cache_add(vdev_t *vd);
extern void spa_l2cache_remove(vdev_t *vd);
extern boolean_t spa_l2cache_exists(uint64_t guid, uint64_t *pool);
extern void spa_l2cache_activate(vdev_t *vd);
extern void spa_l2cache_drop(spa_t *spa);
/* scanning */
extern int spa_scan(spa_t *spa, pool_scan_func_t func);
extern int spa_scan_stop(spa_t *spa);
extern int spa_scrub_pause_resume(spa_t *spa, pool_scrub_cmd_t flag);
/* spa syncing */
extern void spa_sync(spa_t *spa, uint64_t txg); /* only for DMU use */
extern void spa_sync_allpools(void);
extern uint_t zfs_sync_pass_deferred_free;
/* spa namespace global mutex */
extern kmutex_t spa_namespace_lock;
/*
* SPA configuration functions in spa_config.c
*/
#define SPA_CONFIG_UPDATE_POOL 0
#define SPA_CONFIG_UPDATE_VDEVS 1
extern void spa_write_cachefile(spa_t *, boolean_t, boolean_t, boolean_t);
extern void spa_config_load(void);
extern nvlist_t *spa_all_configs(uint64_t *);
extern void spa_config_set(spa_t *spa, nvlist_t *config);
extern nvlist_t *spa_config_generate(spa_t *spa, vdev_t *vd, uint64_t txg,
int getstats);
extern void spa_config_update(spa_t *spa, int what);
extern int spa_config_parse(spa_t *spa, vdev_t **vdp, nvlist_t *nv,
vdev_t *parent, uint_t id, int atype);
/*
* Miscellaneous SPA routines in spa_misc.c
*/
/* Namespace manipulation */
extern spa_t *spa_lookup(const char *name);
extern spa_t *spa_add(const char *name, nvlist_t *config, const char *altroot);
extern void spa_remove(spa_t *spa);
extern spa_t *spa_next(spa_t *prev);
/* Refcount functions */
extern void spa_open_ref(spa_t *spa, const void *tag);
extern void spa_close(spa_t *spa, const void *tag);
extern void spa_async_close(spa_t *spa, const void *tag);
extern boolean_t spa_refcount_zero(spa_t *spa);
#define SCL_NONE 0x00
#define SCL_CONFIG 0x01
#define SCL_STATE 0x02
#define SCL_L2ARC 0x04 /* hack until L2ARC 2.0 */
#define SCL_ALLOC 0x08
#define SCL_ZIO 0x10
#define SCL_FREE 0x20
#define SCL_VDEV 0x40
#define SCL_LOCKS 7
#define SCL_ALL ((1 << SCL_LOCKS) - 1)
#define SCL_STATE_ALL (SCL_STATE | SCL_L2ARC | SCL_ZIO)
/* Historical pool statistics */
typedef struct spa_history_kstat {
kmutex_t lock;
uint64_t count;
uint64_t size;
kstat_t *kstat;
void *priv;
list_t list;
} spa_history_kstat_t;
typedef struct spa_history_list {
uint64_t size;
procfs_list_t procfs_list;
} spa_history_list_t;
typedef struct spa_stats {
spa_history_list_t read_history;
spa_history_list_t txg_history;
spa_history_kstat_t tx_assign_histogram;
spa_history_list_t mmp_history;
spa_history_kstat_t state; /* pool state */
spa_history_kstat_t guid; /* pool guid */
spa_history_kstat_t iostats;
} spa_stats_t;
typedef enum txg_state {
TXG_STATE_BIRTH = 0,
TXG_STATE_OPEN = 1,
TXG_STATE_QUIESCED = 2,
TXG_STATE_WAIT_FOR_SYNC = 3,
TXG_STATE_SYNCED = 4,
TXG_STATE_COMMITTED = 5,
} txg_state_t;
typedef struct txg_stat {
vdev_stat_t vs1;
vdev_stat_t vs2;
uint64_t txg;
uint64_t ndirty;
} txg_stat_t;
/* Assorted pool IO kstats */
typedef struct spa_iostats {
kstat_named_t trim_extents_written;
kstat_named_t trim_bytes_written;
kstat_named_t trim_extents_skipped;
kstat_named_t trim_bytes_skipped;
kstat_named_t trim_extents_failed;
kstat_named_t trim_bytes_failed;
kstat_named_t autotrim_extents_written;
kstat_named_t autotrim_bytes_written;
kstat_named_t autotrim_extents_skipped;
kstat_named_t autotrim_bytes_skipped;
kstat_named_t autotrim_extents_failed;
kstat_named_t autotrim_bytes_failed;
kstat_named_t simple_trim_extents_written;
kstat_named_t simple_trim_bytes_written;
kstat_named_t simple_trim_extents_skipped;
kstat_named_t simple_trim_bytes_skipped;
kstat_named_t simple_trim_extents_failed;
kstat_named_t simple_trim_bytes_failed;
} spa_iostats_t;
extern void spa_stats_init(spa_t *spa);
extern void spa_stats_destroy(spa_t *spa);
extern void spa_read_history_add(spa_t *spa, const zbookmark_phys_t *zb,
uint32_t aflags);
extern void spa_txg_history_add(spa_t *spa, uint64_t txg, hrtime_t birth_time);
extern int spa_txg_history_set(spa_t *spa, uint64_t txg,
txg_state_t completed_state, hrtime_t completed_time);
extern txg_stat_t *spa_txg_history_init_io(spa_t *, uint64_t,
struct dsl_pool *);
extern void spa_txg_history_fini_io(spa_t *, txg_stat_t *);
extern void spa_tx_assign_add_nsecs(spa_t *spa, uint64_t nsecs);
extern int spa_mmp_history_set_skip(spa_t *spa, uint64_t mmp_kstat_id);
extern int spa_mmp_history_set(spa_t *spa, uint64_t mmp_kstat_id, int io_error,
hrtime_t duration);
extern void spa_mmp_history_add(spa_t *spa, uint64_t txg, uint64_t timestamp,
uint64_t mmp_delay, vdev_t *vd, int label, uint64_t mmp_kstat_id,
int error);
extern void spa_iostats_trim_add(spa_t *spa, trim_type_t type,
uint64_t extents_written, uint64_t bytes_written,
uint64_t extents_skipped, uint64_t bytes_skipped,
uint64_t extents_failed, uint64_t bytes_failed);
extern void spa_import_progress_add(spa_t *spa);
extern void spa_import_progress_remove(uint64_t spa_guid);
extern int spa_import_progress_set_mmp_check(uint64_t pool_guid,
uint64_t mmp_sec_remaining);
extern int spa_import_progress_set_max_txg(uint64_t pool_guid,
uint64_t max_txg);
extern int spa_import_progress_set_state(uint64_t pool_guid,
spa_load_state_t spa_load_state);
/* Pool configuration locks */
extern int spa_config_tryenter(spa_t *spa, int locks, const void *tag,
krw_t rw);
extern void spa_config_enter(spa_t *spa, int locks, const void *tag, krw_t rw);
+extern void spa_config_enter_mmp(spa_t *spa, int locks, const void *tag,
+ krw_t rw);
extern void spa_config_exit(spa_t *spa, int locks, const void *tag);
extern int spa_config_held(spa_t *spa, int locks, krw_t rw);
/* Pool vdev add/remove lock */
extern uint64_t spa_vdev_enter(spa_t *spa);
extern uint64_t spa_vdev_detach_enter(spa_t *spa, uint64_t guid);
extern uint64_t spa_vdev_config_enter(spa_t *spa);
extern void spa_vdev_config_exit(spa_t *spa, vdev_t *vd, uint64_t txg,
int error, const char *tag);
extern int spa_vdev_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error);
/* Pool vdev state change lock */
extern void spa_vdev_state_enter(spa_t *spa, int oplock);
extern int spa_vdev_state_exit(spa_t *spa, vdev_t *vd, int error);
/* Log state */
typedef enum spa_log_state {
SPA_LOG_UNKNOWN = 0, /* unknown log state */
SPA_LOG_MISSING, /* missing log(s) */
SPA_LOG_CLEAR, /* clear the log(s) */
SPA_LOG_GOOD, /* log(s) are good */
} spa_log_state_t;
extern spa_log_state_t spa_get_log_state(spa_t *spa);
extern void spa_set_log_state(spa_t *spa, spa_log_state_t state);
extern int spa_reset_logs(spa_t *spa);
/* Log claim callback */
extern void spa_claim_notify(zio_t *zio);
extern void spa_deadman(void *);
/* Accessor functions */
extern boolean_t spa_shutting_down(spa_t *spa);
extern struct dsl_pool *spa_get_dsl(spa_t *spa);
extern boolean_t spa_is_initializing(spa_t *spa);
extern boolean_t spa_indirect_vdevs_loaded(spa_t *spa);
extern blkptr_t *spa_get_rootblkptr(spa_t *spa);
extern void spa_set_rootblkptr(spa_t *spa, const blkptr_t *bp);
extern void spa_altroot(spa_t *, char *, size_t);
extern uint32_t spa_sync_pass(spa_t *spa);
extern char *spa_name(spa_t *spa);
extern uint64_t spa_guid(spa_t *spa);
extern uint64_t spa_load_guid(spa_t *spa);
extern uint64_t spa_last_synced_txg(spa_t *spa);
extern uint64_t spa_first_txg(spa_t *spa);
extern uint64_t spa_syncing_txg(spa_t *spa);
extern uint64_t spa_final_dirty_txg(spa_t *spa);
extern uint64_t spa_version(spa_t *spa);
extern pool_state_t spa_state(spa_t *spa);
extern spa_load_state_t spa_load_state(spa_t *spa);
extern uint64_t spa_freeze_txg(spa_t *spa);
extern uint64_t spa_get_worst_case_asize(spa_t *spa, uint64_t lsize);
extern uint64_t spa_get_dspace(spa_t *spa);
extern uint64_t spa_get_checkpoint_space(spa_t *spa);
extern uint64_t spa_get_slop_space(spa_t *spa);
extern void spa_update_dspace(spa_t *spa);
extern uint64_t spa_version(spa_t *spa);
extern boolean_t spa_deflate(spa_t *spa);
extern metaslab_class_t *spa_normal_class(spa_t *spa);
extern metaslab_class_t *spa_log_class(spa_t *spa);
extern metaslab_class_t *spa_embedded_log_class(spa_t *spa);
extern metaslab_class_t *spa_special_class(spa_t *spa);
extern metaslab_class_t *spa_dedup_class(spa_t *spa);
extern metaslab_class_t *spa_preferred_class(spa_t *spa, uint64_t size,
dmu_object_type_t objtype, uint_t level, uint_t special_smallblk);
extern void spa_evicting_os_register(spa_t *, objset_t *os);
extern void spa_evicting_os_deregister(spa_t *, objset_t *os);
extern void spa_evicting_os_wait(spa_t *spa);
extern int spa_max_replication(spa_t *spa);
extern int spa_prev_software_version(spa_t *spa);
extern uint64_t spa_get_failmode(spa_t *spa);
extern uint64_t spa_get_deadman_failmode(spa_t *spa);
extern void spa_set_deadman_failmode(spa_t *spa, const char *failmode);
extern boolean_t spa_suspended(spa_t *spa);
extern uint64_t spa_bootfs(spa_t *spa);
extern uint64_t spa_delegation(spa_t *spa);
extern objset_t *spa_meta_objset(spa_t *spa);
extern space_map_t *spa_syncing_log_sm(spa_t *spa);
extern uint64_t spa_deadman_synctime(spa_t *spa);
extern uint64_t spa_deadman_ziotime(spa_t *spa);
extern uint64_t spa_dirty_data(spa_t *spa);
extern spa_autotrim_t spa_get_autotrim(spa_t *spa);
/* Miscellaneous support routines */
extern void spa_load_failed(spa_t *spa, const char *fmt, ...)
__attribute__((format(printf, 2, 3)));
extern void spa_load_note(spa_t *spa, const char *fmt, ...)
__attribute__((format(printf, 2, 3)));
extern void spa_activate_mos_feature(spa_t *spa, const char *feature,
dmu_tx_t *tx);
extern void spa_deactivate_mos_feature(spa_t *spa, const char *feature);
extern spa_t *spa_by_guid(uint64_t pool_guid, uint64_t device_guid);
extern boolean_t spa_guid_exists(uint64_t pool_guid, uint64_t device_guid);
extern char *spa_strdup(const char *);
extern void spa_strfree(char *);
extern uint64_t spa_generate_guid(spa_t *spa);
extern void snprintf_blkptr(char *buf, size_t buflen, const blkptr_t *bp);
extern void spa_freeze(spa_t *spa);
extern int spa_change_guid(spa_t *spa);
extern void spa_upgrade(spa_t *spa, uint64_t version);
extern void spa_evict_all(void);
extern vdev_t *spa_lookup_by_guid(spa_t *spa, uint64_t guid,
boolean_t l2cache);
extern boolean_t spa_has_l2cache(spa_t *, uint64_t guid);
extern boolean_t spa_has_spare(spa_t *, uint64_t guid);
extern uint64_t dva_get_dsize_sync(spa_t *spa, const dva_t *dva);
extern uint64_t bp_get_dsize_sync(spa_t *spa, const blkptr_t *bp);
extern uint64_t bp_get_dsize(spa_t *spa, const blkptr_t *bp);
extern boolean_t spa_has_slogs(spa_t *spa);
extern boolean_t spa_is_root(spa_t *spa);
extern boolean_t spa_writeable(spa_t *spa);
extern boolean_t spa_has_pending_synctask(spa_t *spa);
extern int spa_maxblocksize(spa_t *spa);
extern int spa_maxdnodesize(spa_t *spa);
extern boolean_t spa_has_checkpoint(spa_t *spa);
extern boolean_t spa_importing_readonly_checkpoint(spa_t *spa);
extern boolean_t spa_suspend_async_destroy(spa_t *spa);
extern uint64_t spa_min_claim_txg(spa_t *spa);
extern boolean_t zfs_dva_valid(spa_t *spa, const dva_t *dva,
const blkptr_t *bp);
typedef void (*spa_remap_cb_t)(uint64_t vdev, uint64_t offset, uint64_t size,
void *arg);
extern boolean_t spa_remap_blkptr(spa_t *spa, blkptr_t *bp,
spa_remap_cb_t callback, void *arg);
extern uint64_t spa_get_last_removal_txg(spa_t *spa);
extern boolean_t spa_trust_config(spa_t *spa);
extern uint64_t spa_missing_tvds_allowed(spa_t *spa);
extern void spa_set_missing_tvds(spa_t *spa, uint64_t missing);
extern boolean_t spa_top_vdevs_spacemap_addressable(spa_t *spa);
extern uint64_t spa_total_metaslabs(spa_t *spa);
extern boolean_t spa_multihost(spa_t *spa);
extern uint32_t spa_get_hostid(spa_t *spa);
extern void spa_activate_allocation_classes(spa_t *, dmu_tx_t *);
extern boolean_t spa_livelist_delete_check(spa_t *spa);
extern spa_mode_t spa_mode(spa_t *spa);
extern uint64_t zfs_strtonum(const char *str, char **nptr);
extern char *spa_his_ievent_table[];
extern void spa_history_create_obj(spa_t *spa, dmu_tx_t *tx);
extern int spa_history_get(spa_t *spa, uint64_t *offset, uint64_t *len_read,
char *his_buf);
extern int spa_history_log(spa_t *spa, const char *his_buf);
extern int spa_history_log_nvl(spa_t *spa, nvlist_t *nvl);
extern void spa_history_log_version(spa_t *spa, const char *operation,
dmu_tx_t *tx);
extern void spa_history_log_internal(spa_t *spa, const char *operation,
dmu_tx_t *tx, const char *fmt, ...) __printflike(4, 5);
extern void spa_history_log_internal_ds(struct dsl_dataset *ds, const char *op,
dmu_tx_t *tx, const char *fmt, ...) __printflike(4, 5);
extern void spa_history_log_internal_dd(dsl_dir_t *dd, const char *operation,
dmu_tx_t *tx, const char *fmt, ...) __printflike(4, 5);
extern const char *spa_state_to_name(spa_t *spa);
/* error handling */
struct zbookmark_phys;
extern void spa_log_error(spa_t *spa, const zbookmark_phys_t *zb,
const uint64_t *birth);
extern void spa_remove_error(spa_t *spa, zbookmark_phys_t *zb);
extern int zfs_ereport_post(const char *clazz, spa_t *spa, vdev_t *vd,
const zbookmark_phys_t *zb, zio_t *zio, uint64_t state);
extern boolean_t zfs_ereport_is_valid(const char *clazz, spa_t *spa, vdev_t *vd,
zio_t *zio);
extern void zfs_ereport_taskq_fini(void);
extern void zfs_ereport_clear(spa_t *spa, vdev_t *vd);
extern nvlist_t *zfs_event_create(spa_t *spa, vdev_t *vd, const char *type,
const char *name, nvlist_t *aux);
extern void zfs_post_remove(spa_t *spa, vdev_t *vd);
extern void zfs_post_state_change(spa_t *spa, vdev_t *vd, uint64_t laststate);
extern void zfs_post_autoreplace(spa_t *spa, vdev_t *vd);
extern uint64_t spa_approx_errlog_size(spa_t *spa);
extern int spa_get_errlog(spa_t *spa, void *uaddr, uint64_t *count);
extern void spa_errlog_rotate(spa_t *spa);
extern void spa_errlog_drain(spa_t *spa);
extern void spa_errlog_sync(spa_t *spa, uint64_t txg);
extern void spa_get_errlists(spa_t *spa, avl_tree_t *last, avl_tree_t *scrub);
extern void spa_delete_dataset_errlog(spa_t *spa, uint64_t ds, dmu_tx_t *tx);
extern void spa_swap_errlog(spa_t *spa, uint64_t new_head_ds,
uint64_t old_head_ds, dmu_tx_t *tx);
extern void sync_error_list(spa_t *spa, avl_tree_t *t, uint64_t *obj,
dmu_tx_t *tx);
extern void spa_upgrade_errlog(spa_t *spa, dmu_tx_t *tx);
/* vdev cache */
extern void vdev_cache_stat_init(void);
extern void vdev_cache_stat_fini(void);
/* vdev mirror */
extern void vdev_mirror_stat_init(void);
extern void vdev_mirror_stat_fini(void);
/* Initialization and termination */
extern void spa_init(spa_mode_t mode);
extern void spa_fini(void);
extern void spa_boot_init(void);
/* properties */
extern int spa_prop_set(spa_t *spa, nvlist_t *nvp);
extern int spa_prop_get(spa_t *spa, nvlist_t **nvp);
extern void spa_prop_clear_bootfs(spa_t *spa, uint64_t obj, dmu_tx_t *tx);
extern void spa_configfile_set(spa_t *, nvlist_t *, boolean_t);
/* asynchronous event notification */
extern void spa_event_notify(spa_t *spa, vdev_t *vdev, nvlist_t *hist_nvl,
const char *name);
extern void zfs_ereport_zvol_post(const char *subclass, const char *name,
const char *device_name, const char *raw_name);
/* waiting for pool activities to complete */
extern int spa_wait(const char *pool, zpool_wait_activity_t activity,
boolean_t *waited);
extern int spa_wait_tag(const char *name, zpool_wait_activity_t activity,
uint64_t tag, boolean_t *waited);
extern void spa_notify_waiters(spa_t *spa);
extern void spa_wake_waiters(spa_t *spa);
extern void spa_import_os(spa_t *spa);
extern void spa_export_os(spa_t *spa);
extern void spa_activate_os(spa_t *spa);
extern void spa_deactivate_os(spa_t *spa);
/* module param call functions */
int param_set_deadman_ziotime(ZFS_MODULE_PARAM_ARGS);
int param_set_deadman_synctime(ZFS_MODULE_PARAM_ARGS);
int param_set_slop_shift(ZFS_MODULE_PARAM_ARGS);
int param_set_deadman_failmode(ZFS_MODULE_PARAM_ARGS);
#ifdef ZFS_DEBUG
#define dprintf_bp(bp, fmt, ...) do { \
if (zfs_flags & ZFS_DEBUG_DPRINTF) { \
char *__blkbuf = kmem_alloc(BP_SPRINTF_LEN, KM_SLEEP); \
snprintf_blkptr(__blkbuf, BP_SPRINTF_LEN, (bp)); \
dprintf(fmt " %s\n", __VA_ARGS__, __blkbuf); \
kmem_free(__blkbuf, BP_SPRINTF_LEN); \
} \
} while (0)
#else
#define dprintf_bp(bp, fmt, ...)
#endif
extern spa_mode_t spa_mode_global;
extern int zfs_deadman_enabled;
extern uint64_t zfs_deadman_synctime_ms;
extern uint64_t zfs_deadman_ziotime_ms;
extern uint64_t zfs_deadman_checktime_ms;
extern kmem_cache_t *zio_buf_cache[];
extern kmem_cache_t *zio_data_buf_cache[];
#ifdef __cplusplus
}
#endif
#endif /* _SYS_SPA_H */
diff --git a/sys/contrib/openzfs/include/sys/vdev_impl.h b/sys/contrib/openzfs/include/sys/vdev_impl.h
index 7cfffe3b4eed..ea3043c82a39 100644
--- a/sys/contrib/openzfs/include/sys/vdev_impl.h
+++ b/sys/contrib/openzfs/include/sys/vdev_impl.h
@@ -1,671 +1,672 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
*/
#ifndef _SYS_VDEV_IMPL_H
#define _SYS_VDEV_IMPL_H
#include <sys/avl.h>
#include <sys/bpobj.h>
#include <sys/dmu.h>
#include <sys/metaslab.h>
#include <sys/nvpair.h>
#include <sys/space_map.h>
#include <sys/vdev.h>
#include <sys/dkio.h>
#include <sys/uberblock_impl.h>
#include <sys/vdev_indirect_mapping.h>
#include <sys/vdev_indirect_births.h>
#include <sys/vdev_rebuild.h>
#include <sys/vdev_removal.h>
#include <sys/zfs_ratelimit.h>
#ifdef __cplusplus
extern "C" {
#endif
/*
* Virtual device descriptors.
*
* All storage pool operations go through the virtual device framework,
* which provides data replication and I/O scheduling.
*/
/*
* Forward declarations that lots of things need.
*/
typedef struct vdev_queue vdev_queue_t;
typedef struct vdev_cache vdev_cache_t;
typedef struct vdev_cache_entry vdev_cache_entry_t;
struct abd;
extern uint_t zfs_vdev_queue_depth_pct;
extern uint_t zfs_vdev_def_queue_depth;
extern uint_t zfs_vdev_async_write_max_active;
/*
* Virtual device operations
*/
typedef int vdev_init_func_t(spa_t *spa, nvlist_t *nv, void **tsd);
typedef void vdev_kobj_post_evt_func_t(vdev_t *vd);
typedef void vdev_fini_func_t(vdev_t *vd);
typedef int vdev_open_func_t(vdev_t *vd, uint64_t *size, uint64_t *max_size,
uint64_t *ashift, uint64_t *pshift);
typedef void vdev_close_func_t(vdev_t *vd);
typedef uint64_t vdev_asize_func_t(vdev_t *vd, uint64_t psize);
typedef uint64_t vdev_min_asize_func_t(vdev_t *vd);
typedef uint64_t vdev_min_alloc_func_t(vdev_t *vd);
typedef void vdev_io_start_func_t(zio_t *zio);
typedef void vdev_io_done_func_t(zio_t *zio);
typedef void vdev_state_change_func_t(vdev_t *vd, int, int);
typedef boolean_t vdev_need_resilver_func_t(vdev_t *vd, const dva_t *dva,
size_t psize, uint64_t phys_birth);
typedef void vdev_hold_func_t(vdev_t *vd);
typedef void vdev_rele_func_t(vdev_t *vd);
typedef void vdev_remap_cb_t(uint64_t inner_offset, vdev_t *vd,
uint64_t offset, uint64_t size, void *arg);
typedef void vdev_remap_func_t(vdev_t *vd, uint64_t offset, uint64_t size,
vdev_remap_cb_t callback, void *arg);
/*
* Given a target vdev, translates the logical range "in" to the physical
* range "res"
*/
typedef void vdev_xlation_func_t(vdev_t *cvd, const range_seg64_t *logical,
range_seg64_t *physical, range_seg64_t *remain);
typedef uint64_t vdev_rebuild_asize_func_t(vdev_t *vd, uint64_t start,
uint64_t size, uint64_t max_segment);
typedef void vdev_metaslab_init_func_t(vdev_t *vd, uint64_t *startp,
uint64_t *sizep);
typedef void vdev_config_generate_func_t(vdev_t *vd, nvlist_t *nv);
typedef uint64_t vdev_nparity_func_t(vdev_t *vd);
typedef uint64_t vdev_ndisks_func_t(vdev_t *vd);
typedef const struct vdev_ops {
vdev_init_func_t *vdev_op_init;
vdev_fini_func_t *vdev_op_fini;
vdev_open_func_t *vdev_op_open;
vdev_close_func_t *vdev_op_close;
vdev_asize_func_t *vdev_op_asize;
vdev_min_asize_func_t *vdev_op_min_asize;
vdev_min_alloc_func_t *vdev_op_min_alloc;
vdev_io_start_func_t *vdev_op_io_start;
vdev_io_done_func_t *vdev_op_io_done;
vdev_state_change_func_t *vdev_op_state_change;
vdev_need_resilver_func_t *vdev_op_need_resilver;
vdev_hold_func_t *vdev_op_hold;
vdev_rele_func_t *vdev_op_rele;
vdev_remap_func_t *vdev_op_remap;
vdev_xlation_func_t *vdev_op_xlate;
vdev_rebuild_asize_func_t *vdev_op_rebuild_asize;
vdev_metaslab_init_func_t *vdev_op_metaslab_init;
vdev_config_generate_func_t *vdev_op_config_generate;
vdev_nparity_func_t *vdev_op_nparity;
vdev_ndisks_func_t *vdev_op_ndisks;
vdev_kobj_post_evt_func_t *vdev_op_kobj_evt_post;
char vdev_op_type[16];
boolean_t vdev_op_leaf;
} vdev_ops_t;
/*
* Virtual device properties
*/
struct vdev_cache_entry {
struct abd *ve_abd;
uint64_t ve_offset;
clock_t ve_lastused;
avl_node_t ve_offset_node;
avl_node_t ve_lastused_node;
uint32_t ve_hits;
uint16_t ve_missed_update;
zio_t *ve_fill_io;
};
struct vdev_cache {
avl_tree_t vc_offset_tree;
avl_tree_t vc_lastused_tree;
kmutex_t vc_lock;
};
typedef struct vdev_queue_class {
uint32_t vqc_active;
/*
* Sorted by offset or timestamp, depending on if the queue is
* LBA-ordered vs FIFO.
*/
avl_tree_t vqc_queued_tree;
} vdev_queue_class_t;
struct vdev_queue {
vdev_t *vq_vdev;
vdev_queue_class_t vq_class[ZIO_PRIORITY_NUM_QUEUEABLE];
avl_tree_t vq_active_tree;
avl_tree_t vq_read_offset_tree;
avl_tree_t vq_write_offset_tree;
avl_tree_t vq_trim_offset_tree;
uint64_t vq_last_offset;
zio_priority_t vq_last_prio; /* Last sent I/O priority. */
uint32_t vq_ia_active; /* Active interactive I/Os. */
uint32_t vq_nia_credit; /* Non-interactive I/Os credit. */
hrtime_t vq_io_complete_ts; /* time last i/o completed */
hrtime_t vq_io_delta_ts;
zio_t vq_io_search; /* used as local for stack reduction */
kmutex_t vq_lock;
};
typedef enum vdev_alloc_bias {
VDEV_BIAS_NONE,
VDEV_BIAS_LOG, /* dedicated to ZIL data (SLOG) */
VDEV_BIAS_SPECIAL, /* dedicated to ddt, metadata, and small blks */
VDEV_BIAS_DEDUP /* dedicated to dedup metadata */
} vdev_alloc_bias_t;
/*
* On-disk indirect vdev state.
*
* An indirect vdev is described exclusively in the MOS config of a pool.
* The config for an indirect vdev includes several fields, which are
* accessed in memory by a vdev_indirect_config_t.
*/
typedef struct vdev_indirect_config {
/*
* Object (in MOS) which contains the indirect mapping. This object
* contains an array of vdev_indirect_mapping_entry_phys_t ordered by
* vimep_src. The bonus buffer for this object is a
* vdev_indirect_mapping_phys_t. This object is allocated when a vdev
* removal is initiated.
*
* Note that this object can be empty if none of the data on the vdev
* has been copied yet.
*/
uint64_t vic_mapping_object;
/*
* Object (in MOS) which contains the birth times for the mapping
* entries. This object contains an array of
* vdev_indirect_birth_entry_phys_t sorted by vibe_offset. The bonus
* buffer for this object is a vdev_indirect_birth_phys_t. This object
* is allocated when a vdev removal is initiated.
*
* Note that this object can be empty if none of the vdev has yet been
* copied.
*/
uint64_t vic_births_object;
/*
* This is the vdev ID which was removed previous to this vdev, or
* UINT64_MAX if there are no previously removed vdevs.
*/
uint64_t vic_prev_indirect_vdev;
} vdev_indirect_config_t;
/*
* Virtual device descriptor
*/
struct vdev {
/*
* Common to all vdev types.
*/
uint64_t vdev_id; /* child number in vdev parent */
uint64_t vdev_guid; /* unique ID for this vdev */
uint64_t vdev_guid_sum; /* self guid + all child guids */
uint64_t vdev_orig_guid; /* orig. guid prior to remove */
uint64_t vdev_asize; /* allocatable device capacity */
uint64_t vdev_min_asize; /* min acceptable asize */
uint64_t vdev_max_asize; /* max acceptable asize */
uint64_t vdev_ashift; /* block alignment shift */
/*
* Logical block alignment shift
*
* The smallest sized/aligned I/O supported by the device.
*/
uint64_t vdev_logical_ashift;
/*
* Physical block alignment shift
*
* The device supports logical I/Os with vdev_logical_ashift
* size/alignment, but optimum performance will be achieved by
* aligning/sizing requests to vdev_physical_ashift. Smaller
* requests may be inflated or incur device level read-modify-write
* operations.
*
* May be 0 to indicate no preference (i.e. use vdev_logical_ashift).
*/
uint64_t vdev_physical_ashift;
uint64_t vdev_state; /* see VDEV_STATE_* #defines */
uint64_t vdev_prevstate; /* used when reopening a vdev */
vdev_ops_t *vdev_ops; /* vdev operations */
spa_t *vdev_spa; /* spa for this vdev */
void *vdev_tsd; /* type-specific data */
vdev_t *vdev_top; /* top-level vdev */
vdev_t *vdev_parent; /* parent vdev */
vdev_t **vdev_child; /* array of children */
uint64_t vdev_children; /* number of children */
vdev_stat_t vdev_stat; /* virtual device statistics */
vdev_stat_ex_t vdev_stat_ex; /* extended statistics */
boolean_t vdev_expanding; /* expand the vdev? */
boolean_t vdev_reopening; /* reopen in progress? */
boolean_t vdev_nonrot; /* true if solid state */
int vdev_load_error; /* error on last load */
int vdev_open_error; /* error on last open */
int vdev_validate_error; /* error on last validate */
kthread_t *vdev_open_thread; /* thread opening children */
kthread_t *vdev_validate_thread; /* thread validating children */
uint64_t vdev_crtxg; /* txg when top-level was added */
+ uint64_t vdev_root_zap;
/*
* Top-level vdev state.
*/
uint64_t vdev_ms_array; /* metaslab array object */
uint64_t vdev_ms_shift; /* metaslab size shift */
uint64_t vdev_ms_count; /* number of metaslabs */
metaslab_group_t *vdev_mg; /* metaslab group */
metaslab_group_t *vdev_log_mg; /* embedded slog metaslab group */
metaslab_t **vdev_ms; /* metaslab array */
uint64_t vdev_pending_fastwrite; /* allocated fastwrites */
txg_list_t vdev_ms_list; /* per-txg dirty metaslab lists */
txg_list_t vdev_dtl_list; /* per-txg dirty DTL lists */
txg_node_t vdev_txg_node; /* per-txg dirty vdev linkage */
boolean_t vdev_remove_wanted; /* async remove wanted? */
boolean_t vdev_probe_wanted; /* async probe wanted? */
list_node_t vdev_config_dirty_node; /* config dirty list */
list_node_t vdev_state_dirty_node; /* state dirty list */
uint64_t vdev_deflate_ratio; /* deflation ratio (x512) */
uint64_t vdev_islog; /* is an intent log device */
uint64_t vdev_noalloc; /* device is passivated? */
uint64_t vdev_removing; /* device is being removed? */
uint64_t vdev_failfast; /* device failfast setting */
boolean_t vdev_ishole; /* is a hole in the namespace */
uint64_t vdev_top_zap;
vdev_alloc_bias_t vdev_alloc_bias; /* metaslab allocation bias */
/* pool checkpoint related */
space_map_t *vdev_checkpoint_sm; /* contains reserved blocks */
/* Initialize related */
boolean_t vdev_initialize_exit_wanted;
vdev_initializing_state_t vdev_initialize_state;
list_node_t vdev_initialize_node;
kthread_t *vdev_initialize_thread;
/* Protects vdev_initialize_thread and vdev_initialize_state. */
kmutex_t vdev_initialize_lock;
kcondvar_t vdev_initialize_cv;
uint64_t vdev_initialize_offset[TXG_SIZE];
uint64_t vdev_initialize_last_offset;
range_tree_t *vdev_initialize_tree; /* valid while initializing */
uint64_t vdev_initialize_bytes_est;
uint64_t vdev_initialize_bytes_done;
uint64_t vdev_initialize_action_time; /* start and end time */
/* TRIM related */
boolean_t vdev_trim_exit_wanted;
boolean_t vdev_autotrim_exit_wanted;
vdev_trim_state_t vdev_trim_state;
list_node_t vdev_trim_node;
kmutex_t vdev_autotrim_lock;
kcondvar_t vdev_autotrim_cv;
kcondvar_t vdev_autotrim_kick_cv;
kthread_t *vdev_autotrim_thread;
/* Protects vdev_trim_thread and vdev_trim_state. */
kmutex_t vdev_trim_lock;
kcondvar_t vdev_trim_cv;
kthread_t *vdev_trim_thread;
uint64_t vdev_trim_offset[TXG_SIZE];
uint64_t vdev_trim_last_offset;
uint64_t vdev_trim_bytes_est;
uint64_t vdev_trim_bytes_done;
uint64_t vdev_trim_rate; /* requested rate (bytes/sec) */
uint64_t vdev_trim_partial; /* requested partial TRIM */
uint64_t vdev_trim_secure; /* requested secure TRIM */
uint64_t vdev_trim_action_time; /* start and end time */
/* Rebuild related */
boolean_t vdev_rebuilding;
boolean_t vdev_rebuild_exit_wanted;
boolean_t vdev_rebuild_cancel_wanted;
boolean_t vdev_rebuild_reset_wanted;
kmutex_t vdev_rebuild_lock;
kcondvar_t vdev_rebuild_cv;
kthread_t *vdev_rebuild_thread;
vdev_rebuild_t vdev_rebuild_config;
/* For limiting outstanding I/Os (initialize, TRIM) */
kmutex_t vdev_initialize_io_lock;
kcondvar_t vdev_initialize_io_cv;
uint64_t vdev_initialize_inflight;
kmutex_t vdev_trim_io_lock;
kcondvar_t vdev_trim_io_cv;
uint64_t vdev_trim_inflight[3];
/*
* Values stored in the config for an indirect or removing vdev.
*/
vdev_indirect_config_t vdev_indirect_config;
/*
* The vdev_indirect_rwlock protects the vdev_indirect_mapping
* pointer from changing on indirect vdevs (when it is condensed).
* Note that removing (not yet indirect) vdevs have different
* access patterns (the mapping is not accessed from open context,
* e.g. from zio_read) and locking strategy (e.g. svr_lock).
*/
krwlock_t vdev_indirect_rwlock;
vdev_indirect_mapping_t *vdev_indirect_mapping;
vdev_indirect_births_t *vdev_indirect_births;
/*
* In memory data structures used to manage the obsolete sm, for
* indirect or removing vdevs.
*
* The vdev_obsolete_segments is the in-core record of the segments
* that are no longer referenced anywhere in the pool (due to
* being freed or remapped and not referenced by any snapshots).
* During a sync, segments are added to vdev_obsolete_segments
* via vdev_indirect_mark_obsolete(); at the end of each sync
* pass, this is appended to vdev_obsolete_sm via
* vdev_indirect_sync_obsolete(). The vdev_obsolete_lock
* protects against concurrent modifications of vdev_obsolete_segments
* from multiple zio threads.
*/
kmutex_t vdev_obsolete_lock;
range_tree_t *vdev_obsolete_segments;
space_map_t *vdev_obsolete_sm;
/*
* Protects the vdev_scan_io_queue field itself as well as the
* structure's contents (when present).
*/
kmutex_t vdev_scan_io_queue_lock;
struct dsl_scan_io_queue *vdev_scan_io_queue;
/*
* Leaf vdev state.
*/
range_tree_t *vdev_dtl[DTL_TYPES]; /* dirty time logs */
space_map_t *vdev_dtl_sm; /* dirty time log space map */
txg_node_t vdev_dtl_node; /* per-txg dirty DTL linkage */
uint64_t vdev_dtl_object; /* DTL object */
uint64_t vdev_psize; /* physical device capacity */
uint64_t vdev_wholedisk; /* true if this is a whole disk */
uint64_t vdev_offline; /* persistent offline state */
uint64_t vdev_faulted; /* persistent faulted state */
uint64_t vdev_degraded; /* persistent degraded state */
uint64_t vdev_removed; /* persistent removed state */
uint64_t vdev_resilver_txg; /* persistent resilvering state */
uint64_t vdev_rebuild_txg; /* persistent rebuilding state */
char *vdev_path; /* vdev path (if any) */
char *vdev_devid; /* vdev devid (if any) */
char *vdev_physpath; /* vdev device path (if any) */
char *vdev_enc_sysfs_path; /* enclosure sysfs path */
char *vdev_fru; /* physical FRU location */
uint64_t vdev_not_present; /* not present during import */
uint64_t vdev_unspare; /* unspare when resilvering done */
boolean_t vdev_nowritecache; /* true if flushwritecache failed */
boolean_t vdev_has_trim; /* TRIM is supported */
boolean_t vdev_has_securetrim; /* secure TRIM is supported */
boolean_t vdev_checkremove; /* temporary online test */
boolean_t vdev_forcefault; /* force online fault */
boolean_t vdev_splitting; /* split or repair in progress */
boolean_t vdev_delayed_close; /* delayed device close? */
boolean_t vdev_tmpoffline; /* device taken offline temporarily? */
boolean_t vdev_detached; /* device detached? */
boolean_t vdev_cant_read; /* vdev is failing all reads */
boolean_t vdev_cant_write; /* vdev is failing all writes */
boolean_t vdev_isspare; /* was a hot spare */
boolean_t vdev_isl2cache; /* was a l2cache device */
boolean_t vdev_copy_uberblocks; /* post expand copy uberblocks */
boolean_t vdev_resilver_deferred; /* resilver deferred */
boolean_t vdev_kobj_flag; /* kobj event record */
vdev_queue_t vdev_queue; /* I/O deadline schedule queue */
vdev_cache_t vdev_cache; /* physical block cache */
spa_aux_vdev_t *vdev_aux; /* for l2cache and spares vdevs */
zio_t *vdev_probe_zio; /* root of current probe */
vdev_aux_t vdev_label_aux; /* on-disk aux state */
uint64_t vdev_leaf_zap;
hrtime_t vdev_mmp_pending; /* 0 if write finished */
uint64_t vdev_mmp_kstat_id; /* to find kstat entry */
uint64_t vdev_expansion_time; /* vdev's last expansion time */
list_node_t vdev_leaf_node; /* leaf vdev list */
/*
* For DTrace to work in userland (libzpool) context, these fields must
* remain at the end of the structure. DTrace will use the kernel's
* CTF definition for 'struct vdev', and since the size of a kmutex_t is
* larger in userland, the offsets for the rest of the fields would be
* incorrect.
*/
kmutex_t vdev_dtl_lock; /* vdev_dtl_{map,resilver} */
kmutex_t vdev_stat_lock; /* vdev_stat */
kmutex_t vdev_probe_lock; /* protects vdev_probe_zio */
/*
* We rate limit ZIO delay, deadman, and checksum events, since they
* can flood ZED with tons of events when a drive is acting up.
*/
zfs_ratelimit_t vdev_delay_rl;
zfs_ratelimit_t vdev_deadman_rl;
zfs_ratelimit_t vdev_checksum_rl;
/*
* Checksum and IO thresholds for tuning ZED
*/
uint64_t vdev_checksum_n;
uint64_t vdev_checksum_t;
uint64_t vdev_io_n;
uint64_t vdev_io_t;
};
#define VDEV_PAD_SIZE (8 << 10)
/* 2 padding areas (vl_pad1 and vl_be) to skip */
#define VDEV_SKIP_SIZE VDEV_PAD_SIZE * 2
#define VDEV_PHYS_SIZE (112 << 10)
#define VDEV_UBERBLOCK_RING (128 << 10)
/*
* MMP blocks occupy the last MMP_BLOCKS_PER_LABEL slots in the uberblock
* ring when MMP is enabled.
*/
#define MMP_BLOCKS_PER_LABEL 1
/* The largest uberblock we support is 8k. */
#define MAX_UBERBLOCK_SHIFT (13)
#define VDEV_UBERBLOCK_SHIFT(vd) \
MIN(MAX((vd)->vdev_top->vdev_ashift, UBERBLOCK_SHIFT), \
MAX_UBERBLOCK_SHIFT)
#define VDEV_UBERBLOCK_COUNT(vd) \
(VDEV_UBERBLOCK_RING >> VDEV_UBERBLOCK_SHIFT(vd))
#define VDEV_UBERBLOCK_OFFSET(vd, n) \
offsetof(vdev_label_t, vl_uberblock[(n) << VDEV_UBERBLOCK_SHIFT(vd)])
#define VDEV_UBERBLOCK_SIZE(vd) (1ULL << VDEV_UBERBLOCK_SHIFT(vd))
typedef struct vdev_phys {
char vp_nvlist[VDEV_PHYS_SIZE - sizeof (zio_eck_t)];
zio_eck_t vp_zbt;
} vdev_phys_t;
typedef enum vbe_vers {
/*
* The bootenv file is stored as ascii text in the envblock.
* It is used by the GRUB bootloader used on Linux to store the
* contents of the grubenv file. The file is stored as raw ASCII,
* and is protected by an embedded checksum. By default, GRUB will
* check if the boot filesystem supports storing the environment data
* in a special location, and if so, will invoke filesystem specific
* logic to retrieve it. This can be overridden by a variable, should
* the user so desire.
*/
VB_RAW = 0,
/*
* The bootenv file is converted to an nvlist and then packed into the
* envblock.
*/
VB_NVLIST = 1
} vbe_vers_t;
typedef struct vdev_boot_envblock {
uint64_t vbe_version;
char vbe_bootenv[VDEV_PAD_SIZE - sizeof (uint64_t) -
sizeof (zio_eck_t)];
zio_eck_t vbe_zbt;
} vdev_boot_envblock_t;
_Static_assert(sizeof (vdev_boot_envblock_t) == VDEV_PAD_SIZE,
"vdev_boot_envblock_t wrong size");
typedef struct vdev_label {
char vl_pad1[VDEV_PAD_SIZE]; /* 8K */
vdev_boot_envblock_t vl_be; /* 8K */
vdev_phys_t vl_vdev_phys; /* 112K */
char vl_uberblock[VDEV_UBERBLOCK_RING]; /* 128K */
} vdev_label_t; /* 256K total */
/*
* vdev_dirty() flags
*/
#define VDD_METASLAB 0x01
#define VDD_DTL 0x02
/* Offset of embedded boot loader region on each label */
#define VDEV_BOOT_OFFSET (2 * sizeof (vdev_label_t))
/*
* Size of embedded boot loader region on each label.
* The total size of the first two labels plus the boot area is 4MB.
*/
#define VDEV_BOOT_SIZE (7ULL << 19) /* 3.5M */
/*
* Size of label regions at the start and end of each leaf device.
*/
#define VDEV_LABEL_START_SIZE (2 * sizeof (vdev_label_t) + VDEV_BOOT_SIZE)
#define VDEV_LABEL_END_SIZE (2 * sizeof (vdev_label_t))
#define VDEV_LABELS 4
#define VDEV_BEST_LABEL VDEV_LABELS
#define VDEV_OFFSET_IS_LABEL(vd, off) \
(((off) < VDEV_LABEL_START_SIZE) || \
((off) >= ((vd)->vdev_psize - VDEV_LABEL_END_SIZE)))
#define VDEV_ALLOC_LOAD 0
#define VDEV_ALLOC_ADD 1
#define VDEV_ALLOC_SPARE 2
#define VDEV_ALLOC_L2CACHE 3
#define VDEV_ALLOC_ROOTPOOL 4
#define VDEV_ALLOC_SPLIT 5
#define VDEV_ALLOC_ATTACH 6
/*
* Allocate or free a vdev
*/
extern vdev_t *vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid,
vdev_ops_t *ops);
extern int vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *config,
vdev_t *parent, uint_t id, int alloctype);
extern void vdev_free(vdev_t *vd);
/*
* Add or remove children and parents
*/
extern void vdev_add_child(vdev_t *pvd, vdev_t *cvd);
extern void vdev_remove_child(vdev_t *pvd, vdev_t *cvd);
extern void vdev_compact_children(vdev_t *pvd);
extern vdev_t *vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops);
extern void vdev_remove_parent(vdev_t *cvd);
/*
* vdev sync load and sync
*/
extern boolean_t vdev_log_state_valid(vdev_t *vd);
extern int vdev_load(vdev_t *vd);
extern int vdev_dtl_load(vdev_t *vd);
extern void vdev_sync(vdev_t *vd, uint64_t txg);
extern void vdev_sync_done(vdev_t *vd, uint64_t txg);
extern void vdev_dirty(vdev_t *vd, int flags, void *arg, uint64_t txg);
extern void vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg);
/*
* Available vdev types.
*/
extern vdev_ops_t vdev_root_ops;
extern vdev_ops_t vdev_mirror_ops;
extern vdev_ops_t vdev_replacing_ops;
extern vdev_ops_t vdev_raidz_ops;
extern vdev_ops_t vdev_draid_ops;
extern vdev_ops_t vdev_draid_spare_ops;
extern vdev_ops_t vdev_disk_ops;
extern vdev_ops_t vdev_file_ops;
extern vdev_ops_t vdev_missing_ops;
extern vdev_ops_t vdev_hole_ops;
extern vdev_ops_t vdev_spare_ops;
extern vdev_ops_t vdev_indirect_ops;
/*
* Common size functions
*/
extern void vdev_default_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs);
extern uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize);
extern uint64_t vdev_default_min_asize(vdev_t *vd);
extern uint64_t vdev_get_min_asize(vdev_t *vd);
extern void vdev_set_min_asize(vdev_t *vd);
extern uint64_t vdev_get_min_alloc(vdev_t *vd);
extern uint64_t vdev_get_nparity(vdev_t *vd);
extern uint64_t vdev_get_ndisks(vdev_t *vd);
/*
* Global variables
*/
extern int zfs_vdev_standard_sm_blksz;
/*
* Functions from vdev_indirect.c
*/
extern void vdev_indirect_sync_obsolete(vdev_t *vd, dmu_tx_t *tx);
extern boolean_t vdev_indirect_should_condense(vdev_t *vd);
extern void spa_condense_indirect_start_sync(vdev_t *vd, dmu_tx_t *tx);
extern int vdev_obsolete_sm_object(vdev_t *vd, uint64_t *sm_obj);
extern int vdev_obsolete_counts_are_precise(vdev_t *vd, boolean_t *are_precise);
/*
* Other miscellaneous functions
*/
int vdev_checkpoint_sm_object(vdev_t *vd, uint64_t *sm_obj);
void vdev_metaslab_group_create(vdev_t *vd);
uint64_t vdev_best_ashift(uint64_t logical, uint64_t a, uint64_t b);
/*
* Vdev ashift optimization tunables
*/
extern uint_t zfs_vdev_min_auto_ashift;
extern uint_t zfs_vdev_max_auto_ashift;
int param_set_min_auto_ashift(ZFS_MODULE_PARAM_ARGS);
int param_set_max_auto_ashift(ZFS_MODULE_PARAM_ARGS);
#ifdef __cplusplus
}
#endif
#endif /* _SYS_VDEV_IMPL_H */
diff --git a/sys/contrib/openzfs/include/sys/zfs_acl.h b/sys/contrib/openzfs/include/sys/zfs_acl.h
index e5c570c474a3..e19288528849 100644
--- a/sys/contrib/openzfs/include/sys/zfs_acl.h
+++ b/sys/contrib/openzfs/include/sys/zfs_acl.h
@@ -1,249 +1,249 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
*/
#ifndef _SYS_FS_ZFS_ACL_H
#define _SYS_FS_ZFS_ACL_H
#ifdef _KERNEL
#include <sys/isa_defs.h>
#include <sys/types32.h>
#include <sys/xvattr.h>
#endif
#include <sys/acl.h>
#include <sys/dmu.h>
#include <sys/zfs_fuid.h>
#include <sys/sa.h>
#ifdef __cplusplus
extern "C" {
#endif
struct znode_phys;
#define ACE_SLOT_CNT 6
#define ZFS_ACL_VERSION_INITIAL 0ULL
#define ZFS_ACL_VERSION_FUID 1ULL
#define ZFS_ACL_VERSION ZFS_ACL_VERSION_FUID
/*
* ZFS ACLs (Access Control Lists) are stored in various forms.
*
* Files created with ACL version ZFS_ACL_VERSION_INITIAL
* will all be created with fixed length ACEs of type
* zfs_oldace_t.
*
* Files with ACL version ZFS_ACL_VERSION_FUID will be created
* with various sized ACEs. The abstraction entries will utilize
* zfs_ace_hdr_t, normal user/group entries will use zfs_ace_t
* and some specialized CIFS ACEs will use zfs_object_ace_t.
*/
/*
* All ACEs have a common hdr. For
* owner@, group@, and everyone@ this is all
* that's needed.
*/
typedef struct zfs_ace_hdr {
uint16_t z_type;
uint16_t z_flags;
uint32_t z_access_mask;
} zfs_ace_hdr_t;
typedef zfs_ace_hdr_t zfs_ace_abstract_t;
/*
* Standard ACE
*/
typedef struct zfs_ace {
zfs_ace_hdr_t z_hdr;
uint64_t z_fuid;
} zfs_ace_t;
/*
* The following type only applies to ACE_ACCESS_ALLOWED|DENIED_OBJECT_ACE_TYPE
* and will only be set/retrieved in a CIFS context.
*/
typedef struct zfs_object_ace {
zfs_ace_t z_ace;
uint8_t z_object_type[16]; /* object type */
uint8_t z_inherit_type[16]; /* inherited object type */
} zfs_object_ace_t;
typedef struct zfs_oldace {
uint32_t z_fuid; /* "who" */
uint32_t z_access_mask; /* access mask */
uint16_t z_flags; /* flags, i.e inheritance */
uint16_t z_type; /* type of entry allow/deny */
} zfs_oldace_t;
typedef struct zfs_acl_phys_v0 {
uint64_t z_acl_extern_obj; /* ext acl pieces */
uint32_t z_acl_count; /* Number of ACEs */
uint16_t z_acl_version; /* acl version */
uint16_t z_acl_pad; /* pad */
zfs_oldace_t z_ace_data[ACE_SLOT_CNT]; /* 6 standard ACEs */
} zfs_acl_phys_v0_t;
#define ZFS_ACE_SPACE (sizeof (zfs_oldace_t) * ACE_SLOT_CNT)
/*
* Size of ACL count is always 2 bytes.
* Necessary to for dealing with both V0 ACL and V1 ACL layout
*/
#define ZFS_ACL_COUNT_SIZE (sizeof (uint16_t))
typedef struct zfs_acl_phys {
uint64_t z_acl_extern_obj; /* ext acl pieces */
uint32_t z_acl_size; /* Number of bytes in ACL */
uint16_t z_acl_version; /* acl version */
uint16_t z_acl_count; /* ace count */
uint8_t z_ace_data[ZFS_ACE_SPACE]; /* space for embedded ACEs */
} zfs_acl_phys_t;
typedef struct acl_ops {
uint32_t (*ace_mask_get) (void *acep); /* get access mask */
void (*ace_mask_set) (void *acep,
uint32_t mask); /* set access mask */
uint16_t (*ace_flags_get) (void *acep); /* get flags */
void (*ace_flags_set) (void *acep,
uint16_t flags); /* set flags */
uint16_t (*ace_type_get)(void *acep); /* get type */
void (*ace_type_set)(void *acep,
uint16_t type); /* set type */
uint64_t (*ace_who_get)(void *acep); /* get who/fuid */
void (*ace_who_set)(void *acep,
uint64_t who); /* set who/fuid */
size_t (*ace_size)(void *acep); /* how big is this ace */
size_t (*ace_abstract_size)(void); /* sizeof abstract entry */
int (*ace_mask_off)(void); /* off of access mask in ace */
/* ptr to data if any */
int (*ace_data)(void *acep, void **datap);
} acl_ops_t;
/*
* A zfs_acl_t structure is composed of a list of zfs_acl_node_t's.
* Each node will have one or more ACEs associated with it. You will
* only have multiple nodes during a chmod operation. Normally only
* one node is required.
*/
typedef struct zfs_acl_node {
list_node_t z_next; /* Next chunk of ACEs */
void *z_acldata; /* pointer into actual ACE(s) */
void *z_allocdata; /* pointer to kmem allocated memory */
size_t z_allocsize; /* Size of blob in bytes */
size_t z_size; /* length of ACL data */
uint64_t z_ace_count; /* number of ACEs in this acl node */
int z_ace_idx; /* ace iterator positioned on */
} zfs_acl_node_t;
typedef struct zfs_acl {
uint64_t z_acl_count; /* Number of ACEs */
size_t z_acl_bytes; /* Number of bytes in ACL */
uint_t z_version; /* version of ACL */
void *z_next_ace; /* pointer to next ACE */
uint64_t z_hints; /* ACL hints (ZFS_INHERIT_ACE ...) */
zfs_acl_node_t *z_curr_node; /* current node iterator is handling */
list_t z_acl; /* chunks of ACE data */
const acl_ops_t *z_ops; /* ACL operations */
} zfs_acl_t;
typedef struct acl_locator_cb {
zfs_acl_t *cb_aclp;
zfs_acl_node_t *cb_acl_node;
} zfs_acl_locator_cb_t;
#define ACL_DATA_ALLOCED 0x1
#define ZFS_ACL_SIZE(aclcnt) (sizeof (ace_t) * (aclcnt))
struct zfs_fuid_info;
typedef struct zfs_acl_ids {
uint64_t z_fuid; /* file owner fuid */
uint64_t z_fgid; /* file group owner fuid */
uint64_t z_mode; /* mode to set on create */
zfs_acl_t *z_aclp; /* ACL to create with file */
struct zfs_fuid_info *z_fuidp; /* for tracking fuids for log */
} zfs_acl_ids_t;
/*
* Property values for acl_mode and acl_inherit.
*
* acl_mode can take discard, noallow, groupmask and passthrough.
* whereas acl_inherit has secure instead of groupmask.
*/
#define ZFS_ACL_DISCARD 0
#define ZFS_ACL_NOALLOW 1
#define ZFS_ACL_GROUPMASK 2
#define ZFS_ACL_PASSTHROUGH 3
#define ZFS_ACL_RESTRICTED 4
#define ZFS_ACL_PASSTHROUGH_X 5
struct znode;
struct zfsvfs;
#ifdef _KERNEL
int zfs_acl_ids_create(struct znode *, int, vattr_t *,
- cred_t *, vsecattr_t *, zfs_acl_ids_t *, zuserns_t *);
+ cred_t *, vsecattr_t *, zfs_acl_ids_t *, zidmap_t *);
void zfs_acl_ids_free(zfs_acl_ids_t *);
boolean_t zfs_acl_ids_overquota(struct zfsvfs *, zfs_acl_ids_t *, uint64_t);
int zfs_getacl(struct znode *, vsecattr_t *, boolean_t, cred_t *);
int zfs_setacl(struct znode *, vsecattr_t *, boolean_t, cred_t *);
void zfs_acl_rele(void *);
void zfs_oldace_byteswap(ace_t *, int);
void zfs_ace_byteswap(void *, size_t, boolean_t);
extern boolean_t zfs_has_access(struct znode *zp, cred_t *cr);
extern int zfs_zaccess(struct znode *, int, int, boolean_t, cred_t *,
- zuserns_t *);
+ zidmap_t *);
int zfs_fastaccesschk_execute(struct znode *, cred_t *);
-extern int zfs_zaccess_rwx(struct znode *, mode_t, int, cred_t *, zuserns_t *);
+extern int zfs_zaccess_rwx(struct znode *, mode_t, int, cred_t *, zidmap_t *);
extern int zfs_zaccess_unix(void *, int, cred_t *);
extern int zfs_acl_access(struct znode *, int, cred_t *);
int zfs_acl_chmod_setattr(struct znode *, zfs_acl_t **, uint64_t);
-int zfs_zaccess_delete(struct znode *, struct znode *, cred_t *, zuserns_t *);
+int zfs_zaccess_delete(struct znode *, struct znode *, cred_t *, zidmap_t *);
int zfs_zaccess_rename(struct znode *, struct znode *,
- struct znode *, struct znode *, cred_t *cr, zuserns_t *mnt_ns);
+ struct znode *, struct znode *, cred_t *cr, zidmap_t *mnt_ns);
void zfs_acl_free(zfs_acl_t *);
int zfs_vsec_2_aclp(struct zfsvfs *, umode_t, vsecattr_t *, cred_t *,
struct zfs_fuid_info **, zfs_acl_t **);
int zfs_aclset_common(struct znode *, zfs_acl_t *, cred_t *, dmu_tx_t *);
uint64_t zfs_external_acl(struct znode *);
int zfs_znode_acl_version(struct znode *);
int zfs_acl_size(struct znode *, int *);
zfs_acl_t *zfs_acl_alloc(int);
zfs_acl_node_t *zfs_acl_node_alloc(size_t);
void zfs_acl_xform(struct znode *, zfs_acl_t *, cred_t *);
void zfs_acl_data_locator(void **, uint32_t *, uint32_t, boolean_t, void *);
uint64_t zfs_mode_compute(uint64_t, zfs_acl_t *,
uint64_t *, uint64_t, uint64_t);
int zfs_acl_node_read(struct znode *, boolean_t, zfs_acl_t **, boolean_t);
int zfs_acl_chown_setattr(struct znode *);
#endif
#ifdef __cplusplus
}
#endif
#endif /* _SYS_FS_ZFS_ACL_H */
diff --git a/sys/contrib/openzfs/include/zfeature_common.h b/sys/contrib/openzfs/include/zfeature_common.h
index ef915a70952e..7066c699e203 100644
--- a/sys/contrib/openzfs/include/zfeature_common.h
+++ b/sys/contrib/openzfs/include/zfeature_common.h
@@ -1,139 +1,140 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2011, 2018 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
* Copyright (c) 2013, Joyent, Inc. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
*/
#ifndef _ZFEATURE_COMMON_H
#define _ZFEATURE_COMMON_H extern __attribute__((visibility("default")))
#include <sys/fs/zfs.h>
#include <sys/inttypes.h>
#include <sys/types.h>
#ifdef __cplusplus
extern "C" {
#endif
struct zfeature_info;
typedef enum spa_feature {
SPA_FEATURE_NONE = -1,
SPA_FEATURE_ASYNC_DESTROY,
SPA_FEATURE_EMPTY_BPOBJ,
SPA_FEATURE_LZ4_COMPRESS,
SPA_FEATURE_MULTI_VDEV_CRASH_DUMP,
SPA_FEATURE_SPACEMAP_HISTOGRAM,
SPA_FEATURE_ENABLED_TXG,
SPA_FEATURE_HOLE_BIRTH,
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_EMBEDDED_DATA,
SPA_FEATURE_BOOKMARKS,
SPA_FEATURE_FS_SS_LIMIT,
SPA_FEATURE_LARGE_BLOCKS,
SPA_FEATURE_LARGE_DNODE,
SPA_FEATURE_SHA512,
SPA_FEATURE_SKEIN,
SPA_FEATURE_EDONR,
SPA_FEATURE_USEROBJ_ACCOUNTING,
SPA_FEATURE_ENCRYPTION,
SPA_FEATURE_PROJECT_QUOTA,
SPA_FEATURE_DEVICE_REMOVAL,
SPA_FEATURE_OBSOLETE_COUNTS,
SPA_FEATURE_POOL_CHECKPOINT,
SPA_FEATURE_SPACEMAP_V2,
SPA_FEATURE_ALLOCATION_CLASSES,
SPA_FEATURE_RESILVER_DEFER,
SPA_FEATURE_BOOKMARK_V2,
SPA_FEATURE_REDACTION_BOOKMARKS,
SPA_FEATURE_REDACTED_DATASETS,
SPA_FEATURE_BOOKMARK_WRITTEN,
SPA_FEATURE_LOG_SPACEMAP,
SPA_FEATURE_LIVELIST,
SPA_FEATURE_DEVICE_REBUILD,
SPA_FEATURE_ZSTD_COMPRESS,
SPA_FEATURE_DRAID,
SPA_FEATURE_ZILSAXATTR,
SPA_FEATURE_HEAD_ERRLOG,
SPA_FEATURE_BLAKE3,
SPA_FEATURE_BLOCK_CLONING,
+ SPA_FEATURE_AVZ_V2,
SPA_FEATURES
} spa_feature_t;
#define SPA_FEATURE_DISABLED (-1ULL)
typedef enum zfeature_flags {
/* Can open pool readonly even if this feature is not supported. */
ZFEATURE_FLAG_READONLY_COMPAT = (1 << 0),
/*
* Is this feature necessary to load the pool? i.e. do we need this
* feature to read the full feature list out of the MOS?
*/
ZFEATURE_FLAG_MOS = (1 << 1),
/* Activate this feature at the same time it is enabled. */
ZFEATURE_FLAG_ACTIVATE_ON_ENABLE = (1 << 2),
/* Each dataset has a field set if it has ever used this feature. */
ZFEATURE_FLAG_PER_DATASET = (1 << 3)
} zfeature_flags_t;
typedef enum zfeature_type {
ZFEATURE_TYPE_BOOLEAN,
ZFEATURE_TYPE_UINT64_ARRAY,
ZFEATURE_NUM_TYPES
} zfeature_type_t;
typedef struct zfeature_info {
spa_feature_t fi_feature;
const char *fi_uname; /* User-facing feature name */
const char *fi_guid; /* On-disk feature identifier */
const char *fi_desc; /* Feature description */
zfeature_flags_t fi_flags;
boolean_t fi_zfs_mod_supported; /* supported by running zfs module */
zfeature_type_t fi_type; /* Only relevant for PER_DATASET features */
/* array of dependencies, terminated by SPA_FEATURE_NONE */
const spa_feature_t *fi_depends;
} zfeature_info_t;
typedef int (zfeature_func_t)(zfeature_info_t *, void *);
#define ZFS_FEATURE_DEBUG
_ZFEATURE_COMMON_H zfeature_info_t spa_feature_table[SPA_FEATURES];
_ZFEATURE_COMMON_H boolean_t zfeature_checks_disable;
_ZFEATURE_COMMON_H boolean_t zfeature_is_valid_guid(const char *);
_ZFEATURE_COMMON_H boolean_t zfeature_is_supported(const char *);
_ZFEATURE_COMMON_H int zfeature_lookup_guid(const char *, spa_feature_t *);
_ZFEATURE_COMMON_H int zfeature_lookup_name(const char *, spa_feature_t *);
_ZFEATURE_COMMON_H boolean_t zfeature_depends_on(spa_feature_t, spa_feature_t);
_ZFEATURE_COMMON_H void zpool_feature_init(void);
#ifdef __cplusplus
}
#endif
#endif /* _ZFEATURE_COMMON_H */
diff --git a/sys/contrib/openzfs/lib/libspl/include/sys/isa_defs.h b/sys/contrib/openzfs/lib/libspl/include/sys/isa_defs.h
index 114cca4f1545..302f31e989cb 100644
--- a/sys/contrib/openzfs/lib/libspl/include/sys/isa_defs.h
+++ b/sys/contrib/openzfs/lib/libspl/include/sys/isa_defs.h
@@ -1,277 +1,293 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2007 Sun Microsystems, Inc. All rights reserved.
* Use is subject to license terms.
*/
#ifndef _SYS_ISA_DEFS_H
#define _SYS_ISA_DEFS_H
#ifdef __cplusplus
extern "C" {
#endif
/* x86_64 arch specific defines */
#if defined(__x86_64) || defined(__x86_64__)
#if !defined(__x86_64)
#define __x86_64
#endif
#if !defined(__amd64)
#define __amd64
#endif
#if !defined(__x86)
#define __x86
#endif
#if defined(_ILP32)
/* x32-specific defines; careful to *not* define _LP64 here */
#else
#if !defined(_LP64)
#define _LP64
#endif
#endif
#if !defined(_ZFS_LITTLE_ENDIAN)
#define _ZFS_LITTLE_ENDIAN
#endif
#define _SUNOS_VTOC_16
#define HAVE_EFFICIENT_UNALIGNED_ACCESS
/* i386 arch specific defines */
#elif defined(__i386) || defined(__i386__)
#if !defined(__i386)
#define __i386
#endif
#if !defined(__x86)
#define __x86
#endif
#if !defined(_ILP32)
#define _ILP32
#endif
#if !defined(_ZFS_LITTLE_ENDIAN)
#define _ZFS_LITTLE_ENDIAN
#endif
#define _SUNOS_VTOC_16
#define HAVE_EFFICIENT_UNALIGNED_ACCESS
/* powerpc arch specific defines */
#elif defined(__powerpc) || defined(__powerpc__) || defined(__powerpc64__)
#if !defined(__powerpc)
#define __powerpc
#endif
#if !defined(__powerpc__)
#define __powerpc__
#endif
#if defined(__powerpc64__)
#if !defined(_LP64)
#define _LP64
#endif
#else
#if !defined(_ILP32)
#define _ILP32
#endif
#endif
#define _SUNOS_VTOC_16
#define HAVE_EFFICIENT_UNALIGNED_ACCESS
#if defined(__BYTE_ORDER)
#if defined(__BIG_ENDIAN) && __BYTE_ORDER == __BIG_ENDIAN
#define _ZFS_BIG_ENDIAN
#elif defined(__LITTLE_ENDIAN) && __BYTE_ORDER == __LITTLE_ENDIAN
#define _ZFS_LITTLE_ENDIAN
#endif
#elif defined(_BYTE_ORDER)
#if defined(_BIG_ENDIAN) && _BYTE_ORDER == _BIG_ENDIAN
#define _ZFS_BIG_ENDIAN
#elif defined(_LITTLE_ENDIAN) && _BYTE_ORDER == _LITTLE_ENDIAN
#define _ZFS_LITTLE_ENDIAN
#endif
#elif defined(_BIG_ENDIAN) && !defined(_LITTLE_ENDIAN)
#define _ZFS_BIG_ENDIAN
#elif defined(_LITTLE_ENDIAN) && !defined(_BIG_ENDIAN)
#define _ZFS_LITTLE_ENDIAN
#endif
/* arm arch specific defines */
#elif defined(__arm) || defined(__arm__)
#if !defined(__arm)
#define __arm
#endif
#if !defined(__arm__)
#define __arm__
#endif
#if !defined(_ILP32)
#define _ILP32
#endif
#if defined(__ARMEL__)
#define _ZFS_LITTLE_ENDIAN
#else
#define _ZFS_BIG_ENDIAN
#endif
#define _SUNOS_VTOC_16
#if defined(__ARM_FEATURE_UNALIGNED)
#define HAVE_EFFICIENT_UNALIGNED_ACCESS
#endif
/* aarch64 arch specific defines */
#elif defined(__aarch64__)
#if !defined(_LP64)
#define _LP64
#endif
#if defined(__AARCH64EL__)
#define _ZFS_LITTLE_ENDIAN
#else
#define _ZFS_BIG_ENDIAN
#endif
#define _SUNOS_VTOC_16
/* sparc arch specific defines */
#elif defined(__sparc) || defined(__sparc__)
#if !defined(__sparc)
#define __sparc
#endif
#if !defined(__sparc__)
#define __sparc__
#endif
#define _ZFS_BIG_ENDIAN
#define _SUNOS_VTOC_16
#if defined(__arch64__)
#if !defined(_LP64)
#define _LP64
#endif
#else
#if !defined(_ILP32)
#define _ILP32
#endif
#endif
/* s390 arch specific defines */
#elif defined(__s390__)
#if defined(__s390x__)
#if !defined(_LP64)
#define _LP64
#endif
#else
#if !defined(_ILP32)
#define _ILP32
#endif
#endif
#define _ZFS_BIG_ENDIAN
#define _SUNOS_VTOC_16
/* MIPS arch specific defines */
#elif defined(__mips__)
#if defined(__MIPSEB__)
#define _ZFS_BIG_ENDIAN
#elif defined(__MIPSEL__)
#define _ZFS_LITTLE_ENDIAN
#else
#error MIPS no endian specified
#endif
#if !defined(_LP64) && !defined(_ILP32)
#define _ILP32
#endif
#define _SUNOS_VTOC_16
/*
* RISC-V arch specific defines
* only RV64G (including atomic) LP64 is supported yet
*/
#elif defined(__riscv) && defined(__riscv_xlen) && __riscv_xlen == 64 && \
defined(__riscv_atomic) && __riscv_atomic
#if !defined(_LP64)
#define _LP64 1
#endif
#ifndef __riscv__
#define __riscv__
#endif
#ifndef __rv64g__
#define __rv64g__
#endif
#define _ZFS_LITTLE_ENDIAN
#define _SUNOS_VTOC_16
+/*
+ * LoongArch arch specific defines
+ * only LoongArch64 is supported yet
+ */
+#elif defined(__loongarch__) && defined(__loongarch_lp64)
+
+#if !defined(_LP64)
+#define _LP64
+#endif
+
+#define _ZFS_LITTLE_ENDIAN
+#define _SUNOS_VTOC_16
+
+/* not all LoongArch cores support unaligned accesses in hardware */
+#define _ALIGNMENT_REQUIRED 1
+
#else
/*
* Currently supported:
- * x86_64, x32, i386, arm, powerpc, s390, sparc, mips, and RV64G
+ * x86_64, x32, i386, arm, powerpc, s390, sparc, mips, RV64G, and LoongArch64
*/
#error "Unsupported ISA type"
#endif
#if defined(_ILP32) && defined(_LP64)
#error "Both _ILP32 and _LP64 are defined"
#endif
#if !defined(_ILP32) && !defined(_LP64)
#error "Neither _ILP32 or _LP64 are defined"
#endif
#if defined(_ZFS_LITTLE_ENDIAN) && defined(_ZFS_BIG_ENDIAN)
#error "Both _ZFS_LITTLE_ENDIAN and _ZFS_BIG_ENDIAN are defined"
#endif
#if !defined(_ZFS_LITTLE_ENDIAN) && !defined(_ZFS_BIG_ENDIAN)
#error "Neither _ZFS_LITTLE_ENDIAN nor _ZFS_BIG_ENDIAN are defined"
#endif
#ifdef __cplusplus
}
#endif
#endif /* _SYS_ISA_DEFS_H */
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs.abi b/sys/contrib/openzfs/lib/libzfs/libzfs.abi
index f12424106d39..732863dcffc7 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs.abi
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs.abi
@@ -1,9353 +1,9414 @@
<abi-corpus version='2.0' architecture='elf-amd-x86_64' soname='libzfs.so.4'>
<elf-needed>
<dependency name='libzfs_core.so.3'/>
<dependency name='libnvpair.so.3'/>
<dependency name='libuuid.so.1'/>
<dependency name='libblkid.so.1'/>
<dependency name='libudev.so.1'/>
<dependency name='libuutil.so.3'/>
<dependency name='libm.so.6'/>
<dependency name='libcrypto.so.3'/>
<dependency name='libz.so.1'/>
<dependency name='libc.so.6'/>
<dependency name='ld-linux-x86-64.so.2'/>
</elf-needed>
<elf-function-symbols>
<elf-symbol name='_sol_getmntent' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_char' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_char_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_int_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_long' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_long_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_ptr_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_short' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_add_short_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_and_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_cas_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_clear_long_excl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_dec_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_inc_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uchar_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_uint_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ulong_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_or_ushort_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_set_long_excl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_16_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_32_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_64_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_8_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_char' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_char_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_int_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_long' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_long_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_ptr_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_short' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_sub_short_nv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_16' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_32' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_8' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ptr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_uchar' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_uint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ulong' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='atomic_swap_ushort' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_add' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_destroy_nodes' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_find' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_first' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_insert' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_insert_here' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_is_empty' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_last' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_nearest' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_numnodes' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_swap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_update' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_update_gt' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_update_lt' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='avl_walk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='bookmark_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='cityhash4' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='color_end' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='color_start' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='dataset_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='dataset_nestcheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_alloc_and_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_alloc_and_read' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_err_check' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_free' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_rescan' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_use_whole_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='efi_write' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='entity_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_2_byteswap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_2_incremental_byteswap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_2_incremental_native' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_2_native' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_byteswap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_fini' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_impl_set' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_incremental_byteswap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_incremental_native' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_native' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_native_varsize' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='get_dataset_depth' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='get_system_hostid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getexecname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getextmntent' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getmntany' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getprop_uint64' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='getzoneid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='is_mounted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='is_mpath_whole_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libpc_error_description' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libspl_assertf' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libspl_set_assert_ok' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_add_handle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_envvar_is_set' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_errno' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_error_action' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_error_description' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_error_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_fini' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_free_str_array' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_add' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_cache' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_find' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_fini' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_mnttab_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_print_on_error' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_run_process' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_run_process_get_stdout' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_run_process_get_stdout_nopath' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_after' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_before' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_insert_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_is_empty' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_active' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_link_replace' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_move_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_next' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_prev' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove_head' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_remove_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='list_tail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_consumer' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_enter' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_exit' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_producer' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='membar_sync' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='mkdirp' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='mountpoint_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='permset_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='pool_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='print_timestamp' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='printf_color' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_commit_shares' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_disable_share' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_enable_share' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_errorstr' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_is_shared' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_truncate_shares' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_validate_shareopts' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='snapshot_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='spl_pagesize' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='strlcat' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='strlcpy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_abandon' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_dispatch' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_member' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_resume' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_suspend' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_suspended' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='tpool_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='update_vdev_config_dev_strs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='use_color' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_expand_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_name_to_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_align_right' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_column_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_default_numeric' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_default_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_get_table' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_get_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_index_to_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_random_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_readonly' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_string_to_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_user' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='vdev_prop_values' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_depends_on' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_is_supported' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_is_valid_guid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_lookup_guid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_lookup_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_adjust_mount_options' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_allocatable_devs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_append_partition' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_basename' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_bookmark_exists' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_clone' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_close' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_commit_shares' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_component_namecheck' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_create_ancestors' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_attempt_load_keys' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_clone_check' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_get_encryption_root' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_load_key' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_rewrap' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_crypto_unload_key' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dataset_exists' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dataset_name_hidden' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_deleg_canonicalize_perm' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_deleg_verify_nvlist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_deleg_whokey' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_destroy_snaps' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_destroy_snaps_nvl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_destroy_snaps_nvl_os' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dev_flush' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dev_is_dm' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dev_is_whole_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_device_get_devid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_device_get_physical' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_dirnamelen' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_expand_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_foreach_mountpoint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_all_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_clones_nvl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_enclosure_sysfs_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_fsacl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_handle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_holds' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_pool_handle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_pool_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_recvd_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_underlying_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_underlying_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_get_user_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_handle_dup' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_hold' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_hold_nvl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_ioctl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_is_mounted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_is_shared' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_isnumber' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_bookmarks' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='zfs_iter_bookmarks_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_children' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='zfs_iter_children_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_dependents' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='zfs_iter_dependents_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_filesystems' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='zfs_iter_filesystems_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_mounted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_root' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapshots' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapshots_sorted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='zfs_iter_snapshots_sorted_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='zfs_iter_snapshots_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_iter_snapspec' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='zfs_iter_snapspec_v2' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_mod_supported' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_mount' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_mount_at' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_mount_delegation_check' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_name_to_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_name_valid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicebytes' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicenum' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicenum_format' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_niceraw' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicestrtonum' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_nicetime' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_open' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_parent_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_parse_mount_options' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_path_to_zhandle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_promote' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_align_right' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_column_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_default_numeric' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_default_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_delegatable' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_encryption_key_param' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_numeric' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_recvd' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_table' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_userquota' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_userquota_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_written' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_get_written_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_index_to_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_inherit' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_inheritable' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_is_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_random_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_readonly' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_set' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_set_list' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_setonce' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_string_to_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_user' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_userquota' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_valid_for_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_valid_keylocation' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_values' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_visible' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prop_written' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_prune_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_receive' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_refresh_properties' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_release' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_rename' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_resolve_shortname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_rollback' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_save_arguments' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_one' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_progress' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_resume' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_resume_token_to_nvlist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_send_saved' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_set_fsacl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_setproctitle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_setproctitle_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_share' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_show_diffs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_smb_acl_add' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_smb_acl_purge' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_smb_acl_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_smb_acl_rename' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_snapshot' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_snapshot_nvl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_spa_version' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_spa_version_map' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_special_devs' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_standard_error' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_strcmp_pathname' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_strip_partition' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_strip_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_truncate_shares' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_type_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_unmount' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_unmountall' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_unshare' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_unshareall' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_userns' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_userspace' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_valid_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_version_kernel' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_version_print' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_version_userland' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_wait_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_zpl_version_map' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_add' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_checkpoint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_clear' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_clear_label' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_close' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_create' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_default_search_paths' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_destroy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_disable_datasets' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_disable_datasets_os' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_disable_volume_os' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_discard_checkpoint' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_dump_ddt' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_enable_datasets' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_events_clear' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_events_next' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_events_seek' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_expand_proplist' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_explain_recover' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_export' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_export_force' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_feature_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_find_config' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_find_vdev' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_find_vdev_by_physpath' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_free_handles' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_all_vdev_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_bootenv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_config' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_errlog' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_features' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_handle' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_history' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_load_policy' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_prop_int' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_state' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_state_str' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_userprop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_vdev_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_get_vdev_prop_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_history_unpack' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_import' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_import_props' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_import_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_in_use' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_initialize' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_initialize_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_is_draid_spare' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_iter' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_label_disk' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_label_disk_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_load_compat' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_log_history' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_name_to_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_obj_to_path' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_obj_to_path_ds' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_open' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_open_canfail' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_pool_state_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_print_unsup_feat' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_align_right' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_column_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_default_numeric' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_default_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_feature' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_get_feature' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_get_table' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_get_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_index_to_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_init' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_random_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_readonly' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_setonce' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_string_to_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_unsupported' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_values' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_prop_vdev' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_props_refresh' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_read_label' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_refresh_stats' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_reguid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_reopen_one' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_scan' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_search_import' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_set_bootenv' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_set_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_set_vdev_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_skip_pool' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_state_to_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_sync_one' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_trim' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_upgrade' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_attach' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_clear' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_degrade' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_detach' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_fault' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_indirect_size' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_name' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_offline' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_online' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_path_to_guid' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_remove' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_remove_cancel' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_remove_wanted' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_vdev_split' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_wait' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zpool_wait_status' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_free_list' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_get_list' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_index_to_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_iter' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_iter_common' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_name_to_prop' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_print_one_property' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_random_value' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_hidden' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_impl' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_number' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_register_string' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_string_to_index' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_valid_char' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_valid_for_type' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_values' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zprop_width' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zvol_volsize_to_reservation' type='func-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
</elf-function-symbols>
<elf-variable-symbols>
<elf-symbol name='efi_debug' size='4' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_abd_ops' size='24' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_avx2_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_avx512bw_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_avx512f_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_sse2_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_ssse3_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_superscalar4_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='fletcher_4_superscalar_ops' size='128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='libzfs_config_ops' size='16' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='sa_protocol_names' size='16' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
- <elf-symbol name='spa_feature_table' size='2128' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
+ <elf-symbol name='spa_feature_table' size='2184' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfeature_checks_disable' size='4' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_deleg_perm_tab' size='512' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_history_event_names' size='328' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_max_dataset_nesting' size='4' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
<elf-symbol name='zfs_userquota_prop_prefixes' size='96' type='object-type' binding='global-binding' visibility='default-visibility' is-defined='yes'/>
</elf-variable-symbols>
<abi-instr address-size='64' path='lib/libefi/rdwr_efi.c' language='LANG_C99'>
<typedef-decl name='uInt' type-id='f0981eeb' id='09110a74'/>
<var-decl name='efi_debug' type-id='95e97e5e' mangled-name='efi_debug' visibility='default' elf-symbol-id='efi_debug'/>
<function-decl name='uuid_generate' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='cf536864'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uuid_is_null' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='354f7eb9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='crc32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5bbcce85'/>
<parameter type-id='e8cb3e0e'/>
<parameter type-id='09110a74'/>
<return type-id='5bbcce85'/>
</function-decl>
<function-decl name='efi_err_check' mangled-name='efi_err_check' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_err_check'>
<parameter type-id='0d8119a8' name='vtoc'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libshare/libshare.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='b99c00c9' size-in-bits='128' id='2d6895a3'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<var-decl name='sa_protocol_names' type-id='2d6895a3' mangled-name='sa_protocol_names' visibility='default' elf-symbol-id='sa_protocol_names'/>
<type-decl name='unsigned long int' size-in-bits='64' id='7359adad'/>
</abi-instr>
<abi-instr address-size='64' path='lib/libshare/nfs.c' language='LANG_C99'>
<function-decl name='rename' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='memchr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='flock' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fchmod' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='e1c52942'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='mkdir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='e1c52942'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libshare/os/linux/nfs.c' language='LANG_C99'>
<class-decl name='sa_share_impl' size-in-bits='192' is-struct='yes' visibility='default' id='72b09bf8'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='sa_zfsname' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='sa_mountpoint' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='sa_shareopts' type-id='80f4b756' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='sa_share_impl_t' type-id='946a2c6b' id='a48b47d0'/>
<class-decl name='sa_fstype_t' size-in-bits='384' is-struct='yes' naming-typedef-id='639af739' visibility='default' id='944afa86'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='enable_share' type-id='2f78a9c1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='disable_share' type-id='2f78a9c1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='is_shared' type-id='81020bc2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='validate_shareopts' type-id='f194a8fb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='commit_shares' type-id='797ee7da' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='truncate_shares' type-id='5d51038b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='sa_fstype_t' type-id='944afa86' id='639af739'/>
<qualified-type-def type-id='639af739' const='yes' id='d19dbca9'/>
<qualified-type-def type-id='72b09bf8' const='yes' id='484950e3'/>
<pointer-type-def type-id='484950e3' size-in-bits='64' id='946a2c6b'/>
<pointer-type-def type-id='276427e1' size-in-bits='64' id='1db260e5'/>
<qualified-type-def type-id='1db260e5' const='yes' id='797ee7da'/>
<pointer-type-def type-id='5113b296' size-in-bits='64' id='70487b28'/>
<qualified-type-def type-id='70487b28' const='yes' id='f194a8fb'/>
<pointer-type-def type-id='c13578bc' size-in-bits='64' id='fa1f29ce'/>
<qualified-type-def type-id='fa1f29ce' const='yes' id='2f78a9c1'/>
<pointer-type-def type-id='723e6cf2' size-in-bits='64' id='1d99e49c'/>
<pointer-type-def type-id='86373eb1' size-in-bits='64' id='f337456d'/>
<qualified-type-def type-id='f337456d' const='yes' id='81020bc2'/>
<qualified-type-def type-id='953b12f8' const='yes' id='5d51038b'/>
<var-decl name='libshare_nfs_type' type-id='d19dbca9' visibility='default'/>
<function-decl name='nfs_escape_mountpoint' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='9b23c9ad'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nfs_is_shared_impl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='a48b47d0'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nfs_toggle_share' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='a48b47d0'/>
<parameter type-id='1d99e49c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nfs_reset_shares' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='276427e1'>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='5113b296'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='c13578bc'>
<parameter type-id='a48b47d0'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='723e6cf2'>
<parameter type-id='a48b47d0'/>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='86373eb1'>
<parameter type-id='a48b47d0'/>
<return type-id='c19b74c3'/>
</function-type>
<function-type size-in-bits='64' id='ee076206'>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libshare/os/linux/smb.c' language='LANG_C99'>
<var-decl name='libshare_smb_type' type-id='d19dbca9' visibility='default'/>
<function-decl name='__fgets_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='e75a27e9'/>
<return type-id='26a90f95'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/assert.c' language='LANG_C99'>
<function-decl name='libspl_set_assert_ok' mangled-name='libspl_set_assert_ok' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libspl_set_assert_ok'>
<parameter type-id='c19b74c3' name='val'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/atomic.c' language='LANG_C99'>
<typedef-decl name='int8_t' type-id='2171a512' id='ee31ee44'/>
<typedef-decl name='__int8_t' type-id='28577a57' id='2171a512'/>
<qualified-type-def type-id='149c6638' volatile='yes' id='5120c5f7'/>
<pointer-type-def type-id='5120c5f7' size-in-bits='64' id='93977ae7'/>
<qualified-type-def type-id='b96825af' volatile='yes' id='84ff7d66'/>
<pointer-type-def type-id='84ff7d66' size-in-bits='64' id='aa323ea4'/>
<qualified-type-def type-id='ee1f298e' volatile='yes' id='6f7e09cb'/>
<pointer-type-def type-id='6f7e09cb' size-in-bits='64' id='64698d33'/>
<function-decl name='atomic_inc_8' mangled-name='atomic_inc_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_8'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_16' mangled-name='atomic_inc_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_16'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_32' mangled-name='atomic_inc_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_32'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_ulong' mangled-name='atomic_inc_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_ulong'>
<parameter type-id='64698d33' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_8' mangled-name='atomic_dec_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_8'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_16' mangled-name='atomic_dec_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_16'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_32' mangled-name='atomic_dec_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_32'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_dec_ulong' mangled-name='atomic_dec_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_ulong'>
<parameter type-id='64698d33' name='target'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_ptr' mangled-name='atomic_add_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_8' mangled-name='atomic_add_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_16' mangled-name='atomic_add_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_add_32' mangled-name='atomic_add_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_ptr' mangled-name='atomic_sub_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_8' mangled-name='atomic_sub_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_16' mangled-name='atomic_sub_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_sub_32' mangled-name='atomic_sub_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_8' mangled-name='atomic_or_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_16' mangled-name='atomic_or_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_32' mangled-name='atomic_or_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_or_ulong' mangled-name='atomic_or_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_8' mangled-name='atomic_and_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_16' mangled-name='atomic_and_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_32' mangled-name='atomic_and_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_and_ulong' mangled-name='atomic_and_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='atomic_inc_8_nv' mangled-name='atomic_inc_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_inc_16_nv' mangled-name='atomic_inc_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_inc_32_nv' mangled-name='atomic_inc_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_inc_ulong_nv' mangled-name='atomic_inc_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_inc_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_dec_8_nv' mangled-name='atomic_dec_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_dec_16_nv' mangled-name='atomic_dec_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_dec_32_nv' mangled-name='atomic_dec_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_dec_ulong_nv' mangled-name='atomic_dec_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_dec_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_add_ptr_nv' mangled-name='atomic_add_ptr_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_ptr_nv'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_add_8_nv' mangled-name='atomic_add_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_add_16_nv' mangled-name='atomic_add_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_add_32_nv' mangled-name='atomic_add_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_add_long_nv' mangled-name='atomic_add_long_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_add_long_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='bd54fe1a' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_sub_ptr_nv' mangled-name='atomic_sub_ptr_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_ptr_nv'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='79a0948f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_sub_8_nv' mangled-name='atomic_sub_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='ee31ee44' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_sub_16_nv' mangled-name='atomic_sub_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='23bd8cb5' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_sub_32_nv' mangled-name='atomic_sub_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='3ff5601b' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_sub_long_nv' mangled-name='atomic_sub_long_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_sub_long_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='bd54fe1a' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_or_8_nv' mangled-name='atomic_or_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_or_16_nv' mangled-name='atomic_or_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_or_32_nv' mangled-name='atomic_or_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_or_ulong_nv' mangled-name='atomic_or_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_or_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_and_8_nv' mangled-name='atomic_and_8_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_8_nv'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_and_16_nv' mangled-name='atomic_and_16_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_16_nv'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_and_32_nv' mangled-name='atomic_and_32_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_32_nv'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='bits'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_and_ulong_nv' mangled-name='atomic_and_ulong_nv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_and_ulong_nv'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_cas_ptr' mangled-name='atomic_cas_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='eaa32e2f' name='exp'/>
<parameter type-id='eaa32e2f' name='des'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_cas_8' mangled-name='atomic_cas_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='exp'/>
<parameter type-id='b96825af' name='des'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_cas_16' mangled-name='atomic_cas_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='exp'/>
<parameter type-id='149c6638' name='des'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_cas_32' mangled-name='atomic_cas_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_32'>
<parameter type-id='3a147f31' name='target'/>
<parameter type-id='8f92235e' name='exp'/>
<parameter type-id='8f92235e' name='des'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='atomic_cas_ulong' mangled-name='atomic_cas_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_cas_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='exp'/>
<parameter type-id='ee1f298e' name='des'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_swap_8' mangled-name='atomic_swap_8' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_8'>
<parameter type-id='aa323ea4' name='target'/>
<parameter type-id='b96825af' name='bits'/>
<return type-id='b96825af'/>
</function-decl>
<function-decl name='atomic_swap_16' mangled-name='atomic_swap_16' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_16'>
<parameter type-id='93977ae7' name='target'/>
<parameter type-id='149c6638' name='bits'/>
<return type-id='149c6638'/>
</function-decl>
<function-decl name='atomic_swap_ulong' mangled-name='atomic_swap_ulong' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_ulong'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='ee1f298e' name='bits'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='atomic_swap_ptr' mangled-name='atomic_swap_ptr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_ptr'>
<parameter type-id='fe09dd29' name='target'/>
<parameter type-id='eaa32e2f' name='bits'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='atomic_set_long_excl' mangled-name='atomic_set_long_excl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_set_long_excl'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='3502e3ff' name='value'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='atomic_clear_long_excl' mangled-name='atomic_clear_long_excl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_clear_long_excl'>
<parameter type-id='64698d33' name='target'/>
<parameter type-id='3502e3ff' name='value'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='membar_enter' mangled-name='membar_enter' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_enter'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='membar_consumer' mangled-name='membar_consumer' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_consumer'>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/getexecname.c' language='LANG_C99'>
<function-decl name='getexecname' mangled-name='getexecname' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getexecname'>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='getexecname_impl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='26a90f95'/>
<return type-id='79a0948f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/list.c' language='LANG_C99'>
<typedef-decl name='list_node_t' type-id='b0b5e45e' id='b21843b2'/>
<typedef-decl name='list_t' type-id='e824dae9' id='0899125f'/>
<class-decl name='list_node' size-in-bits='128' is-struct='yes' visibility='default' id='b0b5e45e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='next' type-id='b03eadb4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='prev' type-id='b03eadb4' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='list' size-in-bits='256' is-struct='yes' visibility='default' id='e824dae9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='list_size' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='list_offset' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='list_head' type-id='b0b5e45e' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='b0b5e45e' size-in-bits='64' id='b03eadb4'/>
<pointer-type-def type-id='b21843b2' size-in-bits='64' id='ccc38265'/>
<pointer-type-def type-id='0899125f' size-in-bits='64' id='352ec160'/>
<function-decl name='list_create' mangled-name='list_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_create'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='b59d7dce' name='size'/>
<parameter type-id='b59d7dce' name='offset'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_destroy' mangled-name='list_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_destroy'>
<parameter type-id='352ec160' name='list'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_after' mangled-name='list_insert_after' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_after'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<parameter type-id='eaa32e2f' name='nobject'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_before' mangled-name='list_insert_before' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_before'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<parameter type-id='eaa32e2f' name='nobject'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_head' mangled-name='list_insert_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_head'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_insert_tail' mangled-name='list_insert_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_insert_tail'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_remove' mangled-name='list_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_remove_head' mangled-name='list_remove_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove_head'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_remove_tail' mangled-name='list_remove_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_remove_tail'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_head' mangled-name='list_head' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_head'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_tail' mangled-name='list_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_tail'>
<parameter type-id='352ec160' name='list'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_next' mangled-name='list_next' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_next'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_prev' mangled-name='list_prev' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_prev'>
<parameter type-id='352ec160' name='list'/>
<parameter type-id='eaa32e2f' name='object'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='list_move_tail' mangled-name='list_move_tail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_move_tail'>
<parameter type-id='352ec160' name='dst'/>
<parameter type-id='352ec160' name='src'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_replace' mangled-name='list_link_replace' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_replace'>
<parameter type-id='ccc38265' name='lold'/>
<parameter type-id='ccc38265' name='lnew'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_init' mangled-name='list_link_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_init'>
<parameter type-id='ccc38265' name='ln'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='list_link_active' mangled-name='list_link_active' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_link_active'>
<parameter type-id='ccc38265' name='ln'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='list_is_empty' mangled-name='list_is_empty' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='list_is_empty'>
<parameter type-id='352ec160' name='list'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/mkdirp.c' language='LANG_C99'>
<typedef-decl name='wchar_t' type-id='95e97e5e' id='928221d2'/>
<qualified-type-def type-id='928221d2' const='yes' id='effb3702'/>
<pointer-type-def type-id='effb3702' size-in-bits='64' id='f077d3f8'/>
<qualified-type-def type-id='f077d3f8' restrict='yes' id='598aab80'/>
<pointer-type-def type-id='928221d2' size-in-bits='64' id='323d93c1'/>
<qualified-type-def type-id='323d93c1' restrict='yes' id='f1358bc3'/>
<function-decl name='__mbstowcs_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f1358bc3'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='__wcstombs_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='598aab80'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/os/linux/getmntany.c' language='LANG_C99'>
<pointer-type-def type-id='56fe4a37' size-in-bits='64' id='b6b61d2f'/>
<qualified-type-def type-id='b6b61d2f' restrict='yes' id='3cad23cd'/>
<function-decl name='getmntent_r' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='3cad23cd'/>
<parameter type-id='266fe297'/>
<parameter type-id='95e97e5e'/>
<return type-id='b6b61d2f'/>
</function-decl>
<function-decl name='feof' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libspl/timestamp.c' language='LANG_C99'>
<typedef-decl name='nl_item' type-id='95e97e5e' id='03b79a94'/>
<function-decl name='nl_langinfo' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='03b79a94'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='print_timestamp' mangled-name='print_timestamp' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='print_timestamp'>
<parameter type-id='3502e3ff' name='timestamp_fmt'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libtpool/thread_pool.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='8901473c' size-in-bits='576' id='f5da478b'>
<subrange length='1' type-id='7359adad' id='52f813b4'/>
</array-type-def>
<array-type-def dimensions='1' type-id='49ef3ffd' size-in-bits='1024' id='a14403f5'>
<subrange length='16' type-id='7359adad' id='848d0938'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='384' id='36d7f119'>
<subrange length='48' type-id='7359adad' id='8f6d2a81'/>
</array-type-def>
<array-type-def dimensions='1' type-id='bd54fe1a' size-in-bits='512' id='5d4efd44'>
<subrange length='8' type-id='7359adad' id='56e0c0b1'/>
</array-type-def>
<array-type-def dimensions='1' type-id='f0981eeb' size-in-bits='64' id='0d532ec1'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<array-type-def dimensions='1' type-id='eaa32e2f' size-in-bits='256' id='209ef23f'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<class-decl name='__cancel_jmp_buf_tag' size-in-bits='576' is-struct='yes' visibility='default' id='8901473c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__cancel_jmp_buf' type-id='379a1ab7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='__mask_was_saved' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__pthread_unwind_buf_t' size-in-bits='832' is-struct='yes' naming-typedef-id='4423cf7f' visibility='default' id='a0abc656'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__cancel_jmp_buf' type-id='f5da478b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='__pad' type-id='209ef23f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__pthread_unwind_buf_t' type-id='a0abc656' id='4423cf7f'/>
<union-decl name='__atomic_wide_counter' size-in-bits='64' naming-typedef-id='f3b40860' visibility='default' id='613ce450'>
<data-member access='public'>
<var-decl name='__value64' type-id='3a47d82b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__value32' type-id='e7f43f72' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='__anonymous_struct__' size-in-bits='64' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f72'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__low' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='__high' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__atomic_wide_counter' type-id='613ce450' id='f3b40860'/>
<typedef-decl name='__cpu_mask' type-id='7359adad' id='49ef3ffd'/>
<class-decl name='cpu_set_t' size-in-bits='1024' is-struct='yes' naming-typedef-id='8037c762' visibility='default' id='1f20d231'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__bits' type-id='a14403f5' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='cpu_set_t' type-id='1f20d231' id='8037c762'/>
<union-decl name='pthread_condattr_t' size-in-bits='32' naming-typedef-id='836265dd' visibility='default' id='33dd3aad'>
<data-member access='public'>
<var-decl name='__size' type-id='8e0573fd' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='95e97e5e' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_condattr_t' type-id='33dd3aad' id='836265dd'/>
<union-decl name='pthread_cond_t' size-in-bits='384' naming-typedef-id='62fab762' visibility='default' id='cbb12c12'>
<data-member access='public'>
<var-decl name='__data' type-id='c987b47c' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__size' type-id='36d7f119' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='1eb56b1e' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_cond_t' type-id='cbb12c12' id='62fab762'/>
<typedef-decl name='__jmp_buf' type-id='5d4efd44' id='379a1ab7'/>
<class-decl name='__pthread_cond_s' size-in-bits='384' is-struct='yes' visibility='default' id='c987b47c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__wseq' type-id='f3b40860' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='__g1_start' type-id='f3b40860' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='__g_refs' type-id='0d532ec1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='__g_size' type-id='0d532ec1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='__g1_orig_size' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='__wrefs' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='__g_signals' type-id='0d532ec1' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='sched_param' size-in-bits='32' is-struct='yes' visibility='default' id='0897719a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='sched_priority' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='tpool_job_t' type-id='3b8579e5' id='66a0afc9'/>
<class-decl name='tpool_job' size-in-bits='192' is-struct='yes' visibility='default' id='3b8579e5'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tpj_next' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tpj_func' type-id='b7f9d8e6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tpj_arg' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='tpool_active_t' type-id='c8d086f4' id='6fcda10e'/>
<class-decl name='tpool_active' size-in-bits='128' is-struct='yes' visibility='default' id='c8d086f4'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tpa_next' type-id='ad33e5e7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tpa_tid' type-id='4051f5e7' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='8901473c' size-in-bits='64' id='eb91b7ea'/>
<pointer-type-def type-id='4423cf7f' size-in-bits='64' id='ba7c727c'/>
<pointer-type-def type-id='b9c97942' size-in-bits='64' id='bbf06c47'/>
<qualified-type-def type-id='bbf06c47' restrict='yes' id='65e6ec45'/>
<qualified-type-def type-id='b9c97942' const='yes' id='191f6b72'/>
<pointer-type-def type-id='191f6b72' size-in-bits='64' id='e475fb88'/>
<qualified-type-def type-id='e475fb88' restrict='yes' id='5a8729d0'/>
<qualified-type-def type-id='8037c762' const='yes' id='f50ea9b2'/>
<pointer-type-def type-id='f50ea9b2' size-in-bits='64' id='5e14fa48'/>
<qualified-type-def type-id='836265dd' const='yes' id='7d24c58d'/>
<pointer-type-def type-id='7d24c58d' size-in-bits='64' id='a7e325e5'/>
<qualified-type-def type-id='a7e325e5' restrict='yes' id='4c428e67'/>
<qualified-type-def type-id='0897719a' const='yes' id='c4a7b189'/>
<pointer-type-def type-id='c4a7b189' size-in-bits='64' id='36fca399'/>
<qualified-type-def type-id='36fca399' restrict='yes' id='37e4897b'/>
<qualified-type-def type-id='a9c79a1f' const='yes' id='cd087e36'/>
<pointer-type-def type-id='cd087e36' size-in-bits='64' id='e05e8614'/>
<qualified-type-def type-id='e05e8614' restrict='yes' id='0be2e71c'/>
<pointer-type-def type-id='8037c762' size-in-bits='64' id='d74a6869'/>
<qualified-type-def type-id='7292109c' restrict='yes' id='6942f6a4'/>
<qualified-type-def type-id='7347a39e' restrict='yes' id='578ba182'/>
<pointer-type-def type-id='62fab762' size-in-bits='64' id='db285b03'/>
<qualified-type-def type-id='db285b03' restrict='yes' id='2a468b41'/>
<qualified-type-def type-id='18c91f9e' restrict='yes' id='6e745582'/>
<pointer-type-def type-id='0897719a' size-in-bits='64' id='23cbcb08'/>
<qualified-type-def type-id='23cbcb08' restrict='yes' id='b09b2050'/>
<pointer-type-def type-id='6fcda10e' size-in-bits='64' id='ad33e5e7'/>
<pointer-type-def type-id='66a0afc9' size-in-bits='64' id='f32b30e4'/>
<qualified-type-def type-id='63e171df' restrict='yes' id='9e7a3a7d'/>
<function-decl name='pthread_self' visibility='default' binding='global' size-in-bits='64'>
<return type-id='4051f5e7'/>
</function-decl>
<function-decl name='pthread_attr_init' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getdetachstate' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='540db505'/>
<parameter type-id='7292109c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setdetachstate' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getguardsize' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='540db505'/>
<parameter type-id='78c01427'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setguardsize' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getschedparam' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='b09b2050'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setschedparam' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='578ba182'/>
<parameter type-id='37e4897b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getschedpolicy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='6942f6a4'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setschedpolicy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getinheritsched' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='6942f6a4'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setinheritsched' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getscope' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='6942f6a4'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setscope' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getstack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e1815e87'/>
<parameter type-id='9e7a3a7d'/>
<parameter type-id='d19b2c25'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setstack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_setaffinity_np' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7347a39e'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='5e14fa48'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_attr_getaffinity_np' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='540db505'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='d74a6869'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_setcancelstate' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='7292109c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_setcanceltype' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='7292109c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__pthread_register_cancel' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='ba7c727c'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__pthread_unregister_cancel' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='ba7c727c'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__pthread_unwind_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='ba7c727c'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='pthread_cond_init' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2a468b41'/>
<parameter type-id='4c428e67'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cond_signal' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='db285b03'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cond_broadcast' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='db285b03'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cond_wait' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2a468b41'/>
<parameter type-id='6e745582'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cond_timedwait' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2a468b41'/>
<parameter type-id='6e745582'/>
<parameter type-id='0be2e71c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__sysconf' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='bd54fe1a'/>
</function-decl>
<function-decl name='pthread_sigmask' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='5a8729d0'/>
<parameter type-id='65e6ec45'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='tpool_abandon' mangled-name='tpool_abandon' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_abandon'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_suspend' mangled-name='tpool_suspend' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_suspend'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_suspended' mangled-name='tpool_suspended' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_suspended'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='tpool_resume' mangled-name='tpool_resume' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_resume'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_member' mangled-name='tpool_member' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_member'>
<parameter type-id='9cf59a50' name='tpool'/>
<return type-id='95e97e5e'/>
</function-decl>
<pointer-type-def type-id='b1bbf10d' size-in-bits='64' id='9cf59a50'/>
<typedef-decl name='tpool_t' type-id='88d1b7f9' id='b1bbf10d'/>
<class-decl name='tpool' size-in-bits='2496' is-struct='yes' visibility='default' id='88d1b7f9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tp_forw' type-id='9cf59a50' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tp_back' type-id='9cf59a50' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tp_mutex' type-id='7a6844eb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='tp_busycv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='tp_workcv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1216'>
<var-decl name='tp_waitcv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1600'>
<var-decl name='tp_active' type-id='ad33e5e7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1664'>
<var-decl name='tp_head' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1728'>
<var-decl name='tp_tail' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1792'>
<var-decl name='tp_attr' type-id='7d8569fd' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2240'>
<var-decl name='tp_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2272'>
<var-decl name='tp_linger' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2304'>
<var-decl name='tp_njobs' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2336'>
<var-decl name='tp_minimum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2368'>
<var-decl name='tp_maximum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2400'>
<var-decl name='tp_current' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2432'>
<var-decl name='tp_idle' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<function-type size-in-bits='64' id='c5c76c9c'>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_changelist.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='bf311473' size-in-bits='128' id='f0f65199'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<type-decl name='char' size-in-bits='8' id='a84c031d'/>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='8192' id='b54ce520'>
<subrange length='1024' type-id='7359adad' id='c60446f8'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='2048' id='d1617432'>
<subrange length='256' type-id='7359adad' id='36e5b9fa'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='320' id='36c46961'>
<subrange length='40' type-id='7359adad' id='8f80b239'/>
</array-type-def>
<class-decl name='re_dfa_t' is-struct='yes' visibility='default' is-declaration-only='yes' id='b48d2441'/>
<class-decl name='uu_avl' is-struct='yes' visibility='default' is-declaration-only='yes' id='4af029d1'/>
<class-decl name='uu_avl_pool' is-struct='yes' visibility='default' is-declaration-only='yes' id='12a530a8'/>
<class-decl name='uu_avl_walk' is-struct='yes' visibility='default' is-declaration-only='yes' id='e70a39e3'/>
<type-decl name='int' size-in-bits='32' id='95e97e5e'/>
<type-decl name='long int' size-in-bits='64' id='bd54fe1a'/>
<type-decl name='long long int' size-in-bits='64' id='1eb56b1e'/>
<type-decl name='short int' size-in-bits='16' id='a2185560'/>
<array-type-def dimensions='1' type-id='e475ab95' size-in-bits='192' id='0ce65a8b'>
<subrange length='3' type-id='7359adad' id='56f209d2'/>
</array-type-def>
<type-decl name='unnamed-enum-underlying-type-32' is-anonymous='yes' size-in-bits='32' alignment-in-bits='32' id='9cac1fee'/>
<type-decl name='unsigned char' size-in-bits='8' id='002ac4a6'/>
<type-decl name='unsigned int' size-in-bits='32' id='f0981eeb'/>
<type-decl name='unsigned long int' size-in-bits='64' id='7359adad'/>
<type-decl name='void' id='48b5725f'/>
<typedef-decl name='uu_compare_fn_t' type-id='add6e811' id='40f93560'/>
<typedef-decl name='uu_avl_pool_t' type-id='12a530a8' id='7f84e390'/>
<typedef-decl name='uu_avl_t' type-id='4af029d1' id='bb7f0973'/>
<class-decl name='uu_avl_node' size-in-bits='192' is-struct='yes' visibility='default' id='f65f4326'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='uan_opaque' type-id='0ce65a8b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='uu_avl_node_t' type-id='f65f4326' id='73a65116'/>
<typedef-decl name='uu_avl_walk_t' type-id='e70a39e3' id='edd8457b'/>
<typedef-decl name='uu_avl_index_t' type-id='e475ab95' id='5d7f5fc8'/>
<typedef-decl name='zfs_handle_t' type-id='f6ee4445' id='775509eb'/>
<typedef-decl name='zpool_handle_t' type-id='67002a8a' id='b1efc708'/>
<typedef-decl name='libzfs_handle_t' type-id='c8a9d9d8' id='95942d0c'/>
<typedef-decl name='zfs_iter_f' type-id='5571cde4' id='d8e49ab9'/>
<typedef-decl name='avl_tree_t' type-id='b351119f' id='f20fbd51'/>
<class-decl name='avl_node' size-in-bits='192' is-struct='yes' visibility='default' id='428b67b3'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='avl_child' type-id='f0f65199' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='avl_pcb' type-id='e475ab95' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='avl_tree' size-in-bits='320' is-struct='yes' visibility='default' id='b351119f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='avl_root' type-id='bf311473' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='avl_compar' type-id='585e1de9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='avl_offset' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='avl_numnodes' type-id='ee1f298e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='avl_pad' type-id='b59d7dce' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='dmu_objset_stats' size-in-bits='2304' is-struct='yes' visibility='default' id='098f0221'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='dds_num_clones' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='dds_creation_txg' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='dds_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='dds_type' type-id='230f1e16' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='dds_is_snapshot' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='232'>
<var-decl name='dds_inconsistent' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='240'>
<var-decl name='dds_redacted' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='248'>
<var-decl name='dds_origin' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='dmu_objset_stats_t' type-id='098f0221' id='b2c14f17'/>
<enum-decl name='zfs_type_t' naming-typedef-id='2e45de5d' id='5d8f7321'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_TYPE_INVALID' value='0'/>
<enumerator name='ZFS_TYPE_FILESYSTEM' value='1'/>
<enumerator name='ZFS_TYPE_SNAPSHOT' value='2'/>
<enumerator name='ZFS_TYPE_VOLUME' value='4'/>
<enumerator name='ZFS_TYPE_POOL' value='8'/>
<enumerator name='ZFS_TYPE_BOOKMARK' value='16'/>
<enumerator name='ZFS_TYPE_VDEV' value='32'/>
</enum-decl>
<typedef-decl name='zfs_type_t' type-id='5d8f7321' id='2e45de5d'/>
<enum-decl name='dmu_objset_type' id='6b1b19f9'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DMU_OST_NONE' value='0'/>
<enumerator name='DMU_OST_META' value='1'/>
<enumerator name='DMU_OST_ZFS' value='2'/>
<enumerator name='DMU_OST_ZVOL' value='3'/>
<enumerator name='DMU_OST_OTHER' value='4'/>
<enumerator name='DMU_OST_ANY' value='5'/>
<enumerator name='DMU_OST_NUMTYPES' value='6'/>
</enum-decl>
<typedef-decl name='dmu_objset_type_t' type-id='6b1b19f9' id='230f1e16'/>
<enum-decl name='zfs_prop_t' naming-typedef-id='58603c44' id='4b000d60'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPROP_CONT' value='-2'/>
<enumerator name='ZPROP_INVAL' value='-1'/>
<enumerator name='ZPROP_USERPROP' value='-1'/>
<enumerator name='ZFS_PROP_TYPE' value='0'/>
<enumerator name='ZFS_PROP_CREATION' value='1'/>
<enumerator name='ZFS_PROP_USED' value='2'/>
<enumerator name='ZFS_PROP_AVAILABLE' value='3'/>
<enumerator name='ZFS_PROP_REFERENCED' value='4'/>
<enumerator name='ZFS_PROP_COMPRESSRATIO' value='5'/>
<enumerator name='ZFS_PROP_MOUNTED' value='6'/>
<enumerator name='ZFS_PROP_ORIGIN' value='7'/>
<enumerator name='ZFS_PROP_QUOTA' value='8'/>
<enumerator name='ZFS_PROP_RESERVATION' value='9'/>
<enumerator name='ZFS_PROP_VOLSIZE' value='10'/>
<enumerator name='ZFS_PROP_VOLBLOCKSIZE' value='11'/>
<enumerator name='ZFS_PROP_RECORDSIZE' value='12'/>
<enumerator name='ZFS_PROP_MOUNTPOINT' value='13'/>
<enumerator name='ZFS_PROP_SHARENFS' value='14'/>
<enumerator name='ZFS_PROP_CHECKSUM' value='15'/>
<enumerator name='ZFS_PROP_COMPRESSION' value='16'/>
<enumerator name='ZFS_PROP_ATIME' value='17'/>
<enumerator name='ZFS_PROP_DEVICES' value='18'/>
<enumerator name='ZFS_PROP_EXEC' value='19'/>
<enumerator name='ZFS_PROP_SETUID' value='20'/>
<enumerator name='ZFS_PROP_READONLY' value='21'/>
<enumerator name='ZFS_PROP_ZONED' value='22'/>
<enumerator name='ZFS_PROP_SNAPDIR' value='23'/>
<enumerator name='ZFS_PROP_ACLMODE' value='24'/>
<enumerator name='ZFS_PROP_ACLINHERIT' value='25'/>
<enumerator name='ZFS_PROP_CREATETXG' value='26'/>
<enumerator name='ZFS_PROP_NAME' value='27'/>
<enumerator name='ZFS_PROP_CANMOUNT' value='28'/>
<enumerator name='ZFS_PROP_ISCSIOPTIONS' value='29'/>
<enumerator name='ZFS_PROP_XATTR' value='30'/>
<enumerator name='ZFS_PROP_NUMCLONES' value='31'/>
<enumerator name='ZFS_PROP_COPIES' value='32'/>
<enumerator name='ZFS_PROP_VERSION' value='33'/>
<enumerator name='ZFS_PROP_UTF8ONLY' value='34'/>
<enumerator name='ZFS_PROP_NORMALIZE' value='35'/>
<enumerator name='ZFS_PROP_CASE' value='36'/>
<enumerator name='ZFS_PROP_VSCAN' value='37'/>
<enumerator name='ZFS_PROP_NBMAND' value='38'/>
<enumerator name='ZFS_PROP_SHARESMB' value='39'/>
<enumerator name='ZFS_PROP_REFQUOTA' value='40'/>
<enumerator name='ZFS_PROP_REFRESERVATION' value='41'/>
<enumerator name='ZFS_PROP_GUID' value='42'/>
<enumerator name='ZFS_PROP_PRIMARYCACHE' value='43'/>
<enumerator name='ZFS_PROP_SECONDARYCACHE' value='44'/>
<enumerator name='ZFS_PROP_USEDSNAP' value='45'/>
<enumerator name='ZFS_PROP_USEDDS' value='46'/>
<enumerator name='ZFS_PROP_USEDCHILD' value='47'/>
<enumerator name='ZFS_PROP_USEDREFRESERV' value='48'/>
<enumerator name='ZFS_PROP_USERACCOUNTING' value='49'/>
<enumerator name='ZFS_PROP_STMF_SHAREINFO' value='50'/>
<enumerator name='ZFS_PROP_DEFER_DESTROY' value='51'/>
<enumerator name='ZFS_PROP_USERREFS' value='52'/>
<enumerator name='ZFS_PROP_LOGBIAS' value='53'/>
<enumerator name='ZFS_PROP_UNIQUE' value='54'/>
<enumerator name='ZFS_PROP_OBJSETID' value='55'/>
<enumerator name='ZFS_PROP_DEDUP' value='56'/>
<enumerator name='ZFS_PROP_MLSLABEL' value='57'/>
<enumerator name='ZFS_PROP_SYNC' value='58'/>
<enumerator name='ZFS_PROP_DNODESIZE' value='59'/>
<enumerator name='ZFS_PROP_REFRATIO' value='60'/>
<enumerator name='ZFS_PROP_WRITTEN' value='61'/>
<enumerator name='ZFS_PROP_CLONES' value='62'/>
<enumerator name='ZFS_PROP_LOGICALUSED' value='63'/>
<enumerator name='ZFS_PROP_LOGICALREFERENCED' value='64'/>
<enumerator name='ZFS_PROP_INCONSISTENT' value='65'/>
<enumerator name='ZFS_PROP_VOLMODE' value='66'/>
<enumerator name='ZFS_PROP_FILESYSTEM_LIMIT' value='67'/>
<enumerator name='ZFS_PROP_SNAPSHOT_LIMIT' value='68'/>
<enumerator name='ZFS_PROP_FILESYSTEM_COUNT' value='69'/>
<enumerator name='ZFS_PROP_SNAPSHOT_COUNT' value='70'/>
<enumerator name='ZFS_PROP_SNAPDEV' value='71'/>
<enumerator name='ZFS_PROP_ACLTYPE' value='72'/>
<enumerator name='ZFS_PROP_SELINUX_CONTEXT' value='73'/>
<enumerator name='ZFS_PROP_SELINUX_FSCONTEXT' value='74'/>
<enumerator name='ZFS_PROP_SELINUX_DEFCONTEXT' value='75'/>
<enumerator name='ZFS_PROP_SELINUX_ROOTCONTEXT' value='76'/>
<enumerator name='ZFS_PROP_RELATIME' value='77'/>
<enumerator name='ZFS_PROP_REDUNDANT_METADATA' value='78'/>
<enumerator name='ZFS_PROP_OVERLAY' value='79'/>
<enumerator name='ZFS_PROP_PREV_SNAP' value='80'/>
<enumerator name='ZFS_PROP_RECEIVE_RESUME_TOKEN' value='81'/>
<enumerator name='ZFS_PROP_ENCRYPTION' value='82'/>
<enumerator name='ZFS_PROP_KEYLOCATION' value='83'/>
<enumerator name='ZFS_PROP_KEYFORMAT' value='84'/>
<enumerator name='ZFS_PROP_PBKDF2_SALT' value='85'/>
<enumerator name='ZFS_PROP_PBKDF2_ITERS' value='86'/>
<enumerator name='ZFS_PROP_ENCRYPTION_ROOT' value='87'/>
<enumerator name='ZFS_PROP_KEY_GUID' value='88'/>
<enumerator name='ZFS_PROP_KEYSTATUS' value='89'/>
<enumerator name='ZFS_PROP_REMAPTXG' value='90'/>
<enumerator name='ZFS_PROP_SPECIAL_SMALL_BLOCKS' value='91'/>
<enumerator name='ZFS_PROP_IVSET_GUID' value='92'/>
<enumerator name='ZFS_PROP_REDACTED' value='93'/>
<enumerator name='ZFS_PROP_REDACT_SNAPS' value='94'/>
<enumerator name='ZFS_PROP_SNAPSHOTS_CHANGED' value='95'/>
<enumerator name='ZFS_NUM_PROPS' value='96'/>
</enum-decl>
<typedef-decl name='zfs_prop_t' type-id='4b000d60' id='58603c44'/>
<enum-decl name='zprop_source_t' naming-typedef-id='a2256d42' id='5903f80e'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPROP_SRC_NONE' value='1'/>
<enumerator name='ZPROP_SRC_DEFAULT' value='2'/>
<enumerator name='ZPROP_SRC_TEMPORARY' value='4'/>
<enumerator name='ZPROP_SRC_LOCAL' value='8'/>
<enumerator name='ZPROP_SRC_INHERITED' value='16'/>
<enumerator name='ZPROP_SRC_RECEIVED' value='32'/>
</enum-decl>
<typedef-decl name='zprop_source_t' type-id='5903f80e' id='a2256d42'/>
<class-decl name='nvlist' size-in-bits='192' is-struct='yes' visibility='default' id='ac266fd9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='nvl_version' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='nvl_nvflag' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='nvl_priv' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='nvl_flag' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='nvl_pad' type-id='3ff5601b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='nvlist_t' type-id='ac266fd9' id='8e8d4be3'/>
<enum-decl name='sa_protocol' id='9155d4b5'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='SA_PROTOCOL_NFS' value='0'/>
<enumerator name='SA_PROTOCOL_SMB' value='1'/>
<enumerator name='SA_PROTOCOL_COUNT' value='2'/>
</enum-decl>
<enum-decl name='boolean_t' naming-typedef-id='c19b74c3' id='f58c8277'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='B_FALSE' value='0'/>
<enumerator name='B_TRUE' value='1'/>
</enum-decl>
<typedef-decl name='boolean_t' type-id='f58c8277' id='c19b74c3'/>
<typedef-decl name='uint_t' type-id='f0981eeb' id='3502e3ff'/>
<typedef-decl name='ulong_t' type-id='7359adad' id='ee1f298e'/>
<typedef-decl name='longlong_t' type-id='1eb56b1e' id='9b3ff54f'/>
<typedef-decl name='diskaddr_t' type-id='9b3ff54f' id='804dc465'/>
<typedef-decl name='zoneid_t' type-id='3502e3ff' id='4da03624'/>
<typedef-decl name='__re_long_size_t' type-id='7359adad' id='ba516949'/>
<typedef-decl name='reg_syntax_t' type-id='7359adad' id='1b72c3b3'/>
<class-decl name='re_pattern_buffer' size-in-bits='512' is-struct='yes' visibility='default' id='19fc9a8c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='buffer' type-id='33976309' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='allocated' type-id='ba516949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='used' type-id='ba516949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='syntax' type-id='1b72c3b3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='fastmap' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='translate' type-id='cf536864' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='re_nsub' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='can_be_null' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='449'>
<var-decl name='regs_allocated' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='451'>
<var-decl name='fastmap_accurate' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='452'>
<var-decl name='no_sub' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='453'>
<var-decl name='not_bol' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='454'>
<var-decl name='not_eol' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='455'>
<var-decl name='newline_anchor' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='regex_t' type-id='19fc9a8c' id='aca3bac8'/>
<typedef-decl name='uintptr_t' type-id='7359adad' id='e475ab95'/>
<union-decl name='pthread_mutex_t' size-in-bits='320' naming-typedef-id='7a6844eb' visibility='default' id='70681f9b'>
<data-member access='public'>
<var-decl name='__data' type-id='4c734837' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__size' type-id='36c46961' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='bd54fe1a' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_mutex_t' type-id='70681f9b' id='7a6844eb'/>
<typedef-decl name='int32_t' type-id='33f57a65' id='3ff5601b'/>
<typedef-decl name='uint8_t' type-id='c51d6389' id='b96825af'/>
<typedef-decl name='uint32_t' type-id='62f1140c' id='8f92235e'/>
<typedef-decl name='uint64_t' type-id='8910171f' id='9c313c2d'/>
<class-decl name='__pthread_mutex_s' size-in-bits='320' is-struct='yes' visibility='default' id='4c734837'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__lock' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='__count' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='__owner' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='__nusers' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='__kind' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='__spins' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='176'>
<var-decl name='__elision' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='__list' type-id='518fb49c' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__pthread_internal_list' size-in-bits='128' is-struct='yes' visibility='default' id='0e01899c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__prev' type-id='4d98cd5a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='__next' type-id='4d98cd5a' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__pthread_list_t' type-id='0e01899c' id='518fb49c'/>
<typedef-decl name='__uint8_t' type-id='002ac4a6' id='c51d6389'/>
<typedef-decl name='__int32_t' type-id='95e97e5e' id='33f57a65'/>
<typedef-decl name='__uint32_t' type-id='f0981eeb' id='62f1140c'/>
<typedef-decl name='__uint64_t' type-id='7359adad' id='8910171f'/>
<typedef-decl name='size_t' type-id='7359adad' id='b59d7dce'/>
<class-decl name='libzfs_handle' size-in-bits='18240' is-struct='yes' visibility='default' id='c8a9d9d8'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='libzfs_error' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='libzfs_fd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='libzfs_pool_handles' type-id='4c81de99' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='libzfs_ns_avlpool' type-id='de82c773' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='libzfs_ns_avl' type-id='a5c21a38' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='libzfs_ns_gen' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='libzfs_desc_active' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='libzfs_action' type-id='b54ce520' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8544'>
<var-decl name='libzfs_desc' type-id='b54ce520' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='16736'>
<var-decl name='libzfs_printerr' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='16768'>
<var-decl name='libzfs_mnttab_enable' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='16832'>
<var-decl name='libzfs_mnttab_cache_lock' type-id='7a6844eb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='17152'>
<var-decl name='libzfs_mnttab_cache' type-id='f20fbd51' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='17472'>
<var-decl name='libzfs_pool_iter' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='17504'>
<var-decl name='libzfs_prop_debug' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='17536'>
<var-decl name='libzfs_urire' type-id='aca3bac8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='18048'>
<var-decl name='libzfs_max_nvlist' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='18112'>
<var-decl name='libfetch' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='18176'>
<var-decl name='libfetch_load_error' type-id='26a90f95' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zfs_handle' size-in-bits='4928' is-struct='yes' visibility='default' id='f6ee4445'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zfs_hdl' type-id='b0382bb3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zpool_hdl' type-id='4c81de99' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zfs_name' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2176'>
<var-decl name='zfs_type' type-id='2e45de5d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2208'>
<var-decl name='zfs_head_type' type-id='2e45de5d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2240'>
<var-decl name='zfs_dmustats' type-id='b2c14f17' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4544'>
<var-decl name='zfs_props' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4608'>
<var-decl name='zfs_user_props' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4672'>
<var-decl name='zfs_recvd_props' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4736'>
<var-decl name='zfs_mntcheck' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4800'>
<var-decl name='zfs_mntopts' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='4864'>
<var-decl name='zfs_props_table' type-id='ae3e8ca6' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zpool_handle' size-in-bits='2560' is-struct='yes' visibility='default' id='67002a8a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zpool_hdl' type-id='b0382bb3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zpool_next' type-id='4c81de99' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zpool_name' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2176'>
<var-decl name='zpool_state' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2240'>
<var-decl name='zpool_config_size' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2304'>
<var-decl name='zpool_config' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2368'>
<var-decl name='zpool_old_config' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2432'>
<var-decl name='zpool_props' type-id='5ce45b60' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2496'>
<var-decl name='zpool_start_block' type-id='804dc465' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='0e01899c' size-in-bits='64' id='4d98cd5a'/>
<pointer-type-def type-id='428b67b3' size-in-bits='64' id='bf311473'/>
<pointer-type-def type-id='a84c031d' size-in-bits='64' id='26a90f95'/>
<pointer-type-def type-id='26a90f95' size-in-bits='64' id='9b23c9ad'/>
<qualified-type-def type-id='a84c031d' const='yes' id='9b45d938'/>
<pointer-type-def type-id='9b45d938' size-in-bits='64' id='80f4b756'/>
<qualified-type-def type-id='9155d4b5' const='yes' id='9f2c1699'/>
<pointer-type-def type-id='9f2c1699' size-in-bits='64' id='4567bbc9'/>
<qualified-type-def type-id='775509eb' const='yes' id='5eadf2db'/>
<pointer-type-def type-id='5eadf2db' size-in-bits='64' id='fcd57163'/>
<pointer-type-def type-id='96ee24a5' size-in-bits='64' id='585e1de9'/>
<pointer-type-def type-id='cb9628fa' size-in-bits='64' id='5571cde4'/>
<pointer-type-def type-id='95942d0c' size-in-bits='64' id='b0382bb3'/>
<pointer-type-def type-id='8e8d4be3' size-in-bits='64' id='5ce45b60'/>
<pointer-type-def type-id='b48d2441' size-in-bits='64' id='33976309'/>
<pointer-type-def type-id='b96825af' size-in-bits='64' id='ae3e8ca6'/>
<pointer-type-def type-id='002ac4a6' size-in-bits='64' id='cf536864'/>
<pointer-type-def type-id='5d7f5fc8' size-in-bits='64' id='813a2225'/>
<pointer-type-def type-id='73a65116' size-in-bits='64' id='2dc35b9d'/>
<pointer-type-def type-id='7f84e390' size-in-bits='64' id='de82c773'/>
<pointer-type-def type-id='bb7f0973' size-in-bits='64' id='a5c21a38'/>
<pointer-type-def type-id='edd8457b' size-in-bits='64' id='5842d146'/>
<pointer-type-def type-id='40f93560' size-in-bits='64' id='d502b39f'/>
<pointer-type-def type-id='48b5725f' size-in-bits='64' id='eaa32e2f'/>
<pointer-type-def type-id='775509eb' size-in-bits='64' id='9200a744'/>
<pointer-type-def type-id='b1efc708' size-in-bits='64' id='4c81de99'/>
<pointer-type-def type-id='a2256d42' size-in-bits='64' id='debc6aa3'/>
<class-decl name='re_dfa_t' is-struct='yes' visibility='default' is-declaration-only='yes' id='b48d2441'/>
<class-decl name='uu_avl' is-struct='yes' visibility='default' is-declaration-only='yes' id='4af029d1'/>
<class-decl name='uu_avl_pool' is-struct='yes' visibility='default' is-declaration-only='yes' id='12a530a8'/>
<class-decl name='uu_avl_walk' is-struct='yes' visibility='default' is-declaration-only='yes' id='e70a39e3'/>
<function-decl name='uu_avl_pool_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='d502b39f'/>
<parameter type-id='8f92235e'/>
<return type-id='de82c773'/>
</function-decl>
<function-decl name='uu_avl_pool_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='de82c773'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_node_init' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='2dc35b9d'/>
<parameter type-id='de82c773'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='de82c773'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='8f92235e'/>
<return type-id='a5c21a38'/>
</function-decl>
<function-decl name='uu_avl_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_last' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_walk_start' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='8f92235e'/>
<return type-id='5842d146'/>
</function-decl>
<function-decl name='uu_avl_walk_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5842d146'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_walk_end' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5842d146'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_find' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='813a2225'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_insert' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='5d7f5fc8'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='uu_avl_remove' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_get_handle' mangled-name='zfs_get_handle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_handle'>
<parameter type-id='9200a744'/>
<return type-id='b0382bb3'/>
</function-decl>
<function-decl name='zfs_open' mangled-name='zfs_open' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_open'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zfs_close' mangled-name='zfs_close' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_close'>
<parameter type-id='9200a744'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_get_name' mangled-name='zfs_get_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_name'>
<parameter type-id='fcd57163'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_prop_get' mangled-name='zfs_prop_get' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get'>
<parameter type-id='9200a744'/>
<parameter type-id='58603c44'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='debc6aa3'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='c19b74c3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_int' mangled-name='zfs_prop_get_int' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_int'>
<parameter type-id='9200a744'/>
<parameter type-id='58603c44'/>
<return type-id='9c313c2d'/>
</function-decl>
- <function-decl name='zfs_iter_children' mangled-name='zfs_iter_children' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_children'>
+ <function-decl name='zfs_iter_children_v2' mangled-name='zfs_iter_children_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_children_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
- <function-decl name='zfs_iter_dependents' mangled-name='zfs_iter_dependents' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_dependents'>
+ <function-decl name='zfs_iter_dependents_v2' mangled-name='zfs_iter_dependents_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_dependents_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_mounted' mangled-name='zfs_iter_mounted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_mounted'>
<parameter type-id='9200a744'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_refresh_properties' mangled-name='zfs_refresh_properties' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_refresh_properties'>
<parameter type-id='9200a744'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_is_mounted' mangled-name='zfs_is_mounted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_is_mounted'>
<parameter type-id='9200a744'/>
<parameter type-id='9b23c9ad'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_mount' mangled-name='zfs_mount' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_mount'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_unmount' mangled-name='zfs_unmount' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_unmount'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_is_shared' mangled-name='zfs_is_shared' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_is_shared'>
<parameter type-id='9200a744'/>
<parameter type-id='9b23c9ad'/>
<parameter type-id='4567bbc9'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_share' mangled-name='zfs_share' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_share'>
<parameter type-id='9200a744'/>
<parameter type-id='4567bbc9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_unshare' mangled-name='zfs_unshare' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_unshare'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='4567bbc9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_commit_shares' mangled-name='zfs_commit_shares' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_commit_shares'>
<parameter type-id='4567bbc9'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='getzoneid' mangled-name='getzoneid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getzoneid'>
<return type-id='4da03624'/>
</function-decl>
<function-decl name='sa_commit_shares' mangled-name='sa_commit_shares' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_commit_shares'>
<parameter type-id='9155d4b5'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strlcat' mangled-name='strlcat' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='strlcat'>
<parameter type-id='26a90f95'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='strlcpy' mangled-name='strlcpy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='strlcpy'>
<parameter type-id='26a90f95'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strcmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strncmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strlen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='zfs_error' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_alloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='remove_mountpoint' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='96ee24a5'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='add6e811'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='cb9628fa'>
<parameter type-id='9200a744'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_config.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='32768' id='d16c6df4'>
<subrange length='4096' type-id='7359adad' id='bc1b5ddc'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='65536' id='163f6aa5'>
<subrange length='8192' type-id='7359adad' id='c88f397d'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='infinite' id='e84913bd'>
<subrange length='infinite' type-id='7359adad' id='031f2035'/>
</array-type-def>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='128' id='c1c22e6c'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='24' id='d3490169'>
<subrange length='3' type-id='7359adad' id='56f209d2'/>
</array-type-def>
<type-decl name='variadic parameter type' id='2c1145c5'/>
<typedef-decl name='zpool_iter_f' type-id='3aebb66f' id='fa476e62'/>
<enum-decl name='data_type_t' naming-typedef-id='8d0687d2' id='aeeae136'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DATA_TYPE_DONTCARE' value='-1'/>
<enumerator name='DATA_TYPE_UNKNOWN' value='0'/>
<enumerator name='DATA_TYPE_BOOLEAN' value='1'/>
<enumerator name='DATA_TYPE_BYTE' value='2'/>
<enumerator name='DATA_TYPE_INT16' value='3'/>
<enumerator name='DATA_TYPE_UINT16' value='4'/>
<enumerator name='DATA_TYPE_INT32' value='5'/>
<enumerator name='DATA_TYPE_UINT32' value='6'/>
<enumerator name='DATA_TYPE_INT64' value='7'/>
<enumerator name='DATA_TYPE_UINT64' value='8'/>
<enumerator name='DATA_TYPE_STRING' value='9'/>
<enumerator name='DATA_TYPE_BYTE_ARRAY' value='10'/>
<enumerator name='DATA_TYPE_INT16_ARRAY' value='11'/>
<enumerator name='DATA_TYPE_UINT16_ARRAY' value='12'/>
<enumerator name='DATA_TYPE_INT32_ARRAY' value='13'/>
<enumerator name='DATA_TYPE_UINT32_ARRAY' value='14'/>
<enumerator name='DATA_TYPE_INT64_ARRAY' value='15'/>
<enumerator name='DATA_TYPE_UINT64_ARRAY' value='16'/>
<enumerator name='DATA_TYPE_STRING_ARRAY' value='17'/>
<enumerator name='DATA_TYPE_HRTIME' value='18'/>
<enumerator name='DATA_TYPE_NVLIST' value='19'/>
<enumerator name='DATA_TYPE_NVLIST_ARRAY' value='20'/>
<enumerator name='DATA_TYPE_BOOLEAN_VALUE' value='21'/>
<enumerator name='DATA_TYPE_INT8' value='22'/>
<enumerator name='DATA_TYPE_UINT8' value='23'/>
<enumerator name='DATA_TYPE_BOOLEAN_ARRAY' value='24'/>
<enumerator name='DATA_TYPE_INT8_ARRAY' value='25'/>
<enumerator name='DATA_TYPE_UINT8_ARRAY' value='26'/>
<enumerator name='DATA_TYPE_DOUBLE' value='27'/>
</enum-decl>
<typedef-decl name='data_type_t' type-id='aeeae136' id='8d0687d2'/>
<class-decl name='nvpair' size-in-bits='128' is-struct='yes' visibility='default' id='1c34e459'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='nvp_size' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='nvp_name_sz' type-id='23bd8cb5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='48'>
<var-decl name='nvp_reserve' type-id='23bd8cb5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='nvp_value_elem' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='nvp_type' type-id='8d0687d2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='nvp_name' type-id='e84913bd' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='nvpair_t' type-id='1c34e459' id='57928edf'/>
<class-decl name='drr_begin' size-in-bits='2432' is-struct='yes' visibility='default' id='09fcdc01'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_magic' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_versioninfo' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_creation_time' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_type' type-id='230f1e16' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='drr_flags' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_fromguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_toname' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zinject_record' size-in-bits='2816' is-struct='yes' visibility='default' id='3216f820'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zi_objset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zi_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zi_start' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='zi_end' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='zi_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='zi_level' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='zi_error' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='zi_type' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='zi_freq' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='zi_failfast' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='zi_func' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2560'>
<var-decl name='zi_iotype' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2592'>
<var-decl name='zi_duration' type-id='3ff5601b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2624'>
<var-decl name='zi_timer' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2688'>
<var-decl name='zi_nlanes' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2752'>
<var-decl name='zi_cmd' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2784'>
<var-decl name='zi_dvas' type-id='8f92235e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zinject_record_t' type-id='3216f820' id='a4301ca6'/>
<class-decl name='zfs_share' size-in-bits='256' is-struct='yes' visibility='default' id='feb6f2da'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='z_exportdata' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='z_sharedata' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='z_sharetype' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='z_sharemax' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_share_t' type-id='feb6f2da' id='ee5cec36'/>
<class-decl name='zfs_cmd' size-in-bits='109952' is-struct='yes' visibility='default' id='3522cd69'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zc_name' type-id='d16c6df4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32768'>
<var-decl name='zc_nvlist_src' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32832'>
<var-decl name='zc_nvlist_src_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32896'>
<var-decl name='zc_nvlist_dst' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32960'>
<var-decl name='zc_nvlist_dst_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33024'>
<var-decl name='zc_nvlist_dst_filled' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33056'>
<var-decl name='zc_pad2' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33088'>
<var-decl name='zc_history' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='33152'>
<var-decl name='zc_value' type-id='163f6aa5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='98688'>
<var-decl name='zc_string' type-id='d1617432' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100736'>
<var-decl name='zc_guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100800'>
<var-decl name='zc_nvlist_conf' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100864'>
<var-decl name='zc_nvlist_conf_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100928'>
<var-decl name='zc_cookie' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='100992'>
<var-decl name='zc_objset_type' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101056'>
<var-decl name='zc_perm_action' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101120'>
<var-decl name='zc_history_len' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101184'>
<var-decl name='zc_history_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101248'>
<var-decl name='zc_obj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101312'>
<var-decl name='zc_iflags' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101376'>
<var-decl name='zc_share' type-id='ee5cec36' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='101632'>
<var-decl name='zc_objset_stats' type-id='b2c14f17' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='103936'>
<var-decl name='zc_begin_record' type-id='09fcdc01' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='106368'>
<var-decl name='zc_inject_record' type-id='a4301ca6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109184'>
<var-decl name='zc_defer_destroy' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109216'>
<var-decl name='zc_flags' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109248'>
<var-decl name='zc_action_handle' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109312'>
<var-decl name='zc_cleanup_fd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109344'>
<var-decl name='zc_simple' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109352'>
<var-decl name='zc_pad' type-id='d3490169' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109376'>
<var-decl name='zc_sendobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109440'>
<var-decl name='zc_fromobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109504'>
<var-decl name='zc_createtxg' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109568'>
<var-decl name='zc_stat' type-id='0371a9c7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='109888'>
<var-decl name='zc_zoneid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_cmd_t' type-id='3522cd69' id='a5559cdd'/>
<class-decl name='zfs_stat' size-in-bits='320' is-struct='yes' visibility='default' id='6417f0b9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zs_gen' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zs_mode' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zs_links' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='zs_ctime' type-id='c1c22e6c' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_stat_t' type-id='6417f0b9' id='0371a9c7'/>
<typedef-decl name='int16_t' type-id='03896e23' id='23bd8cb5'/>
<typedef-decl name='__int16_t' type-id='a2185560' id='03896e23'/>
<pointer-type-def type-id='c19b74c3' size-in-bits='64' id='37e3bd22'/>
<qualified-type-def type-id='8e8d4be3' const='yes' id='693c3853'/>
<pointer-type-def type-id='693c3853' size-in-bits='64' id='22cce67b'/>
<qualified-type-def type-id='57928edf' const='yes' id='642ee20f'/>
<pointer-type-def type-id='642ee20f' size-in-bits='64' id='dace003f'/>
<pointer-type-def type-id='2bce87e3' size-in-bits='64' id='3aebb66f'/>
<pointer-type-def type-id='95e97e5e' size-in-bits='64' id='7292109c'/>
<pointer-type-def type-id='5ce45b60' size-in-bits='64' id='857bb57e'/>
<pointer-type-def type-id='57928edf' size-in-bits='64' id='3fa542f0'/>
<pointer-type-def type-id='eaa32e2f' size-in-bits='64' id='63e171df'/>
<pointer-type-def type-id='3522cd69' size-in-bits='64' id='b65f7fd1'/>
<pointer-type-def type-id='a5559cdd' size-in-bits='64' id='e4ec4540'/>
<pointer-type-def type-id='4c81de99' size-in-bits='64' id='237193c9'/>
<function-decl name='uu_avl_first' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='eaa32e2f'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='uu_avl_teardown' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a5c21a38'/>
<parameter type-id='63e171df'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='zfs_standard_error' mangled-name='zfs_standard_error' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_standard_error'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_ioctl' mangled-name='zfs_ioctl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_ioctl'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='b65f7fd1'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='nvlist_dup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='857bb57e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_exists' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nvlist_next_nvpair' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='dace003f'/>
<return type-id='3fa542f0'/>
</function-decl>
<function-decl name='nvpair_name' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='fnvpair_value_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3fa542f0'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='libspl_assertf' mangled-name='libspl_assertf' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libspl_assertf'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__errno_location' visibility='default' binding='global' size-in-bits='64'>
<return type-id='7292109c'/>
</function-decl>
<function-decl name='dcgettext' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='getenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strchr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zpool_get_config' mangled-name='zpool_get_config' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_config'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='857bb57e' name='oldconfig'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_get_features' mangled-name='zpool_get_features' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_features'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_refresh_stats' mangled-name='zpool_refresh_stats' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_refresh_stats'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='37e3bd22' name='missing'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_skip_pool' mangled-name='zpool_skip_pool' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_skip_pool'>
<parameter type-id='80f4b756' name='poolname'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_iter' mangled-name='zpool_iter' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_iter'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='fa476e62' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_root' mangled-name='zfs_iter_root' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_root'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_strdup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='no_memory' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zcmd_alloc_dst_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zcmd_expand_dst_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zcmd_read_dst_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zcmd_free_nvlists' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e4ec4540'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='make_dataset_handle' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zpool_open_silent' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='237193c9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='2bce87e3'>
<parameter type-id='4c81de99'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_crypto.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='fb7c6451' size-in-bits='256' id='64177143'>
<subrange length='32' type-id='7359adad' id='ae5bde82'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='8' id='89feb1ec'>
<subrange length='1' type-id='7359adad' id='52f813b4'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='160' id='664ac0b7'>
<subrange length='20' type-id='7359adad' id='fdca39cf'/>
</array-type-def>
<class-decl name='_IO_codecvt' is-struct='yes' visibility='default' is-declaration-only='yes' id='a4036571'/>
<class-decl name='_IO_marker' is-struct='yes' visibility='default' is-declaration-only='yes' id='010ae0b9'/>
<class-decl name='_IO_wide_data' is-struct='yes' visibility='default' is-declaration-only='yes' id='79bd3751'/>
<array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='896' id='47394ee0'>
<subrange length='28' type-id='7359adad' id='3db583d7'/>
</array-type-def>
<type-decl name='signed char' size-in-bits='8' id='28577a57'/>
<array-type-def dimensions='1' type-id='7359adad' size-in-bits='1024' id='d2baa450'>
<subrange length='16' type-id='7359adad' id='848d0938'/>
</array-type-def>
<type-decl name='unsigned short int' size-in-bits='16' id='8efea9e5'/>
<enum-decl name='zpool_prop_t' naming-typedef-id='5d0c23fb' id='af1ba157'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_PROP_INVAL' value='-1'/>
<enumerator name='ZPOOL_PROP_NAME' value='0'/>
<enumerator name='ZPOOL_PROP_SIZE' value='1'/>
<enumerator name='ZPOOL_PROP_CAPACITY' value='2'/>
<enumerator name='ZPOOL_PROP_ALTROOT' value='3'/>
<enumerator name='ZPOOL_PROP_HEALTH' value='4'/>
<enumerator name='ZPOOL_PROP_GUID' value='5'/>
<enumerator name='ZPOOL_PROP_VERSION' value='6'/>
<enumerator name='ZPOOL_PROP_BOOTFS' value='7'/>
<enumerator name='ZPOOL_PROP_DELEGATION' value='8'/>
<enumerator name='ZPOOL_PROP_AUTOREPLACE' value='9'/>
<enumerator name='ZPOOL_PROP_CACHEFILE' value='10'/>
<enumerator name='ZPOOL_PROP_FAILUREMODE' value='11'/>
<enumerator name='ZPOOL_PROP_LISTSNAPS' value='12'/>
<enumerator name='ZPOOL_PROP_AUTOEXPAND' value='13'/>
<enumerator name='ZPOOL_PROP_DEDUPDITTO' value='14'/>
<enumerator name='ZPOOL_PROP_DEDUPRATIO' value='15'/>
<enumerator name='ZPOOL_PROP_FREE' value='16'/>
<enumerator name='ZPOOL_PROP_ALLOCATED' value='17'/>
<enumerator name='ZPOOL_PROP_READONLY' value='18'/>
<enumerator name='ZPOOL_PROP_ASHIFT' value='19'/>
<enumerator name='ZPOOL_PROP_COMMENT' value='20'/>
<enumerator name='ZPOOL_PROP_EXPANDSZ' value='21'/>
<enumerator name='ZPOOL_PROP_FREEING' value='22'/>
<enumerator name='ZPOOL_PROP_FRAGMENTATION' value='23'/>
<enumerator name='ZPOOL_PROP_LEAKED' value='24'/>
<enumerator name='ZPOOL_PROP_MAXBLOCKSIZE' value='25'/>
<enumerator name='ZPOOL_PROP_TNAME' value='26'/>
<enumerator name='ZPOOL_PROP_MAXDNODESIZE' value='27'/>
<enumerator name='ZPOOL_PROP_MULTIHOST' value='28'/>
<enumerator name='ZPOOL_PROP_CHECKPOINT' value='29'/>
<enumerator name='ZPOOL_PROP_LOAD_GUID' value='30'/>
<enumerator name='ZPOOL_PROP_AUTOTRIM' value='31'/>
<enumerator name='ZPOOL_PROP_COMPATIBILITY' value='32'/>
<enumerator name='ZPOOL_PROP_BCLONEUSED' value='33'/>
<enumerator name='ZPOOL_PROP_BCLONESAVED' value='34'/>
<enumerator name='ZPOOL_PROP_BCLONERATIO' value='35'/>
<enumerator name='ZPOOL_NUM_PROPS' value='36'/>
</enum-decl>
<typedef-decl name='zpool_prop_t' type-id='af1ba157' id='5d0c23fb'/>
<typedef-decl name='regoff_t' type-id='95e97e5e' id='54a2a2a8'/>
<class-decl name='regmatch_t' size-in-bits='64' is-struct='yes' naming-typedef-id='1b941664' visibility='default' id='4f932615'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='rm_so' type-id='54a2a2a8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='rm_eo' type-id='54a2a2a8' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='regmatch_t' type-id='4f932615' id='1b941664'/>
<typedef-decl name='__sighandler_t' type-id='03347643' id='8cdd9566'/>
<typedef-decl name='ssize_t' type-id='41060289' id='79a0948f'/>
<class-decl name='sigaction' size-in-bits='1216' is-struct='yes' visibility='default' id='fe391c48'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__sigaction_handler' type-id='ac5ab595' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='sa_mask' type-id='b9c97942' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1088'>
<var-decl name='sa_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1152'>
<var-decl name='sa_restorer' type-id='953b12f8' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='__anonymous_union__' size-in-bits='64' is-anonymous='yes' visibility='default' id='ac5ab595'>
<data-member access='public'>
<var-decl name='sa_handler' type-id='8cdd9566' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='sa_sigaction' type-id='6e756877' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='termios' size-in-bits='480' is-struct='yes' visibility='default' id='ad55d2bc'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='c_iflag' type-id='241ce6f8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='c_oflag' type-id='241ce6f8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='c_cflag' type-id='241ce6f8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='c_lflag' type-id='241ce6f8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='c_line' type-id='fb7c6451' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='136'>
<var-decl name='c_cc' type-id='64177143' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='416'>
<var-decl name='c_ispeed' type-id='6a8e8a14' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='c_ospeed' type-id='6a8e8a14' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='cc_t' type-id='002ac4a6' id='fb7c6451'/>
<typedef-decl name='speed_t' type-id='f0981eeb' id='6a8e8a14'/>
<typedef-decl name='tcflag_t' type-id='f0981eeb' id='241ce6f8'/>
<typedef-decl name='__uid_t' type-id='f0981eeb' id='cc5fcceb'/>
<typedef-decl name='__off_t' type-id='bd54fe1a' id='79989e9c'/>
<typedef-decl name='__off64_t' type-id='bd54fe1a' id='724e4de6'/>
<typedef-decl name='__pid_t' type-id='95e97e5e' id='3629bad8'/>
<typedef-decl name='__clock_t' type-id='bd54fe1a' id='4d66c6d7'/>
<typedef-decl name='__ssize_t' type-id='bd54fe1a' id='41060289'/>
<typedef-decl name='FILE' type-id='ec1ed955' id='aa12d1ba'/>
<class-decl name='__sigset_t' size-in-bits='1024' is-struct='yes' naming-typedef-id='b9c97942' visibility='default' id='2616147f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__val' type-id='d2baa450' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__sigset_t' type-id='2616147f' id='b9c97942'/>
<union-decl name='sigval' size-in-bits='64' visibility='default' id='a094b870'>
<data-member access='public'>
<var-decl name='sival_int' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='sival_ptr' type-id='eaa32e2f' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='__sigval_t' type-id='a094b870' id='eabacd01'/>
<class-decl name='siginfo_t' size-in-bits='1024' is-struct='yes' naming-typedef-id='cb681f62' visibility='default' id='d8149419'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_signo' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_errno' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_code' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='__pad0' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='_sifields' type-id='ac5ab596' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='__anonymous_union__1' size-in-bits='896' is-anonymous='yes' visibility='default' id='ac5ab596'>
<data-member access='public'>
<var-decl name='_pad' type-id='47394ee0' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_kill' type-id='e7f43f73' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_timer' type-id='e7f43f74' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_rt' type-id='e7f43f75' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigchld' type-id='e7f43f76' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigfault' type-id='e7f43f77' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigpoll' type-id='e7f43f78' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigsys' type-id='e7f43f79' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='__anonymous_struct__1' size-in-bits='64' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f73'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_pid' type-id='3629bad8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__2' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f74'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_tid' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_overrun' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_sigval' type-id='eabacd01' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__3' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f75'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_pid' type-id='3629bad8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_sigval' type-id='eabacd01' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__4' size-in-bits='256' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f76'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_pid' type-id='3629bad8' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='si_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_status' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='si_utime' type-id='4d66c6d7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='si_stime' type-id='4d66c6d7' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__5' size-in-bits='256' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f77'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_addr' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_addr_lsb' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='_bounds' type-id='ac5ab597' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='__anonymous_union__2' size-in-bits='128' is-anonymous='yes' visibility='default' id='ac5ab597'>
<data-member access='public'>
<var-decl name='_addr_bnd' type-id='e7f43f7a' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_pkey' type-id='62f1140c' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='__anonymous_struct__6' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f7a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='_lower' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='_upper' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__7' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f78'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='si_band' type-id='bd54fe1a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='si_fd' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='__anonymous_struct__8' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f79'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='_call_addr' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='_syscall' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='_arch' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='siginfo_t' type-id='d8149419' id='cb681f62'/>
<typedef-decl name='sigset_t' type-id='b9c97942' id='daf33c64'/>
<typedef-decl name='_IO_lock_t' type-id='48b5725f' id='bb4788fa'/>
<class-decl name='_IO_FILE' size-in-bits='1728' is-struct='yes' visibility='default' id='ec1ed955'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='_IO_read_ptr' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='_IO_read_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='_IO_read_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='_IO_write_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='_IO_write_ptr' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='_IO_write_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='_IO_buf_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='_IO_buf_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='_IO_save_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='640'>
<var-decl name='_IO_backup_base' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='_IO_save_end' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='768'>
<var-decl name='_markers' type-id='e4c6fa61' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='_chain' type-id='dca988a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='896'>
<var-decl name='_fileno' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='928'>
<var-decl name='_flags2' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='_old_offset' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1024'>
<var-decl name='_cur_column' type-id='8efea9e5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1040'>
<var-decl name='_vtable_offset' type-id='28577a57' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1048'>
<var-decl name='_shortbuf' type-id='89feb1ec' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1088'>
<var-decl name='_lock' type-id='cecf4ea7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1152'>
<var-decl name='_offset' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1216'>
<var-decl name='_codecvt' type-id='570f8c59' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1280'>
<var-decl name='_wide_data' type-id='c65a1f29' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1344'>
<var-decl name='_freeres_list' type-id='dca988a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1408'>
<var-decl name='_freeres_buf' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1472'>
<var-decl name='__pad5' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1536'>
<var-decl name='_mode' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1568'>
<var-decl name='_unused2' type-id='664ac0b7' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='aa12d1ba' size-in-bits='64' id='822cd80b'/>
<qualified-type-def type-id='822cd80b' restrict='yes' id='e75a27e9'/>
<pointer-type-def type-id='ec1ed955' size-in-bits='64' id='dca988a5'/>
<pointer-type-def type-id='a4036571' size-in-bits='64' id='570f8c59'/>
<pointer-type-def type-id='bb4788fa' size-in-bits='64' id='cecf4ea7'/>
<pointer-type-def type-id='010ae0b9' size-in-bits='64' id='e4c6fa61'/>
<pointer-type-def type-id='79bd3751' size-in-bits='64' id='c65a1f29'/>
<qualified-type-def type-id='9b23c9ad' restrict='yes' id='8c85230f'/>
<qualified-type-def type-id='80f4b756' restrict='yes' id='9d26089a'/>
<pointer-type-def type-id='80f4b756' size-in-bits='64' id='7d3cd834'/>
<qualified-type-def type-id='aca3bac8' const='yes' id='2498fd78'/>
<pointer-type-def type-id='2498fd78' size-in-bits='64' id='eed6c816'/>
<qualified-type-def type-id='eed6c816' restrict='yes' id='a431a9da'/>
<qualified-type-def type-id='fe391c48' const='yes' id='14a93b33'/>
<pointer-type-def type-id='14a93b33' size-in-bits='64' id='9f68085b'/>
<qualified-type-def type-id='9f68085b' restrict='yes' id='e2a5e6f9'/>
<qualified-type-def type-id='ad55d2bc' const='yes' id='a46bf13f'/>
<pointer-type-def type-id='a46bf13f' size-in-bits='64' id='eaec840f'/>
<qualified-type-def type-id='002ac4a6' const='yes' id='ea86de29'/>
<pointer-type-def type-id='ea86de29' size-in-bits='64' id='354f7eb9'/>
<qualified-type-def type-id='8efea9e5' const='yes' id='3beb2af4'/>
<pointer-type-def type-id='3beb2af4' size-in-bits='64' id='31347b7a'/>
<pointer-type-def type-id='31347b7a' size-in-bits='64' id='c59e1ef0'/>
<pointer-type-def type-id='1b941664' size-in-bits='64' id='7e2979d5'/>
<qualified-type-def type-id='7e2979d5' restrict='yes' id='fc212857'/>
<pointer-type-def type-id='fe391c48' size-in-bits='64' id='568dd84e'/>
<qualified-type-def type-id='568dd84e' restrict='yes' id='3d8ee6f2'/>
<pointer-type-def type-id='cb681f62' size-in-bits='64' id='185869c1'/>
<pointer-type-def type-id='daf33c64' size-in-bits='64' id='9e80f729'/>
<pointer-type-def type-id='b59d7dce' size-in-bits='64' id='78c01427'/>
<qualified-type-def type-id='78c01427' restrict='yes' id='d19b2c25'/>
<pointer-type-def type-id='ad55d2bc' size-in-bits='64' id='665a4eda'/>
<pointer-type-def type-id='9c313c2d' size-in-bits='64' id='5d6479ae'/>
<pointer-type-def type-id='ae3e8ca6' size-in-bits='64' id='d8774064'/>
<pointer-type-def type-id='3502e3ff' size-in-bits='64' id='4dd26a40'/>
<pointer-type-def type-id='ee076206' size-in-bits='64' id='953b12f8'/>
<pointer-type-def type-id='f712e2b7' size-in-bits='64' id='03347643'/>
<pointer-type-def type-id='ef70d893' size-in-bits='64' id='6e756877'/>
<qualified-type-def type-id='eaa32e2f' restrict='yes' id='1b7446cd'/>
<class-decl name='_IO_codecvt' is-struct='yes' visibility='default' is-declaration-only='yes' id='a4036571'/>
<class-decl name='_IO_marker' is-struct='yes' visibility='default' is-declaration-only='yes' id='010ae0b9'/>
<class-decl name='_IO_wide_data' is-struct='yes' visibility='default' is-declaration-only='yes' id='79bd3751'/>
<function-decl name='zpool_get_prop_int' mangled-name='zpool_get_prop_int' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_prop_int'>
<parameter type-id='4c81de99'/>
<parameter type-id='5d0c23fb'/>
<parameter type-id='debc6aa3'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_handle_dup' mangled-name='zfs_handle_dup' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_handle_dup'>
<parameter type-id='9200a744'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zfs_valid_proplist' mangled-name='zfs_valid_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_valid_proplist'>
<parameter type-id='b0382bb3'/>
<parameter type-id='2e45de5d'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9200a744'/>
<parameter type-id='4c81de99'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='80f4b756'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_prop_to_name' mangled-name='zfs_prop_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_to_name'>
<parameter type-id='58603c44'/>
<return type-id='80f4b756'/>
</function-decl>
- <function-decl name='zfs_iter_filesystems' mangled-name='zfs_iter_filesystems' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_filesystems'>
+ <function-decl name='zfs_iter_filesystems_v2' mangled-name='zfs_iter_filesystems_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_filesystems_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_parent_name' mangled-name='zfs_parent_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_parent_name'>
<parameter type-id='9200a744'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_load_key' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_unload_key' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_change_key' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_name_to_prop' mangled-name='zfs_name_to_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_name_to_prop'>
<parameter type-id='80f4b756'/>
<return type-id='58603c44'/>
</function-decl>
<function-decl name='nvlist_add_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fnvlist_alloc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='__ctype_b_loc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='c59e1ef0'/>
</function-decl>
<function-decl name='dlopen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='dlsym' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='1b7446cd'/>
<parameter type-id='9d26089a'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='dlerror' visibility='default' binding='global' size-in-bits='64'>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='PKCS5_PBKDF2_HMAC_SHA1' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='354f7eb9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='cf536864'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='regexec' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a431a9da'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='fc212857'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='kill' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3629bad8'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sigemptyset' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9e80f729'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sigaction' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='e2a5e6f9'/>
<parameter type-id='3d8ee6f2'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fclose' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fflush' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fdopen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<return type-id='822cd80b'/>
</function-decl>
<function-decl name='fputc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__getdelim' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='8c85230f'/>
<parameter type-id='d19b2c25'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='e75a27e9'/>
<return type-id='41060289'/>
</function-decl>
<function-decl name='rewind' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='ferror' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fileno' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='malloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='calloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='strdup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strerror' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='tcgetattr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='665a4eda'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='tcsetattr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaec840f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='close' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getpid' visibility='default' binding='global' size-in-bits='64'>
<return type-id='3629bad8'/>
</function-decl>
<function-decl name='isatty' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='unlink' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__open_too_many_args' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__open_missing_mode' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__printf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__asprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__fread_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='1b7446cd'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='e75a27e9'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='__read_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='zfs_crypto_get_encryption_root' mangled-name='zfs_crypto_get_encryption_root' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_get_encryption_root'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='37e3bd22' name='is_encroot'/>
<parameter type-id='26a90f95' name='buf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_create' mangled-name='zfs_crypto_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_create'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='parent_name'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='5ce45b60' name='pool_props'/>
<parameter type-id='c19b74c3' name='stdin_available'/>
<parameter type-id='d8774064' name='wkeydata_out'/>
<parameter type-id='4dd26a40' name='wkeylen_out'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_clone_check' mangled-name='zfs_crypto_clone_check' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_clone_check'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='9200a744' name='origin_zhp'/>
<parameter type-id='26a90f95' name='parent_name'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_attempt_load_keys' mangled-name='zfs_crypto_attempt_load_keys' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_attempt_load_keys'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='fsname'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_load_key' mangled-name='zfs_crypto_load_key' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_load_key'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='c19b74c3' name='noop'/>
<parameter type-id='80f4b756' name='alt_keylocation'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_unload_key' mangled-name='zfs_crypto_unload_key' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_unload_key'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_crypto_rewrap' mangled-name='zfs_crypto_rewrap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_crypto_rewrap'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='5ce45b60' name='raw_props'/>
<parameter type-id='c19b74c3' name='inheritkey'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_error_aux' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='f712e2b7'>
<parameter type-id='95e97e5e'/>
<return type-id='48b5725f'/>
</function-type>
<function-type size-in-bits='64' id='ef70d893'>
<parameter type-id='95e97e5e'/>
<parameter type-id='185869c1'/>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_dataset.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='32' id='8e0573fd'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<class-decl name='prop_changelist' is-struct='yes' visibility='default' is-declaration-only='yes' id='d86edc51'/>
<class-decl name='zprop_list' size-in-bits='448' is-struct='yes' visibility='default' id='bd9b4291'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pl_prop' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pl_user_prop' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='pl_next' type-id='9f1a1109' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='pl_all' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='pl_width' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='pl_recvd_width' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='pl_fixed' type-id='c19b74c3' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zprop_list_t' type-id='bd9b4291' id='bdb8ac4f'/>
<class-decl name='renameflags' size-in-bits='32' is-struct='yes' visibility='default' id='7aee5792'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='recursive' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1'>
<var-decl name='nounmount' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2'>
<var-decl name='forceunmount' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='renameflags_t' type-id='7aee5792' id='067170c2'/>
<typedef-decl name='zfs_userspace_cb_t' type-id='ca64ff60' id='16c5f410'/>
<enum-decl name='lzc_dataset_type' id='bc9887f1'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='LZC_DATSET_TYPE_ZFS' value='2'/>
<enumerator name='LZC_DATSET_TYPE_ZVOL' value='3'/>
</enum-decl>
<typedef-decl name='avl_index_t' type-id='e475ab95' id='fba6cb51'/>
<enum-decl name='zfs_userquota_prop_t' naming-typedef-id='279fde6a' id='5258d2f6'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_PROP_USERUSED' value='0'/>
<enumerator name='ZFS_PROP_USERQUOTA' value='1'/>
<enumerator name='ZFS_PROP_GROUPUSED' value='2'/>
<enumerator name='ZFS_PROP_GROUPQUOTA' value='3'/>
<enumerator name='ZFS_PROP_USEROBJUSED' value='4'/>
<enumerator name='ZFS_PROP_USEROBJQUOTA' value='5'/>
<enumerator name='ZFS_PROP_GROUPOBJUSED' value='6'/>
<enumerator name='ZFS_PROP_GROUPOBJQUOTA' value='7'/>
<enumerator name='ZFS_PROP_PROJECTUSED' value='8'/>
<enumerator name='ZFS_PROP_PROJECTQUOTA' value='9'/>
<enumerator name='ZFS_PROP_PROJECTOBJUSED' value='10'/>
<enumerator name='ZFS_PROP_PROJECTOBJQUOTA' value='11'/>
<enumerator name='ZFS_NUM_USERQUOTA_PROPS' value='12'/>
</enum-decl>
<typedef-decl name='zfs_userquota_prop_t' type-id='5258d2f6' id='279fde6a'/>
<enum-decl name='zfs_wait_activity_t' naming-typedef-id='3024501a' id='527d5dc6'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_WAIT_DELETEQ' value='0'/>
<enumerator name='ZFS_WAIT_NUM_ACTIVITIES' value='1'/>
</enum-decl>
<typedef-decl name='zfs_wait_activity_t' type-id='527d5dc6' id='3024501a'/>
<enum-decl name='namecheck_err_t' naming-typedef-id='8e0af06e' id='f43bbcda'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='NAME_ERR_LEADING_SLASH' value='0'/>
<enumerator name='NAME_ERR_EMPTY_COMPONENT' value='1'/>
<enumerator name='NAME_ERR_TRAILING_SLASH' value='2'/>
<enumerator name='NAME_ERR_INVALCHAR' value='3'/>
<enumerator name='NAME_ERR_MULTIPLE_DELIMITERS' value='4'/>
<enumerator name='NAME_ERR_NOLETTER' value='5'/>
<enumerator name='NAME_ERR_RESERVED' value='6'/>
<enumerator name='NAME_ERR_DISKLIKE' value='7'/>
<enumerator name='NAME_ERR_TOOLONG' value='8'/>
<enumerator name='NAME_ERR_SELF_REF' value='9'/>
<enumerator name='NAME_ERR_PARENT_REF' value='10'/>
<enumerator name='NAME_ERR_NO_AT' value='11'/>
<enumerator name='NAME_ERR_NO_POUND' value='12'/>
</enum-decl>
<typedef-decl name='namecheck_err_t' type-id='f43bbcda' id='8e0af06e'/>
<enum-decl name='zprop_type_t' naming-typedef-id='31429eff' id='87676253'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='PROP_TYPE_NUMBER' value='0'/>
<enumerator name='PROP_TYPE_STRING' value='1'/>
<enumerator name='PROP_TYPE_INDEX' value='2'/>
</enum-decl>
<typedef-decl name='zprop_type_t' type-id='87676253' id='31429eff'/>
<class-decl name='mnttab' size-in-bits='256' is-struct='yes' visibility='default' id='1b055409'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_special' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_mountp' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_fstype' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_mntopts' type-id='26a90f95' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='group' size-in-bits='256' is-struct='yes' visibility='default' id='01a1b934'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='gr_name' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='gr_passwd' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='gr_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='gr_mem' type-id='9b23c9ad' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='mntent' size-in-bits='320' is-struct='yes' visibility='default' id='56fe4a37'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_fsname' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_dir' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_type' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_opts' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='mnt_freq' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='mnt_passno' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='passwd' size-in-bits='384' is-struct='yes' visibility='default' id='a63d15a3'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pw_name' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pw_passwd' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='pw_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='pw_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='pw_gecos' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='pw_dir' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='pw_shell' type-id='26a90f95' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='pthread_mutexattr_t' size-in-bits='32' naming-typedef-id='8afd6070' visibility='default' id='7300eb00'>
<data-member access='public'>
<var-decl name='__size' type-id='8e0573fd' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='95e97e5e' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_mutexattr_t' type-id='7300eb00' id='8afd6070'/>
<typedef-decl name='int64_t' type-id='0c9942d2' id='9da381c4'/>
<typedef-decl name='__int64_t' type-id='bd54fe1a' id='0c9942d2'/>
<typedef-decl name='__gid_t' type-id='f0981eeb' id='d94ec6d9'/>
<typedef-decl name='__time_t' type-id='bd54fe1a' id='65eda9c0'/>
<class-decl name='tm' size-in-bits='448' is-struct='yes' visibility='default' id='dddf6ca2'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tm_sec' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='tm_min' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tm_hour' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='tm_mday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tm_mon' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='tm_year' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='tm_wday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='tm_yday' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='tm_isdst' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='tm_gmtoff' type-id='bd54fe1a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='tm_zone' type-id='80f4b756' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='time_t' type-id='65eda9c0' id='c9d12d66'/>
<typedef-decl name='uid_t' type-id='cc5fcceb' id='354978ed'/>
<typedef-decl name='prop_changelist_t' type-id='d86edc51' id='eae6431d'/>
<pointer-type-def type-id='fba6cb51' size-in-bits='64' id='32adbf30'/>
<pointer-type-def type-id='f20fbd51' size-in-bits='64' id='a3681dea'/>
<qualified-type-def type-id='26a90f95' restrict='yes' id='266fe297'/>
<qualified-type-def type-id='56fe4a37' const='yes' id='a75125ce'/>
<pointer-type-def type-id='a75125ce' size-in-bits='64' id='48bea5ec'/>
<qualified-type-def type-id='8afd6070' const='yes' id='1d853360'/>
<pointer-type-def type-id='1d853360' size-in-bits='64' id='c2afbd7e'/>
<qualified-type-def type-id='c9d12d66' const='yes' id='588b3216'/>
<pointer-type-def type-id='588b3216' size-in-bits='64' id='9f201474'/>
<qualified-type-def type-id='9f201474' restrict='yes' id='d6e2847c'/>
<qualified-type-def type-id='dddf6ca2' const='yes' id='e824a34f'/>
<pointer-type-def type-id='e824a34f' size-in-bits='64' id='d6ad37ff'/>
<qualified-type-def type-id='d6ad37ff' restrict='yes' id='f8c6051d'/>
<qualified-type-def type-id='9c313c2d' const='yes' id='c3b7ba7d'/>
<pointer-type-def type-id='c3b7ba7d' size-in-bits='64' id='713a56f5'/>
<pointer-type-def type-id='01a1b934' size-in-bits='64' id='566b3f52'/>
<pointer-type-def type-id='7e291ce6' size-in-bits='64' id='ca64ff60'/>
<pointer-type-def type-id='9da381c4' size-in-bits='64' id='cb785ebf'/>
<pointer-type-def type-id='1b055409' size-in-bits='64' id='9d424d31'/>
<pointer-type-def type-id='8e0af06e' size-in-bits='64' id='053457bd'/>
<pointer-type-def type-id='857bb57e' size-in-bits='64' id='75be733c'/>
<pointer-type-def type-id='a63d15a3' size-in-bits='64' id='a195f4a3'/>
<pointer-type-def type-id='eae6431d' size-in-bits='64' id='0d41d328'/>
<pointer-type-def type-id='7a6844eb' size-in-bits='64' id='18c91f9e'/>
<pointer-type-def type-id='dddf6ca2' size-in-bits='64' id='d915a820'/>
<qualified-type-def type-id='d915a820' restrict='yes' id='f099ad08'/>
<pointer-type-def type-id='5d6479ae' size-in-bits='64' id='892b4acc'/>
<pointer-type-def type-id='bd9b4291' size-in-bits='64' id='9f1a1109'/>
<pointer-type-def type-id='bdb8ac4f' size-in-bits='64' id='3a9b2288'/>
<pointer-type-def type-id='3a9b2288' size-in-bits='64' id='e4378506'/>
<class-decl name='prop_changelist' is-struct='yes' visibility='default' is-declaration-only='yes' id='d86edc51'/>
<function-decl name='zpool_open' mangled-name='zpool_open' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_open'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<return type-id='4c81de99'/>
</function-decl>
<function-decl name='zpool_open_canfail' mangled-name='zpool_open_canfail' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_open_canfail'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<return type-id='4c81de99'/>
</function-decl>
<function-decl name='zpool_close' mangled-name='zpool_close' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_close'>
<parameter type-id='4c81de99'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_get_name' mangled-name='zpool_get_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_name'>
<parameter type-id='4c81de99'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_get_prop' mangled-name='zpool_get_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_prop'>
<parameter type-id='4c81de99'/>
<parameter type-id='5d0c23fb'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='debc6aa3'/>
<parameter type-id='c19b74c3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_default_string' mangled-name='zfs_prop_default_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_default_string'>
<parameter type-id='58603c44'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_prop_default_numeric' mangled-name='zfs_prop_default_numeric' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_default_numeric'>
<parameter type-id='58603c44'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zpool_prop_get_feature' mangled-name='zpool_prop_get_feature' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_get_feature'>
<parameter type-id='4c81de99'/>
<parameter type-id='80f4b756'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
- <function-decl name='zfs_iter_snapshots' mangled-name='zfs_iter_snapshots' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots'>
+ <function-decl name='zfs_iter_snapshots_v2' mangled-name='zfs_iter_snapshots_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<return type-id='95e97e5e'/>
</function-decl>
- <function-decl name='zfs_iter_bookmarks' mangled-name='zfs_iter_bookmarks' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_bookmarks'>
+ <function-decl name='zfs_iter_bookmarks_v2' mangled-name='zfs_iter_bookmarks_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_bookmarks_v2'>
<parameter type-id='9200a744'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='d8e49ab9'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_destroy_snaps_nvl_os' mangled-name='zfs_destroy_snaps_nvl_os' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_destroy_snaps_nvl_os'>
<parameter type-id='b0382bb3'/>
<parameter type-id='5ce45b60'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_nicestrtonum' mangled-name='zfs_nicestrtonum' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicestrtonum'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_snapshot' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='bc9887f1'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_clone' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_promote' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='26a90f95'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy_snaps' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_bookmarks' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy_bookmarks' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_hold' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_release' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_holds' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_exists' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='lzc_rollback_to' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_channel_program_nosync' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait_fs' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='3024501a'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_nicebytes' mangled-name='zfs_nicebytes' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicebytes'>
<parameter type-id='9c313c2d'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_nicenum' mangled-name='zfs_nicenum' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicenum'>
<parameter type-id='9c313c2d'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_create' mangled-name='avl_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_create'>
<parameter type-id='a3681dea'/>
<parameter type-id='585e1de9'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_find' mangled-name='avl_find' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_find'>
<parameter type-id='a3681dea'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='32adbf30'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_add' mangled-name='avl_add' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_add'>
<parameter type-id='a3681dea'/>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_remove' mangled-name='avl_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_remove'>
<parameter type-id='a3681dea'/>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_numnodes' mangled-name='avl_numnodes' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_numnodes'>
<parameter type-id='a3681dea'/>
<return type-id='ee1f298e'/>
</function-decl>
<function-decl name='avl_destroy_nodes' mangled-name='avl_destroy_nodes' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_destroy_nodes'>
<parameter type-id='a3681dea'/>
<parameter type-id='63e171df'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_destroy' mangled-name='avl_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_destroy'>
<parameter type-id='a3681dea'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_prop_readonly' mangled-name='zfs_prop_readonly' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_readonly'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_inheritable' mangled-name='zfs_prop_inheritable' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_inheritable'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_setonce' mangled-name='zfs_prop_setonce' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_setonce'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_encryption_key_param' mangled-name='zfs_prop_encryption_key_param' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_encryption_key_param'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_valid_keylocation' mangled-name='zfs_prop_valid_keylocation' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_valid_keylocation'>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_user' mangled-name='zfs_prop_user' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_user'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_userquota' mangled-name='zfs_prop_userquota' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_userquota'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_written' mangled-name='zfs_prop_written' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_written'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_index_to_string' mangled-name='zfs_prop_index_to_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_index_to_string'>
<parameter type-id='58603c44'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_valid_for_type' mangled-name='zfs_prop_valid_for_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_valid_for_type'>
<parameter type-id='95e97e5e'/>
<parameter type-id='2e45de5d'/>
<parameter type-id='c19b74c3'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nvlist_alloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='857bb57e'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_size' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='78c01427'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_pack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='9b23c9ad'/>
<parameter type-id='78c01427'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_unpack' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='857bb57e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='22cce67b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_uint64_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='713a56f5'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_remove' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='8d0687d2'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_remove_all' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_int64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<parameter type-id='cb785ebf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_uint64_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='892b4acc'/>
<parameter type-id='4dd26a40'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_nvlist_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='75be733c'/>
<parameter type-id='4dd26a40'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_empty' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nvpair_type' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='8d0687d2'/>
</function-decl>
<function-decl name='nvpair_value_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvpair_value_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fnvlist_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_boolean' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_lookup_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='fnvlist_lookup_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='fnvlist_lookup_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='fnvpair_value_int32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='3ff5601b'/>
</function-decl>
<function-decl name='fnvpair_value_uint64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='entity_namecheck' mangled-name='entity_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='entity_namecheck'>
<parameter type-id='80f4b756'/>
<parameter type-id='053457bd'/>
<parameter type-id='26a90f95'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='dataset_nestcheck' mangled-name='dataset_nestcheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='dataset_nestcheck'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='mountpoint_namecheck' mangled-name='mountpoint_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='mountpoint_namecheck'>
<parameter type-id='80f4b756'/>
<parameter type-id='053457bd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_type' mangled-name='zfs_prop_get_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_type'>
<parameter type-id='58603c44'/>
<return type-id='31429eff'/>
</function-decl>
<function-decl name='sa_validate_shareopts' mangled-name='sa_validate_shareopts' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_validate_shareopts'>
<parameter type-id='80f4b756'/>
<parameter type-id='9155d4b5'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getmntany' mangled-name='getmntany' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getmntany'>
<parameter type-id='822cd80b'/>
<parameter type-id='9d424d31'/>
<parameter type-id='9d424d31'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='_sol_getmntent' mangled-name='_sol_getmntent' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='_sol_getmntent'>
<parameter type-id='822cd80b'/>
<parameter type-id='9d424d31'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getgrnam' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='566b3f52'/>
</function-decl>
<function-decl name='hasmntopt' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='48bea5ec'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='pthread_mutex_init' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<parameter type-id='c2afbd7e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_mutex_destroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_mutex_lock' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_mutex_unlock' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='18c91f9e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getpwnam' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='a195f4a3'/>
</function-decl>
<function-decl name='strtol' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<return type-id='bd54fe1a'/>
</function-decl>
<function-decl name='strtoul' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<return type-id='7359adad'/>
</function-decl>
<function-decl name='abort' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strrchr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strcspn' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='strstr' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strsep' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='8c85230f'/>
<parameter type-id='9d26089a'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='strftime' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='9d26089a'/>
<parameter type-id='f8c6051d'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='localtime_r' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='d6e2847c'/>
<parameter type-id='f099ad08'/>
<return type-id='d915a820'/>
</function-decl>
<function-decl name='__fprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='ioctl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='7359adad'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_type_to_name' mangled-name='zfs_type_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_type_to_name'>
<parameter type-id='2e45de5d' name='type'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_name_valid' mangled-name='zfs_name_valid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_name_valid'>
<parameter type-id='80f4b756' name='name'/>
<parameter type-id='2e45de5d' name='type'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_free_handles' mangled-name='zpool_free_handles' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_free_handles'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_bookmark_exists' mangled-name='zfs_bookmark_exists' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_bookmark_exists'>
<parameter type-id='80f4b756' name='path'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='libzfs_mnttab_init' mangled-name='libzfs_mnttab_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_init'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_mnttab_fini' mangled-name='libzfs_mnttab_fini' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_fini'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_mnttab_cache' mangled-name='libzfs_mnttab_cache' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_cache'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='c19b74c3' name='enable'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_mnttab_find' mangled-name='libzfs_mnttab_find' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_find'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='fsname'/>
<parameter type-id='9d424d31' name='entry'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_mnttab_add' mangled-name='libzfs_mnttab_add' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_add'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='special'/>
<parameter type-id='80f4b756' name='mountp'/>
<parameter type-id='80f4b756' name='mntopts'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_mnttab_remove' mangled-name='libzfs_mnttab_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_mnttab_remove'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='fsname'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_spa_version' mangled-name='zfs_spa_version' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_spa_version'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='7292109c' name='spa_version'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_set' mangled-name='zfs_prop_set' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_set'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='80f4b756' name='propval'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_set_list' mangled-name='zfs_prop_set_list' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_set_list'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_inherit' mangled-name='zfs_prop_inherit' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_inherit'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='c19b74c3' name='received'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='getprop_uint64' mangled-name='getprop_uint64' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getprop_uint64'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='58603c44' name='prop'/>
<parameter type-id='7d3cd834' name='source'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_prop_get_recvd' mangled-name='zfs_prop_get_recvd' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_recvd'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='26a90f95' name='propbuf'/>
<parameter type-id='b59d7dce' name='proplen'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_clones_nvl' mangled-name='zfs_get_clones_nvl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_clones_nvl'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_prop_get_numeric' mangled-name='zfs_prop_get_numeric' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_numeric'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='58603c44' name='prop'/>
<parameter type-id='5d6479ae' name='value'/>
<parameter type-id='debc6aa3' name='src'/>
<parameter type-id='26a90f95' name='statbuf'/>
<parameter type-id='b59d7dce' name='statlen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_userquota_int' mangled-name='zfs_prop_get_userquota_int' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_userquota_int'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='5d6479ae' name='propvalue'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_userquota' mangled-name='zfs_prop_get_userquota' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_userquota'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='26a90f95' name='propbuf'/>
<parameter type-id='95e97e5e' name='proplen'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_written_int' mangled-name='zfs_prop_get_written_int' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_written_int'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='5d6479ae' name='propvalue'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_get_written' mangled-name='zfs_prop_get_written' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_written'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='26a90f95' name='propbuf'/>
<parameter type-id='95e97e5e' name='proplen'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_pool_name' mangled-name='zfs_get_pool_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_pool_name'>
<parameter type-id='fcd57163' name='zhp'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_get_type' mangled-name='zfs_get_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_type'>
<parameter type-id='fcd57163' name='zhp'/>
<return type-id='2e45de5d'/>
</function-decl>
<function-decl name='zfs_get_underlying_type' mangled-name='zfs_get_underlying_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_underlying_type'>
<parameter type-id='fcd57163' name='zhp'/>
<return type-id='2e45de5d'/>
</function-decl>
<function-decl name='zfs_dataset_exists' mangled-name='zfs_dataset_exists' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dataset_exists'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='2e45de5d' name='types'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_create_ancestors' mangled-name='zfs_create_ancestors' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_create_ancestors'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_create' mangled-name='zfs_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_create'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='2e45de5d' name='type'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_destroy' mangled-name='zfs_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_destroy'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='c19b74c3' name='defer'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_destroy_snaps' mangled-name='zfs_destroy_snaps' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_destroy_snaps'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='26a90f95' name='snapname'/>
<parameter type-id='c19b74c3' name='defer'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_destroy_snaps_nvl' mangled-name='zfs_destroy_snaps_nvl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_destroy_snaps_nvl'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='5ce45b60' name='snaps'/>
<parameter type-id='c19b74c3' name='defer'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_clone' mangled-name='zfs_clone' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_clone'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='target'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_promote' mangled-name='zfs_promote' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_promote'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_snapshot_nvl' mangled-name='zfs_snapshot_nvl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_snapshot_nvl'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='5ce45b60' name='snaps'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_snapshot' mangled-name='zfs_snapshot' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_snapshot'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='c19b74c3' name='recursive'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_rollback' mangled-name='zfs_rollback' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_rollback'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='9200a744' name='snap'/>
<parameter type-id='c19b74c3' name='force'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_rename' mangled-name='zfs_rename' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_rename'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='target'/>
<parameter type-id='067170c2' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_all_props' mangled-name='zfs_get_all_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_all_props'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_get_recvd_props' mangled-name='zfs_get_recvd_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_recvd_props'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_get_user_props' mangled-name='zfs_get_user_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_user_props'>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_expand_proplist' mangled-name='zfs_expand_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_expand_proplist'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='e4378506' name='plp'/>
<parameter type-id='c19b74c3' name='received'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prune_proplist' mangled-name='zfs_prune_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prune_proplist'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='ae3e8ca6' name='props'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_smb_acl_add' mangled-name='zfs_smb_acl_add' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_smb_acl_add'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='dataset'/>
<parameter type-id='26a90f95' name='path'/>
<parameter type-id='26a90f95' name='resource'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_smb_acl_remove' mangled-name='zfs_smb_acl_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_smb_acl_remove'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='dataset'/>
<parameter type-id='26a90f95' name='path'/>
<parameter type-id='26a90f95' name='resource'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_smb_acl_purge' mangled-name='zfs_smb_acl_purge' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_smb_acl_purge'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='dataset'/>
<parameter type-id='26a90f95' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_smb_acl_rename' mangled-name='zfs_smb_acl_rename' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_smb_acl_rename'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='dataset'/>
<parameter type-id='26a90f95' name='path'/>
<parameter type-id='26a90f95' name='oldname'/>
<parameter type-id='26a90f95' name='newname'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_userspace' mangled-name='zfs_userspace' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_userspace'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='279fde6a' name='type'/>
<parameter type-id='16c5f410' name='func'/>
<parameter type-id='eaa32e2f' name='arg'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_hold' mangled-name='zfs_hold' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_hold'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='tag'/>
<parameter type-id='c19b74c3' name='recursive'/>
<parameter type-id='95e97e5e' name='cleanup_fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_hold_nvl' mangled-name='zfs_hold_nvl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_hold_nvl'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='cleanup_fd'/>
<parameter type-id='5ce45b60' name='holds'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_release' mangled-name='zfs_release' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_release'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='snapname'/>
<parameter type-id='80f4b756' name='tag'/>
<parameter type-id='c19b74c3' name='recursive'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_fsacl' mangled-name='zfs_get_fsacl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_fsacl'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='857bb57e' name='nvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_set_fsacl' mangled-name='zfs_set_fsacl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_set_fsacl'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='c19b74c3' name='un'/>
<parameter type-id='5ce45b60' name='nvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_get_holds' mangled-name='zfs_get_holds' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_holds'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='857bb57e' name='nvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zvol_volsize_to_reservation' mangled-name='zvol_volsize_to_reservation' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zvol_volsize_to_reservation'>
<parameter type-id='4c81de99' name='zph'/>
<parameter type-id='9c313c2d' name='volsize'/>
<parameter type-id='5ce45b60' name='props'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_wait_status' mangled-name='zfs_wait_status' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_wait_status'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='3024501a' name='activity'/>
<parameter type-id='37e3bd22' name='missing'/>
<parameter type-id='37e3bd22' name='waited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_error_fmt' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_standard_error_fmt' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_setprop_error' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='58603c44'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='26a90f95'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_parse_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='3fa542f0'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='2e45de5d'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='7d3cd834'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_expand_list' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4378506'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zcmd_write_src_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='changelist_prefix' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='changelist_postfix' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='changelist_rename' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='changelist_remove' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='changelist_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='changelist_gather' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='58603c44'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<return type-id='0d41d328'/>
</function-decl>
<function-decl name='changelist_haszonedchild' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_name_valid' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-type size-in-bits='64' id='7e291ce6'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='80f4b756'/>
<parameter type-id='354978ed'/>
<parameter type-id='9c313c2d'/>
<return type-id='95e97e5e'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_diff.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='448' id='6093ff7c'>
<subrange length='56' type-id='7359adad' id='f8137894'/>
</array-type-def>
<typedef-decl name='pthread_t' type-id='7359adad' id='4051f5e7'/>
<union-decl name='pthread_attr_t' size-in-bits='448' visibility='default' id='b63afacd'>
<data-member access='public'>
<var-decl name='__size' type-id='6093ff7c' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='__align' type-id='bd54fe1a' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='pthread_attr_t' type-id='b63afacd' id='7d8569fd'/>
<class-decl name='differ_info' size-in-bits='9088' is-struct='yes' visibility='default' id='d41965ee'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zhp' type-id='9200a744' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='fromsnap' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='frommnt' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='tosnap' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='tomnt' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='ds' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='dsmnt' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='tmpsnap' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='errbuf' type-id='b54ce520' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8704'>
<var-decl name='isclone' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8736'>
<var-decl name='scripted' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8768'>
<var-decl name='classify' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8800'>
<var-decl name='timestamped' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8832'>
<var-decl name='no_mangle' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8896'>
<var-decl name='shares' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8960'>
<var-decl name='zerr' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8992'>
<var-decl name='cleanupfd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='9024'>
<var-decl name='outputfd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='9056'>
<var-decl name='datafd' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='differ_info_t' type-id='d41965ee' id='e8525f0e'/>
<qualified-type-def type-id='7d8569fd' const='yes' id='e06dee2d'/>
<pointer-type-def type-id='e06dee2d' size-in-bits='64' id='540db505'/>
<qualified-type-def type-id='540db505' restrict='yes' id='e1815e87'/>
<pointer-type-def type-id='e8525f0e' size-in-bits='64' id='ee78f675'/>
<pointer-type-def type-id='4051f5e7' size-in-bits='64' id='e01b5462'/>
<qualified-type-def type-id='e01b5462' restrict='yes' id='cc338b26'/>
<pointer-type-def type-id='cd5d79f4' size-in-bits='64' id='5ad9edb6'/>
<function-decl name='is_mounted' mangled-name='is_mounted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='is_mounted'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9b23c9ad'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='color_start' mangled-name='color_start' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='color_start'>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='color_end' mangled-name='color_end' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='color_end'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='pthread_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='cc338b26'/>
<parameter type-id='e1815e87'/>
<parameter type-id='5ad9edb6'/>
<parameter type-id='1b7446cd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_join' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4051f5e7'/>
<parameter type-id='63e171df'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pthread_cancel' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4051f5e7'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fputs' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='e75a27e9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pipe2' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='7292109c'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_show_diffs' mangled-name='zfs_show_diffs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_show_diffs'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='outfd'/>
<parameter type-id='80f4b756' name='fromsnap'/>
<parameter type-id='80f4b756' name='tosnap'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_asprintf' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zfs_validate_name' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='c19b74c3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='find_shares_object' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='ee78f675'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='cd5d79f4'>
<parameter type-id='eaa32e2f'/>
<return type-id='eaa32e2f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_import.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='03085adc' size-in-bits='192' id='083f8d58'>
<subrange length='3' type-id='7359adad' id='56f209d2'/>
</array-type-def>
<typedef-decl name='refresh_config_func_t' type-id='29f040d2' id='b7c58eaa'/>
<typedef-decl name='pool_active_func_t' type-id='baa42fef' id='de5d1d8f'/>
<class-decl name='pool_config_ops' size-in-bits='128' is-struct='yes' visibility='default' id='8b092c69'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pco_refresh_config' type-id='e7c00489' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pco_pool_active' type-id='9eadf5e0' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='pool_config_ops_t' type-id='1a21babe' id='b1e62775'/>
<enum-decl name='pool_state' id='4871ac24'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_STATE_ACTIVE' value='0'/>
<enumerator name='POOL_STATE_EXPORTED' value='1'/>
<enumerator name='POOL_STATE_DESTROYED' value='2'/>
<enumerator name='POOL_STATE_SPARE' value='3'/>
<enumerator name='POOL_STATE_L2CACHE' value='4'/>
<enumerator name='POOL_STATE_UNINITIALIZED' value='5'/>
<enumerator name='POOL_STATE_UNAVAIL' value='6'/>
<enumerator name='POOL_STATE_POTENTIALLY_ACTIVE' value='7'/>
</enum-decl>
<typedef-decl name='pool_state_t' type-id='4871ac24' id='084a08a3'/>
<class-decl name='stat64' size-in-bits='1152' is-struct='yes' visibility='default' id='0bbec9cd'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='st_dev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='st_ino' type-id='71288a47' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='st_nlink' type-id='80f0b9df' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='st_mode' type-id='e1c52942' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='st_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='st_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='__pad0' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='st_rdev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='st_size' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='st_blksize' type-id='d3f10a7f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='st_blocks' type-id='4e711bf1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='st_atim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='st_mtim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='st_ctim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='__glibc_reserved' type-id='083f8d58' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__dev_t' type-id='7359adad' id='35ed8932'/>
<typedef-decl name='__ino64_t' type-id='7359adad' id='71288a47'/>
<typedef-decl name='__mode_t' type-id='f0981eeb' id='e1c52942'/>
<typedef-decl name='__nlink_t' type-id='7359adad' id='80f0b9df'/>
<typedef-decl name='__blksize_t' type-id='bd54fe1a' id='d3f10a7f'/>
<typedef-decl name='__blkcnt64_t' type-id='bd54fe1a' id='4e711bf1'/>
<typedef-decl name='__syscall_slong_t' type-id='bd54fe1a' id='03085adc'/>
<class-decl name='timespec' size-in-bits='128' is-struct='yes' visibility='default' id='a9c79a1f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tv_sec' type-id='65eda9c0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tv_nsec' type-id='03085adc' visibility='default'/>
</data-member>
</class-decl>
<qualified-type-def type-id='8b092c69' const='yes' id='1a21babe'/>
<pointer-type-def type-id='de5d1d8f' size-in-bits='64' id='9eadf5e0'/>
<pointer-type-def type-id='084a08a3' size-in-bits='64' id='b9ea57b8'/>
<pointer-type-def type-id='b7c58eaa' size-in-bits='64' id='e7c00489'/>
<pointer-type-def type-id='0bbec9cd' size-in-bits='64' id='62f7a03d'/>
<var-decl name='libzfs_config_ops' type-id='b1e62775' mangled-name='libzfs_config_ops' visibility='default' elf-symbol-id='libzfs_config_ops'/>
<function-decl name='zpool_read_label' mangled-name='zpool_read_label' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_read_label'>
<parameter type-id='95e97e5e'/>
<parameter type-id='857bb57e'/>
<parameter type-id='7292109c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='pwrite64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='724e4de6'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='__pread64_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='724e4de6'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='fstat64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='62f7a03d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zcmd_write_conf_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_clear_label' mangled-name='zpool_clear_label' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_clear_label'>
<parameter type-id='95e97e5e' name='fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_in_use' mangled-name='zpool_in_use' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_in_use'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='b9ea57b8' name='state'/>
<parameter type-id='9b23c9ad' name='namestr'/>
<parameter type-id='37e3bd22' name='inuse'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='baa42fef'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='29f040d2'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='5ce45b60'/>
<return type-id='5ce45b60'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_iter.c' language='LANG_C99'>
<pointer-type-def type-id='b351119f' size-in-bits='64' id='716943c7'/>
<function-decl name='avl_first' mangled-name='avl_first' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_first'>
<parameter type-id='a3681dea'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_walk' mangled-name='avl_walk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_walk'>
<parameter type-id='716943c7'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='95e97e5e'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='make_dataset_handle_zc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='e4ec4540'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='make_dataset_simple_handle_zc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='e4ec4540'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='make_bookmark_handle' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<return type-id='9200a744'/>
</function-decl>
+ <function-decl name='zfs_iter_filesystems' mangled-name='zfs_iter_filesystems' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_filesystems'>
+ <parameter type-id='9200a744' name='zhp'/>
+ <parameter type-id='d8e49ab9' name='func'/>
+ <parameter type-id='eaa32e2f' name='data'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
+ <function-decl name='zfs_iter_snapshots' mangled-name='zfs_iter_snapshots' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots'>
+ <parameter type-id='9200a744' name='zhp'/>
+ <parameter type-id='c19b74c3' name='simple'/>
+ <parameter type-id='d8e49ab9' name='func'/>
+ <parameter type-id='eaa32e2f' name='data'/>
+ <parameter type-id='9c313c2d' name='min_txg'/>
+ <parameter type-id='9c313c2d' name='max_txg'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
+ <function-decl name='zfs_iter_bookmarks' mangled-name='zfs_iter_bookmarks' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_bookmarks'>
+ <parameter type-id='9200a744' name='zhp'/>
+ <parameter type-id='d8e49ab9' name='func'/>
+ <parameter type-id='eaa32e2f' name='data'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
<function-decl name='zfs_iter_snapshots_sorted' mangled-name='zfs_iter_snapshots_sorted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots_sorted'>
+ <parameter type-id='9200a744' name='zhp'/>
+ <parameter type-id='d8e49ab9' name='callback'/>
+ <parameter type-id='eaa32e2f' name='data'/>
+ <parameter type-id='9c313c2d' name='min_txg'/>
+ <parameter type-id='9c313c2d' name='max_txg'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
+ <function-decl name='zfs_iter_snapshots_sorted_v2' mangled-name='zfs_iter_snapshots_sorted_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapshots_sorted_v2'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='flags'/>
<parameter type-id='d8e49ab9' name='callback'/>
<parameter type-id='eaa32e2f' name='data'/>
<parameter type-id='9c313c2d' name='min_txg'/>
<parameter type-id='9c313c2d' name='max_txg'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_iter_snapspec' mangled-name='zfs_iter_snapspec' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapspec'>
+ <parameter type-id='9200a744' name='fs_zhp'/>
+ <parameter type-id='80f4b756' name='spec_orig'/>
+ <parameter type-id='d8e49ab9' name='func'/>
+ <parameter type-id='eaa32e2f' name='arg'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
+ <function-decl name='zfs_iter_snapspec_v2' mangled-name='zfs_iter_snapspec_v2' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_snapspec_v2'>
<parameter type-id='9200a744' name='fs_zhp'/>
<parameter type-id='95e97e5e' name='flags'/>
<parameter type-id='80f4b756' name='spec_orig'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='arg'/>
<return type-id='95e97e5e'/>
</function-decl>
+ <function-decl name='zfs_iter_children' mangled-name='zfs_iter_children' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_children'>
+ <parameter type-id='9200a744' name='zhp'/>
+ <parameter type-id='d8e49ab9' name='func'/>
+ <parameter type-id='eaa32e2f' name='data'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
+ <function-decl name='zfs_iter_dependents' mangled-name='zfs_iter_dependents' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_iter_dependents'>
+ <parameter type-id='9200a744' name='zhp'/>
+ <parameter type-id='c19b74c3' name='allowrecursion'/>
+ <parameter type-id='d8e49ab9' name='func'/>
+ <parameter type-id='eaa32e2f' name='data'/>
+ <return type-id='95e97e5e'/>
+ </function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_mount.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='6028cbfe' size-in-bits='256' id='b39b9aa7'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<class-decl name='__dirstream' is-struct='yes' visibility='default' is-declaration-only='yes' id='20cd73f2'/>
<class-decl name='tpool' size-in-bits='2496' is-struct='yes' visibility='default' id='88d1b7f9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tp_forw' type-id='9cf59a50' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='tp_back' type-id='9cf59a50' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='tp_mutex' type-id='7a6844eb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='tp_busycv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='tp_workcv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1216'>
<var-decl name='tp_waitcv' type-id='62fab762' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1600'>
<var-decl name='tp_active' type-id='ad33e5e7' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1664'>
<var-decl name='tp_head' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1728'>
<var-decl name='tp_tail' type-id='f32b30e4' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1792'>
<var-decl name='tp_attr' type-id='7d8569fd' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2240'>
<var-decl name='tp_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2272'>
<var-decl name='tp_linger' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2304'>
<var-decl name='tp_njobs' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2336'>
<var-decl name='tp_minimum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2368'>
<var-decl name='tp_maximum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2400'>
<var-decl name='tp_current' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2432'>
<var-decl name='tp_idle' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='64' id='e4266c7e'>
<subrange length='2' type-id='7359adad' id='52efc4ef'/>
</array-type-def>
<class-decl name='get_all_cb' size-in-bits='192' is-struct='yes' visibility='default' id='803dac95'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='cb_handles' type-id='4507922a' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='cb_alloc' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='cb_used' type-id='b59d7dce' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='get_all_cb_t' type-id='803dac95' id='9b293607'/>
<typedef-decl name='tpool_t' type-id='88d1b7f9' id='b1bbf10d'/>
<typedef-decl name='DIR' type-id='20cd73f2' id='54a5d683'/>
<typedef-decl name='mode_t' type-id='e1c52942' id='d50d396c'/>
<typedef-decl name='__compar_fn_t' type-id='585e1de9' id='aba7edd8'/>
<class-decl name='dirent64' size-in-bits='2240' is-struct='yes' visibility='default' id='5725d813'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='d_ino' type-id='71288a47' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='d_off' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='d_reclen' type-id='8efea9e5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='144'>
<var-decl name='d_type' type-id='002ac4a6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='152'>
<var-decl name='d_name' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='statfs64' size-in-bits='960' is-struct='yes' visibility='default' id='a2a6be1a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='f_type' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='f_bsize' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='f_blocks' type-id='95fe1a02' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='f_bfree' type-id='95fe1a02' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='f_bavail' type-id='95fe1a02' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='f_files' type-id='0c3a4dde' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='f_ffree' type-id='0c3a4dde' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='f_fsid' type-id='0f35d263' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='f_namelen' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='f_frsize' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='640'>
<var-decl name='f_flags' type-id='6028cbfe' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='f_spare' type-id='b39b9aa7' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='stat' size-in-bits='1152' is-struct='yes' visibility='default' id='aafc373f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='st_dev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='st_ino' type-id='e43e523d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='st_nlink' type-id='80f0b9df' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='st_mode' type-id='e1c52942' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='st_uid' type-id='cc5fcceb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='st_gid' type-id='d94ec6d9' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='__pad0' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='st_rdev' type-id='35ed8932' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='st_size' type-id='79989e9c' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='st_blksize' type-id='d3f10a7f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='st_blocks' type-id='dbc43803' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='st_atim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='st_mtim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='st_ctim' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='__glibc_reserved' type-id='083f8d58' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__ino_t' type-id='7359adad' id='e43e523d'/>
<class-decl name='__fsid_t' size-in-bits='64' is-struct='yes' naming-typedef-id='0f35d263' visibility='default' id='ea35c84a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='__val' type-id='e4266c7e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__fsid_t' type-id='ea35c84a' id='0f35d263'/>
<typedef-decl name='__blkcnt_t' type-id='bd54fe1a' id='dbc43803'/>
<typedef-decl name='__fsblkcnt64_t' type-id='7359adad' id='95fe1a02'/>
<typedef-decl name='__fsfilcnt64_t' type-id='7359adad' id='0c3a4dde'/>
<typedef-decl name='__fsword_t' type-id='bd54fe1a' id='6028cbfe'/>
<pointer-type-def type-id='54a5d683' size-in-bits='64' id='f09217ba'/>
<pointer-type-def type-id='5725d813' size-in-bits='64' id='07b96073'/>
<pointer-type-def type-id='9b293607' size-in-bits='64' id='77bf1784'/>
<pointer-type-def type-id='7d8569fd' size-in-bits='64' id='7347a39e'/>
<pointer-type-def type-id='aafc373f' size-in-bits='64' id='4330df87'/>
<qualified-type-def type-id='4330df87' restrict='yes' id='73665405'/>
<pointer-type-def type-id='a2a6be1a' size-in-bits='64' id='7fd094c8'/>
<pointer-type-def type-id='b1bbf10d' size-in-bits='64' id='9cf59a50'/>
<pointer-type-def type-id='c5c76c9c' size-in-bits='64' id='b7f9d8e6'/>
<pointer-type-def type-id='9200a744' size-in-bits='64' id='4507922a'/>
<class-decl name='__dirstream' is-struct='yes' visibility='default' is-declaration-only='yes' id='20cd73f2'/>
<function-decl name='zpool_disable_datasets_os' mangled-name='zpool_disable_datasets_os' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_disable_datasets_os'>
<parameter type-id='4c81de99'/>
<parameter type-id='c19b74c3'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_disable_volume_os' mangled-name='zpool_disable_volume_os' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_disable_volume_os'>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_create' mangled-name='tpool_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_create'>
<parameter type-id='3502e3ff'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='7347a39e'/>
<return type-id='9cf59a50'/>
</function-decl>
<function-decl name='tpool_dispatch' mangled-name='tpool_dispatch' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_dispatch'>
<parameter type-id='9cf59a50'/>
<parameter type-id='b7f9d8e6'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='tpool_destroy' mangled-name='tpool_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_destroy'>
<parameter type-id='9cf59a50'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='tpool_wait' mangled-name='tpool_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='tpool_wait'>
<parameter type-id='9cf59a50'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='mkdirp' mangled-name='mkdirp' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='mkdirp'>
<parameter type-id='80f4b756'/>
<parameter type-id='d50d396c'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sa_errorstr' mangled-name='sa_errorstr' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_errorstr'>
<parameter type-id='95e97e5e'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='sa_enable_share' mangled-name='sa_enable_share' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_enable_share'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9155d4b5'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sa_disable_share' mangled-name='sa_disable_share' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_disable_share'>
<parameter type-id='80f4b756'/>
<parameter type-id='9155d4b5'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sa_is_shared' mangled-name='sa_is_shared' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_is_shared'>
<parameter type-id='80f4b756'/>
<parameter type-id='9155d4b5'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='sa_truncate_shares' mangled-name='sa_truncate_shares' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='sa_truncate_shares'>
<parameter type-id='9155d4b5'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fdopendir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='f09217ba'/>
</function-decl>
<function-decl name='closedir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f09217ba'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='readdir64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f09217ba'/>
<return type-id='07b96073'/>
</function-decl>
<function-decl name='qsort' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='aba7edd8'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='rmdir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__openat_too_many_args' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='__openat_missing_mode' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='statfs64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='7fd094c8'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_realloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='changelist_unshare' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0d41d328'/>
<parameter type-id='4567bbc9'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='do_mount' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='do_unmount' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9200a744'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_mount_at' mangled-name='zfs_mount_at' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_mount_at'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='options'/>
<parameter type-id='95e97e5e' name='flags'/>
<parameter type-id='80f4b756' name='mountpoint'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_unmountall' mangled-name='zfs_unmountall' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_unmountall'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_truncate_shares' mangled-name='zfs_truncate_shares' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_truncate_shares'>
<parameter type-id='4567bbc9' name='proto'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_unshareall' mangled-name='zfs_unshareall' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_unshareall'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='4567bbc9' name='proto'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_add_handle' mangled-name='libzfs_add_handle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_add_handle'>
<parameter type-id='77bf1784' name='cbp'/>
<parameter type-id='9200a744' name='zhp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_foreach_mountpoint' mangled-name='zfs_foreach_mountpoint' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_foreach_mountpoint'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='4507922a' name='handles'/>
<parameter type-id='b59d7dce' name='num_handles'/>
<parameter type-id='d8e49ab9' name='func'/>
<parameter type-id='eaa32e2f' name='data'/>
<parameter type-id='c19b74c3' name='parallel'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_enable_datasets' mangled-name='zpool_enable_datasets' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_enable_datasets'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='mntopts'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_disable_datasets' mangled-name='zpool_disable_datasets' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_disable_datasets'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='c19b74c3' name='force'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_pool.c' language='LANG_C99'>
<type-decl name='long long unsigned int' size-in-bits='64' id='3a47d82b'/>
<class-decl name='splitflags' size-in-bits='64' is-struct='yes' visibility='default' id='dc01bf52'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='dryrun' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1'>
<var-decl name='import' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='name_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='splitflags_t' type-id='dc01bf52' id='325c1e34'/>
<class-decl name='trimflags' size-in-bits='192' is-struct='yes' visibility='default' id='8ef58008'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='fullpool' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='secure' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='wait' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='rate' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='trimflags_t' type-id='8ef58008' id='a093cbb8'/>
<enum-decl name='zpool_status_t' naming-typedef-id='d3dd6294' id='5e770b40'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_CACHE' value='0'/>
<enumerator name='ZPOOL_STATUS_MISSING_DEV_R' value='1'/>
<enumerator name='ZPOOL_STATUS_MISSING_DEV_NR' value='2'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_LABEL_R' value='3'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_LABEL_NR' value='4'/>
<enumerator name='ZPOOL_STATUS_BAD_GUID_SUM' value='5'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_POOL' value='6'/>
<enumerator name='ZPOOL_STATUS_CORRUPT_DATA' value='7'/>
<enumerator name='ZPOOL_STATUS_FAILING_DEV' value='8'/>
<enumerator name='ZPOOL_STATUS_VERSION_NEWER' value='9'/>
<enumerator name='ZPOOL_STATUS_HOSTID_MISMATCH' value='10'/>
<enumerator name='ZPOOL_STATUS_HOSTID_ACTIVE' value='11'/>
<enumerator name='ZPOOL_STATUS_HOSTID_REQUIRED' value='12'/>
<enumerator name='ZPOOL_STATUS_IO_FAILURE_WAIT' value='13'/>
<enumerator name='ZPOOL_STATUS_IO_FAILURE_CONTINUE' value='14'/>
<enumerator name='ZPOOL_STATUS_IO_FAILURE_MMP' value='15'/>
<enumerator name='ZPOOL_STATUS_BAD_LOG' value='16'/>
<enumerator name='ZPOOL_STATUS_ERRATA' value='17'/>
<enumerator name='ZPOOL_STATUS_UNSUP_FEAT_READ' value='18'/>
<enumerator name='ZPOOL_STATUS_UNSUP_FEAT_WRITE' value='19'/>
<enumerator name='ZPOOL_STATUS_FAULTED_DEV_R' value='20'/>
<enumerator name='ZPOOL_STATUS_FAULTED_DEV_NR' value='21'/>
<enumerator name='ZPOOL_STATUS_VERSION_OLDER' value='22'/>
<enumerator name='ZPOOL_STATUS_FEAT_DISABLED' value='23'/>
<enumerator name='ZPOOL_STATUS_RESILVERING' value='24'/>
<enumerator name='ZPOOL_STATUS_OFFLINE_DEV' value='25'/>
<enumerator name='ZPOOL_STATUS_REMOVED_DEV' value='26'/>
<enumerator name='ZPOOL_STATUS_REBUILDING' value='27'/>
<enumerator name='ZPOOL_STATUS_REBUILD_SCRUB' value='28'/>
<enumerator name='ZPOOL_STATUS_NON_NATIVE_ASHIFT' value='29'/>
<enumerator name='ZPOOL_STATUS_COMPATIBILITY_ERR' value='30'/>
<enumerator name='ZPOOL_STATUS_INCOMPATIBLE_FEAT' value='31'/>
<enumerator name='ZPOOL_STATUS_OK' value='32'/>
</enum-decl>
<typedef-decl name='zpool_status_t' type-id='5e770b40' id='d3dd6294'/>
<enum-decl name='zpool_compat_status_t' naming-typedef-id='901b78d1' id='20676925'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_COMPATIBILITY_OK' value='0'/>
<enumerator name='ZPOOL_COMPATIBILITY_WARNTOKEN' value='1'/>
<enumerator name='ZPOOL_COMPATIBILITY_BADTOKEN' value='2'/>
<enumerator name='ZPOOL_COMPATIBILITY_BADFILE' value='3'/>
<enumerator name='ZPOOL_COMPATIBILITY_NOFILES' value='4'/>
</enum-decl>
<typedef-decl name='zpool_compat_status_t' type-id='20676925' id='901b78d1'/>
<enum-decl name='vdev_prop_t' naming-typedef-id='5aa5c90c' id='1573bec8'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='VDEV_PROP_INVAL' value='-1'/>
<enumerator name='VDEV_PROP_USERPROP' value='-1'/>
<enumerator name='VDEV_PROP_NAME' value='0'/>
<enumerator name='VDEV_PROP_CAPACITY' value='1'/>
<enumerator name='VDEV_PROP_STATE' value='2'/>
<enumerator name='VDEV_PROP_GUID' value='3'/>
<enumerator name='VDEV_PROP_ASIZE' value='4'/>
<enumerator name='VDEV_PROP_PSIZE' value='5'/>
<enumerator name='VDEV_PROP_ASHIFT' value='6'/>
<enumerator name='VDEV_PROP_SIZE' value='7'/>
<enumerator name='VDEV_PROP_FREE' value='8'/>
<enumerator name='VDEV_PROP_ALLOCATED' value='9'/>
<enumerator name='VDEV_PROP_COMMENT' value='10'/>
<enumerator name='VDEV_PROP_EXPANDSZ' value='11'/>
<enumerator name='VDEV_PROP_FRAGMENTATION' value='12'/>
<enumerator name='VDEV_PROP_BOOTSIZE' value='13'/>
<enumerator name='VDEV_PROP_PARITY' value='14'/>
<enumerator name='VDEV_PROP_PATH' value='15'/>
<enumerator name='VDEV_PROP_DEVID' value='16'/>
<enumerator name='VDEV_PROP_PHYS_PATH' value='17'/>
<enumerator name='VDEV_PROP_ENC_PATH' value='18'/>
<enumerator name='VDEV_PROP_FRU' value='19'/>
<enumerator name='VDEV_PROP_PARENT' value='20'/>
<enumerator name='VDEV_PROP_CHILDREN' value='21'/>
<enumerator name='VDEV_PROP_NUMCHILDREN' value='22'/>
<enumerator name='VDEV_PROP_READ_ERRORS' value='23'/>
<enumerator name='VDEV_PROP_WRITE_ERRORS' value='24'/>
<enumerator name='VDEV_PROP_CHECKSUM_ERRORS' value='25'/>
<enumerator name='VDEV_PROP_INITIALIZE_ERRORS' value='26'/>
<enumerator name='VDEV_PROP_OPS_NULL' value='27'/>
<enumerator name='VDEV_PROP_OPS_READ' value='28'/>
<enumerator name='VDEV_PROP_OPS_WRITE' value='29'/>
<enumerator name='VDEV_PROP_OPS_FREE' value='30'/>
<enumerator name='VDEV_PROP_OPS_CLAIM' value='31'/>
<enumerator name='VDEV_PROP_OPS_TRIM' value='32'/>
<enumerator name='VDEV_PROP_BYTES_NULL' value='33'/>
<enumerator name='VDEV_PROP_BYTES_READ' value='34'/>
<enumerator name='VDEV_PROP_BYTES_WRITE' value='35'/>
<enumerator name='VDEV_PROP_BYTES_FREE' value='36'/>
<enumerator name='VDEV_PROP_BYTES_CLAIM' value='37'/>
<enumerator name='VDEV_PROP_BYTES_TRIM' value='38'/>
<enumerator name='VDEV_PROP_REMOVING' value='39'/>
<enumerator name='VDEV_PROP_ALLOCATING' value='40'/>
<enumerator name='VDEV_PROP_FAILFAST' value='41'/>
<enumerator name='VDEV_PROP_CHECKSUM_N' value='42'/>
<enumerator name='VDEV_PROP_CHECKSUM_T' value='43'/>
<enumerator name='VDEV_PROP_IO_N' value='44'/>
<enumerator name='VDEV_PROP_IO_T' value='45'/>
<enumerator name='VDEV_NUM_PROPS' value='46'/>
</enum-decl>
<typedef-decl name='vdev_prop_t' type-id='1573bec8' id='5aa5c90c'/>
<class-decl name='zpool_load_policy' size-in-bits='256' is-struct='yes' visibility='default' id='2f65b36f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zlp_rewind' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='zlp_maxmeta' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='zlp_maxdata' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='zlp_txg' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zpool_load_policy_t' type-id='2f65b36f' id='d11b7617'/>
<enum-decl name='vdev_state' id='21566197'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='VDEV_STATE_UNKNOWN' value='0'/>
<enumerator name='VDEV_STATE_CLOSED' value='1'/>
<enumerator name='VDEV_STATE_OFFLINE' value='2'/>
<enumerator name='VDEV_STATE_REMOVED' value='3'/>
<enumerator name='VDEV_STATE_CANT_OPEN' value='4'/>
<enumerator name='VDEV_STATE_FAULTED' value='5'/>
<enumerator name='VDEV_STATE_DEGRADED' value='6'/>
<enumerator name='VDEV_STATE_HEALTHY' value='7'/>
</enum-decl>
<typedef-decl name='vdev_state_t' type-id='21566197' id='35acf840'/>
<enum-decl name='vdev_aux' id='7f5bcca4'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='VDEV_AUX_NONE' value='0'/>
<enumerator name='VDEV_AUX_OPEN_FAILED' value='1'/>
<enumerator name='VDEV_AUX_CORRUPT_DATA' value='2'/>
<enumerator name='VDEV_AUX_NO_REPLICAS' value='3'/>
<enumerator name='VDEV_AUX_BAD_GUID_SUM' value='4'/>
<enumerator name='VDEV_AUX_TOO_SMALL' value='5'/>
<enumerator name='VDEV_AUX_BAD_LABEL' value='6'/>
<enumerator name='VDEV_AUX_VERSION_NEWER' value='7'/>
<enumerator name='VDEV_AUX_VERSION_OLDER' value='8'/>
<enumerator name='VDEV_AUX_UNSUP_FEAT' value='9'/>
<enumerator name='VDEV_AUX_SPARED' value='10'/>
<enumerator name='VDEV_AUX_ERR_EXCEEDED' value='11'/>
<enumerator name='VDEV_AUX_IO_FAILURE' value='12'/>
<enumerator name='VDEV_AUX_BAD_LOG' value='13'/>
<enumerator name='VDEV_AUX_EXTERNAL' value='14'/>
<enumerator name='VDEV_AUX_SPLIT_POOL' value='15'/>
<enumerator name='VDEV_AUX_BAD_ASHIFT' value='16'/>
<enumerator name='VDEV_AUX_EXTERNAL_PERSIST' value='17'/>
<enumerator name='VDEV_AUX_ACTIVE' value='18'/>
<enumerator name='VDEV_AUX_CHILDREN_OFFLINE' value='19'/>
<enumerator name='VDEV_AUX_ASHIFT_TOO_BIG' value='20'/>
</enum-decl>
<typedef-decl name='vdev_aux_t' type-id='7f5bcca4' id='9d774e0b'/>
<enum-decl name='pool_scan_func' id='1b092565'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_SCAN_NONE' value='0'/>
<enumerator name='POOL_SCAN_SCRUB' value='1'/>
<enumerator name='POOL_SCAN_RESILVER' value='2'/>
<enumerator name='POOL_SCAN_FUNCS' value='3'/>
</enum-decl>
<typedef-decl name='pool_scan_func_t' type-id='1b092565' id='7313fbe2'/>
<enum-decl name='pool_scrub_cmd' id='a1474cbd'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_SCRUB_NORMAL' value='0'/>
<enumerator name='POOL_SCRUB_PAUSE' value='1'/>
<enumerator name='POOL_SCRUB_FLAGS_END' value='2'/>
</enum-decl>
<typedef-decl name='pool_scrub_cmd_t' type-id='a1474cbd' id='b51cf3c2'/>
<enum-decl name='zpool_errata' id='d9abbf54'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_ERRATA_NONE' value='0'/>
<enumerator name='ZPOOL_ERRATA_ZOL_2094_SCRUB' value='1'/>
<enumerator name='ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY' value='2'/>
<enumerator name='ZPOOL_ERRATA_ZOL_6845_ENCRYPTION' value='3'/>
<enumerator name='ZPOOL_ERRATA_ZOL_8308_ENCRYPTION' value='4'/>
</enum-decl>
<typedef-decl name='zpool_errata_t' type-id='d9abbf54' id='688c495b'/>
<enum-decl name='pool_initialize_func' id='5c246ad4'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_INITIALIZE_START' value='0'/>
<enumerator name='POOL_INITIALIZE_CANCEL' value='1'/>
<enumerator name='POOL_INITIALIZE_SUSPEND' value='2'/>
<enumerator name='POOL_INITIALIZE_FUNCS' value='3'/>
</enum-decl>
<typedef-decl name='pool_initialize_func_t' type-id='5c246ad4' id='7063e1ab'/>
<enum-decl name='pool_trim_func' id='54ed608a'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='POOL_TRIM_START' value='0'/>
<enumerator name='POOL_TRIM_CANCEL' value='1'/>
<enumerator name='POOL_TRIM_SUSPEND' value='2'/>
<enumerator name='POOL_TRIM_FUNCS' value='3'/>
</enum-decl>
<typedef-decl name='pool_trim_func_t' type-id='54ed608a' id='b1146b8d'/>
<enum-decl name='zpool_wait_activity_t' naming-typedef-id='73446457' id='849338e3'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZPOOL_WAIT_CKPT_DISCARD' value='0'/>
<enumerator name='ZPOOL_WAIT_FREE' value='1'/>
<enumerator name='ZPOOL_WAIT_INITIALIZE' value='2'/>
<enumerator name='ZPOOL_WAIT_REPLACE' value='3'/>
<enumerator name='ZPOOL_WAIT_REMOVE' value='4'/>
<enumerator name='ZPOOL_WAIT_RESILVER' value='5'/>
<enumerator name='ZPOOL_WAIT_SCRUB' value='6'/>
<enumerator name='ZPOOL_WAIT_TRIM' value='7'/>
<enumerator name='ZPOOL_WAIT_NUM_ACTIVITIES' value='8'/>
</enum-decl>
<typedef-decl name='zpool_wait_activity_t' type-id='849338e3' id='73446457'/>
<enum-decl name='spa_feature' id='33ecb627'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='SPA_FEATURE_NONE' value='-1'/>
<enumerator name='SPA_FEATURE_ASYNC_DESTROY' value='0'/>
<enumerator name='SPA_FEATURE_EMPTY_BPOBJ' value='1'/>
<enumerator name='SPA_FEATURE_LZ4_COMPRESS' value='2'/>
<enumerator name='SPA_FEATURE_MULTI_VDEV_CRASH_DUMP' value='3'/>
<enumerator name='SPA_FEATURE_SPACEMAP_HISTOGRAM' value='4'/>
<enumerator name='SPA_FEATURE_ENABLED_TXG' value='5'/>
<enumerator name='SPA_FEATURE_HOLE_BIRTH' value='6'/>
<enumerator name='SPA_FEATURE_EXTENSIBLE_DATASET' value='7'/>
<enumerator name='SPA_FEATURE_EMBEDDED_DATA' value='8'/>
<enumerator name='SPA_FEATURE_BOOKMARKS' value='9'/>
<enumerator name='SPA_FEATURE_FS_SS_LIMIT' value='10'/>
<enumerator name='SPA_FEATURE_LARGE_BLOCKS' value='11'/>
<enumerator name='SPA_FEATURE_LARGE_DNODE' value='12'/>
<enumerator name='SPA_FEATURE_SHA512' value='13'/>
<enumerator name='SPA_FEATURE_SKEIN' value='14'/>
<enumerator name='SPA_FEATURE_EDONR' value='15'/>
<enumerator name='SPA_FEATURE_USEROBJ_ACCOUNTING' value='16'/>
<enumerator name='SPA_FEATURE_ENCRYPTION' value='17'/>
<enumerator name='SPA_FEATURE_PROJECT_QUOTA' value='18'/>
<enumerator name='SPA_FEATURE_DEVICE_REMOVAL' value='19'/>
<enumerator name='SPA_FEATURE_OBSOLETE_COUNTS' value='20'/>
<enumerator name='SPA_FEATURE_POOL_CHECKPOINT' value='21'/>
<enumerator name='SPA_FEATURE_SPACEMAP_V2' value='22'/>
<enumerator name='SPA_FEATURE_ALLOCATION_CLASSES' value='23'/>
<enumerator name='SPA_FEATURE_RESILVER_DEFER' value='24'/>
<enumerator name='SPA_FEATURE_BOOKMARK_V2' value='25'/>
<enumerator name='SPA_FEATURE_REDACTION_BOOKMARKS' value='26'/>
<enumerator name='SPA_FEATURE_REDACTED_DATASETS' value='27'/>
<enumerator name='SPA_FEATURE_BOOKMARK_WRITTEN' value='28'/>
<enumerator name='SPA_FEATURE_LOG_SPACEMAP' value='29'/>
<enumerator name='SPA_FEATURE_LIVELIST' value='30'/>
<enumerator name='SPA_FEATURE_DEVICE_REBUILD' value='31'/>
<enumerator name='SPA_FEATURE_ZSTD_COMPRESS' value='32'/>
<enumerator name='SPA_FEATURE_DRAID' value='33'/>
<enumerator name='SPA_FEATURE_ZILSAXATTR' value='34'/>
<enumerator name='SPA_FEATURE_HEAD_ERRLOG' value='35'/>
<enumerator name='SPA_FEATURE_BLAKE3' value='36'/>
<enumerator name='SPA_FEATURE_BLOCK_CLONING' value='37'/>
- <enumerator name='SPA_FEATURES' value='38'/>
+ <enumerator name='SPA_FEATURE_AVZ_V2' value='38'/>
+ <enumerator name='SPA_FEATURES' value='39'/>
</enum-decl>
<typedef-decl name='spa_feature_t' type-id='33ecb627' id='d6618c78'/>
<qualified-type-def type-id='22cce67b' const='yes' id='d2816df0'/>
<pointer-type-def type-id='d2816df0' size-in-bits='64' id='3bbfee2e'/>
<qualified-type-def type-id='b96825af' const='yes' id='2b61797f'/>
<pointer-type-def type-id='2b61797f' size-in-bits='64' id='9f7200cf'/>
<pointer-type-def type-id='d6618c78' size-in-bits='64' id='a8425263'/>
<qualified-type-def type-id='62f7a03d' restrict='yes' id='f1cadedf'/>
<pointer-type-def type-id='a093cbb8' size-in-bits='64' id='b13f38c3'/>
<pointer-type-def type-id='35acf840' size-in-bits='64' id='17f3480d'/>
<pointer-type-def type-id='688c495b' size-in-bits='64' id='cec6f2e4'/>
<pointer-type-def type-id='d11b7617' size-in-bits='64' id='23432aaa'/>
<function-decl name='zpool_get_handle' mangled-name='zpool_get_handle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_handle'>
<parameter type-id='4c81de99'/>
<return type-id='b0382bb3'/>
</function-decl>
<function-decl name='zpool_prop_to_name' mangled-name='zpool_prop_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_to_name'>
<parameter type-id='5d0c23fb'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_to_name' mangled-name='vdev_prop_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_to_name'>
<parameter type-id='5aa5c90c'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_user' mangled-name='vdev_prop_user' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_user'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_get_status' mangled-name='zpool_get_status' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_status'>
<parameter type-id='4c81de99'/>
<parameter type-id='7d3cd834'/>
<parameter type-id='cec6f2e4'/>
<return type-id='d3dd6294'/>
</function-decl>
<function-decl name='zpool_prop_default_string' mangled-name='zpool_prop_default_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_default_string'>
<parameter type-id='5d0c23fb'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_prop_default_numeric' mangled-name='zpool_prop_default_numeric' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_default_numeric'>
<parameter type-id='5d0c23fb'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='libzfs_envvar_is_set' mangled-name='libzfs_envvar_is_set' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_envvar_is_set'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='lzc_initialize' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='7063e1ab'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_trim' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='b1146b8d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_sync' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_reopen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_pool_checkpoint' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_pool_checkpoint_discard' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='73446457'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_wait_tag' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='73446457'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='37e3bd22'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_set_bootenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='22cce67b'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_bootenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_get_vdev_prop' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_set_vdev_prop' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_resolve_shortname' mangled-name='zfs_resolve_shortname' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_resolve_shortname'>
<parameter type-id='80f4b756'/>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_strip_partition' mangled-name='zfs_strip_partition' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_strip_partition'>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zfs_strip_path' mangled-name='zfs_strip_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_strip_path'>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_strcmp_pathname' mangled-name='zfs_strcmp_pathname' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_strcmp_pathname'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_history_unpack' mangled-name='zpool_history_unpack' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_history_unpack'>
<parameter type-id='26a90f95'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='75be733c'/>
<parameter type-id='4dd26a40'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_basename' mangled-name='zfs_basename' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_basename'>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_name_to_prop' mangled-name='zpool_name_to_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_name_to_prop'>
<parameter type-id='80f4b756'/>
<return type-id='5d0c23fb'/>
</function-decl>
<function-decl name='zpool_prop_readonly' mangled-name='zpool_prop_readonly' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_readonly'>
<parameter type-id='5d0c23fb'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_prop_setonce' mangled-name='zpool_prop_setonce' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_setonce'>
<parameter type-id='5d0c23fb'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_prop_feature' mangled-name='zpool_prop_feature' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_feature'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_prop_index_to_string' mangled-name='zpool_prop_index_to_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_index_to_string'>
<parameter type-id='5d0c23fb'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='vdev_name_to_prop' mangled-name='vdev_name_to_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_name_to_prop'>
<parameter type-id='80f4b756'/>
<return type-id='5aa5c90c'/>
</function-decl>
<function-decl name='vdev_prop_default_string' mangled-name='vdev_prop_default_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_default_string'>
<parameter type-id='5aa5c90c'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_default_numeric' mangled-name='vdev_prop_default_numeric' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_default_numeric'>
<parameter type-id='5aa5c90c'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='vdev_prop_readonly' mangled-name='vdev_prop_readonly' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_readonly'>
<parameter type-id='5aa5c90c'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='vdev_prop_index_to_string' mangled-name='vdev_prop_index_to_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_index_to_string'>
<parameter type-id='5aa5c90c'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_prop_vdev' mangled-name='zpool_prop_vdev' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_vdev'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='nvlist_add_nvpair' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='3fa542f0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_uint8_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9f7200cf'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_add_nvlist_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='3bbfee2e'/>
<parameter type-id='3502e3ff'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvpair_value_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3fa542f0'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fnvlist_add_boolean_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_int64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9da381c4'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_add_nvlist_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='3bbfee2e'/>
<parameter type-id='3502e3ff'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_lookup_uint64_array' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<parameter type-id='4dd26a40'/>
<return type-id='5d6479ae'/>
</function-decl>
<function-decl name='fnvpair_value_int64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='9da381c4'/>
</function-decl>
<function-decl name='fnvpair_value_string' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfeature_is_supported' mangled-name='zfeature_is_supported' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_is_supported'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfeature_lookup_guid' mangled-name='zfeature_lookup_guid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_lookup_guid'>
<parameter type-id='80f4b756'/>
<parameter type-id='a8425263'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfeature_lookup_name' mangled-name='zfeature_lookup_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_lookup_name'>
<parameter type-id='80f4b756'/>
<parameter type-id='a8425263'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_load_policy' mangled-name='zpool_get_load_policy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_load_policy'>
<parameter type-id='5ce45b60'/>
<parameter type-id='23432aaa'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='pool_namecheck' mangled-name='pool_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='pool_namecheck'>
<parameter type-id='80f4b756'/>
<parameter type-id='053457bd'/>
<parameter type-id='26a90f95'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_prop_get_type' mangled-name='zpool_prop_get_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_get_type'>
<parameter type-id='5d0c23fb'/>
<return type-id='31429eff'/>
</function-decl>
<function-decl name='vdev_prop_get_type' mangled-name='vdev_prop_get_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_get_type'>
<parameter type-id='5aa5c90c'/>
<return type-id='31429eff'/>
</function-decl>
<function-decl name='get_system_hostid' mangled-name='get_system_hostid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='get_system_hostid'>
<return type-id='7359adad'/>
</function-decl>
<function-decl name='strtoull' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<return type-id='3a47d82b'/>
</function-decl>
<function-decl name='memcmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strtok_r' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='266fe297'/>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='__realpath_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='munmap' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='stat64' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='f1cadedf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_standard_error' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_standard_error_fmt' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_relabel_disk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_props_refresh' mangled-name='zpool_props_refresh' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_props_refresh'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_state_to_name' mangled-name='zpool_state_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_state_to_name'>
<parameter type-id='35acf840' name='state'/>
<parameter type-id='9d774e0b' name='aux'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_pool_state_to_name' mangled-name='zpool_pool_state_to_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_pool_state_to_name'>
<parameter type-id='084a08a3' name='state'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_get_state_str' mangled-name='zpool_get_state_str' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_state_str'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_get_userprop' mangled-name='zpool_get_userprop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_userprop'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='len'/>
<parameter type-id='debc6aa3' name='srctype'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_set_prop' mangled-name='zpool_set_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_set_prop'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='80f4b756' name='propval'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_expand_proplist' mangled-name='zpool_expand_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_expand_proplist'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='e4378506' name='plp'/>
<parameter type-id='2e45de5d' name='type'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='vdev_expand_proplist' mangled-name='vdev_expand_proplist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_expand_proplist'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='vdevname'/>
<parameter type-id='e4378506' name='plp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_state' mangled-name='zpool_get_state' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_state'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_is_draid_spare' mangled-name='zpool_is_draid_spare' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_is_draid_spare'>
<parameter type-id='80f4b756' name='name'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_create' mangled-name='zpool_create' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_create'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='pool'/>
<parameter type-id='5ce45b60' name='nvroot'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='5ce45b60' name='fsprops'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_destroy' mangled-name='zpool_destroy' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_destroy'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='log_str'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_checkpoint' mangled-name='zpool_checkpoint' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_checkpoint'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_discard_checkpoint' mangled-name='zpool_discard_checkpoint' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_discard_checkpoint'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_add' mangled-name='zpool_add' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_add'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='5ce45b60' name='nvroot'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_export' mangled-name='zpool_export' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_export'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='c19b74c3' name='force'/>
<parameter type-id='80f4b756' name='log_str'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_export_force' mangled-name='zpool_export_force' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_export_force'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='log_str'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_explain_recover' mangled-name='zpool_explain_recover' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_explain_recover'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='name'/>
<parameter type-id='95e97e5e' name='reason'/>
<parameter type-id='5ce45b60' name='config'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_import' mangled-name='zpool_import' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_import'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='5ce45b60' name='config'/>
<parameter type-id='80f4b756' name='newname'/>
<parameter type-id='26a90f95' name='altroot'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_print_unsup_feat' mangled-name='zpool_print_unsup_feat' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_print_unsup_feat'>
<parameter type-id='5ce45b60' name='config'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_import_props' mangled-name='zpool_import_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_import_props'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='5ce45b60' name='config'/>
<parameter type-id='80f4b756' name='newname'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_initialize' mangled-name='zpool_initialize' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_initialize'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='7063e1ab' name='cmd_type'/>
<parameter type-id='5ce45b60' name='vds'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_initialize_wait' mangled-name='zpool_initialize_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_initialize_wait'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='7063e1ab' name='cmd_type'/>
<parameter type-id='5ce45b60' name='vds'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_trim' mangled-name='zpool_trim' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_trim'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='b1146b8d' name='cmd_type'/>
<parameter type-id='5ce45b60' name='vds'/>
<parameter type-id='b13f38c3' name='trim_flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_scan' mangled-name='zpool_scan' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_scan'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='7313fbe2' name='func'/>
<parameter type-id='b51cf3c2' name='cmd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_find_vdev_by_physpath' mangled-name='zpool_find_vdev_by_physpath' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_find_vdev_by_physpath'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='ppath'/>
<parameter type-id='37e3bd22' name='avail_spare'/>
<parameter type-id='37e3bd22' name='l2cache'/>
<parameter type-id='37e3bd22' name='log'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_find_vdev' mangled-name='zpool_find_vdev' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_find_vdev'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='37e3bd22' name='avail_spare'/>
<parameter type-id='37e3bd22' name='l2cache'/>
<parameter type-id='37e3bd22' name='log'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_vdev_path_to_guid' mangled-name='zpool_vdev_path_to_guid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_path_to_guid'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zpool_vdev_online' mangled-name='zpool_vdev_online' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_online'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='95e97e5e' name='flags'/>
<parameter type-id='17f3480d' name='newstate'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_offline' mangled-name='zpool_vdev_offline' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_offline'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='c19b74c3' name='istmp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_remove_wanted' mangled-name='zpool_vdev_remove_wanted' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_remove_wanted'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_fault' mangled-name='zpool_vdev_fault' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_fault'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='guid'/>
<parameter type-id='9d774e0b' name='aux'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_degrade' mangled-name='zpool_vdev_degrade' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_degrade'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='guid'/>
<parameter type-id='9d774e0b' name='aux'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_attach' mangled-name='zpool_vdev_attach' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_attach'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='old_disk'/>
<parameter type-id='80f4b756' name='new_disk'/>
<parameter type-id='5ce45b60' name='nvroot'/>
<parameter type-id='95e97e5e' name='replacing'/>
<parameter type-id='c19b74c3' name='rebuild'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_detach' mangled-name='zpool_vdev_detach' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_detach'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_split' mangled-name='zpool_vdev_split' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_split'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='26a90f95' name='newname'/>
<parameter type-id='857bb57e' name='newroot'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='325c1e34' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_remove' mangled-name='zpool_vdev_remove' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_remove'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_remove_cancel' mangled-name='zpool_vdev_remove_cancel' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_remove_cancel'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_indirect_size' mangled-name='zpool_vdev_indirect_size' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_indirect_size'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='5d6479ae' name='sizep'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_clear' mangled-name='zpool_clear' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_clear'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='5ce45b60' name='rewindnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_clear' mangled-name='zpool_vdev_clear' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_clear'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='guid'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_reguid' mangled-name='zpool_reguid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_reguid'>
<parameter type-id='4c81de99' name='zhp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_reopen_one' mangled-name='zpool_reopen_one' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_reopen_one'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_sync_one' mangled-name='zpool_sync_one' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_sync_one'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_vdev_name' mangled-name='zpool_vdev_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_vdev_name'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='5ce45b60' name='nv'/>
<parameter type-id='95e97e5e' name='name_flags'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zpool_get_errlog' mangled-name='zpool_get_errlog' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_errlog'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='857bb57e' name='nverrlistp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_upgrade' mangled-name='zpool_upgrade' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_upgrade'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='new_version'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_save_arguments' mangled-name='zfs_save_arguments' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_save_arguments'>
<parameter type-id='95e97e5e' name='argc'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='26a90f95' name='string'/>
<parameter type-id='95e97e5e' name='len'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_log_history' mangled-name='zpool_log_history' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_log_history'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='message'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_history' mangled-name='zpool_get_history' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_history'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='857bb57e' name='nvhisp'/>
<parameter type-id='5d6479ae' name='off'/>
<parameter type-id='37e3bd22' name='eof'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_events_next' mangled-name='zpool_events_next' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_events_next'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='857bb57e' name='nvp'/>
<parameter type-id='7292109c' name='dropped'/>
<parameter type-id='f0981eeb' name='flags'/>
<parameter type-id='95e97e5e' name='zevent_fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_events_clear' mangled-name='zpool_events_clear' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_events_clear'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='7292109c' name='count'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_events_seek' mangled-name='zpool_events_seek' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_events_seek'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='9c313c2d' name='eid'/>
<parameter type-id='95e97e5e' name='zevent_fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_obj_to_path' mangled-name='zpool_obj_to_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_obj_to_path'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='dsobj'/>
<parameter type-id='9c313c2d' name='obj'/>
<parameter type-id='26a90f95' name='pathname'/>
<parameter type-id='b59d7dce' name='len'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_obj_to_path_ds' mangled-name='zpool_obj_to_path_ds' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_obj_to_path_ds'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='9c313c2d' name='dsobj'/>
<parameter type-id='9c313c2d' name='obj'/>
<parameter type-id='26a90f95' name='pathname'/>
<parameter type-id='b59d7dce' name='len'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_wait' mangled-name='zpool_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_wait'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='73446457' name='activity'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_wait_status' mangled-name='zpool_wait_status' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_wait_status'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='73446457' name='activity'/>
<parameter type-id='37e3bd22' name='missing'/>
<parameter type-id='37e3bd22' name='waited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_set_bootenv' mangled-name='zpool_set_bootenv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_set_bootenv'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='22cce67b' name='envmap'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_bootenv' mangled-name='zpool_get_bootenv' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_bootenv'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='857bb57e' name='nvlp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_load_compat' mangled-name='zpool_load_compat' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_load_compat'>
<parameter type-id='80f4b756' name='compat'/>
<parameter type-id='37e3bd22' name='features'/>
<parameter type-id='26a90f95' name='report'/>
<parameter type-id='b59d7dce' name='rlen'/>
<return type-id='901b78d1'/>
</function-decl>
<function-decl name='zpool_get_vdev_prop_value' mangled-name='zpool_get_vdev_prop_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_vdev_prop_value'>
<parameter type-id='5ce45b60' name='nvprop'/>
<parameter type-id='5aa5c90c' name='prop'/>
<parameter type-id='26a90f95' name='prop_name'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='len'/>
<parameter type-id='debc6aa3' name='srctype'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_vdev_prop' mangled-name='zpool_get_vdev_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_vdev_prop'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='vdevname'/>
<parameter type-id='5aa5c90c' name='prop'/>
<parameter type-id='26a90f95' name='prop_name'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='len'/>
<parameter type-id='debc6aa3' name='srctype'/>
<parameter type-id='c19b74c3' name='literal'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_get_all_vdev_props' mangled-name='zpool_get_all_vdev_props' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_get_all_vdev_props'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='vdevname'/>
<parameter type-id='857bb57e' name='outnvl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_set_vdev_prop' mangled-name='zpool_set_vdev_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_set_vdev_prop'>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='vdevname'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='80f4b756' name='propval'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_sendrecv.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='2176' id='8c2bcad1'>
<subrange length='34' type-id='7359adad' id='6a6a7e00'/>
</array-type-def>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='256' id='85c64d26'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='96' id='fa8ef949'>
<subrange length='12' type-id='7359adad' id='84827bdc'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='128' id='fa9986a5'>
<subrange length='16' type-id='7359adad' id='848d0938'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='40' id='0f4ddd0b'>
<subrange length='5' type-id='7359adad' id='53010e10'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='48' id='0f562bd0'>
<subrange length='6' type-id='7359adad' id='52fa524b'/>
</array-type-def>
<array-type-def dimensions='1' type-id='b96825af' size-in-bits='64' id='13339fda'>
<subrange length='8' type-id='7359adad' id='56e0c0b1'/>
</array-type-def>
<class-decl name='sendflags' size-in-bits='576' is-struct='yes' visibility='default' id='f6aa15be'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='verbosity' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='replicate' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='skipmissing' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='doall' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='fromorigin' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='pad' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='props' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='dryrun' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='parsable' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='progress' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='progressastitle' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='largeblock' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='embed_data' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='416'>
<var-decl name='compress' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='raw' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='backup' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='holds' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='544'>
<var-decl name='saved' type-id='c19b74c3' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='sendflags_t' type-id='f6aa15be' id='945467e6'/>
<typedef-decl name='snapfilter_cb_t' type-id='d2a5e211' id='3d3ffb69'/>
<class-decl name='recvflags' size-in-bits='448' is-struct='yes' visibility='default' id='34a384dc'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='verbose' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='isprefix' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='istail' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='dryrun' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='force' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='canmountoff' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='resumable' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='byteswap' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='nomount' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='holds' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='skipholds' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='domount' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='forceunmount' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='416'>
<var-decl name='heal' type-id='c19b74c3' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='recvflags_t' type-id='34a384dc' id='9e59d1d4'/>
<enum-decl name='lzc_send_flags' id='bfbd3c8e'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='LZC_SEND_FLAG_EMBED_DATA' value='1'/>
<enumerator name='LZC_SEND_FLAG_LARGE_BLOCK' value='2'/>
<enumerator name='LZC_SEND_FLAG_COMPRESS' value='4'/>
<enumerator name='LZC_SEND_FLAG_RAW' value='8'/>
<enumerator name='LZC_SEND_FLAG_SAVED' value='16'/>
</enum-decl>
<class-decl name='ddt_key' size-in-bits='320' is-struct='yes' visibility='default' id='e0a4a1cb'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='ddk_cksum' type-id='39730d0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='ddk_prop' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ddt_key_t' type-id='e0a4a1cb' id='67f6d2cf'/>
<enum-decl name='dmu_object_type' id='04b3b0b9'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DMU_OT_NONE' value='0'/>
<enumerator name='DMU_OT_OBJECT_DIRECTORY' value='1'/>
<enumerator name='DMU_OT_OBJECT_ARRAY' value='2'/>
<enumerator name='DMU_OT_PACKED_NVLIST' value='3'/>
<enumerator name='DMU_OT_PACKED_NVLIST_SIZE' value='4'/>
<enumerator name='DMU_OT_BPOBJ' value='5'/>
<enumerator name='DMU_OT_BPOBJ_HDR' value='6'/>
<enumerator name='DMU_OT_SPACE_MAP_HEADER' value='7'/>
<enumerator name='DMU_OT_SPACE_MAP' value='8'/>
<enumerator name='DMU_OT_INTENT_LOG' value='9'/>
<enumerator name='DMU_OT_DNODE' value='10'/>
<enumerator name='DMU_OT_OBJSET' value='11'/>
<enumerator name='DMU_OT_DSL_DIR' value='12'/>
<enumerator name='DMU_OT_DSL_DIR_CHILD_MAP' value='13'/>
<enumerator name='DMU_OT_DSL_DS_SNAP_MAP' value='14'/>
<enumerator name='DMU_OT_DSL_PROPS' value='15'/>
<enumerator name='DMU_OT_DSL_DATASET' value='16'/>
<enumerator name='DMU_OT_ZNODE' value='17'/>
<enumerator name='DMU_OT_OLDACL' value='18'/>
<enumerator name='DMU_OT_PLAIN_FILE_CONTENTS' value='19'/>
<enumerator name='DMU_OT_DIRECTORY_CONTENTS' value='20'/>
<enumerator name='DMU_OT_MASTER_NODE' value='21'/>
<enumerator name='DMU_OT_UNLINKED_SET' value='22'/>
<enumerator name='DMU_OT_ZVOL' value='23'/>
<enumerator name='DMU_OT_ZVOL_PROP' value='24'/>
<enumerator name='DMU_OT_PLAIN_OTHER' value='25'/>
<enumerator name='DMU_OT_UINT64_OTHER' value='26'/>
<enumerator name='DMU_OT_ZAP_OTHER' value='27'/>
<enumerator name='DMU_OT_ERROR_LOG' value='28'/>
<enumerator name='DMU_OT_SPA_HISTORY' value='29'/>
<enumerator name='DMU_OT_SPA_HISTORY_OFFSETS' value='30'/>
<enumerator name='DMU_OT_POOL_PROPS' value='31'/>
<enumerator name='DMU_OT_DSL_PERMS' value='32'/>
<enumerator name='DMU_OT_ACL' value='33'/>
<enumerator name='DMU_OT_SYSACL' value='34'/>
<enumerator name='DMU_OT_FUID' value='35'/>
<enumerator name='DMU_OT_FUID_SIZE' value='36'/>
<enumerator name='DMU_OT_NEXT_CLONES' value='37'/>
<enumerator name='DMU_OT_SCAN_QUEUE' value='38'/>
<enumerator name='DMU_OT_USERGROUP_USED' value='39'/>
<enumerator name='DMU_OT_USERGROUP_QUOTA' value='40'/>
<enumerator name='DMU_OT_USERREFS' value='41'/>
<enumerator name='DMU_OT_DDT_ZAP' value='42'/>
<enumerator name='DMU_OT_DDT_STATS' value='43'/>
<enumerator name='DMU_OT_SA' value='44'/>
<enumerator name='DMU_OT_SA_MASTER_NODE' value='45'/>
<enumerator name='DMU_OT_SA_ATTR_REGISTRATION' value='46'/>
<enumerator name='DMU_OT_SA_ATTR_LAYOUTS' value='47'/>
<enumerator name='DMU_OT_SCAN_XLATE' value='48'/>
<enumerator name='DMU_OT_DEDUP' value='49'/>
<enumerator name='DMU_OT_DEADLIST' value='50'/>
<enumerator name='DMU_OT_DEADLIST_HDR' value='51'/>
<enumerator name='DMU_OT_DSL_CLONES' value='52'/>
<enumerator name='DMU_OT_BPOBJ_SUBOBJ' value='53'/>
<enumerator name='DMU_OT_NUMTYPES' value='54'/>
<enumerator name='DMU_OTN_UINT8_DATA' value='128'/>
<enumerator name='DMU_OTN_UINT8_METADATA' value='192'/>
<enumerator name='DMU_OTN_UINT16_DATA' value='129'/>
<enumerator name='DMU_OTN_UINT16_METADATA' value='193'/>
<enumerator name='DMU_OTN_UINT32_DATA' value='130'/>
<enumerator name='DMU_OTN_UINT32_METADATA' value='194'/>
<enumerator name='DMU_OTN_UINT64_DATA' value='131'/>
<enumerator name='DMU_OTN_UINT64_METADATA' value='195'/>
<enumerator name='DMU_OTN_ZAP_DATA' value='132'/>
<enumerator name='DMU_OTN_ZAP_METADATA' value='196'/>
<enumerator name='DMU_OTN_UINT8_ENC_DATA' value='160'/>
<enumerator name='DMU_OTN_UINT8_ENC_METADATA' value='224'/>
<enumerator name='DMU_OTN_UINT16_ENC_DATA' value='161'/>
<enumerator name='DMU_OTN_UINT16_ENC_METADATA' value='225'/>
<enumerator name='DMU_OTN_UINT32_ENC_DATA' value='162'/>
<enumerator name='DMU_OTN_UINT32_ENC_METADATA' value='226'/>
<enumerator name='DMU_OTN_UINT64_ENC_DATA' value='163'/>
<enumerator name='DMU_OTN_UINT64_ENC_METADATA' value='227'/>
<enumerator name='DMU_OTN_ZAP_ENC_DATA' value='164'/>
<enumerator name='DMU_OTN_ZAP_ENC_METADATA' value='228'/>
</enum-decl>
<typedef-decl name='dmu_object_type_t' type-id='04b3b0b9' id='5c9d8906'/>
<class-decl name='zio_cksum' size-in-bits='256' is-struct='yes' visibility='default' id='1d53e28b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='zc_word' type-id='85c64d26' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zio_cksum_t' type-id='1d53e28b' id='39730d0b'/>
<class-decl name='dmu_replay_record' size-in-bits='2496' is-struct='yes' visibility='default' id='781a52d7'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_type' type-id='08f5ca17' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='drr_payloadlen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_u' type-id='ac5ab598' visibility='default'/>
</data-member>
</class-decl>
<enum-decl name='__anonymous_enum__' is-anonymous='yes' id='08f5ca17'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='DRR_BEGIN' value='0'/>
<enumerator name='DRR_OBJECT' value='1'/>
<enumerator name='DRR_FREEOBJECTS' value='2'/>
<enumerator name='DRR_WRITE' value='3'/>
<enumerator name='DRR_FREE' value='4'/>
<enumerator name='DRR_END' value='5'/>
<enumerator name='DRR_WRITE_BYREF' value='6'/>
<enumerator name='DRR_SPILL' value='7'/>
<enumerator name='DRR_WRITE_EMBEDDED' value='8'/>
<enumerator name='DRR_OBJECT_RANGE' value='9'/>
<enumerator name='DRR_REDACT' value='10'/>
<enumerator name='DRR_NUMTYPES' value='11'/>
</enum-decl>
<union-decl name='__anonymous_union__' size-in-bits='2432' is-anonymous='yes' visibility='default' id='ac5ab598'>
<data-member access='public'>
<var-decl name='drr_begin' type-id='09fcdc01' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_end' type-id='6ee25631' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_object' type-id='f9ad530b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_freeobjects' type-id='a27d958e' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write' type-id='4cc69e4b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_free' type-id='c836cfd2' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write_byref' type-id='e511cdce' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_spill' type-id='1e69a80a' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_write_embedded' type-id='98b1345e' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_object_range' type-id='aba1f9e1' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_redact' type-id='50389039' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='drr_checksum' type-id='a5fe3647' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='drr_end' size-in-bits='320' is-struct='yes' visibility='default' id='6ee25631'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_checksum' type-id='39730d0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_object' size-in-bits='448' is-struct='yes' visibility='default' id='f9ad530b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='drr_bonustype' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_blksz' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='160'>
<var-decl name='drr_bonuslen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='200'>
<var-decl name='drr_compress' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='208'>
<var-decl name='drr_dn_slots' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='216'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='224'>
<var-decl name='drr_raw_bonuslen' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_indblkshift' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='328'>
<var-decl name='drr_nlevels' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='336'>
<var-decl name='drr_nblkptr' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='344'>
<var-decl name='drr_pad' type-id='0f4ddd0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_maxblkid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_freeobjects' size-in-bits='192' is-struct='yes' visibility='default' id='a27d958e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_firstobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_numobjs' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write' size-in-bits='1088' is-struct='yes' visibility='default' id='4cc69e4b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='drr_pad' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_logical_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='328'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='336'>
<var-decl name='drr_compressiontype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='344'>
<var-decl name='drr_pad2' type-id='0f4ddd0b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_key' type-id='67f6d2cf' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='drr_compressed_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='768'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='928'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_free' size-in-bits='256' is-struct='yes' visibility='default' id='c836cfd2'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write_byref' size-in-bits='832' is-struct='yes' visibility='default' id='e511cdce'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_refguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_refobject' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_refoffset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='drr_checksumtype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='456'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='464'>
<var-decl name='drr_pad2' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='drr_key' type-id='67f6d2cf' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_spill' size-in-bits='640' is-struct='yes' visibility='default' id='1e69a80a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='200'>
<var-decl name='drr_compressiontype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='208'>
<var-decl name='drr_pad' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_compressed_size' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='608'>
<var-decl name='drr_type' type-id='5c9d8906' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_write_embedded' size-in-bits='384' is-struct='yes' visibility='default' id='98b1345e'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_compression' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='264'>
<var-decl name='drr_etype' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='272'>
<var-decl name='drr_pad' type-id='0f562bd0' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='drr_lsize' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='drr_psize' type-id='8f92235e' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_object_range' size-in-bits='512' is-struct='yes' visibility='default' id='aba1f9e1'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_firstobj' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_numslots' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_salt' type-id='13339fda' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='drr_iv' type-id='fa8ef949' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='drr_mac' type-id='fa9986a5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='drr_flags' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='488'>
<var-decl name='drr_pad' type-id='d3490169' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_redact' size-in-bits='256' is-struct='yes' visibility='default' id='50389039'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_object' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='drr_offset' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='drr_length' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='drr_toguid' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='drr_checksum' size-in-bits='2432' is-struct='yes' visibility='default' id='a5fe3647'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='drr_pad' type-id='8c2bcad1' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='2176'>
<var-decl name='drr_checksum' type-id='39730d0b' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='__clockid_t' type-id='95e97e5e' id='08f9a87a'/>
<typedef-decl name='clockid_t' type-id='08f9a87a' id='a1c3b834'/>
<typedef-decl name='Byte' type-id='002ac4a6' id='efb9ba06'/>
<typedef-decl name='uLong' type-id='7359adad' id='5bbcce85'/>
<typedef-decl name='Bytef' type-id='efb9ba06' id='c1606520'/>
<typedef-decl name='uLongf' type-id='5bbcce85' id='4d39af59'/>
<pointer-type-def type-id='c1606520' size-in-bits='64' id='4c667223'/>
<qualified-type-def type-id='c1606520' const='yes' id='a6124a50'/>
<pointer-type-def type-id='a6124a50' size-in-bits='64' id='e8cb3e0e'/>
<qualified-type-def type-id='781a52d7' const='yes' id='413ab2b8'/>
<pointer-type-def type-id='413ab2b8' size-in-bits='64' id='41671bd6'/>
<pointer-type-def type-id='c70fa2e8' size-in-bits='64' id='2e711a2a'/>
<pointer-type-def type-id='3ff5601b' size-in-bits='64' id='4aafb922'/>
<pointer-type-def type-id='9e59d1d4' size-in-bits='64' id='4ea84b4f'/>
<pointer-type-def type-id='945467e6' size-in-bits='64' id='8def7735'/>
<pointer-type-def type-id='3d3ffb69' size-in-bits='64' id='72a26210'/>
<pointer-type-def type-id='c9d12d66' size-in-bits='64' id='b2eb2c3f'/>
<pointer-type-def type-id='a9c79a1f' size-in-bits='64' id='3d83ba87'/>
<pointer-type-def type-id='4d39af59' size-in-bits='64' id='60db3356'/>
<pointer-type-def type-id='39730d0b' size-in-bits='64' id='c24fc2ee'/>
<function-decl name='nvlist_print' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='822cd80b'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_get_pool_handle' mangled-name='zfs_get_pool_handle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_pool_handle'>
<parameter type-id='fcd57163'/>
<return type-id='4c81de99'/>
</function-decl>
<function-decl name='lzc_send_wrapper' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2e711a2a'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_redacted' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='bfbd3c8e'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_resume_redacted' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='bfbd3c8e'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_with_cmdprops' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='41671bd6'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_receive_with_heal' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='ae3e8ca6'/>
<parameter type-id='3502e3ff'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='41671bd6'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='857bb57e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_space' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='bfbd3c8e'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_send_space_resume_redacted' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='bfbd3c8e'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='5d6479ae'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='lzc_rename' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_setproctitle' mangled-name='zfs_setproctitle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_setproctitle'>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_insert' mangled-name='avl_insert' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_insert'>
<parameter type-id='a3681dea'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='fba6cb51'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='nvlist_add_boolean' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvlist_lookup_boolean' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='nvpair_value_int32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<parameter type-id='4aafb922'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fnvlist_size' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='fnvlist_merge' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_remove' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_lookup_boolean_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='fletcher_4_native_varsize' mangled-name='fletcher_4_native_varsize' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_native_varsize'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='c24fc2ee'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_incremental_native' mangled-name='fletcher_4_incremental_native' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_incremental_native'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fletcher_4_incremental_byteswap' mangled-name='fletcher_4_incremental_byteswap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_incremental_byteswap'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='perror' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strndup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='time' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b2eb2c3f'/>
<return type-id='c9d12d66'/>
</function-decl>
<function-decl name='clock_gettime' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a1c3b834'/>
<parameter type-id='3d83ba87'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='write' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='sleep' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f0981eeb'/>
<return type-id='f0981eeb'/>
</function-decl>
<function-decl name='uncompress' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4c667223'/>
<parameter type-id='60db3356'/>
<parameter type-id='e8cb3e0e'/>
<parameter type-id='5bbcce85'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='create_parents' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<parameter type-id='26a90f95'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send_progress' mangled-name='zfs_send_progress' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_progress'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='5d6479ae' name='bytes_written'/>
<parameter type-id='5d6479ae' name='blocks_visited'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send_resume_token_to_nvlist' mangled-name='zfs_send_resume_token_to_nvlist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_resume_token_to_nvlist'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='token'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zfs_send_resume' mangled-name='zfs_send_resume' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_resume'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='8def7735' name='flags'/>
<parameter type-id='95e97e5e' name='outfd'/>
<parameter type-id='80f4b756' name='resume_token'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send_saved' mangled-name='zfs_send_saved' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_saved'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='8def7735' name='flags'/>
<parameter type-id='95e97e5e' name='outfd'/>
<parameter type-id='80f4b756' name='resume_token'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send' mangled-name='zfs_send' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='fromsnap'/>
<parameter type-id='80f4b756' name='tosnap'/>
<parameter type-id='8def7735' name='flags'/>
<parameter type-id='95e97e5e' name='outfd'/>
<parameter type-id='72a26210' name='filter_func'/>
<parameter type-id='eaa32e2f' name='cb_arg'/>
<parameter type-id='857bb57e' name='debugnvp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_send_one' mangled-name='zfs_send_one' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_send_one'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='from'/>
<parameter type-id='95e97e5e' name='fd'/>
<parameter type-id='8def7735' name='flags'/>
<parameter type-id='80f4b756' name='redactbook'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_receive' mangled-name='zfs_receive' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_receive'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='tosnap'/>
<parameter type-id='5ce45b60' name='props'/>
<parameter type-id='4ea84b4f' name='flags'/>
<parameter type-id='95e97e5e' name='infd'/>
<parameter type-id='a3681dea' name='stream_avl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='c70fa2e8'>
<parameter type-id='95e97e5e'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='d2a5e211'>
<parameter type-id='9200a744'/>
<parameter type-id='eaa32e2f'/>
<return type-id='c19b74c3'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_status.c' language='LANG_C99'>
<function-decl name='zpool_import_status' mangled-name='zpool_import_status' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_import_status'>
<parameter type-id='5ce45b60' name='config'/>
<parameter type-id='7d3cd834' name='msgid'/>
<parameter type-id='cec6f2e4' name='errata'/>
<return type-id='d3dd6294'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/libzfs_util.c' language='LANG_C99'>
<class-decl name='__va_list_tag' size-in-bits='192' is-struct='yes' visibility='default' id='d5027220'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='gp_offset' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='fp_offset' type-id='f0981eeb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='overflow_arg_area' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='reg_save_area' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<type-decl name='double' size-in-bits='64' id='a0eb0f08'/>
<array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='192' id='e41bdf22'>
<subrange length='6' type-id='7359adad' id='52fa524b'/>
</array-type-def>
<array-type-def dimensions='1' type-id='19cefcee' size-in-bits='160' alignment-in-bits='32' id='3fcf57d2'>
<subrange length='5' type-id='7359adad' id='53010e10'/>
</array-type-def>
<enum-decl name='zfs_get_column_t' naming-typedef-id='19cefcee' id='223bdcaa'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='GET_COL_NONE' value='0'/>
<enumerator name='GET_COL_NAME' value='1'/>
<enumerator name='GET_COL_PROPERTY' value='2'/>
<enumerator name='GET_COL_VALUE' value='3'/>
<enumerator name='GET_COL_RECVD' value='4'/>
<enumerator name='GET_COL_SOURCE' value='5'/>
</enum-decl>
<typedef-decl name='zfs_get_column_t' type-id='223bdcaa' id='19cefcee'/>
<class-decl name='vdev_cbdata' size-in-bits='192' is-struct='yes' visibility='default' id='b8006be8'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='cb_name_flags' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='cb_names' type-id='9b23c9ad' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='cb_names_count' type-id='f0981eeb' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='vdev_cbdata_t' type-id='b8006be8' id='a9679c94'/>
<class-decl name='zprop_get_cbdata' size-in-bits='832' is-struct='yes' visibility='default' id='f3d3c319'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='cb_sources' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='cb_columns' type-id='3fcf57d2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='cb_colwidths' type-id='e41bdf22' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='cb_scripted' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='416'>
<var-decl name='cb_literal' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='cb_first' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='cb_proplist' type-id='3a9b2288' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='cb_type' type-id='2e45de5d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='640'>
<var-decl name='cb_vdevs' type-id='a9679c94' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zprop_get_cbdata_t' type-id='f3d3c319' id='f3d87113'/>
<typedef-decl name='zprop_func' type-id='2e711a2a' id='1ec3747a'/>
<enum-decl name='zprop_attr_t' naming-typedef-id='999701cc' id='77d05200'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='PROP_DEFAULT' value='0'/>
<enumerator name='PROP_READONLY' value='1'/>
<enumerator name='PROP_INHERIT' value='2'/>
<enumerator name='PROP_ONETIME' value='3'/>
<enumerator name='PROP_ONETIME_DEFAULT' value='4'/>
</enum-decl>
<typedef-decl name='zprop_attr_t' type-id='77d05200' id='999701cc'/>
<class-decl name='zfs_index' size-in-bits='128' is-struct='yes' visibility='default' id='87957af9'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pi_name' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pi_value' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zprop_index_t' type-id='87957af9' id='64636ce3'/>
<class-decl name='zprop_desc_t' size-in-bits='640' is-struct='yes' naming-typedef-id='ffa52b96' visibility='default' id='bbff5e4b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='pd_name' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='pd_propnum' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='pd_proptype' type-id='31429eff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='pd_strdefault' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='pd_numdefault' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='pd_attr' type-id='999701cc' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='pd_types' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='pd_values' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='pd_colname' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='pd_rightalign' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='449'>
<var-decl name='pd_visible' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='450'>
<var-decl name='pd_zfs_mod_supported' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='451'>
<var-decl name='pd_always_flex' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='pd_table' type-id='c8bc397b' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='pd_table_size' type-id='b59d7dce' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zprop_desc_t' type-id='bbff5e4b' id='ffa52b96'/>
<class-decl name='extmnttab' size-in-bits='320' is-struct='yes' visibility='default' id='0c544dc0'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='mnt_special' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='mnt_mountp' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='mnt_fstype' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='mnt_mntopts' type-id='26a90f95' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='mnt_major' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='mnt_minor' type-id='3502e3ff' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='d5027220' size-in-bits='64' id='b7f2d5e6'/>
<qualified-type-def type-id='26a90f95' const='yes' id='57de658a'/>
<pointer-type-def type-id='57de658a' size-in-bits='64' id='f319fae0'/>
<pointer-type-def type-id='9b23c9ad' size-in-bits='64' id='c0563f85'/>
<qualified-type-def type-id='33f57a65' const='yes' id='21fd6035'/>
<pointer-type-def type-id='21fd6035' size-in-bits='64' id='a0de50cd'/>
<pointer-type-def type-id='a0de50cd' size-in-bits='64' id='24f95ba5'/>
<qualified-type-def type-id='64636ce3' const='yes' id='072f7953'/>
<pointer-type-def type-id='072f7953' size-in-bits='64' id='c8bc397b'/>
<pointer-type-def type-id='0c544dc0' size-in-bits='64' id='394fc496'/>
<pointer-type-def type-id='aca3bac8' size-in-bits='64' id='d33f11cb'/>
<qualified-type-def type-id='d33f11cb' restrict='yes' id='5c53ba29'/>
<pointer-type-def type-id='ffa52b96' size-in-bits='64' id='76c8174b'/>
<pointer-type-def type-id='f3d87113' size-in-bits='64' id='0d2a0670'/>
<function-decl name='zfs_version_kernel' mangled-name='zfs_version_kernel' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_version_kernel'>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='libzfs_core_init' visibility='default' binding='global' size-in-bits='64'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_core_fini' visibility='default' binding='global' size-in-bits='64'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_prop_unsupported' mangled-name='zpool_prop_unsupported' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_unsupported'>
<parameter type-id='80f4b756'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zpool_feature_init' mangled-name='zpool_feature_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_feature_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_init' mangled-name='fletcher_4_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_fini' mangled-name='fletcher_4_fini' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_fini'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_prop_init' mangled-name='zfs_prop_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_prop_get_table' mangled-name='zfs_prop_get_table' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_get_table'>
<return type-id='76c8174b'/>
</function-decl>
<function-decl name='zpool_prop_init' mangled-name='zpool_prop_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zpool_prop_get_table' mangled-name='zpool_prop_get_table' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_get_table'>
<return type-id='76c8174b'/>
</function-decl>
<function-decl name='vdev_prop_init' mangled-name='vdev_prop_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_init'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_iter_common' mangled-name='zprop_iter_common' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_iter_common'>
<parameter type-id='1ec3747a'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_name_to_prop' mangled-name='zprop_name_to_prop' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_name_to_prop'>
<parameter type-id='80f4b756'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_string_to_index' mangled-name='zprop_string_to_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_string_to_index'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='5d6479ae'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_values' mangled-name='zprop_values' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_values'>
<parameter type-id='95e97e5e'/>
<parameter type-id='2e45de5d'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zprop_width' mangled-name='zprop_width' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_width'>
<parameter type-id='95e97e5e'/>
<parameter type-id='37e3bd22'/>
<parameter type-id='2e45de5d'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='zprop_valid_for_type' mangled-name='zprop_valid_for_type' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_valid_for_type'>
<parameter type-id='95e97e5e'/>
<parameter type-id='2e45de5d'/>
<parameter type-id='c19b74c3'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='getextmntent' mangled-name='getextmntent' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='getextmntent'>
<parameter type-id='80f4b756'/>
<parameter type-id='394fc496'/>
<parameter type-id='62f7a03d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__ctype_toupper_loc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='24f95ba5'/>
</function-decl>
<function-decl name='dlclose' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='regcomp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5c53ba29'/>
<parameter type-id='9d26089a'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='regfree' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='d33f11cb'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='puts' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='strtod' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='8c85230f'/>
<return type-id='a0eb0f08'/>
</function-decl>
<function-decl name='realloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='exit' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='strnlen' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='strncasecmp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='dup2' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='execve' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='f319fae0'/>
<parameter type-id='f319fae0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='execv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='f319fae0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='execvp' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='f319fae0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='execvpe' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='f319fae0'/>
<parameter type-id='f319fae0'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='_exit' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fork' visibility='default' binding='global' size-in-bits='64'>
<return type-id='3629bad8'/>
</function-decl>
<function-decl name='pow' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a0eb0f08'/>
<parameter type-id='a0eb0f08'/>
<return type-id='a0eb0f08'/>
</function-decl>
<function-decl name='__vfprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e75a27e9'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b7f2d5e6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__vasprintf_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='8c85230f'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='9d26089a'/>
<parameter type-id='b7f2d5e6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='waitpid' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3629bad8'/>
<parameter type-id='7292109c'/>
<parameter type-id='95e97e5e'/>
<return type-id='3629bad8'/>
</function-decl>
<function-decl name='namespace_clear' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b0382bb3'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_load_module' visibility='default' binding='global' size-in-bits='64'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_errno' mangled-name='libzfs_errno' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_errno'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_error_action' mangled-name='libzfs_error_action' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_error_action'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='libzfs_error_description' mangled-name='libzfs_error_description' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_error_description'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='libzfs_print_on_error' mangled-name='libzfs_print_on_error' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_print_on_error'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='c19b74c3' name='printerr'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_run_process' mangled-name='libzfs_run_process' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_run_process'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='95e97e5e' name='flags'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_run_process_get_stdout' mangled-name='libzfs_run_process_get_stdout' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_run_process_get_stdout'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='9b23c9ad' name='env'/>
<parameter type-id='c0563f85' name='lines'/>
<parameter type-id='7292109c' name='lines_cnt'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_run_process_get_stdout_nopath' mangled-name='libzfs_run_process_get_stdout_nopath' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_run_process_get_stdout_nopath'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='9b23c9ad' name='env'/>
<parameter type-id='c0563f85' name='lines'/>
<parameter type-id='7292109c' name='lines_cnt'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_free_str_array' mangled-name='libzfs_free_str_array' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_free_str_array'>
<parameter type-id='9b23c9ad' name='strs'/>
<parameter type-id='95e97e5e' name='count'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='libzfs_init' mangled-name='libzfs_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_init'>
<return type-id='b0382bb3'/>
</function-decl>
<function-decl name='libzfs_fini' mangled-name='libzfs_fini' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_fini'>
<parameter type-id='b0382bb3' name='hdl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_path_to_zhandle' mangled-name='zfs_path_to_zhandle' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_path_to_zhandle'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='2e45de5d' name='argtype'/>
<return type-id='9200a744'/>
</function-decl>
<function-decl name='zprop_print_one_property' mangled-name='zprop_print_one_property' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_print_one_property'>
<parameter type-id='80f4b756' name='name'/>
<parameter type-id='0d2a0670' name='cbp'/>
<parameter type-id='80f4b756' name='propname'/>
<parameter type-id='80f4b756' name='value'/>
<parameter type-id='a2256d42' name='sourcetype'/>
<parameter type-id='80f4b756' name='source'/>
<parameter type-id='80f4b756' name='recvd_value'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_get_list' mangled-name='zprop_get_list' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_get_list'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='26a90f95' name='props'/>
<parameter type-id='e4378506' name='listp'/>
<parameter type-id='2e45de5d' name='type'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_free_list' mangled-name='zprop_free_list' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_free_list'>
<parameter type-id='3a9b2288' name='pl'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_iter' mangled-name='zprop_iter' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_iter'>
<parameter type-id='1ec3747a' name='func'/>
<parameter type-id='eaa32e2f' name='cb'/>
<parameter type-id='c19b74c3' name='show_all'/>
<parameter type-id='c19b74c3' name='ordered'/>
<parameter type-id='2e45de5d' name='type'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_version_userland' mangled-name='zfs_version_userland' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_version_userland'>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_version_print' mangled-name='zfs_version_print' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_version_print'>
<return type-id='95e97e5e'/>
</function-decl>
+ <function-decl name='use_color' mangled-name='use_color' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='use_color'>
+ <return type-id='95e97e5e'/>
+ </function-decl>
<function-decl name='printf_color' mangled-name='printf_color' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='printf_color'>
<parameter type-id='80f4b756' name='color'/>
<parameter type-id='80f4b756' name='format'/>
<parameter is-variadic='yes'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/os/linux/libzfs_mount_os.c' language='LANG_C99'>
<pointer-type-def type-id='7359adad' size-in-bits='64' id='1d2c2b85'/>
<function-decl name='geteuid' visibility='default' binding='global' size-in-bits='64'>
<return type-id='cc5fcceb'/>
</function-decl>
<function-decl name='mount' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='7359adad'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='umount2' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_parse_mount_options' mangled-name='zfs_parse_mount_options' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_parse_mount_options'>
<parameter type-id='80f4b756' name='mntopts'/>
<parameter type-id='1d2c2b85' name='mntflags'/>
<parameter type-id='1d2c2b85' name='zfsflags'/>
<parameter type-id='95e97e5e' name='sloppy'/>
<parameter type-id='26a90f95' name='badopt'/>
<parameter type-id='26a90f95' name='mtabopt'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_adjust_mount_options' mangled-name='zfs_adjust_mount_options' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_adjust_mount_options'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='mntpoint'/>
<parameter type-id='26a90f95' name='mntopts'/>
<parameter type-id='26a90f95' name='mtabopt'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_mount_delegation_check' mangled-name='zfs_mount_delegation_check' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_mount_delegation_check'>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/os/linux/libzfs_pool_os.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='288' id='16e6f2c6'>
<subrange length='36' type-id='7359adad' id='ae666bde'/>
</array-type-def>
<array-type-def dimensions='1' type-id='a65ae39c' size-in-bits='960' id='fa198beb'>
<subrange length='1' type-id='7359adad' id='52f813b4'/>
</array-type-def>
<array-type-def dimensions='1' type-id='3502e3ff' size-in-bits='384' id='dba89ba3'>
<subrange length='12' type-id='7359adad' id='84827bdc'/>
</array-type-def>
<array-type-def dimensions='1' type-id='3502e3ff' size-in-bits='256' id='01d84ed4'>
<subrange length='8' type-id='7359adad' id='56e0c0b1'/>
</array-type-def>
<class-decl name='dk_part' size-in-bits='960' is-struct='yes' visibility='default' id='a65ae39c'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='p_start' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='p_size' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='p_guid' type-id='214f32ea' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='p_tag' type-id='d908a348' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='272'>
<var-decl name='p_flag' type-id='d908a348' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='p_name' type-id='16e6f2c6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='p_uguid' type-id='214f32ea' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='704'>
<var-decl name='p_resv' type-id='01d84ed4' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='dk_gpt' size-in-bits='1920' is-struct='yes' visibility='default' id='dd4a2e5a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='efi_version' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='efi_nparts' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='efi_part_size' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='efi_lbasize' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='efi_last_lba' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='efi_first_u_lba' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='efi_last_u_lba' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='efi_disk_uguid' type-id='214f32ea' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='efi_flags' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='480'>
<var-decl name='efi_reserved1' type-id='3502e3ff' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='efi_altern_lba' type-id='804dc465' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='576'>
<var-decl name='efi_reserved' type-id='dba89ba3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='efi_parts' type-id='fa198beb' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='uuid' size-in-bits='128' is-struct='yes' visibility='default' id='214f32ea'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='time_low' type-id='8f92235e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='time_mid' type-id='149c6638' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='48'>
<var-decl name='time_hi_and_version' type-id='149c6638' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='clock_seq_hi_and_reserved' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='72'>
<var-decl name='clock_seq_low' type-id='b96825af' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='80'>
<var-decl name='node_addr' type-id='0f562bd0' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ushort_t' type-id='8efea9e5' id='d908a348'/>
<typedef-decl name='uint16_t' type-id='253c2d2a' id='149c6638'/>
<typedef-decl name='__uint16_t' type-id='8efea9e5' id='253c2d2a'/>
<pointer-type-def type-id='dd4a2e5a' size-in-bits='64' id='0d8119a8'/>
<pointer-type-def type-id='0d8119a8' size-in-bits='64' id='c43b27a6'/>
<function-decl name='zpool_label_disk_wait' mangled-name='zpool_label_disk_wait' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_label_disk_wait'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_append_partition' mangled-name='zfs_append_partition' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_append_partition'>
<parameter type-id='26a90f95'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_alloc_and_init' mangled-name='efi_alloc_and_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_alloc_and_init'>
<parameter type-id='95e97e5e'/>
<parameter type-id='8f92235e'/>
<parameter type-id='c43b27a6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_alloc_and_read' mangled-name='efi_alloc_and_read' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_alloc_and_read'>
<parameter type-id='95e97e5e'/>
<parameter type-id='c43b27a6'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_write' mangled-name='efi_write' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_write'>
<parameter type-id='95e97e5e'/>
<parameter type-id='0d8119a8'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_rescan' mangled-name='efi_rescan' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_rescan'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='efi_free' mangled-name='efi_free' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_free'>
<parameter type-id='0d8119a8'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='efi_use_whole_disk' mangled-name='efi_use_whole_disk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='efi_use_whole_disk'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='rand' visibility='default' binding='global' size-in-bits='64'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fsync' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_label_disk' mangled-name='zpool_label_disk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_label_disk'>
<parameter type-id='b0382bb3' name='hdl'/>
<parameter type-id='4c81de99' name='zhp'/>
<parameter type-id='80f4b756' name='name'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzfs/os/linux/libzfs_util_os.c' language='LANG_C99'>
<class-decl name='itimerspec' size-in-bits='256' is-struct='yes' visibility='default' id='acbdbcc6'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='it_interval' type-id='a9c79a1f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='it_value' type-id='a9c79a1f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='nfds_t' type-id='7359adad' id='555eef66'/>
<class-decl name='pollfd' size-in-bits='64' is-struct='yes' visibility='default' id='b440e872'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='fd' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='events' type-id='a2185560' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='48'>
<var-decl name='revents' type-id='a2185560' visibility='default'/>
</data-member>
</class-decl>
<qualified-type-def type-id='acbdbcc6' const='yes' id='4ba62af7'/>
<pointer-type-def type-id='4ba62af7' size-in-bits='64' id='f39579e7'/>
<pointer-type-def type-id='acbdbcc6' size-in-bits='64' id='116842ac'/>
<pointer-type-def type-id='b440e872' size-in-bits='64' id='3ac36db0'/>
<function-decl name='access' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='__poll_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='3ac36db0'/>
<parameter type-id='555eef66'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='7359adad'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='inotify_init1' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='inotify_add_watch' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='8f92235e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='timerfd_create' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='08f9a87a'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='timerfd_settime' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='f39579e7'/>
<parameter type-id='116842ac'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='libzfs_error_init' mangled-name='libzfs_error_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libzfs_error_init'>
<parameter type-id='95e97e5e' name='error'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_userns' mangled-name='zfs_userns' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_userns'>
<parameter type-id='9200a744' name='zhp'/>
<parameter type-id='80f4b756' name='nspath'/>
<parameter type-id='95e97e5e' name='attach'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/os/linux/zutil_device_path_os.c' language='LANG_C99'>
<class-decl name='udev' is-struct='yes' visibility='default' is-declaration-only='yes' id='e4a7fb7f'/>
<class-decl name='udev_device' is-struct='yes' visibility='default' is-declaration-only='yes' id='640b33ca'/>
<pointer-type-def type-id='e4a7fb7f' size-in-bits='64' id='025eefe7'/>
<pointer-type-def type-id='640b33ca' size-in-bits='64' id='b32bae08'/>
<class-decl name='udev' is-struct='yes' visibility='default' is-declaration-only='yes' id='e4a7fb7f'/>
<class-decl name='udev_device' is-struct='yes' visibility='default' is-declaration-only='yes' id='640b33ca'/>
<function-decl name='udev_new' visibility='default' binding='global' size-in-bits='64'>
<return type-id='025eefe7'/>
</function-decl>
<function-decl name='udev_device_unref' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b32bae08'/>
<return type-id='b32bae08'/>
</function-decl>
<function-decl name='udev_device_new_from_subsystem_sysname' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='025eefe7'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b32bae08'/>
</function-decl>
<function-decl name='udev_device_get_property_value' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b32bae08'/>
<parameter type-id='80f4b756'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='__readlink_chk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='9d26089a'/>
<parameter type-id='266fe297'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='79a0948f'/>
</function-decl>
<function-decl name='zfs_get_enclosure_sysfs_path' mangled-name='zfs_get_enclosure_sysfs_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_enclosure_sysfs_path'>
<parameter type-id='80f4b756' name='dev_name'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zfs_dev_is_dm' mangled-name='zfs_dev_is_dm' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dev_is_dm'>
<parameter type-id='80f4b756' name='dev_name'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_dev_is_whole_disk' mangled-name='zfs_dev_is_whole_disk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dev_is_whole_disk'>
<parameter type-id='80f4b756' name='dev_name'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_get_underlying_path' mangled-name='zfs_get_underlying_path' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_get_underlying_path'>
<parameter type-id='80f4b756' name='dev_name'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='is_mpath_whole_disk' mangled-name='is_mpath_whole_disk' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='is_mpath_whole_disk'>
<parameter type-id='80f4b756' name='path'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/os/linux/zutil_import_os.c' language='LANG_C99'>
<class-decl name='blkid_struct_cache' is-struct='yes' visibility='default' is-declaration-only='yes' id='09286066'/>
<class-decl name='blkid_struct_dev' is-struct='yes' visibility='default' is-declaration-only='yes' id='86223623'/>
<class-decl name='blkid_struct_dev_iterate' is-struct='yes' visibility='default' is-declaration-only='yes' id='d88420d6'/>
<class-decl name='udev_list_entry' is-struct='yes' visibility='default' is-declaration-only='yes' id='e7dbdca3'/>
<typedef-decl name='pool_vdev_iter_f' type-id='6c16a6c8' id='dff793e0'/>
<typedef-decl name='blkid_dev' type-id='8433f053' id='f47b023a'/>
<typedef-decl name='blkid_cache' type-id='940e3afc' id='0882dfdf'/>
<typedef-decl name='blkid_dev_iterate' type-id='b8fa2efc' id='f4760fa7'/>
<typedef-decl name='__useconds_t' type-id='f0981eeb' id='4e80d4b1'/>
<pointer-type-def type-id='0882dfdf' size-in-bits='64' id='2e3e7caa'/>
<pointer-type-def type-id='f47b023a' size-in-bits='64' id='d87f9b75'/>
<pointer-type-def type-id='09286066' size-in-bits='64' id='940e3afc'/>
<pointer-type-def type-id='86223623' size-in-bits='64' id='8433f053'/>
<pointer-type-def type-id='d88420d6' size-in-bits='64' id='b8fa2efc'/>
<pointer-type-def type-id='2ec2411e' size-in-bits='64' id='6c16a6c8'/>
<pointer-type-def type-id='e7dbdca3' size-in-bits='64' id='deabd0d3'/>
<class-decl name='blkid_struct_cache' is-struct='yes' visibility='default' is-declaration-only='yes' id='09286066'/>
<class-decl name='blkid_struct_dev' is-struct='yes' visibility='default' is-declaration-only='yes' id='86223623'/>
<class-decl name='blkid_struct_dev_iterate' is-struct='yes' visibility='default' is-declaration-only='yes' id='d88420d6'/>
<class-decl name='udev_list_entry' is-struct='yes' visibility='default' is-declaration-only='yes' id='e7dbdca3'/>
<function-decl name='for_each_vdev_in_nvlist' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<parameter type-id='dff793e0'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='label_paths' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5507783b'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='7d3cd834'/>
<parameter type-id='7d3cd834'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zutil_alloc' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5507783b'/>
<parameter type-id='b59d7dce'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='zutil_strdup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5507783b'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='slice_cache_compare' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='blkid_put_cache' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0882dfdf'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='blkid_get_cache' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='2e3e7caa'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='blkid_dev_devname' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f47b023a'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='blkid_dev_iterate_begin' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0882dfdf'/>
<return type-id='f4760fa7'/>
</function-decl>
<function-decl name='blkid_dev_set_search' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f4760fa7'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='blkid_dev_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f4760fa7'/>
<parameter type-id='d87f9b75'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='blkid_dev_iterate_end' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='f4760fa7'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='blkid_probe_all_new' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='0882dfdf'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='udev_unref' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='025eefe7'/>
<return type-id='025eefe7'/>
</function-decl>
<function-decl name='udev_list_entry_get_next' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='deabd0d3'/>
<return type-id='deabd0d3'/>
</function-decl>
<function-decl name='udev_list_entry_get_name' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='deabd0d3'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='udev_device_get_parent_with_subsystem_devtype' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b32bae08'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b32bae08'/>
</function-decl>
<function-decl name='udev_device_get_devlinks_list_entry' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='b32bae08'/>
<return type-id='deabd0d3'/>
</function-decl>
<function-decl name='sched_yield' visibility='default' binding='global' size-in-bits='64'>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='usleep' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='4e80d4b1'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_dev_flush' mangled-name='zfs_dev_flush' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dev_flush'>
<parameter type-id='95e97e5e' name='fd'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_device_get_devid' mangled-name='zfs_device_get_devid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_device_get_devid'>
<parameter type-id='b32bae08' name='dev'/>
<parameter type-id='26a90f95' name='bufptr'/>
<parameter type-id='b59d7dce' name='buflen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_device_get_physical' mangled-name='zfs_device_get_physical' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_device_get_physical'>
<parameter type-id='b32bae08' name='dev'/>
<parameter type-id='26a90f95' name='bufptr'/>
<parameter type-id='b59d7dce' name='buflen'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-type size-in-bits='64' id='2ec2411e'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='5ce45b60'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/os/linux/zutil_setproctitle.c' language='LANG_C99'>
<function-decl name='warnx' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter is-variadic='yes'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='setenv' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='95e97e5e'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_setproctitle_init' mangled-name='zfs_setproctitle_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_setproctitle_init'>
<parameter type-id='95e97e5e' name='argc'/>
<parameter type-id='9b23c9ad' name='argv'/>
<parameter type-id='9b23c9ad' name='envp'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/zutil_device_path.c' language='LANG_C99'>
<pointer-type-def type-id='b99c00c9' size-in-bits='64' id='13956559'/>
<function-decl name='zpool_default_search_paths' mangled-name='zpool_default_search_paths' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_default_search_paths'>
<parameter type-id='78c01427'/>
<return type-id='13956559'/>
</function-decl>
<function-decl name='strspn' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='zfs_dirnamelen' mangled-name='zfs_dirnamelen' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dirnamelen'>
<parameter type-id='80f4b756' name='path'/>
<return type-id='79a0948f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/zutil_import.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='a84c031d' size-in-bits='256' id='16dc656a'>
<subrange length='32' type-id='7359adad' id='ae5bde82'/>
</array-type-def>
<array-type-def dimensions='1' type-id='95e97e5e' size-in-bits='384' id='73b82f0f'>
<subrange length='12' type-id='7359adad' id='84827bdc'/>
</array-type-def>
<class-decl name='importargs' size-in-bits='448' is-struct='yes' visibility='default' id='7ac83801'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='path' type-id='9b23c9ad' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='paths' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='poolname' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='guid' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='cachefile' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='can_be_active' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='352'>
<var-decl name='scan' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='policy' type-id='5ce45b60' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='importargs_t' type-id='7ac83801' id='7a842a6b'/>
<class-decl name='libpc_handle' size-in-bits='8448' is-struct='yes' visibility='default' id='7c8737f0'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='lpc_error' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='lpc_printerr' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='lpc_open_access_error' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='lpc_desc_active' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='lpc_desc' type-id='b54ce520' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8320'>
<var-decl name='lpc_ops' type-id='f095e320' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='8384'>
<var-decl name='lpc_lib_handle' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='libpc_handle_t' type-id='7c8737f0' id='8a70a786'/>
<class-decl name='aiocb' size-in-bits='1344' is-struct='yes' visibility='default' id='e4957c49'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='aio_fildes' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='32'>
<var-decl name='aio_lio_opcode' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='aio_reqprio' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='aio_buf' type-id='fe09dd29' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='aio_nbytes' type-id='b59d7dce' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='aio_sigevent' type-id='519bc206' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='768'>
<var-decl name='__next_prio' type-id='924bbc81' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='832'>
<var-decl name='__abs_prio' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='864'>
<var-decl name='__policy' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='896'>
<var-decl name='__error_code' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='960'>
<var-decl name='__return_value' type-id='41060289' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1024'>
<var-decl name='aio_offset' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='1088'>
<var-decl name='__glibc_reserved' type-id='16dc656a' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='sigevent' size-in-bits='512' is-struct='yes' visibility='default' id='519bc206'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='sigev_value' type-id='eabacd01' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='sigev_signo' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='96'>
<var-decl name='sigev_notify' type-id='95e97e5e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='_sigev_un' type-id='ac5ab599' visibility='default'/>
</data-member>
</class-decl>
<union-decl name='__anonymous_union__' size-in-bits='384' is-anonymous='yes' visibility='default' id='ac5ab599'>
<data-member access='public'>
<var-decl name='_pad' type-id='73b82f0f' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_tid' type-id='3629bad8' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='_sigev_thread' type-id='e7f43f7b' visibility='default'/>
</data-member>
</union-decl>
<class-decl name='__anonymous_struct__' size-in-bits='128' is-struct='yes' is-anonymous='yes' visibility='default' id='e7f43f7b'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='_function' type-id='5f147c28' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='_attribute' type-id='7347a39e' visibility='default'/>
</data-member>
</class-decl>
<pointer-type-def type-id='e4957c49' size-in-bits='64' id='924bbc81'/>
<qualified-type-def type-id='924bbc81' const='yes' id='5499dcde'/>
<pointer-type-def type-id='5499dcde' size-in-bits='64' id='2236d41c'/>
<qualified-type-def type-id='2236d41c' restrict='yes' id='31488924'/>
<pointer-type-def type-id='a3681dea' size-in-bits='64' id='fce6d540'/>
<qualified-type-def type-id='e4957c49' const='yes' id='fced9da2'/>
<pointer-type-def type-id='fced9da2' size-in-bits='64' id='b20efd18'/>
<pointer-type-def type-id='7a842a6b' size-in-bits='64' id='07ee4a58'/>
<pointer-type-def type-id='8a70a786' size-in-bits='64' id='5507783b'/>
<pointer-type-def type-id='b1e62775' size-in-bits='64' id='f095e320'/>
<pointer-type-def type-id='519bc206' size-in-bits='64' id='ef2f159c'/>
<qualified-type-def type-id='ef2f159c' restrict='yes' id='de0eb5a4'/>
<pointer-type-def type-id='f1abb096' size-in-bits='64' id='5f147c28'/>
<qualified-type-def type-id='48b5725f' volatile='yes' id='b0b3cbf9'/>
<pointer-type-def type-id='b0b3cbf9' size-in-bits='64' id='fe09dd29'/>
<function-decl name='update_vdev_config_dev_strs' mangled-name='update_vdev_config_dev_strs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='update_vdev_config_dev_strs'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='update_vdevs_config_dev_sysfs_path' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5ce45b60'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fnvlist_dup' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='22cce67b'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='spl_pagesize' mangled-name='spl_pagesize' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='spl_pagesize'>
<return type-id='b59d7dce'/>
</function-decl>
<function-decl name='posix_memalign' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='63e171df'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='b59d7dce'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='sysconf' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='95e97e5e'/>
<return type-id='bd54fe1a'/>
</function-decl>
<function-decl name='libpc_error_description' mangled-name='libpc_error_description' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='libpc_error_description'>
<parameter type-id='5507783b' name='hdl'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_search_import' mangled-name='zpool_search_import' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_search_import'>
<parameter type-id='5507783b' name='hdl'/>
<parameter type-id='07ee4a58' name='import'/>
<return type-id='5ce45b60'/>
</function-decl>
<function-decl name='zpool_find_config' mangled-name='zpool_find_config' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_find_config'>
<parameter type-id='5507783b' name='hdl'/>
<parameter type-id='80f4b756' name='target'/>
<parameter type-id='857bb57e' name='configp'/>
<parameter type-id='07ee4a58' name='args'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_find_import_blkid' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='5507783b'/>
<parameter type-id='18c91f9e'/>
<parameter type-id='fce6d540'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_open_func' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='f1abb096'>
<parameter type-id='eabacd01'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/zutil_nicenum.c' language='LANG_C99'>
<type-decl name='long double' size-in-bits='128' id='e095c704'/>
<enum-decl name='zfs_nicenum_format' id='29cf1969'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_NICENUM_1024' value='0'/>
<enumerator name='ZFS_NICENUM_BYTES' value='1'/>
<enumerator name='ZFS_NICENUM_TIME' value='2'/>
<enumerator name='ZFS_NICENUM_RAW' value='3'/>
<enumerator name='ZFS_NICENUM_RAWTIME' value='4'/>
</enum-decl>
<function-decl name='powl' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='e095c704'/>
<parameter type-id='e095c704'/>
<return type-id='e095c704'/>
</function-decl>
<function-decl name='floor' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='a0eb0f08'/>
<return type-id='a0eb0f08'/>
</function-decl>
<function-decl name='zfs_isnumber' mangled-name='zfs_isnumber' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_isnumber'>
<parameter type-id='80f4b756' name='str'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_nicenum_format' mangled-name='zfs_nicenum_format' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicenum_format'>
<parameter type-id='9c313c2d' name='num'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='buflen'/>
<parameter type-id='29cf1969' name='format'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_nicetime' mangled-name='zfs_nicetime' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_nicetime'>
<parameter type-id='9c313c2d' name='num'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='buflen'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfs_niceraw' mangled-name='zfs_niceraw' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_niceraw'>
<parameter type-id='9c313c2d' name='num'/>
<parameter type-id='26a90f95' name='buf'/>
<parameter type-id='b59d7dce' name='buflen'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='lib/libzutil/zutil_pool.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='853fd5dc' size-in-bits='32768' id='b505fc2f'>
<subrange length='64' type-id='7359adad' id='b10be967'/>
</array-type-def>
<class-decl name='ddt_stat' size-in-bits='512' is-struct='yes' visibility='default' id='65242dfe'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='dds_blocks' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='dds_lsize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='dds_psize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='dds_dsize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='dds_ref_blocks' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='dds_ref_lsize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='dds_ref_psize' type-id='9c313c2d' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='dds_ref_dsize' type-id='9c313c2d' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ddt_stat_t' type-id='65242dfe' id='853fd5dc'/>
<class-decl name='ddt_histogram' size-in-bits='32768' is-struct='yes' visibility='default' id='bc2b3086'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='ddh_stat' type-id='b505fc2f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='ddt_histogram_t' type-id='bc2b3086' id='2d7fe832'/>
<qualified-type-def type-id='2d7fe832' const='yes' id='ec92d602'/>
<pointer-type-def type-id='ec92d602' size-in-bits='64' id='932720f8'/>
<qualified-type-def type-id='853fd5dc' const='yes' id='764c298c'/>
<pointer-type-def type-id='764c298c' size-in-bits='64' id='dfe59052'/>
<function-decl name='zpool_dump_ddt' mangled-name='zpool_dump_ddt' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_dump_ddt'>
<parameter type-id='dfe59052' name='dds_total'/>
<parameter type-id='932720f8' name='ddh'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/avl/avl.c' language='LANG_C99'>
<function-decl name='avl_last' mangled-name='avl_last' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_last'>
<parameter type-id='a3681dea' name='tree'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_nearest' mangled-name='avl_nearest' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_nearest'>
<parameter type-id='a3681dea' name='tree'/>
<parameter type-id='fba6cb51' name='where'/>
<parameter type-id='95e97e5e' name='direction'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='avl_insert_here' mangled-name='avl_insert_here' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_insert_here'>
<parameter type-id='a3681dea' name='tree'/>
<parameter type-id='eaa32e2f' name='new_data'/>
<parameter type-id='eaa32e2f' name='here'/>
<parameter type-id='95e97e5e' name='direction'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_update_lt' mangled-name='avl_update_lt' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_update_lt'>
<parameter type-id='a3681dea' name='t'/>
<parameter type-id='eaa32e2f' name='obj'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='avl_update_gt' mangled-name='avl_update_gt' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_update_gt'>
<parameter type-id='a3681dea' name='t'/>
<parameter type-id='eaa32e2f' name='obj'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='avl_update' mangled-name='avl_update' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_update'>
<parameter type-id='a3681dea' name='t'/>
<parameter type-id='eaa32e2f' name='obj'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='avl_swap' mangled-name='avl_swap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_swap'>
<parameter type-id='a3681dea' name='tree1'/>
<parameter type-id='a3681dea' name='tree2'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='avl_is_empty' mangled-name='avl_is_empty' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='avl_is_empty'>
<parameter type-id='a3681dea' name='tree'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/cityhash.c' language='LANG_C99'>
<function-decl name='cityhash4' mangled-name='cityhash4' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='cityhash4'>
<parameter type-id='9c313c2d' name='w1'/>
<parameter type-id='9c313c2d' name='w2'/>
<parameter type-id='9c313c2d' name='w3'/>
<parameter type-id='9c313c2d' name='w4'/>
<return type-id='9c313c2d'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfeature_common.c' language='LANG_C99'>
- <array-type-def dimensions='1' type-id='83f29ca2' size-in-bits='17024' id='9944fffc'>
- <subrange length='38' type-id='7359adad' id='aa4ccdac'/>
+ <array-type-def dimensions='1' type-id='83f29ca2' size-in-bits='17472' id='dd432c71'>
+ <subrange length='39' type-id='7359adad' id='ae4a9561'/>
</array-type-def>
<enum-decl name='zfeature_flags' id='6db816a4'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFEATURE_FLAG_READONLY_COMPAT' value='1'/>
<enumerator name='ZFEATURE_FLAG_MOS' value='2'/>
<enumerator name='ZFEATURE_FLAG_ACTIVATE_ON_ENABLE' value='4'/>
<enumerator name='ZFEATURE_FLAG_PER_DATASET' value='8'/>
</enum-decl>
<typedef-decl name='zfeature_flags_t' type-id='6db816a4' id='fc329033'/>
<enum-decl name='zfeature_type' id='c4fa2355'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFEATURE_TYPE_BOOLEAN' value='0'/>
<enumerator name='ZFEATURE_TYPE_UINT64_ARRAY' value='1'/>
<enumerator name='ZFEATURE_NUM_TYPES' value='2'/>
</enum-decl>
<typedef-decl name='zfeature_type_t' type-id='c4fa2355' id='732d2bb2'/>
<class-decl name='zfeature_info' size-in-bits='448' is-struct='yes' visibility='default' id='1178d146'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='fi_feature' type-id='d6618c78' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='fi_uname' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='fi_guid' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='fi_desc' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='fi_flags' type-id='fc329033' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='288'>
<var-decl name='fi_zfs_mod_supported' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='fi_type' type-id='732d2bb2' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='fi_depends' type-id='1acff326' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfeature_info_t' type-id='1178d146' id='83f29ca2'/>
<typedef-decl name='__free_fn_t' type-id='b7f9d8e6' id='3ff5e51e'/>
<class-decl name='dirent' size-in-bits='2240' is-struct='yes' visibility='default' id='611586a1'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='d_ino' type-id='71288a47' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='d_off' type-id='724e4de6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='d_reclen' type-id='8efea9e5' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='144'>
<var-decl name='d_type' type-id='002ac4a6' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='152'>
<var-decl name='d_name' type-id='d1617432' visibility='default'/>
</data-member>
</class-decl>
<class-decl name='zfs_mod_supported_features' size-in-bits='128' is-struct='yes' visibility='default' id='3eee3342'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='tree' type-id='eaa32e2f' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='all_features' type-id='c19b74c3' visibility='default'/>
</data-member>
</class-decl>
<qualified-type-def type-id='d6618c78' const='yes' id='81a65028'/>
<pointer-type-def type-id='81a65028' size-in-bits='64' id='1acff326'/>
<qualified-type-def type-id='3eee3342' const='yes' id='0c1d5bbb'/>
<pointer-type-def type-id='0c1d5bbb' size-in-bits='64' id='a3372543'/>
<pointer-type-def type-id='611586a1' size-in-bits='64' id='2e243169'/>
<qualified-type-def type-id='eaa32e2f' const='yes' id='83be723c'/>
<pointer-type-def type-id='83be723c' size-in-bits='64' id='7acd98a2'/>
- <var-decl name='spa_feature_table' type-id='9944fffc' mangled-name='spa_feature_table' visibility='default' elf-symbol-id='spa_feature_table'/>
+ <var-decl name='spa_feature_table' type-id='dd432c71' mangled-name='spa_feature_table' visibility='default' elf-symbol-id='spa_feature_table'/>
<var-decl name='zfeature_checks_disable' type-id='c19b74c3' mangled-name='zfeature_checks_disable' visibility='default' elf-symbol-id='zfeature_checks_disable'/>
<function-decl name='opendir' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='f09217ba'/>
</function-decl>
<function-decl name='tsearch' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='63e171df'/>
<parameter type-id='aba7edd8'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='tfind' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='7acd98a2'/>
<parameter type-id='aba7edd8'/>
<return type-id='eaa32e2f'/>
</function-decl>
<function-decl name='tdestroy' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='3ff5e51e'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zfeature_is_valid_guid' mangled-name='zfeature_is_valid_guid' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_is_valid_guid'>
<parameter type-id='80f4b756' name='name'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfeature_depends_on' mangled-name='zfeature_depends_on' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfeature_depends_on'>
<parameter type-id='d6618c78' name='fid'/>
<parameter type-id='d6618c78' name='check'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_mod_supported' mangled-name='zfs_mod_supported' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_mod_supported'>
<parameter type-id='80f4b756' name='scope'/>
<parameter type-id='80f4b756' name='name'/>
<parameter type-id='a3372543' name='sfeatures'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_comutil.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='b99c00c9' size-in-bits='2624' id='5ce15418'>
<subrange length='41' type-id='7359adad' id='cb834f44'/>
</array-type-def>
<qualified-type-def type-id='80f4b756' const='yes' id='b99c00c9'/>
<pointer-type-def type-id='8f92235e' size-in-bits='64' id='90421557'/>
<function-decl name='nvpair_value_uint32' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='dace003f'/>
<parameter type-id='90421557'/>
<return type-id='95e97e5e'/>
</function-decl>
<var-decl name='zfs_history_event_names' type-id='5ce15418' mangled-name='zfs_history_event_names' visibility='default' elf-symbol-id='zfs_history_event_names'/>
<function-decl name='strpbrk' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<return type-id='26a90f95'/>
</function-decl>
<function-decl name='zfs_allocatable_devs' mangled-name='zfs_allocatable_devs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_allocatable_devs'>
<parameter type-id='5ce45b60' name='nv'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_special_devs' mangled-name='zfs_special_devs' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_special_devs'>
<parameter type-id='5ce45b60' name='nv'/>
<parameter type-id='80f4b756' name='type'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_zpl_version_map' mangled-name='zfs_zpl_version_map' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_zpl_version_map'>
<parameter type-id='95e97e5e' name='spa_version'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_spa_version_map' mangled-name='zfs_spa_version_map' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_spa_version_map'>
<parameter type-id='95e97e5e' name='zpl_version'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_dataset_name_hidden' mangled-name='zfs_dataset_name_hidden' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_dataset_name_hidden'>
<parameter type-id='80f4b756' name='name'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_deleg.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='fa1870fd' size-in-bits='4096' id='59e94aca'>
<subrange length='32' type-id='7359adad' id='ae5bde82'/>
</array-type-def>
<array-type-def dimensions='1' type-id='fa1870fd' size-in-bits='infinite' id='7c00e69d'>
<subrange length='infinite' id='031f2035'/>
</array-type-def>
<enum-decl name='zfs_deleg_who_type_t' naming-typedef-id='36d4bd5a' id='b5fa5816'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_DELEG_WHO_UNKNOWN' value='0'/>
<enumerator name='ZFS_DELEG_USER' value='117'/>
<enumerator name='ZFS_DELEG_USER_SETS' value='85'/>
<enumerator name='ZFS_DELEG_GROUP' value='103'/>
<enumerator name='ZFS_DELEG_GROUP_SETS' value='71'/>
<enumerator name='ZFS_DELEG_EVERYONE' value='101'/>
<enumerator name='ZFS_DELEG_EVERYONE_SETS' value='69'/>
<enumerator name='ZFS_DELEG_CREATE' value='99'/>
<enumerator name='ZFS_DELEG_CREATE_SETS' value='67'/>
<enumerator name='ZFS_DELEG_NAMED_SET' value='115'/>
<enumerator name='ZFS_DELEG_NAMED_SET_SETS' value='83'/>
</enum-decl>
<typedef-decl name='zfs_deleg_who_type_t' type-id='b5fa5816' id='36d4bd5a'/>
<enum-decl name='zfs_deleg_note_t' naming-typedef-id='4613c173' id='729d4547'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZFS_DELEG_NOTE_CREATE' value='0'/>
<enumerator name='ZFS_DELEG_NOTE_DESTROY' value='1'/>
<enumerator name='ZFS_DELEG_NOTE_SNAPSHOT' value='2'/>
<enumerator name='ZFS_DELEG_NOTE_ROLLBACK' value='3'/>
<enumerator name='ZFS_DELEG_NOTE_CLONE' value='4'/>
<enumerator name='ZFS_DELEG_NOTE_PROMOTE' value='5'/>
<enumerator name='ZFS_DELEG_NOTE_RENAME' value='6'/>
<enumerator name='ZFS_DELEG_NOTE_SEND' value='7'/>
<enumerator name='ZFS_DELEG_NOTE_RECEIVE' value='8'/>
<enumerator name='ZFS_DELEG_NOTE_ALLOW' value='9'/>
<enumerator name='ZFS_DELEG_NOTE_USERPROP' value='10'/>
<enumerator name='ZFS_DELEG_NOTE_MOUNT' value='11'/>
<enumerator name='ZFS_DELEG_NOTE_SHARE' value='12'/>
<enumerator name='ZFS_DELEG_NOTE_USERQUOTA' value='13'/>
<enumerator name='ZFS_DELEG_NOTE_GROUPQUOTA' value='14'/>
<enumerator name='ZFS_DELEG_NOTE_USERUSED' value='15'/>
<enumerator name='ZFS_DELEG_NOTE_GROUPUSED' value='16'/>
<enumerator name='ZFS_DELEG_NOTE_USEROBJQUOTA' value='17'/>
<enumerator name='ZFS_DELEG_NOTE_GROUPOBJQUOTA' value='18'/>
<enumerator name='ZFS_DELEG_NOTE_USEROBJUSED' value='19'/>
<enumerator name='ZFS_DELEG_NOTE_GROUPOBJUSED' value='20'/>
<enumerator name='ZFS_DELEG_NOTE_HOLD' value='21'/>
<enumerator name='ZFS_DELEG_NOTE_RELEASE' value='22'/>
<enumerator name='ZFS_DELEG_NOTE_DIFF' value='23'/>
<enumerator name='ZFS_DELEG_NOTE_BOOKMARK' value='24'/>
<enumerator name='ZFS_DELEG_NOTE_LOAD_KEY' value='25'/>
<enumerator name='ZFS_DELEG_NOTE_CHANGE_KEY' value='26'/>
<enumerator name='ZFS_DELEG_NOTE_PROJECTUSED' value='27'/>
<enumerator name='ZFS_DELEG_NOTE_PROJECTQUOTA' value='28'/>
<enumerator name='ZFS_DELEG_NOTE_PROJECTOBJUSED' value='29'/>
<enumerator name='ZFS_DELEG_NOTE_PROJECTOBJQUOTA' value='30'/>
<enumerator name='ZFS_DELEG_NOTE_NONE' value='31'/>
</enum-decl>
<typedef-decl name='zfs_deleg_note_t' type-id='729d4547' id='4613c173'/>
<class-decl name='zfs_deleg_perm_tab' size-in-bits='128' is-struct='yes' visibility='default' id='5aa05c1f'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='z_perm' type-id='80f4b756' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='z_note' type-id='4613c173' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_deleg_perm_tab_t' type-id='5aa05c1f' id='f3f851ad'/>
<qualified-type-def type-id='f3f851ad' const='yes' id='fa1870fd'/>
<var-decl name='zfs_deleg_perm_tab' type-id='7c00e69d' mangled-name='zfs_deleg_perm_tab' visibility='default' elf-symbol-id='zfs_deleg_perm_tab'/>
<function-decl name='permset_namecheck' mangled-name='permset_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='permset_namecheck'>
<parameter type-id='80f4b756'/>
<parameter type-id='053457bd'/>
<parameter type-id='26a90f95'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_delegatable' mangled-name='zfs_prop_delegatable' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_delegatable'>
<parameter type-id='58603c44'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_deleg_canonicalize_perm' mangled-name='zfs_deleg_canonicalize_perm' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_deleg_canonicalize_perm'>
<parameter type-id='80f4b756' name='perm'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_deleg_verify_nvlist' mangled-name='zfs_deleg_verify_nvlist' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_deleg_verify_nvlist'>
<parameter type-id='5ce45b60' name='nvp'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_deleg_whokey' mangled-name='zfs_deleg_whokey' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_deleg_whokey'>
<parameter type-id='26a90f95' name='attr'/>
<parameter type-id='36d4bd5a' name='type'/>
<parameter type-id='a84c031d' name='inheritchr'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='48b5725f'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='9c313c2d' size-in-bits='512' id='c5d13f42'>
<subrange length='8' type-id='7359adad' id='56e0c0b1'/>
</array-type-def>
<array-type-def dimensions='1' type-id='90dbb6d6' size-in-bits='2048' id='16582e69'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='8240361c' size-in-bits='1024' id='481f90b1'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='7c1ab40c' size-in-bits='512' id='cbd91ec1'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<array-type-def dimensions='1' type-id='6d059eaa' size-in-bits='1024' id='729b6ebb'>
<subrange length='4' type-id='7359adad' id='16fe7105'/>
</array-type-def>
<enum-decl name='zio_byteorder_t' naming-typedef-id='595a65ec' id='fc861be0'>
<underlying-type type-id='9cac1fee'/>
<enumerator name='ZIO_CHECKSUM_NATIVE' value='0'/>
<enumerator name='ZIO_CHECKSUM_BYTESWAP' value='1'/>
</enum-decl>
<typedef-decl name='zio_byteorder_t' type-id='fc861be0' id='595a65ec'/>
<class-decl name='zio_abd_checksum_data' size-in-bits='256' is-struct='yes' visibility='default' id='4bf4b004'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='acd_byteorder' type-id='595a65ec' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='acd_ctx' type-id='0f7df99e' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='acd_zcp' type-id='c24fc2ee' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='acd_private' type-id='eaa32e2f' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zio_abd_checksum_data_t' type-id='4bf4b004' id='74e39470'/>
<typedef-decl name='zio_abd_checksum_init_t' type-id='a5444274' id='029a8ebe'/>
<typedef-decl name='zio_abd_checksum_fini_t' type-id='a5444274' id='d6fd5c6c'/>
<typedef-decl name='zio_abd_checksum_iter_t' type-id='f4a1892e' id='cefa0f4a'/>
<class-decl name='zio_abd_checksum_func' size-in-bits='192' is-struct='yes' visibility='default' id='aa14691a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='acf_init' type-id='0bcca125' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='acf_fini' type-id='bfe36153' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='acf_iter' type-id='1e276399' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zio_abd_checksum_func_t' type-id='3f8e8d11' id='c2eb138a'/>
<class-decl name='zfs_fletcher_superscalar' size-in-bits='256' is-struct='yes' visibility='default' id='28efb250'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='v' type-id='85c64d26' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_fletcher_superscalar_t' type-id='28efb250' id='6d059eaa'/>
<class-decl name='zfs_fletcher_sse' size-in-bits='128' is-struct='yes' visibility='default' id='acd4019a'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='v' type-id='c1c22e6c' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_fletcher_sse_t' type-id='acd4019a' id='7c1ab40c'/>
<class-decl name='zfs_fletcher_avx' size-in-bits='256' is-struct='yes' visibility='default' id='8c208dfa'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='v' type-id='85c64d26' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_fletcher_avx_t' type-id='8c208dfa' id='8240361c'/>
<class-decl name='zfs_fletcher_avx512' size-in-bits='512' is-struct='yes' visibility='default' id='c6d0c382'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='v' type-id='c5d13f42' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='zfs_fletcher_avx512_t' type-id='c6d0c382' id='90dbb6d6'/>
<union-decl name='fletcher_4_ctx' size-in-bits='2048' visibility='default' id='1f951ade'>
<data-member access='public'>
<var-decl name='scalar' type-id='39730d0b' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='superscalar' type-id='729b6ebb' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='sse' type-id='cbd91ec1' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='avx' type-id='481f90b1' visibility='default'/>
</data-member>
<data-member access='public'>
<var-decl name='avx512' type-id='16582e69' visibility='default'/>
</data-member>
</union-decl>
<typedef-decl name='fletcher_4_ctx_t' type-id='1f951ade' id='4b675395'/>
<qualified-type-def type-id='aa14691a' const='yes' id='3f8e8d11'/>
<pointer-type-def type-id='4b675395' size-in-bits='64' id='0f7df99e'/>
<qualified-type-def type-id='8f92235e' volatile='yes' id='430e0681'/>
<pointer-type-def type-id='430e0681' size-in-bits='64' id='3a147f31'/>
<pointer-type-def type-id='74e39470' size-in-bits='64' id='eefe7427'/>
<pointer-type-def type-id='d6fd5c6c' size-in-bits='64' id='bfe36153'/>
<pointer-type-def type-id='029a8ebe' size-in-bits='64' id='0bcca125'/>
<pointer-type-def type-id='cefa0f4a' size-in-bits='64' id='1e276399'/>
<var-decl name='fletcher_4_abd_ops' type-id='c2eb138a' mangled-name='fletcher_4_abd_ops' visibility='default' elf-symbol-id='fletcher_4_abd_ops'/>
<function-decl name='atomic_swap_32' mangled-name='atomic_swap_32' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='atomic_swap_32'>
<parameter type-id='3a147f31'/>
<parameter type-id='8f92235e'/>
<return type-id='8f92235e'/>
</function-decl>
<function-decl name='membar_producer' mangled-name='membar_producer' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='membar_producer'>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_init' mangled-name='fletcher_init' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_init'>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_2_incremental_native' mangled-name='fletcher_2_incremental_native' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_2_incremental_native'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='b59d7dce' name='size'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fletcher_2_native' mangled-name='fletcher_2_native' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_2_native'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='9c313c2d' name='size'/>
<parameter type-id='eaa32e2f' name='ctx_template'/>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_2_incremental_byteswap' mangled-name='fletcher_2_incremental_byteswap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_2_incremental_byteswap'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='b59d7dce' name='size'/>
<parameter type-id='eaa32e2f' name='data'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fletcher_2_byteswap' mangled-name='fletcher_2_byteswap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_2_byteswap'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='9c313c2d' name='size'/>
<parameter type-id='eaa32e2f' name='ctx_template'/>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_impl_set' mangled-name='fletcher_4_impl_set' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_impl_set'>
<parameter type-id='80f4b756' name='val'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='fletcher_4_native' mangled-name='fletcher_4_native' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_native'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='9c313c2d' name='size'/>
<parameter type-id='eaa32e2f' name='ctx_template'/>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='fletcher_4_byteswap' mangled-name='fletcher_4_byteswap' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='fletcher_4_byteswap'>
<parameter type-id='eaa32e2f' name='buf'/>
<parameter type-id='9c313c2d' name='size'/>
<parameter type-id='eaa32e2f' name='ctx_template'/>
<parameter type-id='c24fc2ee' name='zcp'/>
<return type-id='48b5725f'/>
</function-decl>
<function-type size-in-bits='64' id='f4a1892e'>
<parameter type-id='eaa32e2f'/>
<parameter type-id='b59d7dce'/>
<parameter type-id='eaa32e2f'/>
<return type-id='95e97e5e'/>
</function-type>
<function-type size-in-bits='64' id='a5444274'>
<parameter type-id='eefe7427'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_avx512.c' language='LANG_C99'>
<typedef-decl name='fletcher_4_init_f' type-id='173aa527' id='b9ae1656'/>
<typedef-decl name='fletcher_4_fini_f' type-id='0ad5b8a8' id='c4c1f4fc'/>
<typedef-decl name='fletcher_4_compute_f' type-id='38147eff' id='ad1dc4cb'/>
<class-decl name='fletcher_4_func' size-in-bits='1024' is-struct='yes' visibility='default' id='57f479a0'>
<data-member access='public' layout-offset-in-bits='0'>
<var-decl name='init_native' type-id='b9ae1656' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='64'>
<var-decl name='fini_native' type-id='c4c1f4fc' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='128'>
<var-decl name='compute_native' type-id='ad1dc4cb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='192'>
<var-decl name='init_byteswap' type-id='b9ae1656' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='256'>
<var-decl name='fini_byteswap' type-id='c4c1f4fc' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='320'>
<var-decl name='compute_byteswap' type-id='ad1dc4cb' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='384'>
<var-decl name='valid' type-id='297d38bc' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='448'>
<var-decl name='uses_fpu' type-id='c19b74c3' visibility='default'/>
</data-member>
<data-member access='public' layout-offset-in-bits='512'>
<var-decl name='name' type-id='80f4b756' visibility='default'/>
</data-member>
</class-decl>
<typedef-decl name='fletcher_4_ops_t' type-id='57f479a0' id='eba91718'/>
<qualified-type-def type-id='eba91718' const='yes' id='9eeabdc8'/>
<pointer-type-def type-id='e9e61702' size-in-bits='64' id='297d38bc'/>
<pointer-type-def type-id='fe40251b' size-in-bits='64' id='173aa527'/>
<pointer-type-def type-id='17fb1f83' size-in-bits='64' id='38147eff'/>
<pointer-type-def type-id='fb39e25e' size-in-bits='64' id='0ad5b8a8'/>
<var-decl name='fletcher_4_avx512f_ops' type-id='9eeabdc8' mangled-name='fletcher_4_avx512f_ops' visibility='default' elf-symbol-id='fletcher_4_avx512f_ops'/>
<var-decl name='fletcher_4_avx512bw_ops' type-id='9eeabdc8' mangled-name='fletcher_4_avx512bw_ops' visibility='default' elf-symbol-id='fletcher_4_avx512bw_ops'/>
<function-type size-in-bits='64' id='e9e61702'>
<return type-id='c19b74c3'/>
</function-type>
<function-type size-in-bits='64' id='fe40251b'>
<parameter type-id='0f7df99e'/>
<return type-id='48b5725f'/>
</function-type>
<function-type size-in-bits='64' id='17fb1f83'>
<parameter type-id='0f7df99e'/>
<parameter type-id='eaa32e2f'/>
<parameter type-id='9c313c2d'/>
<return type-id='48b5725f'/>
</function-type>
<function-type size-in-bits='64' id='fb39e25e'>
<parameter type-id='0f7df99e'/>
<parameter type-id='c24fc2ee'/>
<return type-id='48b5725f'/>
</function-type>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_intel.c' language='LANG_C99'>
<var-decl name='fletcher_4_avx2_ops' type-id='9eeabdc8' mangled-name='fletcher_4_avx2_ops' visibility='default' elf-symbol-id='fletcher_4_avx2_ops'/>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_sse.c' language='LANG_C99'>
<var-decl name='fletcher_4_sse2_ops' type-id='9eeabdc8' mangled-name='fletcher_4_sse2_ops' visibility='default' elf-symbol-id='fletcher_4_sse2_ops'/>
<var-decl name='fletcher_4_ssse3_ops' type-id='9eeabdc8' mangled-name='fletcher_4_ssse3_ops' visibility='default' elf-symbol-id='fletcher_4_ssse3_ops'/>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_superscalar.c' language='LANG_C99'>
<var-decl name='fletcher_4_superscalar_ops' type-id='9eeabdc8' mangled-name='fletcher_4_superscalar_ops' visibility='default' elf-symbol-id='fletcher_4_superscalar_ops'/>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_fletcher_superscalar4.c' language='LANG_C99'>
<var-decl name='fletcher_4_superscalar4_ops' type-id='9eeabdc8' mangled-name='fletcher_4_superscalar4_ops' visibility='default' elf-symbol-id='fletcher_4_superscalar4_ops'/>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_namecheck.c' language='LANG_C99'>
<var-decl name='zfs_max_dataset_nesting' type-id='95e97e5e' mangled-name='zfs_max_dataset_nesting' visibility='default' elf-symbol-id='zfs_max_dataset_nesting'/>
<function-decl name='get_dataset_depth' mangled-name='get_dataset_depth' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='get_dataset_depth'>
<parameter type-id='80f4b756' name='path'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_component_namecheck' mangled-name='zfs_component_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_component_namecheck'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='053457bd' name='why'/>
<parameter type-id='26a90f95' name='what'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='dataset_namecheck' mangled-name='dataset_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='dataset_namecheck'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='053457bd' name='why'/>
<parameter type-id='26a90f95' name='what'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='bookmark_namecheck' mangled-name='bookmark_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='bookmark_namecheck'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='053457bd' name='why'/>
<parameter type-id='26a90f95' name='what'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='snapshot_namecheck' mangled-name='snapshot_namecheck' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='snapshot_namecheck'>
<parameter type-id='80f4b756' name='path'/>
<parameter type-id='053457bd' name='why'/>
<parameter type-id='26a90f95' name='what'/>
<return type-id='95e97e5e'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zfs_prop.c' language='LANG_C99'>
<array-type-def dimensions='1' type-id='b99c00c9' size-in-bits='768' id='bcc77e38'>
<subrange length='12' type-id='7359adad' id='84827bdc'/>
</array-type-def>
<pointer-type-def type-id='3eee3342' size-in-bits='64' id='73f8e240'/>
<var-decl name='zfs_userquota_prop_prefixes' type-id='bcc77e38' mangled-name='zfs_userquota_prop_prefixes' visibility='default' elf-symbol-id='zfs_userquota_prop_prefixes'/>
<function-decl name='zfs_mod_list_supported' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='80f4b756'/>
<return type-id='73f8e240'/>
</function-decl>
<function-decl name='zfs_mod_list_supported_free' visibility='default' binding='global' size-in-bits='64'>
<parameter type-id='73f8e240'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_impl' mangled-name='zprop_register_impl' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_impl'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='31429eff'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='80f4b756'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='c8bc397b'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_string' mangled-name='zprop_register_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_string'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_number' mangled-name='zprop_register_number' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_number'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_index' mangled-name='zprop_register_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_index'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c8bc397b'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_register_hidden' mangled-name='zprop_register_hidden' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_register_hidden'>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='31429eff'/>
<parameter type-id='999701cc'/>
<parameter type-id='95e97e5e'/>
<parameter type-id='80f4b756'/>
<parameter type-id='c19b74c3'/>
<parameter type-id='a3372543'/>
<return type-id='48b5725f'/>
</function-decl>
<function-decl name='zprop_index_to_string' mangled-name='zprop_index_to_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_index_to_string'>
<parameter type-id='95e97e5e'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='7d3cd834'/>
<parameter type-id='2e45de5d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zprop_random_value' mangled-name='zprop_random_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_random_value'>
<parameter type-id='95e97e5e'/>
<parameter type-id='9c313c2d'/>
<parameter type-id='2e45de5d'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zprop_valid_char' mangled-name='zprop_valid_char' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zprop_valid_char'>
<parameter type-id='a84c031d'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_string_to_index' mangled-name='zfs_prop_string_to_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_string_to_index'>
<parameter type-id='58603c44' name='prop'/>
<parameter type-id='80f4b756' name='string'/>
<parameter type-id='5d6479ae' name='index'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_random_value' mangled-name='zfs_prop_random_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_random_value'>
<parameter type-id='58603c44' name='prop'/>
<parameter type-id='9c313c2d' name='seed'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zfs_prop_visible' mangled-name='zfs_prop_visible' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_visible'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='zfs_prop_values' mangled-name='zfs_prop_values' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_values'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_prop_is_string' mangled-name='zfs_prop_is_string' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_is_string'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zfs_prop_column_name' mangled-name='zfs_prop_column_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_column_name'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zfs_prop_align_right' mangled-name='zfs_prop_align_right' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zfs_prop_align_right'>
<parameter type-id='58603c44' name='prop'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zpool_prop.c' language='LANG_C99'>
<function-decl name='zpool_prop_string_to_index' mangled-name='zpool_prop_string_to_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_string_to_index'>
<parameter type-id='5d0c23fb' name='prop'/>
<parameter type-id='80f4b756' name='string'/>
<parameter type-id='5d6479ae' name='index'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='zpool_prop_random_value' mangled-name='zpool_prop_random_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_random_value'>
<parameter type-id='5d0c23fb' name='prop'/>
<parameter type-id='9c313c2d' name='seed'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='zpool_prop_values' mangled-name='zpool_prop_values' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_values'>
<parameter type-id='5d0c23fb' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_prop_column_name' mangled-name='zpool_prop_column_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_column_name'>
<parameter type-id='5d0c23fb' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='zpool_prop_align_right' mangled-name='zpool_prop_align_right' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='zpool_prop_align_right'>
<parameter type-id='5d0c23fb' name='prop'/>
<return type-id='c19b74c3'/>
</function-decl>
<function-decl name='vdev_prop_get_table' mangled-name='vdev_prop_get_table' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_get_table'>
<return type-id='76c8174b'/>
</function-decl>
<function-decl name='vdev_prop_string_to_index' mangled-name='vdev_prop_string_to_index' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_string_to_index'>
<parameter type-id='5aa5c90c' name='prop'/>
<parameter type-id='80f4b756' name='string'/>
<parameter type-id='5d6479ae' name='index'/>
<return type-id='95e97e5e'/>
</function-decl>
<function-decl name='vdev_prop_random_value' mangled-name='vdev_prop_random_value' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_random_value'>
<parameter type-id='5aa5c90c' name='prop'/>
<parameter type-id='9c313c2d' name='seed'/>
<return type-id='9c313c2d'/>
</function-decl>
<function-decl name='vdev_prop_values' mangled-name='vdev_prop_values' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_values'>
<parameter type-id='5aa5c90c' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_column_name' mangled-name='vdev_prop_column_name' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_column_name'>
<parameter type-id='5aa5c90c' name='prop'/>
<return type-id='80f4b756'/>
</function-decl>
<function-decl name='vdev_prop_align_right' mangled-name='vdev_prop_align_right' visibility='default' binding='global' size-in-bits='64' elf-symbol-id='vdev_prop_align_right'>
<parameter type-id='5aa5c90c' name='prop'/>
<return type-id='c19b74c3'/>
</function-decl>
</abi-instr>
<abi-instr address-size='64' path='module/zcommon/zprop_common.c' language='LANG_C99'>
<function-decl name='__ctype_tolower_loc' visibility='default' binding='global' size-in-bits='64'>
<return type-id='24f95ba5'/>
</function-decl>
</abi-instr>
</abi-corpus>
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_changelist.c b/sys/contrib/openzfs/lib/libzfs/libzfs_changelist.c
index d7ea60822419..dd14c570ec03 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_changelist.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_changelist.c
@@ -1,779 +1,780 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2010 Sun Microsystems, Inc. All rights reserved.
* Use is subject to license terms.
*
* Portions Copyright 2007 Ramprakash Jelari
* Copyright (c) 2014, 2020 by Delphix. All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>
* Copyright (c) 2018 Datto Inc.
*/
#include <libintl.h>
#include <libuutil.h>
#include <stddef.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <zone.h>
#include <libzfs.h>
#include "libzfs_impl.h"
/*
* Structure to keep track of dataset state. Before changing the 'sharenfs' or
* 'mountpoint' property, we record whether the filesystem was previously
* mounted/shared. This prior state dictates whether we remount/reshare the
* dataset after the property has been changed.
*
* The interface consists of the following sequence of functions:
*
* changelist_gather()
* changelist_prefix()
* < change property >
* changelist_postfix()
* changelist_free()
*
* Other interfaces:
*
* changelist_remove() - remove a node from a gathered list
* changelist_rename() - renames all datasets appropriately when doing a rename
* changelist_unshare() - unshares all the nodes in a given changelist
* changelist_haszonedchild() - check if there is any child exported to
* a local zone
*/
typedef struct prop_changenode {
zfs_handle_t *cn_handle;
int cn_shared;
int cn_mounted;
int cn_zoned;
boolean_t cn_needpost; /* is postfix() needed? */
uu_avl_node_t cn_treenode;
} prop_changenode_t;
struct prop_changelist {
zfs_prop_t cl_prop;
zfs_prop_t cl_realprop;
zfs_prop_t cl_shareprop; /* used with sharenfs/sharesmb */
uu_avl_pool_t *cl_pool;
uu_avl_t *cl_tree;
boolean_t cl_waslegacy;
boolean_t cl_allchildren;
boolean_t cl_alldependents;
int cl_mflags; /* Mount flags */
int cl_gflags; /* Gather request flags */
boolean_t cl_haszonedchild;
};
/*
* If the property is 'mountpoint', go through and unmount filesystems as
* necessary. We don't do the same for 'sharenfs', because we can just re-share
* with different options without interrupting service. We do handle 'sharesmb'
* since there may be old resource names that need to be removed.
*/
int
changelist_prefix(prop_changelist_t *clp)
{
prop_changenode_t *cn;
uu_avl_walk_t *walk;
int ret = 0;
const enum sa_protocol smb[] = {SA_PROTOCOL_SMB, SA_NO_PROTOCOL};
boolean_t commit_smb_shares = B_FALSE;
if (clp->cl_prop != ZFS_PROP_MOUNTPOINT &&
clp->cl_prop != ZFS_PROP_SHARESMB)
return (0);
if ((walk = uu_avl_walk_start(clp->cl_tree, UU_WALK_ROBUST)) == NULL)
return (-1);
while ((cn = uu_avl_walk_next(walk)) != NULL) {
/* if a previous loop failed, set the remaining to false */
if (ret == -1) {
cn->cn_needpost = B_FALSE;
continue;
}
/*
* If we are in the global zone, but this dataset is exported
* to a local zone, do nothing.
*/
if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned)
continue;
if (!ZFS_IS_VOLUME(cn->cn_handle)) {
/*
* Do the property specific processing.
*/
switch (clp->cl_prop) {
case ZFS_PROP_MOUNTPOINT:
if (clp->cl_gflags & CL_GATHER_DONT_UNMOUNT)
break;
if (zfs_unmount(cn->cn_handle, NULL,
clp->cl_mflags) != 0) {
ret = -1;
cn->cn_needpost = B_FALSE;
}
break;
case ZFS_PROP_SHARESMB:
(void) zfs_unshare(cn->cn_handle, NULL,
smb);
commit_smb_shares = B_TRUE;
break;
default:
break;
}
}
}
if (commit_smb_shares)
zfs_commit_shares(smb);
uu_avl_walk_end(walk);
if (ret == -1)
(void) changelist_postfix(clp);
return (ret);
}
/*
* If the property is 'mountpoint' or 'sharenfs', go through and remount and/or
* reshare the filesystems as necessary. In changelist_gather() we recorded
* whether the filesystem was previously shared or mounted. The action we take
* depends on the previous state, and whether the value was previously 'legacy'.
* For non-legacy properties, we only remount/reshare the filesystem if it was
* previously mounted/shared. Otherwise, we always remount/reshare the
* filesystem.
*/
int
changelist_postfix(prop_changelist_t *clp)
{
prop_changenode_t *cn;
uu_avl_walk_t *walk;
char shareopts[ZFS_MAXPROPLEN];
int errors = 0;
boolean_t commit_smb_shares = B_FALSE;
boolean_t commit_nfs_shares = B_FALSE;
/*
* If we're changing the mountpoint, attempt to destroy the underlying
* mountpoint. All other datasets will have inherited from this dataset
* (in which case their mountpoints exist in the filesystem in the new
* location), or have explicit mountpoints set (in which case they won't
* be in the changelist).
*/
if ((cn = uu_avl_last(clp->cl_tree)) == NULL)
return (0);
if (clp->cl_prop == ZFS_PROP_MOUNTPOINT &&
!(clp->cl_gflags & CL_GATHER_DONT_UNMOUNT))
remove_mountpoint(cn->cn_handle);
/*
* We walk the datasets in reverse, because we want to mount any parent
* datasets before mounting the children. We walk all datasets even if
* there are errors.
*/
if ((walk = uu_avl_walk_start(clp->cl_tree,
UU_WALK_REVERSE | UU_WALK_ROBUST)) == NULL)
return (-1);
while ((cn = uu_avl_walk_next(walk)) != NULL) {
boolean_t sharenfs;
boolean_t sharesmb;
boolean_t mounted;
boolean_t needs_key;
/*
* If we are in the global zone, but this dataset is exported
* to a local zone, do nothing.
*/
if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned)
continue;
/* Only do post-processing if it's required */
if (!cn->cn_needpost)
continue;
cn->cn_needpost = B_FALSE;
zfs_refresh_properties(cn->cn_handle);
if (ZFS_IS_VOLUME(cn->cn_handle))
continue;
/*
* Remount if previously mounted or mountpoint was legacy,
* or sharenfs or sharesmb property is set.
*/
sharenfs = ((zfs_prop_get(cn->cn_handle, ZFS_PROP_SHARENFS,
shareopts, sizeof (shareopts), NULL, NULL, 0,
B_FALSE) == 0) && (strcmp(shareopts, "off") != 0));
sharesmb = ((zfs_prop_get(cn->cn_handle, ZFS_PROP_SHARESMB,
shareopts, sizeof (shareopts), NULL, NULL, 0,
B_FALSE) == 0) && (strcmp(shareopts, "off") != 0));
needs_key = (zfs_prop_get_int(cn->cn_handle,
ZFS_PROP_KEYSTATUS) == ZFS_KEYSTATUS_UNAVAILABLE);
mounted = (clp->cl_gflags & CL_GATHER_DONT_UNMOUNT) ||
zfs_is_mounted(cn->cn_handle, NULL);
if (!mounted && !needs_key && (cn->cn_mounted ||
((sharenfs || sharesmb || clp->cl_waslegacy) &&
(zfs_prop_get_int(cn->cn_handle,
ZFS_PROP_CANMOUNT) == ZFS_CANMOUNT_ON)))) {
if (zfs_mount(cn->cn_handle, NULL, 0) != 0)
errors++;
else
mounted = TRUE;
}
/*
* If the file system is mounted we always re-share even
* if the filesystem is currently shared, so that we can
* adopt any new options.
*/
const enum sa_protocol nfs[] =
{SA_PROTOCOL_NFS, SA_NO_PROTOCOL};
if (sharenfs && mounted) {
errors += zfs_share(cn->cn_handle, nfs);
commit_nfs_shares = B_TRUE;
} else if (cn->cn_shared || clp->cl_waslegacy) {
errors += zfs_unshare(cn->cn_handle, NULL, nfs);
commit_nfs_shares = B_TRUE;
}
const enum sa_protocol smb[] =
{SA_PROTOCOL_SMB, SA_NO_PROTOCOL};
if (sharesmb && mounted) {
errors += zfs_share(cn->cn_handle, smb);
commit_smb_shares = B_TRUE;
} else if (cn->cn_shared || clp->cl_waslegacy) {
errors += zfs_unshare(cn->cn_handle, NULL, smb);
commit_smb_shares = B_TRUE;
}
}
enum sa_protocol proto[SA_PROTOCOL_COUNT + 1], *p = proto;
if (commit_nfs_shares)
*p++ = SA_PROTOCOL_NFS;
if (commit_smb_shares)
*p++ = SA_PROTOCOL_SMB;
*p++ = SA_NO_PROTOCOL;
zfs_commit_shares(proto);
uu_avl_walk_end(walk);
return (errors ? -1 : 0);
}
/*
* Is this "dataset" a child of "parent"?
*/
static boolean_t
isa_child_of(const char *dataset, const char *parent)
{
int len;
len = strlen(parent);
if (strncmp(dataset, parent, len) == 0 &&
(dataset[len] == '@' || dataset[len] == '/' ||
dataset[len] == '\0'))
return (B_TRUE);
else
return (B_FALSE);
}
/*
* If we rename a filesystem, child filesystem handles are no longer valid
* since we identify each dataset by its name in the ZFS namespace. As a
* result, we have to go through and fix up all the names appropriately. We
* could do this automatically if libzfs kept track of all open handles, but
* this is a lot less work.
*/
void
changelist_rename(prop_changelist_t *clp, const char *src, const char *dst)
{
prop_changenode_t *cn;
uu_avl_walk_t *walk;
char newname[ZFS_MAX_DATASET_NAME_LEN];
if ((walk = uu_avl_walk_start(clp->cl_tree, UU_WALK_ROBUST)) == NULL)
return;
while ((cn = uu_avl_walk_next(walk)) != NULL) {
/*
* Do not rename a clone that's not in the source hierarchy.
*/
if (!isa_child_of(cn->cn_handle->zfs_name, src))
continue;
/*
* Destroy the previous mountpoint if needed.
*/
remove_mountpoint(cn->cn_handle);
(void) strlcpy(newname, dst, sizeof (newname));
(void) strlcat(newname, cn->cn_handle->zfs_name + strlen(src),
sizeof (newname));
(void) strlcpy(cn->cn_handle->zfs_name, newname,
sizeof (cn->cn_handle->zfs_name));
}
uu_avl_walk_end(walk);
}
/*
* Given a gathered changelist for the 'sharenfs' or 'sharesmb' property,
* unshare all the datasets in the list.
*/
int
changelist_unshare(prop_changelist_t *clp, const enum sa_protocol *proto)
{
prop_changenode_t *cn;
uu_avl_walk_t *walk;
int ret = 0;
if (clp->cl_prop != ZFS_PROP_SHARENFS &&
clp->cl_prop != ZFS_PROP_SHARESMB)
return (0);
if ((walk = uu_avl_walk_start(clp->cl_tree, UU_WALK_ROBUST)) == NULL)
return (-1);
while ((cn = uu_avl_walk_next(walk)) != NULL) {
if (zfs_unshare(cn->cn_handle, NULL, proto) != 0)
ret = -1;
}
for (const enum sa_protocol *p = proto; *p != SA_NO_PROTOCOL; ++p)
sa_commit_shares(*p);
uu_avl_walk_end(walk);
return (ret);
}
/*
* Check if there is any child exported to a local zone in a given changelist.
* This information has already been recorded while gathering the changelist
* via changelist_gather().
*/
int
changelist_haszonedchild(prop_changelist_t *clp)
{
return (clp->cl_haszonedchild);
}
/*
* Remove a node from a gathered list.
*/
void
changelist_remove(prop_changelist_t *clp, const char *name)
{
prop_changenode_t *cn;
uu_avl_walk_t *walk;
if ((walk = uu_avl_walk_start(clp->cl_tree, UU_WALK_ROBUST)) == NULL)
return;
while ((cn = uu_avl_walk_next(walk)) != NULL) {
if (strcmp(cn->cn_handle->zfs_name, name) == 0) {
uu_avl_remove(clp->cl_tree, cn);
zfs_close(cn->cn_handle);
free(cn);
uu_avl_walk_end(walk);
return;
}
}
uu_avl_walk_end(walk);
}
/*
* Release any memory associated with a changelist.
*/
void
changelist_free(prop_changelist_t *clp)
{
prop_changenode_t *cn;
if (clp->cl_tree) {
uu_avl_walk_t *walk;
if ((walk = uu_avl_walk_start(clp->cl_tree,
UU_WALK_ROBUST)) == NULL)
return;
while ((cn = uu_avl_walk_next(walk)) != NULL) {
uu_avl_remove(clp->cl_tree, cn);
zfs_close(cn->cn_handle);
free(cn);
}
uu_avl_walk_end(walk);
uu_avl_destroy(clp->cl_tree);
}
if (clp->cl_pool)
uu_avl_pool_destroy(clp->cl_pool);
free(clp);
}
/*
* Add one dataset to changelist
*/
static int
changelist_add_mounted(zfs_handle_t *zhp, void *data)
{
prop_changelist_t *clp = data;
prop_changenode_t *cn;
uu_avl_index_t idx;
ASSERT3U(clp->cl_prop, ==, ZFS_PROP_MOUNTPOINT);
cn = zfs_alloc(zfs_get_handle(zhp), sizeof (prop_changenode_t));
cn->cn_handle = zhp;
cn->cn_mounted = zfs_is_mounted(zhp, NULL);
ASSERT3U(cn->cn_mounted, ==, B_TRUE);
cn->cn_shared = zfs_is_shared(zhp, NULL, NULL);
cn->cn_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
cn->cn_needpost = B_TRUE;
/* Indicate if any child is exported to a local zone. */
if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned)
clp->cl_haszonedchild = B_TRUE;
uu_avl_node_init(cn, &cn->cn_treenode, clp->cl_pool);
if (uu_avl_find(clp->cl_tree, cn, NULL, &idx) == NULL) {
uu_avl_insert(clp->cl_tree, cn, idx);
} else {
free(cn);
zfs_close(zhp);
}
return (0);
}
static int
change_one(zfs_handle_t *zhp, void *data)
{
prop_changelist_t *clp = data;
char property[ZFS_MAXPROPLEN];
char where[64];
prop_changenode_t *cn = NULL;
zprop_source_t sourcetype = ZPROP_SRC_NONE;
zprop_source_t share_sourcetype = ZPROP_SRC_NONE;
int ret = 0;
/*
* We only want to unmount/unshare those filesystems that may inherit
* from the target filesystem. If we find any filesystem with a
* locally set mountpoint, we ignore any children since changing the
* property will not affect them. If this is a rename, we iterate
* over all children regardless, since we need them unmounted in
* order to do the rename. Also, if this is a volume and we're doing
* a rename, then always add it to the changelist.
*/
if (!(ZFS_IS_VOLUME(zhp) && clp->cl_realprop == ZFS_PROP_NAME) &&
zfs_prop_get(zhp, clp->cl_prop, property,
sizeof (property), &sourcetype, where, sizeof (where),
B_FALSE) != 0) {
goto out;
}
/*
* If we are "watching" sharenfs or sharesmb
* then check out the companion property which is tracked
* in cl_shareprop
*/
if (clp->cl_shareprop != ZPROP_INVAL &&
zfs_prop_get(zhp, clp->cl_shareprop, property,
sizeof (property), &share_sourcetype, where, sizeof (where),
B_FALSE) != 0) {
goto out;
}
if (clp->cl_alldependents || clp->cl_allchildren ||
sourcetype == ZPROP_SRC_DEFAULT ||
sourcetype == ZPROP_SRC_INHERITED ||
(clp->cl_shareprop != ZPROP_INVAL &&
(share_sourcetype == ZPROP_SRC_DEFAULT ||
share_sourcetype == ZPROP_SRC_INHERITED))) {
cn = zfs_alloc(zfs_get_handle(zhp), sizeof (prop_changenode_t));
cn->cn_handle = zhp;
cn->cn_mounted = (clp->cl_gflags & CL_GATHER_MOUNT_ALWAYS) ||
zfs_is_mounted(zhp, NULL);
cn->cn_shared = zfs_is_shared(zhp, NULL, NULL);
cn->cn_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
cn->cn_needpost = B_TRUE;
/* Indicate if any child is exported to a local zone. */
if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned)
clp->cl_haszonedchild = B_TRUE;
uu_avl_node_init(cn, &cn->cn_treenode, clp->cl_pool);
uu_avl_index_t idx;
if (uu_avl_find(clp->cl_tree, cn, NULL, &idx) == NULL) {
uu_avl_insert(clp->cl_tree, cn, idx);
} else {
free(cn);
cn = NULL;
}
if (!clp->cl_alldependents)
- ret = zfs_iter_children(zhp, 0, change_one, data);
+ ret = zfs_iter_children_v2(zhp, 0, change_one, data);
/*
* If we added the handle to the changelist, we will re-use it
* later so return without closing it.
*/
if (cn != NULL)
return (ret);
}
out:
zfs_close(zhp);
return (ret);
}
static int
compare_props(const void *a, const void *b, zfs_prop_t prop)
{
const prop_changenode_t *ca = a;
const prop_changenode_t *cb = b;
char propa[MAXPATHLEN];
char propb[MAXPATHLEN];
boolean_t haspropa, haspropb;
haspropa = (zfs_prop_get(ca->cn_handle, prop, propa, sizeof (propa),
NULL, NULL, 0, B_FALSE) == 0);
haspropb = (zfs_prop_get(cb->cn_handle, prop, propb, sizeof (propb),
NULL, NULL, 0, B_FALSE) == 0);
if (!haspropa && haspropb)
return (-1);
else if (haspropa && !haspropb)
return (1);
else if (!haspropa && !haspropb)
return (0);
else
return (strcmp(propb, propa));
}
static int
compare_mountpoints(const void *a, const void *b, void *unused)
{
/*
* When unsharing or unmounting filesystems, we need to do it in
* mountpoint order. This allows the user to have a mountpoint
* hierarchy that is different from the dataset hierarchy, and still
* allow it to be changed.
*/
(void) unused;
return (compare_props(a, b, ZFS_PROP_MOUNTPOINT));
}
static int
compare_dataset_names(const void *a, const void *b, void *unused)
{
(void) unused;
return (compare_props(a, b, ZFS_PROP_NAME));
}
/*
* Given a ZFS handle and a property, construct a complete list of datasets
* that need to be modified as part of this process. For anything but the
* 'mountpoint' and 'sharenfs' properties, this just returns an empty list.
* Otherwise, we iterate over all children and look for any datasets that
* inherit the property. For each such dataset, we add it to the list and
* mark whether it was shared beforehand.
*/
prop_changelist_t *
changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int gather_flags,
int mnt_flags)
{
prop_changelist_t *clp;
prop_changenode_t *cn;
zfs_handle_t *temp;
char property[ZFS_MAXPROPLEN];
boolean_t legacy = B_FALSE;
clp = zfs_alloc(zhp->zfs_hdl, sizeof (prop_changelist_t));
/*
* For mountpoint-related tasks, we want to sort everything by
* mountpoint, so that we mount and unmount them in the appropriate
* order, regardless of their position in the hierarchy.
*/
if (prop == ZFS_PROP_NAME || prop == ZFS_PROP_ZONED ||
prop == ZFS_PROP_MOUNTPOINT || prop == ZFS_PROP_SHARENFS ||
prop == ZFS_PROP_SHARESMB) {
if (zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT,
property, sizeof (property),
NULL, NULL, 0, B_FALSE) == 0 &&
(strcmp(property, "legacy") == 0 ||
strcmp(property, "none") == 0)) {
legacy = B_TRUE;
}
}
clp->cl_pool = uu_avl_pool_create("changelist_pool",
sizeof (prop_changenode_t),
offsetof(prop_changenode_t, cn_treenode),
legacy ? compare_dataset_names : compare_mountpoints, 0);
if (clp->cl_pool == NULL) {
assert(uu_error() == UU_ERROR_NO_MEMORY);
(void) zfs_error(zhp->zfs_hdl, EZFS_NOMEM, "internal error");
changelist_free(clp);
return (NULL);
}
clp->cl_tree = uu_avl_create(clp->cl_pool, NULL, UU_DEFAULT);
clp->cl_gflags = gather_flags;
clp->cl_mflags = mnt_flags;
if (clp->cl_tree == NULL) {
assert(uu_error() == UU_ERROR_NO_MEMORY);
(void) zfs_error(zhp->zfs_hdl, EZFS_NOMEM, "internal error");
changelist_free(clp);
return (NULL);
}
/*
* If this is a rename or the 'zoned' property, we pretend we're
* changing the mountpoint and flag it so we can catch all children in
* change_one().
*
* Flag cl_alldependents to catch all children plus the dependents
* (clones) that are not in the hierarchy.
*/
if (prop == ZFS_PROP_NAME) {
clp->cl_prop = ZFS_PROP_MOUNTPOINT;
clp->cl_alldependents = B_TRUE;
} else if (prop == ZFS_PROP_ZONED) {
clp->cl_prop = ZFS_PROP_MOUNTPOINT;
clp->cl_allchildren = B_TRUE;
} else if (prop == ZFS_PROP_CANMOUNT) {
clp->cl_prop = ZFS_PROP_MOUNTPOINT;
} else if (prop == ZFS_PROP_VOLSIZE) {
clp->cl_prop = ZFS_PROP_MOUNTPOINT;
} else {
clp->cl_prop = prop;
}
clp->cl_realprop = prop;
if (clp->cl_prop != ZFS_PROP_MOUNTPOINT &&
clp->cl_prop != ZFS_PROP_SHARENFS &&
clp->cl_prop != ZFS_PROP_SHARESMB)
return (clp);
/*
* If watching SHARENFS or SHARESMB then
* also watch its companion property.
*/
if (clp->cl_prop == ZFS_PROP_SHARENFS)
clp->cl_shareprop = ZFS_PROP_SHARESMB;
else if (clp->cl_prop == ZFS_PROP_SHARESMB)
clp->cl_shareprop = ZFS_PROP_SHARENFS;
if (clp->cl_prop == ZFS_PROP_MOUNTPOINT &&
(clp->cl_gflags & CL_GATHER_ITER_MOUNTED)) {
/*
* Instead of iterating through all of the dataset children we
* gather mounted dataset children from MNTTAB
*/
if (zfs_iter_mounted(zhp, changelist_add_mounted, clp) != 0) {
changelist_free(clp);
return (NULL);
}
} else if (clp->cl_alldependents) {
- if (zfs_iter_dependents(zhp, 0, B_TRUE, change_one, clp) != 0) {
+ if (zfs_iter_dependents_v2(zhp, 0, B_TRUE, change_one,
+ clp) != 0) {
changelist_free(clp);
return (NULL);
}
- } else if (zfs_iter_children(zhp, 0, change_one, clp) != 0) {
+ } else if (zfs_iter_children_v2(zhp, 0, change_one, clp) != 0) {
changelist_free(clp);
return (NULL);
}
/*
* We have to re-open ourselves because we auto-close all the handles
* and can't tell the difference.
*/
if ((temp = zfs_open(zhp->zfs_hdl, zfs_get_name(zhp),
ZFS_TYPE_DATASET)) == NULL) {
changelist_free(clp);
return (NULL);
}
/*
* Always add ourself to the list. We add ourselves to the end so that
* we're the last to be unmounted.
*/
cn = zfs_alloc(zhp->zfs_hdl, sizeof (prop_changenode_t));
cn->cn_handle = temp;
cn->cn_mounted = (clp->cl_gflags & CL_GATHER_MOUNT_ALWAYS) ||
zfs_is_mounted(temp, NULL);
cn->cn_shared = zfs_is_shared(temp, NULL, NULL);
cn->cn_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
cn->cn_needpost = B_TRUE;
uu_avl_node_init(cn, &cn->cn_treenode, clp->cl_pool);
uu_avl_index_t idx;
if (uu_avl_find(clp->cl_tree, cn, NULL, &idx) == NULL) {
uu_avl_insert(clp->cl_tree, cn, idx);
} else {
free(cn);
zfs_close(temp);
}
/*
* If the mountpoint property was previously 'legacy', or 'none',
* record it as the behavior of changelist_postfix() will be different.
*/
if ((clp->cl_prop == ZFS_PROP_MOUNTPOINT) && legacy) {
/*
* do not automatically mount ex-legacy datasets if
* we specifically set canmount to noauto
*/
if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) !=
ZFS_CANMOUNT_NOAUTO)
clp->cl_waslegacy = B_TRUE;
}
return (clp);
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_crypto.c b/sys/contrib/openzfs/lib/libzfs/libzfs_crypto.c
index 40059063e21a..8f2a50d55e87 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_crypto.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_crypto.c
@@ -1,1818 +1,1818 @@
/*
* CDDL HEADER START
*
* This file and its contents are supplied under the terms of the
* Common Development and Distribution License ("CDDL"), version 1.0.
* You may only use this file in accordance with the terms of version
* 1.0 of the CDDL.
*
* A full copy of the text of the CDDL should have accompanied this
* source. A copy of the CDDL is also available via the Internet at
* http://www.illumos.org/license/CDDL.
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2017, Datto, Inc. All rights reserved.
* Copyright 2020 Joyent, Inc.
*/
#include <sys/zfs_context.h>
#include <sys/fs/zfs.h>
#include <sys/dsl_crypt.h>
#include <libintl.h>
#include <termios.h>
#include <signal.h>
#include <errno.h>
#include <openssl/evp.h>
#if LIBFETCH_DYNAMIC
#include <dlfcn.h>
#endif
#if LIBFETCH_IS_FETCH
#include <sys/param.h>
#include <stdio.h>
#include <fetch.h>
#elif LIBFETCH_IS_LIBCURL
#include <curl/curl.h>
#endif
#include <libzfs.h>
#include "libzfs_impl.h"
#include "zfeature_common.h"
/*
* User keys are used to decrypt the master encryption keys of a dataset. This
* indirection allows a user to change his / her access key without having to
* re-encrypt the entire dataset. User keys can be provided in one of several
* ways. Raw keys are simply given to the kernel as is. Similarly, hex keys
* are converted to binary and passed into the kernel. Password based keys are
* a bit more complicated. Passwords alone do not provide suitable entropy for
* encryption and may be too short or too long to be used. In order to derive
* a more appropriate key we use a PBKDF2 function. This function is designed
* to take a (relatively) long time to calculate in order to discourage
* attackers from guessing from a list of common passwords. PBKDF2 requires
* 2 additional parameters. The first is the number of iterations to run, which
* will ultimately determine how long it takes to derive the resulting key from
* the password. The second parameter is a salt that is randomly generated for
* each dataset. The salt is used to "tweak" PBKDF2 such that a group of
* attackers cannot reasonably generate a table of commonly known passwords to
* their output keys and expect it work for all past and future PBKDF2 users.
* We store the salt as a hidden property of the dataset (although it is
* technically ok if the salt is known to the attacker).
*/
#define MIN_PASSPHRASE_LEN 8
#define MAX_PASSPHRASE_LEN 512
#define MAX_KEY_PROMPT_ATTEMPTS 3
static int caught_interrupt;
static int get_key_material_file(libzfs_handle_t *, const char *, const char *,
zfs_keyformat_t, boolean_t, uint8_t **, size_t *);
static int get_key_material_https(libzfs_handle_t *, const char *, const char *,
zfs_keyformat_t, boolean_t, uint8_t **, size_t *);
static zfs_uri_handler_t uri_handlers[] = {
{ "file", get_key_material_file },
{ "https", get_key_material_https },
{ "http", get_key_material_https },
{ NULL, NULL }
};
static int
pkcs11_get_urandom(uint8_t *buf, size_t bytes)
{
int rand;
ssize_t bytes_read = 0;
rand = open("/dev/urandom", O_RDONLY | O_CLOEXEC);
if (rand < 0)
return (rand);
while (bytes_read < bytes) {
ssize_t rc = read(rand, buf + bytes_read, bytes - bytes_read);
if (rc < 0)
break;
bytes_read += rc;
}
(void) close(rand);
return (bytes_read);
}
static int
zfs_prop_parse_keylocation(libzfs_handle_t *restrict hdl, const char *str,
zfs_keylocation_t *restrict locp, char **restrict schemep)
{
*locp = ZFS_KEYLOCATION_NONE;
*schemep = NULL;
if (strcmp("prompt", str) == 0) {
*locp = ZFS_KEYLOCATION_PROMPT;
return (0);
}
regmatch_t pmatch[2];
if (regexec(&hdl->libzfs_urire, str, ARRAY_SIZE(pmatch),
pmatch, 0) == 0) {
size_t scheme_len;
if (pmatch[1].rm_so == -1) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Invalid URI"));
return (EINVAL);
}
scheme_len = pmatch[1].rm_eo - pmatch[1].rm_so;
*schemep = calloc(1, scheme_len + 1);
if (*schemep == NULL) {
int ret = errno;
errno = 0;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Invalid URI"));
return (ret);
}
(void) memcpy(*schemep, str + pmatch[1].rm_so, scheme_len);
*locp = ZFS_KEYLOCATION_URI;
return (0);
}
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Invalid keylocation"));
return (EINVAL);
}
static int
hex_key_to_raw(char *hex, int hexlen, uint8_t *out)
{
int ret, i;
unsigned int c;
for (i = 0; i < hexlen; i += 2) {
if (!isxdigit(hex[i]) || !isxdigit(hex[i + 1])) {
ret = EINVAL;
goto error;
}
ret = sscanf(&hex[i], "%02x", &c);
if (ret != 1) {
ret = EINVAL;
goto error;
}
out[i / 2] = c;
}
return (0);
error:
return (ret);
}
static void
catch_signal(int sig)
{
caught_interrupt = sig;
}
static const char *
get_format_prompt_string(zfs_keyformat_t format)
{
switch (format) {
case ZFS_KEYFORMAT_RAW:
return ("raw key");
case ZFS_KEYFORMAT_HEX:
return ("hex key");
case ZFS_KEYFORMAT_PASSPHRASE:
return ("passphrase");
default:
/* shouldn't happen */
return (NULL);
}
}
/* do basic validation of the key material */
static int
validate_key(libzfs_handle_t *hdl, zfs_keyformat_t keyformat,
const char *key, size_t keylen, boolean_t do_verify)
{
switch (keyformat) {
case ZFS_KEYFORMAT_RAW:
/* verify the key length is correct */
if (keylen < WRAPPING_KEY_LEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Raw key too short (expected %u)."),
WRAPPING_KEY_LEN);
return (EINVAL);
}
if (keylen > WRAPPING_KEY_LEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Raw key too long (expected %u)."),
WRAPPING_KEY_LEN);
return (EINVAL);
}
break;
case ZFS_KEYFORMAT_HEX:
/* verify the key length is correct */
if (keylen < WRAPPING_KEY_LEN * 2) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Hex key too short (expected %u)."),
WRAPPING_KEY_LEN * 2);
return (EINVAL);
}
if (keylen > WRAPPING_KEY_LEN * 2) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Hex key too long (expected %u)."),
WRAPPING_KEY_LEN * 2);
return (EINVAL);
}
/* check for invalid hex digits */
for (size_t i = 0; i < WRAPPING_KEY_LEN * 2; i++) {
if (!isxdigit(key[i])) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Invalid hex character detected."));
return (EINVAL);
}
}
break;
case ZFS_KEYFORMAT_PASSPHRASE:
/*
* Verify the length is within bounds when setting a new key,
* but not when loading an existing key.
*/
if (!do_verify)
break;
if (keylen > MAX_PASSPHRASE_LEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Passphrase too long (max %u)."),
MAX_PASSPHRASE_LEN);
return (EINVAL);
}
if (keylen < MIN_PASSPHRASE_LEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Passphrase too short (min %u)."),
MIN_PASSPHRASE_LEN);
return (EINVAL);
}
break;
default:
/* can't happen, checked above */
break;
}
return (0);
}
static int
libzfs_getpassphrase(zfs_keyformat_t keyformat, boolean_t is_reenter,
boolean_t new_key, const char *fsname,
char **restrict res, size_t *restrict reslen)
{
FILE *f = stdin;
size_t buflen = 0;
ssize_t bytes;
int ret = 0;
struct termios old_term, new_term;
struct sigaction act, osigint, osigtstp;
*res = NULL;
*reslen = 0;
/*
* handle SIGINT and ignore SIGSTP. This is necessary to
* restore the state of the terminal.
*/
caught_interrupt = 0;
act.sa_flags = 0;
(void) sigemptyset(&act.sa_mask);
act.sa_handler = catch_signal;
(void) sigaction(SIGINT, &act, &osigint);
act.sa_handler = SIG_IGN;
(void) sigaction(SIGTSTP, &act, &osigtstp);
(void) printf("%s %s%s",
is_reenter ? "Re-enter" : "Enter",
new_key ? "new " : "",
get_format_prompt_string(keyformat));
if (fsname != NULL)
(void) printf(" for '%s'", fsname);
(void) fputc(':', stdout);
(void) fflush(stdout);
/* disable the terminal echo for key input */
(void) tcgetattr(fileno(f), &old_term);
new_term = old_term;
new_term.c_lflag &= ~(ECHO | ECHOE | ECHOK | ECHONL);
ret = tcsetattr(fileno(f), TCSAFLUSH, &new_term);
if (ret != 0) {
ret = errno;
errno = 0;
goto out;
}
bytes = getline(res, &buflen, f);
if (bytes < 0) {
ret = errno;
errno = 0;
goto out;
}
/* trim the ending newline if it exists */
if (bytes > 0 && (*res)[bytes - 1] == '\n') {
(*res)[bytes - 1] = '\0';
bytes--;
}
*reslen = bytes;
out:
/* reset the terminal */
(void) tcsetattr(fileno(f), TCSAFLUSH, &old_term);
(void) sigaction(SIGINT, &osigint, NULL);
(void) sigaction(SIGTSTP, &osigtstp, NULL);
/* if we caught a signal, re-throw it now */
if (caught_interrupt != 0)
(void) kill(getpid(), caught_interrupt);
/* print the newline that was not echo'd */
(void) printf("\n");
return (ret);
}
static int
get_key_interactive(libzfs_handle_t *restrict hdl, const char *fsname,
zfs_keyformat_t keyformat, boolean_t confirm_key, boolean_t newkey,
uint8_t **restrict outbuf, size_t *restrict len_out)
{
char *buf = NULL, *buf2 = NULL;
size_t buflen = 0, buf2len = 0;
int ret = 0;
ASSERT(isatty(fileno(stdin)));
/* raw keys cannot be entered on the terminal */
if (keyformat == ZFS_KEYFORMAT_RAW) {
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Cannot enter raw keys on the terminal"));
goto out;
}
/* prompt for the key */
if ((ret = libzfs_getpassphrase(keyformat, B_FALSE, newkey, fsname,
&buf, &buflen)) != 0) {
free(buf);
buf = NULL;
buflen = 0;
goto out;
}
if (!confirm_key)
goto out;
if ((ret = validate_key(hdl, keyformat, buf, buflen, confirm_key)) !=
0) {
free(buf);
return (ret);
}
ret = libzfs_getpassphrase(keyformat, B_TRUE, newkey, fsname, &buf2,
&buf2len);
if (ret != 0) {
free(buf);
free(buf2);
buf = buf2 = NULL;
buflen = buf2len = 0;
goto out;
}
if (buflen != buf2len || strcmp(buf, buf2) != 0) {
free(buf);
buf = NULL;
buflen = 0;
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Provided keys do not match."));
}
free(buf2);
out:
*outbuf = (uint8_t *)buf;
*len_out = buflen;
return (ret);
}
static int
get_key_material_raw(FILE *fd, zfs_keyformat_t keyformat,
uint8_t **buf, size_t *len_out)
{
int ret = 0;
size_t buflen = 0;
*len_out = 0;
/* read the key material */
if (keyformat != ZFS_KEYFORMAT_RAW) {
ssize_t bytes;
bytes = getline((char **)buf, &buflen, fd);
if (bytes < 0) {
ret = errno;
errno = 0;
goto out;
}
/* trim the ending newline if it exists */
if (bytes > 0 && (*buf)[bytes - 1] == '\n') {
(*buf)[bytes - 1] = '\0';
bytes--;
}
*len_out = bytes;
} else {
size_t n;
/*
* Raw keys may have newline characters in them and so can't
* use getline(). Here we attempt to read 33 bytes so that we
* can properly check the key length (the file should only have
* 32 bytes).
*/
*buf = malloc((WRAPPING_KEY_LEN + 1) * sizeof (uint8_t));
if (*buf == NULL) {
ret = ENOMEM;
goto out;
}
n = fread(*buf, 1, WRAPPING_KEY_LEN + 1, fd);
if (n == 0 || ferror(fd)) {
/* size errors are handled by the calling function */
free(*buf);
*buf = NULL;
ret = errno;
errno = 0;
goto out;
}
*len_out = n;
}
out:
return (ret);
}
static int
get_key_material_file(libzfs_handle_t *hdl, const char *uri,
const char *fsname, zfs_keyformat_t keyformat, boolean_t newkey,
uint8_t **restrict buf, size_t *restrict len_out)
{
(void) fsname, (void) newkey;
FILE *f = NULL;
int ret = 0;
if (strlen(uri) < 7)
return (EINVAL);
if ((f = fopen(uri + 7, "re")) == NULL) {
ret = errno;
errno = 0;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Failed to open key material file: %s"), strerror(ret));
return (ret);
}
ret = get_key_material_raw(f, keyformat, buf, len_out);
(void) fclose(f);
return (ret);
}
static int
get_key_material_https(libzfs_handle_t *hdl, const char *uri,
const char *fsname, zfs_keyformat_t keyformat, boolean_t newkey,
uint8_t **restrict buf, size_t *restrict len_out)
{
(void) fsname, (void) newkey;
int ret = 0;
FILE *key = NULL;
boolean_t is_http = strncmp(uri, "http:", strlen("http:")) == 0;
if (strlen(uri) < (is_http ? 7 : 8)) {
ret = EINVAL;
goto end;
}
#if LIBFETCH_DYNAMIC
#define LOAD_FUNCTION(func) \
__typeof__(func) *func = dlsym(hdl->libfetch, #func);
if (hdl->libfetch == NULL)
hdl->libfetch = dlopen(LIBFETCH_SONAME, RTLD_LAZY);
if (hdl->libfetch == NULL) {
hdl->libfetch = (void *)-1;
char *err = dlerror();
if (err)
hdl->libfetch_load_error = strdup(err);
}
if (hdl->libfetch == (void *)-1) {
ret = ENOSYS;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Couldn't load %s: %s"),
LIBFETCH_SONAME, hdl->libfetch_load_error ?: "(?)");
goto end;
}
boolean_t ok;
#if LIBFETCH_IS_FETCH
LOAD_FUNCTION(fetchGetURL);
char *fetchLastErrString = dlsym(hdl->libfetch, "fetchLastErrString");
ok = fetchGetURL && fetchLastErrString;
#elif LIBFETCH_IS_LIBCURL
LOAD_FUNCTION(curl_easy_init);
LOAD_FUNCTION(curl_easy_setopt);
LOAD_FUNCTION(curl_easy_perform);
LOAD_FUNCTION(curl_easy_cleanup);
LOAD_FUNCTION(curl_easy_strerror);
LOAD_FUNCTION(curl_easy_getinfo);
ok = curl_easy_init && curl_easy_setopt && curl_easy_perform &&
curl_easy_cleanup && curl_easy_strerror && curl_easy_getinfo;
#endif
if (!ok) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"keylocation=%s back-end %s missing symbols."),
is_http ? "http://" : "https://", LIBFETCH_SONAME);
ret = ENOSYS;
goto end;
}
#endif
#if LIBFETCH_IS_FETCH
key = fetchGetURL(uri, "");
if (key == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Couldn't GET %s: %s"),
uri, fetchLastErrString);
ret = ENETDOWN;
}
#elif LIBFETCH_IS_LIBCURL
CURL *curl = curl_easy_init();
if (curl == NULL) {
ret = ENOTSUP;
goto end;
}
int kfd = -1;
#ifdef O_TMPFILE
kfd = open(getenv("TMPDIR") ?: "/tmp",
O_RDWR | O_TMPFILE | O_EXCL | O_CLOEXEC, 0600);
if (kfd != -1)
goto kfdok;
#endif
char *path;
if (asprintf(&path,
"%s/libzfs-XXXXXXXX.https", getenv("TMPDIR") ?: "/tmp") == -1) {
ret = ENOMEM;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "%s"),
strerror(ret));
goto end;
}
kfd = mkostemps(path, strlen(".https"), O_CLOEXEC);
if (kfd == -1) {
ret = errno;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Couldn't create temporary file %s: %s"),
path, strerror(ret));
free(path);
goto end;
}
(void) unlink(path);
free(path);
kfdok:
if ((key = fdopen(kfd, "r+")) == NULL) {
ret = errno;
(void) close(kfd);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Couldn't reopen temporary file: %s"), strerror(ret));
goto end;
}
char errbuf[CURL_ERROR_SIZE] = "";
char *cainfo = getenv("SSL_CA_CERT_FILE"); /* matches fetch(3) */
char *capath = getenv("SSL_CA_CERT_PATH"); /* matches fetch(3) */
char *clcert = getenv("SSL_CLIENT_CERT_FILE"); /* matches fetch(3) */
char *clkey = getenv("SSL_CLIENT_KEY_FILE"); /* matches fetch(3) */
(void) curl_easy_setopt(curl, CURLOPT_URL, uri);
(void) curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L);
(void) curl_easy_setopt(curl, CURLOPT_TIMEOUT_MS, 30000L);
(void) curl_easy_setopt(curl, CURLOPT_WRITEDATA, key);
(void) curl_easy_setopt(curl, CURLOPT_ERRORBUFFER, errbuf);
if (cainfo != NULL)
(void) curl_easy_setopt(curl, CURLOPT_CAINFO, cainfo);
if (capath != NULL)
(void) curl_easy_setopt(curl, CURLOPT_CAPATH, capath);
if (clcert != NULL)
(void) curl_easy_setopt(curl, CURLOPT_SSLCERT, clcert);
if (clkey != NULL)
(void) curl_easy_setopt(curl, CURLOPT_SSLKEY, clkey);
CURLcode res = curl_easy_perform(curl);
if (res != CURLE_OK) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Failed to connect to %s: %s"),
uri, strlen(errbuf) ? errbuf : curl_easy_strerror(res));
ret = ENETDOWN;
} else {
long resp = 200;
(void) curl_easy_getinfo(curl, CURLINFO_RESPONSE_CODE, &resp);
if (resp < 200 || resp >= 300) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Couldn't GET %s: %ld"),
uri, resp);
ret = ENOENT;
} else
rewind(key);
}
curl_easy_cleanup(curl);
#else
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"No keylocation=%s back-end."), is_http ? "http://" : "https://");
ret = ENOSYS;
#endif
end:
if (ret == 0)
ret = get_key_material_raw(key, keyformat, buf, len_out);
if (key != NULL)
fclose(key);
return (ret);
}
/*
* Attempts to fetch key material, no matter where it might live. The key
* material is allocated and returned in km_out. *can_retry_out will be set
* to B_TRUE if the user is providing the key material interactively, allowing
* for re-entry attempts.
*/
static int
get_key_material(libzfs_handle_t *hdl, boolean_t do_verify, boolean_t newkey,
zfs_keyformat_t keyformat, const char *keylocation, const char *fsname,
uint8_t **km_out, size_t *kmlen_out, boolean_t *can_retry_out)
{
int ret;
zfs_keylocation_t keyloc = ZFS_KEYLOCATION_NONE;
uint8_t *km = NULL;
size_t kmlen = 0;
char *uri_scheme = NULL;
zfs_uri_handler_t *handler = NULL;
boolean_t can_retry = B_FALSE;
/* verify and parse the keylocation */
ret = zfs_prop_parse_keylocation(hdl, keylocation, &keyloc,
&uri_scheme);
if (ret != 0)
goto error;
/* open the appropriate file descriptor */
switch (keyloc) {
case ZFS_KEYLOCATION_PROMPT:
if (isatty(fileno(stdin))) {
can_retry = keyformat != ZFS_KEYFORMAT_RAW;
ret = get_key_interactive(hdl, fsname, keyformat,
do_verify, newkey, &km, &kmlen);
} else {
/* fetch the key material into the buffer */
ret = get_key_material_raw(stdin, keyformat, &km,
&kmlen);
}
if (ret != 0)
goto error;
break;
case ZFS_KEYLOCATION_URI:
ret = ENOTSUP;
for (handler = uri_handlers; handler->zuh_scheme != NULL;
handler++) {
if (strcmp(handler->zuh_scheme, uri_scheme) != 0)
continue;
if ((ret = handler->zuh_handler(hdl, keylocation,
fsname, keyformat, newkey, &km, &kmlen)) != 0)
goto error;
break;
}
if (ret == ENOTSUP) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"URI scheme is not supported"));
goto error;
}
break;
default:
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Invalid keylocation."));
goto error;
}
if ((ret = validate_key(hdl, keyformat, (const char *)km, kmlen,
do_verify)) != 0)
goto error;
*km_out = km;
*kmlen_out = kmlen;
if (can_retry_out != NULL)
*can_retry_out = can_retry;
free(uri_scheme);
return (0);
error:
free(km);
*km_out = NULL;
*kmlen_out = 0;
if (can_retry_out != NULL)
*can_retry_out = can_retry;
free(uri_scheme);
return (ret);
}
static int
derive_key(libzfs_handle_t *hdl, zfs_keyformat_t format, uint64_t iters,
uint8_t *key_material, uint64_t salt,
uint8_t **key_out)
{
int ret;
uint8_t *key;
*key_out = NULL;
key = zfs_alloc(hdl, WRAPPING_KEY_LEN);
switch (format) {
case ZFS_KEYFORMAT_RAW:
memcpy(key, key_material, WRAPPING_KEY_LEN);
break;
case ZFS_KEYFORMAT_HEX:
ret = hex_key_to_raw((char *)key_material,
WRAPPING_KEY_LEN * 2, key);
if (ret != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Invalid hex key provided."));
goto error;
}
break;
case ZFS_KEYFORMAT_PASSPHRASE:
salt = LE_64(salt);
ret = PKCS5_PBKDF2_HMAC_SHA1((char *)key_material,
strlen((char *)key_material), ((uint8_t *)&salt),
sizeof (uint64_t), iters, WRAPPING_KEY_LEN, key);
if (ret != 1) {
ret = EIO;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Failed to generate key from passphrase."));
goto error;
}
break;
default:
ret = EINVAL;
goto error;
}
*key_out = key;
return (0);
error:
free(key);
*key_out = NULL;
return (ret);
}
static boolean_t
encryption_feature_is_enabled(zpool_handle_t *zph)
{
nvlist_t *features;
uint64_t feat_refcount;
/* check that features can be enabled */
if (zpool_get_prop_int(zph, ZPOOL_PROP_VERSION, NULL)
< SPA_VERSION_FEATURES)
return (B_FALSE);
/* check for crypto feature */
features = zpool_get_features(zph);
if (!features || nvlist_lookup_uint64(features,
spa_feature_table[SPA_FEATURE_ENCRYPTION].fi_guid,
&feat_refcount) != 0)
return (B_FALSE);
return (B_TRUE);
}
static int
populate_create_encryption_params_nvlists(libzfs_handle_t *hdl,
zfs_handle_t *zhp, boolean_t newkey, zfs_keyformat_t keyformat,
const char *keylocation, nvlist_t *props, uint8_t **wkeydata,
uint_t *wkeylen)
{
int ret;
uint64_t iters = 0, salt = 0;
uint8_t *key_material = NULL;
size_t key_material_len = 0;
uint8_t *key_data = NULL;
const char *fsname = (zhp) ? zfs_get_name(zhp) : NULL;
/* get key material from keyformat and keylocation */
ret = get_key_material(hdl, B_TRUE, newkey, keyformat, keylocation,
fsname, &key_material, &key_material_len, NULL);
if (ret != 0)
goto error;
/* passphrase formats require a salt and pbkdf2 iters property */
if (keyformat == ZFS_KEYFORMAT_PASSPHRASE) {
/* always generate a new salt */
ret = pkcs11_get_urandom((uint8_t *)&salt, sizeof (uint64_t));
if (ret != sizeof (uint64_t)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Failed to generate salt."));
goto error;
}
ret = nvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_PBKDF2_SALT), salt);
if (ret != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Failed to add salt to properties."));
goto error;
}
/*
* If not otherwise specified, use the default number of
* pbkdf2 iterations. If specified, we have already checked
* that the given value is greater than MIN_PBKDF2_ITERATIONS
* during zfs_valid_proplist().
*/
ret = nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS), &iters);
if (ret == ENOENT) {
iters = DEFAULT_PBKDF2_ITERATIONS;
ret = nvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS), iters);
if (ret != 0)
goto error;
} else if (ret != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Failed to get pbkdf2 iterations."));
goto error;
}
} else {
/* check that pbkdf2iters was not specified by the user */
ret = nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS), &iters);
if (ret == 0) {
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Cannot specify pbkdf2iters with a non-passphrase "
"keyformat."));
goto error;
}
}
/* derive a key from the key material */
ret = derive_key(hdl, keyformat, iters, key_material, salt, &key_data);
if (ret != 0)
goto error;
free(key_material);
*wkeydata = key_data;
*wkeylen = WRAPPING_KEY_LEN;
return (0);
error:
if (key_material != NULL)
free(key_material);
if (key_data != NULL)
free(key_data);
*wkeydata = NULL;
*wkeylen = 0;
return (ret);
}
static boolean_t
proplist_has_encryption_props(nvlist_t *props)
{
int ret;
uint64_t intval;
const char *strval;
ret = nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_ENCRYPTION), &intval);
if (ret == 0 && intval != ZIO_CRYPT_OFF)
return (B_TRUE);
ret = nvlist_lookup_string(props,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION), &strval);
if (ret == 0 && strcmp(strval, "none") != 0)
return (B_TRUE);
ret = nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_KEYFORMAT), &intval);
if (ret == 0)
return (B_TRUE);
ret = nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS), &intval);
if (ret == 0)
return (B_TRUE);
return (B_FALSE);
}
int
zfs_crypto_get_encryption_root(zfs_handle_t *zhp, boolean_t *is_encroot,
char *buf)
{
int ret;
char prop_encroot[MAXNAMELEN];
/* if the dataset isn't encrypted, just return */
if (zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION) == ZIO_CRYPT_OFF) {
*is_encroot = B_FALSE;
if (buf != NULL)
buf[0] = '\0';
return (0);
}
ret = zfs_prop_get(zhp, ZFS_PROP_ENCRYPTION_ROOT, prop_encroot,
sizeof (prop_encroot), NULL, NULL, 0, B_TRUE);
if (ret != 0) {
*is_encroot = B_FALSE;
if (buf != NULL)
buf[0] = '\0';
return (ret);
}
*is_encroot = strcmp(prop_encroot, zfs_get_name(zhp)) == 0;
if (buf != NULL)
strcpy(buf, prop_encroot);
return (0);
}
int
zfs_crypto_create(libzfs_handle_t *hdl, char *parent_name, nvlist_t *props,
nvlist_t *pool_props, boolean_t stdin_available, uint8_t **wkeydata_out,
uint_t *wkeylen_out)
{
int ret;
char errbuf[ERRBUFLEN];
uint64_t crypt = ZIO_CRYPT_INHERIT, pcrypt = ZIO_CRYPT_INHERIT;
uint64_t keyformat = ZFS_KEYFORMAT_NONE;
const char *keylocation = NULL;
zfs_handle_t *pzhp = NULL;
uint8_t *wkeydata = NULL;
uint_t wkeylen = 0;
boolean_t local_crypt = B_TRUE;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "Encryption create error"));
/* lookup crypt from props */
ret = nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_ENCRYPTION), &crypt);
if (ret != 0)
local_crypt = B_FALSE;
/* lookup key location and format from props */
(void) nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_KEYFORMAT), &keyformat);
(void) nvlist_lookup_string(props,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION), &keylocation);
if (parent_name != NULL) {
/* get a reference to parent dataset */
pzhp = make_dataset_handle(hdl, parent_name);
if (pzhp == NULL) {
ret = ENOENT;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Failed to lookup parent."));
goto out;
}
/* Lookup parent's crypt */
pcrypt = zfs_prop_get_int(pzhp, ZFS_PROP_ENCRYPTION);
/* Params require the encryption feature */
if (!encryption_feature_is_enabled(pzhp->zpool_hdl)) {
if (proplist_has_encryption_props(props)) {
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Encryption feature not enabled."));
goto out;
}
ret = 0;
goto out;
}
} else {
/*
* special case for root dataset where encryption feature
* feature won't be on disk yet
*/
if (!nvlist_exists(pool_props, "feature@encryption")) {
if (proplist_has_encryption_props(props)) {
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Encryption feature not enabled."));
goto out;
}
ret = 0;
goto out;
}
pcrypt = ZIO_CRYPT_OFF;
}
/* Get the inherited encryption property if we don't have it locally */
if (!local_crypt)
crypt = pcrypt;
/*
* At this point crypt should be the actual encryption value. If
* encryption is off just verify that no encryption properties have
* been specified and return.
*/
if (crypt == ZIO_CRYPT_OFF) {
if (proplist_has_encryption_props(props)) {
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Encryption must be turned on to set encryption "
"properties."));
goto out;
}
ret = 0;
goto out;
}
/*
* If we have a parent crypt it is valid to specify encryption alone.
* This will result in a child that is encrypted with the chosen
* encryption suite that will also inherit the parent's key. If
* the parent is not encrypted we need an encryption suite provided.
*/
if (pcrypt == ZIO_CRYPT_OFF && keylocation == NULL &&
keyformat == ZFS_KEYFORMAT_NONE) {
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Keyformat required for new encryption root."));
goto out;
}
/*
* Specifying a keylocation implies this will be a new encryption root.
* Check that a keyformat is also specified.
*/
if (keylocation != NULL && keyformat == ZFS_KEYFORMAT_NONE) {
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Keyformat required for new encryption root."));
goto out;
}
/* default to prompt if no keylocation is specified */
if (keyformat != ZFS_KEYFORMAT_NONE && keylocation == NULL) {
keylocation = (char *)"prompt";
ret = nvlist_add_string(props,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION), keylocation);
if (ret != 0)
goto out;
}
/*
* If a local key is provided, this dataset will be a new
* encryption root. Populate the encryption params.
*/
if (keylocation != NULL) {
/*
* 'zfs recv -o keylocation=prompt' won't work because stdin
* is being used by the send stream, so we disallow it.
*/
if (!stdin_available && strcmp(keylocation, "prompt") == 0) {
ret = EINVAL;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Cannot use "
"'prompt' keylocation because stdin is in use."));
goto out;
}
ret = populate_create_encryption_params_nvlists(hdl, NULL,
B_TRUE, keyformat, keylocation, props, &wkeydata,
&wkeylen);
if (ret != 0)
goto out;
}
if (pzhp != NULL)
zfs_close(pzhp);
*wkeydata_out = wkeydata;
*wkeylen_out = wkeylen;
return (0);
out:
if (pzhp != NULL)
zfs_close(pzhp);
if (wkeydata != NULL)
free(wkeydata);
*wkeydata_out = NULL;
*wkeylen_out = 0;
return (ret);
}
int
zfs_crypto_clone_check(libzfs_handle_t *hdl, zfs_handle_t *origin_zhp,
char *parent_name, nvlist_t *props)
{
(void) origin_zhp, (void) parent_name;
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "Encryption clone error"));
/*
* No encryption properties should be specified. They will all be
* inherited from the origin dataset.
*/
if (nvlist_exists(props, zfs_prop_to_name(ZFS_PROP_KEYFORMAT)) ||
nvlist_exists(props, zfs_prop_to_name(ZFS_PROP_KEYLOCATION)) ||
nvlist_exists(props, zfs_prop_to_name(ZFS_PROP_ENCRYPTION)) ||
nvlist_exists(props, zfs_prop_to_name(ZFS_PROP_PBKDF2_ITERS))) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Encryption properties must inherit from origin dataset."));
return (EINVAL);
}
return (0);
}
typedef struct loadkeys_cbdata {
uint64_t cb_numfailed;
uint64_t cb_numattempted;
} loadkey_cbdata_t;
static int
load_keys_cb(zfs_handle_t *zhp, void *arg)
{
int ret;
boolean_t is_encroot;
loadkey_cbdata_t *cb = arg;
uint64_t keystatus = zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS);
/* only attempt to load keys for encryption roots */
ret = zfs_crypto_get_encryption_root(zhp, &is_encroot, NULL);
if (ret != 0 || !is_encroot)
goto out;
/* don't attempt to load already loaded keys */
if (keystatus == ZFS_KEYSTATUS_AVAILABLE)
goto out;
/* Attempt to load the key. Record status in cb. */
cb->cb_numattempted++;
ret = zfs_crypto_load_key(zhp, B_FALSE, NULL);
if (ret)
cb->cb_numfailed++;
out:
- (void) zfs_iter_filesystems(zhp, 0, load_keys_cb, cb);
+ (void) zfs_iter_filesystems_v2(zhp, 0, load_keys_cb, cb);
zfs_close(zhp);
/* always return 0, since this function is best effort */
return (0);
}
/*
* This function is best effort. It attempts to load all the keys for the given
* filesystem and all of its children.
*/
int
zfs_crypto_attempt_load_keys(libzfs_handle_t *hdl, const char *fsname)
{
int ret;
zfs_handle_t *zhp = NULL;
loadkey_cbdata_t cb = { 0 };
zhp = zfs_open(hdl, fsname, ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL) {
ret = ENOENT;
goto error;
}
ret = load_keys_cb(zfs_handle_dup(zhp), &cb);
if (ret)
goto error;
(void) printf(gettext("%llu / %llu keys successfully loaded\n"),
(u_longlong_t)(cb.cb_numattempted - cb.cb_numfailed),
(u_longlong_t)cb.cb_numattempted);
if (cb.cb_numfailed != 0) {
ret = -1;
goto error;
}
zfs_close(zhp);
return (0);
error:
if (zhp != NULL)
zfs_close(zhp);
return (ret);
}
int
zfs_crypto_load_key(zfs_handle_t *zhp, boolean_t noop,
const char *alt_keylocation)
{
int ret, attempts = 0;
char errbuf[ERRBUFLEN];
uint64_t keystatus, iters = 0, salt = 0;
uint64_t keyformat = ZFS_KEYFORMAT_NONE;
char prop_keylocation[MAXNAMELEN];
char prop_encroot[MAXNAMELEN];
const char *keylocation = NULL;
uint8_t *key_material = NULL, *key_data = NULL;
size_t key_material_len;
boolean_t is_encroot, can_retry = B_FALSE, correctible = B_FALSE;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "Key load error"));
/* check that encryption is enabled for the pool */
if (!encryption_feature_is_enabled(zhp->zpool_hdl)) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Encryption feature not enabled."));
ret = EINVAL;
goto error;
}
/* Fetch the keyformat. Check that the dataset is encrypted. */
keyformat = zfs_prop_get_int(zhp, ZFS_PROP_KEYFORMAT);
if (keyformat == ZFS_KEYFORMAT_NONE) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"'%s' is not encrypted."), zfs_get_name(zhp));
ret = EINVAL;
goto error;
}
/*
* Fetch the key location. Check that we are working with an
* encryption root.
*/
ret = zfs_crypto_get_encryption_root(zhp, &is_encroot, prop_encroot);
if (ret != 0) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Failed to get encryption root for '%s'."),
zfs_get_name(zhp));
goto error;
} else if (!is_encroot) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Keys must be loaded for encryption root of '%s' (%s)."),
zfs_get_name(zhp), prop_encroot);
ret = EINVAL;
goto error;
}
/*
* if the caller has elected to override the keylocation property
* use that instead
*/
if (alt_keylocation != NULL) {
keylocation = alt_keylocation;
} else {
ret = zfs_prop_get(zhp, ZFS_PROP_KEYLOCATION, prop_keylocation,
sizeof (prop_keylocation), NULL, NULL, 0, B_TRUE);
if (ret != 0) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Failed to get keylocation for '%s'."),
zfs_get_name(zhp));
goto error;
}
keylocation = prop_keylocation;
}
/* check that the key is unloaded unless this is a noop */
if (!noop) {
keystatus = zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS);
if (keystatus == ZFS_KEYSTATUS_AVAILABLE) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Key already loaded for '%s'."), zfs_get_name(zhp));
ret = EEXIST;
goto error;
}
}
/* passphrase formats require a salt and pbkdf2_iters property */
if (keyformat == ZFS_KEYFORMAT_PASSPHRASE) {
salt = zfs_prop_get_int(zhp, ZFS_PROP_PBKDF2_SALT);
iters = zfs_prop_get_int(zhp, ZFS_PROP_PBKDF2_ITERS);
}
try_again:
/* fetching and deriving the key are correctable errors. set the flag */
correctible = B_TRUE;
/* get key material from key format and location */
ret = get_key_material(zhp->zfs_hdl, B_FALSE, B_FALSE, keyformat,
keylocation, zfs_get_name(zhp), &key_material, &key_material_len,
&can_retry);
if (ret != 0)
goto error;
/* derive a key from the key material */
ret = derive_key(zhp->zfs_hdl, keyformat, iters, key_material, salt,
&key_data);
if (ret != 0)
goto error;
correctible = B_FALSE;
/* pass the wrapping key and noop flag to the ioctl */
ret = lzc_load_key(zhp->zfs_name, noop, key_data, WRAPPING_KEY_LEN);
if (ret != 0) {
switch (ret) {
case EPERM:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Permission denied."));
break;
case EINVAL:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Invalid parameters provided for dataset %s."),
zfs_get_name(zhp));
break;
case EEXIST:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Key already loaded for '%s'."), zfs_get_name(zhp));
break;
case EBUSY:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"'%s' is busy."), zfs_get_name(zhp));
break;
case EACCES:
correctible = B_TRUE;
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Incorrect key provided for '%s'."),
zfs_get_name(zhp));
break;
case ZFS_ERR_CRYPTO_NOTSUP:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"'%s' uses an unsupported encryption suite."),
zfs_get_name(zhp));
break;
}
goto error;
}
free(key_material);
free(key_data);
return (0);
error:
zfs_error(zhp->zfs_hdl, EZFS_CRYPTOFAILED, errbuf);
if (key_material != NULL) {
free(key_material);
key_material = NULL;
}
if (key_data != NULL) {
free(key_data);
key_data = NULL;
}
/*
* Here we decide if it is ok to allow the user to retry entering their
* key. The can_retry flag will be set if the user is entering their
* key from an interactive prompt. The correctable flag will only be
* set if an error that occurred could be corrected by retrying. Both
* flags are needed to allow the user to attempt key entry again
*/
attempts++;
if (can_retry && correctible && attempts < MAX_KEY_PROMPT_ATTEMPTS)
goto try_again;
return (ret);
}
int
zfs_crypto_unload_key(zfs_handle_t *zhp)
{
int ret;
char errbuf[ERRBUFLEN];
char prop_encroot[MAXNAMELEN];
uint64_t keystatus, keyformat;
boolean_t is_encroot;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "Key unload error"));
/* check that encryption is enabled for the pool */
if (!encryption_feature_is_enabled(zhp->zpool_hdl)) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Encryption feature not enabled."));
ret = EINVAL;
goto error;
}
/* Fetch the keyformat. Check that the dataset is encrypted. */
keyformat = zfs_prop_get_int(zhp, ZFS_PROP_KEYFORMAT);
if (keyformat == ZFS_KEYFORMAT_NONE) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"'%s' is not encrypted."), zfs_get_name(zhp));
ret = EINVAL;
goto error;
}
/*
* Fetch the key location. Check that we are working with an
* encryption root.
*/
ret = zfs_crypto_get_encryption_root(zhp, &is_encroot, prop_encroot);
if (ret != 0) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Failed to get encryption root for '%s'."),
zfs_get_name(zhp));
goto error;
} else if (!is_encroot) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Keys must be unloaded for encryption root of '%s' (%s)."),
zfs_get_name(zhp), prop_encroot);
ret = EINVAL;
goto error;
}
/* check that the key is loaded */
keystatus = zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS);
if (keystatus == ZFS_KEYSTATUS_UNAVAILABLE) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Key already unloaded for '%s'."), zfs_get_name(zhp));
ret = EACCES;
goto error;
}
/* call the ioctl */
ret = lzc_unload_key(zhp->zfs_name);
if (ret != 0) {
switch (ret) {
case EPERM:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Permission denied."));
break;
case EACCES:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Key already unloaded for '%s'."),
zfs_get_name(zhp));
break;
case EBUSY:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"'%s' is busy."), zfs_get_name(zhp));
break;
}
zfs_error(zhp->zfs_hdl, EZFS_CRYPTOFAILED, errbuf);
}
return (ret);
error:
zfs_error(zhp->zfs_hdl, EZFS_CRYPTOFAILED, errbuf);
return (ret);
}
static int
zfs_crypto_verify_rewrap_nvlist(zfs_handle_t *zhp, nvlist_t *props,
nvlist_t **props_out, char *errbuf)
{
int ret;
nvpair_t *elem = NULL;
zfs_prop_t prop;
nvlist_t *new_props = NULL;
new_props = fnvlist_alloc();
/*
* loop through all provided properties, we should only have
* keyformat, keylocation and pbkdf2iters. The actual validation of
* values is done by zfs_valid_proplist().
*/
while ((elem = nvlist_next_nvpair(props, elem)) != NULL) {
const char *propname = nvpair_name(elem);
prop = zfs_name_to_prop(propname);
switch (prop) {
case ZFS_PROP_PBKDF2_ITERS:
case ZFS_PROP_KEYFORMAT:
case ZFS_PROP_KEYLOCATION:
break;
default:
ret = EINVAL;
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Only keyformat, keylocation and pbkdf2iters may "
"be set with this command."));
goto error;
}
}
new_props = zfs_valid_proplist(zhp->zfs_hdl, zhp->zfs_type, props,
zfs_prop_get_int(zhp, ZFS_PROP_ZONED), NULL, zhp->zpool_hdl,
B_TRUE, errbuf);
if (new_props == NULL) {
ret = EINVAL;
goto error;
}
*props_out = new_props;
return (0);
error:
nvlist_free(new_props);
*props_out = NULL;
return (ret);
}
int
zfs_crypto_rewrap(zfs_handle_t *zhp, nvlist_t *raw_props, boolean_t inheritkey)
{
int ret;
char errbuf[ERRBUFLEN];
boolean_t is_encroot;
nvlist_t *props = NULL;
uint8_t *wkeydata = NULL;
uint_t wkeylen = 0;
dcp_cmd_t cmd = (inheritkey) ? DCP_CMD_INHERIT : DCP_CMD_NEW_KEY;
uint64_t crypt, pcrypt, keystatus, pkeystatus;
uint64_t keyformat = ZFS_KEYFORMAT_NONE;
zfs_handle_t *pzhp = NULL;
const char *keylocation = NULL;
char origin_name[MAXNAMELEN];
char prop_keylocation[MAXNAMELEN];
char parent_name[ZFS_MAX_DATASET_NAME_LEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "Key change error"));
/* check that encryption is enabled for the pool */
if (!encryption_feature_is_enabled(zhp->zpool_hdl)) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Encryption feature not enabled."));
ret = EINVAL;
goto error;
}
/* get crypt from dataset */
crypt = zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION);
if (crypt == ZIO_CRYPT_OFF) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Dataset not encrypted."));
ret = EINVAL;
goto error;
}
/* get the encryption root of the dataset */
ret = zfs_crypto_get_encryption_root(zhp, &is_encroot, NULL);
if (ret != 0) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Failed to get encryption root for '%s'."),
zfs_get_name(zhp));
goto error;
}
/* Clones use their origin's key and cannot rewrap it */
ret = zfs_prop_get(zhp, ZFS_PROP_ORIGIN, origin_name,
sizeof (origin_name), NULL, NULL, 0, B_TRUE);
if (ret == 0 && strcmp(origin_name, "") != 0) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Keys cannot be changed on clones."));
ret = EINVAL;
goto error;
}
/*
* If the user wants to use the inheritkey variant of this function
* we don't need to collect any crypto arguments.
*/
if (!inheritkey) {
/* validate the provided properties */
ret = zfs_crypto_verify_rewrap_nvlist(zhp, raw_props, &props,
errbuf);
if (ret != 0)
goto error;
/*
* Load keyformat and keylocation from the nvlist. Fetch from
* the dataset properties if not specified.
*/
(void) nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_KEYFORMAT), &keyformat);
(void) nvlist_lookup_string(props,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION), &keylocation);
if (is_encroot) {
/*
* If this is already an encryption root, just keep
* any properties not set by the user.
*/
if (keyformat == ZFS_KEYFORMAT_NONE) {
keyformat = zfs_prop_get_int(zhp,
ZFS_PROP_KEYFORMAT);
ret = nvlist_add_uint64(props,
zfs_prop_to_name(ZFS_PROP_KEYFORMAT),
keyformat);
if (ret != 0) {
zfs_error_aux(zhp->zfs_hdl,
dgettext(TEXT_DOMAIN, "Failed to "
"get existing keyformat "
"property."));
goto error;
}
}
if (keylocation == NULL) {
ret = zfs_prop_get(zhp, ZFS_PROP_KEYLOCATION,
prop_keylocation, sizeof (prop_keylocation),
NULL, NULL, 0, B_TRUE);
if (ret != 0) {
zfs_error_aux(zhp->zfs_hdl,
dgettext(TEXT_DOMAIN, "Failed to "
"get existing keylocation "
"property."));
goto error;
}
keylocation = prop_keylocation;
}
} else {
/* need a new key for non-encryption roots */
if (keyformat == ZFS_KEYFORMAT_NONE) {
ret = EINVAL;
zfs_error_aux(zhp->zfs_hdl,
dgettext(TEXT_DOMAIN, "Keyformat required "
"for new encryption root."));
goto error;
}
/* default to prompt if no keylocation is specified */
if (keylocation == NULL) {
keylocation = "prompt";
ret = nvlist_add_string(props,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION),
keylocation);
if (ret != 0)
goto error;
}
}
/* fetch the new wrapping key and associated properties */
ret = populate_create_encryption_params_nvlists(zhp->zfs_hdl,
zhp, B_TRUE, keyformat, keylocation, props, &wkeydata,
&wkeylen);
if (ret != 0)
goto error;
} else {
/* check that zhp is an encryption root */
if (!is_encroot) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Key inheritting can only be performed on "
"encryption roots."));
ret = EINVAL;
goto error;
}
/* get the parent's name */
ret = zfs_parent_name(zhp, parent_name, sizeof (parent_name));
if (ret != 0) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Root dataset cannot inherit key."));
ret = EINVAL;
goto error;
}
/* get a handle to the parent */
pzhp = make_dataset_handle(zhp->zfs_hdl, parent_name);
if (pzhp == NULL) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Failed to lookup parent."));
ret = ENOENT;
goto error;
}
/* parent must be encrypted */
pcrypt = zfs_prop_get_int(pzhp, ZFS_PROP_ENCRYPTION);
if (pcrypt == ZIO_CRYPT_OFF) {
zfs_error_aux(pzhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Parent must be encrypted."));
ret = EINVAL;
goto error;
}
/* check that the parent's key is loaded */
pkeystatus = zfs_prop_get_int(pzhp, ZFS_PROP_KEYSTATUS);
if (pkeystatus == ZFS_KEYSTATUS_UNAVAILABLE) {
zfs_error_aux(pzhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Parent key must be loaded."));
ret = EACCES;
goto error;
}
}
/* check that the key is loaded */
keystatus = zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS);
if (keystatus == ZFS_KEYSTATUS_UNAVAILABLE) {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Key must be loaded."));
ret = EACCES;
goto error;
}
/* call the ioctl */
ret = lzc_change_key(zhp->zfs_name, cmd, props, wkeydata, wkeylen);
if (ret != 0) {
switch (ret) {
case EPERM:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Permission denied."));
break;
case EINVAL:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Invalid properties for key change."));
break;
case EACCES:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"Key is not currently loaded."));
break;
}
zfs_error(zhp->zfs_hdl, EZFS_CRYPTOFAILED, errbuf);
}
if (pzhp != NULL)
zfs_close(pzhp);
if (props != NULL)
nvlist_free(props);
if (wkeydata != NULL)
free(wkeydata);
return (ret);
error:
if (pzhp != NULL)
zfs_close(pzhp);
if (props != NULL)
nvlist_free(props);
if (wkeydata != NULL)
free(wkeydata);
zfs_error(zhp->zfs_hdl, EZFS_CRYPTOFAILED, errbuf);
return (ret);
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_dataset.c b/sys/contrib/openzfs/lib/libzfs/libzfs_dataset.c
index 8fa36fa95a17..138eca19acc3 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_dataset.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_dataset.c
@@ -1,5634 +1,5634 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2019 Joyent, Inc.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2012 DEY Storage Systems, Inc. All rights reserved.
* Copyright (c) 2012 Pawel Jakub Dawidek <pawel@dawidek.net>.
* Copyright (c) 2013 Martin Matuska. All rights reserved.
* Copyright (c) 2013 Steven Hartland. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>
* Copyright 2017-2018 RackTop Systems.
* Copyright (c) 2019 Datto Inc.
* Copyright (c) 2019, loli10K <ezomori.nozomu@gmail.com>
* Copyright (c) 2021 Matt Fiddaman
*/
#include <ctype.h>
#include <errno.h>
#include <libintl.h>
#include <stdio.h>
#include <stdlib.h>
#include <strings.h>
#include <unistd.h>
#include <stddef.h>
#include <zone.h>
#include <fcntl.h>
#include <sys/mntent.h>
#include <sys/mount.h>
#include <pwd.h>
#include <grp.h>
#ifdef HAVE_IDMAP
#include <idmap.h>
#include <aclutils.h>
#include <directory.h>
#endif /* HAVE_IDMAP */
#include <sys/dnode.h>
#include <sys/spa.h>
#include <sys/zap.h>
#include <sys/dsl_crypt.h>
#include <libzfs.h>
#include <libzutil.h>
#include "zfs_namecheck.h"
#include "zfs_prop.h"
#include "libzfs_impl.h"
#include "zfs_deleg.h"
static int userquota_propname_decode(const char *propname, boolean_t zoned,
zfs_userquota_prop_t *typep, char *domain, int domainlen, uint64_t *ridp);
/*
* Given a single type (not a mask of types), return the type in a human
* readable form.
*/
const char *
zfs_type_to_name(zfs_type_t type)
{
switch (type) {
case ZFS_TYPE_FILESYSTEM:
return (dgettext(TEXT_DOMAIN, "filesystem"));
case ZFS_TYPE_SNAPSHOT:
return (dgettext(TEXT_DOMAIN, "snapshot"));
case ZFS_TYPE_VOLUME:
return (dgettext(TEXT_DOMAIN, "volume"));
case ZFS_TYPE_POOL:
return (dgettext(TEXT_DOMAIN, "pool"));
case ZFS_TYPE_BOOKMARK:
return (dgettext(TEXT_DOMAIN, "bookmark"));
default:
assert(!"unhandled zfs_type_t");
}
return (NULL);
}
/*
* Validate a ZFS path. This is used even before trying to open the dataset, to
* provide a more meaningful error message. We call zfs_error_aux() to
* explain exactly why the name was not valid.
*/
int
zfs_validate_name(libzfs_handle_t *hdl, const char *path, int type,
boolean_t modifying)
{
namecheck_err_t why;
char what;
if (!(type & ZFS_TYPE_SNAPSHOT) && strchr(path, '@') != NULL) {
if (hdl != NULL)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"snapshot delimiter '@' is not expected here"));
return (0);
}
if (type == ZFS_TYPE_SNAPSHOT && strchr(path, '@') == NULL) {
if (hdl != NULL)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"missing '@' delimiter in snapshot name"));
return (0);
}
if (!(type & ZFS_TYPE_BOOKMARK) && strchr(path, '#') != NULL) {
if (hdl != NULL)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"bookmark delimiter '#' is not expected here"));
return (0);
}
if (type == ZFS_TYPE_BOOKMARK && strchr(path, '#') == NULL) {
if (hdl != NULL)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"missing '#' delimiter in bookmark name"));
return (0);
}
if (modifying && strchr(path, '%') != NULL) {
if (hdl != NULL)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid character %c in name"), '%');
return (0);
}
if (entity_namecheck(path, &why, &what) != 0) {
if (hdl != NULL) {
switch (why) {
case NAME_ERR_TOOLONG:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"name is too long"));
break;
case NAME_ERR_LEADING_SLASH:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"leading slash in name"));
break;
case NAME_ERR_EMPTY_COMPONENT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"empty component or misplaced '@'"
" or '#' delimiter in name"));
break;
case NAME_ERR_TRAILING_SLASH:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"trailing slash in name"));
break;
case NAME_ERR_INVALCHAR:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "invalid character "
"'%c' in name"), what);
break;
case NAME_ERR_MULTIPLE_DELIMITERS:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"multiple '@' and/or '#' delimiters in "
"name"));
break;
case NAME_ERR_NOLETTER:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool doesn't begin with a letter"));
break;
case NAME_ERR_RESERVED:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"name is reserved"));
break;
case NAME_ERR_DISKLIKE:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"reserved disk name"));
break;
case NAME_ERR_SELF_REF:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"self reference, '.' is found in name"));
break;
case NAME_ERR_PARENT_REF:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"parent reference, '..' is found in name"));
break;
default:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"(%d) not defined"), why);
break;
}
}
return (0);
}
return (-1);
}
int
zfs_name_valid(const char *name, zfs_type_t type)
{
if (type == ZFS_TYPE_POOL)
return (zpool_name_valid(NULL, B_FALSE, name));
return (zfs_validate_name(NULL, name, type, B_FALSE));
}
/*
* This function takes the raw DSL properties, and filters out the user-defined
* properties into a separate nvlist.
*/
static nvlist_t *
process_user_props(zfs_handle_t *zhp, nvlist_t *props)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
nvpair_t *elem;
nvlist_t *nvl;
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) {
(void) no_memory(hdl);
return (NULL);
}
elem = NULL;
while ((elem = nvlist_next_nvpair(props, elem)) != NULL) {
if (!zfs_prop_user(nvpair_name(elem)))
continue;
nvlist_t *propval = fnvpair_value_nvlist(elem);
if (nvlist_add_nvlist(nvl, nvpair_name(elem), propval) != 0) {
nvlist_free(nvl);
(void) no_memory(hdl);
return (NULL);
}
}
return (nvl);
}
static zpool_handle_t *
zpool_add_handle(zfs_handle_t *zhp, const char *pool_name)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
zpool_handle_t *zph;
if ((zph = zpool_open_canfail(hdl, pool_name)) != NULL) {
if (hdl->libzfs_pool_handles != NULL)
zph->zpool_next = hdl->libzfs_pool_handles;
hdl->libzfs_pool_handles = zph;
}
return (zph);
}
static zpool_handle_t *
zpool_find_handle(zfs_handle_t *zhp, const char *pool_name, int len)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
zpool_handle_t *zph = hdl->libzfs_pool_handles;
while ((zph != NULL) &&
(strncmp(pool_name, zpool_get_name(zph), len) != 0))
zph = zph->zpool_next;
return (zph);
}
/*
* Returns a handle to the pool that contains the provided dataset.
* If a handle to that pool already exists then that handle is returned.
* Otherwise, a new handle is created and added to the list of handles.
*/
static zpool_handle_t *
zpool_handle(zfs_handle_t *zhp)
{
char *pool_name;
int len;
zpool_handle_t *zph;
len = strcspn(zhp->zfs_name, "/@#") + 1;
pool_name = zfs_alloc(zhp->zfs_hdl, len);
(void) strlcpy(pool_name, zhp->zfs_name, len);
zph = zpool_find_handle(zhp, pool_name, len);
if (zph == NULL)
zph = zpool_add_handle(zhp, pool_name);
free(pool_name);
return (zph);
}
void
zpool_free_handles(libzfs_handle_t *hdl)
{
zpool_handle_t *next, *zph = hdl->libzfs_pool_handles;
while (zph != NULL) {
next = zph->zpool_next;
zpool_close(zph);
zph = next;
}
hdl->libzfs_pool_handles = NULL;
}
/*
* Utility function to gather stats (objset and zpl) for the given object.
*/
static int
get_stats_ioctl(zfs_handle_t *zhp, zfs_cmd_t *zc)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
(void) strlcpy(zc->zc_name, zhp->zfs_name, sizeof (zc->zc_name));
while (zfs_ioctl(hdl, ZFS_IOC_OBJSET_STATS, zc) != 0) {
if (errno == ENOMEM)
zcmd_expand_dst_nvlist(hdl, zc);
else
return (-1);
}
return (0);
}
/*
* Utility function to get the received properties of the given object.
*/
static int
get_recvd_props_ioctl(zfs_handle_t *zhp)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
nvlist_t *recvdprops;
zfs_cmd_t zc = {"\0"};
int err;
zcmd_alloc_dst_nvlist(hdl, &zc, 0);
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
while (zfs_ioctl(hdl, ZFS_IOC_OBJSET_RECVD_PROPS, &zc) != 0) {
if (errno == ENOMEM)
zcmd_expand_dst_nvlist(hdl, &zc);
else {
zcmd_free_nvlists(&zc);
return (-1);
}
}
err = zcmd_read_dst_nvlist(zhp->zfs_hdl, &zc, &recvdprops);
zcmd_free_nvlists(&zc);
if (err != 0)
return (-1);
nvlist_free(zhp->zfs_recvd_props);
zhp->zfs_recvd_props = recvdprops;
return (0);
}
static int
put_stats_zhdl(zfs_handle_t *zhp, zfs_cmd_t *zc)
{
nvlist_t *allprops, *userprops;
zhp->zfs_dmustats = zc->zc_objset_stats; /* structure assignment */
if (zcmd_read_dst_nvlist(zhp->zfs_hdl, zc, &allprops) != 0) {
return (-1);
}
/*
* XXX Why do we store the user props separately, in addition to
* storing them in zfs_props?
*/
if ((userprops = process_user_props(zhp, allprops)) == NULL) {
nvlist_free(allprops);
return (-1);
}
nvlist_free(zhp->zfs_props);
nvlist_free(zhp->zfs_user_props);
zhp->zfs_props = allprops;
zhp->zfs_user_props = userprops;
return (0);
}
static int
get_stats(zfs_handle_t *zhp)
{
int rc = 0;
zfs_cmd_t zc = {"\0"};
zcmd_alloc_dst_nvlist(zhp->zfs_hdl, &zc, 0);
if (get_stats_ioctl(zhp, &zc) != 0)
rc = -1;
else if (put_stats_zhdl(zhp, &zc) != 0)
rc = -1;
zcmd_free_nvlists(&zc);
return (rc);
}
/*
* Refresh the properties currently stored in the handle.
*/
void
zfs_refresh_properties(zfs_handle_t *zhp)
{
(void) get_stats(zhp);
}
/*
* Makes a handle from the given dataset name. Used by zfs_open() and
* zfs_iter_* to create child handles on the fly.
*/
static int
make_dataset_handle_common(zfs_handle_t *zhp, zfs_cmd_t *zc)
{
if (put_stats_zhdl(zhp, zc) != 0)
return (-1);
/*
* We've managed to open the dataset and gather statistics. Determine
* the high-level type.
*/
if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL) {
zhp->zfs_head_type = ZFS_TYPE_VOLUME;
} else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZFS) {
zhp->zfs_head_type = ZFS_TYPE_FILESYSTEM;
} else if (zhp->zfs_dmustats.dds_type == DMU_OST_OTHER) {
errno = EINVAL;
return (-1);
} else if (zhp->zfs_dmustats.dds_inconsistent) {
errno = EBUSY;
return (-1);
} else {
abort();
}
if (zhp->zfs_dmustats.dds_is_snapshot)
zhp->zfs_type = ZFS_TYPE_SNAPSHOT;
else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL)
zhp->zfs_type = ZFS_TYPE_VOLUME;
else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZFS)
zhp->zfs_type = ZFS_TYPE_FILESYSTEM;
else
abort(); /* we should never see any other types */
if ((zhp->zpool_hdl = zpool_handle(zhp)) == NULL)
return (-1);
return (0);
}
zfs_handle_t *
make_dataset_handle(libzfs_handle_t *hdl, const char *path)
{
zfs_cmd_t zc = {"\0"};
zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t));
if (zhp == NULL)
return (NULL);
zhp->zfs_hdl = hdl;
(void) strlcpy(zhp->zfs_name, path, sizeof (zhp->zfs_name));
zcmd_alloc_dst_nvlist(hdl, &zc, 0);
if (get_stats_ioctl(zhp, &zc) == -1) {
zcmd_free_nvlists(&zc);
free(zhp);
return (NULL);
}
if (make_dataset_handle_common(zhp, &zc) == -1) {
free(zhp);
zhp = NULL;
}
zcmd_free_nvlists(&zc);
return (zhp);
}
zfs_handle_t *
make_dataset_handle_zc(libzfs_handle_t *hdl, zfs_cmd_t *zc)
{
zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t));
if (zhp == NULL)
return (NULL);
zhp->zfs_hdl = hdl;
(void) strlcpy(zhp->zfs_name, zc->zc_name, sizeof (zhp->zfs_name));
if (make_dataset_handle_common(zhp, zc) == -1) {
free(zhp);
return (NULL);
}
return (zhp);
}
zfs_handle_t *
make_dataset_simple_handle_zc(zfs_handle_t *pzhp, zfs_cmd_t *zc)
{
zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t));
if (zhp == NULL)
return (NULL);
zhp->zfs_hdl = pzhp->zfs_hdl;
(void) strlcpy(zhp->zfs_name, zc->zc_name, sizeof (zhp->zfs_name));
zhp->zfs_head_type = pzhp->zfs_type;
zhp->zfs_type = ZFS_TYPE_SNAPSHOT;
zhp->zpool_hdl = zpool_handle(zhp);
if (zc->zc_objset_stats.dds_creation_txg != 0) {
/* structure assignment */
zhp->zfs_dmustats = zc->zc_objset_stats;
} else {
if (get_stats_ioctl(zhp, zc) == -1) {
zcmd_free_nvlists(zc);
free(zhp);
return (NULL);
}
if (make_dataset_handle_common(zhp, zc) == -1) {
zcmd_free_nvlists(zc);
free(zhp);
return (NULL);
}
}
if (zhp->zfs_dmustats.dds_is_snapshot ||
strchr(zc->zc_name, '@') != NULL)
zhp->zfs_type = ZFS_TYPE_SNAPSHOT;
else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL)
zhp->zfs_type = ZFS_TYPE_VOLUME;
else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZFS)
zhp->zfs_type = ZFS_TYPE_FILESYSTEM;
return (zhp);
}
zfs_handle_t *
zfs_handle_dup(zfs_handle_t *zhp_orig)
{
zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t));
if (zhp == NULL)
return (NULL);
zhp->zfs_hdl = zhp_orig->zfs_hdl;
zhp->zpool_hdl = zhp_orig->zpool_hdl;
(void) strlcpy(zhp->zfs_name, zhp_orig->zfs_name,
sizeof (zhp->zfs_name));
zhp->zfs_type = zhp_orig->zfs_type;
zhp->zfs_head_type = zhp_orig->zfs_head_type;
zhp->zfs_dmustats = zhp_orig->zfs_dmustats;
if (zhp_orig->zfs_props != NULL) {
if (nvlist_dup(zhp_orig->zfs_props, &zhp->zfs_props, 0) != 0) {
(void) no_memory(zhp->zfs_hdl);
zfs_close(zhp);
return (NULL);
}
}
if (zhp_orig->zfs_user_props != NULL) {
if (nvlist_dup(zhp_orig->zfs_user_props,
&zhp->zfs_user_props, 0) != 0) {
(void) no_memory(zhp->zfs_hdl);
zfs_close(zhp);
return (NULL);
}
}
if (zhp_orig->zfs_recvd_props != NULL) {
if (nvlist_dup(zhp_orig->zfs_recvd_props,
&zhp->zfs_recvd_props, 0)) {
(void) no_memory(zhp->zfs_hdl);
zfs_close(zhp);
return (NULL);
}
}
zhp->zfs_mntcheck = zhp_orig->zfs_mntcheck;
if (zhp_orig->zfs_mntopts != NULL) {
zhp->zfs_mntopts = zfs_strdup(zhp_orig->zfs_hdl,
zhp_orig->zfs_mntopts);
}
zhp->zfs_props_table = zhp_orig->zfs_props_table;
return (zhp);
}
boolean_t
zfs_bookmark_exists(const char *path)
{
nvlist_t *bmarks;
nvlist_t *props;
char fsname[ZFS_MAX_DATASET_NAME_LEN];
char *bmark_name;
char *pound;
int err;
boolean_t rv;
(void) strlcpy(fsname, path, sizeof (fsname));
pound = strchr(fsname, '#');
if (pound == NULL)
return (B_FALSE);
*pound = '\0';
bmark_name = pound + 1;
props = fnvlist_alloc();
err = lzc_get_bookmarks(fsname, props, &bmarks);
nvlist_free(props);
if (err != 0) {
nvlist_free(bmarks);
return (B_FALSE);
}
rv = nvlist_exists(bmarks, bmark_name);
nvlist_free(bmarks);
return (rv);
}
zfs_handle_t *
make_bookmark_handle(zfs_handle_t *parent, const char *path,
nvlist_t *bmark_props)
{
zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t));
if (zhp == NULL)
return (NULL);
/* Fill in the name. */
zhp->zfs_hdl = parent->zfs_hdl;
(void) strlcpy(zhp->zfs_name, path, sizeof (zhp->zfs_name));
/* Set the property lists. */
if (nvlist_dup(bmark_props, &zhp->zfs_props, 0) != 0) {
free(zhp);
return (NULL);
}
/* Set the types. */
zhp->zfs_head_type = parent->zfs_head_type;
zhp->zfs_type = ZFS_TYPE_BOOKMARK;
if ((zhp->zpool_hdl = zpool_handle(zhp)) == NULL) {
nvlist_free(zhp->zfs_props);
free(zhp);
return (NULL);
}
return (zhp);
}
struct zfs_open_bookmarks_cb_data {
const char *path;
zfs_handle_t *zhp;
};
static int
zfs_open_bookmarks_cb(zfs_handle_t *zhp, void *data)
{
struct zfs_open_bookmarks_cb_data *dp = data;
/*
* Is it the one we are looking for?
*/
if (strcmp(dp->path, zfs_get_name(zhp)) == 0) {
/*
* We found it. Save it and let the caller know we are done.
*/
dp->zhp = zhp;
return (EEXIST);
}
/*
* Not found. Close the handle and ask for another one.
*/
zfs_close(zhp);
return (0);
}
/*
* Opens the given snapshot, bookmark, filesystem, or volume. The 'types'
* argument is a mask of acceptable types. The function will print an
* appropriate error message and return NULL if it can't be opened.
*/
zfs_handle_t *
zfs_open(libzfs_handle_t *hdl, const char *path, int types)
{
zfs_handle_t *zhp;
char errbuf[ERRBUFLEN];
char *bookp;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot open '%s'"), path);
/*
* Validate the name before we even try to open it.
*/
if (!zfs_validate_name(hdl, path, types, B_FALSE)) {
(void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
errno = EINVAL;
return (NULL);
}
/*
* Bookmarks needs to be handled separately.
*/
bookp = strchr(path, '#');
if (bookp == NULL) {
/*
* Try to get stats for the dataset, which will tell us if it
* exists.
*/
errno = 0;
if ((zhp = make_dataset_handle(hdl, path)) == NULL) {
(void) zfs_standard_error(hdl, errno, errbuf);
return (NULL);
}
} else {
char dsname[ZFS_MAX_DATASET_NAME_LEN];
zfs_handle_t *pzhp;
struct zfs_open_bookmarks_cb_data cb_data = {path, NULL};
/*
* We need to cut out '#' and everything after '#'
* to get the parent dataset name only.
*/
assert(bookp - path < sizeof (dsname));
(void) strlcpy(dsname, path,
MIN(sizeof (dsname), bookp - path + 1));
/*
* Create handle for the parent dataset.
*/
errno = 0;
if ((pzhp = make_dataset_handle(hdl, dsname)) == NULL) {
(void) zfs_standard_error(hdl, errno, errbuf);
return (NULL);
}
/*
* Iterate bookmarks to find the right one.
*/
errno = 0;
- if ((zfs_iter_bookmarks(pzhp, 0, zfs_open_bookmarks_cb,
+ if ((zfs_iter_bookmarks_v2(pzhp, 0, zfs_open_bookmarks_cb,
&cb_data) == 0) && (cb_data.zhp == NULL)) {
(void) zfs_error(hdl, EZFS_NOENT, errbuf);
zfs_close(pzhp);
errno = ENOENT;
return (NULL);
}
if (cb_data.zhp == NULL) {
(void) zfs_standard_error(hdl, errno, errbuf);
zfs_close(pzhp);
return (NULL);
}
zhp = cb_data.zhp;
/*
* Cleanup.
*/
zfs_close(pzhp);
}
if (!(types & zhp->zfs_type)) {
(void) zfs_error(hdl, EZFS_BADTYPE, errbuf);
zfs_close(zhp);
errno = EINVAL;
return (NULL);
}
return (zhp);
}
/*
* Release a ZFS handle. Nothing to do but free the associated memory.
*/
void
zfs_close(zfs_handle_t *zhp)
{
if (zhp->zfs_mntopts)
free(zhp->zfs_mntopts);
nvlist_free(zhp->zfs_props);
nvlist_free(zhp->zfs_user_props);
nvlist_free(zhp->zfs_recvd_props);
free(zhp);
}
typedef struct mnttab_node {
struct mnttab mtn_mt;
avl_node_t mtn_node;
} mnttab_node_t;
static int
libzfs_mnttab_cache_compare(const void *arg1, const void *arg2)
{
const mnttab_node_t *mtn1 = (const mnttab_node_t *)arg1;
const mnttab_node_t *mtn2 = (const mnttab_node_t *)arg2;
int rv;
rv = strcmp(mtn1->mtn_mt.mnt_special, mtn2->mtn_mt.mnt_special);
return (TREE_ISIGN(rv));
}
void
libzfs_mnttab_init(libzfs_handle_t *hdl)
{
pthread_mutex_init(&hdl->libzfs_mnttab_cache_lock, NULL);
assert(avl_numnodes(&hdl->libzfs_mnttab_cache) == 0);
avl_create(&hdl->libzfs_mnttab_cache, libzfs_mnttab_cache_compare,
sizeof (mnttab_node_t), offsetof(mnttab_node_t, mtn_node));
}
static int
libzfs_mnttab_update(libzfs_handle_t *hdl)
{
FILE *mnttab;
struct mnttab entry;
if ((mnttab = fopen(MNTTAB, "re")) == NULL)
return (ENOENT);
while (getmntent(mnttab, &entry) == 0) {
mnttab_node_t *mtn;
avl_index_t where;
if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0)
continue;
mtn = zfs_alloc(hdl, sizeof (mnttab_node_t));
mtn->mtn_mt.mnt_special = zfs_strdup(hdl, entry.mnt_special);
mtn->mtn_mt.mnt_mountp = zfs_strdup(hdl, entry.mnt_mountp);
mtn->mtn_mt.mnt_fstype = zfs_strdup(hdl, entry.mnt_fstype);
mtn->mtn_mt.mnt_mntopts = zfs_strdup(hdl, entry.mnt_mntopts);
/* Exclude duplicate mounts */
if (avl_find(&hdl->libzfs_mnttab_cache, mtn, &where) != NULL) {
free(mtn->mtn_mt.mnt_special);
free(mtn->mtn_mt.mnt_mountp);
free(mtn->mtn_mt.mnt_fstype);
free(mtn->mtn_mt.mnt_mntopts);
free(mtn);
continue;
}
avl_add(&hdl->libzfs_mnttab_cache, mtn);
}
(void) fclose(mnttab);
return (0);
}
void
libzfs_mnttab_fini(libzfs_handle_t *hdl)
{
void *cookie = NULL;
mnttab_node_t *mtn;
while ((mtn = avl_destroy_nodes(&hdl->libzfs_mnttab_cache, &cookie))
!= NULL) {
free(mtn->mtn_mt.mnt_special);
free(mtn->mtn_mt.mnt_mountp);
free(mtn->mtn_mt.mnt_fstype);
free(mtn->mtn_mt.mnt_mntopts);
free(mtn);
}
avl_destroy(&hdl->libzfs_mnttab_cache);
(void) pthread_mutex_destroy(&hdl->libzfs_mnttab_cache_lock);
}
void
libzfs_mnttab_cache(libzfs_handle_t *hdl, boolean_t enable)
{
hdl->libzfs_mnttab_enable = enable;
}
int
libzfs_mnttab_find(libzfs_handle_t *hdl, const char *fsname,
struct mnttab *entry)
{
FILE *mnttab;
mnttab_node_t find;
mnttab_node_t *mtn;
int ret = ENOENT;
if (!hdl->libzfs_mnttab_enable) {
struct mnttab srch = { 0 };
if (avl_numnodes(&hdl->libzfs_mnttab_cache))
libzfs_mnttab_fini(hdl);
if ((mnttab = fopen(MNTTAB, "re")) == NULL)
return (ENOENT);
srch.mnt_special = (char *)fsname;
srch.mnt_fstype = (char *)MNTTYPE_ZFS;
ret = getmntany(mnttab, entry, &srch) ? ENOENT : 0;
(void) fclose(mnttab);
return (ret);
}
pthread_mutex_lock(&hdl->libzfs_mnttab_cache_lock);
if (avl_numnodes(&hdl->libzfs_mnttab_cache) == 0) {
int error;
if ((error = libzfs_mnttab_update(hdl)) != 0) {
pthread_mutex_unlock(&hdl->libzfs_mnttab_cache_lock);
return (error);
}
}
find.mtn_mt.mnt_special = (char *)fsname;
mtn = avl_find(&hdl->libzfs_mnttab_cache, &find, NULL);
if (mtn) {
*entry = mtn->mtn_mt;
ret = 0;
}
pthread_mutex_unlock(&hdl->libzfs_mnttab_cache_lock);
return (ret);
}
void
libzfs_mnttab_add(libzfs_handle_t *hdl, const char *special,
const char *mountp, const char *mntopts)
{
mnttab_node_t *mtn;
pthread_mutex_lock(&hdl->libzfs_mnttab_cache_lock);
if (avl_numnodes(&hdl->libzfs_mnttab_cache) != 0) {
mtn = zfs_alloc(hdl, sizeof (mnttab_node_t));
mtn->mtn_mt.mnt_special = zfs_strdup(hdl, special);
mtn->mtn_mt.mnt_mountp = zfs_strdup(hdl, mountp);
mtn->mtn_mt.mnt_fstype = zfs_strdup(hdl, MNTTYPE_ZFS);
mtn->mtn_mt.mnt_mntopts = zfs_strdup(hdl, mntopts);
/*
* Another thread may have already added this entry
* via libzfs_mnttab_update. If so we should skip it.
*/
if (avl_find(&hdl->libzfs_mnttab_cache, mtn, NULL) != NULL) {
free(mtn->mtn_mt.mnt_special);
free(mtn->mtn_mt.mnt_mountp);
free(mtn->mtn_mt.mnt_fstype);
free(mtn->mtn_mt.mnt_mntopts);
free(mtn);
} else {
avl_add(&hdl->libzfs_mnttab_cache, mtn);
}
}
pthread_mutex_unlock(&hdl->libzfs_mnttab_cache_lock);
}
void
libzfs_mnttab_remove(libzfs_handle_t *hdl, const char *fsname)
{
mnttab_node_t find;
mnttab_node_t *ret;
pthread_mutex_lock(&hdl->libzfs_mnttab_cache_lock);
find.mtn_mt.mnt_special = (char *)fsname;
if ((ret = avl_find(&hdl->libzfs_mnttab_cache, (void *)&find, NULL))
!= NULL) {
avl_remove(&hdl->libzfs_mnttab_cache, ret);
free(ret->mtn_mt.mnt_special);
free(ret->mtn_mt.mnt_mountp);
free(ret->mtn_mt.mnt_fstype);
free(ret->mtn_mt.mnt_mntopts);
free(ret);
}
pthread_mutex_unlock(&hdl->libzfs_mnttab_cache_lock);
}
int
zfs_spa_version(zfs_handle_t *zhp, int *spa_version)
{
zpool_handle_t *zpool_handle = zhp->zpool_hdl;
if (zpool_handle == NULL)
return (-1);
*spa_version = zpool_get_prop_int(zpool_handle,
ZPOOL_PROP_VERSION, NULL);
return (0);
}
/*
* The choice of reservation property depends on the SPA version.
*/
static int
zfs_which_resv_prop(zfs_handle_t *zhp, zfs_prop_t *resv_prop)
{
int spa_version;
if (zfs_spa_version(zhp, &spa_version) < 0)
return (-1);
if (spa_version >= SPA_VERSION_REFRESERVATION)
*resv_prop = ZFS_PROP_REFRESERVATION;
else
*resv_prop = ZFS_PROP_RESERVATION;
return (0);
}
/*
* Given an nvlist of properties to set, validates that they are correct, and
* parses any numeric properties (index, boolean, etc) if they are specified as
* strings.
*/
nvlist_t *
zfs_valid_proplist(libzfs_handle_t *hdl, zfs_type_t type, nvlist_t *nvl,
uint64_t zoned, zfs_handle_t *zhp, zpool_handle_t *zpool_hdl,
boolean_t key_params_ok, const char *errbuf)
{
nvpair_t *elem;
uint64_t intval;
const char *strval;
zfs_prop_t prop;
nvlist_t *ret;
int chosen_normal = -1;
int chosen_utf = -1;
if (nvlist_alloc(&ret, NV_UNIQUE_NAME, 0) != 0) {
(void) no_memory(hdl);
return (NULL);
}
/*
* Make sure this property is valid and applies to this type.
*/
elem = NULL;
while ((elem = nvlist_next_nvpair(nvl, elem)) != NULL) {
const char *propname = nvpair_name(elem);
prop = zfs_name_to_prop(propname);
if (prop == ZPROP_USERPROP && zfs_prop_user(propname)) {
/*
* This is a user property: make sure it's a
* string, and that it's less than ZAP_MAXNAMELEN.
*/
if (nvpair_type(elem) != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (strlen(nvpair_name(elem)) >= ZAP_MAXNAMELEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property name '%s' is too long"),
propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
(void) nvpair_value_string(elem, &strval);
if (nvlist_add_string(ret, propname, strval) != 0) {
(void) no_memory(hdl);
goto error;
}
continue;
}
/*
* Currently, only user properties can be modified on
* snapshots.
*/
if (type == ZFS_TYPE_SNAPSHOT) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"this property can not be modified for snapshots"));
(void) zfs_error(hdl, EZFS_PROPTYPE, errbuf);
goto error;
}
if (prop == ZPROP_USERPROP && zfs_prop_userquota(propname)) {
zfs_userquota_prop_t uqtype;
char *newpropname = NULL;
char domain[128];
uint64_t rid;
uint64_t valary[3];
int rc;
if (userquota_propname_decode(propname, zoned,
&uqtype, domain, sizeof (domain), &rid) != 0) {
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN,
"'%s' has an invalid user/group name"),
propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (uqtype != ZFS_PROP_USERQUOTA &&
uqtype != ZFS_PROP_GROUPQUOTA &&
uqtype != ZFS_PROP_USEROBJQUOTA &&
uqtype != ZFS_PROP_GROUPOBJQUOTA &&
uqtype != ZFS_PROP_PROJECTQUOTA &&
uqtype != ZFS_PROP_PROJECTOBJQUOTA) {
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "'%s' is readonly"),
propname);
(void) zfs_error(hdl, EZFS_PROPREADONLY,
errbuf);
goto error;
}
if (nvpair_type(elem) == DATA_TYPE_STRING) {
(void) nvpair_value_string(elem, &strval);
if (strcmp(strval, "none") == 0) {
intval = 0;
} else if (zfs_nicestrtonum(hdl,
strval, &intval) != 0) {
(void) zfs_error(hdl,
EZFS_BADPROP, errbuf);
goto error;
}
} else if (nvpair_type(elem) ==
DATA_TYPE_UINT64) {
(void) nvpair_value_uint64(elem, &intval);
if (intval == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"use 'none' to disable "
"{user|group|project}quota"));
goto error;
}
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a number"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
/*
* Encode the prop name as
* userquota@<hex-rid>-domain, to make it easy
* for the kernel to decode.
*/
rc = asprintf(&newpropname, "%s%llx-%s",
zfs_userquota_prop_prefixes[uqtype],
(longlong_t)rid, domain);
if (rc == -1 || newpropname == NULL) {
(void) no_memory(hdl);
goto error;
}
valary[0] = uqtype;
valary[1] = rid;
valary[2] = intval;
if (nvlist_add_uint64_array(ret, newpropname,
valary, 3) != 0) {
free(newpropname);
(void) no_memory(hdl);
goto error;
}
free(newpropname);
continue;
} else if (prop == ZPROP_USERPROP &&
zfs_prop_written(propname)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is readonly"),
propname);
(void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf);
goto error;
}
if (prop == ZPROP_INVAL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property '%s'"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (!zfs_prop_valid_for_type(prop, type, B_FALSE)) {
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "'%s' does not "
"apply to datasets of this type"), propname);
(void) zfs_error(hdl, EZFS_PROPTYPE, errbuf);
goto error;
}
if (zfs_prop_readonly(prop) &&
!(zfs_prop_setonce(prop) && zhp == NULL) &&
!(zfs_prop_encryption_key_param(prop) && key_params_ok)) {
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "'%s' is readonly"),
propname);
(void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf);
goto error;
}
if (zprop_parse_value(hdl, elem, prop, type, ret,
&strval, &intval, errbuf) != 0)
goto error;
/*
* Perform some additional checks for specific properties.
*/
switch (prop) {
case ZFS_PROP_VERSION:
{
int version;
if (zhp == NULL)
break;
version = zfs_prop_get_int(zhp, ZFS_PROP_VERSION);
if (intval < version) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Can not downgrade; already at version %u"),
version);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
}
case ZFS_PROP_VOLBLOCKSIZE:
case ZFS_PROP_RECORDSIZE:
{
int maxbs = SPA_MAXBLOCKSIZE;
char buf[64];
if (zpool_hdl != NULL) {
maxbs = zpool_get_prop_int(zpool_hdl,
ZPOOL_PROP_MAXBLOCKSIZE, NULL);
}
/*
* The value must be a power of two between
* SPA_MINBLOCKSIZE and maxbs.
*/
if (intval < SPA_MINBLOCKSIZE ||
intval > maxbs || !ISP2(intval)) {
zfs_nicebytes(maxbs, buf, sizeof (buf));
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be power of 2 from 512B "
"to %s"), propname, buf);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
}
case ZFS_PROP_SPECIAL_SMALL_BLOCKS:
{
int maxbs = SPA_OLD_MAXBLOCKSIZE;
char buf[64];
if (zpool_hdl != NULL) {
char state[64] = "";
maxbs = zpool_get_prop_int(zpool_hdl,
ZPOOL_PROP_MAXBLOCKSIZE, NULL);
/*
* Issue a warning but do not fail so that
* tests for settable properties succeed.
*/
if (zpool_prop_get_feature(zpool_hdl,
"feature@allocation_classes", state,
sizeof (state)) != 0 ||
strcmp(state, ZFS_FEATURE_ACTIVE) != 0) {
(void) fprintf(stderr, gettext(
"%s: property requires a special "
"device in the pool\n"), propname);
}
}
if (intval != 0 &&
(intval < SPA_MINBLOCKSIZE ||
intval > maxbs || !ISP2(intval))) {
zfs_nicebytes(maxbs, buf, sizeof (buf));
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid '%s=%llu' property: must be zero "
"or a power of 2 from 512B to %s"),
propname, (unsigned long long)intval, buf);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
}
case ZFS_PROP_MLSLABEL:
{
#ifdef HAVE_MLSLABEL
/*
* Verify the mlslabel string and convert to
* internal hex label string.
*/
m_label_t *new_sl;
char *hex = NULL; /* internal label string */
/* Default value is already OK. */
if (strcasecmp(strval, ZFS_MLSLABEL_DEFAULT) == 0)
break;
/* Verify the label can be converted to binary form */
if (((new_sl = m_label_alloc(MAC_LABEL)) == NULL) ||
(str_to_label(strval, &new_sl, MAC_LABEL,
L_NO_CORRECTION, NULL) == -1)) {
goto badlabel;
}
/* Now translate to hex internal label string */
if (label_to_str(new_sl, &hex, M_INTERNAL,
DEF_NAMES) != 0) {
if (hex)
free(hex);
goto badlabel;
}
m_label_free(new_sl);
/* If string is already in internal form, we're done. */
if (strcmp(strval, hex) == 0) {
free(hex);
break;
}
/* Replace the label string with the internal form. */
(void) nvlist_remove(ret, zfs_prop_to_name(prop),
DATA_TYPE_STRING);
fnvlist_add_string(ret, zfs_prop_to_name(prop), hex);
free(hex);
break;
badlabel:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid mlslabel '%s'"), strval);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
m_label_free(new_sl); /* OK if null */
goto error;
#else
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"mlslabels are unsupported"));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
#endif /* HAVE_MLSLABEL */
}
case ZFS_PROP_MOUNTPOINT:
{
namecheck_err_t why;
if (strcmp(strval, ZFS_MOUNTPOINT_NONE) == 0 ||
strcmp(strval, ZFS_MOUNTPOINT_LEGACY) == 0)
break;
if (mountpoint_namecheck(strval, &why)) {
switch (why) {
case NAME_ERR_LEADING_SLASH:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN,
"'%s' must be an absolute path, "
"'none', or 'legacy'"), propname);
break;
case NAME_ERR_TOOLONG:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN,
"component of '%s' is too long"),
propname);
break;
default:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN,
"(%d) not defined"),
why);
break;
}
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
zfs_fallthrough;
}
case ZFS_PROP_SHARESMB:
case ZFS_PROP_SHARENFS:
/*
* For the mountpoint and sharenfs or sharesmb
* properties, check if it can be set in a
* global/non-global zone based on
* the zoned property value:
*
* global zone non-global zone
* --------------------------------------------------
* zoned=on mountpoint (no) mountpoint (yes)
* sharenfs (no) sharenfs (no)
* sharesmb (no) sharesmb (no)
*
* zoned=off mountpoint (yes) N/A
* sharenfs (yes)
* sharesmb (yes)
*/
if (zoned) {
if (getzoneid() == GLOBAL_ZONEID) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' cannot be set on "
"dataset in a non-global zone"),
propname);
(void) zfs_error(hdl, EZFS_ZONED,
errbuf);
goto error;
} else if (prop == ZFS_PROP_SHARENFS ||
prop == ZFS_PROP_SHARESMB) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' cannot be set in "
"a non-global zone"), propname);
(void) zfs_error(hdl, EZFS_ZONED,
errbuf);
goto error;
}
} else if (getzoneid() != GLOBAL_ZONEID) {
/*
* If zoned property is 'off', this must be in
* a global zone. If not, something is wrong.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' cannot be set while dataset "
"'zoned' property is set"), propname);
(void) zfs_error(hdl, EZFS_ZONED, errbuf);
goto error;
}
/*
* At this point, it is legitimate to set the
* property. Now we want to make sure that the
* property value is valid if it is sharenfs.
*/
if ((prop == ZFS_PROP_SHARENFS ||
prop == ZFS_PROP_SHARESMB) &&
strcmp(strval, "on") != 0 &&
strcmp(strval, "off") != 0) {
enum sa_protocol proto;
if (prop == ZFS_PROP_SHARESMB)
proto = SA_PROTOCOL_SMB;
else
proto = SA_PROTOCOL_NFS;
if (sa_validate_shareopts(strval, proto) !=
SA_OK) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' cannot be set to invalid "
"options"), propname);
(void) zfs_error(hdl, EZFS_BADPROP,
errbuf);
goto error;
}
}
break;
case ZFS_PROP_KEYLOCATION:
if (!zfs_prop_valid_keylocation(strval, B_FALSE)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid keylocation"));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (zhp != NULL) {
uint64_t crypt =
zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION);
if (crypt == ZIO_CRYPT_OFF &&
strcmp(strval, "none") != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"keylocation must be 'none' "
"for unencrypted datasets"));
(void) zfs_error(hdl, EZFS_BADPROP,
errbuf);
goto error;
} else if (crypt != ZIO_CRYPT_OFF &&
strcmp(strval, "none") == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"keylocation must not be 'none' "
"for encrypted datasets"));
(void) zfs_error(hdl, EZFS_BADPROP,
errbuf);
goto error;
}
}
break;
case ZFS_PROP_PBKDF2_ITERS:
if (intval < MIN_PBKDF2_ITERATIONS) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"minimum pbkdf2 iterations is %u"),
MIN_PBKDF2_ITERATIONS);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZFS_PROP_UTF8ONLY:
chosen_utf = (int)intval;
break;
case ZFS_PROP_NORMALIZE:
chosen_normal = (int)intval;
break;
default:
break;
}
/*
* For changes to existing volumes, we have some additional
* checks to enforce.
*/
if (type == ZFS_TYPE_VOLUME && zhp != NULL) {
uint64_t blocksize = zfs_prop_get_int(zhp,
ZFS_PROP_VOLBLOCKSIZE);
char buf[64];
switch (prop) {
case ZFS_PROP_VOLSIZE:
if (intval % blocksize != 0) {
zfs_nicebytes(blocksize, buf,
sizeof (buf));
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a multiple of "
"volume block size (%s)"),
propname, buf);
(void) zfs_error(hdl, EZFS_BADPROP,
errbuf);
goto error;
}
if (intval == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' cannot be zero"),
propname);
(void) zfs_error(hdl, EZFS_BADPROP,
errbuf);
goto error;
}
break;
default:
break;
}
}
/* check encryption properties */
if (zhp != NULL) {
int64_t crypt = zfs_prop_get_int(zhp,
ZFS_PROP_ENCRYPTION);
switch (prop) {
case ZFS_PROP_COPIES:
if (crypt != ZIO_CRYPT_OFF && intval > 2) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"encrypted datasets cannot have "
"3 copies"));
(void) zfs_error(hdl, EZFS_BADPROP,
errbuf);
goto error;
}
break;
default:
break;
}
}
}
/*
* If normalization was chosen, but no UTF8 choice was made,
* enforce rejection of non-UTF8 names.
*
* If normalization was chosen, but rejecting non-UTF8 names
* was explicitly not chosen, it is an error.
*
* If utf8only was turned off, but the parent has normalization,
* turn off normalization.
*/
if (chosen_normal > 0 && chosen_utf < 0) {
if (nvlist_add_uint64(ret,
zfs_prop_to_name(ZFS_PROP_UTF8ONLY), 1) != 0) {
(void) no_memory(hdl);
goto error;
}
} else if (chosen_normal > 0 && chosen_utf == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be set 'on' if normalization chosen"),
zfs_prop_to_name(ZFS_PROP_UTF8ONLY));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
} else if (chosen_normal < 0 && chosen_utf == 0) {
if (nvlist_add_uint64(ret,
zfs_prop_to_name(ZFS_PROP_NORMALIZE), 0) != 0) {
(void) no_memory(hdl);
goto error;
}
}
return (ret);
error:
nvlist_free(ret);
return (NULL);
}
static int
zfs_add_synthetic_resv(zfs_handle_t *zhp, nvlist_t *nvl)
{
uint64_t old_volsize;
uint64_t new_volsize;
uint64_t old_reservation;
uint64_t new_reservation;
zfs_prop_t resv_prop;
nvlist_t *props;
zpool_handle_t *zph = zpool_handle(zhp);
/*
* If this is an existing volume, and someone is setting the volsize,
* make sure that it matches the reservation, or add it if necessary.
*/
old_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE);
if (zfs_which_resv_prop(zhp, &resv_prop) < 0)
return (-1);
old_reservation = zfs_prop_get_int(zhp, resv_prop);
props = fnvlist_alloc();
fnvlist_add_uint64(props, zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE),
zfs_prop_get_int(zhp, ZFS_PROP_VOLBLOCKSIZE));
if ((zvol_volsize_to_reservation(zph, old_volsize, props) !=
old_reservation) || nvlist_exists(nvl,
zfs_prop_to_name(resv_prop))) {
fnvlist_free(props);
return (0);
}
if (nvlist_lookup_uint64(nvl, zfs_prop_to_name(ZFS_PROP_VOLSIZE),
&new_volsize) != 0) {
fnvlist_free(props);
return (-1);
}
new_reservation = zvol_volsize_to_reservation(zph, new_volsize, props);
fnvlist_free(props);
if (nvlist_add_uint64(nvl, zfs_prop_to_name(resv_prop),
new_reservation) != 0) {
(void) no_memory(zhp->zfs_hdl);
return (-1);
}
return (1);
}
/*
* Helper for 'zfs {set|clone} refreservation=auto'. Must be called after
* zfs_valid_proplist(), as it is what sets the UINT64_MAX sentinel value.
* Return codes must match zfs_add_synthetic_resv().
*/
static int
zfs_fix_auto_resv(zfs_handle_t *zhp, nvlist_t *nvl)
{
uint64_t volsize;
uint64_t resvsize;
zfs_prop_t prop;
nvlist_t *props;
if (!ZFS_IS_VOLUME(zhp)) {
return (0);
}
if (zfs_which_resv_prop(zhp, &prop) != 0) {
return (-1);
}
if (prop != ZFS_PROP_REFRESERVATION) {
return (0);
}
if (nvlist_lookup_uint64(nvl, zfs_prop_to_name(prop), &resvsize) != 0) {
/* No value being set, so it can't be "auto" */
return (0);
}
if (resvsize != UINT64_MAX) {
/* Being set to a value other than "auto" */
return (0);
}
props = fnvlist_alloc();
fnvlist_add_uint64(props, zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE),
zfs_prop_get_int(zhp, ZFS_PROP_VOLBLOCKSIZE));
if (nvlist_lookup_uint64(nvl, zfs_prop_to_name(ZFS_PROP_VOLSIZE),
&volsize) != 0) {
volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE);
}
resvsize = zvol_volsize_to_reservation(zpool_handle(zhp), volsize,
props);
fnvlist_free(props);
(void) nvlist_remove_all(nvl, zfs_prop_to_name(prop));
if (nvlist_add_uint64(nvl, zfs_prop_to_name(prop), resvsize) != 0) {
(void) no_memory(zhp->zfs_hdl);
return (-1);
}
return (1);
}
static boolean_t
zfs_is_namespace_prop(zfs_prop_t prop)
{
switch (prop) {
case ZFS_PROP_ATIME:
case ZFS_PROP_RELATIME:
case ZFS_PROP_DEVICES:
case ZFS_PROP_EXEC:
case ZFS_PROP_SETUID:
case ZFS_PROP_READONLY:
case ZFS_PROP_XATTR:
case ZFS_PROP_NBMAND:
return (B_TRUE);
default:
return (B_FALSE);
}
}
/*
* Given a property name and value, set the property for the given dataset.
*/
int
zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval)
{
int ret = -1;
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zfs_hdl;
nvlist_t *nvl = NULL;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot set property for '%s'"),
zhp->zfs_name);
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0 ||
nvlist_add_string(nvl, propname, propval) != 0) {
(void) no_memory(hdl);
goto error;
}
ret = zfs_prop_set_list(zhp, nvl);
error:
nvlist_free(nvl);
return (ret);
}
/*
* Given an nvlist of property names and values, set the properties for the
* given dataset.
*/
int
zfs_prop_set_list(zfs_handle_t *zhp, nvlist_t *props)
{
zfs_cmd_t zc = {"\0"};
int ret = -1;
prop_changelist_t **cls = NULL;
int cl_idx;
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zfs_hdl;
nvlist_t *nvl;
int nvl_len = 0;
int added_resv = 0;
zfs_prop_t prop = 0;
nvpair_t *elem;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot set property for '%s'"),
zhp->zfs_name);
if ((nvl = zfs_valid_proplist(hdl, zhp->zfs_type, props,
zfs_prop_get_int(zhp, ZFS_PROP_ZONED), zhp, zhp->zpool_hdl,
B_FALSE, errbuf)) == NULL)
goto error;
/*
* We have to check for any extra properties which need to be added
* before computing the length of the nvlist.
*/
for (elem = nvlist_next_nvpair(nvl, NULL);
elem != NULL;
elem = nvlist_next_nvpair(nvl, elem)) {
if (zfs_name_to_prop(nvpair_name(elem)) == ZFS_PROP_VOLSIZE &&
(added_resv = zfs_add_synthetic_resv(zhp, nvl)) == -1) {
goto error;
}
}
if (added_resv != 1 &&
(added_resv = zfs_fix_auto_resv(zhp, nvl)) == -1) {
goto error;
}
/*
* Check how many properties we're setting and allocate an array to
* store changelist pointers for postfix().
*/
for (elem = nvlist_next_nvpair(nvl, NULL);
elem != NULL;
elem = nvlist_next_nvpair(nvl, elem))
nvl_len++;
if ((cls = calloc(nvl_len, sizeof (prop_changelist_t *))) == NULL)
goto error;
cl_idx = 0;
for (elem = nvlist_next_nvpair(nvl, NULL);
elem != NULL;
elem = nvlist_next_nvpair(nvl, elem)) {
prop = zfs_name_to_prop(nvpair_name(elem));
assert(cl_idx < nvl_len);
/*
* We don't want to unmount & remount the dataset when changing
* its canmount property to 'on' or 'noauto'. We only use
* the changelist logic to unmount when setting canmount=off.
*/
if (prop != ZFS_PROP_CANMOUNT ||
(fnvpair_value_uint64(elem) == ZFS_CANMOUNT_OFF &&
zfs_is_mounted(zhp, NULL))) {
cls[cl_idx] = changelist_gather(zhp, prop, 0, 0);
if (cls[cl_idx] == NULL)
goto error;
}
if (prop == ZFS_PROP_MOUNTPOINT &&
changelist_haszonedchild(cls[cl_idx])) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"child dataset with inherited mountpoint is used "
"in a non-global zone"));
ret = zfs_error(hdl, EZFS_ZONED, errbuf);
goto error;
}
if (cls[cl_idx] != NULL &&
(ret = changelist_prefix(cls[cl_idx])) != 0)
goto error;
cl_idx++;
}
assert(cl_idx == nvl_len);
/*
* Execute the corresponding ioctl() to set this list of properties.
*/
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
zcmd_write_src_nvlist(hdl, &zc, nvl);
zcmd_alloc_dst_nvlist(hdl, &zc, 0);
ret = zfs_ioctl(hdl, ZFS_IOC_SET_PROP, &zc);
if (ret != 0) {
if (zc.zc_nvlist_dst_filled == B_FALSE) {
(void) zfs_standard_error(hdl, errno, errbuf);
goto error;
}
/* Get the list of unset properties back and report them. */
nvlist_t *errorprops = NULL;
if (zcmd_read_dst_nvlist(hdl, &zc, &errorprops) != 0)
goto error;
for (nvpair_t *elem = nvlist_next_nvpair(errorprops, NULL);
elem != NULL;
elem = nvlist_next_nvpair(errorprops, elem)) {
prop = zfs_name_to_prop(nvpair_name(elem));
zfs_setprop_error(hdl, prop, errno, errbuf);
}
nvlist_free(errorprops);
if (added_resv && errno == ENOSPC) {
/* clean up the volsize property we tried to set */
uint64_t old_volsize = zfs_prop_get_int(zhp,
ZFS_PROP_VOLSIZE);
nvlist_free(nvl);
nvl = NULL;
zcmd_free_nvlists(&zc);
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0)
goto error;
if (nvlist_add_uint64(nvl,
zfs_prop_to_name(ZFS_PROP_VOLSIZE),
old_volsize) != 0)
goto error;
zcmd_write_src_nvlist(hdl, &zc, nvl);
(void) zfs_ioctl(hdl, ZFS_IOC_SET_PROP, &zc);
}
} else {
for (cl_idx = 0; cl_idx < nvl_len; cl_idx++) {
if (cls[cl_idx] != NULL) {
int clp_err = changelist_postfix(cls[cl_idx]);
if (clp_err != 0)
ret = clp_err;
}
}
if (ret == 0) {
/*
* Refresh the statistics so the new property
* value is reflected.
*/
(void) get_stats(zhp);
/*
* Remount the filesystem to propagate the change
* if one of the options handled by the generic
* Linux namespace layer has been modified.
*/
if (zfs_is_namespace_prop(prop) &&
zfs_is_mounted(zhp, NULL))
ret = zfs_mount(zhp, MNTOPT_REMOUNT, 0);
}
}
error:
nvlist_free(nvl);
zcmd_free_nvlists(&zc);
if (cls != NULL) {
for (cl_idx = 0; cl_idx < nvl_len; cl_idx++) {
if (cls[cl_idx] != NULL)
changelist_free(cls[cl_idx]);
}
free(cls);
}
return (ret);
}
/*
* Given a property, inherit the value from the parent dataset, or if received
* is TRUE, revert to the received value, if any.
*/
int
zfs_prop_inherit(zfs_handle_t *zhp, const char *propname, boolean_t received)
{
zfs_cmd_t zc = {"\0"};
int ret;
prop_changelist_t *cl;
libzfs_handle_t *hdl = zhp->zfs_hdl;
char errbuf[ERRBUFLEN];
zfs_prop_t prop;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot inherit %s for '%s'"), propname, zhp->zfs_name);
zc.zc_cookie = received;
if ((prop = zfs_name_to_prop(propname)) == ZPROP_USERPROP) {
/*
* For user properties, the amount of work we have to do is very
* small, so just do it here.
*/
if (!zfs_prop_user(propname)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
}
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
(void) strlcpy(zc.zc_value, propname, sizeof (zc.zc_value));
if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_INHERIT_PROP, &zc) != 0)
return (zfs_standard_error(hdl, errno, errbuf));
(void) get_stats(zhp);
return (0);
}
/*
* Verify that this property is inheritable.
*/
if (zfs_prop_readonly(prop))
return (zfs_error(hdl, EZFS_PROPREADONLY, errbuf));
if (!zfs_prop_inheritable(prop) && !received)
return (zfs_error(hdl, EZFS_PROPNONINHERIT, errbuf));
/*
* Check to see if the value applies to this type
*/
if (!zfs_prop_valid_for_type(prop, zhp->zfs_type, B_FALSE))
return (zfs_error(hdl, EZFS_PROPTYPE, errbuf));
/*
* Normalize the name, to get rid of shorthand abbreviations.
*/
propname = zfs_prop_to_name(prop);
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
(void) strlcpy(zc.zc_value, propname, sizeof (zc.zc_value));
if (prop == ZFS_PROP_MOUNTPOINT && getzoneid() == GLOBAL_ZONEID &&
zfs_prop_get_int(zhp, ZFS_PROP_ZONED)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset is used in a non-global zone"));
return (zfs_error(hdl, EZFS_ZONED, errbuf));
}
/*
* Determine datasets which will be affected by this change, if any.
*/
if ((cl = changelist_gather(zhp, prop, 0, 0)) == NULL)
return (-1);
if (prop == ZFS_PROP_MOUNTPOINT && changelist_haszonedchild(cl)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"child dataset with inherited mountpoint is used "
"in a non-global zone"));
ret = zfs_error(hdl, EZFS_ZONED, errbuf);
goto error;
}
if ((ret = changelist_prefix(cl)) != 0)
goto error;
if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_INHERIT_PROP, &zc) != 0) {
changelist_free(cl);
return (zfs_standard_error(hdl, errno, errbuf));
} else {
if ((ret = changelist_postfix(cl)) != 0)
goto error;
/*
* Refresh the statistics so the new property is reflected.
*/
(void) get_stats(zhp);
/*
* Remount the filesystem to propagate the change
* if one of the options handled by the generic
* Linux namespace layer has been modified.
*/
if (zfs_is_namespace_prop(prop) &&
zfs_is_mounted(zhp, NULL))
ret = zfs_mount(zhp, MNTOPT_REMOUNT, 0);
}
error:
changelist_free(cl);
return (ret);
}
/*
* True DSL properties are stored in an nvlist. The following two functions
* extract them appropriately.
*/
uint64_t
getprop_uint64(zfs_handle_t *zhp, zfs_prop_t prop, const char **source)
{
nvlist_t *nv;
uint64_t value;
*source = NULL;
if (nvlist_lookup_nvlist(zhp->zfs_props,
zfs_prop_to_name(prop), &nv) == 0) {
value = fnvlist_lookup_uint64(nv, ZPROP_VALUE);
(void) nvlist_lookup_string(nv, ZPROP_SOURCE, source);
} else {
verify(!zhp->zfs_props_table ||
zhp->zfs_props_table[prop] == B_TRUE);
value = zfs_prop_default_numeric(prop);
*source = "";
}
return (value);
}
static const char *
getprop_string(zfs_handle_t *zhp, zfs_prop_t prop, const char **source)
{
nvlist_t *nv;
const char *value;
*source = NULL;
if (nvlist_lookup_nvlist(zhp->zfs_props,
zfs_prop_to_name(prop), &nv) == 0) {
value = fnvlist_lookup_string(nv, ZPROP_VALUE);
(void) nvlist_lookup_string(nv, ZPROP_SOURCE, source);
} else {
verify(!zhp->zfs_props_table ||
zhp->zfs_props_table[prop] == B_TRUE);
value = zfs_prop_default_string(prop);
*source = "";
}
return (value);
}
static boolean_t
zfs_is_recvd_props_mode(zfs_handle_t *zhp)
{
return (zhp->zfs_props != NULL &&
zhp->zfs_props == zhp->zfs_recvd_props);
}
static void
zfs_set_recvd_props_mode(zfs_handle_t *zhp, uintptr_t *cookie)
{
*cookie = (uintptr_t)zhp->zfs_props;
zhp->zfs_props = zhp->zfs_recvd_props;
}
static void
zfs_unset_recvd_props_mode(zfs_handle_t *zhp, uintptr_t *cookie)
{
zhp->zfs_props = (nvlist_t *)*cookie;
*cookie = 0;
}
/*
* Internal function for getting a numeric property. Both zfs_prop_get() and
* zfs_prop_get_int() are built using this interface.
*
* Certain properties can be overridden using 'mount -o'. In this case, scan
* the contents of the /proc/self/mounts entry, searching for the
* appropriate options. If they differ from the on-disk values, report the
* current values and mark the source "temporary".
*/
static int
get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zprop_source_t *src,
const char **source, uint64_t *val)
{
zfs_cmd_t zc = {"\0"};
nvlist_t *zplprops = NULL;
struct mnttab mnt;
const char *mntopt_on = NULL;
const char *mntopt_off = NULL;
boolean_t received = zfs_is_recvd_props_mode(zhp);
*source = NULL;
/*
* If the property is being fetched for a snapshot, check whether
* the property is valid for the snapshot's head dataset type.
*/
if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT &&
!zfs_prop_valid_for_type(prop, zhp->zfs_head_type, B_TRUE)) {
*val = zfs_prop_default_numeric(prop);
return (-1);
}
switch (prop) {
case ZFS_PROP_ATIME:
mntopt_on = MNTOPT_ATIME;
mntopt_off = MNTOPT_NOATIME;
break;
case ZFS_PROP_RELATIME:
mntopt_on = MNTOPT_RELATIME;
mntopt_off = MNTOPT_NORELATIME;
break;
case ZFS_PROP_DEVICES:
mntopt_on = MNTOPT_DEVICES;
mntopt_off = MNTOPT_NODEVICES;
break;
case ZFS_PROP_EXEC:
mntopt_on = MNTOPT_EXEC;
mntopt_off = MNTOPT_NOEXEC;
break;
case ZFS_PROP_READONLY:
mntopt_on = MNTOPT_RO;
mntopt_off = MNTOPT_RW;
break;
case ZFS_PROP_SETUID:
mntopt_on = MNTOPT_SETUID;
mntopt_off = MNTOPT_NOSETUID;
break;
case ZFS_PROP_XATTR:
mntopt_on = MNTOPT_XATTR;
mntopt_off = MNTOPT_NOXATTR;
break;
case ZFS_PROP_NBMAND:
mntopt_on = MNTOPT_NBMAND;
mntopt_off = MNTOPT_NONBMAND;
break;
default:
break;
}
/*
* Because looking up the mount options is potentially expensive
* (iterating over all of /proc/self/mounts), we defer its
* calculation until we're looking up a property which requires
* its presence.
*/
if (!zhp->zfs_mntcheck &&
(mntopt_on != NULL || prop == ZFS_PROP_MOUNTED)) {
libzfs_handle_t *hdl = zhp->zfs_hdl;
struct mnttab entry;
if (libzfs_mnttab_find(hdl, zhp->zfs_name, &entry) == 0)
zhp->zfs_mntopts = zfs_strdup(hdl,
entry.mnt_mntopts);
zhp->zfs_mntcheck = B_TRUE;
}
if (zhp->zfs_mntopts == NULL)
mnt.mnt_mntopts = (char *)"";
else
mnt.mnt_mntopts = zhp->zfs_mntopts;
switch (prop) {
case ZFS_PROP_ATIME:
case ZFS_PROP_RELATIME:
case ZFS_PROP_DEVICES:
case ZFS_PROP_EXEC:
case ZFS_PROP_READONLY:
case ZFS_PROP_SETUID:
#ifndef __FreeBSD__
case ZFS_PROP_XATTR:
#endif
case ZFS_PROP_NBMAND:
*val = getprop_uint64(zhp, prop, source);
if (received)
break;
if (hasmntopt(&mnt, mntopt_on) && !*val) {
*val = B_TRUE;
if (src)
*src = ZPROP_SRC_TEMPORARY;
} else if (hasmntopt(&mnt, mntopt_off) && *val) {
*val = B_FALSE;
if (src)
*src = ZPROP_SRC_TEMPORARY;
}
break;
case ZFS_PROP_CANMOUNT:
case ZFS_PROP_VOLSIZE:
case ZFS_PROP_QUOTA:
case ZFS_PROP_REFQUOTA:
case ZFS_PROP_RESERVATION:
case ZFS_PROP_REFRESERVATION:
case ZFS_PROP_FILESYSTEM_LIMIT:
case ZFS_PROP_SNAPSHOT_LIMIT:
case ZFS_PROP_FILESYSTEM_COUNT:
case ZFS_PROP_SNAPSHOT_COUNT:
*val = getprop_uint64(zhp, prop, source);
if (*source == NULL) {
/* not default, must be local */
*source = zhp->zfs_name;
}
break;
case ZFS_PROP_MOUNTED:
*val = (zhp->zfs_mntopts != NULL);
break;
case ZFS_PROP_NUMCLONES:
*val = zhp->zfs_dmustats.dds_num_clones;
break;
case ZFS_PROP_VERSION:
case ZFS_PROP_NORMALIZE:
case ZFS_PROP_UTF8ONLY:
case ZFS_PROP_CASE:
zcmd_alloc_dst_nvlist(zhp->zfs_hdl, &zc, 0);
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_OBJSET_ZPLPROPS, &zc)) {
zcmd_free_nvlists(&zc);
if (prop == ZFS_PROP_VERSION &&
zhp->zfs_type == ZFS_TYPE_VOLUME)
*val = zfs_prop_default_numeric(prop);
return (-1);
}
if (zcmd_read_dst_nvlist(zhp->zfs_hdl, &zc, &zplprops) != 0 ||
nvlist_lookup_uint64(zplprops, zfs_prop_to_name(prop),
val) != 0) {
zcmd_free_nvlists(&zc);
return (-1);
}
nvlist_free(zplprops);
zcmd_free_nvlists(&zc);
break;
case ZFS_PROP_INCONSISTENT:
*val = zhp->zfs_dmustats.dds_inconsistent;
break;
case ZFS_PROP_REDACTED:
*val = zhp->zfs_dmustats.dds_redacted;
break;
case ZFS_PROP_GUID:
if (zhp->zfs_dmustats.dds_guid != 0)
*val = zhp->zfs_dmustats.dds_guid;
else
*val = getprop_uint64(zhp, prop, source);
break;
case ZFS_PROP_CREATETXG:
/*
* We can directly read createtxg property from zfs
* handle for Filesystem, Snapshot and ZVOL types.
*/
if (((zhp->zfs_type == ZFS_TYPE_FILESYSTEM) ||
(zhp->zfs_type == ZFS_TYPE_SNAPSHOT) ||
(zhp->zfs_type == ZFS_TYPE_VOLUME)) &&
(zhp->zfs_dmustats.dds_creation_txg != 0)) {
*val = zhp->zfs_dmustats.dds_creation_txg;
break;
} else {
*val = getprop_uint64(zhp, prop, source);
}
zfs_fallthrough;
default:
switch (zfs_prop_get_type(prop)) {
case PROP_TYPE_NUMBER:
case PROP_TYPE_INDEX:
*val = getprop_uint64(zhp, prop, source);
/*
* If we tried to use a default value for a
* readonly property, it means that it was not
* present. Note this only applies to "truly"
* readonly properties, not set-once properties
* like volblocksize.
*/
if (zfs_prop_readonly(prop) &&
!zfs_prop_setonce(prop) &&
*source != NULL && (*source)[0] == '\0') {
*source = NULL;
return (-1);
}
break;
case PROP_TYPE_STRING:
default:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"cannot get non-numeric property"));
return (zfs_error(zhp->zfs_hdl, EZFS_BADPROP,
dgettext(TEXT_DOMAIN, "internal error")));
}
}
return (0);
}
/*
* Calculate the source type, given the raw source string.
*/
static void
get_source(zfs_handle_t *zhp, zprop_source_t *srctype, const char *source,
char *statbuf, size_t statlen)
{
if (statbuf == NULL ||
srctype == NULL || *srctype == ZPROP_SRC_TEMPORARY) {
return;
}
if (source == NULL) {
*srctype = ZPROP_SRC_NONE;
} else if (source[0] == '\0') {
*srctype = ZPROP_SRC_DEFAULT;
} else if (strstr(source, ZPROP_SOURCE_VAL_RECVD) != NULL) {
*srctype = ZPROP_SRC_RECEIVED;
} else {
if (strcmp(source, zhp->zfs_name) == 0) {
*srctype = ZPROP_SRC_LOCAL;
} else {
(void) strlcpy(statbuf, source, statlen);
*srctype = ZPROP_SRC_INHERITED;
}
}
}
int
zfs_prop_get_recvd(zfs_handle_t *zhp, const char *propname, char *propbuf,
size_t proplen, boolean_t literal)
{
zfs_prop_t prop;
int err = 0;
if (zhp->zfs_recvd_props == NULL)
if (get_recvd_props_ioctl(zhp) != 0)
return (-1);
prop = zfs_name_to_prop(propname);
if (prop != ZPROP_USERPROP) {
uintptr_t cookie;
if (!nvlist_exists(zhp->zfs_recvd_props, propname))
return (-1);
zfs_set_recvd_props_mode(zhp, &cookie);
err = zfs_prop_get(zhp, prop, propbuf, proplen,
NULL, NULL, 0, literal);
zfs_unset_recvd_props_mode(zhp, &cookie);
} else {
nvlist_t *propval;
const char *recvdval;
if (nvlist_lookup_nvlist(zhp->zfs_recvd_props,
propname, &propval) != 0)
return (-1);
recvdval = fnvlist_lookup_string(propval, ZPROP_VALUE);
(void) strlcpy(propbuf, recvdval, proplen);
}
return (err == 0 ? 0 : -1);
}
static int
get_clones_string(zfs_handle_t *zhp, char *propbuf, size_t proplen)
{
nvlist_t *value;
nvpair_t *pair;
value = zfs_get_clones_nvl(zhp);
if (value == NULL || nvlist_empty(value))
return (-1);
propbuf[0] = '\0';
for (pair = nvlist_next_nvpair(value, NULL); pair != NULL;
pair = nvlist_next_nvpair(value, pair)) {
if (propbuf[0] != '\0')
(void) strlcat(propbuf, ",", proplen);
(void) strlcat(propbuf, nvpair_name(pair), proplen);
}
return (0);
}
struct get_clones_arg {
uint64_t numclones;
nvlist_t *value;
const char *origin;
char buf[ZFS_MAX_DATASET_NAME_LEN];
};
static int
get_clones_cb(zfs_handle_t *zhp, void *arg)
{
struct get_clones_arg *gca = arg;
if (gca->numclones == 0) {
zfs_close(zhp);
return (0);
}
if (zfs_prop_get(zhp, ZFS_PROP_ORIGIN, gca->buf, sizeof (gca->buf),
NULL, NULL, 0, B_TRUE) != 0)
goto out;
if (strcmp(gca->buf, gca->origin) == 0) {
fnvlist_add_boolean(gca->value, zfs_get_name(zhp));
gca->numclones--;
}
out:
- (void) zfs_iter_children(zhp, 0, get_clones_cb, gca);
+ (void) zfs_iter_children_v2(zhp, 0, get_clones_cb, gca);
zfs_close(zhp);
return (0);
}
nvlist_t *
zfs_get_clones_nvl(zfs_handle_t *zhp)
{
nvlist_t *nv, *value;
if (nvlist_lookup_nvlist(zhp->zfs_props,
zfs_prop_to_name(ZFS_PROP_CLONES), &nv) != 0) {
struct get_clones_arg gca;
/*
* if this is a snapshot, then the kernel wasn't able
* to get the clones. Do it by slowly iterating.
*/
if (zhp->zfs_type != ZFS_TYPE_SNAPSHOT)
return (NULL);
if (nvlist_alloc(&nv, NV_UNIQUE_NAME, 0) != 0)
return (NULL);
if (nvlist_alloc(&value, NV_UNIQUE_NAME, 0) != 0) {
nvlist_free(nv);
return (NULL);
}
gca.numclones = zfs_prop_get_int(zhp, ZFS_PROP_NUMCLONES);
gca.value = value;
gca.origin = zhp->zfs_name;
if (gca.numclones != 0) {
zfs_handle_t *root;
char pool[ZFS_MAX_DATASET_NAME_LEN];
char *cp = pool;
/* get the pool name */
(void) strlcpy(pool, zhp->zfs_name, sizeof (pool));
(void) strsep(&cp, "/@");
root = zfs_open(zhp->zfs_hdl, pool,
ZFS_TYPE_FILESYSTEM);
if (root == NULL) {
nvlist_free(nv);
nvlist_free(value);
return (NULL);
}
(void) get_clones_cb(root, &gca);
}
if (gca.numclones != 0 ||
nvlist_add_nvlist(nv, ZPROP_VALUE, value) != 0 ||
nvlist_add_nvlist(zhp->zfs_props,
zfs_prop_to_name(ZFS_PROP_CLONES), nv) != 0) {
nvlist_free(nv);
nvlist_free(value);
return (NULL);
}
nvlist_free(nv);
nvlist_free(value);
nv = fnvlist_lookup_nvlist(zhp->zfs_props,
zfs_prop_to_name(ZFS_PROP_CLONES));
}
return (fnvlist_lookup_nvlist(nv, ZPROP_VALUE));
}
static int
get_rsnaps_string(zfs_handle_t *zhp, char *propbuf, size_t proplen)
{
nvlist_t *value;
uint64_t *snaps;
uint_t nsnaps;
if (nvlist_lookup_nvlist(zhp->zfs_props,
zfs_prop_to_name(ZFS_PROP_REDACT_SNAPS), &value) != 0)
return (-1);
if (nvlist_lookup_uint64_array(value, ZPROP_VALUE, &snaps,
&nsnaps) != 0)
return (-1);
if (nsnaps == 0) {
/* There's no redaction snapshots; pass a special value back */
(void) snprintf(propbuf, proplen, "none");
return (0);
}
propbuf[0] = '\0';
for (int i = 0; i < nsnaps; i++) {
char buf[128];
if (propbuf[0] != '\0')
(void) strlcat(propbuf, ",", proplen);
(void) snprintf(buf, sizeof (buf), "%llu",
(u_longlong_t)snaps[i]);
(void) strlcat(propbuf, buf, proplen);
}
return (0);
}
/*
* Accepts a property and value and checks that the value
* matches the one found by the channel program. If they are
* not equal, print both of them.
*/
static void
zcp_check(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t intval,
const char *strval)
{
if (!zhp->zfs_hdl->libzfs_prop_debug)
return;
int error;
char *poolname = zhp->zpool_hdl->zpool_name;
const char *prop_name = zfs_prop_to_name(prop);
const char *program =
"args = ...\n"
"ds = args['dataset']\n"
"prop = args['property']\n"
"value, setpoint = zfs.get_prop(ds, prop)\n"
"return {value=value, setpoint=setpoint}\n";
nvlist_t *outnvl;
nvlist_t *retnvl;
nvlist_t *argnvl = fnvlist_alloc();
fnvlist_add_string(argnvl, "dataset", zhp->zfs_name);
fnvlist_add_string(argnvl, "property", zfs_prop_to_name(prop));
error = lzc_channel_program_nosync(poolname, program,
10 * 1000 * 1000, 10 * 1024 * 1024, argnvl, &outnvl);
if (error == 0) {
retnvl = fnvlist_lookup_nvlist(outnvl, "return");
if (zfs_prop_get_type(prop) == PROP_TYPE_NUMBER) {
int64_t ans;
error = nvlist_lookup_int64(retnvl, "value", &ans);
if (error != 0) {
(void) fprintf(stderr, "%s: zcp check error: "
"%u\n", prop_name, error);
return;
}
if (ans != intval) {
(void) fprintf(stderr, "%s: zfs found %llu, "
"but zcp found %llu\n", prop_name,
(u_longlong_t)intval, (u_longlong_t)ans);
}
} else {
const char *str_ans;
error = nvlist_lookup_string(retnvl, "value", &str_ans);
if (error != 0) {
(void) fprintf(stderr, "%s: zcp check error: "
"%u\n", prop_name, error);
return;
}
if (strcmp(strval, str_ans) != 0) {
(void) fprintf(stderr,
"%s: zfs found '%s', but zcp found '%s'\n",
prop_name, strval, str_ans);
}
}
} else {
(void) fprintf(stderr, "%s: zcp check failed, channel program "
"error: %u\n", prop_name, error);
}
nvlist_free(argnvl);
nvlist_free(outnvl);
}
/*
* Retrieve a property from the given object. If 'literal' is specified, then
* numbers are left as exact values. Otherwise, numbers are converted to a
* human-readable form.
*
* Returns 0 on success, or -1 on error.
*/
int
zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen,
zprop_source_t *src, char *statbuf, size_t statlen, boolean_t literal)
{
const char *source = NULL;
uint64_t val;
const char *str;
const char *strval;
boolean_t received = zfs_is_recvd_props_mode(zhp);
/*
* Check to see if this property applies to our object
*/
if (!zfs_prop_valid_for_type(prop, zhp->zfs_type, B_FALSE))
return (-1);
if (received && zfs_prop_readonly(prop))
return (-1);
if (src)
*src = ZPROP_SRC_NONE;
switch (prop) {
case ZFS_PROP_CREATION:
/*
* 'creation' is a time_t stored in the statistics. We convert
* this into a string unless 'literal' is specified.
*/
{
val = getprop_uint64(zhp, prop, &source);
time_t time = (time_t)val;
struct tm t;
if (literal ||
localtime_r(&time, &t) == NULL ||
strftime(propbuf, proplen, "%a %b %e %k:%M %Y",
&t) == 0)
(void) snprintf(propbuf, proplen, "%llu",
(u_longlong_t)val);
}
zcp_check(zhp, prop, val, NULL);
break;
case ZFS_PROP_MOUNTPOINT:
/*
* Getting the precise mountpoint can be tricky.
*
* - for 'none' or 'legacy', return those values.
* - for inherited mountpoints, we want to take everything
* after our ancestor and append it to the inherited value.
*
* If the pool has an alternate root, we want to prepend that
* root to any values we return.
*/
str = getprop_string(zhp, prop, &source);
if (str[0] == '/') {
char buf[MAXPATHLEN];
char *root = buf;
const char *relpath;
/*
* If we inherit the mountpoint, even from a dataset
* with a received value, the source will be the path of
* the dataset we inherit from. If source is
* ZPROP_SOURCE_VAL_RECVD, the received value is not
* inherited.
*/
if (strcmp(source, ZPROP_SOURCE_VAL_RECVD) == 0) {
relpath = "";
} else {
relpath = zhp->zfs_name + strlen(source);
if (relpath[0] == '/')
relpath++;
}
if ((zpool_get_prop(zhp->zpool_hdl,
ZPOOL_PROP_ALTROOT, buf, MAXPATHLEN, NULL,
B_FALSE)) || (strcmp(root, "-") == 0))
root[0] = '\0';
/*
* Special case an alternate root of '/'. This will
* avoid having multiple leading slashes in the
* mountpoint path.
*/
if (strcmp(root, "/") == 0)
root++;
/*
* If the mountpoint is '/' then skip over this
* if we are obtaining either an alternate root or
* an inherited mountpoint.
*/
if (str[1] == '\0' && (root[0] != '\0' ||
relpath[0] != '\0'))
str++;
if (relpath[0] == '\0')
(void) snprintf(propbuf, proplen, "%s%s",
root, str);
else
(void) snprintf(propbuf, proplen, "%s%s%s%s",
root, str, relpath[0] == '@' ? "" : "/",
relpath);
} else {
/* 'legacy' or 'none' */
(void) strlcpy(propbuf, str, proplen);
}
zcp_check(zhp, prop, 0, propbuf);
break;
case ZFS_PROP_ORIGIN:
if (*zhp->zfs_dmustats.dds_origin != '\0') {
str = (char *)&zhp->zfs_dmustats.dds_origin;
} else {
str = getprop_string(zhp, prop, &source);
}
if (str == NULL || *str == '\0')
str = zfs_prop_default_string(prop);
if (str == NULL)
return (-1);
(void) strlcpy(propbuf, str, proplen);
zcp_check(zhp, prop, 0, str);
break;
case ZFS_PROP_REDACT_SNAPS:
if (get_rsnaps_string(zhp, propbuf, proplen) != 0)
return (-1);
break;
case ZFS_PROP_CLONES:
if (get_clones_string(zhp, propbuf, proplen) != 0)
return (-1);
break;
case ZFS_PROP_QUOTA:
case ZFS_PROP_REFQUOTA:
case ZFS_PROP_RESERVATION:
case ZFS_PROP_REFRESERVATION:
if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
return (-1);
/*
* If quota or reservation is 0, we translate this into 'none'
* (unless literal is set), and indicate that it's the default
* value. Otherwise, we print the number nicely and indicate
* that its set locally.
*/
if (val == 0) {
if (literal)
(void) strlcpy(propbuf, "0", proplen);
else
(void) strlcpy(propbuf, "none", proplen);
} else {
if (literal)
(void) snprintf(propbuf, proplen, "%llu",
(u_longlong_t)val);
else
zfs_nicebytes(val, propbuf, proplen);
}
zcp_check(zhp, prop, val, NULL);
break;
case ZFS_PROP_FILESYSTEM_LIMIT:
case ZFS_PROP_SNAPSHOT_LIMIT:
case ZFS_PROP_FILESYSTEM_COUNT:
case ZFS_PROP_SNAPSHOT_COUNT:
if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
return (-1);
/*
* If limit is UINT64_MAX, we translate this into 'none', and
* indicate that it's the default value. Otherwise, we print
* the number nicely and indicate that it's set locally.
*/
if (val == UINT64_MAX) {
(void) strlcpy(propbuf, "none", proplen);
} else if (literal) {
(void) snprintf(propbuf, proplen, "%llu",
(u_longlong_t)val);
} else {
zfs_nicenum(val, propbuf, proplen);
}
zcp_check(zhp, prop, val, NULL);
break;
case ZFS_PROP_REFRATIO:
case ZFS_PROP_COMPRESSRATIO:
if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
return (-1);
if (literal)
(void) snprintf(propbuf, proplen, "%llu.%02llu",
(u_longlong_t)(val / 100),
(u_longlong_t)(val % 100));
else
(void) snprintf(propbuf, proplen, "%llu.%02llux",
(u_longlong_t)(val / 100),
(u_longlong_t)(val % 100));
zcp_check(zhp, prop, val, NULL);
break;
case ZFS_PROP_TYPE:
switch (zhp->zfs_type) {
case ZFS_TYPE_FILESYSTEM:
str = "filesystem";
break;
case ZFS_TYPE_VOLUME:
str = "volume";
break;
case ZFS_TYPE_SNAPSHOT:
str = "snapshot";
break;
case ZFS_TYPE_BOOKMARK:
str = "bookmark";
break;
default:
abort();
}
(void) snprintf(propbuf, proplen, "%s", str);
zcp_check(zhp, prop, 0, propbuf);
break;
case ZFS_PROP_MOUNTED:
/*
* The 'mounted' property is a pseudo-property that described
* whether the filesystem is currently mounted. Even though
* it's a boolean value, the typical values of "on" and "off"
* don't make sense, so we translate to "yes" and "no".
*/
if (get_numeric_property(zhp, ZFS_PROP_MOUNTED,
src, &source, &val) != 0)
return (-1);
if (val)
(void) strlcpy(propbuf, "yes", proplen);
else
(void) strlcpy(propbuf, "no", proplen);
break;
case ZFS_PROP_NAME:
/*
* The 'name' property is a pseudo-property derived from the
* dataset name. It is presented as a real property to simplify
* consumers.
*/
(void) strlcpy(propbuf, zhp->zfs_name, proplen);
zcp_check(zhp, prop, 0, propbuf);
break;
case ZFS_PROP_MLSLABEL:
{
#ifdef HAVE_MLSLABEL
m_label_t *new_sl = NULL;
char *ascii = NULL; /* human readable label */
(void) strlcpy(propbuf,
getprop_string(zhp, prop, &source), proplen);
if (literal || (strcasecmp(propbuf,
ZFS_MLSLABEL_DEFAULT) == 0))
break;
/*
* Try to translate the internal hex string to
* human-readable output. If there are any
* problems just use the hex string.
*/
if (str_to_label(propbuf, &new_sl, MAC_LABEL,
L_NO_CORRECTION, NULL) == -1) {
m_label_free(new_sl);
break;
}
if (label_to_str(new_sl, &ascii, M_LABEL,
DEF_NAMES) != 0) {
if (ascii)
free(ascii);
m_label_free(new_sl);
break;
}
m_label_free(new_sl);
(void) strlcpy(propbuf, ascii, proplen);
free(ascii);
#else
(void) strlcpy(propbuf,
getprop_string(zhp, prop, &source), proplen);
#endif /* HAVE_MLSLABEL */
}
break;
case ZFS_PROP_GUID:
case ZFS_PROP_KEY_GUID:
case ZFS_PROP_IVSET_GUID:
case ZFS_PROP_CREATETXG:
case ZFS_PROP_OBJSETID:
case ZFS_PROP_PBKDF2_ITERS:
/*
* These properties are stored as numbers, but they are
* identifiers or counters.
* We don't want them to be pretty printed, because pretty
* printing truncates their values making them useless.
*/
if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
return (-1);
(void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)val);
zcp_check(zhp, prop, val, NULL);
break;
case ZFS_PROP_REFERENCED:
case ZFS_PROP_AVAILABLE:
case ZFS_PROP_USED:
case ZFS_PROP_USEDSNAP:
case ZFS_PROP_USEDDS:
case ZFS_PROP_USEDREFRESERV:
case ZFS_PROP_USEDCHILD:
if (get_numeric_property(zhp, prop, src, &source, &val) != 0)
return (-1);
if (literal) {
(void) snprintf(propbuf, proplen, "%llu",
(u_longlong_t)val);
} else {
zfs_nicebytes(val, propbuf, proplen);
}
zcp_check(zhp, prop, val, NULL);
break;
case ZFS_PROP_SNAPSHOTS_CHANGED:
{
if ((get_numeric_property(zhp, prop, src, &source,
&val) != 0) || val == 0) {
return (-1);
}
time_t time = (time_t)val;
struct tm t;
if (literal ||
localtime_r(&time, &t) == NULL ||
strftime(propbuf, proplen, "%a %b %e %k:%M:%S %Y",
&t) == 0)
(void) snprintf(propbuf, proplen, "%llu",
(u_longlong_t)val);
}
zcp_check(zhp, prop, val, NULL);
break;
default:
switch (zfs_prop_get_type(prop)) {
case PROP_TYPE_NUMBER:
if (get_numeric_property(zhp, prop, src,
&source, &val) != 0) {
return (-1);
}
if (literal) {
(void) snprintf(propbuf, proplen, "%llu",
(u_longlong_t)val);
} else {
zfs_nicenum(val, propbuf, proplen);
}
zcp_check(zhp, prop, val, NULL);
break;
case PROP_TYPE_STRING:
str = getprop_string(zhp, prop, &source);
if (str == NULL)
return (-1);
(void) strlcpy(propbuf, str, proplen);
zcp_check(zhp, prop, 0, str);
break;
case PROP_TYPE_INDEX:
if (get_numeric_property(zhp, prop, src,
&source, &val) != 0)
return (-1);
if (zfs_prop_index_to_string(prop, val, &strval) != 0)
return (-1);
(void) strlcpy(propbuf, strval, proplen);
zcp_check(zhp, prop, 0, strval);
break;
default:
abort();
}
}
get_source(zhp, src, source, statbuf, statlen);
return (0);
}
/*
* Utility function to get the given numeric property. Does no validation that
* the given property is the appropriate type; should only be used with
* hard-coded property types.
*/
uint64_t
zfs_prop_get_int(zfs_handle_t *zhp, zfs_prop_t prop)
{
const char *source;
uint64_t val = 0;
(void) get_numeric_property(zhp, prop, NULL, &source, &val);
return (val);
}
static int
zfs_prop_set_int(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t val)
{
char buf[64];
(void) snprintf(buf, sizeof (buf), "%llu", (longlong_t)val);
return (zfs_prop_set(zhp, zfs_prop_to_name(prop), buf));
}
/*
* Similar to zfs_prop_get(), but returns the value as an integer.
*/
int
zfs_prop_get_numeric(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t *value,
zprop_source_t *src, char *statbuf, size_t statlen)
{
const char *source;
/*
* Check to see if this property applies to our object
*/
if (!zfs_prop_valid_for_type(prop, zhp->zfs_type, B_FALSE)) {
return (zfs_error_fmt(zhp->zfs_hdl, EZFS_PROPTYPE,
dgettext(TEXT_DOMAIN, "cannot get property '%s'"),
zfs_prop_to_name(prop)));
}
if (src)
*src = ZPROP_SRC_NONE;
if (get_numeric_property(zhp, prop, src, &source, value) != 0)
return (-1);
get_source(zhp, src, source, statbuf, statlen);
return (0);
}
#ifdef HAVE_IDMAP
static int
idmap_id_to_numeric_domain_rid(uid_t id, boolean_t isuser,
char **domainp, idmap_rid_t *ridp)
{
idmap_get_handle_t *get_hdl = NULL;
idmap_stat status;
int err = EINVAL;
if (idmap_get_create(&get_hdl) != IDMAP_SUCCESS)
goto out;
if (isuser) {
err = idmap_get_sidbyuid(get_hdl, id,
IDMAP_REQ_FLG_USE_CACHE, domainp, ridp, &status);
} else {
err = idmap_get_sidbygid(get_hdl, id,
IDMAP_REQ_FLG_USE_CACHE, domainp, ridp, &status);
}
if (err == IDMAP_SUCCESS &&
idmap_get_mappings(get_hdl) == IDMAP_SUCCESS &&
status == IDMAP_SUCCESS)
err = 0;
else
err = EINVAL;
out:
if (get_hdl)
idmap_get_destroy(get_hdl);
return (err);
}
#endif /* HAVE_IDMAP */
/*
* convert the propname into parameters needed by kernel
* Eg: userquota@ahrens -> ZFS_PROP_USERQUOTA, "", 126829
* Eg: userused@matt@domain -> ZFS_PROP_USERUSED, "S-1-123-456", 789
* Eg: groupquota@staff -> ZFS_PROP_GROUPQUOTA, "", 1234
* Eg: groupused@staff -> ZFS_PROP_GROUPUSED, "", 1234
* Eg: projectquota@123 -> ZFS_PROP_PROJECTQUOTA, "", 123
* Eg: projectused@789 -> ZFS_PROP_PROJECTUSED, "", 789
*/
static int
userquota_propname_decode(const char *propname, boolean_t zoned,
zfs_userquota_prop_t *typep, char *domain, int domainlen, uint64_t *ridp)
{
zfs_userquota_prop_t type;
char *cp;
boolean_t isuser;
boolean_t isgroup;
boolean_t isproject;
struct passwd *pw;
struct group *gr;
domain[0] = '\0';
/* Figure out the property type ({user|group|project}{quota|space}) */
for (type = 0; type < ZFS_NUM_USERQUOTA_PROPS; type++) {
if (strncmp(propname, zfs_userquota_prop_prefixes[type],
strlen(zfs_userquota_prop_prefixes[type])) == 0)
break;
}
if (type == ZFS_NUM_USERQUOTA_PROPS)
return (EINVAL);
*typep = type;
isuser = (type == ZFS_PROP_USERQUOTA || type == ZFS_PROP_USERUSED ||
type == ZFS_PROP_USEROBJQUOTA ||
type == ZFS_PROP_USEROBJUSED);
isgroup = (type == ZFS_PROP_GROUPQUOTA || type == ZFS_PROP_GROUPUSED ||
type == ZFS_PROP_GROUPOBJQUOTA ||
type == ZFS_PROP_GROUPOBJUSED);
isproject = (type == ZFS_PROP_PROJECTQUOTA ||
type == ZFS_PROP_PROJECTUSED || type == ZFS_PROP_PROJECTOBJQUOTA ||
type == ZFS_PROP_PROJECTOBJUSED);
cp = strchr(propname, '@') + 1;
if (isuser && (pw = getpwnam(cp)) != NULL) {
if (zoned && getzoneid() == GLOBAL_ZONEID)
return (ENOENT);
*ridp = pw->pw_uid;
} else if (isgroup && (gr = getgrnam(cp)) != NULL) {
if (zoned && getzoneid() == GLOBAL_ZONEID)
return (ENOENT);
*ridp = gr->gr_gid;
} else if (!isproject && strchr(cp, '@')) {
#ifdef HAVE_IDMAP
/*
* It's a SID name (eg "user@domain") that needs to be
* turned into S-1-domainID-RID.
*/
directory_error_t e;
char *numericsid = NULL;
char *end;
if (zoned && getzoneid() == GLOBAL_ZONEID)
return (ENOENT);
if (isuser) {
e = directory_sid_from_user_name(NULL,
cp, &numericsid);
} else {
e = directory_sid_from_group_name(NULL,
cp, &numericsid);
}
if (e != NULL) {
directory_error_free(e);
return (ENOENT);
}
if (numericsid == NULL)
return (ENOENT);
cp = numericsid;
(void) strlcpy(domain, cp, domainlen);
cp = strrchr(domain, '-');
*cp = '\0';
cp++;
errno = 0;
*ridp = strtoull(cp, &end, 10);
free(numericsid);
if (errno != 0 || *end != '\0')
return (EINVAL);
#else
(void) domainlen;
return (ENOSYS);
#endif /* HAVE_IDMAP */
} else {
/* It's a user/group/project ID (eg "12345"). */
uid_t id;
char *end;
id = strtoul(cp, &end, 10);
if (*end != '\0')
return (EINVAL);
if (id > MAXUID && !isproject) {
#ifdef HAVE_IDMAP
/* It's an ephemeral ID. */
idmap_rid_t rid;
char *mapdomain;
if (idmap_id_to_numeric_domain_rid(id, isuser,
&mapdomain, &rid) != 0)
return (ENOENT);
(void) strlcpy(domain, mapdomain, domainlen);
*ridp = rid;
#else
return (ENOSYS);
#endif /* HAVE_IDMAP */
} else {
*ridp = id;
}
}
return (0);
}
static int
zfs_prop_get_userquota_common(zfs_handle_t *zhp, const char *propname,
uint64_t *propvalue, zfs_userquota_prop_t *typep)
{
int err;
zfs_cmd_t zc = {"\0"};
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
err = userquota_propname_decode(propname,
zfs_prop_get_int(zhp, ZFS_PROP_ZONED),
typep, zc.zc_value, sizeof (zc.zc_value), &zc.zc_guid);
zc.zc_objset_type = *typep;
if (err)
return (err);
err = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_USERSPACE_ONE, &zc);
if (err)
return (err);
*propvalue = zc.zc_cookie;
return (0);
}
int
zfs_prop_get_userquota_int(zfs_handle_t *zhp, const char *propname,
uint64_t *propvalue)
{
zfs_userquota_prop_t type;
return (zfs_prop_get_userquota_common(zhp, propname, propvalue,
&type));
}
int
zfs_prop_get_userquota(zfs_handle_t *zhp, const char *propname,
char *propbuf, int proplen, boolean_t literal)
{
int err;
uint64_t propvalue;
zfs_userquota_prop_t type;
err = zfs_prop_get_userquota_common(zhp, propname, &propvalue,
&type);
if (err)
return (err);
if (literal) {
(void) snprintf(propbuf, proplen, "%llu",
(u_longlong_t)propvalue);
} else if (propvalue == 0 &&
(type == ZFS_PROP_USERQUOTA || type == ZFS_PROP_GROUPQUOTA ||
type == ZFS_PROP_USEROBJQUOTA || type == ZFS_PROP_GROUPOBJQUOTA ||
type == ZFS_PROP_PROJECTQUOTA ||
type == ZFS_PROP_PROJECTOBJQUOTA)) {
(void) strlcpy(propbuf, "none", proplen);
} else if (type == ZFS_PROP_USERQUOTA || type == ZFS_PROP_GROUPQUOTA ||
type == ZFS_PROP_USERUSED || type == ZFS_PROP_GROUPUSED ||
type == ZFS_PROP_PROJECTUSED || type == ZFS_PROP_PROJECTQUOTA) {
zfs_nicebytes(propvalue, propbuf, proplen);
} else {
zfs_nicenum(propvalue, propbuf, proplen);
}
return (0);
}
/*
* propname must start with "written@" or "written#".
*/
int
zfs_prop_get_written_int(zfs_handle_t *zhp, const char *propname,
uint64_t *propvalue)
{
int err;
zfs_cmd_t zc = {"\0"};
const char *snapname;
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
assert(zfs_prop_written(propname));
snapname = propname + strlen("written@");
if (strchr(snapname, '@') != NULL || strchr(snapname, '#') != NULL) {
/* full snapshot or bookmark name specified */
(void) strlcpy(zc.zc_value, snapname, sizeof (zc.zc_value));
} else {
/* snapname is the short name, append it to zhp's fsname */
char *cp;
(void) strlcpy(zc.zc_value, zhp->zfs_name,
sizeof (zc.zc_value));
cp = strchr(zc.zc_value, '@');
if (cp != NULL)
*cp = '\0';
(void) strlcat(zc.zc_value, snapname - 1, sizeof (zc.zc_value));
}
err = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_SPACE_WRITTEN, &zc);
if (err)
return (err);
*propvalue = zc.zc_cookie;
return (0);
}
int
zfs_prop_get_written(zfs_handle_t *zhp, const char *propname,
char *propbuf, int proplen, boolean_t literal)
{
int err;
uint64_t propvalue;
err = zfs_prop_get_written_int(zhp, propname, &propvalue);
if (err)
return (err);
if (literal) {
(void) snprintf(propbuf, proplen, "%llu",
(u_longlong_t)propvalue);
} else {
zfs_nicebytes(propvalue, propbuf, proplen);
}
return (0);
}
/*
* Returns the name of the given zfs handle.
*/
const char *
zfs_get_name(const zfs_handle_t *zhp)
{
return (zhp->zfs_name);
}
/*
* Returns the name of the parent pool for the given zfs handle.
*/
const char *
zfs_get_pool_name(const zfs_handle_t *zhp)
{
return (zhp->zpool_hdl->zpool_name);
}
/*
* Returns the type of the given zfs handle.
*/
zfs_type_t
zfs_get_type(const zfs_handle_t *zhp)
{
return (zhp->zfs_type);
}
/*
* Returns the type of the given zfs handle,
* or, if a snapshot, the type of the snapshotted dataset.
*/
zfs_type_t
zfs_get_underlying_type(const zfs_handle_t *zhp)
{
return (zhp->zfs_head_type);
}
/*
* Is one dataset name a child dataset of another?
*
* Needs to handle these cases:
* Dataset 1 "a/foo" "a/foo" "a/foo" "a/foo"
* Dataset 2 "a/fo" "a/foobar" "a/bar/baz" "a/foo/bar"
* Descendant? No. No. No. Yes.
*/
static boolean_t
is_descendant(const char *ds1, const char *ds2)
{
size_t d1len = strlen(ds1);
/* ds2 can't be a descendant if it's smaller */
if (strlen(ds2) < d1len)
return (B_FALSE);
/* otherwise, compare strings and verify that there's a '/' char */
return (ds2[d1len] == '/' && (strncmp(ds1, ds2, d1len) == 0));
}
/*
* Given a complete name, return just the portion that refers to the parent.
* Will return -1 if there is no parent (path is just the name of the
* pool).
*/
static int
parent_name(const char *path, char *buf, size_t buflen)
{
char *slashp;
(void) strlcpy(buf, path, buflen);
if ((slashp = strrchr(buf, '/')) == NULL)
return (-1);
*slashp = '\0';
return (0);
}
int
zfs_parent_name(zfs_handle_t *zhp, char *buf, size_t buflen)
{
return (parent_name(zfs_get_name(zhp), buf, buflen));
}
/*
* If accept_ancestor is false, then check to make sure that the given path has
* a parent, and that it exists. If accept_ancestor is true, then find the
* closest existing ancestor for the given path. In prefixlen return the
* length of already existing prefix of the given path. We also fetch the
* 'zoned' property, which is used to validate property settings when creating
* new datasets.
*/
static int
check_parents(libzfs_handle_t *hdl, const char *path, uint64_t *zoned,
boolean_t accept_ancestor, int *prefixlen)
{
zfs_cmd_t zc = {"\0"};
char parent[ZFS_MAX_DATASET_NAME_LEN];
char *slash;
zfs_handle_t *zhp;
char errbuf[ERRBUFLEN];
uint64_t is_zoned;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot create '%s'"), path);
/* get parent, and check to see if this is just a pool */
if (parent_name(path, parent, sizeof (parent)) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"missing dataset name"));
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
}
/* check to see if the pool exists */
if ((slash = strchr(parent, '/')) == NULL)
slash = parent + strlen(parent);
(void) strlcpy(zc.zc_name, parent,
MIN(sizeof (zc.zc_name), slash - parent + 1));
if (zfs_ioctl(hdl, ZFS_IOC_OBJSET_STATS, &zc) != 0 &&
errno == ENOENT) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"no such pool '%s'"), zc.zc_name);
return (zfs_error(hdl, EZFS_NOENT, errbuf));
}
/* check to see if the parent dataset exists */
while ((zhp = make_dataset_handle(hdl, parent)) == NULL) {
if (errno == ENOENT && accept_ancestor) {
/*
* Go deeper to find an ancestor, give up on top level.
*/
if (parent_name(parent, parent, sizeof (parent)) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"no such pool '%s'"), zc.zc_name);
return (zfs_error(hdl, EZFS_NOENT, errbuf));
}
} else if (errno == ENOENT) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"parent does not exist"));
return (zfs_error(hdl, EZFS_NOENT, errbuf));
} else
return (zfs_standard_error(hdl, errno, errbuf));
}
is_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
if (zoned != NULL)
*zoned = is_zoned;
/* we are in a non-global zone, but parent is in the global zone */
if (getzoneid() != GLOBAL_ZONEID && !is_zoned) {
(void) zfs_standard_error(hdl, EPERM, errbuf);
zfs_close(zhp);
return (-1);
}
/* make sure parent is a filesystem */
if (zfs_get_type(zhp) != ZFS_TYPE_FILESYSTEM) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"parent is not a filesystem"));
(void) zfs_error(hdl, EZFS_BADTYPE, errbuf);
zfs_close(zhp);
return (-1);
}
zfs_close(zhp);
if (prefixlen != NULL)
*prefixlen = strlen(parent);
return (0);
}
/*
* Finds whether the dataset of the given type(s) exists.
*/
boolean_t
zfs_dataset_exists(libzfs_handle_t *hdl, const char *path, zfs_type_t types)
{
zfs_handle_t *zhp;
if (!zfs_validate_name(hdl, path, types, B_FALSE))
return (B_FALSE);
/*
* Try to get stats for the dataset, which will tell us if it exists.
*/
if ((zhp = make_dataset_handle(hdl, path)) != NULL) {
int ds_type = zhp->zfs_type;
zfs_close(zhp);
if (types & ds_type)
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Given a path to 'target', create all the ancestors between
* the prefixlen portion of the path, and the target itself.
* Fail if the initial prefixlen-ancestor does not already exist.
*/
int
create_parents(libzfs_handle_t *hdl, char *target, int prefixlen)
{
zfs_handle_t *h;
char *cp;
const char *opname;
/* make sure prefix exists */
cp = target + prefixlen;
if (*cp != '/') {
assert(strchr(cp, '/') == NULL);
h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM);
} else {
*cp = '\0';
h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM);
*cp = '/';
}
if (h == NULL)
return (-1);
zfs_close(h);
/*
* Attempt to create, mount, and share any ancestor filesystems,
* up to the prefixlen-long one.
*/
for (cp = target + prefixlen + 1;
(cp = strchr(cp, '/')) != NULL; *cp = '/', cp++) {
*cp = '\0';
h = make_dataset_handle(hdl, target);
if (h) {
/* it already exists, nothing to do here */
zfs_close(h);
continue;
}
if (zfs_create(hdl, target, ZFS_TYPE_FILESYSTEM,
NULL) != 0) {
opname = dgettext(TEXT_DOMAIN, "create");
goto ancestorerr;
}
h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM);
if (h == NULL) {
opname = dgettext(TEXT_DOMAIN, "open");
goto ancestorerr;
}
if (zfs_mount(h, NULL, 0) != 0) {
opname = dgettext(TEXT_DOMAIN, "mount");
goto ancestorerr;
}
if (zfs_share(h, NULL) != 0) {
opname = dgettext(TEXT_DOMAIN, "share");
goto ancestorerr;
}
zfs_close(h);
}
zfs_commit_shares(NULL);
return (0);
ancestorerr:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"failed to %s ancestor '%s'"), opname, target);
return (-1);
}
/*
* Creates non-existing ancestors of the given path.
*/
int
zfs_create_ancestors(libzfs_handle_t *hdl, const char *path)
{
int prefix;
char *path_copy;
char errbuf[ERRBUFLEN];
int rc = 0;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot create '%s'"), path);
/*
* Check that we are not passing the nesting limit
* before we start creating any ancestors.
*/
if (dataset_nestcheck(path) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"maximum name nesting depth exceeded"));
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
}
if (check_parents(hdl, path, NULL, B_TRUE, &prefix) != 0)
return (-1);
if ((path_copy = strdup(path)) != NULL) {
rc = create_parents(hdl, path_copy, prefix);
free(path_copy);
}
if (path_copy == NULL || rc != 0)
return (-1);
return (0);
}
/*
* Create a new filesystem or volume.
*/
int
zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type,
nvlist_t *props)
{
int ret;
uint64_t size = 0;
uint64_t blocksize = zfs_prop_default_numeric(ZFS_PROP_VOLBLOCKSIZE);
uint64_t zoned;
enum lzc_dataset_type ost;
zpool_handle_t *zpool_handle;
uint8_t *wkeydata = NULL;
uint_t wkeylen = 0;
char errbuf[ERRBUFLEN];
char parent[ZFS_MAX_DATASET_NAME_LEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot create '%s'"), path);
/* validate the path, taking care to note the extended error message */
if (!zfs_validate_name(hdl, path, type, B_TRUE))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
if (dataset_nestcheck(path) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"maximum name nesting depth exceeded"));
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
}
/* validate parents exist */
if (check_parents(hdl, path, &zoned, B_FALSE, NULL) != 0)
return (-1);
/*
* The failure modes when creating a dataset of a different type over
* one that already exists is a little strange. In particular, if you
* try to create a dataset on top of an existing dataset, the ioctl()
* will return ENOENT, not EEXIST. To prevent this from happening, we
* first try to see if the dataset exists.
*/
if (zfs_dataset_exists(hdl, path, ZFS_TYPE_DATASET)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset already exists"));
return (zfs_error(hdl, EZFS_EXISTS, errbuf));
}
if (type == ZFS_TYPE_VOLUME)
ost = LZC_DATSET_TYPE_ZVOL;
else
ost = LZC_DATSET_TYPE_ZFS;
/* open zpool handle for prop validation */
char pool_path[ZFS_MAX_DATASET_NAME_LEN];
(void) strlcpy(pool_path, path, sizeof (pool_path));
/* truncate pool_path at first slash */
char *p = strchr(pool_path, '/');
if (p != NULL)
*p = '\0';
if ((zpool_handle = zpool_open(hdl, pool_path)) == NULL)
return (-1);
if (props && (props = zfs_valid_proplist(hdl, type, props,
zoned, NULL, zpool_handle, B_TRUE, errbuf)) == 0) {
zpool_close(zpool_handle);
return (-1);
}
zpool_close(zpool_handle);
if (type == ZFS_TYPE_VOLUME) {
/*
* If we are creating a volume, the size and block size must
* satisfy a few restraints. First, the blocksize must be a
* valid block size between SPA_{MIN,MAX}BLOCKSIZE. Second, the
* volsize must be a multiple of the block size, and cannot be
* zero.
*/
if (props == NULL || nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_VOLSIZE), &size) != 0) {
nvlist_free(props);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"missing volume size"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
}
if ((ret = nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE),
&blocksize)) != 0) {
if (ret == ENOENT) {
blocksize = zfs_prop_default_numeric(
ZFS_PROP_VOLBLOCKSIZE);
} else {
nvlist_free(props);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"missing volume block size"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
}
}
if (size == 0) {
nvlist_free(props);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"volume size cannot be zero"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
}
if (size % blocksize != 0) {
nvlist_free(props);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"volume size must be a multiple of volume block "
"size"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
}
}
(void) parent_name(path, parent, sizeof (parent));
if (zfs_crypto_create(hdl, parent, props, NULL, B_TRUE,
&wkeydata, &wkeylen) != 0) {
nvlist_free(props);
return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf));
}
/* create the dataset */
ret = lzc_create(path, ost, props, wkeydata, wkeylen);
nvlist_free(props);
if (wkeydata != NULL)
free(wkeydata);
/* check for failure */
if (ret != 0) {
switch (errno) {
case ENOENT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"no such parent '%s'"), parent);
return (zfs_error(hdl, EZFS_NOENT, errbuf));
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded to set this "
"property or value"));
return (zfs_error(hdl, EZFS_BADVERSION, errbuf));
case EACCES:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"encryption root's key is not loaded "
"or provided"));
return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf));
case ERANGE:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property value(s) specified"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
#ifdef _ILP32
case EOVERFLOW:
/*
* This platform can't address a volume this big.
*/
if (type == ZFS_TYPE_VOLUME)
return (zfs_error(hdl, EZFS_VOLTOOBIG,
errbuf));
zfs_fallthrough;
#endif
default:
return (zfs_standard_error(hdl, errno, errbuf));
}
}
return (0);
}
/*
* Destroys the given dataset. The caller must make sure that the filesystem
* isn't mounted, and that there are no active dependents. If the file system
* does not exist this function does nothing.
*/
int
zfs_destroy(zfs_handle_t *zhp, boolean_t defer)
{
int error;
if (zhp->zfs_type != ZFS_TYPE_SNAPSHOT && defer)
return (EINVAL);
if (zhp->zfs_type == ZFS_TYPE_BOOKMARK) {
nvlist_t *nv = fnvlist_alloc();
fnvlist_add_boolean(nv, zhp->zfs_name);
error = lzc_destroy_bookmarks(nv, NULL);
fnvlist_free(nv);
if (error != 0) {
return (zfs_standard_error_fmt(zhp->zfs_hdl, error,
dgettext(TEXT_DOMAIN, "cannot destroy '%s'"),
zhp->zfs_name));
}
return (0);
}
if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) {
nvlist_t *nv = fnvlist_alloc();
fnvlist_add_boolean(nv, zhp->zfs_name);
error = lzc_destroy_snaps(nv, defer, NULL);
fnvlist_free(nv);
} else {
error = lzc_destroy(zhp->zfs_name);
}
if (error != 0 && error != ENOENT) {
return (zfs_standard_error_fmt(zhp->zfs_hdl, errno,
dgettext(TEXT_DOMAIN, "cannot destroy '%s'"),
zhp->zfs_name));
}
remove_mountpoint(zhp);
return (0);
}
struct destroydata {
nvlist_t *nvl;
const char *snapname;
};
static int
zfs_check_snap_cb(zfs_handle_t *zhp, void *arg)
{
struct destroydata *dd = arg;
char name[ZFS_MAX_DATASET_NAME_LEN];
int rv = 0;
if (snprintf(name, sizeof (name), "%s@%s", zhp->zfs_name,
dd->snapname) >= sizeof (name))
return (EINVAL);
if (lzc_exists(name))
fnvlist_add_boolean(dd->nvl, name);
- rv = zfs_iter_filesystems(zhp, 0, zfs_check_snap_cb, dd);
+ rv = zfs_iter_filesystems_v2(zhp, 0, zfs_check_snap_cb, dd);
zfs_close(zhp);
return (rv);
}
/*
* Destroys all snapshots with the given name in zhp & descendants.
*/
int
zfs_destroy_snaps(zfs_handle_t *zhp, char *snapname, boolean_t defer)
{
int ret;
struct destroydata dd = { 0 };
dd.snapname = snapname;
dd.nvl = fnvlist_alloc();
(void) zfs_check_snap_cb(zfs_handle_dup(zhp), &dd);
if (nvlist_empty(dd.nvl)) {
ret = zfs_standard_error_fmt(zhp->zfs_hdl, ENOENT,
dgettext(TEXT_DOMAIN, "cannot destroy '%s@%s'"),
zhp->zfs_name, snapname);
} else {
ret = zfs_destroy_snaps_nvl(zhp->zfs_hdl, dd.nvl, defer);
}
fnvlist_free(dd.nvl);
return (ret);
}
/*
* Destroys all the snapshots named in the nvlist.
*/
int
zfs_destroy_snaps_nvl(libzfs_handle_t *hdl, nvlist_t *snaps, boolean_t defer)
{
nvlist_t *errlist = NULL;
nvpair_t *pair;
int ret = zfs_destroy_snaps_nvl_os(hdl, snaps);
if (ret != 0)
return (ret);
ret = lzc_destroy_snaps(snaps, defer, &errlist);
if (ret == 0) {
nvlist_free(errlist);
return (0);
}
if (nvlist_empty(errlist)) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot destroy snapshots"));
ret = zfs_standard_error(hdl, ret, errbuf);
}
for (pair = nvlist_next_nvpair(errlist, NULL);
pair != NULL; pair = nvlist_next_nvpair(errlist, pair)) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot destroy snapshot %s"),
nvpair_name(pair));
switch (fnvpair_value_int32(pair)) {
case EEXIST:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "snapshot is cloned"));
ret = zfs_error(hdl, EZFS_EXISTS, errbuf);
break;
default:
ret = zfs_standard_error(hdl, errno, errbuf);
break;
}
}
nvlist_free(errlist);
return (ret);
}
/*
* Clones the given dataset. The target must be of the same type as the source.
*/
int
zfs_clone(zfs_handle_t *zhp, const char *target, nvlist_t *props)
{
char parent[ZFS_MAX_DATASET_NAME_LEN];
int ret;
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zfs_hdl;
uint64_t zoned;
assert(zhp->zfs_type == ZFS_TYPE_SNAPSHOT);
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot create '%s'"), target);
/* validate the target/clone name */
if (!zfs_validate_name(hdl, target, ZFS_TYPE_FILESYSTEM, B_TRUE))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
/* validate parents exist */
if (check_parents(hdl, target, &zoned, B_FALSE, NULL) != 0)
return (-1);
(void) parent_name(target, parent, sizeof (parent));
/* do the clone */
if (props) {
zfs_type_t type = ZFS_TYPE_FILESYSTEM;
if (ZFS_IS_VOLUME(zhp))
type = ZFS_TYPE_VOLUME;
if ((props = zfs_valid_proplist(hdl, type, props, zoned,
zhp, zhp->zpool_hdl, B_TRUE, errbuf)) == NULL)
return (-1);
if (zfs_fix_auto_resv(zhp, props) == -1) {
nvlist_free(props);
return (-1);
}
}
if (zfs_crypto_clone_check(hdl, zhp, parent, props) != 0) {
nvlist_free(props);
return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf));
}
ret = lzc_clone(target, zhp->zfs_name, props);
nvlist_free(props);
if (ret != 0) {
switch (errno) {
case ENOENT:
/*
* The parent doesn't exist. We should have caught this
* above, but there may a race condition that has since
* destroyed the parent.
*
* At this point, we don't know whether it's the source
* that doesn't exist anymore, or whether the target
* dataset doesn't exist.
*/
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"no such parent '%s'"), parent);
return (zfs_error(zhp->zfs_hdl, EZFS_NOENT, errbuf));
case EXDEV:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"source and target pools differ"));
return (zfs_error(zhp->zfs_hdl, EZFS_CROSSTARGET,
errbuf));
default:
return (zfs_standard_error(zhp->zfs_hdl, errno,
errbuf));
}
}
return (ret);
}
/*
* Promotes the given clone fs to be the clone parent.
*/
int
zfs_promote(zfs_handle_t *zhp)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
char snapname[ZFS_MAX_DATASET_NAME_LEN];
int ret;
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot promote '%s'"), zhp->zfs_name);
if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"snapshots can not be promoted"));
return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
}
if (zhp->zfs_dmustats.dds_origin[0] == '\0') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"not a cloned filesystem"));
return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
}
if (!zfs_validate_name(hdl, zhp->zfs_name, zhp->zfs_type, B_TRUE))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
ret = lzc_promote(zhp->zfs_name, snapname, sizeof (snapname));
if (ret != 0) {
switch (ret) {
case EACCES:
/*
* Promoting encrypted dataset outside its
* encryption root.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot promote dataset outside its "
"encryption root"));
return (zfs_error(hdl, EZFS_EXISTS, errbuf));
case EEXIST:
/* There is a conflicting snapshot name. */
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"conflicting snapshot '%s' from parent '%s'"),
snapname, zhp->zfs_dmustats.dds_origin);
return (zfs_error(hdl, EZFS_EXISTS, errbuf));
default:
return (zfs_standard_error(hdl, ret, errbuf));
}
}
return (ret);
}
typedef struct snapdata {
nvlist_t *sd_nvl;
const char *sd_snapname;
} snapdata_t;
static int
zfs_snapshot_cb(zfs_handle_t *zhp, void *arg)
{
snapdata_t *sd = arg;
char name[ZFS_MAX_DATASET_NAME_LEN];
int rv = 0;
if (zfs_prop_get_int(zhp, ZFS_PROP_INCONSISTENT) == 0) {
if (snprintf(name, sizeof (name), "%s@%s", zfs_get_name(zhp),
sd->sd_snapname) >= sizeof (name))
return (EINVAL);
fnvlist_add_boolean(sd->sd_nvl, name);
- rv = zfs_iter_filesystems(zhp, 0, zfs_snapshot_cb, sd);
+ rv = zfs_iter_filesystems_v2(zhp, 0, zfs_snapshot_cb, sd);
}
zfs_close(zhp);
return (rv);
}
/*
* Creates snapshots. The keys in the snaps nvlist are the snapshots to be
* created.
*/
int
zfs_snapshot_nvl(libzfs_handle_t *hdl, nvlist_t *snaps, nvlist_t *props)
{
int ret;
char errbuf[ERRBUFLEN];
nvpair_t *elem;
nvlist_t *errors;
zpool_handle_t *zpool_hdl;
char pool[ZFS_MAX_DATASET_NAME_LEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot create snapshots "));
elem = NULL;
while ((elem = nvlist_next_nvpair(snaps, elem)) != NULL) {
const char *snapname = nvpair_name(elem);
/* validate the target name */
if (!zfs_validate_name(hdl, snapname, ZFS_TYPE_SNAPSHOT,
B_TRUE)) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot create snapshot '%s'"), snapname);
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
}
}
/*
* get pool handle for prop validation. assumes all snaps are in the
* same pool, as does lzc_snapshot (below).
*/
elem = nvlist_next_nvpair(snaps, NULL);
if (elem == NULL)
return (-1);
(void) strlcpy(pool, nvpair_name(elem), sizeof (pool));
pool[strcspn(pool, "/@")] = '\0';
zpool_hdl = zpool_open(hdl, pool);
if (zpool_hdl == NULL)
return (-1);
if (props != NULL &&
(props = zfs_valid_proplist(hdl, ZFS_TYPE_SNAPSHOT,
props, B_FALSE, NULL, zpool_hdl, B_FALSE, errbuf)) == NULL) {
zpool_close(zpool_hdl);
return (-1);
}
zpool_close(zpool_hdl);
ret = lzc_snapshot(snaps, props, &errors);
if (ret != 0) {
boolean_t printed = B_FALSE;
for (elem = nvlist_next_nvpair(errors, NULL);
elem != NULL;
elem = nvlist_next_nvpair(errors, elem)) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot create snapshot '%s'"), nvpair_name(elem));
(void) zfs_standard_error(hdl,
fnvpair_value_int32(elem), errbuf);
printed = B_TRUE;
}
if (!printed) {
switch (ret) {
case EXDEV:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"multiple snapshots of same "
"fs not allowed"));
(void) zfs_error(hdl, EZFS_EXISTS, errbuf);
break;
default:
(void) zfs_standard_error(hdl, ret, errbuf);
}
}
}
nvlist_free(props);
nvlist_free(errors);
return (ret);
}
int
zfs_snapshot(libzfs_handle_t *hdl, const char *path, boolean_t recursive,
nvlist_t *props)
{
int ret;
snapdata_t sd = { 0 };
char fsname[ZFS_MAX_DATASET_NAME_LEN];
char *cp;
zfs_handle_t *zhp;
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot snapshot %s"), path);
if (!zfs_validate_name(hdl, path, ZFS_TYPE_SNAPSHOT, B_TRUE))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
(void) strlcpy(fsname, path, sizeof (fsname));
cp = strchr(fsname, '@');
*cp = '\0';
sd.sd_snapname = cp + 1;
if ((zhp = zfs_open(hdl, fsname, ZFS_TYPE_FILESYSTEM |
ZFS_TYPE_VOLUME)) == NULL) {
return (-1);
}
sd.sd_nvl = fnvlist_alloc();
if (recursive) {
(void) zfs_snapshot_cb(zfs_handle_dup(zhp), &sd);
} else {
fnvlist_add_boolean(sd.sd_nvl, path);
}
ret = zfs_snapshot_nvl(hdl, sd.sd_nvl, props);
fnvlist_free(sd.sd_nvl);
zfs_close(zhp);
return (ret);
}
/*
* Destroy any more recent snapshots. We invoke this callback on any dependents
* of the snapshot first. If the 'cb_dependent' member is non-zero, then this
* is a dependent and we should just destroy it without checking the transaction
* group.
*/
typedef struct rollback_data {
const char *cb_target; /* the snapshot */
uint64_t cb_create; /* creation time reference */
boolean_t cb_error;
boolean_t cb_force;
} rollback_data_t;
static int
rollback_destroy_dependent(zfs_handle_t *zhp, void *data)
{
rollback_data_t *cbp = data;
prop_changelist_t *clp;
/* We must destroy this clone; first unmount it */
clp = changelist_gather(zhp, ZFS_PROP_NAME, 0,
cbp->cb_force ? MS_FORCE: 0);
if (clp == NULL || changelist_prefix(clp) != 0) {
cbp->cb_error = B_TRUE;
zfs_close(zhp);
return (0);
}
if (zfs_destroy(zhp, B_FALSE) != 0)
cbp->cb_error = B_TRUE;
else
changelist_remove(clp, zhp->zfs_name);
(void) changelist_postfix(clp);
changelist_free(clp);
zfs_close(zhp);
return (0);
}
static int
rollback_destroy(zfs_handle_t *zhp, void *data)
{
rollback_data_t *cbp = data;
if (zfs_prop_get_int(zhp, ZFS_PROP_CREATETXG) > cbp->cb_create) {
- cbp->cb_error |= zfs_iter_dependents(zhp, 0, B_FALSE,
+ cbp->cb_error |= zfs_iter_dependents_v2(zhp, 0, B_FALSE,
rollback_destroy_dependent, cbp);
cbp->cb_error |= zfs_destroy(zhp, B_FALSE);
}
zfs_close(zhp);
return (0);
}
/*
* Given a dataset, rollback to a specific snapshot, discarding any
* data changes since then and making it the active dataset.
*
* Any snapshots and bookmarks more recent than the target are
* destroyed, along with their dependents (i.e. clones).
*/
int
zfs_rollback(zfs_handle_t *zhp, zfs_handle_t *snap, boolean_t force)
{
rollback_data_t cb = { 0 };
int err;
boolean_t restore_resv = 0;
uint64_t old_volsize = 0, new_volsize;
zfs_prop_t resv_prop = { 0 };
uint64_t min_txg = 0;
assert(zhp->zfs_type == ZFS_TYPE_FILESYSTEM ||
zhp->zfs_type == ZFS_TYPE_VOLUME);
/*
* Destroy all recent snapshots and their dependents.
*/
cb.cb_force = force;
cb.cb_target = snap->zfs_name;
cb.cb_create = zfs_prop_get_int(snap, ZFS_PROP_CREATETXG);
if (cb.cb_create > 0)
min_txg = cb.cb_create;
- (void) zfs_iter_snapshots(zhp, 0, rollback_destroy, &cb,
+ (void) zfs_iter_snapshots_v2(zhp, 0, rollback_destroy, &cb,
min_txg, 0);
- (void) zfs_iter_bookmarks(zhp, 0, rollback_destroy, &cb);
+ (void) zfs_iter_bookmarks_v2(zhp, 0, rollback_destroy, &cb);
if (cb.cb_error)
return (-1);
/*
* Now that we have verified that the snapshot is the latest,
* rollback to the given snapshot.
*/
if (zhp->zfs_type == ZFS_TYPE_VOLUME) {
if (zfs_which_resv_prop(zhp, &resv_prop) < 0)
return (-1);
old_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE);
restore_resv =
(old_volsize == zfs_prop_get_int(zhp, resv_prop));
}
/*
* Pass both the filesystem and the wanted snapshot names,
* we would get an error back if the snapshot is destroyed or
* a new snapshot is created before this request is processed.
*/
err = lzc_rollback_to(zhp->zfs_name, snap->zfs_name);
if (err != 0) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot rollback '%s'"),
zhp->zfs_name);
switch (err) {
case EEXIST:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"there is a snapshot or bookmark more recent "
"than '%s'"), snap->zfs_name);
(void) zfs_error(zhp->zfs_hdl, EZFS_EXISTS, errbuf);
break;
case ESRCH:
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"'%s' is not found among snapshots of '%s'"),
snap->zfs_name, zhp->zfs_name);
(void) zfs_error(zhp->zfs_hdl, EZFS_NOENT, errbuf);
break;
case EINVAL:
(void) zfs_error(zhp->zfs_hdl, EZFS_BADTYPE, errbuf);
break;
default:
(void) zfs_standard_error(zhp->zfs_hdl, err, errbuf);
}
return (err);
}
/*
* For volumes, if the pre-rollback volsize matched the pre-
* rollback reservation and the volsize has changed then set
* the reservation property to the post-rollback volsize.
* Make a new handle since the rollback closed the dataset.
*/
if ((zhp->zfs_type == ZFS_TYPE_VOLUME) &&
(zhp = make_dataset_handle(zhp->zfs_hdl, zhp->zfs_name))) {
if (restore_resv) {
new_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE);
if (old_volsize != new_volsize)
err = zfs_prop_set_int(zhp, resv_prop,
new_volsize);
}
zfs_close(zhp);
}
return (err);
}
/*
* Renames the given dataset.
*/
int
zfs_rename(zfs_handle_t *zhp, const char *target, renameflags_t flags)
{
int ret = 0;
zfs_cmd_t zc = {"\0"};
char *delim;
prop_changelist_t *cl = NULL;
char parent[ZFS_MAX_DATASET_NAME_LEN];
char property[ZFS_MAXPROPLEN];
libzfs_handle_t *hdl = zhp->zfs_hdl;
char errbuf[ERRBUFLEN];
/* if we have the same exact name, just return success */
if (strcmp(zhp->zfs_name, target) == 0)
return (0);
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot rename to '%s'"), target);
/* make sure source name is valid */
if (!zfs_validate_name(hdl, zhp->zfs_name, zhp->zfs_type, B_TRUE))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
/*
* Make sure the target name is valid
*/
if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) {
if ((strchr(target, '@') == NULL) ||
*target == '@') {
/*
* Snapshot target name is abbreviated,
* reconstruct full dataset name
*/
(void) strlcpy(parent, zhp->zfs_name,
sizeof (parent));
delim = strchr(parent, '@');
if (strchr(target, '@') == NULL)
*(++delim) = '\0';
else
*delim = '\0';
(void) strlcat(parent, target, sizeof (parent));
target = parent;
} else {
/*
* Make sure we're renaming within the same dataset.
*/
delim = strchr(target, '@');
if (strncmp(zhp->zfs_name, target, delim - target)
!= 0 || zhp->zfs_name[delim - target] != '@') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"snapshots must be part of same "
"dataset"));
return (zfs_error(hdl, EZFS_CROSSTARGET,
errbuf));
}
}
if (!zfs_validate_name(hdl, target, zhp->zfs_type, B_TRUE))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
} else {
if (flags.recursive) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"recursive rename must be a snapshot"));
return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
}
if (!zfs_validate_name(hdl, target, zhp->zfs_type, B_TRUE))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
/* validate parents */
if (check_parents(hdl, target, NULL, B_FALSE, NULL) != 0)
return (-1);
/* make sure we're in the same pool */
verify((delim = strchr(target, '/')) != NULL);
if (strncmp(zhp->zfs_name, target, delim - target) != 0 ||
zhp->zfs_name[delim - target] != '/') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"datasets must be within same pool"));
return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf));
}
/* new name cannot be a child of the current dataset name */
if (is_descendant(zhp->zfs_name, target)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"New dataset name cannot be a descendant of "
"current dataset name"));
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
}
}
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot rename '%s'"), zhp->zfs_name);
if (getzoneid() == GLOBAL_ZONEID &&
zfs_prop_get_int(zhp, ZFS_PROP_ZONED)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset is used in a non-global zone"));
return (zfs_error(hdl, EZFS_ZONED, errbuf));
}
/*
* Avoid unmounting file systems with mountpoint property set to
* 'legacy' or 'none' even if -u option is not given.
*/
if (zhp->zfs_type == ZFS_TYPE_FILESYSTEM &&
!flags.recursive && !flags.nounmount &&
zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, property,
sizeof (property), NULL, NULL, 0, B_FALSE) == 0 &&
(strcmp(property, "legacy") == 0 ||
strcmp(property, "none") == 0)) {
flags.nounmount = B_TRUE;
}
if (flags.recursive) {
char *parentname = zfs_strdup(zhp->zfs_hdl, zhp->zfs_name);
delim = strchr(parentname, '@');
*delim = '\0';
zfs_handle_t *zhrp = zfs_open(zhp->zfs_hdl, parentname,
ZFS_TYPE_DATASET);
free(parentname);
if (zhrp == NULL) {
ret = -1;
goto error;
}
zfs_close(zhrp);
} else if (zhp->zfs_type != ZFS_TYPE_SNAPSHOT) {
if ((cl = changelist_gather(zhp, ZFS_PROP_NAME,
flags.nounmount ? CL_GATHER_DONT_UNMOUNT :
CL_GATHER_ITER_MOUNTED,
flags.forceunmount ? MS_FORCE : 0)) == NULL)
return (-1);
if (changelist_haszonedchild(cl)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"child dataset with inherited mountpoint is used "
"in a non-global zone"));
(void) zfs_error(hdl, EZFS_ZONED, errbuf);
ret = -1;
goto error;
}
if ((ret = changelist_prefix(cl)) != 0)
goto error;
}
if (ZFS_IS_VOLUME(zhp))
zc.zc_objset_type = DMU_OST_ZVOL;
else
zc.zc_objset_type = DMU_OST_ZFS;
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
(void) strlcpy(zc.zc_value, target, sizeof (zc.zc_value));
zc.zc_cookie = !!flags.recursive;
zc.zc_cookie |= (!!flags.nounmount) << 1;
if ((ret = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_RENAME, &zc)) != 0) {
/*
* if it was recursive, the one that actually failed will
* be in zc.zc_name
*/
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot rename '%s'"), zc.zc_name);
if (flags.recursive && errno == EEXIST) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"a child dataset already has a snapshot "
"with the new name"));
(void) zfs_error(hdl, EZFS_EXISTS, errbuf);
} else if (errno == EACCES) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot move encrypted child outside of "
"its encryption root"));
(void) zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf);
} else {
(void) zfs_standard_error(zhp->zfs_hdl, errno, errbuf);
}
/*
* On failure, we still want to remount any filesystems that
* were previously mounted, so we don't alter the system state.
*/
if (cl != NULL)
(void) changelist_postfix(cl);
} else {
if (cl != NULL) {
changelist_rename(cl, zfs_get_name(zhp), target);
ret = changelist_postfix(cl);
}
}
error:
if (cl != NULL) {
changelist_free(cl);
}
return (ret);
}
nvlist_t *
zfs_get_all_props(zfs_handle_t *zhp)
{
return (zhp->zfs_props);
}
nvlist_t *
zfs_get_recvd_props(zfs_handle_t *zhp)
{
if (zhp->zfs_recvd_props == NULL)
if (get_recvd_props_ioctl(zhp) != 0)
return (NULL);
return (zhp->zfs_recvd_props);
}
nvlist_t *
zfs_get_user_props(zfs_handle_t *zhp)
{
return (zhp->zfs_user_props);
}
/*
* This function is used by 'zfs list' to determine the exact set of columns to
* display, and their maximum widths. This does two main things:
*
* - If this is a list of all properties, then expand the list to include
* all native properties, and set a flag so that for each dataset we look
* for new unique user properties and add them to the list.
*
* - For non fixed-width properties, keep track of the maximum width seen
* so that we can size the column appropriately. If the user has
* requested received property values, we also need to compute the width
* of the RECEIVED column.
*/
int
zfs_expand_proplist(zfs_handle_t *zhp, zprop_list_t **plp, boolean_t received,
boolean_t literal)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
zprop_list_t *entry;
zprop_list_t **last, **start;
nvlist_t *userprops, *propval;
nvpair_t *elem;
const char *strval;
char buf[ZFS_MAXPROPLEN];
if (zprop_expand_list(hdl, plp, ZFS_TYPE_DATASET) != 0)
return (-1);
userprops = zfs_get_user_props(zhp);
entry = *plp;
if (entry->pl_all && nvlist_next_nvpair(userprops, NULL) != NULL) {
/*
* Go through and add any user properties as necessary. We
* start by incrementing our list pointer to the first
* non-native property.
*/
start = plp;
while (*start != NULL) {
if ((*start)->pl_prop == ZPROP_USERPROP)
break;
start = &(*start)->pl_next;
}
elem = NULL;
while ((elem = nvlist_next_nvpair(userprops, elem)) != NULL) {
/*
* See if we've already found this property in our list.
*/
for (last = start; *last != NULL;
last = &(*last)->pl_next) {
if (strcmp((*last)->pl_user_prop,
nvpair_name(elem)) == 0)
break;
}
if (*last == NULL) {
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_user_prop =
zfs_strdup(hdl, nvpair_name(elem));
entry->pl_prop = ZPROP_USERPROP;
entry->pl_width = strlen(nvpair_name(elem));
entry->pl_all = B_TRUE;
*last = entry;
}
}
}
/*
* Now go through and check the width of any non-fixed columns
*/
for (entry = *plp; entry != NULL; entry = entry->pl_next) {
if (entry->pl_fixed && !literal)
continue;
if (entry->pl_prop != ZPROP_USERPROP) {
if (zfs_prop_get(zhp, entry->pl_prop,
buf, sizeof (buf), NULL, NULL, 0, literal) == 0) {
if (strlen(buf) > entry->pl_width)
entry->pl_width = strlen(buf);
}
if (received && zfs_prop_get_recvd(zhp,
zfs_prop_to_name(entry->pl_prop),
buf, sizeof (buf), literal) == 0)
if (strlen(buf) > entry->pl_recvd_width)
entry->pl_recvd_width = strlen(buf);
} else {
if (nvlist_lookup_nvlist(userprops, entry->pl_user_prop,
&propval) == 0) {
strval = fnvlist_lookup_string(propval,
ZPROP_VALUE);
if (strlen(strval) > entry->pl_width)
entry->pl_width = strlen(strval);
}
if (received && zfs_prop_get_recvd(zhp,
entry->pl_user_prop,
buf, sizeof (buf), literal) == 0)
if (strlen(buf) > entry->pl_recvd_width)
entry->pl_recvd_width = strlen(buf);
}
}
return (0);
}
void
zfs_prune_proplist(zfs_handle_t *zhp, uint8_t *props)
{
nvpair_t *curr;
nvpair_t *next;
/*
* Keep a reference to the props-table against which we prune the
* properties.
*/
zhp->zfs_props_table = props;
curr = nvlist_next_nvpair(zhp->zfs_props, NULL);
while (curr) {
zfs_prop_t zfs_prop = zfs_name_to_prop(nvpair_name(curr));
next = nvlist_next_nvpair(zhp->zfs_props, curr);
/*
* User properties will result in ZPROP_USERPROP (an alias
* for ZPROP_INVAL), and since we
* only know how to prune standard ZFS properties, we always
* leave these in the list. This can also happen if we
* encounter an unknown DSL property (when running older
* software, for example).
*/
if (zfs_prop != ZPROP_USERPROP && props[zfs_prop] == B_FALSE)
(void) nvlist_remove(zhp->zfs_props,
nvpair_name(curr), nvpair_type(curr));
curr = next;
}
}
static int
zfs_smb_acl_mgmt(libzfs_handle_t *hdl, char *dataset, char *path,
zfs_smb_acl_op_t cmd, char *resource1, char *resource2)
{
zfs_cmd_t zc = {"\0"};
nvlist_t *nvlist = NULL;
int error;
(void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name));
(void) strlcpy(zc.zc_value, path, sizeof (zc.zc_value));
zc.zc_cookie = (uint64_t)cmd;
if (cmd == ZFS_SMB_ACL_RENAME) {
if (nvlist_alloc(&nvlist, NV_UNIQUE_NAME, 0) != 0) {
(void) no_memory(hdl);
return (0);
}
}
switch (cmd) {
case ZFS_SMB_ACL_ADD:
case ZFS_SMB_ACL_REMOVE:
(void) strlcpy(zc.zc_string, resource1, sizeof (zc.zc_string));
break;
case ZFS_SMB_ACL_RENAME:
if (nvlist_add_string(nvlist, ZFS_SMB_ACL_SRC,
resource1) != 0) {
(void) no_memory(hdl);
return (-1);
}
if (nvlist_add_string(nvlist, ZFS_SMB_ACL_TARGET,
resource2) != 0) {
(void) no_memory(hdl);
return (-1);
}
zcmd_write_src_nvlist(hdl, &zc, nvlist);
break;
case ZFS_SMB_ACL_PURGE:
break;
default:
return (-1);
}
error = ioctl(hdl->libzfs_fd, ZFS_IOC_SMB_ACL, &zc);
nvlist_free(nvlist);
return (error);
}
int
zfs_smb_acl_add(libzfs_handle_t *hdl, char *dataset,
char *path, char *resource)
{
return (zfs_smb_acl_mgmt(hdl, dataset, path, ZFS_SMB_ACL_ADD,
resource, NULL));
}
int
zfs_smb_acl_remove(libzfs_handle_t *hdl, char *dataset,
char *path, char *resource)
{
return (zfs_smb_acl_mgmt(hdl, dataset, path, ZFS_SMB_ACL_REMOVE,
resource, NULL));
}
int
zfs_smb_acl_purge(libzfs_handle_t *hdl, char *dataset, char *path)
{
return (zfs_smb_acl_mgmt(hdl, dataset, path, ZFS_SMB_ACL_PURGE,
NULL, NULL));
}
int
zfs_smb_acl_rename(libzfs_handle_t *hdl, char *dataset, char *path,
char *oldname, char *newname)
{
return (zfs_smb_acl_mgmt(hdl, dataset, path, ZFS_SMB_ACL_RENAME,
oldname, newname));
}
int
zfs_userspace(zfs_handle_t *zhp, zfs_userquota_prop_t type,
zfs_userspace_cb_t func, void *arg)
{
zfs_cmd_t zc = {"\0"};
zfs_useracct_t buf[100];
libzfs_handle_t *hdl = zhp->zfs_hdl;
int ret;
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
zc.zc_objset_type = type;
zc.zc_nvlist_dst = (uintptr_t)buf;
for (;;) {
zfs_useracct_t *zua = buf;
zc.zc_nvlist_dst_size = sizeof (buf);
if (zfs_ioctl(hdl, ZFS_IOC_USERSPACE_MANY, &zc) != 0) {
if ((errno == ENOTSUP &&
(type == ZFS_PROP_USEROBJUSED ||
type == ZFS_PROP_GROUPOBJUSED ||
type == ZFS_PROP_USEROBJQUOTA ||
type == ZFS_PROP_GROUPOBJQUOTA ||
type == ZFS_PROP_PROJECTOBJUSED ||
type == ZFS_PROP_PROJECTOBJQUOTA ||
type == ZFS_PROP_PROJECTUSED ||
type == ZFS_PROP_PROJECTQUOTA)))
break;
return (zfs_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN,
"cannot get used/quota for %s"), zc.zc_name));
}
if (zc.zc_nvlist_dst_size == 0)
break;
while (zc.zc_nvlist_dst_size > 0) {
if ((ret = func(arg, zua->zu_domain, zua->zu_rid,
zua->zu_space)) != 0)
return (ret);
zua++;
zc.zc_nvlist_dst_size -= sizeof (zfs_useracct_t);
}
}
return (0);
}
struct holdarg {
nvlist_t *nvl;
const char *snapname;
const char *tag;
boolean_t recursive;
int error;
};
static int
zfs_hold_one(zfs_handle_t *zhp, void *arg)
{
struct holdarg *ha = arg;
char name[ZFS_MAX_DATASET_NAME_LEN];
int rv = 0;
if (snprintf(name, sizeof (name), "%s@%s", zhp->zfs_name,
ha->snapname) >= sizeof (name))
return (EINVAL);
if (lzc_exists(name))
fnvlist_add_string(ha->nvl, name, ha->tag);
if (ha->recursive)
- rv = zfs_iter_filesystems(zhp, 0, zfs_hold_one, ha);
+ rv = zfs_iter_filesystems_v2(zhp, 0, zfs_hold_one, ha);
zfs_close(zhp);
return (rv);
}
int
zfs_hold(zfs_handle_t *zhp, const char *snapname, const char *tag,
boolean_t recursive, int cleanup_fd)
{
int ret;
struct holdarg ha;
ha.nvl = fnvlist_alloc();
ha.snapname = snapname;
ha.tag = tag;
ha.recursive = recursive;
(void) zfs_hold_one(zfs_handle_dup(zhp), &ha);
if (nvlist_empty(ha.nvl)) {
char errbuf[ERRBUFLEN];
fnvlist_free(ha.nvl);
ret = ENOENT;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot hold snapshot '%s@%s'"),
zhp->zfs_name, snapname);
(void) zfs_standard_error(zhp->zfs_hdl, ret, errbuf);
return (ret);
}
ret = zfs_hold_nvl(zhp, cleanup_fd, ha.nvl);
fnvlist_free(ha.nvl);
return (ret);
}
int
zfs_hold_nvl(zfs_handle_t *zhp, int cleanup_fd, nvlist_t *holds)
{
int ret;
nvlist_t *errors;
libzfs_handle_t *hdl = zhp->zfs_hdl;
char errbuf[ERRBUFLEN];
nvpair_t *elem;
errors = NULL;
ret = lzc_hold(holds, cleanup_fd, &errors);
if (ret == 0) {
/* There may be errors even in the success case. */
fnvlist_free(errors);
return (0);
}
if (nvlist_empty(errors)) {
/* no hold-specific errors */
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot hold"));
switch (ret) {
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded"));
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
case EINVAL:
(void) zfs_error(hdl, EZFS_BADTYPE, errbuf);
break;
default:
(void) zfs_standard_error(hdl, ret, errbuf);
}
}
for (elem = nvlist_next_nvpair(errors, NULL);
elem != NULL;
elem = nvlist_next_nvpair(errors, elem)) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot hold snapshot '%s'"), nvpair_name(elem));
switch (fnvpair_value_int32(elem)) {
case E2BIG:
/*
* Temporary tags wind up having the ds object id
* prepended. So even if we passed the length check
* above, it's still possible for the tag to wind
* up being slightly too long.
*/
(void) zfs_error(hdl, EZFS_TAGTOOLONG, errbuf);
break;
case EINVAL:
(void) zfs_error(hdl, EZFS_BADTYPE, errbuf);
break;
case EEXIST:
(void) zfs_error(hdl, EZFS_REFTAG_HOLD, errbuf);
break;
default:
(void) zfs_standard_error(hdl,
fnvpair_value_int32(elem), errbuf);
}
}
fnvlist_free(errors);
return (ret);
}
static int
zfs_release_one(zfs_handle_t *zhp, void *arg)
{
struct holdarg *ha = arg;
char name[ZFS_MAX_DATASET_NAME_LEN];
int rv = 0;
nvlist_t *existing_holds;
if (snprintf(name, sizeof (name), "%s@%s", zhp->zfs_name,
ha->snapname) >= sizeof (name)) {
ha->error = EINVAL;
rv = EINVAL;
}
if (lzc_get_holds(name, &existing_holds) != 0) {
ha->error = ENOENT;
} else if (!nvlist_exists(existing_holds, ha->tag)) {
ha->error = ESRCH;
} else {
nvlist_t *torelease = fnvlist_alloc();
fnvlist_add_boolean(torelease, ha->tag);
fnvlist_add_nvlist(ha->nvl, name, torelease);
fnvlist_free(torelease);
}
if (ha->recursive)
- rv = zfs_iter_filesystems(zhp, 0, zfs_release_one, ha);
+ rv = zfs_iter_filesystems_v2(zhp, 0, zfs_release_one, ha);
zfs_close(zhp);
return (rv);
}
int
zfs_release(zfs_handle_t *zhp, const char *snapname, const char *tag,
boolean_t recursive)
{
int ret;
struct holdarg ha;
nvlist_t *errors = NULL;
nvpair_t *elem;
libzfs_handle_t *hdl = zhp->zfs_hdl;
char errbuf[ERRBUFLEN];
ha.nvl = fnvlist_alloc();
ha.snapname = snapname;
ha.tag = tag;
ha.recursive = recursive;
ha.error = 0;
(void) zfs_release_one(zfs_handle_dup(zhp), &ha);
if (nvlist_empty(ha.nvl)) {
fnvlist_free(ha.nvl);
ret = ha.error;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot release hold from snapshot '%s@%s'"),
zhp->zfs_name, snapname);
if (ret == ESRCH) {
(void) zfs_error(hdl, EZFS_REFTAG_RELE, errbuf);
} else {
(void) zfs_standard_error(hdl, ret, errbuf);
}
return (ret);
}
ret = lzc_release(ha.nvl, &errors);
fnvlist_free(ha.nvl);
if (ret == 0) {
/* There may be errors even in the success case. */
fnvlist_free(errors);
return (0);
}
if (nvlist_empty(errors)) {
/* no hold-specific errors */
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot release"));
switch (errno) {
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded"));
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
default:
(void) zfs_standard_error(hdl, errno, errbuf);
}
}
for (elem = nvlist_next_nvpair(errors, NULL);
elem != NULL;
elem = nvlist_next_nvpair(errors, elem)) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot release hold from snapshot '%s'"),
nvpair_name(elem));
switch (fnvpair_value_int32(elem)) {
case ESRCH:
(void) zfs_error(hdl, EZFS_REFTAG_RELE, errbuf);
break;
case EINVAL:
(void) zfs_error(hdl, EZFS_BADTYPE, errbuf);
break;
default:
(void) zfs_standard_error(hdl,
fnvpair_value_int32(elem), errbuf);
}
}
fnvlist_free(errors);
return (ret);
}
int
zfs_get_fsacl(zfs_handle_t *zhp, nvlist_t **nvl)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zfs_hdl;
int nvsz = 2048;
void *nvbuf;
int err = 0;
char errbuf[ERRBUFLEN];
assert(zhp->zfs_type == ZFS_TYPE_VOLUME ||
zhp->zfs_type == ZFS_TYPE_FILESYSTEM);
tryagain:
nvbuf = malloc(nvsz);
if (nvbuf == NULL) {
err = (zfs_error(hdl, EZFS_NOMEM, strerror(errno)));
goto out;
}
zc.zc_nvlist_dst_size = nvsz;
zc.zc_nvlist_dst = (uintptr_t)nvbuf;
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
if (zfs_ioctl(hdl, ZFS_IOC_GET_FSACL, &zc) != 0) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot get permissions on '%s'"),
zc.zc_name);
switch (errno) {
case ENOMEM:
free(nvbuf);
nvsz = zc.zc_nvlist_dst_size;
goto tryagain;
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded"));
err = zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
case EINVAL:
err = zfs_error(hdl, EZFS_BADTYPE, errbuf);
break;
case ENOENT:
err = zfs_error(hdl, EZFS_NOENT, errbuf);
break;
default:
err = zfs_standard_error(hdl, errno, errbuf);
break;
}
} else {
/* success */
int rc = nvlist_unpack(nvbuf, zc.zc_nvlist_dst_size, nvl, 0);
if (rc) {
err = zfs_standard_error_fmt(hdl, rc, dgettext(
TEXT_DOMAIN, "cannot get permissions on '%s'"),
zc.zc_name);
}
}
free(nvbuf);
out:
return (err);
}
int
zfs_set_fsacl(zfs_handle_t *zhp, boolean_t un, nvlist_t *nvl)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zfs_hdl;
char *nvbuf;
char errbuf[ERRBUFLEN];
size_t nvsz;
int err;
assert(zhp->zfs_type == ZFS_TYPE_VOLUME ||
zhp->zfs_type == ZFS_TYPE_FILESYSTEM);
err = nvlist_size(nvl, &nvsz, NV_ENCODE_NATIVE);
assert(err == 0);
nvbuf = malloc(nvsz);
err = nvlist_pack(nvl, &nvbuf, &nvsz, NV_ENCODE_NATIVE, 0);
assert(err == 0);
zc.zc_nvlist_src_size = nvsz;
zc.zc_nvlist_src = (uintptr_t)nvbuf;
zc.zc_perm_action = un;
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
if (zfs_ioctl(hdl, ZFS_IOC_SET_FSACL, &zc) != 0) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot set permissions on '%s'"),
zc.zc_name);
switch (errno) {
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded"));
err = zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
case EINVAL:
err = zfs_error(hdl, EZFS_BADTYPE, errbuf);
break;
case ENOENT:
err = zfs_error(hdl, EZFS_NOENT, errbuf);
break;
default:
err = zfs_standard_error(hdl, errno, errbuf);
break;
}
}
free(nvbuf);
return (err);
}
int
zfs_get_holds(zfs_handle_t *zhp, nvlist_t **nvl)
{
int err;
char errbuf[ERRBUFLEN];
err = lzc_get_holds(zhp->zfs_name, nvl);
if (err != 0) {
libzfs_handle_t *hdl = zhp->zfs_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot get holds for '%s'"),
zhp->zfs_name);
switch (err) {
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded"));
err = zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
case EINVAL:
err = zfs_error(hdl, EZFS_BADTYPE, errbuf);
break;
case ENOENT:
err = zfs_error(hdl, EZFS_NOENT, errbuf);
break;
default:
err = zfs_standard_error(hdl, errno, errbuf);
break;
}
}
return (err);
}
/*
* The theory of raidz space accounting
*
* The "referenced" property of RAIDZ vdevs is scaled such that a 128KB block
* will "reference" 128KB, even though it allocates more than that, to store the
* parity information (and perhaps skip sectors). This concept of the
* "referenced" (and other DMU space accounting) being lower than the allocated
* space by a constant factor is called "raidz deflation."
*
* As mentioned above, the constant factor for raidz deflation assumes a 128KB
* block size. However, zvols typically have a much smaller block size (default
* 8KB). These smaller blocks may require proportionally much more parity
* information (and perhaps skip sectors). In this case, the change to the
* "referenced" property may be much more than the logical block size.
*
* Suppose a raidz vdev has 5 disks with ashift=12. A 128k block may be written
* as follows.
*
* +-------+-------+-------+-------+-------+
* | disk1 | disk2 | disk3 | disk4 | disk5 |
* +-------+-------+-------+-------+-------+
* | P0 | D0 | D8 | D16 | D24 |
* | P1 | D1 | D9 | D17 | D25 |
* | P2 | D2 | D10 | D18 | D26 |
* | P3 | D3 | D11 | D19 | D27 |
* | P4 | D4 | D12 | D20 | D28 |
* | P5 | D5 | D13 | D21 | D29 |
* | P6 | D6 | D14 | D22 | D30 |
* | P7 | D7 | D15 | D23 | D31 |
* +-------+-------+-------+-------+-------+
*
* Above, notice that 160k was allocated: 8 x 4k parity sectors + 32 x 4k data
* sectors. The dataset's referenced will increase by 128k and the pool's
* allocated and free properties will be adjusted by 160k.
*
* A 4k block written to the same raidz vdev will require two 4k sectors. The
* blank cells represent unallocated space.
*
* +-------+-------+-------+-------+-------+
* | disk1 | disk2 | disk3 | disk4 | disk5 |
* +-------+-------+-------+-------+-------+
* | P0 | D0 | | | |
* +-------+-------+-------+-------+-------+
*
* Above, notice that the 4k block required one sector for parity and another
* for data. vdev_raidz_asize() will return 8k and as such the pool's allocated
* and free properties will be adjusted by 8k. The dataset will not be charged
* 8k. Rather, it will be charged a value that is scaled according to the
* overhead of the 128k block on the same vdev. This 8k allocation will be
* charged 8k * 128k / 160k. 128k is from SPA_OLD_MAXBLOCKSIZE and 160k is as
* calculated in the 128k block example above.
*
* Every raidz allocation is sized to be a multiple of nparity+1 sectors. That
* is, every raidz1 allocation will be a multiple of 2 sectors, raidz2
* allocations are a multiple of 3 sectors, and raidz3 allocations are a
* multiple of of 4 sectors. When a block does not fill the required number of
* sectors, skip blocks (sectors) are used.
*
* An 8k block being written to a raidz vdev may be written as follows:
*
* +-------+-------+-------+-------+-------+
* | disk1 | disk2 | disk3 | disk4 | disk5 |
* +-------+-------+-------+-------+-------+
* | P0 | D0 | D1 | S0 | |
* +-------+-------+-------+-------+-------+
*
* In order to maintain the nparity+1 allocation size, a skip block (S0) was
* added. For this 8k block, the pool's allocated and free properties are
* adjusted by 16k and the dataset's referenced is increased by 16k * 128k /
* 160k. Again, 128k is from SPA_OLD_MAXBLOCKSIZE and 160k is as calculated in
* the 128k block example above.
*
* The situation is slightly different for dRAID since the minimum allocation
* size is the full group width. The same 8K block above would be written as
* follows in a dRAID group:
*
* +-------+-------+-------+-------+-------+
* | disk1 | disk2 | disk3 | disk4 | disk5 |
* +-------+-------+-------+-------+-------+
* | P0 | D0 | D1 | S0 | S1 |
* +-------+-------+-------+-------+-------+
*
* Compression may lead to a variety of block sizes being written for the same
* volume or file. There is no clear way to reserve just the amount of space
* that will be required, so the worst case (no compression) is assumed.
* Note that metadata blocks will typically be compressed, so the reservation
* size returned by zvol_volsize_to_reservation() will generally be slightly
* larger than the maximum that the volume can reference.
*/
/*
* Derived from function of same name in module/zfs/vdev_raidz.c. Returns the
* amount of space (in bytes) that will be allocated for the specified block
* size. Note that the "referenced" space accounted will be less than this, but
* not necessarily equal to "blksize", due to RAIDZ deflation.
*/
static uint64_t
vdev_raidz_asize(uint64_t ndisks, uint64_t nparity, uint64_t ashift,
uint64_t blksize)
{
uint64_t asize, ndata;
ASSERT3U(ndisks, >, nparity);
ndata = ndisks - nparity;
asize = ((blksize - 1) >> ashift) + 1;
asize += nparity * ((asize + ndata - 1) / ndata);
asize = roundup(asize, nparity + 1) << ashift;
return (asize);
}
/*
* Derived from function of same name in module/zfs/vdev_draid.c. Returns the
* amount of space (in bytes) that will be allocated for the specified block
* size.
*/
static uint64_t
vdev_draid_asize(uint64_t ndisks, uint64_t nparity, uint64_t ashift,
uint64_t blksize)
{
ASSERT3U(ndisks, >, nparity);
uint64_t ndata = ndisks - nparity;
uint64_t rows = ((blksize - 1) / (ndata << ashift)) + 1;
uint64_t asize = (rows * ndisks) << ashift;
return (asize);
}
/*
* Determine how much space will be allocated if it lands on the most space-
* inefficient top-level vdev. Returns the size in bytes required to store one
* copy of the volume data. See theory comment above.
*/
static uint64_t
volsize_from_vdevs(zpool_handle_t *zhp, uint64_t nblocks, uint64_t blksize)
{
nvlist_t *config, *tree, **vdevs;
uint_t nvdevs;
uint64_t ret = 0;
config = zpool_get_config(zhp, NULL);
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &tree) != 0 ||
nvlist_lookup_nvlist_array(tree, ZPOOL_CONFIG_CHILDREN,
&vdevs, &nvdevs) != 0) {
return (nblocks * blksize);
}
for (int v = 0; v < nvdevs; v++) {
const char *type;
uint64_t nparity, ashift, asize, tsize;
uint64_t volsize;
if (nvlist_lookup_string(vdevs[v], ZPOOL_CONFIG_TYPE,
&type) != 0)
continue;
if (strcmp(type, VDEV_TYPE_RAIDZ) != 0 &&
strcmp(type, VDEV_TYPE_DRAID) != 0)
continue;
if (nvlist_lookup_uint64(vdevs[v],
ZPOOL_CONFIG_NPARITY, &nparity) != 0)
continue;
if (nvlist_lookup_uint64(vdevs[v],
ZPOOL_CONFIG_ASHIFT, &ashift) != 0)
continue;
if (strcmp(type, VDEV_TYPE_RAIDZ) == 0) {
nvlist_t **disks;
uint_t ndisks;
if (nvlist_lookup_nvlist_array(vdevs[v],
ZPOOL_CONFIG_CHILDREN, &disks, &ndisks) != 0)
continue;
/* allocation size for the "typical" 128k block */
tsize = vdev_raidz_asize(ndisks, nparity, ashift,
SPA_OLD_MAXBLOCKSIZE);
/* allocation size for the blksize block */
asize = vdev_raidz_asize(ndisks, nparity, ashift,
blksize);
} else {
uint64_t ndata;
if (nvlist_lookup_uint64(vdevs[v],
ZPOOL_CONFIG_DRAID_NDATA, &ndata) != 0)
continue;
/* allocation size for the "typical" 128k block */
tsize = vdev_draid_asize(ndata + nparity, nparity,
ashift, SPA_OLD_MAXBLOCKSIZE);
/* allocation size for the blksize block */
asize = vdev_draid_asize(ndata + nparity, nparity,
ashift, blksize);
}
/*
* Scale this size down as a ratio of 128k / tsize.
* See theory statement above.
*/
volsize = nblocks * asize * SPA_OLD_MAXBLOCKSIZE / tsize;
if (volsize > ret) {
ret = volsize;
}
}
if (ret == 0) {
ret = nblocks * blksize;
}
return (ret);
}
/*
* Convert the zvol's volume size to an appropriate reservation. See theory
* comment above.
*
* Note: If this routine is updated, it is necessary to update the ZFS test
* suite's shell version in reservation.shlib.
*/
uint64_t
zvol_volsize_to_reservation(zpool_handle_t *zph, uint64_t volsize,
nvlist_t *props)
{
uint64_t numdb;
uint64_t nblocks, volblocksize;
int ncopies;
const char *strval;
if (nvlist_lookup_string(props,
zfs_prop_to_name(ZFS_PROP_COPIES), &strval) == 0)
ncopies = atoi(strval);
else
ncopies = 1;
if (nvlist_lookup_uint64(props,
zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE),
&volblocksize) != 0)
volblocksize = ZVOL_DEFAULT_BLOCKSIZE;
nblocks = volsize / volblocksize;
/*
* Metadata defaults to using 128k blocks, not volblocksize blocks. For
* this reason, only the data blocks are scaled based on vdev config.
*/
volsize = volsize_from_vdevs(zph, nblocks, volblocksize);
/* start with metadnode L0-L6 */
numdb = 7;
/* calculate number of indirects */
while (nblocks > 1) {
nblocks += DNODES_PER_LEVEL - 1;
nblocks /= DNODES_PER_LEVEL;
numdb += nblocks;
}
numdb *= MIN(SPA_DVAS_PER_BP, ncopies + 1);
volsize *= ncopies;
/*
* this is exactly DN_MAX_INDBLKSHIFT when metadata isn't
* compressed, but in practice they compress down to about
* 1100 bytes
*/
numdb *= 1ULL << DN_MAX_INDBLKSHIFT;
volsize += numdb;
return (volsize);
}
/*
* Wait for the given activity and return the status of the wait (whether or not
* any waiting was done) in the 'waited' parameter. Non-existent fses are
* reported via the 'missing' parameter, rather than by printing an error
* message. This is convenient when this function is called in a loop over a
* long period of time (as it is, for example, by zfs's wait cmd). In that
* scenario, a fs being exported or destroyed should be considered a normal
* event, so we don't want to print an error when we find that the fs doesn't
* exist.
*/
int
zfs_wait_status(zfs_handle_t *zhp, zfs_wait_activity_t activity,
boolean_t *missing, boolean_t *waited)
{
int error = lzc_wait_fs(zhp->zfs_name, activity, waited);
*missing = (error == ENOENT);
if (*missing)
return (0);
if (error != 0) {
(void) zfs_standard_error_fmt(zhp->zfs_hdl, error,
dgettext(TEXT_DOMAIN, "error waiting in fs '%s'"),
zhp->zfs_name);
}
return (error);
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_iter.c b/sys/contrib/openzfs/lib/libzfs/libzfs_iter.c
index 681fe5b4748d..452d8fd6ab71 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_iter.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_iter.c
@@ -1,605 +1,654 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013, 2019 by Delphix. All rights reserved.
* Copyright 2014 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2019 Datto Inc.
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <stddef.h>
#include <libintl.h>
#include <libzfs.h>
#include <libzutil.h>
#include <sys/mntent.h>
#include "libzfs_impl.h"
static int
zfs_iter_clones(zfs_handle_t *zhp, int flags __maybe_unused, zfs_iter_f func,
void *data)
{
nvlist_t *nvl = zfs_get_clones_nvl(zhp);
nvpair_t *pair;
if (nvl == NULL)
return (0);
for (pair = nvlist_next_nvpair(nvl, NULL); pair != NULL;
pair = nvlist_next_nvpair(nvl, pair)) {
zfs_handle_t *clone = zfs_open(zhp->zfs_hdl, nvpair_name(pair),
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (clone != NULL) {
int err = func(clone, data);
if (err != 0)
return (err);
}
}
return (0);
}
static int
zfs_do_list_ioctl(zfs_handle_t *zhp, int arg, zfs_cmd_t *zc)
{
int rc;
uint64_t orig_cookie;
orig_cookie = zc->zc_cookie;
top:
(void) strlcpy(zc->zc_name, zhp->zfs_name, sizeof (zc->zc_name));
zc->zc_objset_stats.dds_creation_txg = 0;
rc = zfs_ioctl(zhp->zfs_hdl, arg, zc);
if (rc == -1) {
switch (errno) {
case ENOMEM:
/* expand nvlist memory and try again */
zcmd_expand_dst_nvlist(zhp->zfs_hdl, zc);
zc->zc_cookie = orig_cookie;
goto top;
/*
* An errno value of ESRCH indicates normal completion.
* If ENOENT is returned, then the underlying dataset
* has been removed since we obtained the handle.
*/
case ESRCH:
case ENOENT:
rc = 1;
break;
default:
rc = zfs_standard_error(zhp->zfs_hdl, errno,
dgettext(TEXT_DOMAIN,
"cannot iterate filesystems"));
break;
}
}
return (rc);
}
/*
* Iterate over all child filesystems
*/
int
-zfs_iter_filesystems(zfs_handle_t *zhp, int flags, zfs_iter_f func, void *data)
+zfs_iter_filesystems(zfs_handle_t *zhp, zfs_iter_f func, void *data)
+{
+ return (zfs_iter_filesystems_v2(zhp, 0, func, data));
+}
+
+int
+zfs_iter_filesystems_v2(zfs_handle_t *zhp, int flags, zfs_iter_f func,
+ void *data)
{
zfs_cmd_t zc = {"\0"};
zfs_handle_t *nzhp;
int ret;
if (zhp->zfs_type != ZFS_TYPE_FILESYSTEM)
return (0);
zcmd_alloc_dst_nvlist(zhp->zfs_hdl, &zc, 0);
if ((flags & ZFS_ITER_SIMPLE) == ZFS_ITER_SIMPLE)
zc.zc_simple = B_TRUE;
while ((ret = zfs_do_list_ioctl(zhp, ZFS_IOC_DATASET_LIST_NEXT,
&zc)) == 0) {
if (zc.zc_simple)
nzhp = make_dataset_simple_handle_zc(zhp, &zc);
else
nzhp = make_dataset_handle_zc(zhp->zfs_hdl, &zc);
/*
* Silently ignore errors, as the only plausible explanation is
* that the pool has since been removed.
*/
if (nzhp == NULL)
continue;
if ((ret = func(nzhp, data)) != 0) {
zcmd_free_nvlists(&zc);
return (ret);
}
}
zcmd_free_nvlists(&zc);
return ((ret < 0) ? ret : 0);
}
/*
* Iterate over all snapshots
*/
int
-zfs_iter_snapshots(zfs_handle_t *zhp, int flags, zfs_iter_f func,
+zfs_iter_snapshots(zfs_handle_t *zhp, boolean_t simple, zfs_iter_f func,
+ void *data, uint64_t min_txg, uint64_t max_txg)
+{
+ return (zfs_iter_snapshots_v2(zhp, simple ? ZFS_ITER_SIMPLE : 0, func,
+ data, min_txg, max_txg));
+}
+
+int
+zfs_iter_snapshots_v2(zfs_handle_t *zhp, int flags, zfs_iter_f func,
void *data, uint64_t min_txg, uint64_t max_txg)
{
zfs_cmd_t zc = {"\0"};
zfs_handle_t *nzhp;
int ret;
nvlist_t *range_nvl = NULL;
if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT ||
zhp->zfs_type == ZFS_TYPE_BOOKMARK)
return (0);
zc.zc_simple = (flags & ZFS_ITER_SIMPLE) != 0;
zcmd_alloc_dst_nvlist(zhp->zfs_hdl, &zc, 0);
if (min_txg != 0) {
range_nvl = fnvlist_alloc();
fnvlist_add_uint64(range_nvl, SNAP_ITER_MIN_TXG, min_txg);
}
if (max_txg != 0) {
if (range_nvl == NULL)
range_nvl = fnvlist_alloc();
fnvlist_add_uint64(range_nvl, SNAP_ITER_MAX_TXG, max_txg);
}
if (range_nvl != NULL)
zcmd_write_src_nvlist(zhp->zfs_hdl, &zc, range_nvl);
while ((ret = zfs_do_list_ioctl(zhp, ZFS_IOC_SNAPSHOT_LIST_NEXT,
&zc)) == 0) {
if (zc.zc_simple)
nzhp = make_dataset_simple_handle_zc(zhp, &zc);
else
nzhp = make_dataset_handle_zc(zhp->zfs_hdl, &zc);
if (nzhp == NULL)
continue;
if ((ret = func(nzhp, data)) != 0) {
zcmd_free_nvlists(&zc);
fnvlist_free(range_nvl);
return (ret);
}
}
zcmd_free_nvlists(&zc);
fnvlist_free(range_nvl);
return ((ret < 0) ? ret : 0);
}
/*
* Iterate over all bookmarks
*/
int
-zfs_iter_bookmarks(zfs_handle_t *zhp, int flags __maybe_unused,
+zfs_iter_bookmarks(zfs_handle_t *zhp, zfs_iter_f func, void *data)
+{
+ return (zfs_iter_bookmarks_v2(zhp, 0, func, data));
+}
+
+int
+zfs_iter_bookmarks_v2(zfs_handle_t *zhp, int flags __maybe_unused,
zfs_iter_f func, void *data)
{
zfs_handle_t *nzhp;
nvlist_t *props = NULL;
nvlist_t *bmarks = NULL;
int err;
nvpair_t *pair;
if ((zfs_get_type(zhp) & (ZFS_TYPE_SNAPSHOT | ZFS_TYPE_BOOKMARK)) != 0)
return (0);
/* Setup the requested properties nvlist. */
props = fnvlist_alloc();
for (zfs_prop_t p = 0; p < ZFS_NUM_PROPS; p++) {
if (zfs_prop_valid_for_type(p, ZFS_TYPE_BOOKMARK, B_FALSE)) {
fnvlist_add_boolean(props, zfs_prop_to_name(p));
}
}
fnvlist_add_boolean(props, "redact_complete");
if ((err = lzc_get_bookmarks(zhp->zfs_name, props, &bmarks)) != 0)
goto out;
for (pair = nvlist_next_nvpair(bmarks, NULL);
pair != NULL; pair = nvlist_next_nvpair(bmarks, pair)) {
char name[ZFS_MAX_DATASET_NAME_LEN];
const char *bmark_name;
nvlist_t *bmark_props;
bmark_name = nvpair_name(pair);
bmark_props = fnvpair_value_nvlist(pair);
if (snprintf(name, sizeof (name), "%s#%s", zhp->zfs_name,
bmark_name) >= sizeof (name)) {
err = EINVAL;
goto out;
}
nzhp = make_bookmark_handle(zhp, name, bmark_props);
if (nzhp == NULL)
continue;
if ((err = func(nzhp, data)) != 0)
goto out;
}
out:
fnvlist_free(props);
fnvlist_free(bmarks);
return (err);
}
/*
* Routines for dealing with the sorted snapshot functionality
*/
typedef struct zfs_node {
zfs_handle_t *zn_handle;
avl_node_t zn_avlnode;
} zfs_node_t;
static int
zfs_sort_snaps(zfs_handle_t *zhp, void *data)
{
avl_tree_t *avl = data;
zfs_node_t *node;
zfs_node_t search;
search.zn_handle = zhp;
node = avl_find(avl, &search, NULL);
if (node) {
/*
* If this snapshot was renamed while we were creating the
* AVL tree, it's possible that we already inserted it under
* its old name. Remove the old handle before adding the new
* one.
*/
zfs_close(node->zn_handle);
avl_remove(avl, node);
free(node);
}
node = zfs_alloc(zhp->zfs_hdl, sizeof (zfs_node_t));
node->zn_handle = zhp;
avl_add(avl, node);
return (0);
}
static int
zfs_snapshot_compare(const void *larg, const void *rarg)
{
zfs_handle_t *l = ((zfs_node_t *)larg)->zn_handle;
zfs_handle_t *r = ((zfs_node_t *)rarg)->zn_handle;
uint64_t lcreate, rcreate;
/*
* Sort them according to creation time. We use the hidden
* CREATETXG property to get an absolute ordering of snapshots.
*/
lcreate = zfs_prop_get_int(l, ZFS_PROP_CREATETXG);
rcreate = zfs_prop_get_int(r, ZFS_PROP_CREATETXG);
return (TREE_CMP(lcreate, rcreate));
}
int
-zfs_iter_snapshots_sorted(zfs_handle_t *zhp, int flags, zfs_iter_f callback,
+zfs_iter_snapshots_sorted(zfs_handle_t *zhp, zfs_iter_f callback,
+ void *data, uint64_t min_txg, uint64_t max_txg)
+{
+ return (zfs_iter_snapshots_sorted_v2(zhp, 0, callback, data,
+ min_txg, max_txg));
+}
+
+int
+zfs_iter_snapshots_sorted_v2(zfs_handle_t *zhp, int flags, zfs_iter_f callback,
void *data, uint64_t min_txg, uint64_t max_txg)
{
int ret = 0;
zfs_node_t *node;
avl_tree_t avl;
void *cookie = NULL;
avl_create(&avl, zfs_snapshot_compare,
sizeof (zfs_node_t), offsetof(zfs_node_t, zn_avlnode));
- ret = zfs_iter_snapshots(zhp, flags, zfs_sort_snaps, &avl, min_txg,
+ ret = zfs_iter_snapshots_v2(zhp, flags, zfs_sort_snaps, &avl, min_txg,
max_txg);
for (node = avl_first(&avl); node != NULL; node = AVL_NEXT(&avl, node))
ret |= callback(node->zn_handle, data);
while ((node = avl_destroy_nodes(&avl, &cookie)) != NULL)
free(node);
avl_destroy(&avl);
return (ret);
}
typedef struct {
char *ssa_first;
char *ssa_last;
boolean_t ssa_seenfirst;
boolean_t ssa_seenlast;
zfs_iter_f ssa_func;
void *ssa_arg;
} snapspec_arg_t;
static int
snapspec_cb(zfs_handle_t *zhp, void *arg)
{
snapspec_arg_t *ssa = arg;
const char *shortsnapname;
int err = 0;
if (ssa->ssa_seenlast)
return (0);
shortsnapname = strchr(zfs_get_name(zhp), '@') + 1;
if (!ssa->ssa_seenfirst && strcmp(shortsnapname, ssa->ssa_first) == 0)
ssa->ssa_seenfirst = B_TRUE;
if (strcmp(shortsnapname, ssa->ssa_last) == 0)
ssa->ssa_seenlast = B_TRUE;
if (ssa->ssa_seenfirst) {
err = ssa->ssa_func(zhp, ssa->ssa_arg);
} else {
zfs_close(zhp);
}
return (err);
}
/*
* spec is a string like "A,B%C,D"
*
* <snaps>, where <snaps> can be:
* <snap> (single snapshot)
* <snap>%<snap> (range of snapshots, inclusive)
* %<snap> (range of snapshots, starting with earliest)
* <snap>% (range of snapshots, ending with last)
* % (all snapshots)
* <snaps>[,...] (comma separated list of the above)
*
* If a snapshot can not be opened, continue trying to open the others, but
* return ENOENT at the end.
*/
int
-zfs_iter_snapspec(zfs_handle_t *fs_zhp, int flags, const char *spec_orig,
+zfs_iter_snapspec(zfs_handle_t *fs_zhp, const char *spec_orig,
+ zfs_iter_f func, void *arg)
+{
+ return (zfs_iter_snapspec_v2(fs_zhp, 0, spec_orig, func, arg));
+}
+
+int
+zfs_iter_snapspec_v2(zfs_handle_t *fs_zhp, int flags, const char *spec_orig,
zfs_iter_f func, void *arg)
{
char *buf, *comma_separated, *cp;
int err = 0;
int ret = 0;
buf = zfs_strdup(fs_zhp->zfs_hdl, spec_orig);
cp = buf;
while ((comma_separated = strsep(&cp, ",")) != NULL) {
char *pct = strchr(comma_separated, '%');
if (pct != NULL) {
snapspec_arg_t ssa = { 0 };
ssa.ssa_func = func;
ssa.ssa_arg = arg;
if (pct == comma_separated)
ssa.ssa_seenfirst = B_TRUE;
else
ssa.ssa_first = comma_separated;
*pct = '\0';
ssa.ssa_last = pct + 1;
/*
* If there is a lastname specified, make sure it
* exists.
*/
if (ssa.ssa_last[0] != '\0') {
char snapname[ZFS_MAX_DATASET_NAME_LEN];
(void) snprintf(snapname, sizeof (snapname),
"%s@%s", zfs_get_name(fs_zhp),
ssa.ssa_last);
if (!zfs_dataset_exists(fs_zhp->zfs_hdl,
snapname, ZFS_TYPE_SNAPSHOT)) {
ret = ENOENT;
continue;
}
}
- err = zfs_iter_snapshots_sorted(fs_zhp, flags,
+ err = zfs_iter_snapshots_sorted_v2(fs_zhp, flags,
snapspec_cb, &ssa, 0, 0);
if (ret == 0)
ret = err;
if (ret == 0 && (!ssa.ssa_seenfirst ||
(ssa.ssa_last[0] != '\0' && !ssa.ssa_seenlast))) {
ret = ENOENT;
}
} else {
char snapname[ZFS_MAX_DATASET_NAME_LEN];
zfs_handle_t *snap_zhp;
(void) snprintf(snapname, sizeof (snapname), "%s@%s",
zfs_get_name(fs_zhp), comma_separated);
snap_zhp = make_dataset_handle(fs_zhp->zfs_hdl,
snapname);
if (snap_zhp == NULL) {
ret = ENOENT;
continue;
}
err = func(snap_zhp, arg);
if (ret == 0)
ret = err;
}
}
free(buf);
return (ret);
}
/*
* Iterate over all children, snapshots and filesystems
* Process snapshots before filesystems because they are nearer the input
* handle: this is extremely important when used with zfs_iter_f functions
* looking for data, following the logic that we would like to find it as soon
* and as close as possible.
*/
int
-zfs_iter_children(zfs_handle_t *zhp, int flags, zfs_iter_f func, void *data)
+zfs_iter_children(zfs_handle_t *zhp, zfs_iter_f func, void *data)
+{
+ return (zfs_iter_children_v2(zhp, 0, func, data));
+}
+
+int
+zfs_iter_children_v2(zfs_handle_t *zhp, int flags, zfs_iter_f func, void *data)
{
int ret;
- if ((ret = zfs_iter_snapshots(zhp, flags, func, data, 0, 0)) != 0)
+ if ((ret = zfs_iter_snapshots_v2(zhp, flags, func, data, 0, 0)) != 0)
return (ret);
- return (zfs_iter_filesystems(zhp, flags, func, data));
+ return (zfs_iter_filesystems_v2(zhp, flags, func, data));
}
typedef struct iter_stack_frame {
struct iter_stack_frame *next;
zfs_handle_t *zhp;
} iter_stack_frame_t;
typedef struct iter_dependents_arg {
boolean_t first;
int flags;
boolean_t allowrecursion;
iter_stack_frame_t *stack;
zfs_iter_f func;
void *data;
} iter_dependents_arg_t;
static int
iter_dependents_cb(zfs_handle_t *zhp, void *arg)
{
iter_dependents_arg_t *ida = arg;
int err = 0;
boolean_t first = ida->first;
ida->first = B_FALSE;
if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) {
err = zfs_iter_clones(zhp, ida->flags, iter_dependents_cb, ida);
} else if (zhp->zfs_type != ZFS_TYPE_BOOKMARK) {
iter_stack_frame_t isf;
iter_stack_frame_t *f;
/*
* check if there is a cycle by seeing if this fs is already
* on the stack.
*/
for (f = ida->stack; f != NULL; f = f->next) {
if (f->zhp->zfs_dmustats.dds_guid ==
zhp->zfs_dmustats.dds_guid) {
if (ida->allowrecursion) {
zfs_close(zhp);
return (0);
} else {
zfs_error_aux(zhp->zfs_hdl,
dgettext(TEXT_DOMAIN,
"recursive dependency at '%s'"),
zfs_get_name(zhp));
err = zfs_error(zhp->zfs_hdl,
EZFS_RECURSIVE,
dgettext(TEXT_DOMAIN,
"cannot determine dependent "
"datasets"));
zfs_close(zhp);
return (err);
}
}
}
isf.zhp = zhp;
isf.next = ida->stack;
ida->stack = &isf;
- err = zfs_iter_filesystems(zhp, ida->flags,
+ err = zfs_iter_filesystems_v2(zhp, ida->flags,
iter_dependents_cb, ida);
if (err == 0)
- err = zfs_iter_snapshots(zhp, ida->flags,
+ err = zfs_iter_snapshots_v2(zhp, ida->flags,
iter_dependents_cb, ida, 0, 0);
ida->stack = isf.next;
}
if (!first && err == 0)
err = ida->func(zhp, ida->data);
else
zfs_close(zhp);
return (err);
}
int
-zfs_iter_dependents(zfs_handle_t *zhp, int flags, boolean_t allowrecursion,
+zfs_iter_dependents(zfs_handle_t *zhp, boolean_t allowrecursion,
+ zfs_iter_f func, void *data)
+{
+ return (zfs_iter_dependents_v2(zhp, 0, allowrecursion, func, data));
+}
+
+int
+zfs_iter_dependents_v2(zfs_handle_t *zhp, int flags, boolean_t allowrecursion,
zfs_iter_f func, void *data)
{
iter_dependents_arg_t ida;
ida.flags = flags;
ida.allowrecursion = allowrecursion;
ida.stack = NULL;
ida.func = func;
ida.data = data;
ida.first = B_TRUE;
return (iter_dependents_cb(zfs_handle_dup(zhp), &ida));
}
/*
* Iterate over mounted children of the specified dataset
*/
int
zfs_iter_mounted(zfs_handle_t *zhp, zfs_iter_f func, void *data)
{
char mnt_prop[ZFS_MAXPROPLEN];
struct mnttab entry;
zfs_handle_t *mtab_zhp;
size_t namelen = strlen(zhp->zfs_name);
FILE *mnttab;
int err = 0;
if ((mnttab = fopen(MNTTAB, "re")) == NULL)
return (ENOENT);
while (err == 0 && getmntent(mnttab, &entry) == 0) {
/* Ignore non-ZFS entries */
if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0)
continue;
/* Ignore datasets not within the provided dataset */
if (strncmp(entry.mnt_special, zhp->zfs_name, namelen) != 0 ||
entry.mnt_special[namelen] != '/')
continue;
/* Skip snapshot of any child dataset */
if (strchr(entry.mnt_special, '@') != NULL)
continue;
if ((mtab_zhp = zfs_open(zhp->zfs_hdl, entry.mnt_special,
ZFS_TYPE_FILESYSTEM)) == NULL)
continue;
/* Ignore legacy mounts as they are user managed */
verify(zfs_prop_get(mtab_zhp, ZFS_PROP_MOUNTPOINT, mnt_prop,
sizeof (mnt_prop), NULL, NULL, 0, B_FALSE) == 0);
if (strcmp(mnt_prop, "legacy") == 0) {
zfs_close(mtab_zhp);
continue;
}
err = func(mtab_zhp, data);
}
fclose(mnttab);
return (err);
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_mount.c b/sys/contrib/openzfs/lib/libzfs/libzfs_mount.c
index 024f449baa0b..5d1fe651c97e 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_mount.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_mount.c
@@ -1,1462 +1,1462 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2015 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2014, 2022 by Delphix. All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>
* Copyright 2017 RackTop Systems.
* Copyright (c) 2018 Datto Inc.
* Copyright 2018 OmniOS Community Edition (OmniOSce) Association.
*/
/*
* Routines to manage ZFS mounts. We separate all the nasty routines that have
* to deal with the OS. The following functions are the main entry points --
* they are used by mount and unmount and when changing a filesystem's
* mountpoint.
*
* zfs_is_mounted()
* zfs_mount()
* zfs_mount_at()
* zfs_unmount()
* zfs_unmountall()
*
* This file also contains the functions used to manage sharing filesystems:
*
* zfs_is_shared()
* zfs_share()
* zfs_unshare()
* zfs_unshareall()
* zfs_commit_shares()
*
* The following functions are available for pool consumers, and will
* mount/unmount and share/unshare all datasets within pool:
*
* zpool_enable_datasets()
* zpool_disable_datasets()
*/
#include <dirent.h>
#include <dlfcn.h>
#include <errno.h>
#include <fcntl.h>
#include <libgen.h>
#include <libintl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <zone.h>
#include <sys/mntent.h>
#include <sys/mount.h>
#include <sys/stat.h>
#include <sys/vfs.h>
#include <sys/dsl_crypt.h>
#include <libzfs.h>
#include "libzfs_impl.h"
#include <thread_pool.h>
#include <libshare.h>
#include <sys/systeminfo.h>
#define MAXISALEN 257 /* based on sysinfo(2) man page */
static int mount_tp_nthr = 512; /* tpool threads for multi-threaded mounting */
static void zfs_mount_task(void *);
static const proto_table_t proto_table[SA_PROTOCOL_COUNT] = {
[SA_PROTOCOL_NFS] =
{ZFS_PROP_SHARENFS, EZFS_SHARENFSFAILED, EZFS_UNSHARENFSFAILED},
[SA_PROTOCOL_SMB] =
{ZFS_PROP_SHARESMB, EZFS_SHARESMBFAILED, EZFS_UNSHARESMBFAILED},
};
static const enum sa_protocol share_all_proto[SA_PROTOCOL_COUNT + 1] = {
SA_PROTOCOL_NFS,
SA_PROTOCOL_SMB,
SA_NO_PROTOCOL
};
static boolean_t
dir_is_empty_stat(const char *dirname)
{
struct stat st;
/*
* We only want to return false if the given path is a non empty
* directory, all other errors are handled elsewhere.
*/
if (stat(dirname, &st) < 0 || !S_ISDIR(st.st_mode)) {
return (B_TRUE);
}
/*
* An empty directory will still have two entries in it, one
* entry for each of "." and "..".
*/
if (st.st_size > 2) {
return (B_FALSE);
}
return (B_TRUE);
}
static boolean_t
dir_is_empty_readdir(const char *dirname)
{
DIR *dirp;
struct dirent64 *dp;
int dirfd;
if ((dirfd = openat(AT_FDCWD, dirname,
O_RDONLY | O_NDELAY | O_LARGEFILE | O_CLOEXEC, 0)) < 0) {
return (B_TRUE);
}
if ((dirp = fdopendir(dirfd)) == NULL) {
(void) close(dirfd);
return (B_TRUE);
}
while ((dp = readdir64(dirp)) != NULL) {
if (strcmp(dp->d_name, ".") == 0 ||
strcmp(dp->d_name, "..") == 0)
continue;
(void) closedir(dirp);
return (B_FALSE);
}
(void) closedir(dirp);
return (B_TRUE);
}
/*
* Returns true if the specified directory is empty. If we can't open the
* directory at all, return true so that the mount can fail with a more
* informative error message.
*/
static boolean_t
dir_is_empty(const char *dirname)
{
struct statfs64 st;
/*
* If the statvfs call fails or the filesystem is not a ZFS
* filesystem, fall back to the slow path which uses readdir.
*/
if ((statfs64(dirname, &st) != 0) ||
(st.f_type != ZFS_SUPER_MAGIC)) {
return (dir_is_empty_readdir(dirname));
}
/*
* At this point, we know the provided path is on a ZFS
* filesystem, so we can use stat instead of readdir to
* determine if the directory is empty or not. We try to avoid
* using readdir because that requires opening "dirname"; this
* open file descriptor can potentially end up in a child
* process if there's a concurrent fork, thus preventing the
* zfs_mount() from otherwise succeeding (the open file
* descriptor inherited by the child process will cause the
* parent's mount to fail with EBUSY). The performance
* implications of replacing the open, read, and close with a
* single stat is nice; but is not the main motivation for the
* added complexity.
*/
return (dir_is_empty_stat(dirname));
}
/*
* Checks to see if the mount is active. If the filesystem is mounted, we fill
* in 'where' with the current mountpoint, and return 1. Otherwise, we return
* 0.
*/
boolean_t
is_mounted(libzfs_handle_t *zfs_hdl, const char *special, char **where)
{
struct mnttab entry;
if (libzfs_mnttab_find(zfs_hdl, special, &entry) != 0)
return (B_FALSE);
if (where != NULL)
*where = zfs_strdup(zfs_hdl, entry.mnt_mountp);
return (B_TRUE);
}
boolean_t
zfs_is_mounted(zfs_handle_t *zhp, char **where)
{
return (is_mounted(zhp->zfs_hdl, zfs_get_name(zhp), where));
}
/*
* Checks any higher order concerns about whether the given dataset is
* mountable, false otherwise. zfs_is_mountable_internal specifically assumes
* that the caller has verified the sanity of mounting the dataset at
* its mountpoint to the extent the caller wants.
*/
static boolean_t
zfs_is_mountable_internal(zfs_handle_t *zhp)
{
if (zfs_prop_get_int(zhp, ZFS_PROP_ZONED) &&
getzoneid() == GLOBAL_ZONEID)
return (B_FALSE);
return (B_TRUE);
}
/*
* Returns true if the given dataset is mountable, false otherwise. Returns the
* mountpoint in 'buf'.
*/
static boolean_t
zfs_is_mountable(zfs_handle_t *zhp, char *buf, size_t buflen,
zprop_source_t *source, int flags)
{
char sourceloc[MAXNAMELEN];
zprop_source_t sourcetype;
if (!zfs_prop_valid_for_type(ZFS_PROP_MOUNTPOINT, zhp->zfs_type,
B_FALSE))
return (B_FALSE);
verify(zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, buf, buflen,
&sourcetype, sourceloc, sizeof (sourceloc), B_FALSE) == 0);
if (strcmp(buf, ZFS_MOUNTPOINT_NONE) == 0 ||
strcmp(buf, ZFS_MOUNTPOINT_LEGACY) == 0)
return (B_FALSE);
if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) == ZFS_CANMOUNT_OFF)
return (B_FALSE);
if (!zfs_is_mountable_internal(zhp))
return (B_FALSE);
if (zfs_prop_get_int(zhp, ZFS_PROP_REDACTED) && !(flags & MS_FORCE))
return (B_FALSE);
if (source)
*source = sourcetype;
return (B_TRUE);
}
/*
* The filesystem is mounted by invoking the system mount utility rather
* than by the system call mount(2). This ensures that the /etc/mtab
* file is correctly locked for the update. Performing our own locking
* and /etc/mtab update requires making an unsafe assumption about how
* the mount utility performs its locking. Unfortunately, this also means
* in the case of a mount failure we do not have the exact errno. We must
* make due with return value from the mount process.
*
* In the long term a shared library called libmount is under development
* which provides a common API to address the locking and errno issues.
* Once the standard mount utility has been updated to use this library
* we can add an autoconf check to conditionally use it.
*
* http://www.kernel.org/pub/linux/utils/util-linux/libmount-docs/index.html
*/
static int
zfs_add_option(zfs_handle_t *zhp, char *options, int len,
zfs_prop_t prop, const char *on, const char *off)
{
const char *source;
uint64_t value;
/* Skip adding duplicate default options */
if ((strstr(options, on) != NULL) || (strstr(options, off) != NULL))
return (0);
/*
* zfs_prop_get_int() is not used to ensure our mount options
* are not influenced by the current /proc/self/mounts contents.
*/
value = getprop_uint64(zhp, prop, &source);
(void) strlcat(options, ",", len);
(void) strlcat(options, value ? on : off, len);
return (0);
}
static int
zfs_add_options(zfs_handle_t *zhp, char *options, int len)
{
int error = 0;
error = zfs_add_option(zhp, options, len,
ZFS_PROP_ATIME, MNTOPT_ATIME, MNTOPT_NOATIME);
/*
* don't add relatime/strictatime when atime=off, otherwise strictatime
* will force atime=on
*/
if (strstr(options, MNTOPT_NOATIME) == NULL) {
error = zfs_add_option(zhp, options, len,
ZFS_PROP_RELATIME, MNTOPT_RELATIME, MNTOPT_STRICTATIME);
}
error = error ? error : zfs_add_option(zhp, options, len,
ZFS_PROP_DEVICES, MNTOPT_DEVICES, MNTOPT_NODEVICES);
error = error ? error : zfs_add_option(zhp, options, len,
ZFS_PROP_EXEC, MNTOPT_EXEC, MNTOPT_NOEXEC);
error = error ? error : zfs_add_option(zhp, options, len,
ZFS_PROP_READONLY, MNTOPT_RO, MNTOPT_RW);
error = error ? error : zfs_add_option(zhp, options, len,
ZFS_PROP_SETUID, MNTOPT_SETUID, MNTOPT_NOSETUID);
error = error ? error : zfs_add_option(zhp, options, len,
ZFS_PROP_NBMAND, MNTOPT_NBMAND, MNTOPT_NONBMAND);
return (error);
}
int
zfs_mount(zfs_handle_t *zhp, const char *options, int flags)
{
char mountpoint[ZFS_MAXPROPLEN];
if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint), NULL,
flags))
return (0);
return (zfs_mount_at(zhp, options, flags, mountpoint));
}
/*
* Mount the given filesystem.
*/
int
zfs_mount_at(zfs_handle_t *zhp, const char *options, int flags,
const char *mountpoint)
{
struct stat buf;
char mntopts[MNT_LINE_MAX];
char overlay[ZFS_MAXPROPLEN];
char prop_encroot[MAXNAMELEN];
boolean_t is_encroot;
zfs_handle_t *encroot_hp = zhp;
libzfs_handle_t *hdl = zhp->zfs_hdl;
uint64_t keystatus;
int remount = 0, rc;
if (options == NULL) {
(void) strlcpy(mntopts, MNTOPT_DEFAULTS, sizeof (mntopts));
} else {
(void) strlcpy(mntopts, options, sizeof (mntopts));
}
if (strstr(mntopts, MNTOPT_REMOUNT) != NULL)
remount = 1;
/* Potentially duplicates some checks if invoked by zfs_mount(). */
if (!zfs_is_mountable_internal(zhp))
return (0);
/*
* If the pool is imported read-only then all mounts must be read-only
*/
if (zpool_get_prop_int(zhp->zpool_hdl, ZPOOL_PROP_READONLY, NULL))
(void) strlcat(mntopts, "," MNTOPT_RO, sizeof (mntopts));
/*
* Append default mount options which apply to the mount point.
* This is done because under Linux (unlike Solaris) multiple mount
* points may reference a single super block. This means that just
* given a super block there is no back reference to update the per
* mount point options.
*/
rc = zfs_add_options(zhp, mntopts, sizeof (mntopts));
if (rc) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"default options unavailable"));
return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
dgettext(TEXT_DOMAIN, "cannot mount '%s'"),
mountpoint));
}
/*
* If the filesystem is encrypted the key must be loaded in order to
* mount. If the key isn't loaded, the MS_CRYPT flag decides whether
* or not we attempt to load the keys. Note: we must call
* zfs_refresh_properties() here since some callers of this function
* (most notably zpool_enable_datasets()) may implicitly load our key
* by loading the parent's key first.
*/
if (zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION) != ZIO_CRYPT_OFF) {
zfs_refresh_properties(zhp);
keystatus = zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS);
/*
* If the key is unavailable and MS_CRYPT is set give the
* user a chance to enter the key. Otherwise just fail
* immediately.
*/
if (keystatus == ZFS_KEYSTATUS_UNAVAILABLE) {
if (flags & MS_CRYPT) {
rc = zfs_crypto_get_encryption_root(zhp,
&is_encroot, prop_encroot);
if (rc) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Failed to get encryption root for "
"'%s'."), zfs_get_name(zhp));
return (rc);
}
if (!is_encroot) {
encroot_hp = zfs_open(hdl, prop_encroot,
ZFS_TYPE_DATASET);
if (encroot_hp == NULL)
return (hdl->libzfs_error);
}
rc = zfs_crypto_load_key(encroot_hp,
B_FALSE, NULL);
if (!is_encroot)
zfs_close(encroot_hp);
if (rc)
return (rc);
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"encryption key not loaded"));
return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
dgettext(TEXT_DOMAIN, "cannot mount '%s'"),
mountpoint));
}
}
}
/*
* Append zfsutil option so the mount helper allow the mount
*/
strlcat(mntopts, "," MNTOPT_ZFSUTIL, sizeof (mntopts));
/* Create the directory if it doesn't already exist */
if (lstat(mountpoint, &buf) != 0) {
if (mkdirp(mountpoint, 0755) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"failed to create mountpoint: %s"),
strerror(errno));
return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
dgettext(TEXT_DOMAIN, "cannot mount '%s'"),
mountpoint));
}
}
/*
* Overlay mounts are enabled by default but may be disabled
* via the 'overlay' property. The -O flag remains for compatibility.
*/
if (!(flags & MS_OVERLAY)) {
if (zfs_prop_get(zhp, ZFS_PROP_OVERLAY, overlay,
sizeof (overlay), NULL, NULL, 0, B_FALSE) == 0) {
if (strcmp(overlay, "on") == 0) {
flags |= MS_OVERLAY;
}
}
}
/*
* Determine if the mountpoint is empty. If so, refuse to perform the
* mount. We don't perform this check if 'remount' is
* specified or if overlay option (-O) is given
*/
if ((flags & MS_OVERLAY) == 0 && !remount &&
!dir_is_empty(mountpoint)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"directory is not empty"));
return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
dgettext(TEXT_DOMAIN, "cannot mount '%s'"), mountpoint));
}
/* perform the mount */
rc = do_mount(zhp, mountpoint, mntopts, flags);
if (rc) {
/*
* Generic errors are nasty, but there are just way too many
* from mount(), and they're well-understood. We pick a few
* common ones to improve upon.
*/
if (rc == EBUSY) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"mountpoint or dataset is busy"));
} else if (rc == EPERM) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Insufficient privileges"));
} else if (rc == ENOTSUP) {
int spa_version;
VERIFY(zfs_spa_version(zhp, &spa_version) == 0);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Can't mount a version %llu "
"file system on a version %d pool. Pool must be"
" upgraded to mount this file system."),
(u_longlong_t)zfs_prop_get_int(zhp,
ZFS_PROP_VERSION), spa_version);
} else {
zfs_error_aux(hdl, "%s", strerror(rc));
}
return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED,
dgettext(TEXT_DOMAIN, "cannot mount '%s'"),
zhp->zfs_name));
}
/* remove the mounted entry before re-adding on remount */
if (remount)
libzfs_mnttab_remove(hdl, zhp->zfs_name);
/* add the mounted entry into our cache */
libzfs_mnttab_add(hdl, zfs_get_name(zhp), mountpoint, mntopts);
return (0);
}
/*
* Unmount a single filesystem.
*/
static int
unmount_one(zfs_handle_t *zhp, const char *mountpoint, int flags)
{
int error;
error = do_unmount(zhp, mountpoint, flags);
if (error != 0) {
int libzfs_err;
switch (error) {
case EBUSY:
libzfs_err = EZFS_BUSY;
break;
case EIO:
libzfs_err = EZFS_IO;
break;
case ENOENT:
libzfs_err = EZFS_NOENT;
break;
case ENOMEM:
libzfs_err = EZFS_NOMEM;
break;
case EPERM:
libzfs_err = EZFS_PERM;
break;
default:
libzfs_err = EZFS_UMOUNTFAILED;
}
if (zhp) {
return (zfs_error_fmt(zhp->zfs_hdl, libzfs_err,
dgettext(TEXT_DOMAIN, "cannot unmount '%s'"),
mountpoint));
} else {
return (-1);
}
}
return (0);
}
/*
* Unmount the given filesystem.
*/
int
zfs_unmount(zfs_handle_t *zhp, const char *mountpoint, int flags)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
struct mnttab entry;
char *mntpt = NULL;
boolean_t encroot, unmounted = B_FALSE;
/* check to see if we need to unmount the filesystem */
if (mountpoint != NULL || ((zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) &&
libzfs_mnttab_find(hdl, zhp->zfs_name, &entry) == 0)) {
/*
* mountpoint may have come from a call to
* getmnt/getmntany if it isn't NULL. If it is NULL,
* we know it comes from libzfs_mnttab_find which can
* then get freed later. We strdup it to play it safe.
*/
if (mountpoint == NULL)
mntpt = zfs_strdup(hdl, entry.mnt_mountp);
else
mntpt = zfs_strdup(hdl, mountpoint);
/*
* Unshare and unmount the filesystem
*/
if (zfs_unshare(zhp, mntpt, share_all_proto) != 0) {
free(mntpt);
return (-1);
}
zfs_commit_shares(NULL);
if (unmount_one(zhp, mntpt, flags) != 0) {
free(mntpt);
(void) zfs_share(zhp, NULL);
zfs_commit_shares(NULL);
return (-1);
}
libzfs_mnttab_remove(hdl, zhp->zfs_name);
free(mntpt);
unmounted = B_TRUE;
}
/*
* If the MS_CRYPT flag is provided we must ensure we attempt to
* unload the dataset's key regardless of whether we did any work
* to unmount it. We only do this for encryption roots.
*/
if ((flags & MS_CRYPT) != 0 &&
zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION) != ZIO_CRYPT_OFF) {
zfs_refresh_properties(zhp);
if (zfs_crypto_get_encryption_root(zhp, &encroot, NULL) != 0 &&
unmounted) {
(void) zfs_mount(zhp, NULL, 0);
return (-1);
}
if (encroot && zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS) ==
ZFS_KEYSTATUS_AVAILABLE &&
zfs_crypto_unload_key(zhp) != 0) {
(void) zfs_mount(zhp, NULL, 0);
return (-1);
}
}
zpool_disable_volume_os(zhp->zfs_name);
return (0);
}
/*
* Unmount this filesystem and any children inheriting the mountpoint property.
* To do this, just act like we're changing the mountpoint property, but don't
* remount the filesystems afterwards.
*/
int
zfs_unmountall(zfs_handle_t *zhp, int flags)
{
prop_changelist_t *clp;
int ret;
clp = changelist_gather(zhp, ZFS_PROP_MOUNTPOINT,
CL_GATHER_ITER_MOUNTED, flags);
if (clp == NULL)
return (-1);
ret = changelist_prefix(clp);
changelist_free(clp);
return (ret);
}
/*
* Unshare a filesystem by mountpoint.
*/
static int
unshare_one(libzfs_handle_t *hdl, const char *name, const char *mountpoint,
enum sa_protocol proto)
{
int err = sa_disable_share(mountpoint, proto);
if (err != SA_OK)
return (zfs_error_fmt(hdl, proto_table[proto].p_unshare_err,
dgettext(TEXT_DOMAIN, "cannot unshare '%s': %s"),
name, sa_errorstr(err)));
return (0);
}
/*
* Share the given filesystem according to the options in the specified
* protocol specific properties (sharenfs, sharesmb). We rely
* on "libshare" to do the dirty work for us.
*/
int
zfs_share(zfs_handle_t *zhp, const enum sa_protocol *proto)
{
char mountpoint[ZFS_MAXPROPLEN];
char shareopts[ZFS_MAXPROPLEN];
char sourcestr[ZFS_MAXPROPLEN];
const enum sa_protocol *curr_proto;
zprop_source_t sourcetype;
int err = 0;
if (proto == NULL)
proto = share_all_proto;
if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint), NULL, 0))
return (0);
for (curr_proto = proto; *curr_proto != SA_NO_PROTOCOL; curr_proto++) {
/*
* Return success if there are no share options.
*/
if (zfs_prop_get(zhp, proto_table[*curr_proto].p_prop,
shareopts, sizeof (shareopts), &sourcetype, sourcestr,
ZFS_MAXPROPLEN, B_FALSE) != 0 ||
strcmp(shareopts, "off") == 0)
continue;
/*
* If the 'zoned' property is set, then zfs_is_mountable()
* will have already bailed out if we are in the global zone.
* But local zones cannot be NFS servers, so we ignore it for
* local zones as well.
*/
if (zfs_prop_get_int(zhp, ZFS_PROP_ZONED))
continue;
err = sa_enable_share(zfs_get_name(zhp), mountpoint, shareopts,
*curr_proto);
if (err != SA_OK) {
return (zfs_error_fmt(zhp->zfs_hdl,
proto_table[*curr_proto].p_share_err,
dgettext(TEXT_DOMAIN, "cannot share '%s: %s'"),
zfs_get_name(zhp), sa_errorstr(err)));
}
}
return (0);
}
/*
* Check to see if the filesystem is currently shared.
*/
boolean_t
zfs_is_shared(zfs_handle_t *zhp, char **where,
const enum sa_protocol *proto)
{
char *mountpoint;
if (proto == NULL)
proto = share_all_proto;
if (ZFS_IS_VOLUME(zhp))
return (B_FALSE);
if (!zfs_is_mounted(zhp, &mountpoint))
return (B_FALSE);
for (const enum sa_protocol *p = proto; *p != SA_NO_PROTOCOL; ++p)
if (sa_is_shared(mountpoint, *p)) {
if (where != NULL)
*where = mountpoint;
else
free(mountpoint);
return (B_TRUE);
}
free(mountpoint);
return (B_FALSE);
}
void
zfs_commit_shares(const enum sa_protocol *proto)
{
if (proto == NULL)
proto = share_all_proto;
for (const enum sa_protocol *p = proto; *p != SA_NO_PROTOCOL; ++p)
sa_commit_shares(*p);
}
void
zfs_truncate_shares(const enum sa_protocol *proto)
{
if (proto == NULL)
proto = share_all_proto;
for (const enum sa_protocol *p = proto; *p != SA_NO_PROTOCOL; ++p)
sa_truncate_shares(*p);
}
/*
* Unshare the given filesystem.
*/
int
zfs_unshare(zfs_handle_t *zhp, const char *mountpoint,
const enum sa_protocol *proto)
{
libzfs_handle_t *hdl = zhp->zfs_hdl;
struct mnttab entry;
if (proto == NULL)
proto = share_all_proto;
if (mountpoint != NULL || ((zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) &&
libzfs_mnttab_find(hdl, zfs_get_name(zhp), &entry) == 0)) {
/* check to see if need to unmount the filesystem */
const char *mntpt = mountpoint ?: entry.mnt_mountp;
for (const enum sa_protocol *curr_proto = proto;
*curr_proto != SA_NO_PROTOCOL; curr_proto++)
if (sa_is_shared(mntpt, *curr_proto) &&
unshare_one(hdl, zhp->zfs_name,
mntpt, *curr_proto) != 0)
return (-1);
}
return (0);
}
/*
* Same as zfs_unmountall(), but for NFS and SMB unshares.
*/
int
zfs_unshareall(zfs_handle_t *zhp, const enum sa_protocol *proto)
{
prop_changelist_t *clp;
int ret;
if (proto == NULL)
proto = share_all_proto;
clp = changelist_gather(zhp, ZFS_PROP_SHARENFS, 0, 0);
if (clp == NULL)
return (-1);
ret = changelist_unshare(clp, proto);
changelist_free(clp);
return (ret);
}
/*
* Remove the mountpoint associated with the current dataset, if necessary.
* We only remove the underlying directory if:
*
* - The mountpoint is not 'none' or 'legacy'
* - The mountpoint is non-empty
* - The mountpoint is the default or inherited
* - The 'zoned' property is set, or we're in a local zone
*
* Any other directories we leave alone.
*/
void
remove_mountpoint(zfs_handle_t *zhp)
{
char mountpoint[ZFS_MAXPROPLEN];
zprop_source_t source;
if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint),
&source, 0))
return;
if (source == ZPROP_SRC_DEFAULT ||
source == ZPROP_SRC_INHERITED) {
/*
* Try to remove the directory, silently ignoring any errors.
* The filesystem may have since been removed or moved around,
* and this error isn't really useful to the administrator in
* any way.
*/
(void) rmdir(mountpoint);
}
}
/*
* Add the given zfs handle to the cb_handles array, dynamically reallocating
* the array if it is out of space.
*/
void
libzfs_add_handle(get_all_cb_t *cbp, zfs_handle_t *zhp)
{
if (cbp->cb_alloc == cbp->cb_used) {
size_t newsz;
zfs_handle_t **newhandles;
newsz = cbp->cb_alloc != 0 ? cbp->cb_alloc * 2 : 64;
newhandles = zfs_realloc(zhp->zfs_hdl,
cbp->cb_handles, cbp->cb_alloc * sizeof (zfs_handle_t *),
newsz * sizeof (zfs_handle_t *));
cbp->cb_handles = newhandles;
cbp->cb_alloc = newsz;
}
cbp->cb_handles[cbp->cb_used++] = zhp;
}
/*
* Recursive helper function used during file system enumeration
*/
static int
zfs_iter_cb(zfs_handle_t *zhp, void *data)
{
get_all_cb_t *cbp = data;
if (!(zfs_get_type(zhp) & ZFS_TYPE_FILESYSTEM)) {
zfs_close(zhp);
return (0);
}
if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) == ZFS_CANMOUNT_NOAUTO) {
zfs_close(zhp);
return (0);
}
if (zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS) ==
ZFS_KEYSTATUS_UNAVAILABLE) {
zfs_close(zhp);
return (0);
}
/*
* If this filesystem is inconsistent and has a receive resume
* token, we can not mount it.
*/
if (zfs_prop_get_int(zhp, ZFS_PROP_INCONSISTENT) &&
zfs_prop_get(zhp, ZFS_PROP_RECEIVE_RESUME_TOKEN,
NULL, 0, NULL, NULL, 0, B_TRUE) == 0) {
zfs_close(zhp);
return (0);
}
libzfs_add_handle(cbp, zhp);
- if (zfs_iter_filesystems(zhp, 0, zfs_iter_cb, cbp) != 0) {
+ if (zfs_iter_filesystems_v2(zhp, 0, zfs_iter_cb, cbp) != 0) {
zfs_close(zhp);
return (-1);
}
return (0);
}
/*
* Sort comparator that compares two mountpoint paths. We sort these paths so
* that subdirectories immediately follow their parents. This means that we
* effectively treat the '/' character as the lowest value non-nul char.
* Since filesystems from non-global zones can have the same mountpoint
* as other filesystems, the comparator sorts global zone filesystems to
* the top of the list. This means that the global zone will traverse the
* filesystem list in the correct order and can stop when it sees the
* first zoned filesystem. In a non-global zone, only the delegated
* filesystems are seen.
*
* An example sorted list using this comparator would look like:
*
* /foo
* /foo/bar
* /foo/bar/baz
* /foo/baz
* /foo.bar
* /foo (NGZ1)
* /foo (NGZ2)
*
* The mounting code depends on this ordering to deterministically iterate
* over filesystems in order to spawn parallel mount tasks.
*/
static int
mountpoint_cmp(const void *arga, const void *argb)
{
zfs_handle_t *const *zap = arga;
zfs_handle_t *za = *zap;
zfs_handle_t *const *zbp = argb;
zfs_handle_t *zb = *zbp;
char mounta[MAXPATHLEN];
char mountb[MAXPATHLEN];
const char *a = mounta;
const char *b = mountb;
boolean_t gota, gotb;
uint64_t zoneda, zonedb;
zoneda = zfs_prop_get_int(za, ZFS_PROP_ZONED);
zonedb = zfs_prop_get_int(zb, ZFS_PROP_ZONED);
if (zoneda && !zonedb)
return (1);
if (!zoneda && zonedb)
return (-1);
gota = (zfs_get_type(za) == ZFS_TYPE_FILESYSTEM);
if (gota) {
verify(zfs_prop_get(za, ZFS_PROP_MOUNTPOINT, mounta,
sizeof (mounta), NULL, NULL, 0, B_FALSE) == 0);
}
gotb = (zfs_get_type(zb) == ZFS_TYPE_FILESYSTEM);
if (gotb) {
verify(zfs_prop_get(zb, ZFS_PROP_MOUNTPOINT, mountb,
sizeof (mountb), NULL, NULL, 0, B_FALSE) == 0);
}
if (gota && gotb) {
while (*a != '\0' && (*a == *b)) {
a++;
b++;
}
if (*a == *b)
return (0);
if (*a == '\0')
return (-1);
if (*b == '\0')
return (1);
if (*a == '/')
return (-1);
if (*b == '/')
return (1);
return (*a < *b ? -1 : *a > *b);
}
if (gota)
return (-1);
if (gotb)
return (1);
/*
* If neither filesystem has a mountpoint, revert to sorting by
* dataset name.
*/
return (strcmp(zfs_get_name(za), zfs_get_name(zb)));
}
/*
* Return true if path2 is a child of path1 or path2 equals path1 or
* path1 is "/" (path2 is always a child of "/").
*/
static boolean_t
libzfs_path_contains(const char *path1, const char *path2)
{
return (strcmp(path1, path2) == 0 || strcmp(path1, "/") == 0 ||
(strstr(path2, path1) == path2 && path2[strlen(path1)] == '/'));
}
/*
* Given a mountpoint specified by idx in the handles array, find the first
* non-descendent of that mountpoint and return its index. Descendant paths
* start with the parent's path. This function relies on the ordering
* enforced by mountpoint_cmp().
*/
static int
non_descendant_idx(zfs_handle_t **handles, size_t num_handles, int idx)
{
char parent[ZFS_MAXPROPLEN];
char child[ZFS_MAXPROPLEN];
int i;
verify(zfs_prop_get(handles[idx], ZFS_PROP_MOUNTPOINT, parent,
sizeof (parent), NULL, NULL, 0, B_FALSE) == 0);
for (i = idx + 1; i < num_handles; i++) {
verify(zfs_prop_get(handles[i], ZFS_PROP_MOUNTPOINT, child,
sizeof (child), NULL, NULL, 0, B_FALSE) == 0);
if (!libzfs_path_contains(parent, child))
break;
}
return (i);
}
typedef struct mnt_param {
libzfs_handle_t *mnt_hdl;
tpool_t *mnt_tp;
zfs_handle_t **mnt_zhps; /* filesystems to mount */
size_t mnt_num_handles;
int mnt_idx; /* Index of selected entry to mount */
zfs_iter_f mnt_func;
void *mnt_data;
} mnt_param_t;
/*
* Allocate and populate the parameter struct for mount function, and
* schedule mounting of the entry selected by idx.
*/
static void
zfs_dispatch_mount(libzfs_handle_t *hdl, zfs_handle_t **handles,
size_t num_handles, int idx, zfs_iter_f func, void *data, tpool_t *tp)
{
mnt_param_t *mnt_param = zfs_alloc(hdl, sizeof (mnt_param_t));
mnt_param->mnt_hdl = hdl;
mnt_param->mnt_tp = tp;
mnt_param->mnt_zhps = handles;
mnt_param->mnt_num_handles = num_handles;
mnt_param->mnt_idx = idx;
mnt_param->mnt_func = func;
mnt_param->mnt_data = data;
(void) tpool_dispatch(tp, zfs_mount_task, (void*)mnt_param);
}
/*
* This is the structure used to keep state of mounting or sharing operations
* during a call to zpool_enable_datasets().
*/
typedef struct mount_state {
/*
* ms_mntstatus is set to -1 if any mount fails. While multiple threads
* could update this variable concurrently, no synchronization is
* needed as it's only ever set to -1.
*/
int ms_mntstatus;
int ms_mntflags;
const char *ms_mntopts;
} mount_state_t;
static int
zfs_mount_one(zfs_handle_t *zhp, void *arg)
{
mount_state_t *ms = arg;
int ret = 0;
/*
* don't attempt to mount encrypted datasets with
* unloaded keys
*/
if (zfs_prop_get_int(zhp, ZFS_PROP_KEYSTATUS) ==
ZFS_KEYSTATUS_UNAVAILABLE)
return (0);
if (zfs_mount(zhp, ms->ms_mntopts, ms->ms_mntflags) != 0)
ret = ms->ms_mntstatus = -1;
return (ret);
}
static int
zfs_share_one(zfs_handle_t *zhp, void *arg)
{
mount_state_t *ms = arg;
int ret = 0;
if (zfs_share(zhp, NULL) != 0)
ret = ms->ms_mntstatus = -1;
return (ret);
}
/*
* Thread pool function to mount one file system. On completion, it finds and
* schedules its children to be mounted. This depends on the sorting done in
* zfs_foreach_mountpoint(). Note that the degenerate case (chain of entries
* each descending from the previous) will have no parallelism since we always
* have to wait for the parent to finish mounting before we can schedule
* its children.
*/
static void
zfs_mount_task(void *arg)
{
mnt_param_t *mp = arg;
int idx = mp->mnt_idx;
zfs_handle_t **handles = mp->mnt_zhps;
size_t num_handles = mp->mnt_num_handles;
char mountpoint[ZFS_MAXPROPLEN];
verify(zfs_prop_get(handles[idx], ZFS_PROP_MOUNTPOINT, mountpoint,
sizeof (mountpoint), NULL, NULL, 0, B_FALSE) == 0);
if (mp->mnt_func(handles[idx], mp->mnt_data) != 0)
goto out;
/*
* We dispatch tasks to mount filesystems with mountpoints underneath
* this one. We do this by dispatching the next filesystem with a
* descendant mountpoint of the one we just mounted, then skip all of
* its descendants, dispatch the next descendant mountpoint, and so on.
* The non_descendant_idx() function skips over filesystems that are
* descendants of the filesystem we just dispatched.
*/
for (int i = idx + 1; i < num_handles;
i = non_descendant_idx(handles, num_handles, i)) {
char child[ZFS_MAXPROPLEN];
verify(zfs_prop_get(handles[i], ZFS_PROP_MOUNTPOINT,
child, sizeof (child), NULL, NULL, 0, B_FALSE) == 0);
if (!libzfs_path_contains(mountpoint, child))
break; /* not a descendant, return */
zfs_dispatch_mount(mp->mnt_hdl, handles, num_handles, i,
mp->mnt_func, mp->mnt_data, mp->mnt_tp);
}
out:
free(mp);
}
/*
* Issue the func callback for each ZFS handle contained in the handles
* array. This function is used to mount all datasets, and so this function
* guarantees that filesystems for parent mountpoints are called before their
* children. As such, before issuing any callbacks, we first sort the array
* of handles by mountpoint.
*
* Callbacks are issued in one of two ways:
*
* 1. Sequentially: If the parallel argument is B_FALSE or the ZFS_SERIAL_MOUNT
* environment variable is set, then we issue callbacks sequentially.
*
* 2. In parallel: If the parallel argument is B_TRUE and the ZFS_SERIAL_MOUNT
* environment variable is not set, then we use a tpool to dispatch threads
* to mount filesystems in parallel. This function dispatches tasks to mount
* the filesystems at the top-level mountpoints, and these tasks in turn
* are responsible for recursively mounting filesystems in their children
* mountpoints.
*/
void
zfs_foreach_mountpoint(libzfs_handle_t *hdl, zfs_handle_t **handles,
size_t num_handles, zfs_iter_f func, void *data, boolean_t parallel)
{
zoneid_t zoneid = getzoneid();
/*
* The ZFS_SERIAL_MOUNT environment variable is an undocumented
* variable that can be used as a convenience to do a/b comparison
* of serial vs. parallel mounting.
*/
boolean_t serial_mount = !parallel ||
(getenv("ZFS_SERIAL_MOUNT") != NULL);
/*
* Sort the datasets by mountpoint. See mountpoint_cmp for details
* of how these are sorted.
*/
qsort(handles, num_handles, sizeof (zfs_handle_t *), mountpoint_cmp);
if (serial_mount) {
for (int i = 0; i < num_handles; i++) {
func(handles[i], data);
}
return;
}
/*
* Issue the callback function for each dataset using a parallel
* algorithm that uses a thread pool to manage threads.
*/
tpool_t *tp = tpool_create(1, mount_tp_nthr, 0, NULL);
/*
* There may be multiple "top level" mountpoints outside of the pool's
* root mountpoint, e.g.: /foo /bar. Dispatch a mount task for each of
* these.
*/
for (int i = 0; i < num_handles;
i = non_descendant_idx(handles, num_handles, i)) {
/*
* Since the mountpoints have been sorted so that the zoned
* filesystems are at the end, a zoned filesystem seen from
* the global zone means that we're done.
*/
if (zoneid == GLOBAL_ZONEID &&
zfs_prop_get_int(handles[i], ZFS_PROP_ZONED))
break;
zfs_dispatch_mount(hdl, handles, num_handles, i, func, data,
tp);
}
tpool_wait(tp); /* wait for all scheduled mounts to complete */
tpool_destroy(tp);
}
/*
* Mount and share all datasets within the given pool. This assumes that no
* datasets within the pool are currently mounted.
*/
int
zpool_enable_datasets(zpool_handle_t *zhp, const char *mntopts, int flags)
{
get_all_cb_t cb = { 0 };
mount_state_t ms = { 0 };
zfs_handle_t *zfsp;
int ret = 0;
if ((zfsp = zfs_open(zhp->zpool_hdl, zhp->zpool_name,
ZFS_TYPE_DATASET)) == NULL)
goto out;
/*
* Gather all non-snapshot datasets within the pool. Start by adding
* the root filesystem for this pool to the list, and then iterate
* over all child filesystems.
*/
libzfs_add_handle(&cb, zfsp);
- if (zfs_iter_filesystems(zfsp, 0, zfs_iter_cb, &cb) != 0)
+ if (zfs_iter_filesystems_v2(zfsp, 0, zfs_iter_cb, &cb) != 0)
goto out;
/*
* Mount all filesystems
*/
ms.ms_mntopts = mntopts;
ms.ms_mntflags = flags;
zfs_foreach_mountpoint(zhp->zpool_hdl, cb.cb_handles, cb.cb_used,
zfs_mount_one, &ms, B_TRUE);
if (ms.ms_mntstatus != 0)
ret = ms.ms_mntstatus;
/*
* Share all filesystems that need to be shared. This needs to be
* a separate pass because libshare is not mt-safe, and so we need
* to share serially.
*/
ms.ms_mntstatus = 0;
zfs_foreach_mountpoint(zhp->zpool_hdl, cb.cb_handles, cb.cb_used,
zfs_share_one, &ms, B_FALSE);
if (ms.ms_mntstatus != 0)
ret = ms.ms_mntstatus;
else
zfs_commit_shares(NULL);
out:
for (int i = 0; i < cb.cb_used; i++)
zfs_close(cb.cb_handles[i]);
free(cb.cb_handles);
return (ret);
}
struct sets_s {
char *mountpoint;
zfs_handle_t *dataset;
};
static int
mountpoint_compare(const void *a, const void *b)
{
const struct sets_s *mounta = (struct sets_s *)a;
const struct sets_s *mountb = (struct sets_s *)b;
return (strcmp(mountb->mountpoint, mounta->mountpoint));
}
/*
* Unshare and unmount all datasets within the given pool. We don't want to
* rely on traversing the DSL to discover the filesystems within the pool,
* because this may be expensive (if not all of them are mounted), and can fail
* arbitrarily (on I/O error, for example). Instead, we walk /proc/self/mounts
* and gather all the filesystems that are currently mounted.
*/
int
zpool_disable_datasets(zpool_handle_t *zhp, boolean_t force)
{
int used, alloc;
FILE *mnttab;
struct mnttab entry;
size_t namelen;
struct sets_s *sets = NULL;
libzfs_handle_t *hdl = zhp->zpool_hdl;
int i;
int ret = -1;
int flags = (force ? MS_FORCE : 0);
namelen = strlen(zhp->zpool_name);
if ((mnttab = fopen(MNTTAB, "re")) == NULL)
return (ENOENT);
used = alloc = 0;
while (getmntent(mnttab, &entry) == 0) {
/*
* Ignore non-ZFS entries.
*/
if (entry.mnt_fstype == NULL ||
strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0)
continue;
/*
* Ignore filesystems not within this pool.
*/
if (entry.mnt_mountp == NULL ||
strncmp(entry.mnt_special, zhp->zpool_name, namelen) != 0 ||
(entry.mnt_special[namelen] != '/' &&
entry.mnt_special[namelen] != '\0'))
continue;
/*
* At this point we've found a filesystem within our pool. Add
* it to our growing list.
*/
if (used == alloc) {
if (alloc == 0) {
sets = zfs_alloc(hdl,
8 * sizeof (struct sets_s));
alloc = 8;
} else {
sets = zfs_realloc(hdl, sets,
alloc * sizeof (struct sets_s),
alloc * 2 * sizeof (struct sets_s));
alloc *= 2;
}
}
sets[used].mountpoint = zfs_strdup(hdl, entry.mnt_mountp);
/*
* This is allowed to fail, in case there is some I/O error. It
* is only used to determine if we need to remove the underlying
* mountpoint, so failure is not fatal.
*/
sets[used].dataset = make_dataset_handle(hdl,
entry.mnt_special);
used++;
}
/*
* At this point, we have the entire list of filesystems, so sort it by
* mountpoint.
*/
if (used != 0)
qsort(sets, used, sizeof (struct sets_s), mountpoint_compare);
/*
* Walk through and first unshare everything.
*/
for (i = 0; i < used; i++) {
for (enum sa_protocol p = 0; p < SA_PROTOCOL_COUNT; ++p) {
if (sa_is_shared(sets[i].mountpoint, p) &&
unshare_one(hdl, sets[i].mountpoint,
sets[i].mountpoint, p) != 0)
goto out;
}
}
zfs_commit_shares(NULL);
/*
* Now unmount everything, removing the underlying directories as
* appropriate.
*/
for (i = 0; i < used; i++) {
if (unmount_one(sets[i].dataset, sets[i].mountpoint,
flags) != 0)
goto out;
}
for (i = 0; i < used; i++) {
if (sets[i].dataset)
remove_mountpoint(sets[i].dataset);
}
zpool_disable_datasets_os(zhp, force);
ret = 0;
out:
(void) fclose(mnttab);
for (i = 0; i < used; i++) {
if (sets[i].dataset)
zfs_close(sets[i].dataset);
free(sets[i].mountpoint);
}
free(sets);
return (ret);
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c b/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c
index 6281044723a1..4fb71b4e0dc8 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c
@@ -1,5320 +1,5321 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2015 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>
* Copyright (c) 2018 Datto Inc.
* Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2018, loli10K <ezomori.nozomu@gmail.com>
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
* Copyright (c) 2021, Klara Inc.
*/
#include <errno.h>
#include <libintl.h>
#include <stdio.h>
#include <stdlib.h>
#include <strings.h>
#include <unistd.h>
#include <libgen.h>
#include <zone.h>
#include <sys/stat.h>
#include <sys/efi_partition.h>
#include <sys/systeminfo.h>
#include <sys/zfs_ioctl.h>
#include <sys/zfs_sysfs.h>
#include <sys/vdev_disk.h>
#include <sys/types.h>
#include <dlfcn.h>
#include <libzutil.h>
#include <fcntl.h>
#include "zfs_namecheck.h"
#include "zfs_prop.h"
#include "libzfs_impl.h"
#include "zfs_comutil.h"
#include "zfeature_common.h"
static boolean_t zpool_vdev_is_interior(const char *name);
typedef struct prop_flags {
unsigned int create:1; /* Validate property on creation */
unsigned int import:1; /* Validate property on import */
unsigned int vdevprop:1; /* Validate property as a VDEV property */
} prop_flags_t;
/*
* ====================================================================
* zpool property functions
* ====================================================================
*/
static int
zpool_get_all_props(zpool_handle_t *zhp)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zcmd_alloc_dst_nvlist(hdl, &zc, 0);
while (zfs_ioctl(hdl, ZFS_IOC_POOL_GET_PROPS, &zc) != 0) {
if (errno == ENOMEM)
zcmd_expand_dst_nvlist(hdl, &zc);
else {
zcmd_free_nvlists(&zc);
return (-1);
}
}
if (zcmd_read_dst_nvlist(hdl, &zc, &zhp->zpool_props) != 0) {
zcmd_free_nvlists(&zc);
return (-1);
}
zcmd_free_nvlists(&zc);
return (0);
}
int
zpool_props_refresh(zpool_handle_t *zhp)
{
nvlist_t *old_props;
old_props = zhp->zpool_props;
if (zpool_get_all_props(zhp) != 0)
return (-1);
nvlist_free(old_props);
return (0);
}
static const char *
zpool_get_prop_string(zpool_handle_t *zhp, zpool_prop_t prop,
zprop_source_t *src)
{
nvlist_t *nv, *nvl;
const char *value;
zprop_source_t source;
nvl = zhp->zpool_props;
if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) {
source = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
value = fnvlist_lookup_string(nv, ZPROP_VALUE);
} else {
source = ZPROP_SRC_DEFAULT;
if ((value = zpool_prop_default_string(prop)) == NULL)
value = "-";
}
if (src)
*src = source;
return (value);
}
uint64_t
zpool_get_prop_int(zpool_handle_t *zhp, zpool_prop_t prop, zprop_source_t *src)
{
nvlist_t *nv, *nvl;
uint64_t value;
zprop_source_t source;
if (zhp->zpool_props == NULL && zpool_get_all_props(zhp)) {
/*
* zpool_get_all_props() has most likely failed because
* the pool is faulted, but if all we need is the top level
* vdev's guid then get it from the zhp config nvlist.
*/
if ((prop == ZPOOL_PROP_GUID) &&
(nvlist_lookup_nvlist(zhp->zpool_config,
ZPOOL_CONFIG_VDEV_TREE, &nv) == 0) &&
(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &value)
== 0)) {
return (value);
}
return (zpool_prop_default_numeric(prop));
}
nvl = zhp->zpool_props;
if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) {
source = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
value = fnvlist_lookup_uint64(nv, ZPROP_VALUE);
} else {
source = ZPROP_SRC_DEFAULT;
value = zpool_prop_default_numeric(prop);
}
if (src)
*src = source;
return (value);
}
/*
* Map VDEV STATE to printed strings.
*/
const char *
zpool_state_to_name(vdev_state_t state, vdev_aux_t aux)
{
switch (state) {
case VDEV_STATE_CLOSED:
case VDEV_STATE_OFFLINE:
return (gettext("OFFLINE"));
case VDEV_STATE_REMOVED:
return (gettext("REMOVED"));
case VDEV_STATE_CANT_OPEN:
if (aux == VDEV_AUX_CORRUPT_DATA || aux == VDEV_AUX_BAD_LOG)
return (gettext("FAULTED"));
else if (aux == VDEV_AUX_SPLIT_POOL)
return (gettext("SPLIT"));
else
return (gettext("UNAVAIL"));
case VDEV_STATE_FAULTED:
return (gettext("FAULTED"));
case VDEV_STATE_DEGRADED:
return (gettext("DEGRADED"));
case VDEV_STATE_HEALTHY:
return (gettext("ONLINE"));
default:
break;
}
return (gettext("UNKNOWN"));
}
/*
* Map POOL STATE to printed strings.
*/
const char *
zpool_pool_state_to_name(pool_state_t state)
{
switch (state) {
default:
break;
case POOL_STATE_ACTIVE:
return (gettext("ACTIVE"));
case POOL_STATE_EXPORTED:
return (gettext("EXPORTED"));
case POOL_STATE_DESTROYED:
return (gettext("DESTROYED"));
case POOL_STATE_SPARE:
return (gettext("SPARE"));
case POOL_STATE_L2CACHE:
return (gettext("L2CACHE"));
case POOL_STATE_UNINITIALIZED:
return (gettext("UNINITIALIZED"));
case POOL_STATE_UNAVAIL:
return (gettext("UNAVAIL"));
case POOL_STATE_POTENTIALLY_ACTIVE:
return (gettext("POTENTIALLY_ACTIVE"));
}
return (gettext("UNKNOWN"));
}
/*
* Given a pool handle, return the pool health string ("ONLINE", "DEGRADED",
* "SUSPENDED", etc).
*/
const char *
zpool_get_state_str(zpool_handle_t *zhp)
{
zpool_errata_t errata;
zpool_status_t status;
const char *str;
status = zpool_get_status(zhp, NULL, &errata);
if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) {
str = gettext("FAULTED");
} else if (status == ZPOOL_STATUS_IO_FAILURE_WAIT ||
status == ZPOOL_STATUS_IO_FAILURE_MMP) {
str = gettext("SUSPENDED");
} else {
nvlist_t *nvroot = fnvlist_lookup_nvlist(
zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE);
uint_t vsc;
vdev_stat_t *vs = (vdev_stat_t *)fnvlist_lookup_uint64_array(
nvroot, ZPOOL_CONFIG_VDEV_STATS, &vsc);
str = zpool_state_to_name(vs->vs_state, vs->vs_aux);
}
return (str);
}
/*
* Get a zpool property value for 'prop' and return the value in
* a pre-allocated buffer.
*/
int
zpool_get_prop(zpool_handle_t *zhp, zpool_prop_t prop, char *buf,
size_t len, zprop_source_t *srctype, boolean_t literal)
{
uint64_t intval;
const char *strval;
zprop_source_t src = ZPROP_SRC_NONE;
if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) {
switch (prop) {
case ZPOOL_PROP_NAME:
(void) strlcpy(buf, zpool_get_name(zhp), len);
break;
case ZPOOL_PROP_HEALTH:
(void) strlcpy(buf, zpool_get_state_str(zhp), len);
break;
case ZPOOL_PROP_GUID:
intval = zpool_get_prop_int(zhp, prop, &src);
(void) snprintf(buf, len, "%llu", (u_longlong_t)intval);
break;
case ZPOOL_PROP_ALTROOT:
case ZPOOL_PROP_CACHEFILE:
case ZPOOL_PROP_COMMENT:
case ZPOOL_PROP_COMPATIBILITY:
if (zhp->zpool_props != NULL ||
zpool_get_all_props(zhp) == 0) {
(void) strlcpy(buf,
zpool_get_prop_string(zhp, prop, &src),
len);
break;
}
zfs_fallthrough;
default:
(void) strlcpy(buf, "-", len);
break;
}
if (srctype != NULL)
*srctype = src;
return (0);
}
if (zhp->zpool_props == NULL && zpool_get_all_props(zhp) &&
prop != ZPOOL_PROP_NAME)
return (-1);
switch (zpool_prop_get_type(prop)) {
case PROP_TYPE_STRING:
(void) strlcpy(buf, zpool_get_prop_string(zhp, prop, &src),
len);
break;
case PROP_TYPE_NUMBER:
intval = zpool_get_prop_int(zhp, prop, &src);
switch (prop) {
case ZPOOL_PROP_SIZE:
case ZPOOL_PROP_ALLOCATED:
case ZPOOL_PROP_FREE:
case ZPOOL_PROP_FREEING:
case ZPOOL_PROP_LEAKED:
case ZPOOL_PROP_ASHIFT:
case ZPOOL_PROP_MAXBLOCKSIZE:
case ZPOOL_PROP_MAXDNODESIZE:
case ZPOOL_PROP_BCLONESAVED:
case ZPOOL_PROP_BCLONEUSED:
if (literal)
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
else
(void) zfs_nicenum(intval, buf, len);
break;
case ZPOOL_PROP_EXPANDSZ:
case ZPOOL_PROP_CHECKPOINT:
if (intval == 0) {
(void) strlcpy(buf, "-", len);
} else if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) zfs_nicebytes(intval, buf, len);
}
break;
case ZPOOL_PROP_CAPACITY:
if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) snprintf(buf, len, "%llu%%",
(u_longlong_t)intval);
}
break;
case ZPOOL_PROP_FRAGMENTATION:
if (intval == UINT64_MAX) {
(void) strlcpy(buf, "-", len);
} else if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) snprintf(buf, len, "%llu%%",
(u_longlong_t)intval);
}
break;
case ZPOOL_PROP_BCLONERATIO:
case ZPOOL_PROP_DEDUPRATIO:
if (literal)
(void) snprintf(buf, len, "%llu.%02llu",
(u_longlong_t)(intval / 100),
(u_longlong_t)(intval % 100));
else
(void) snprintf(buf, len, "%llu.%02llux",
(u_longlong_t)(intval / 100),
(u_longlong_t)(intval % 100));
break;
case ZPOOL_PROP_HEALTH:
(void) strlcpy(buf, zpool_get_state_str(zhp), len);
break;
case ZPOOL_PROP_VERSION:
if (intval >= SPA_VERSION_FEATURES) {
(void) snprintf(buf, len, "-");
break;
}
zfs_fallthrough;
default:
(void) snprintf(buf, len, "%llu", (u_longlong_t)intval);
}
break;
case PROP_TYPE_INDEX:
intval = zpool_get_prop_int(zhp, prop, &src);
if (zpool_prop_index_to_string(prop, intval, &strval)
!= 0)
return (-1);
(void) strlcpy(buf, strval, len);
break;
default:
abort();
}
if (srctype)
*srctype = src;
return (0);
}
/*
* Get a zpool property value for 'propname' and return the value in
* a pre-allocated buffer.
*/
int
zpool_get_userprop(zpool_handle_t *zhp, const char *propname, char *buf,
size_t len, zprop_source_t *srctype)
{
nvlist_t *nv, *nvl;
uint64_t ival;
const char *value;
zprop_source_t source = ZPROP_SRC_LOCAL;
nvl = zhp->zpool_props;
if (nvlist_lookup_nvlist(nvl, propname, &nv) == 0) {
if (nvlist_lookup_uint64(nv, ZPROP_SOURCE, &ival) == 0)
source = ival;
verify(nvlist_lookup_string(nv, ZPROP_VALUE, &value) == 0);
} else {
source = ZPROP_SRC_DEFAULT;
value = "-";
}
if (srctype)
*srctype = source;
(void) strlcpy(buf, value, len);
return (0);
}
/*
* Check if the bootfs name has the same pool name as it is set to.
* Assuming bootfs is a valid dataset name.
*/
static boolean_t
bootfs_name_valid(const char *pool, const char *bootfs)
{
int len = strlen(pool);
if (bootfs[0] == '\0')
return (B_TRUE);
if (!zfs_name_valid(bootfs, ZFS_TYPE_FILESYSTEM|ZFS_TYPE_SNAPSHOT))
return (B_FALSE);
if (strncmp(pool, bootfs, len) == 0 &&
(bootfs[len] == '/' || bootfs[len] == '\0'))
return (B_TRUE);
return (B_FALSE);
}
/*
* Given an nvlist of zpool properties to be set, validate that they are
* correct, and parse any numeric properties (index, boolean, etc) if they are
* specified as strings.
*/
static nvlist_t *
zpool_valid_proplist(libzfs_handle_t *hdl, const char *poolname,
nvlist_t *props, uint64_t version, prop_flags_t flags, char *errbuf)
{
nvpair_t *elem;
nvlist_t *retprops;
zpool_prop_t prop;
const char *strval;
uint64_t intval;
const char *slash, *check;
struct stat64 statbuf;
zpool_handle_t *zhp;
char report[1024];
if (nvlist_alloc(&retprops, NV_UNIQUE_NAME, 0) != 0) {
(void) no_memory(hdl);
return (NULL);
}
elem = NULL;
while ((elem = nvlist_next_nvpair(props, elem)) != NULL) {
const char *propname = nvpair_name(elem);
if (flags.vdevprop && zpool_prop_vdev(propname)) {
vdev_prop_t vprop = vdev_name_to_prop(propname);
if (vdev_prop_readonly(vprop)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' "
"is readonly"), propname);
(void) zfs_error(hdl, EZFS_PROPREADONLY,
errbuf);
goto error;
}
if (zprop_parse_value(hdl, elem, vprop, ZFS_TYPE_VDEV,
retprops, &strval, &intval, errbuf) != 0)
goto error;
continue;
} else if (flags.vdevprop && vdev_prop_user(propname)) {
if (nvlist_add_nvpair(retprops, elem) != 0) {
(void) no_memory(hdl);
goto error;
}
continue;
} else if (flags.vdevprop) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property: '%s'"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
prop = zpool_name_to_prop(propname);
if (prop == ZPOOL_PROP_INVAL && zpool_prop_feature(propname)) {
int err;
char *fname = strchr(propname, '@') + 1;
err = zfeature_lookup_name(fname, NULL);
if (err != 0) {
ASSERT3U(err, ==, ENOENT);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"feature '%s' unsupported by kernel"),
fname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (nvpair_type(elem) != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
(void) nvpair_value_string(elem, &strval);
if (strcmp(strval, ZFS_FEATURE_ENABLED) != 0 &&
strcmp(strval, ZFS_FEATURE_DISABLED) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set to "
"'enabled' or 'disabled'"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (!flags.create &&
strcmp(strval, ZFS_FEATURE_DISABLED) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set to "
"'disabled' at creation time"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (nvlist_add_uint64(retprops, propname, 0) != 0) {
(void) no_memory(hdl);
goto error;
}
continue;
} else if (prop == ZPOOL_PROP_INVAL &&
zfs_prop_user(propname)) {
/*
* This is a user property: make sure it's a
* string, and that it's less than ZAP_MAXNAMELEN.
*/
if (nvpair_type(elem) != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (strlen(nvpair_name(elem)) >= ZAP_MAXNAMELEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property name '%s' is too long"),
propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
(void) nvpair_value_string(elem, &strval);
if (strlen(strval) >= ZFS_MAXPROPLEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property value '%s' is too long"),
strval);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (nvlist_add_string(retprops, propname,
strval) != 0) {
(void) no_memory(hdl);
goto error;
}
continue;
}
/*
* Make sure this property is valid and applies to this type.
*/
if (prop == ZPOOL_PROP_INVAL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property '%s'"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (zpool_prop_readonly(prop)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' "
"is readonly"), propname);
(void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf);
goto error;
}
if (!flags.create && zpool_prop_setonce(prop)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set at "
"creation time"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (zprop_parse_value(hdl, elem, prop, ZFS_TYPE_POOL, retprops,
&strval, &intval, errbuf) != 0)
goto error;
/*
* Perform additional checking for specific properties.
*/
switch (prop) {
case ZPOOL_PROP_VERSION:
if (intval < version ||
!SPA_VERSION_IS_SUPPORTED(intval)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' number %llu is invalid."),
propname, (unsigned long long)intval);
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
goto error;
}
break;
case ZPOOL_PROP_ASHIFT:
if (intval != 0 &&
(intval < ASHIFT_MIN || intval > ASHIFT_MAX)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' number %llu is invalid, "
"only values between %" PRId32 " and %"
PRId32 " are allowed."),
propname, (unsigned long long)intval,
ASHIFT_MIN, ASHIFT_MAX);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_BOOTFS:
if (flags.create || flags.import) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' cannot be set at creation "
"or import time"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (version < SPA_VERSION_BOOTFS) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded to support "
"'%s' property"), propname);
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
goto error;
}
/*
* bootfs property value has to be a dataset name and
* the dataset has to be in the same pool as it sets to.
*/
if (!bootfs_name_valid(poolname, strval)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' "
"is an invalid name"), strval);
(void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
goto error;
}
if ((zhp = zpool_open_canfail(hdl, poolname)) == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"could not open pool '%s'"), poolname);
(void) zfs_error(hdl, EZFS_OPENFAILED, errbuf);
goto error;
}
zpool_close(zhp);
break;
case ZPOOL_PROP_ALTROOT:
if (!flags.create && !flags.import) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set during pool "
"creation or import"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
if (strval[0] != '/') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"bad alternate root '%s'"), strval);
(void) zfs_error(hdl, EZFS_BADPATH, errbuf);
goto error;
}
break;
case ZPOOL_PROP_CACHEFILE:
if (strval[0] == '\0')
break;
if (strcmp(strval, "none") == 0)
break;
if (strval[0] != '/') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' must be empty, an "
"absolute path, or 'none'"), propname);
(void) zfs_error(hdl, EZFS_BADPATH, errbuf);
goto error;
}
slash = strrchr(strval, '/');
if (slash[1] == '\0' || strcmp(slash, "/.") == 0 ||
strcmp(slash, "/..") == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is not a valid file"), strval);
(void) zfs_error(hdl, EZFS_BADPATH, errbuf);
goto error;
}
*(char *)slash = '\0';
if (strval[0] != '\0' &&
(stat64(strval, &statbuf) != 0 ||
!S_ISDIR(statbuf.st_mode))) {
*(char *)slash = '/';
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is not a valid directory"),
strval);
(void) zfs_error(hdl, EZFS_BADPATH, errbuf);
goto error;
}
*(char *)slash = '/';
break;
case ZPOOL_PROP_COMPATIBILITY:
switch (zpool_load_compat(strval, NULL, report, 1024)) {
case ZPOOL_COMPATIBILITY_OK:
case ZPOOL_COMPATIBILITY_WARNTOKEN:
break;
case ZPOOL_COMPATIBILITY_BADFILE:
case ZPOOL_COMPATIBILITY_BADTOKEN:
case ZPOOL_COMPATIBILITY_NOFILES:
zfs_error_aux(hdl, "%s", report);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_COMMENT:
for (check = strval; *check != '\0'; check++) {
if (!isprint(*check)) {
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN,
"comment may only have printable "
"characters"));
(void) zfs_error(hdl, EZFS_BADPROP,
errbuf);
goto error;
}
}
if (strlen(strval) > ZPROP_MAX_COMMENT) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"comment must not exceed %d characters"),
ZPROP_MAX_COMMENT);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_READONLY:
if (!flags.import) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' can only be set at "
"import time"), propname);
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_MULTIHOST:
if (get_system_hostid() == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"requires a non-zero system hostid"));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
break;
case ZPOOL_PROP_DEDUPDITTO:
printf("Note: property '%s' no longer has "
"any effect\n", propname);
break;
default:
break;
}
}
return (retprops);
error:
nvlist_free(retprops);
return (NULL);
}
/*
* Set zpool property : propname=propval.
*/
int
zpool_set_prop(zpool_handle_t *zhp, const char *propname, const char *propval)
{
zfs_cmd_t zc = {"\0"};
int ret = -1;
char errbuf[ERRBUFLEN];
nvlist_t *nvl = NULL;
nvlist_t *realprops;
uint64_t version;
prop_flags_t flags = { 0 };
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot set property for '%s'"),
zhp->zpool_name);
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
if (nvlist_add_string(nvl, propname, propval) != 0) {
nvlist_free(nvl);
return (no_memory(zhp->zpool_hdl));
}
version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL);
if ((realprops = zpool_valid_proplist(zhp->zpool_hdl,
zhp->zpool_name, nvl, version, flags, errbuf)) == NULL) {
nvlist_free(nvl);
return (-1);
}
nvlist_free(nvl);
nvl = realprops;
/*
* Execute the corresponding ioctl() to set this property.
*/
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zcmd_write_src_nvlist(zhp->zpool_hdl, &zc, nvl);
ret = zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_SET_PROPS, &zc);
zcmd_free_nvlists(&zc);
nvlist_free(nvl);
if (ret)
(void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf);
else
(void) zpool_props_refresh(zhp);
return (ret);
}
int
zpool_expand_proplist(zpool_handle_t *zhp, zprop_list_t **plp,
zfs_type_t type, boolean_t literal)
{
libzfs_handle_t *hdl = zhp->zpool_hdl;
zprop_list_t *entry;
char buf[ZFS_MAXPROPLEN];
nvlist_t *features = NULL;
nvpair_t *nvp;
zprop_list_t **last;
boolean_t firstexpand = (NULL == *plp);
int i;
if (zprop_expand_list(hdl, plp, type) != 0)
return (-1);
if (type == ZFS_TYPE_VDEV)
return (0);
last = plp;
while (*last != NULL)
last = &(*last)->pl_next;
if ((*plp)->pl_all)
features = zpool_get_features(zhp);
if ((*plp)->pl_all && firstexpand) {
/* Handle userprops in the all properties case */
if (zhp->zpool_props == NULL && zpool_props_refresh(zhp))
return (-1);
nvp = NULL;
while ((nvp = nvlist_next_nvpair(zhp->zpool_props, nvp)) !=
NULL) {
const char *propname = nvpair_name(nvp);
if (!zfs_prop_user(propname))
continue;
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_prop = ZPROP_USERPROP;
entry->pl_user_prop = zfs_strdup(hdl, propname);
entry->pl_width = strlen(entry->pl_user_prop);
entry->pl_all = B_TRUE;
*last = entry;
last = &entry->pl_next;
}
for (i = 0; i < SPA_FEATURES; i++) {
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_prop = ZPROP_USERPROP;
entry->pl_user_prop = zfs_asprintf(hdl, "feature@%s",
spa_feature_table[i].fi_uname);
entry->pl_width = strlen(entry->pl_user_prop);
entry->pl_all = B_TRUE;
*last = entry;
last = &entry->pl_next;
}
}
/* add any unsupported features */
for (nvp = nvlist_next_nvpair(features, NULL);
nvp != NULL; nvp = nvlist_next_nvpair(features, nvp)) {
char *propname;
boolean_t found;
if (zfeature_is_supported(nvpair_name(nvp)))
continue;
propname = zfs_asprintf(hdl, "unsupported@%s",
nvpair_name(nvp));
/*
* Before adding the property to the list make sure that no
* other pool already added the same property.
*/
found = B_FALSE;
entry = *plp;
while (entry != NULL) {
if (entry->pl_user_prop != NULL &&
strcmp(propname, entry->pl_user_prop) == 0) {
found = B_TRUE;
break;
}
entry = entry->pl_next;
}
if (found) {
free(propname);
continue;
}
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_prop = ZPROP_USERPROP;
entry->pl_user_prop = propname;
entry->pl_width = strlen(entry->pl_user_prop);
entry->pl_all = B_TRUE;
*last = entry;
last = &entry->pl_next;
}
for (entry = *plp; entry != NULL; entry = entry->pl_next) {
if (entry->pl_fixed && !literal)
continue;
if (entry->pl_prop != ZPROP_USERPROP &&
zpool_get_prop(zhp, entry->pl_prop, buf, sizeof (buf),
NULL, literal) == 0) {
if (strlen(buf) > entry->pl_width)
entry->pl_width = strlen(buf);
} else if (entry->pl_prop == ZPROP_INVAL &&
zfs_prop_user(entry->pl_user_prop) &&
zpool_get_userprop(zhp, entry->pl_user_prop, buf,
sizeof (buf), NULL) == 0) {
if (strlen(buf) > entry->pl_width)
entry->pl_width = strlen(buf);
}
}
return (0);
}
int
vdev_expand_proplist(zpool_handle_t *zhp, const char *vdevname,
zprop_list_t **plp)
{
zprop_list_t *entry;
char buf[ZFS_MAXPROPLEN];
const char *strval = NULL;
int err = 0;
nvpair_t *elem = NULL;
nvlist_t *vprops = NULL;
nvlist_t *propval = NULL;
const char *propname;
vdev_prop_t prop;
zprop_list_t **last;
for (entry = *plp; entry != NULL; entry = entry->pl_next) {
if (entry->pl_fixed)
continue;
if (zpool_get_vdev_prop(zhp, vdevname, entry->pl_prop,
entry->pl_user_prop, buf, sizeof (buf), NULL,
B_FALSE) == 0) {
if (strlen(buf) > entry->pl_width)
entry->pl_width = strlen(buf);
}
if (entry->pl_prop == VDEV_PROP_NAME &&
strlen(vdevname) > entry->pl_width)
entry->pl_width = strlen(vdevname);
}
/* Handle the all properties case */
last = plp;
if (*last != NULL && (*last)->pl_all == B_TRUE) {
while (*last != NULL)
last = &(*last)->pl_next;
err = zpool_get_all_vdev_props(zhp, vdevname, &vprops);
if (err != 0)
return (err);
while ((elem = nvlist_next_nvpair(vprops, elem)) != NULL) {
propname = nvpair_name(elem);
/* Skip properties that are not user defined */
if ((prop = vdev_name_to_prop(propname)) !=
VDEV_PROP_USERPROP)
continue;
if (nvpair_value_nvlist(elem, &propval) != 0)
continue;
strval = fnvlist_lookup_string(propval, ZPROP_VALUE);
entry = zfs_alloc(zhp->zpool_hdl,
sizeof (zprop_list_t));
entry->pl_prop = prop;
entry->pl_user_prop = zfs_strdup(zhp->zpool_hdl,
propname);
entry->pl_width = strlen(strval);
entry->pl_all = B_TRUE;
*last = entry;
last = &entry->pl_next;
}
}
return (0);
}
/*
* Get the state for the given feature on the given ZFS pool.
*/
int
zpool_prop_get_feature(zpool_handle_t *zhp, const char *propname, char *buf,
size_t len)
{
uint64_t refcount;
boolean_t found = B_FALSE;
nvlist_t *features = zpool_get_features(zhp);
boolean_t supported;
const char *feature = strchr(propname, '@') + 1;
supported = zpool_prop_feature(propname);
ASSERT(supported || zpool_prop_unsupported(propname));
/*
* Convert from feature name to feature guid. This conversion is
* unnecessary for unsupported@... properties because they already
* use guids.
*/
if (supported) {
int ret;
spa_feature_t fid;
ret = zfeature_lookup_name(feature, &fid);
if (ret != 0) {
(void) strlcpy(buf, "-", len);
return (ENOTSUP);
}
feature = spa_feature_table[fid].fi_guid;
}
if (nvlist_lookup_uint64(features, feature, &refcount) == 0)
found = B_TRUE;
if (supported) {
if (!found) {
(void) strlcpy(buf, ZFS_FEATURE_DISABLED, len);
} else {
if (refcount == 0)
(void) strlcpy(buf, ZFS_FEATURE_ENABLED, len);
else
(void) strlcpy(buf, ZFS_FEATURE_ACTIVE, len);
}
} else {
if (found) {
if (refcount == 0) {
(void) strcpy(buf, ZFS_UNSUPPORTED_INACTIVE);
} else {
(void) strcpy(buf, ZFS_UNSUPPORTED_READONLY);
}
} else {
(void) strlcpy(buf, "-", len);
return (ENOTSUP);
}
}
return (0);
}
/*
* Validate the given pool name, optionally putting an extended error message in
* 'buf'.
*/
boolean_t
zpool_name_valid(libzfs_handle_t *hdl, boolean_t isopen, const char *pool)
{
namecheck_err_t why;
char what;
int ret;
ret = pool_namecheck(pool, &why, &what);
/*
* The rules for reserved pool names were extended at a later point.
* But we need to support users with existing pools that may now be
* invalid. So we only check for this expanded set of names during a
* create (or import), and only in userland.
*/
if (ret == 0 && !isopen &&
(strncmp(pool, "mirror", 6) == 0 ||
strncmp(pool, "raidz", 5) == 0 ||
strncmp(pool, "draid", 5) == 0 ||
strncmp(pool, "spare", 5) == 0 ||
strcmp(pool, "log") == 0)) {
if (hdl != NULL)
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "name is reserved"));
return (B_FALSE);
}
if (ret != 0) {
if (hdl != NULL) {
switch (why) {
case NAME_ERR_TOOLONG:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "name is too long"));
break;
case NAME_ERR_INVALCHAR:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "invalid character "
"'%c' in pool name"), what);
break;
case NAME_ERR_NOLETTER:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"name must begin with a letter"));
break;
case NAME_ERR_RESERVED:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"name is reserved"));
break;
case NAME_ERR_DISKLIKE:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool name is reserved"));
break;
case NAME_ERR_LEADING_SLASH:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"leading slash in name"));
break;
case NAME_ERR_EMPTY_COMPONENT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"empty component in name"));
break;
case NAME_ERR_TRAILING_SLASH:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"trailing slash in name"));
break;
case NAME_ERR_MULTIPLE_DELIMITERS:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"multiple '@' and/or '#' delimiters in "
"name"));
break;
case NAME_ERR_NO_AT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"permission set is missing '@'"));
break;
default:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"(%d) not defined"), why);
break;
}
}
return (B_FALSE);
}
return (B_TRUE);
}
/*
* Open a handle to the given pool, even if the pool is currently in the FAULTED
* state.
*/
zpool_handle_t *
zpool_open_canfail(libzfs_handle_t *hdl, const char *pool)
{
zpool_handle_t *zhp;
boolean_t missing;
/*
* Make sure the pool name is valid.
*/
if (!zpool_name_valid(hdl, B_TRUE, pool)) {
(void) zfs_error_fmt(hdl, EZFS_INVALIDNAME,
dgettext(TEXT_DOMAIN, "cannot open '%s'"),
pool);
return (NULL);
}
zhp = zfs_alloc(hdl, sizeof (zpool_handle_t));
zhp->zpool_hdl = hdl;
(void) strlcpy(zhp->zpool_name, pool, sizeof (zhp->zpool_name));
if (zpool_refresh_stats(zhp, &missing) != 0) {
zpool_close(zhp);
return (NULL);
}
if (missing) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no such pool"));
(void) zfs_error_fmt(hdl, EZFS_NOENT,
dgettext(TEXT_DOMAIN, "cannot open '%s'"), pool);
zpool_close(zhp);
return (NULL);
}
return (zhp);
}
/*
* Like the above, but silent on error. Used when iterating over pools (because
* the configuration cache may be out of date).
*/
int
zpool_open_silent(libzfs_handle_t *hdl, const char *pool, zpool_handle_t **ret)
{
zpool_handle_t *zhp;
boolean_t missing;
zhp = zfs_alloc(hdl, sizeof (zpool_handle_t));
zhp->zpool_hdl = hdl;
(void) strlcpy(zhp->zpool_name, pool, sizeof (zhp->zpool_name));
if (zpool_refresh_stats(zhp, &missing) != 0) {
zpool_close(zhp);
return (-1);
}
if (missing) {
zpool_close(zhp);
*ret = NULL;
return (0);
}
*ret = zhp;
return (0);
}
/*
* Similar to zpool_open_canfail(), but refuses to open pools in the faulted
* state.
*/
zpool_handle_t *
zpool_open(libzfs_handle_t *hdl, const char *pool)
{
zpool_handle_t *zhp;
if ((zhp = zpool_open_canfail(hdl, pool)) == NULL)
return (NULL);
if (zhp->zpool_state == POOL_STATE_UNAVAIL) {
(void) zfs_error_fmt(hdl, EZFS_POOLUNAVAIL,
dgettext(TEXT_DOMAIN, "cannot open '%s'"), zhp->zpool_name);
zpool_close(zhp);
return (NULL);
}
return (zhp);
}
/*
* Close the handle. Simply frees the memory associated with the handle.
*/
void
zpool_close(zpool_handle_t *zhp)
{
nvlist_free(zhp->zpool_config);
nvlist_free(zhp->zpool_old_config);
nvlist_free(zhp->zpool_props);
free(zhp);
}
/*
* Return the name of the pool.
*/
const char *
zpool_get_name(zpool_handle_t *zhp)
{
return (zhp->zpool_name);
}
/*
* Return the state of the pool (ACTIVE or UNAVAILABLE)
*/
int
zpool_get_state(zpool_handle_t *zhp)
{
return (zhp->zpool_state);
}
/*
* Check if vdev list contains a special vdev
*/
static boolean_t
zpool_has_special_vdev(nvlist_t *nvroot)
{
nvlist_t **child;
uint_t children;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child,
&children) == 0) {
for (uint_t c = 0; c < children; c++) {
const char *bias;
if (nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias) == 0 &&
strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0) {
return (B_TRUE);
}
}
}
return (B_FALSE);
}
/*
* Check if vdev list contains a dRAID vdev
*/
static boolean_t
zpool_has_draid_vdev(nvlist_t *nvroot)
{
nvlist_t **child;
uint_t children;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0) {
for (uint_t c = 0; c < children; c++) {
const char *type;
if (nvlist_lookup_string(child[c],
ZPOOL_CONFIG_TYPE, &type) == 0 &&
strcmp(type, VDEV_TYPE_DRAID) == 0) {
return (B_TRUE);
}
}
}
return (B_FALSE);
}
/*
* Output a dRAID top-level vdev name in to the provided buffer.
*/
static char *
zpool_draid_name(char *name, int len, uint64_t data, uint64_t parity,
uint64_t spares, uint64_t children)
{
snprintf(name, len, "%s%llu:%llud:%lluc:%llus",
VDEV_TYPE_DRAID, (u_longlong_t)parity, (u_longlong_t)data,
(u_longlong_t)children, (u_longlong_t)spares);
return (name);
}
/*
* Return B_TRUE if the provided name is a dRAID spare name.
*/
boolean_t
zpool_is_draid_spare(const char *name)
{
uint64_t spare_id, parity, vdev_id;
if (sscanf(name, VDEV_TYPE_DRAID "%llu-%llu-%llu",
(u_longlong_t *)&parity, (u_longlong_t *)&vdev_id,
(u_longlong_t *)&spare_id) == 3) {
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Create the named pool, using the provided vdev list. It is assumed
* that the consumer has already validated the contents of the nvlist, so we
* don't have to worry about error semantics.
*/
int
zpool_create(libzfs_handle_t *hdl, const char *pool, nvlist_t *nvroot,
nvlist_t *props, nvlist_t *fsprops)
{
zfs_cmd_t zc = {"\0"};
nvlist_t *zc_fsprops = NULL;
nvlist_t *zc_props = NULL;
nvlist_t *hidden_args = NULL;
uint8_t *wkeydata = NULL;
uint_t wkeylen = 0;
char errbuf[ERRBUFLEN];
int ret = -1;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot create '%s'"), pool);
if (!zpool_name_valid(hdl, B_FALSE, pool))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
zcmd_write_conf_nvlist(hdl, &zc, nvroot);
if (props) {
prop_flags_t flags = { .create = B_TRUE, .import = B_FALSE };
if ((zc_props = zpool_valid_proplist(hdl, pool, props,
SPA_VERSION_1, flags, errbuf)) == NULL) {
goto create_failed;
}
}
if (fsprops) {
uint64_t zoned;
const char *zonestr;
zoned = ((nvlist_lookup_string(fsprops,
zfs_prop_to_name(ZFS_PROP_ZONED), &zonestr) == 0) &&
strcmp(zonestr, "on") == 0);
if ((zc_fsprops = zfs_valid_proplist(hdl, ZFS_TYPE_FILESYSTEM,
fsprops, zoned, NULL, NULL, B_TRUE, errbuf)) == NULL) {
goto create_failed;
}
if (nvlist_exists(zc_fsprops,
zfs_prop_to_name(ZFS_PROP_SPECIAL_SMALL_BLOCKS)) &&
!zpool_has_special_vdev(nvroot)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"%s property requires a special vdev"),
zfs_prop_to_name(ZFS_PROP_SPECIAL_SMALL_BLOCKS));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
goto create_failed;
}
if (!zc_props &&
(nvlist_alloc(&zc_props, NV_UNIQUE_NAME, 0) != 0)) {
goto create_failed;
}
if (zfs_crypto_create(hdl, NULL, zc_fsprops, props, B_TRUE,
&wkeydata, &wkeylen) != 0) {
zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf);
goto create_failed;
}
if (nvlist_add_nvlist(zc_props,
ZPOOL_ROOTFS_PROPS, zc_fsprops) != 0) {
goto create_failed;
}
if (wkeydata != NULL) {
if (nvlist_alloc(&hidden_args, NV_UNIQUE_NAME, 0) != 0)
goto create_failed;
if (nvlist_add_uint8_array(hidden_args, "wkeydata",
wkeydata, wkeylen) != 0)
goto create_failed;
if (nvlist_add_nvlist(zc_props, ZPOOL_HIDDEN_ARGS,
hidden_args) != 0)
goto create_failed;
}
}
if (zc_props)
zcmd_write_src_nvlist(hdl, &zc, zc_props);
(void) strlcpy(zc.zc_name, pool, sizeof (zc.zc_name));
if ((ret = zfs_ioctl(hdl, ZFS_IOC_POOL_CREATE, &zc)) != 0) {
zcmd_free_nvlists(&zc);
nvlist_free(zc_props);
nvlist_free(zc_fsprops);
nvlist_free(hidden_args);
if (wkeydata != NULL)
free(wkeydata);
switch (errno) {
case EBUSY:
/*
* This can happen if the user has specified the same
* device multiple times. We can't reliably detect this
* until we try to add it and see we already have a
* label. This can also happen under if the device is
* part of an active md or lvm device.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more vdevs refer to the same device, or "
"one of\nthe devices is part of an active md or "
"lvm device"));
return (zfs_error(hdl, EZFS_BADDEV, errbuf));
case ERANGE:
/*
* This happens if the record size is smaller or larger
* than the allowed size range, or not a power of 2.
*
* NOTE: although zfs_valid_proplist is called earlier,
* this case may have slipped through since the
* pool does not exist yet and it is therefore
* impossible to read properties e.g. max blocksize
* from the pool.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"record size invalid"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
case EOVERFLOW:
/*
* This occurs when one of the devices is below
* SPA_MINDEVSIZE. Unfortunately, we can't detect which
* device was the problem device since there's no
* reliable way to determine device size from userland.
*/
{
char buf[64];
zfs_nicebytes(SPA_MINDEVSIZE, buf,
sizeof (buf));
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is less than the "
"minimum size (%s)"), buf);
}
return (zfs_error(hdl, EZFS_BADDEV, errbuf));
case ENOSPC:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is out of space"));
return (zfs_error(hdl, EZFS_BADDEV, errbuf));
case EINVAL:
if (zpool_has_draid_vdev(nvroot) &&
zfeature_lookup_name("draid", NULL) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dRAID vdevs are unsupported by the "
"kernel"));
return (zfs_error(hdl, EZFS_BADDEV, errbuf));
} else {
return (zpool_standard_error(hdl, errno,
errbuf));
}
default:
return (zpool_standard_error(hdl, errno, errbuf));
}
}
create_failed:
zcmd_free_nvlists(&zc);
nvlist_free(zc_props);
nvlist_free(zc_fsprops);
nvlist_free(hidden_args);
if (wkeydata != NULL)
free(wkeydata);
return (ret);
}
/*
* Destroy the given pool. It is up to the caller to ensure that there are no
* datasets left in the pool.
*/
int
zpool_destroy(zpool_handle_t *zhp, const char *log_str)
{
zfs_cmd_t zc = {"\0"};
zfs_handle_t *zfp = NULL;
libzfs_handle_t *hdl = zhp->zpool_hdl;
char errbuf[ERRBUFLEN];
if (zhp->zpool_state == POOL_STATE_ACTIVE &&
(zfp = zfs_open(hdl, zhp->zpool_name, ZFS_TYPE_FILESYSTEM)) == NULL)
return (-1);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_history = (uint64_t)(uintptr_t)log_str;
if (zfs_ioctl(hdl, ZFS_IOC_POOL_DESTROY, &zc) != 0) {
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot destroy '%s'"), zhp->zpool_name);
if (errno == EROFS) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is read only"));
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
} else {
(void) zpool_standard_error(hdl, errno, errbuf);
}
if (zfp)
zfs_close(zfp);
return (-1);
}
if (zfp) {
remove_mountpoint(zfp);
zfs_close(zfp);
}
return (0);
}
/*
* Create a checkpoint in the given pool.
*/
int
zpool_checkpoint(zpool_handle_t *zhp)
{
libzfs_handle_t *hdl = zhp->zpool_hdl;
char errbuf[ERRBUFLEN];
int error;
error = lzc_pool_checkpoint(zhp->zpool_name);
if (error != 0) {
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot checkpoint '%s'"), zhp->zpool_name);
(void) zpool_standard_error(hdl, error, errbuf);
return (-1);
}
return (0);
}
/*
* Discard the checkpoint from the given pool.
*/
int
zpool_discard_checkpoint(zpool_handle_t *zhp)
{
libzfs_handle_t *hdl = zhp->zpool_hdl;
char errbuf[ERRBUFLEN];
int error;
error = lzc_pool_checkpoint_discard(zhp->zpool_name);
if (error != 0) {
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot discard checkpoint in '%s'"), zhp->zpool_name);
(void) zpool_standard_error(hdl, error, errbuf);
return (-1);
}
return (0);
}
/*
* Add the given vdevs to the pool. The caller must have already performed the
* necessary verification to ensure that the vdev specification is well-formed.
*/
int
zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot)
{
zfs_cmd_t zc = {"\0"};
int ret;
libzfs_handle_t *hdl = zhp->zpool_hdl;
char errbuf[ERRBUFLEN];
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot add to '%s'"), zhp->zpool_name);
if (zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL) <
SPA_VERSION_SPARES &&
nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be "
"upgraded to add hot spares"));
return (zfs_error(hdl, EZFS_BADVERSION, errbuf));
}
if (zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL) <
SPA_VERSION_L2CACHE &&
nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be "
"upgraded to add cache devices"));
return (zfs_error(hdl, EZFS_BADVERSION, errbuf));
}
zcmd_write_conf_nvlist(hdl, &zc, nvroot);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_ADD, &zc) != 0) {
switch (errno) {
case EBUSY:
/*
* This can happen if the user has specified the same
* device multiple times. We can't reliably detect this
* until we try to add it and see we already have a
* label.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more vdevs refer to the same device"));
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case EINVAL:
if (zpool_has_draid_vdev(nvroot) &&
zfeature_lookup_name("draid", NULL) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dRAID vdevs are unsupported by the "
"kernel"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid config; a pool with removing/"
"removed vdevs does not support adding "
"raidz or dRAID vdevs"));
}
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case EOVERFLOW:
/*
* This occurs when one of the devices is below
* SPA_MINDEVSIZE. Unfortunately, we can't detect which
* device was the problem device since there's no
* reliable way to determine device size from userland.
*/
{
char buf[64];
zfs_nicebytes(SPA_MINDEVSIZE, buf,
sizeof (buf));
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"device is less than the minimum "
"size (%s)"), buf);
}
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded to add these vdevs"));
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
default:
(void) zpool_standard_error(hdl, errno, errbuf);
}
ret = -1;
} else {
ret = 0;
}
zcmd_free_nvlists(&zc);
return (ret);
}
/*
* Exports the pool from the system. The caller must ensure that there are no
* mounted datasets in the pool.
*/
static int
zpool_export_common(zpool_handle_t *zhp, boolean_t force, boolean_t hardforce,
const char *log_str)
{
zfs_cmd_t zc = {"\0"};
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_cookie = force;
zc.zc_guid = hardforce;
zc.zc_history = (uint64_t)(uintptr_t)log_str;
if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_EXPORT, &zc) != 0) {
switch (errno) {
case EXDEV:
zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN,
"use '-f' to override the following errors:\n"
"'%s' has an active shared spare which could be"
" used by other pools once '%s' is exported."),
zhp->zpool_name, zhp->zpool_name);
return (zfs_error_fmt(zhp->zpool_hdl, EZFS_ACTIVE_SPARE,
dgettext(TEXT_DOMAIN, "cannot export '%s'"),
zhp->zpool_name));
default:
return (zpool_standard_error_fmt(zhp->zpool_hdl, errno,
dgettext(TEXT_DOMAIN, "cannot export '%s'"),
zhp->zpool_name));
}
}
return (0);
}
int
zpool_export(zpool_handle_t *zhp, boolean_t force, const char *log_str)
{
return (zpool_export_common(zhp, force, B_FALSE, log_str));
}
int
zpool_export_force(zpool_handle_t *zhp, const char *log_str)
{
return (zpool_export_common(zhp, B_TRUE, B_TRUE, log_str));
}
static void
zpool_rewind_exclaim(libzfs_handle_t *hdl, const char *name, boolean_t dryrun,
nvlist_t *config)
{
nvlist_t *nv = NULL;
uint64_t rewindto;
int64_t loss = -1;
struct tm t;
char timestr[128];
if (!hdl->libzfs_printerr || config == NULL)
return;
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nv) != 0 ||
nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_REWIND_INFO, &nv) != 0) {
return;
}
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_TIME, &rewindto) != 0)
return;
(void) nvlist_lookup_int64(nv, ZPOOL_CONFIG_REWIND_TIME, &loss);
if (localtime_r((time_t *)&rewindto, &t) != NULL &&
strftime(timestr, 128, "%c", &t) != 0) {
if (dryrun) {
(void) printf(dgettext(TEXT_DOMAIN,
"Would be able to return %s "
"to its state as of %s.\n"),
name, timestr);
} else {
(void) printf(dgettext(TEXT_DOMAIN,
"Pool %s returned to its state as of %s.\n"),
name, timestr);
}
if (loss > 120) {
(void) printf(dgettext(TEXT_DOMAIN,
"%s approximately %lld "),
dryrun ? "Would discard" : "Discarded",
((longlong_t)loss + 30) / 60);
(void) printf(dgettext(TEXT_DOMAIN,
"minutes of transactions.\n"));
} else if (loss > 0) {
(void) printf(dgettext(TEXT_DOMAIN,
"%s approximately %lld "),
dryrun ? "Would discard" : "Discarded",
(longlong_t)loss);
(void) printf(dgettext(TEXT_DOMAIN,
"seconds of transactions.\n"));
}
}
}
void
zpool_explain_recover(libzfs_handle_t *hdl, const char *name, int reason,
nvlist_t *config)
{
nvlist_t *nv = NULL;
int64_t loss = -1;
uint64_t edata = UINT64_MAX;
uint64_t rewindto;
struct tm t;
char timestr[128];
if (!hdl->libzfs_printerr)
return;
if (reason >= 0)
(void) printf(dgettext(TEXT_DOMAIN, "action: "));
else
(void) printf(dgettext(TEXT_DOMAIN, "\t"));
/* All attempted rewinds failed if ZPOOL_CONFIG_LOAD_TIME missing */
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nv) != 0 ||
nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_REWIND_INFO, &nv) != 0 ||
nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_TIME, &rewindto) != 0)
goto no_info;
(void) nvlist_lookup_int64(nv, ZPOOL_CONFIG_REWIND_TIME, &loss);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_DATA_ERRORS,
&edata);
(void) printf(dgettext(TEXT_DOMAIN,
"Recovery is possible, but will result in some data loss.\n"));
if (localtime_r((time_t *)&rewindto, &t) != NULL &&
strftime(timestr, 128, "%c", &t) != 0) {
(void) printf(dgettext(TEXT_DOMAIN,
"\tReturning the pool to its state as of %s\n"
"\tshould correct the problem. "),
timestr);
} else {
(void) printf(dgettext(TEXT_DOMAIN,
"\tReverting the pool to an earlier state "
"should correct the problem.\n\t"));
}
if (loss > 120) {
(void) printf(dgettext(TEXT_DOMAIN,
"Approximately %lld minutes of data\n"
"\tmust be discarded, irreversibly. "),
((longlong_t)loss + 30) / 60);
} else if (loss > 0) {
(void) printf(dgettext(TEXT_DOMAIN,
"Approximately %lld seconds of data\n"
"\tmust be discarded, irreversibly. "),
(longlong_t)loss);
}
if (edata != 0 && edata != UINT64_MAX) {
if (edata == 1) {
(void) printf(dgettext(TEXT_DOMAIN,
"After rewind, at least\n"
"\tone persistent user-data error will remain. "));
} else {
(void) printf(dgettext(TEXT_DOMAIN,
"After rewind, several\n"
"\tpersistent user-data errors will remain. "));
}
}
(void) printf(dgettext(TEXT_DOMAIN,
"Recovery can be attempted\n\tby executing 'zpool %s -F %s'. "),
reason >= 0 ? "clear" : "import", name);
(void) printf(dgettext(TEXT_DOMAIN,
"A scrub of the pool\n"
"\tis strongly recommended after recovery.\n"));
return;
no_info:
(void) printf(dgettext(TEXT_DOMAIN,
"Destroy and re-create the pool from\n\ta backup source.\n"));
}
/*
* zpool_import() is a contracted interface. Should be kept the same
* if possible.
*
* Applications should use zpool_import_props() to import a pool with
* new properties value to be set.
*/
int
zpool_import(libzfs_handle_t *hdl, nvlist_t *config, const char *newname,
char *altroot)
{
nvlist_t *props = NULL;
int ret;
if (altroot != NULL) {
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0) {
return (zfs_error_fmt(hdl, EZFS_NOMEM,
dgettext(TEXT_DOMAIN, "cannot import '%s'"),
newname));
}
if (nvlist_add_string(props,
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), altroot) != 0 ||
nvlist_add_string(props,
zpool_prop_to_name(ZPOOL_PROP_CACHEFILE), "none") != 0) {
nvlist_free(props);
return (zfs_error_fmt(hdl, EZFS_NOMEM,
dgettext(TEXT_DOMAIN, "cannot import '%s'"),
newname));
}
}
ret = zpool_import_props(hdl, config, newname, props,
ZFS_IMPORT_NORMAL);
nvlist_free(props);
return (ret);
}
static void
print_vdev_tree(libzfs_handle_t *hdl, const char *name, nvlist_t *nv,
int indent)
{
nvlist_t **child;
uint_t c, children;
char *vname;
uint64_t is_log = 0;
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG,
&is_log);
if (name != NULL)
(void) printf("\t%*s%s%s\n", indent, "", name,
is_log ? " [log]" : "");
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
return;
for (c = 0; c < children; c++) {
vname = zpool_vdev_name(hdl, NULL, child[c], VDEV_NAME_TYPE_ID);
print_vdev_tree(hdl, vname, child[c], indent + 2);
free(vname);
}
}
void
zpool_print_unsup_feat(nvlist_t *config)
{
nvlist_t *nvinfo, *unsup_feat;
nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
unsup_feat = fnvlist_lookup_nvlist(nvinfo, ZPOOL_CONFIG_UNSUP_FEAT);
for (nvpair_t *nvp = nvlist_next_nvpair(unsup_feat, NULL);
nvp != NULL; nvp = nvlist_next_nvpair(unsup_feat, nvp)) {
const char *desc = fnvpair_value_string(nvp);
if (strlen(desc) > 0)
(void) printf("\t%s (%s)\n", nvpair_name(nvp), desc);
else
(void) printf("\t%s\n", nvpair_name(nvp));
}
}
/*
* Import the given pool using the known configuration and a list of
* properties to be set. The configuration should have come from
* zpool_find_import(). The 'newname' parameters control whether the pool
* is imported with a different name.
*/
int
zpool_import_props(libzfs_handle_t *hdl, nvlist_t *config, const char *newname,
nvlist_t *props, int flags)
{
zfs_cmd_t zc = {"\0"};
zpool_load_policy_t policy;
nvlist_t *nv = NULL;
nvlist_t *nvinfo = NULL;
nvlist_t *missing = NULL;
const char *thename;
const char *origname;
int ret;
int error = 0;
char errbuf[ERRBUFLEN];
origname = fnvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME);
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot import pool '%s'"), origname);
if (newname != NULL) {
if (!zpool_name_valid(hdl, B_FALSE, newname))
return (zfs_error_fmt(hdl, EZFS_INVALIDNAME,
dgettext(TEXT_DOMAIN, "cannot import '%s'"),
newname));
thename = newname;
} else {
thename = origname;
}
if (props != NULL) {
uint64_t version;
prop_flags_t flags = { .create = B_FALSE, .import = B_TRUE };
version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION);
if ((props = zpool_valid_proplist(hdl, origname,
props, version, flags, errbuf)) == NULL)
return (-1);
zcmd_write_src_nvlist(hdl, &zc, props);
nvlist_free(props);
}
(void) strlcpy(zc.zc_name, thename, sizeof (zc.zc_name));
zc.zc_guid = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID);
zcmd_write_conf_nvlist(hdl, &zc, config);
zcmd_alloc_dst_nvlist(hdl, &zc, zc.zc_nvlist_conf_size * 2);
zc.zc_cookie = flags;
while ((ret = zfs_ioctl(hdl, ZFS_IOC_POOL_IMPORT, &zc)) != 0 &&
errno == ENOMEM)
zcmd_expand_dst_nvlist(hdl, &zc);
if (ret != 0)
error = errno;
(void) zcmd_read_dst_nvlist(hdl, &zc, &nv);
zcmd_free_nvlists(&zc);
zpool_get_load_policy(config, &policy);
if (error) {
char desc[1024];
char aux[256];
/*
* Dry-run failed, but we print out what success
* looks like if we found a best txg
*/
if (policy.zlp_rewind & ZPOOL_TRY_REWIND) {
zpool_rewind_exclaim(hdl, newname ? origname : thename,
B_TRUE, nv);
nvlist_free(nv);
return (-1);
}
if (newname == NULL)
(void) snprintf(desc, sizeof (desc),
dgettext(TEXT_DOMAIN, "cannot import '%s'"),
thename);
else
(void) snprintf(desc, sizeof (desc),
dgettext(TEXT_DOMAIN, "cannot import '%s' as '%s'"),
origname, thename);
switch (error) {
case ENOTSUP:
if (nv != NULL && nvlist_lookup_nvlist(nv,
ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0 &&
nvlist_exists(nvinfo, ZPOOL_CONFIG_UNSUP_FEAT)) {
(void) printf(dgettext(TEXT_DOMAIN, "This "
"pool uses the following feature(s) not "
"supported by this system:\n"));
zpool_print_unsup_feat(nv);
if (nvlist_exists(nvinfo,
ZPOOL_CONFIG_CAN_RDONLY)) {
(void) printf(dgettext(TEXT_DOMAIN,
"All unsupported features are only "
"required for writing to the pool."
"\nThe pool can be imported using "
"'-o readonly=on'.\n"));
}
}
/*
* Unsupported version.
*/
(void) zfs_error(hdl, EZFS_BADVERSION, desc);
break;
case EREMOTEIO:
if (nv != NULL && nvlist_lookup_nvlist(nv,
ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0) {
const char *hostname = "<unknown>";
uint64_t hostid = 0;
mmp_state_t mmp_state;
mmp_state = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_STATE);
if (nvlist_exists(nvinfo,
ZPOOL_CONFIG_MMP_HOSTNAME))
hostname = fnvlist_lookup_string(nvinfo,
ZPOOL_CONFIG_MMP_HOSTNAME);
if (nvlist_exists(nvinfo,
ZPOOL_CONFIG_MMP_HOSTID))
hostid = fnvlist_lookup_uint64(nvinfo,
ZPOOL_CONFIG_MMP_HOSTID);
if (mmp_state == MMP_STATE_ACTIVE) {
(void) snprintf(aux, sizeof (aux),
dgettext(TEXT_DOMAIN, "pool is imp"
"orted on host '%s' (hostid=%lx).\n"
"Export the pool on the other "
"system, then run 'zpool import'."),
hostname, (unsigned long) hostid);
} else if (mmp_state == MMP_STATE_NO_HOSTID) {
(void) snprintf(aux, sizeof (aux),
dgettext(TEXT_DOMAIN, "pool has "
"the multihost property on and "
"the\nsystem's hostid is not set. "
"Set a unique system hostid with "
"the zgenhostid(8) command.\n"));
}
(void) zfs_error_aux(hdl, "%s", aux);
}
(void) zfs_error(hdl, EZFS_ACTIVE_POOL, desc);
break;
case EINVAL:
(void) zfs_error(hdl, EZFS_INVALCONFIG, desc);
break;
case EROFS:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is read only"));
(void) zfs_error(hdl, EZFS_BADDEV, desc);
break;
case ENXIO:
if (nv && nvlist_lookup_nvlist(nv,
ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0 &&
nvlist_lookup_nvlist(nvinfo,
ZPOOL_CONFIG_MISSING_DEVICES, &missing) == 0) {
(void) printf(dgettext(TEXT_DOMAIN,
"The devices below are missing or "
"corrupted, use '-m' to import the pool "
"anyway:\n"));
print_vdev_tree(hdl, NULL, missing, 2);
(void) printf("\n");
}
(void) zpool_standard_error(hdl, error, desc);
break;
case EEXIST:
(void) zpool_standard_error(hdl, error, desc);
break;
case EBUSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices are already in use\n"));
(void) zfs_error(hdl, EZFS_BADDEV, desc);
break;
case ENAMETOOLONG:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"new name of at least one dataset is longer than "
"the maximum allowable length"));
(void) zfs_error(hdl, EZFS_NAMETOOLONG, desc);
break;
default:
(void) zpool_standard_error(hdl, error, desc);
zpool_explain_recover(hdl,
newname ? origname : thename, -error, nv);
break;
}
nvlist_free(nv);
ret = -1;
} else {
zpool_handle_t *zhp;
/*
* This should never fail, but play it safe anyway.
*/
if (zpool_open_silent(hdl, thename, &zhp) != 0)
ret = -1;
else if (zhp != NULL)
zpool_close(zhp);
if (policy.zlp_rewind &
(ZPOOL_DO_REWIND | ZPOOL_TRY_REWIND)) {
zpool_rewind_exclaim(hdl, newname ? origname : thename,
((policy.zlp_rewind & ZPOOL_TRY_REWIND) != 0), nv);
}
nvlist_free(nv);
}
return (ret);
}
/*
* Translate vdev names to guids. If a vdev_path is determined to be
* unsuitable then a vd_errlist is allocated and the vdev path and errno
* are added to it.
*/
static int
zpool_translate_vdev_guids(zpool_handle_t *zhp, nvlist_t *vds,
nvlist_t *vdev_guids, nvlist_t *guids_to_paths, nvlist_t **vd_errlist)
{
nvlist_t *errlist = NULL;
int error = 0;
for (nvpair_t *elem = nvlist_next_nvpair(vds, NULL); elem != NULL;
elem = nvlist_next_nvpair(vds, elem)) {
boolean_t spare, cache;
const char *vd_path = nvpair_name(elem);
nvlist_t *tgt = zpool_find_vdev(zhp, vd_path, &spare, &cache,
NULL);
if ((tgt == NULL) || cache || spare) {
if (errlist == NULL) {
errlist = fnvlist_alloc();
error = EINVAL;
}
uint64_t err = (tgt == NULL) ? EZFS_NODEVICE :
(spare ? EZFS_ISSPARE : EZFS_ISL2CACHE);
fnvlist_add_int64(errlist, vd_path, err);
continue;
}
uint64_t guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
fnvlist_add_uint64(vdev_guids, vd_path, guid);
char msg[MAXNAMELEN];
(void) snprintf(msg, sizeof (msg), "%llu", (u_longlong_t)guid);
fnvlist_add_string(guids_to_paths, msg, vd_path);
}
if (error != 0) {
verify(errlist != NULL);
if (vd_errlist != NULL)
*vd_errlist = errlist;
else
fnvlist_free(errlist);
}
return (error);
}
static int
xlate_init_err(int err)
{
switch (err) {
case ENODEV:
return (EZFS_NODEVICE);
case EINVAL:
case EROFS:
return (EZFS_BADDEV);
case EBUSY:
return (EZFS_INITIALIZING);
case ESRCH:
return (EZFS_NO_INITIALIZE);
}
return (err);
}
/*
* Begin, suspend, or cancel the initialization (initializing of all free
* blocks) for the given vdevs in the given pool.
*/
static int
zpool_initialize_impl(zpool_handle_t *zhp, pool_initialize_func_t cmd_type,
nvlist_t *vds, boolean_t wait)
{
int err;
nvlist_t *vdev_guids = fnvlist_alloc();
nvlist_t *guids_to_paths = fnvlist_alloc();
nvlist_t *vd_errlist = NULL;
nvlist_t *errlist;
nvpair_t *elem;
err = zpool_translate_vdev_guids(zhp, vds, vdev_guids,
guids_to_paths, &vd_errlist);
if (err != 0) {
verify(vd_errlist != NULL);
goto list_errors;
}
err = lzc_initialize(zhp->zpool_name, cmd_type,
vdev_guids, &errlist);
if (err != 0) {
if (errlist != NULL) {
vd_errlist = fnvlist_lookup_nvlist(errlist,
ZPOOL_INITIALIZE_VDEVS);
goto list_errors;
}
(void) zpool_standard_error(zhp->zpool_hdl, err,
dgettext(TEXT_DOMAIN, "operation failed"));
goto out;
}
if (wait) {
for (elem = nvlist_next_nvpair(vdev_guids, NULL); elem != NULL;
elem = nvlist_next_nvpair(vdev_guids, elem)) {
uint64_t guid = fnvpair_value_uint64(elem);
err = lzc_wait_tag(zhp->zpool_name,
ZPOOL_WAIT_INITIALIZE, guid, NULL);
if (err != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl,
err, dgettext(TEXT_DOMAIN, "error "
"waiting for '%s' to initialize"),
nvpair_name(elem));
goto out;
}
}
}
goto out;
list_errors:
for (elem = nvlist_next_nvpair(vd_errlist, NULL); elem != NULL;
elem = nvlist_next_nvpair(vd_errlist, elem)) {
int64_t vd_error = xlate_init_err(fnvpair_value_int64(elem));
const char *path;
if (nvlist_lookup_string(guids_to_paths, nvpair_name(elem),
&path) != 0)
path = nvpair_name(elem);
(void) zfs_error_fmt(zhp->zpool_hdl, vd_error,
"cannot initialize '%s'", path);
}
out:
fnvlist_free(vdev_guids);
fnvlist_free(guids_to_paths);
if (vd_errlist != NULL)
fnvlist_free(vd_errlist);
return (err == 0 ? 0 : -1);
}
int
zpool_initialize(zpool_handle_t *zhp, pool_initialize_func_t cmd_type,
nvlist_t *vds)
{
return (zpool_initialize_impl(zhp, cmd_type, vds, B_FALSE));
}
int
zpool_initialize_wait(zpool_handle_t *zhp, pool_initialize_func_t cmd_type,
nvlist_t *vds)
{
return (zpool_initialize_impl(zhp, cmd_type, vds, B_TRUE));
}
static int
xlate_trim_err(int err)
{
switch (err) {
case ENODEV:
return (EZFS_NODEVICE);
case EINVAL:
case EROFS:
return (EZFS_BADDEV);
case EBUSY:
return (EZFS_TRIMMING);
case ESRCH:
return (EZFS_NO_TRIM);
case EOPNOTSUPP:
return (EZFS_TRIM_NOTSUP);
}
return (err);
}
static int
zpool_trim_wait(zpool_handle_t *zhp, nvlist_t *vdev_guids)
{
int err;
nvpair_t *elem;
for (elem = nvlist_next_nvpair(vdev_guids, NULL); elem != NULL;
elem = nvlist_next_nvpair(vdev_guids, elem)) {
uint64_t guid = fnvpair_value_uint64(elem);
err = lzc_wait_tag(zhp->zpool_name,
ZPOOL_WAIT_TRIM, guid, NULL);
if (err != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl,
err, dgettext(TEXT_DOMAIN, "error "
"waiting to trim '%s'"), nvpair_name(elem));
return (err);
}
}
return (0);
}
/*
* Check errlist and report any errors, omitting ones which should be
* suppressed. Returns B_TRUE if any errors were reported.
*/
static boolean_t
check_trim_errs(zpool_handle_t *zhp, trimflags_t *trim_flags,
nvlist_t *guids_to_paths, nvlist_t *vds, nvlist_t *errlist)
{
nvpair_t *elem;
boolean_t reported_errs = B_FALSE;
int num_vds = 0;
int num_suppressed_errs = 0;
for (elem = nvlist_next_nvpair(vds, NULL);
elem != NULL; elem = nvlist_next_nvpair(vds, elem)) {
num_vds++;
}
for (elem = nvlist_next_nvpair(errlist, NULL);
elem != NULL; elem = nvlist_next_nvpair(errlist, elem)) {
int64_t vd_error = xlate_trim_err(fnvpair_value_int64(elem));
const char *path;
/*
* If only the pool was specified, and it was not a secure
* trim then suppress warnings for individual vdevs which
* do not support trimming.
*/
if (vd_error == EZFS_TRIM_NOTSUP &&
trim_flags->fullpool &&
!trim_flags->secure) {
num_suppressed_errs++;
continue;
}
reported_errs = B_TRUE;
if (nvlist_lookup_string(guids_to_paths, nvpair_name(elem),
&path) != 0)
path = nvpair_name(elem);
(void) zfs_error_fmt(zhp->zpool_hdl, vd_error,
"cannot trim '%s'", path);
}
if (num_suppressed_errs == num_vds) {
(void) zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN,
"no devices in pool support trim operations"));
(void) (zfs_error(zhp->zpool_hdl, EZFS_TRIM_NOTSUP,
dgettext(TEXT_DOMAIN, "cannot trim")));
reported_errs = B_TRUE;
}
return (reported_errs);
}
/*
* Begin, suspend, or cancel the TRIM (discarding of all free blocks) for
* the given vdevs in the given pool.
*/
int
zpool_trim(zpool_handle_t *zhp, pool_trim_func_t cmd_type, nvlist_t *vds,
trimflags_t *trim_flags)
{
int err;
int retval = 0;
nvlist_t *vdev_guids = fnvlist_alloc();
nvlist_t *guids_to_paths = fnvlist_alloc();
nvlist_t *errlist = NULL;
err = zpool_translate_vdev_guids(zhp, vds, vdev_guids,
guids_to_paths, &errlist);
if (err != 0) {
check_trim_errs(zhp, trim_flags, guids_to_paths, vds, errlist);
retval = -1;
goto out;
}
err = lzc_trim(zhp->zpool_name, cmd_type, trim_flags->rate,
trim_flags->secure, vdev_guids, &errlist);
if (err != 0) {
nvlist_t *vd_errlist;
if (errlist != NULL && nvlist_lookup_nvlist(errlist,
ZPOOL_TRIM_VDEVS, &vd_errlist) == 0) {
if (check_trim_errs(zhp, trim_flags, guids_to_paths,
vds, vd_errlist)) {
retval = -1;
goto out;
}
} else {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "operation failed"));
zpool_standard_error(zhp->zpool_hdl, err, errbuf);
retval = -1;
goto out;
}
}
if (trim_flags->wait)
retval = zpool_trim_wait(zhp, vdev_guids);
out:
if (errlist != NULL)
fnvlist_free(errlist);
fnvlist_free(vdev_guids);
fnvlist_free(guids_to_paths);
return (retval);
}
/*
* Scan the pool.
*/
int
zpool_scan(zpool_handle_t *zhp, pool_scan_func_t func, pool_scrub_cmd_t cmd)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
int err;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_cookie = func;
zc.zc_flags = cmd;
if (zfs_ioctl(hdl, ZFS_IOC_POOL_SCAN, &zc) == 0)
return (0);
err = errno;
/* ECANCELED on a scrub means we resumed a paused scrub */
if (err == ECANCELED && func == POOL_SCAN_SCRUB &&
cmd == POOL_SCRUB_NORMAL)
return (0);
if (err == ENOENT && func != POOL_SCAN_NONE && cmd == POOL_SCRUB_NORMAL)
return (0);
if (func == POOL_SCAN_SCRUB) {
if (cmd == POOL_SCRUB_PAUSE) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot pause scrubbing %s"),
zc.zc_name);
} else {
assert(cmd == POOL_SCRUB_NORMAL);
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot scrub %s"),
zc.zc_name);
}
} else if (func == POOL_SCAN_RESILVER) {
assert(cmd == POOL_SCRUB_NORMAL);
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot restart resilver on %s"), zc.zc_name);
} else if (func == POOL_SCAN_NONE) {
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot cancel scrubbing %s"), zc.zc_name);
} else {
assert(!"unexpected result");
}
if (err == EBUSY) {
nvlist_t *nvroot;
pool_scan_stat_t *ps = NULL;
uint_t psc;
nvroot = fnvlist_lookup_nvlist(zhp->zpool_config,
ZPOOL_CONFIG_VDEV_TREE);
(void) nvlist_lookup_uint64_array(nvroot,
ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &psc);
if (ps && ps->pss_func == POOL_SCAN_SCRUB &&
ps->pss_state == DSS_SCANNING) {
if (cmd == POOL_SCRUB_PAUSE)
return (zfs_error(hdl, EZFS_SCRUB_PAUSED,
errbuf));
else
return (zfs_error(hdl, EZFS_SCRUBBING, errbuf));
} else {
return (zfs_error(hdl, EZFS_RESILVERING, errbuf));
}
} else if (err == ENOENT) {
return (zfs_error(hdl, EZFS_NO_SCRUB, errbuf));
} else if (err == ENOTSUP && func == POOL_SCAN_RESILVER) {
return (zfs_error(hdl, EZFS_NO_RESILVER_DEFER, errbuf));
} else {
return (zpool_standard_error(hdl, err, errbuf));
}
}
/*
* Find a vdev that matches the search criteria specified. We use the
* the nvpair name to determine how we should look for the device.
* 'avail_spare' is set to TRUE if the provided guid refers to an AVAIL
* spare; but FALSE if its an INUSE spare.
*/
static nvlist_t *
vdev_to_nvlist_iter(nvlist_t *nv, nvlist_t *search, boolean_t *avail_spare,
boolean_t *l2cache, boolean_t *log)
{
uint_t c, children;
nvlist_t **child;
nvlist_t *ret;
uint64_t is_log;
const char *srchkey;
nvpair_t *pair = nvlist_next_nvpair(search, NULL);
/* Nothing to look for */
if (search == NULL || pair == NULL)
return (NULL);
/* Obtain the key we will use to search */
srchkey = nvpair_name(pair);
switch (nvpair_type(pair)) {
case DATA_TYPE_UINT64:
if (strcmp(srchkey, ZPOOL_CONFIG_GUID) == 0) {
uint64_t srchval = fnvpair_value_uint64(pair);
uint64_t theguid = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_GUID);
if (theguid == srchval)
return (nv);
}
break;
case DATA_TYPE_STRING: {
const char *srchval, *val;
srchval = fnvpair_value_string(pair);
if (nvlist_lookup_string(nv, srchkey, &val) != 0)
break;
/*
* Search for the requested value. Special cases:
*
* - ZPOOL_CONFIG_PATH for whole disk entries. These end in
* "-part1", or "p1". The suffix is hidden from the user,
* but included in the string, so this matches around it.
* - ZPOOL_CONFIG_PATH for short names zfs_strcmp_shortname()
* is used to check all possible expanded paths.
* - looking for a top-level vdev name (i.e. ZPOOL_CONFIG_TYPE).
*
* Otherwise, all other searches are simple string compares.
*/
if (strcmp(srchkey, ZPOOL_CONFIG_PATH) == 0) {
uint64_t wholedisk = 0;
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
&wholedisk);
if (zfs_strcmp_pathname(srchval, val, wholedisk) == 0)
return (nv);
} else if (strcmp(srchkey, ZPOOL_CONFIG_TYPE) == 0) {
char *type, *idx, *end, *p;
uint64_t id, vdev_id;
/*
* Determine our vdev type, keeping in mind
* that the srchval is composed of a type and
* vdev id pair (i.e. mirror-4).
*/
if ((type = strdup(srchval)) == NULL)
return (NULL);
if ((p = strrchr(type, '-')) == NULL) {
free(type);
break;
}
idx = p + 1;
*p = '\0';
/*
* If the types don't match then keep looking.
*/
if (strncmp(val, type, strlen(val)) != 0) {
free(type);
break;
}
verify(zpool_vdev_is_interior(type));
id = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID);
errno = 0;
vdev_id = strtoull(idx, &end, 10);
/*
* If we are looking for a raidz and a parity is
* specified, make sure it matches.
*/
int rzlen = strlen(VDEV_TYPE_RAIDZ);
assert(rzlen == strlen(VDEV_TYPE_DRAID));
int typlen = strlen(type);
if ((strncmp(type, VDEV_TYPE_RAIDZ, rzlen) == 0 ||
strncmp(type, VDEV_TYPE_DRAID, rzlen) == 0) &&
typlen != rzlen) {
uint64_t vdev_parity;
int parity = *(type + rzlen) - '0';
if (parity <= 0 || parity > 3 ||
(typlen - rzlen) != 1) {
/*
* Nonsense parity specified, can
* never match
*/
free(type);
return (NULL);
}
vdev_parity = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_NPARITY);
if ((int)vdev_parity != parity) {
free(type);
break;
}
}
free(type);
if (errno != 0)
return (NULL);
/*
* Now verify that we have the correct vdev id.
*/
if (vdev_id == id)
return (nv);
}
/*
* Common case
*/
if (strcmp(srchval, val) == 0)
return (nv);
break;
}
default:
break;
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0)
return (NULL);
for (c = 0; c < children; c++) {
if ((ret = vdev_to_nvlist_iter(child[c], search,
avail_spare, l2cache, NULL)) != NULL) {
/*
* The 'is_log' value is only set for the toplevel
* vdev, not the leaf vdevs. So we always lookup the
* log device from the root of the vdev tree (where
* 'log' is non-NULL).
*/
if (log != NULL &&
nvlist_lookup_uint64(child[c],
ZPOOL_CONFIG_IS_LOG, &is_log) == 0 &&
is_log) {
*log = B_TRUE;
}
return (ret);
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES,
&child, &children) == 0) {
for (c = 0; c < children; c++) {
if ((ret = vdev_to_nvlist_iter(child[c], search,
avail_spare, l2cache, NULL)) != NULL) {
*avail_spare = B_TRUE;
return (ret);
}
}
}
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE,
&child, &children) == 0) {
for (c = 0; c < children; c++) {
if ((ret = vdev_to_nvlist_iter(child[c], search,
avail_spare, l2cache, NULL)) != NULL) {
*l2cache = B_TRUE;
return (ret);
}
}
}
return (NULL);
}
/*
* Given a physical path or guid, find the associated vdev.
*/
nvlist_t *
zpool_find_vdev_by_physpath(zpool_handle_t *zhp, const char *ppath,
boolean_t *avail_spare, boolean_t *l2cache, boolean_t *log)
{
nvlist_t *search, *nvroot, *ret;
uint64_t guid;
char *end;
search = fnvlist_alloc();
guid = strtoull(ppath, &end, 0);
if (guid != 0 && *end == '\0') {
fnvlist_add_uint64(search, ZPOOL_CONFIG_GUID, guid);
} else {
fnvlist_add_string(search, ZPOOL_CONFIG_PHYS_PATH, ppath);
}
nvroot = fnvlist_lookup_nvlist(zhp->zpool_config,
ZPOOL_CONFIG_VDEV_TREE);
*avail_spare = B_FALSE;
*l2cache = B_FALSE;
if (log != NULL)
*log = B_FALSE;
ret = vdev_to_nvlist_iter(nvroot, search, avail_spare, l2cache, log);
fnvlist_free(search);
return (ret);
}
/*
* Determine if we have an "interior" top-level vdev (i.e mirror/raidz).
*/
static boolean_t
zpool_vdev_is_interior(const char *name)
{
if (strncmp(name, VDEV_TYPE_RAIDZ, strlen(VDEV_TYPE_RAIDZ)) == 0 ||
strncmp(name, VDEV_TYPE_SPARE, strlen(VDEV_TYPE_SPARE)) == 0 ||
strncmp(name,
VDEV_TYPE_REPLACING, strlen(VDEV_TYPE_REPLACING)) == 0 ||
+ strncmp(name, VDEV_TYPE_ROOT, strlen(VDEV_TYPE_ROOT)) == 0 ||
strncmp(name, VDEV_TYPE_MIRROR, strlen(VDEV_TYPE_MIRROR)) == 0)
return (B_TRUE);
if (strncmp(name, VDEV_TYPE_DRAID, strlen(VDEV_TYPE_DRAID)) == 0 &&
!zpool_is_draid_spare(name))
return (B_TRUE);
return (B_FALSE);
}
nvlist_t *
zpool_find_vdev(zpool_handle_t *zhp, const char *path, boolean_t *avail_spare,
boolean_t *l2cache, boolean_t *log)
{
char *end;
nvlist_t *nvroot, *search, *ret;
uint64_t guid;
search = fnvlist_alloc();
guid = strtoull(path, &end, 0);
if (guid != 0 && *end == '\0') {
fnvlist_add_uint64(search, ZPOOL_CONFIG_GUID, guid);
} else if (zpool_vdev_is_interior(path)) {
fnvlist_add_string(search, ZPOOL_CONFIG_TYPE, path);
} else {
fnvlist_add_string(search, ZPOOL_CONFIG_PATH, path);
}
nvroot = fnvlist_lookup_nvlist(zhp->zpool_config,
ZPOOL_CONFIG_VDEV_TREE);
*avail_spare = B_FALSE;
*l2cache = B_FALSE;
if (log != NULL)
*log = B_FALSE;
ret = vdev_to_nvlist_iter(nvroot, search, avail_spare, l2cache, log);
fnvlist_free(search);
return (ret);
}
/*
* Convert a vdev path to a GUID. Returns GUID or 0 on error.
*
* If is_spare, is_l2cache, or is_log is non-NULL, then store within it
* if the VDEV is a spare, l2cache, or log device. If they're NULL then
* ignore them.
*/
static uint64_t
zpool_vdev_path_to_guid_impl(zpool_handle_t *zhp, const char *path,
boolean_t *is_spare, boolean_t *is_l2cache, boolean_t *is_log)
{
boolean_t spare = B_FALSE, l2cache = B_FALSE, log = B_FALSE;
nvlist_t *tgt;
if ((tgt = zpool_find_vdev(zhp, path, &spare, &l2cache,
&log)) == NULL)
return (0);
if (is_spare != NULL)
*is_spare = spare;
if (is_l2cache != NULL)
*is_l2cache = l2cache;
if (is_log != NULL)
*is_log = log;
return (fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID));
}
/* Convert a vdev path to a GUID. Returns GUID or 0 on error. */
uint64_t
zpool_vdev_path_to_guid(zpool_handle_t *zhp, const char *path)
{
return (zpool_vdev_path_to_guid_impl(zhp, path, NULL, NULL, NULL));
}
/*
* Bring the specified vdev online. The 'flags' parameter is a set of the
* ZFS_ONLINE_* flags.
*/
int
zpool_vdev_online(zpool_handle_t *zhp, const char *path, int flags,
vdev_state_t *newstate)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache, islog;
libzfs_handle_t *hdl = zhp->zpool_hdl;
if (flags & ZFS_ONLINE_EXPAND) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot expand %s"), path);
} else {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot online %s"), path);
}
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
&islog)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
if (!(flags & ZFS_ONLINE_SPARE) && avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
#ifndef __FreeBSD__
const char *pathname;
if ((flags & ZFS_ONLINE_EXPAND ||
zpool_get_prop_int(zhp, ZPOOL_PROP_AUTOEXPAND, NULL)) &&
nvlist_lookup_string(tgt, ZPOOL_CONFIG_PATH, &pathname) == 0) {
uint64_t wholedisk = 0;
(void) nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_WHOLE_DISK,
&wholedisk);
/*
* XXX - L2ARC 1.0 devices can't support expansion.
*/
if (l2cache) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot expand cache devices"));
return (zfs_error(hdl, EZFS_VDEVNOTSUP, errbuf));
}
if (wholedisk) {
const char *fullpath = path;
char buf[MAXPATHLEN];
int error;
if (path[0] != '/') {
error = zfs_resolve_shortname(path, buf,
sizeof (buf));
if (error != 0)
return (zfs_error(hdl, EZFS_NODEVICE,
errbuf));
fullpath = buf;
}
error = zpool_relabel_disk(hdl, fullpath, errbuf);
if (error != 0)
return (error);
}
}
#endif
zc.zc_cookie = VDEV_STATE_ONLINE;
zc.zc_obj = flags;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) != 0) {
if (errno == EINVAL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "was split "
"from this pool into a new one. Use '%s' "
"instead"), "zpool detach");
return (zfs_error(hdl, EZFS_POSTSPLIT_ONLINE, errbuf));
}
return (zpool_standard_error(hdl, errno, errbuf));
}
*newstate = zc.zc_cookie;
return (0);
}
/*
* Take the specified vdev offline
*/
int
zpool_vdev_offline(zpool_handle_t *zhp, const char *path, boolean_t istmp)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot offline %s"), path);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
NULL)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
if (avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
zc.zc_cookie = VDEV_STATE_OFFLINE;
zc.zc_obj = istmp ? ZFS_OFFLINE_TEMPORARY : 0;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
return (0);
switch (errno) {
case EBUSY:
/*
* There are no other replicas of this device.
*/
return (zfs_error(hdl, EZFS_NOREPLICAS, errbuf));
case EEXIST:
/*
* The log device has unplayed logs
*/
return (zfs_error(hdl, EZFS_UNPLAYED_LOGS, errbuf));
default:
return (zpool_standard_error(hdl, errno, errbuf));
}
}
/*
* Remove the specified vdev asynchronously from the configuration, so
* that it may come ONLINE if reinserted. This is called from zed on
* Udev remove event.
* Note: We also have a similar function zpool_vdev_remove() that
* removes the vdev from the pool.
*/
int
zpool_vdev_remove_wanted(zpool_handle_t *zhp, const char *path)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot remove %s"), path);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
NULL)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
zc.zc_cookie = VDEV_STATE_REMOVED;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Mark the given vdev faulted.
*/
int
zpool_vdev_fault(zpool_handle_t *zhp, uint64_t guid, vdev_aux_t aux)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot fault %llu"), (u_longlong_t)guid);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_guid = guid;
zc.zc_cookie = VDEV_STATE_FAULTED;
zc.zc_obj = aux;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
return (0);
switch (errno) {
case EBUSY:
/*
* There are no other replicas of this device.
*/
return (zfs_error(hdl, EZFS_NOREPLICAS, errbuf));
default:
return (zpool_standard_error(hdl, errno, errbuf));
}
}
/*
* Mark the given vdev degraded.
*/
int
zpool_vdev_degrade(zpool_handle_t *zhp, uint64_t guid, vdev_aux_t aux)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot degrade %llu"), (u_longlong_t)guid);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_guid = guid;
zc.zc_cookie = VDEV_STATE_DEGRADED;
zc.zc_obj = aux;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Returns TRUE if the given nvlist is a vdev that was originally swapped in as
* a hot spare.
*/
static boolean_t
is_replacing_spare(nvlist_t *search, nvlist_t *tgt, int which)
{
nvlist_t **child;
uint_t c, children;
if (nvlist_lookup_nvlist_array(search, ZPOOL_CONFIG_CHILDREN, &child,
&children) == 0) {
const char *type = fnvlist_lookup_string(search,
ZPOOL_CONFIG_TYPE);
if ((strcmp(type, VDEV_TYPE_SPARE) == 0 ||
strcmp(type, VDEV_TYPE_DRAID_SPARE) == 0) &&
children == 2 && child[which] == tgt)
return (B_TRUE);
for (c = 0; c < children; c++)
if (is_replacing_spare(child[c], tgt, which))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Attach new_disk (fully described by nvroot) to old_disk.
* If 'replacing' is specified, the new disk will replace the old one.
*/
int
zpool_vdev_attach(zpool_handle_t *zhp, const char *old_disk,
const char *new_disk, nvlist_t *nvroot, int replacing, boolean_t rebuild)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
int ret;
nvlist_t *tgt;
boolean_t avail_spare, l2cache, islog;
uint64_t val;
char *newname;
nvlist_t **child;
uint_t children;
nvlist_t *config_root;
libzfs_handle_t *hdl = zhp->zpool_hdl;
if (replacing)
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot replace %s with %s"), old_disk, new_disk);
else
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot attach %s to %s"), new_disk, old_disk);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, old_disk, &avail_spare, &l2cache,
&islog)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
if (avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
if (l2cache)
return (zfs_error(hdl, EZFS_ISL2CACHE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
zc.zc_cookie = replacing;
zc.zc_simple = rebuild;
if (rebuild &&
zfeature_lookup_guid("org.openzfs:device_rebuild", NULL) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"the loaded zfs module doesn't support device rebuilds"));
return (zfs_error(hdl, EZFS_POOL_NOTSUP, errbuf));
}
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) != 0 || children != 1) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"new device must be a single disk"));
return (zfs_error(hdl, EZFS_INVALCONFIG, errbuf));
}
config_root = fnvlist_lookup_nvlist(zpool_get_config(zhp, NULL),
ZPOOL_CONFIG_VDEV_TREE);
if ((newname = zpool_vdev_name(NULL, NULL, child[0], 0)) == NULL)
return (-1);
/*
* If the target is a hot spare that has been swapped in, we can only
* replace it with another hot spare.
*/
if (replacing &&
nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_IS_SPARE, &val) == 0 &&
(zpool_find_vdev(zhp, newname, &avail_spare, &l2cache,
NULL) == NULL || !avail_spare) &&
is_replacing_spare(config_root, tgt, 1)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"can only be replaced by another hot spare"));
free(newname);
return (zfs_error(hdl, EZFS_BADTARGET, errbuf));
}
free(newname);
zcmd_write_conf_nvlist(hdl, &zc, nvroot);
ret = zfs_ioctl(hdl, ZFS_IOC_VDEV_ATTACH, &zc);
zcmd_free_nvlists(&zc);
if (ret == 0)
return (0);
switch (errno) {
case ENOTSUP:
/*
* Can't attach to or replace this type of vdev.
*/
if (replacing) {
uint64_t version = zpool_get_prop_int(zhp,
ZPOOL_PROP_VERSION, NULL);
if (islog) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot replace a log with a spare"));
} else if (rebuild) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"only mirror and dRAID vdevs support "
"sequential reconstruction"));
} else if (zpool_is_draid_spare(new_disk)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dRAID spares can only replace child "
"devices in their parent's dRAID vdev"));
} else if (version >= SPA_VERSION_MULTI_REPLACE) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"already in replacing/spare config; wait "
"for completion or use 'zpool detach'"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot replace a replacing device"));
}
} else {
char status[64] = {0};
zpool_prop_get_feature(zhp,
"feature@device_rebuild", status, 63);
if (rebuild &&
strncmp(status, ZFS_FEATURE_DISABLED, 64) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"device_rebuild feature must be enabled "
"in order to use sequential "
"reconstruction"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"can only attach to mirrors and top-level "
"disks"));
}
}
(void) zfs_error(hdl, EZFS_BADTARGET, errbuf);
break;
case EINVAL:
/*
* The new device must be a single disk.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"new device must be a single disk"));
(void) zfs_error(hdl, EZFS_INVALCONFIG, errbuf);
break;
case EBUSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "%s is busy, "
"or device removal is in progress"),
new_disk);
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case EOVERFLOW:
/*
* The new device is too small.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"device is too small"));
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case EDOM:
/*
* The new device has a different optimal sector size.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"new device has a different optimal sector size; use the "
"option '-o ashift=N' to override the optimal size"));
(void) zfs_error(hdl, EZFS_BADDEV, errbuf);
break;
case ENAMETOOLONG:
/*
* The resulting top-level vdev spec won't fit in the label.
*/
(void) zfs_error(hdl, EZFS_DEVOVERFLOW, errbuf);
break;
default:
(void) zpool_standard_error(hdl, errno, errbuf);
}
return (-1);
}
/*
* Detach the specified device.
*/
int
zpool_vdev_detach(zpool_handle_t *zhp, const char *path)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot detach %s"), path);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
NULL)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
if (avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
if (l2cache)
return (zfs_error(hdl, EZFS_ISL2CACHE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_DETACH, &zc) == 0)
return (0);
switch (errno) {
case ENOTSUP:
/*
* Can't detach from this type of vdev.
*/
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "only "
"applicable to mirror and replacing vdevs"));
(void) zfs_error(hdl, EZFS_BADTARGET, errbuf);
break;
case EBUSY:
/*
* There are no other replicas of this device.
*/
(void) zfs_error(hdl, EZFS_NOREPLICAS, errbuf);
break;
default:
(void) zpool_standard_error(hdl, errno, errbuf);
}
return (-1);
}
/*
* Find a mirror vdev in the source nvlist.
*
* The mchild array contains a list of disks in one of the top-level mirrors
* of the source pool. The schild array contains a list of disks that the
* user specified on the command line. We loop over the mchild array to
* see if any entry in the schild array matches.
*
* If a disk in the mchild array is found in the schild array, we return
* the index of that entry. Otherwise we return -1.
*/
static int
find_vdev_entry(zpool_handle_t *zhp, nvlist_t **mchild, uint_t mchildren,
nvlist_t **schild, uint_t schildren)
{
uint_t mc;
for (mc = 0; mc < mchildren; mc++) {
uint_t sc;
char *mpath = zpool_vdev_name(zhp->zpool_hdl, zhp,
mchild[mc], 0);
for (sc = 0; sc < schildren; sc++) {
char *spath = zpool_vdev_name(zhp->zpool_hdl, zhp,
schild[sc], 0);
boolean_t result = (strcmp(mpath, spath) == 0);
free(spath);
if (result) {
free(mpath);
return (mc);
}
}
free(mpath);
}
return (-1);
}
/*
* Split a mirror pool. If newroot points to null, then a new nvlist
* is generated and it is the responsibility of the caller to free it.
*/
int
zpool_vdev_split(zpool_handle_t *zhp, char *newname, nvlist_t **newroot,
nvlist_t *props, splitflags_t flags)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
const char *bias;
nvlist_t *tree, *config, **child, **newchild, *newconfig = NULL;
nvlist_t **varray = NULL, *zc_props = NULL;
uint_t c, children, newchildren, lastlog = 0, vcount, found = 0;
libzfs_handle_t *hdl = zhp->zpool_hdl;
uint64_t vers, readonly = B_FALSE;
boolean_t freelist = B_FALSE, memory_err = B_TRUE;
int retval = 0;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "Unable to split %s"), zhp->zpool_name);
if (!zpool_name_valid(hdl, B_FALSE, newname))
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
if ((config = zpool_get_config(zhp, NULL)) == NULL) {
(void) fprintf(stderr, gettext("Internal error: unable to "
"retrieve pool configuration\n"));
return (-1);
}
tree = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
vers = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION);
if (props) {
prop_flags_t flags = { .create = B_FALSE, .import = B_TRUE };
if ((zc_props = zpool_valid_proplist(hdl, zhp->zpool_name,
props, vers, flags, errbuf)) == NULL)
return (-1);
(void) nvlist_lookup_uint64(zc_props,
zpool_prop_to_name(ZPOOL_PROP_READONLY), &readonly);
if (readonly) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property %s can only be set at import time"),
zpool_prop_to_name(ZPOOL_PROP_READONLY));
return (-1);
}
}
if (nvlist_lookup_nvlist_array(tree, ZPOOL_CONFIG_CHILDREN, &child,
&children) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Source pool is missing vdev tree"));
nvlist_free(zc_props);
return (-1);
}
varray = zfs_alloc(hdl, children * sizeof (nvlist_t *));
vcount = 0;
if (*newroot == NULL ||
nvlist_lookup_nvlist_array(*newroot, ZPOOL_CONFIG_CHILDREN,
&newchild, &newchildren) != 0)
newchildren = 0;
for (c = 0; c < children; c++) {
uint64_t is_log = B_FALSE, is_hole = B_FALSE;
boolean_t is_special = B_FALSE, is_dedup = B_FALSE;
const char *type;
nvlist_t **mchild, *vdev;
uint_t mchildren;
int entry;
/*
* Unlike cache & spares, slogs are stored in the
* ZPOOL_CONFIG_CHILDREN array. We filter them out here.
*/
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG,
&is_log);
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE,
&is_hole);
if (is_log || is_hole) {
/*
* Create a hole vdev and put it in the config.
*/
if (nvlist_alloc(&vdev, NV_UNIQUE_NAME, 0) != 0)
goto out;
if (nvlist_add_string(vdev, ZPOOL_CONFIG_TYPE,
VDEV_TYPE_HOLE) != 0)
goto out;
if (nvlist_add_uint64(vdev, ZPOOL_CONFIG_IS_HOLE,
1) != 0)
goto out;
if (lastlog == 0)
lastlog = vcount;
varray[vcount++] = vdev;
continue;
}
lastlog = 0;
type = fnvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE);
if (strcmp(type, VDEV_TYPE_INDIRECT) == 0) {
vdev = child[c];
if (nvlist_dup(vdev, &varray[vcount++], 0) != 0)
goto out;
continue;
} else if (strcmp(type, VDEV_TYPE_MIRROR) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Source pool must be composed only of mirrors\n"));
retval = zfs_error(hdl, EZFS_INVALCONFIG, errbuf);
goto out;
}
if (nvlist_lookup_string(child[c],
ZPOOL_CONFIG_ALLOCATION_BIAS, &bias) == 0) {
if (strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0)
is_special = B_TRUE;
else if (strcmp(bias, VDEV_ALLOC_BIAS_DEDUP) == 0)
is_dedup = B_TRUE;
}
verify(nvlist_lookup_nvlist_array(child[c],
ZPOOL_CONFIG_CHILDREN, &mchild, &mchildren) == 0);
/* find or add an entry for this top-level vdev */
if (newchildren > 0 &&
(entry = find_vdev_entry(zhp, mchild, mchildren,
newchild, newchildren)) >= 0) {
/* We found a disk that the user specified. */
vdev = mchild[entry];
++found;
} else {
/* User didn't specify a disk for this vdev. */
vdev = mchild[mchildren - 1];
}
if (nvlist_dup(vdev, &varray[vcount++], 0) != 0)
goto out;
if (flags.dryrun != 0) {
if (is_dedup == B_TRUE) {
if (nvlist_add_string(varray[vcount - 1],
ZPOOL_CONFIG_ALLOCATION_BIAS,
VDEV_ALLOC_BIAS_DEDUP) != 0)
goto out;
} else if (is_special == B_TRUE) {
if (nvlist_add_string(varray[vcount - 1],
ZPOOL_CONFIG_ALLOCATION_BIAS,
VDEV_ALLOC_BIAS_SPECIAL) != 0)
goto out;
}
}
}
/* did we find every disk the user specified? */
if (found != newchildren) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Device list must "
"include at most one disk from each mirror"));
retval = zfs_error(hdl, EZFS_INVALCONFIG, errbuf);
goto out;
}
/* Prepare the nvlist for populating. */
if (*newroot == NULL) {
if (nvlist_alloc(newroot, NV_UNIQUE_NAME, 0) != 0)
goto out;
freelist = B_TRUE;
if (nvlist_add_string(*newroot, ZPOOL_CONFIG_TYPE,
VDEV_TYPE_ROOT) != 0)
goto out;
} else {
verify(nvlist_remove_all(*newroot, ZPOOL_CONFIG_CHILDREN) == 0);
}
/* Add all the children we found */
if (nvlist_add_nvlist_array(*newroot, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t **)varray, lastlog == 0 ? vcount : lastlog) != 0)
goto out;
/*
* If we're just doing a dry run, exit now with success.
*/
if (flags.dryrun) {
memory_err = B_FALSE;
freelist = B_FALSE;
goto out;
}
/* now build up the config list & call the ioctl */
if (nvlist_alloc(&newconfig, NV_UNIQUE_NAME, 0) != 0)
goto out;
if (nvlist_add_nvlist(newconfig,
ZPOOL_CONFIG_VDEV_TREE, *newroot) != 0 ||
nvlist_add_string(newconfig,
ZPOOL_CONFIG_POOL_NAME, newname) != 0 ||
nvlist_add_uint64(newconfig, ZPOOL_CONFIG_VERSION, vers) != 0)
goto out;
/*
* The new pool is automatically part of the namespace unless we
* explicitly export it.
*/
if (!flags.import)
zc.zc_cookie = ZPOOL_EXPORT_AFTER_SPLIT;
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
(void) strlcpy(zc.zc_string, newname, sizeof (zc.zc_string));
zcmd_write_conf_nvlist(hdl, &zc, newconfig);
if (zc_props != NULL)
zcmd_write_src_nvlist(hdl, &zc, zc_props);
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SPLIT, &zc) != 0) {
retval = zpool_standard_error(hdl, errno, errbuf);
goto out;
}
freelist = B_FALSE;
memory_err = B_FALSE;
out:
if (varray != NULL) {
int v;
for (v = 0; v < vcount; v++)
nvlist_free(varray[v]);
free(varray);
}
zcmd_free_nvlists(&zc);
nvlist_free(zc_props);
nvlist_free(newconfig);
if (freelist) {
nvlist_free(*newroot);
*newroot = NULL;
}
if (retval != 0)
return (retval);
if (memory_err)
return (no_memory(hdl));
return (0);
}
/*
* Remove the given device.
*/
int
zpool_vdev_remove(zpool_handle_t *zhp, const char *path)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache, islog;
libzfs_handle_t *hdl = zhp->zpool_hdl;
uint64_t version;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot remove %s"), path);
if (zpool_is_draid_spare(path)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dRAID spares cannot be removed"));
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
}
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
&islog)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL);
if (islog && version < SPA_VERSION_HOLES) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded to support log removal"));
return (zfs_error(hdl, EZFS_BADVERSION, errbuf));
}
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_REMOVE, &zc) == 0)
return (0);
switch (errno) {
case EINVAL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid config; all top-level vdevs must "
"have the same sector size and not be raidz."));
(void) zfs_error(hdl, EZFS_INVALCONFIG, errbuf);
break;
case EBUSY:
if (islog) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Mount encrypted datasets to replay logs."));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Pool busy; removal may already be in progress"));
}
(void) zfs_error(hdl, EZFS_BUSY, errbuf);
break;
case EACCES:
if (islog) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Mount encrypted datasets to replay logs."));
(void) zfs_error(hdl, EZFS_BUSY, errbuf);
} else {
(void) zpool_standard_error(hdl, errno, errbuf);
}
break;
default:
(void) zpool_standard_error(hdl, errno, errbuf);
}
return (-1);
}
int
zpool_vdev_remove_cancel(zpool_handle_t *zhp)
{
zfs_cmd_t zc = {{0}};
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot cancel removal"));
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_cookie = 1;
if (zfs_ioctl(hdl, ZFS_IOC_VDEV_REMOVE, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
int
zpool_vdev_indirect_size(zpool_handle_t *zhp, const char *path,
uint64_t *sizep)
{
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
boolean_t avail_spare, l2cache, islog;
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot determine indirect size of %s"),
path);
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache,
&islog)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
if (avail_spare || l2cache || islog) {
*sizep = 0;
return (0);
}
if (nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_INDIRECT_SIZE, sizep) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"indirect size not available"));
return (zfs_error(hdl, EINVAL, errbuf));
}
return (0);
}
/*
* Clear the errors for the pool, or the particular device if specified.
*/
int
zpool_clear(zpool_handle_t *zhp, const char *path, nvlist_t *rewindnvl)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
nvlist_t *tgt;
zpool_load_policy_t policy;
boolean_t avail_spare, l2cache;
libzfs_handle_t *hdl = zhp->zpool_hdl;
nvlist_t *nvi = NULL;
int error;
if (path)
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot clear errors for %s"),
path);
else
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot clear errors for %s"),
zhp->zpool_name);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if (path) {
if ((tgt = zpool_find_vdev(zhp, path, &avail_spare,
&l2cache, NULL)) == NULL)
return (zfs_error(hdl, EZFS_NODEVICE, errbuf));
/*
* Don't allow error clearing for hot spares. Do allow
* error clearing for l2cache devices.
*/
if (avail_spare)
return (zfs_error(hdl, EZFS_ISSPARE, errbuf));
zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
}
zpool_get_load_policy(rewindnvl, &policy);
zc.zc_cookie = policy.zlp_rewind;
zcmd_alloc_dst_nvlist(hdl, &zc, zhp->zpool_config_size * 2);
zcmd_write_src_nvlist(hdl, &zc, rewindnvl);
while ((error = zfs_ioctl(hdl, ZFS_IOC_CLEAR, &zc)) != 0 &&
errno == ENOMEM)
zcmd_expand_dst_nvlist(hdl, &zc);
if (!error || ((policy.zlp_rewind & ZPOOL_TRY_REWIND) &&
errno != EPERM && errno != EACCES)) {
if (policy.zlp_rewind &
(ZPOOL_DO_REWIND | ZPOOL_TRY_REWIND)) {
(void) zcmd_read_dst_nvlist(hdl, &zc, &nvi);
zpool_rewind_exclaim(hdl, zc.zc_name,
((policy.zlp_rewind & ZPOOL_TRY_REWIND) != 0),
nvi);
nvlist_free(nvi);
}
zcmd_free_nvlists(&zc);
return (0);
}
zcmd_free_nvlists(&zc);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Similar to zpool_clear(), but takes a GUID (used by fmd).
*/
int
zpool_vdev_clear(zpool_handle_t *zhp, uint64_t guid)
{
zfs_cmd_t zc = {"\0"};
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot clear errors for %llx"),
(u_longlong_t)guid);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_guid = guid;
zc.zc_cookie = ZPOOL_NO_REWIND;
if (zfs_ioctl(hdl, ZFS_IOC_CLEAR, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Change the GUID for a pool.
*/
int
zpool_reguid(zpool_handle_t *zhp)
{
char errbuf[ERRBUFLEN];
libzfs_handle_t *hdl = zhp->zpool_hdl;
zfs_cmd_t zc = {"\0"};
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot reguid '%s'"), zhp->zpool_name);
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
if (zfs_ioctl(hdl, ZFS_IOC_POOL_REGUID, &zc) == 0)
return (0);
return (zpool_standard_error(hdl, errno, errbuf));
}
/*
* Reopen the pool.
*/
int
zpool_reopen_one(zpool_handle_t *zhp, void *data)
{
libzfs_handle_t *hdl = zpool_get_handle(zhp);
const char *pool_name = zpool_get_name(zhp);
boolean_t *scrub_restart = data;
int error;
error = lzc_reopen(pool_name, *scrub_restart);
if (error) {
return (zpool_standard_error_fmt(hdl, error,
dgettext(TEXT_DOMAIN, "cannot reopen '%s'"), pool_name));
}
return (0);
}
/* call into libzfs_core to execute the sync IOCTL per pool */
int
zpool_sync_one(zpool_handle_t *zhp, void *data)
{
int ret;
libzfs_handle_t *hdl = zpool_get_handle(zhp);
const char *pool_name = zpool_get_name(zhp);
boolean_t *force = data;
nvlist_t *innvl = fnvlist_alloc();
fnvlist_add_boolean_value(innvl, "force", *force);
if ((ret = lzc_sync(pool_name, innvl, NULL)) != 0) {
nvlist_free(innvl);
return (zpool_standard_error_fmt(hdl, ret,
dgettext(TEXT_DOMAIN, "sync '%s' failed"), pool_name));
}
nvlist_free(innvl);
return (0);
}
#define PATH_BUF_LEN 64
/*
* Given a vdev, return the name to display in iostat. If the vdev has a path,
* we use that, stripping off any leading "/dev/dsk/"; if not, we use the type.
* We also check if this is a whole disk, in which case we strip off the
* trailing 's0' slice name.
*
* This routine is also responsible for identifying when disks have been
* reconfigured in a new location. The kernel will have opened the device by
* devid, but the path will still refer to the old location. To catch this, we
* first do a path -> devid translation (which is fast for the common case). If
* the devid matches, we're done. If not, we do a reverse devid -> path
* translation and issue the appropriate ioctl() to update the path of the vdev.
* If 'zhp' is NULL, then this is an exported pool, and we don't need to do any
* of these checks.
*/
char *
zpool_vdev_name(libzfs_handle_t *hdl, zpool_handle_t *zhp, nvlist_t *nv,
int name_flags)
{
const char *type, *tpath;
const char *path;
uint64_t value;
char buf[PATH_BUF_LEN];
char tmpbuf[PATH_BUF_LEN * 2];
/*
* vdev_name will be "root"/"root-0" for the root vdev, but it is the
* zpool name that will be displayed to the user.
*/
type = fnvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE);
if (zhp != NULL && strcmp(type, "root") == 0)
return (zfs_strdup(hdl, zpool_get_name(zhp)));
if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_PATH"))
name_flags |= VDEV_NAME_PATH;
if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_GUID"))
name_flags |= VDEV_NAME_GUID;
if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_FOLLOW_LINKS"))
name_flags |= VDEV_NAME_FOLLOW_LINKS;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &value) == 0 ||
name_flags & VDEV_NAME_GUID) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &value);
(void) snprintf(buf, sizeof (buf), "%llu", (u_longlong_t)value);
path = buf;
} else if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &tpath) == 0) {
path = tpath;
if (name_flags & VDEV_NAME_FOLLOW_LINKS) {
char *rp = realpath(path, NULL);
if (rp) {
strlcpy(buf, rp, sizeof (buf));
path = buf;
free(rp);
}
}
/*
* For a block device only use the name.
*/
if ((strcmp(type, VDEV_TYPE_DISK) == 0) &&
!(name_flags & VDEV_NAME_PATH)) {
path = zfs_strip_path(path);
}
/*
* Remove the partition from the path if this is a whole disk.
*/
if (strcmp(type, VDEV_TYPE_DRAID_SPARE) != 0 &&
nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &value)
== 0 && value && !(name_flags & VDEV_NAME_PATH)) {
return (zfs_strip_partition(path));
}
} else {
path = type;
/*
* If it's a raidz device, we need to stick in the parity level.
*/
if (strcmp(path, VDEV_TYPE_RAIDZ) == 0) {
value = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY);
(void) snprintf(buf, sizeof (buf), "%s%llu", path,
(u_longlong_t)value);
path = buf;
}
/*
* If it's a dRAID device, we add parity, groups, and spares.
*/
if (strcmp(path, VDEV_TYPE_DRAID) == 0) {
uint64_t ndata, nparity, nspares;
nvlist_t **child;
uint_t children;
verify(nvlist_lookup_nvlist_array(nv,
ZPOOL_CONFIG_CHILDREN, &child, &children) == 0);
nparity = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_NPARITY);
ndata = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_DRAID_NDATA);
nspares = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_DRAID_NSPARES);
path = zpool_draid_name(buf, sizeof (buf), ndata,
nparity, nspares, children);
}
/*
* We identify each top-level vdev by using a <type-id>
* naming convention.
*/
if (name_flags & VDEV_NAME_TYPE_ID) {
uint64_t id = fnvlist_lookup_uint64(nv,
ZPOOL_CONFIG_ID);
(void) snprintf(tmpbuf, sizeof (tmpbuf), "%s-%llu",
path, (u_longlong_t)id);
path = tmpbuf;
}
}
return (zfs_strdup(hdl, path));
}
static int
zbookmark_mem_compare(const void *a, const void *b)
{
return (memcmp(a, b, sizeof (zbookmark_phys_t)));
}
/*
* Retrieve the persistent error log, uniquify the members, and return to the
* caller.
*/
int
zpool_get_errlog(zpool_handle_t *zhp, nvlist_t **nverrlistp)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zpool_hdl;
zbookmark_phys_t *buf;
uint64_t buflen = 10000; /* approx. 1MB of RAM */
if (fnvlist_lookup_uint64(zhp->zpool_config,
ZPOOL_CONFIG_ERRCOUNT) == 0)
return (0);
/*
* Retrieve the raw error list from the kernel. If it doesn't fit,
* allocate a larger buffer and retry.
*/
(void) strcpy(zc.zc_name, zhp->zpool_name);
for (;;) {
buf = zfs_alloc(zhp->zpool_hdl,
buflen * sizeof (zbookmark_phys_t));
zc.zc_nvlist_dst = (uintptr_t)buf;
zc.zc_nvlist_dst_size = buflen;
if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_ERROR_LOG,
&zc) != 0) {
free(buf);
if (errno == ENOMEM) {
buflen *= 2;
} else {
return (zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "errors: List of "
"errors unavailable")));
}
} else {
break;
}
}
/*
* Sort the resulting bookmarks. This is a little confusing due to the
* implementation of ZFS_IOC_ERROR_LOG. The bookmarks are copied last
* to first, and 'zc_nvlist_dst_size' indicates the number of bookmarks
* _not_ copied as part of the process. So we point the start of our
* array appropriate and decrement the total number of elements.
*/
zbookmark_phys_t *zb = buf + zc.zc_nvlist_dst_size;
uint64_t zblen = buflen - zc.zc_nvlist_dst_size;
qsort(zb, zblen, sizeof (zbookmark_phys_t), zbookmark_mem_compare);
verify(nvlist_alloc(nverrlistp, 0, KM_SLEEP) == 0);
/*
* Fill in the nverrlistp with nvlist's of dataset and object numbers.
*/
for (uint64_t i = 0; i < zblen; i++) {
nvlist_t *nv;
/* ignoring zb_blkid and zb_level for now */
if (i > 0 && zb[i-1].zb_objset == zb[i].zb_objset &&
zb[i-1].zb_object == zb[i].zb_object)
continue;
if (nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP) != 0)
goto nomem;
if (nvlist_add_uint64(nv, ZPOOL_ERR_DATASET,
zb[i].zb_objset) != 0) {
nvlist_free(nv);
goto nomem;
}
if (nvlist_add_uint64(nv, ZPOOL_ERR_OBJECT,
zb[i].zb_object) != 0) {
nvlist_free(nv);
goto nomem;
}
if (nvlist_add_nvlist(*nverrlistp, "ejk", nv) != 0) {
nvlist_free(nv);
goto nomem;
}
nvlist_free(nv);
}
free(buf);
return (0);
nomem:
free(buf);
return (no_memory(zhp->zpool_hdl));
}
/*
* Upgrade a ZFS pool to the latest on-disk version.
*/
int
zpool_upgrade(zpool_handle_t *zhp, uint64_t new_version)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) strcpy(zc.zc_name, zhp->zpool_name);
zc.zc_cookie = new_version;
if (zfs_ioctl(hdl, ZFS_IOC_POOL_UPGRADE, &zc) != 0)
return (zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot upgrade '%s'"),
zhp->zpool_name));
return (0);
}
void
zfs_save_arguments(int argc, char **argv, char *string, int len)
{
int i;
(void) strlcpy(string, zfs_basename(argv[0]), len);
for (i = 1; i < argc; i++) {
(void) strlcat(string, " ", len);
(void) strlcat(string, argv[i], len);
}
}
int
zpool_log_history(libzfs_handle_t *hdl, const char *message)
{
zfs_cmd_t zc = {"\0"};
nvlist_t *args;
args = fnvlist_alloc();
fnvlist_add_string(args, "message", message);
zcmd_write_src_nvlist(hdl, &zc, args);
int err = zfs_ioctl(hdl, ZFS_IOC_LOG_HISTORY, &zc);
nvlist_free(args);
zcmd_free_nvlists(&zc);
return (err);
}
/*
* Perform ioctl to get some command history of a pool.
*
* 'buf' is the buffer to fill up to 'len' bytes. 'off' is the
* logical offset of the history buffer to start reading from.
*
* Upon return, 'off' is the next logical offset to read from and
* 'len' is the actual amount of bytes read into 'buf'.
*/
static int
get_history(zpool_handle_t *zhp, char *buf, uint64_t *off, uint64_t *len)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zpool_hdl;
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_history = (uint64_t)(uintptr_t)buf;
zc.zc_history_len = *len;
zc.zc_history_offset = *off;
if (zfs_ioctl(hdl, ZFS_IOC_POOL_GET_HISTORY, &zc) != 0) {
switch (errno) {
case EPERM:
return (zfs_error_fmt(hdl, EZFS_PERM,
dgettext(TEXT_DOMAIN,
"cannot show history for pool '%s'"),
zhp->zpool_name));
case ENOENT:
return (zfs_error_fmt(hdl, EZFS_NOHISTORY,
dgettext(TEXT_DOMAIN, "cannot get history for pool "
"'%s'"), zhp->zpool_name));
case ENOTSUP:
return (zfs_error_fmt(hdl, EZFS_BADVERSION,
dgettext(TEXT_DOMAIN, "cannot get history for pool "
"'%s', pool must be upgraded"), zhp->zpool_name));
default:
return (zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN,
"cannot get history for '%s'"), zhp->zpool_name));
}
}
*len = zc.zc_history_len;
*off = zc.zc_history_offset;
return (0);
}
/*
* Retrieve the command history of a pool.
*/
int
zpool_get_history(zpool_handle_t *zhp, nvlist_t **nvhisp, uint64_t *off,
boolean_t *eof)
{
libzfs_handle_t *hdl = zhp->zpool_hdl;
char *buf;
int buflen = 128 * 1024;
nvlist_t **records = NULL;
uint_t numrecords = 0;
int err = 0, i;
uint64_t start = *off;
buf = zfs_alloc(hdl, buflen);
/* process about 1MiB a time */
while (*off - start < 1024 * 1024) {
uint64_t bytes_read = buflen;
uint64_t leftover;
if ((err = get_history(zhp, buf, off, &bytes_read)) != 0)
break;
/* if nothing else was read in, we're at EOF, just return */
if (!bytes_read) {
*eof = B_TRUE;
break;
}
if ((err = zpool_history_unpack(buf, bytes_read,
&leftover, &records, &numrecords)) != 0) {
zpool_standard_error_fmt(hdl, err,
dgettext(TEXT_DOMAIN,
"cannot get history for '%s'"), zhp->zpool_name);
break;
}
*off -= leftover;
if (leftover == bytes_read) {
/*
* no progress made, because buffer is not big enough
* to hold this record; resize and retry.
*/
buflen *= 2;
free(buf);
buf = zfs_alloc(hdl, buflen);
}
}
free(buf);
if (!err) {
*nvhisp = fnvlist_alloc();
fnvlist_add_nvlist_array(*nvhisp, ZPOOL_HIST_RECORD,
(const nvlist_t **)records, numrecords);
}
for (i = 0; i < numrecords; i++)
nvlist_free(records[i]);
free(records);
return (err);
}
/*
* Retrieve the next event given the passed 'zevent_fd' file descriptor.
* If there is a new event available 'nvp' will contain a newly allocated
* nvlist and 'dropped' will be set to the number of missed events since
* the last call to this function. When 'nvp' is set to NULL it indicates
* no new events are available. In either case the function returns 0 and
* it is up to the caller to free 'nvp'. In the case of a fatal error the
* function will return a non-zero value. When the function is called in
* blocking mode (the default, unless the ZEVENT_NONBLOCK flag is passed),
* it will not return until a new event is available.
*/
int
zpool_events_next(libzfs_handle_t *hdl, nvlist_t **nvp,
int *dropped, unsigned flags, int zevent_fd)
{
zfs_cmd_t zc = {"\0"};
int error = 0;
*nvp = NULL;
*dropped = 0;
zc.zc_cleanup_fd = zevent_fd;
if (flags & ZEVENT_NONBLOCK)
zc.zc_guid = ZEVENT_NONBLOCK;
zcmd_alloc_dst_nvlist(hdl, &zc, ZEVENT_SIZE);
retry:
if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_NEXT, &zc) != 0) {
switch (errno) {
case ESHUTDOWN:
error = zfs_error_fmt(hdl, EZFS_POOLUNAVAIL,
dgettext(TEXT_DOMAIN, "zfs shutdown"));
goto out;
case ENOENT:
/* Blocking error case should not occur */
if (!(flags & ZEVENT_NONBLOCK))
error = zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot get event"));
goto out;
case ENOMEM:
zcmd_expand_dst_nvlist(hdl, &zc);
goto retry;
default:
error = zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot get event"));
goto out;
}
}
error = zcmd_read_dst_nvlist(hdl, &zc, nvp);
if (error != 0)
goto out;
*dropped = (int)zc.zc_cookie;
out:
zcmd_free_nvlists(&zc);
return (error);
}
/*
* Clear all events.
*/
int
zpool_events_clear(libzfs_handle_t *hdl, int *count)
{
zfs_cmd_t zc = {"\0"};
if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_CLEAR, &zc) != 0)
return (zpool_standard_error(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot clear events")));
if (count != NULL)
*count = (int)zc.zc_cookie; /* # of events cleared */
return (0);
}
/*
* Seek to a specific EID, ZEVENT_SEEK_START, or ZEVENT_SEEK_END for
* the passed zevent_fd file handle. On success zero is returned,
* otherwise -1 is returned and hdl->libzfs_error is set to the errno.
*/
int
zpool_events_seek(libzfs_handle_t *hdl, uint64_t eid, int zevent_fd)
{
zfs_cmd_t zc = {"\0"};
int error = 0;
zc.zc_guid = eid;
zc.zc_cleanup_fd = zevent_fd;
if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_SEEK, &zc) != 0) {
switch (errno) {
case ENOENT:
error = zfs_error_fmt(hdl, EZFS_NOENT,
dgettext(TEXT_DOMAIN, "cannot get event"));
break;
case ENOMEM:
error = zfs_error_fmt(hdl, EZFS_NOMEM,
dgettext(TEXT_DOMAIN, "cannot get event"));
break;
default:
error = zpool_standard_error_fmt(hdl, errno,
dgettext(TEXT_DOMAIN, "cannot get event"));
break;
}
}
return (error);
}
static void
zpool_obj_to_path_impl(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj,
char *pathname, size_t len, boolean_t always_unmounted)
{
zfs_cmd_t zc = {"\0"};
boolean_t mounted = B_FALSE;
char *mntpnt = NULL;
char dsname[ZFS_MAX_DATASET_NAME_LEN];
if (dsobj == 0) {
/* special case for the MOS */
(void) snprintf(pathname, len, "<metadata>:<0x%llx>",
(longlong_t)obj);
return;
}
/* get the dataset's name */
(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
zc.zc_obj = dsobj;
if (zfs_ioctl(zhp->zpool_hdl,
ZFS_IOC_DSOBJ_TO_DSNAME, &zc) != 0) {
/* just write out a path of two object numbers */
(void) snprintf(pathname, len, "<0x%llx>:<0x%llx>",
(longlong_t)dsobj, (longlong_t)obj);
return;
}
(void) strlcpy(dsname, zc.zc_value, sizeof (dsname));
/* find out if the dataset is mounted */
mounted = !always_unmounted && is_mounted(zhp->zpool_hdl, dsname,
&mntpnt);
/* get the corrupted object's path */
(void) strlcpy(zc.zc_name, dsname, sizeof (zc.zc_name));
zc.zc_obj = obj;
if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_OBJ_TO_PATH,
&zc) == 0) {
if (mounted) {
(void) snprintf(pathname, len, "%s%s", mntpnt,
zc.zc_value);
} else {
(void) snprintf(pathname, len, "%s:%s",
dsname, zc.zc_value);
}
} else {
(void) snprintf(pathname, len, "%s:<0x%llx>", dsname,
(longlong_t)obj);
}
free(mntpnt);
}
void
zpool_obj_to_path(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj,
char *pathname, size_t len)
{
zpool_obj_to_path_impl(zhp, dsobj, obj, pathname, len, B_FALSE);
}
void
zpool_obj_to_path_ds(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj,
char *pathname, size_t len)
{
zpool_obj_to_path_impl(zhp, dsobj, obj, pathname, len, B_TRUE);
}
/*
* Wait while the specified activity is in progress in the pool.
*/
int
zpool_wait(zpool_handle_t *zhp, zpool_wait_activity_t activity)
{
boolean_t missing;
int error = zpool_wait_status(zhp, activity, &missing, NULL);
if (missing) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl, ENOENT,
dgettext(TEXT_DOMAIN, "error waiting in pool '%s'"),
zhp->zpool_name);
return (ENOENT);
} else {
return (error);
}
}
/*
* Wait for the given activity and return the status of the wait (whether or not
* any waiting was done) in the 'waited' parameter. Non-existent pools are
* reported via the 'missing' parameter, rather than by printing an error
* message. This is convenient when this function is called in a loop over a
* long period of time (as it is, for example, by zpool's wait cmd). In that
* scenario, a pool being exported or destroyed should be considered a normal
* event, so we don't want to print an error when we find that the pool doesn't
* exist.
*/
int
zpool_wait_status(zpool_handle_t *zhp, zpool_wait_activity_t activity,
boolean_t *missing, boolean_t *waited)
{
int error = lzc_wait(zhp->zpool_name, activity, waited);
*missing = (error == ENOENT);
if (*missing)
return (0);
if (error != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl, error,
dgettext(TEXT_DOMAIN, "error waiting in pool '%s'"),
zhp->zpool_name);
}
return (error);
}
int
zpool_set_bootenv(zpool_handle_t *zhp, const nvlist_t *envmap)
{
int error = lzc_set_bootenv(zhp->zpool_name, envmap);
if (error != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl, error,
dgettext(TEXT_DOMAIN,
"error setting bootenv in pool '%s'"), zhp->zpool_name);
}
return (error);
}
int
zpool_get_bootenv(zpool_handle_t *zhp, nvlist_t **nvlp)
{
nvlist_t *nvl;
int error;
nvl = NULL;
error = lzc_get_bootenv(zhp->zpool_name, &nvl);
if (error != 0) {
(void) zpool_standard_error_fmt(zhp->zpool_hdl, error,
dgettext(TEXT_DOMAIN,
"error getting bootenv in pool '%s'"), zhp->zpool_name);
} else {
*nvlp = nvl;
}
return (error);
}
/*
* Attempt to read and parse feature file(s) (from "compatibility" property).
* Files contain zpool feature names, comma or whitespace-separated.
* Comments (# character to next newline) are discarded.
*
* Arguments:
* compatibility : string containing feature filenames
* features : either NULL or pointer to array of boolean
* report : either NULL or pointer to string buffer
* rlen : length of "report" buffer
*
* compatibility is NULL (unset), "", "off", "legacy", or list of
* comma-separated filenames. filenames should either be absolute,
* or relative to:
* 1) ZPOOL_SYSCONF_COMPAT_D (eg: /etc/zfs/compatibility.d) or
* 2) ZPOOL_DATA_COMPAT_D (eg: /usr/share/zfs/compatibility.d).
* (Unset), "" or "off" => enable all features
* "legacy" => disable all features
*
* Any feature names read from files which match unames in spa_feature_table
* will have the corresponding boolean set in the features array (if non-NULL).
* If more than one feature set specified, only features present in *all* of
* them will be set.
*
* "report" if not NULL will be populated with a suitable status message.
*
* Return values:
* ZPOOL_COMPATIBILITY_OK : files read and parsed ok
* ZPOOL_COMPATIBILITY_BADFILE : file too big or not a text file
* ZPOOL_COMPATIBILITY_BADTOKEN : SYSCONF file contains invalid feature name
* ZPOOL_COMPATIBILITY_WARNTOKEN : DATA file contains invalid feature name
* ZPOOL_COMPATIBILITY_NOFILES : no feature files found
*/
zpool_compat_status_t
zpool_load_compat(const char *compat, boolean_t *features, char *report,
size_t rlen)
{
int sdirfd, ddirfd, featfd;
struct stat fs;
char *fc;
char *ps, *ls, *ws;
char *file, *line, *word;
char l_compat[ZFS_MAXPROPLEN];
boolean_t ret_nofiles = B_TRUE;
boolean_t ret_badfile = B_FALSE;
boolean_t ret_badtoken = B_FALSE;
boolean_t ret_warntoken = B_FALSE;
/* special cases (unset), "" and "off" => enable all features */
if (compat == NULL || compat[0] == '\0' ||
strcmp(compat, ZPOOL_COMPAT_OFF) == 0) {
if (features != NULL)
for (uint_t i = 0; i < SPA_FEATURES; i++)
features[i] = B_TRUE;
if (report != NULL)
strlcpy(report, gettext("all features enabled"), rlen);
return (ZPOOL_COMPATIBILITY_OK);
}
/* Final special case "legacy" => disable all features */
if (strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0) {
if (features != NULL)
for (uint_t i = 0; i < SPA_FEATURES; i++)
features[i] = B_FALSE;
if (report != NULL)
strlcpy(report, gettext("all features disabled"), rlen);
return (ZPOOL_COMPATIBILITY_OK);
}
/*
* Start with all true; will be ANDed with results from each file
*/
if (features != NULL)
for (uint_t i = 0; i < SPA_FEATURES; i++)
features[i] = B_TRUE;
char err_badfile[ZFS_MAXPROPLEN] = "";
char err_badtoken[ZFS_MAXPROPLEN] = "";
/*
* We ignore errors from the directory open()
* as they're only needed if the filename is relative
* which will be checked during the openat().
*/
/* O_PATH safer than O_RDONLY if system allows it */
#if defined(O_PATH)
#define ZC_DIR_FLAGS (O_DIRECTORY | O_CLOEXEC | O_PATH)
#else
#define ZC_DIR_FLAGS (O_DIRECTORY | O_CLOEXEC | O_RDONLY)
#endif
sdirfd = open(ZPOOL_SYSCONF_COMPAT_D, ZC_DIR_FLAGS);
ddirfd = open(ZPOOL_DATA_COMPAT_D, ZC_DIR_FLAGS);
(void) strlcpy(l_compat, compat, ZFS_MAXPROPLEN);
for (file = strtok_r(l_compat, ",", &ps);
file != NULL;
file = strtok_r(NULL, ",", &ps)) {
boolean_t l_features[SPA_FEATURES];
enum { Z_SYSCONF, Z_DATA } source;
/* try sysconfdir first, then datadir */
source = Z_SYSCONF;
if ((featfd = openat(sdirfd, file, O_RDONLY | O_CLOEXEC)) < 0) {
featfd = openat(ddirfd, file, O_RDONLY | O_CLOEXEC);
source = Z_DATA;
}
/* File readable and correct size? */
if (featfd < 0 ||
fstat(featfd, &fs) < 0 ||
fs.st_size < 1 ||
fs.st_size > ZPOOL_COMPAT_MAXSIZE) {
(void) close(featfd);
strlcat(err_badfile, file, ZFS_MAXPROPLEN);
strlcat(err_badfile, " ", ZFS_MAXPROPLEN);
ret_badfile = B_TRUE;
continue;
}
/* Prefault the file if system allows */
#if defined(MAP_POPULATE)
#define ZC_MMAP_FLAGS (MAP_PRIVATE | MAP_POPULATE)
#elif defined(MAP_PREFAULT_READ)
#define ZC_MMAP_FLAGS (MAP_PRIVATE | MAP_PREFAULT_READ)
#else
#define ZC_MMAP_FLAGS (MAP_PRIVATE)
#endif
/* private mmap() so we can strtok safely */
fc = (char *)mmap(NULL, fs.st_size, PROT_READ | PROT_WRITE,
ZC_MMAP_FLAGS, featfd, 0);
(void) close(featfd);
/* map ok, and last character == newline? */
if (fc == MAP_FAILED || fc[fs.st_size - 1] != '\n') {
(void) munmap((void *) fc, fs.st_size);
strlcat(err_badfile, file, ZFS_MAXPROPLEN);
strlcat(err_badfile, " ", ZFS_MAXPROPLEN);
ret_badfile = B_TRUE;
continue;
}
ret_nofiles = B_FALSE;
for (uint_t i = 0; i < SPA_FEATURES; i++)
l_features[i] = B_FALSE;
/* replace final newline with NULL to ensure string ends */
fc[fs.st_size - 1] = '\0';
for (line = strtok_r(fc, "\n", &ls);
line != NULL;
line = strtok_r(NULL, "\n", &ls)) {
/* discard comments */
char *r = strchr(line, '#');
if (r != NULL)
*r = '\0';
for (word = strtok_r(line, ", \t", &ws);
word != NULL;
word = strtok_r(NULL, ", \t", &ws)) {
/* Find matching feature name */
uint_t f;
for (f = 0; f < SPA_FEATURES; f++) {
zfeature_info_t *fi =
&spa_feature_table[f];
if (strcmp(word, fi->fi_uname) == 0) {
l_features[f] = B_TRUE;
break;
}
}
if (f < SPA_FEATURES)
continue;
/* found an unrecognized word */
/* lightly sanitize it */
if (strlen(word) > 32)
word[32] = '\0';
for (char *c = word; *c != '\0'; c++)
if (!isprint(*c))
*c = '?';
strlcat(err_badtoken, word, ZFS_MAXPROPLEN);
strlcat(err_badtoken, " ", ZFS_MAXPROPLEN);
if (source == Z_SYSCONF)
ret_badtoken = B_TRUE;
else
ret_warntoken = B_TRUE;
}
}
(void) munmap((void *) fc, fs.st_size);
if (features != NULL)
for (uint_t i = 0; i < SPA_FEATURES; i++)
features[i] &= l_features[i];
}
(void) close(sdirfd);
(void) close(ddirfd);
/* Return the most serious error */
if (ret_badfile) {
if (report != NULL)
snprintf(report, rlen, gettext("could not read/"
"parse feature file(s): %s"), err_badfile);
return (ZPOOL_COMPATIBILITY_BADFILE);
}
if (ret_nofiles) {
if (report != NULL)
strlcpy(report,
gettext("no valid compatibility files specified"),
rlen);
return (ZPOOL_COMPATIBILITY_NOFILES);
}
if (ret_badtoken) {
if (report != NULL)
snprintf(report, rlen, gettext("invalid feature "
"name(s) in local compatibility files: %s"),
err_badtoken);
return (ZPOOL_COMPATIBILITY_BADTOKEN);
}
if (ret_warntoken) {
if (report != NULL)
snprintf(report, rlen, gettext("unrecognized feature "
"name(s) in distribution compatibility files: %s"),
err_badtoken);
return (ZPOOL_COMPATIBILITY_WARNTOKEN);
}
if (report != NULL)
strlcpy(report, gettext("compatibility set ok"), rlen);
return (ZPOOL_COMPATIBILITY_OK);
}
static int
zpool_vdev_guid(zpool_handle_t *zhp, const char *vdevname, uint64_t *vdev_guid)
{
nvlist_t *tgt;
boolean_t avail_spare, l2cache;
verify(zhp != NULL);
if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "pool is in an unavailable state"));
return (zfs_error(zhp->zpool_hdl, EZFS_POOLUNAVAIL, errbuf));
}
if ((tgt = zpool_find_vdev(zhp, vdevname, &avail_spare, &l2cache,
NULL)) == NULL) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "can not find %s in %s"),
vdevname, zhp->zpool_name);
return (zfs_error(zhp->zpool_hdl, EZFS_NODEVICE, errbuf));
}
*vdev_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID);
return (0);
}
/*
* Get a vdev property value for 'prop' and return the value in
* a pre-allocated buffer.
*/
int
zpool_get_vdev_prop_value(nvlist_t *nvprop, vdev_prop_t prop, char *prop_name,
char *buf, size_t len, zprop_source_t *srctype, boolean_t literal)
{
nvlist_t *nv;
const char *strval;
uint64_t intval;
zprop_source_t src = ZPROP_SRC_NONE;
if (prop == VDEV_PROP_USERPROP) {
/* user property, prop_name must contain the property name */
assert(prop_name != NULL);
if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) {
src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
strval = fnvlist_lookup_string(nv, ZPROP_VALUE);
} else {
/* user prop not found */
return (-1);
}
(void) strlcpy(buf, strval, len);
if (srctype)
*srctype = src;
return (0);
}
if (prop_name == NULL)
prop_name = (char *)vdev_prop_to_name(prop);
switch (vdev_prop_get_type(prop)) {
case PROP_TYPE_STRING:
if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) {
src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
strval = fnvlist_lookup_string(nv, ZPROP_VALUE);
} else {
src = ZPROP_SRC_DEFAULT;
if ((strval = vdev_prop_default_string(prop)) == NULL)
strval = "-";
}
(void) strlcpy(buf, strval, len);
break;
case PROP_TYPE_NUMBER:
if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) {
src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
intval = fnvlist_lookup_uint64(nv, ZPROP_VALUE);
} else {
src = ZPROP_SRC_DEFAULT;
intval = vdev_prop_default_numeric(prop);
}
switch (prop) {
case VDEV_PROP_ASIZE:
case VDEV_PROP_PSIZE:
case VDEV_PROP_SIZE:
case VDEV_PROP_BOOTSIZE:
case VDEV_PROP_ALLOCATED:
case VDEV_PROP_FREE:
case VDEV_PROP_READ_ERRORS:
case VDEV_PROP_WRITE_ERRORS:
case VDEV_PROP_CHECKSUM_ERRORS:
case VDEV_PROP_INITIALIZE_ERRORS:
case VDEV_PROP_OPS_NULL:
case VDEV_PROP_OPS_READ:
case VDEV_PROP_OPS_WRITE:
case VDEV_PROP_OPS_FREE:
case VDEV_PROP_OPS_CLAIM:
case VDEV_PROP_OPS_TRIM:
case VDEV_PROP_BYTES_NULL:
case VDEV_PROP_BYTES_READ:
case VDEV_PROP_BYTES_WRITE:
case VDEV_PROP_BYTES_FREE:
case VDEV_PROP_BYTES_CLAIM:
case VDEV_PROP_BYTES_TRIM:
if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) zfs_nicenum(intval, buf, len);
}
break;
case VDEV_PROP_EXPANDSZ:
if (intval == 0) {
(void) strlcpy(buf, "-", len);
} else if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) zfs_nicenum(intval, buf, len);
}
break;
case VDEV_PROP_CAPACITY:
if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) snprintf(buf, len, "%llu%%",
(u_longlong_t)intval);
}
break;
case VDEV_PROP_CHECKSUM_N:
case VDEV_PROP_CHECKSUM_T:
case VDEV_PROP_IO_N:
case VDEV_PROP_IO_T:
if (intval == UINT64_MAX) {
(void) strlcpy(buf, "-", len);
} else {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
}
break;
case VDEV_PROP_FRAGMENTATION:
if (intval == UINT64_MAX) {
(void) strlcpy(buf, "-", len);
} else {
(void) snprintf(buf, len, "%llu%%",
(u_longlong_t)intval);
}
break;
case VDEV_PROP_STATE:
if (literal) {
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
} else {
(void) strlcpy(buf, zpool_state_to_name(intval,
VDEV_AUX_NONE), len);
}
break;
default:
(void) snprintf(buf, len, "%llu",
(u_longlong_t)intval);
}
break;
case PROP_TYPE_INDEX:
if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) {
src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE);
intval = fnvlist_lookup_uint64(nv, ZPROP_VALUE);
} else {
src = ZPROP_SRC_DEFAULT;
intval = vdev_prop_default_numeric(prop);
}
if (vdev_prop_index_to_string(prop, intval,
(const char **)&strval) != 0)
return (-1);
(void) strlcpy(buf, strval, len);
break;
default:
abort();
}
if (srctype)
*srctype = src;
return (0);
}
/*
* Get a vdev property value for 'prop_name' and return the value in
* a pre-allocated buffer.
*/
int
zpool_get_vdev_prop(zpool_handle_t *zhp, const char *vdevname, vdev_prop_t prop,
char *prop_name, char *buf, size_t len, zprop_source_t *srctype,
boolean_t literal)
{
nvlist_t *reqnvl, *reqprops;
nvlist_t *retprops = NULL;
uint64_t vdev_guid = 0;
int ret;
if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0)
return (ret);
if (nvlist_alloc(&reqnvl, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
if (nvlist_alloc(&reqprops, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
fnvlist_add_uint64(reqnvl, ZPOOL_VDEV_PROPS_GET_VDEV, vdev_guid);
if (prop != VDEV_PROP_USERPROP) {
/* prop_name overrides prop value */
if (prop_name != NULL)
prop = vdev_name_to_prop(prop_name);
else
prop_name = (char *)vdev_prop_to_name(prop);
assert(prop < VDEV_NUM_PROPS);
}
assert(prop_name != NULL);
if (nvlist_add_uint64(reqprops, prop_name, prop) != 0) {
nvlist_free(reqnvl);
nvlist_free(reqprops);
return (no_memory(zhp->zpool_hdl));
}
fnvlist_add_nvlist(reqnvl, ZPOOL_VDEV_PROPS_GET_PROPS, reqprops);
ret = lzc_get_vdev_prop(zhp->zpool_name, reqnvl, &retprops);
if (ret == 0) {
ret = zpool_get_vdev_prop_value(retprops, prop, prop_name, buf,
len, srctype, literal);
} else {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot get vdev property %s from"
" %s in %s"), prop_name, vdevname, zhp->zpool_name);
(void) zpool_standard_error(zhp->zpool_hdl, ret, errbuf);
}
nvlist_free(reqnvl);
nvlist_free(reqprops);
nvlist_free(retprops);
return (ret);
}
/*
* Get all vdev properties
*/
int
zpool_get_all_vdev_props(zpool_handle_t *zhp, const char *vdevname,
nvlist_t **outnvl)
{
nvlist_t *nvl = NULL;
uint64_t vdev_guid = 0;
int ret;
if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0)
return (ret);
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
fnvlist_add_uint64(nvl, ZPOOL_VDEV_PROPS_GET_VDEV, vdev_guid);
ret = lzc_get_vdev_prop(zhp->zpool_name, nvl, outnvl);
nvlist_free(nvl);
if (ret) {
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot get vdev properties for"
" %s in %s"), vdevname, zhp->zpool_name);
(void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf);
}
return (ret);
}
/*
* Set vdev property
*/
int
zpool_set_vdev_prop(zpool_handle_t *zhp, const char *vdevname,
const char *propname, const char *propval)
{
int ret;
nvlist_t *nvl = NULL;
nvlist_t *outnvl = NULL;
nvlist_t *props;
nvlist_t *realprops;
prop_flags_t flags = { 0 };
uint64_t version;
uint64_t vdev_guid;
if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0)
return (ret);
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0)
return (no_memory(zhp->zpool_hdl));
fnvlist_add_uint64(nvl, ZPOOL_VDEV_PROPS_SET_VDEV, vdev_guid);
if (nvlist_add_string(props, propname, propval) != 0) {
nvlist_free(props);
return (no_memory(zhp->zpool_hdl));
}
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "cannot set property %s for %s on %s"),
propname, vdevname, zhp->zpool_name);
flags.vdevprop = 1;
version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL);
if ((realprops = zpool_valid_proplist(zhp->zpool_hdl,
zhp->zpool_name, props, version, flags, errbuf)) == NULL) {
nvlist_free(props);
nvlist_free(nvl);
return (-1);
}
nvlist_free(props);
props = realprops;
fnvlist_add_nvlist(nvl, ZPOOL_VDEV_PROPS_SET_PROPS, props);
ret = lzc_set_vdev_prop(zhp->zpool_name, nvl, &outnvl);
nvlist_free(props);
nvlist_free(nvl);
nvlist_free(outnvl);
if (ret)
(void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf);
return (ret);
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c b/sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c
index 23402f86a8ae..87a30f54fea8 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_sendrecv.c
@@ -1,5563 +1,5564 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2012, Joyent, Inc. All rights reserved.
* Copyright (c) 2012 Pawel Jakub Dawidek <pawel@dawidek.net>.
* All rights reserved
* Copyright (c) 2013 Steven Hartland. All rights reserved.
* Copyright 2015, OmniTI Computer Consulting, Inc. All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>
* Copyright (c) 2018, loli10K <ezomori.nozomu@gmail.com>. All rights reserved.
* Copyright (c) 2019 Datto Inc.
*/
#include <assert.h>
#include <ctype.h>
#include <errno.h>
#include <libintl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <stddef.h>
#include <fcntl.h>
#include <sys/mount.h>
#include <sys/mntent.h>
#include <sys/mnttab.h>
#include <sys/avl.h>
#include <sys/debug.h>
#include <sys/stat.h>
#include <pthread.h>
#include <umem.h>
#include <time.h>
#include <libzfs.h>
#include <libzfs_core.h>
#include <libzutil.h>
#include "zfs_namecheck.h"
#include "zfs_prop.h"
#include "zfs_fletcher.h"
#include "libzfs_impl.h"
#include <cityhash.h>
#include <zlib.h>
#include <sys/zio_checksum.h>
#include <sys/dsl_crypt.h>
#include <sys/ddt.h>
#include <sys/socket.h>
#include <sys/sha2.h>
static int zfs_receive_impl(libzfs_handle_t *, const char *, const char *,
recvflags_t *, int, const char *, nvlist_t *, avl_tree_t *, char **,
const char *, nvlist_t *);
static int guid_to_name_redact_snaps(libzfs_handle_t *hdl, const char *parent,
uint64_t guid, boolean_t bookmark_ok, uint64_t *redact_snap_guids,
uint64_t num_redact_snaps, char *name);
static int guid_to_name(libzfs_handle_t *, const char *,
uint64_t, boolean_t, char *);
typedef struct progress_arg {
zfs_handle_t *pa_zhp;
int pa_fd;
boolean_t pa_parsable;
boolean_t pa_estimate;
int pa_verbosity;
boolean_t pa_astitle;
boolean_t pa_progress;
uint64_t pa_size;
} progress_arg_t;
static int
dump_record(dmu_replay_record_t *drr, void *payload, size_t payload_len,
zio_cksum_t *zc, int outfd)
{
ASSERT3U(offsetof(dmu_replay_record_t, drr_u.drr_checksum.drr_checksum),
==, sizeof (dmu_replay_record_t) - sizeof (zio_cksum_t));
fletcher_4_incremental_native(drr,
offsetof(dmu_replay_record_t, drr_u.drr_checksum.drr_checksum), zc);
if (drr->drr_type != DRR_BEGIN) {
ASSERT(ZIO_CHECKSUM_IS_ZERO(&drr->drr_u.
drr_checksum.drr_checksum));
drr->drr_u.drr_checksum.drr_checksum = *zc;
}
fletcher_4_incremental_native(&drr->drr_u.drr_checksum.drr_checksum,
sizeof (zio_cksum_t), zc);
if (write(outfd, drr, sizeof (*drr)) == -1)
return (errno);
if (payload_len != 0) {
fletcher_4_incremental_native(payload, payload_len, zc);
if (write(outfd, payload, payload_len) == -1)
return (errno);
}
return (0);
}
/*
* Routines for dealing with the AVL tree of fs-nvlists
*/
typedef struct fsavl_node {
avl_node_t fn_node;
nvlist_t *fn_nvfs;
const char *fn_snapname;
uint64_t fn_guid;
} fsavl_node_t;
static int
fsavl_compare(const void *arg1, const void *arg2)
{
const fsavl_node_t *fn1 = (const fsavl_node_t *)arg1;
const fsavl_node_t *fn2 = (const fsavl_node_t *)arg2;
return (TREE_CMP(fn1->fn_guid, fn2->fn_guid));
}
/*
* Given the GUID of a snapshot, find its containing filesystem and
* (optionally) name.
*/
static nvlist_t *
fsavl_find(avl_tree_t *avl, uint64_t snapguid, const char **snapname)
{
fsavl_node_t fn_find;
fsavl_node_t *fn;
fn_find.fn_guid = snapguid;
fn = avl_find(avl, &fn_find, NULL);
if (fn) {
if (snapname)
*snapname = fn->fn_snapname;
return (fn->fn_nvfs);
}
return (NULL);
}
static void
fsavl_destroy(avl_tree_t *avl)
{
fsavl_node_t *fn;
void *cookie;
if (avl == NULL)
return;
cookie = NULL;
while ((fn = avl_destroy_nodes(avl, &cookie)) != NULL)
free(fn);
avl_destroy(avl);
free(avl);
}
/*
* Given an nvlist, produce an avl tree of snapshots, ordered by guid
*/
static avl_tree_t *
fsavl_create(nvlist_t *fss)
{
avl_tree_t *fsavl;
nvpair_t *fselem = NULL;
if ((fsavl = malloc(sizeof (avl_tree_t))) == NULL)
return (NULL);
avl_create(fsavl, fsavl_compare, sizeof (fsavl_node_t),
offsetof(fsavl_node_t, fn_node));
while ((fselem = nvlist_next_nvpair(fss, fselem)) != NULL) {
nvlist_t *nvfs, *snaps;
nvpair_t *snapelem = NULL;
nvfs = fnvpair_value_nvlist(fselem);
snaps = fnvlist_lookup_nvlist(nvfs, "snaps");
while ((snapelem =
nvlist_next_nvpair(snaps, snapelem)) != NULL) {
fsavl_node_t *fn;
if ((fn = malloc(sizeof (fsavl_node_t))) == NULL) {
fsavl_destroy(fsavl);
return (NULL);
}
fn->fn_nvfs = nvfs;
fn->fn_snapname = nvpair_name(snapelem);
fn->fn_guid = fnvpair_value_uint64(snapelem);
/*
* Note: if there are multiple snaps with the
* same GUID, we ignore all but one.
*/
avl_index_t where = 0;
if (avl_find(fsavl, fn, &where) == NULL)
avl_insert(fsavl, fn, where);
else
free(fn);
}
}
return (fsavl);
}
/*
* Routines for dealing with the giant nvlist of fs-nvlists, etc.
*/
typedef struct send_data {
/*
* assigned inside every recursive call,
* restored from *_save on return:
*
* guid of fromsnap snapshot in parent dataset
* txg of fromsnap snapshot in current dataset
* txg of tosnap snapshot in current dataset
*/
uint64_t parent_fromsnap_guid;
uint64_t fromsnap_txg;
uint64_t tosnap_txg;
/* the nvlists get accumulated during depth-first traversal */
nvlist_t *parent_snaps;
nvlist_t *fss;
nvlist_t *snapprops;
nvlist_t *snapholds; /* user holds */
/* send-receive configuration, does not change during traversal */
const char *fsname;
const char *fromsnap;
const char *tosnap;
boolean_t recursive;
boolean_t raw;
boolean_t doall;
boolean_t replicate;
boolean_t skipmissing;
boolean_t verbose;
boolean_t backup;
boolean_t seenfrom;
boolean_t seento;
boolean_t holds; /* were holds requested with send -h */
boolean_t props;
/*
* The header nvlist is of the following format:
* {
* "tosnap" -> string
* "fromsnap" -> string (if incremental)
* "fss" -> {
* id -> {
*
* "name" -> string (full name; for debugging)
* "parentfromsnap" -> number (guid of fromsnap in parent)
*
* "props" -> { name -> value (only if set here) }
* "snaps" -> { name (lastname) -> number (guid) }
* "snapprops" -> { name (lastname) -> { name -> value } }
* "snapholds" -> { name (lastname) -> { holdname -> crtime } }
*
* "origin" -> number (guid) (if clone)
* "is_encroot" -> boolean
* "sent" -> boolean (not on-disk)
* }
* }
* }
*
*/
} send_data_t;
static void
send_iterate_prop(zfs_handle_t *zhp, boolean_t received_only, nvlist_t *nv);
/*
* Collect guid, valid props, optionally holds, etc. of a snapshot.
- * This interface is intended for use as a zfs_iter_snapshots_sorted visitor.
+ * This interface is intended for use as a zfs_iter_snapshots_v2_sorted visitor.
*/
static int
send_iterate_snap(zfs_handle_t *zhp, void *arg)
{
send_data_t *sd = arg;
uint64_t guid = zhp->zfs_dmustats.dds_guid;
uint64_t txg = zhp->zfs_dmustats.dds_creation_txg;
boolean_t isfromsnap, istosnap, istosnapwithnofrom;
char *snapname;
const char *from = sd->fromsnap;
const char *to = sd->tosnap;
snapname = strrchr(zhp->zfs_name, '@');
assert(snapname != NULL);
++snapname;
isfromsnap = (from != NULL && strcmp(from, snapname) == 0);
istosnap = (to != NULL && strcmp(to, snapname) == 0);
istosnapwithnofrom = (istosnap && from == NULL);
if (sd->tosnap_txg != 0 && txg > sd->tosnap_txg) {
if (sd->verbose) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"skipping snapshot %s because it was created "
"after the destination snapshot (%s)\n"),
zhp->zfs_name, to);
}
zfs_close(zhp);
return (0);
}
fnvlist_add_uint64(sd->parent_snaps, snapname, guid);
/*
* NB: if there is no fromsnap here (it's a newly created fs in
* an incremental replication), we will substitute the tosnap.
*/
if (isfromsnap || (sd->parent_fromsnap_guid == 0 && istosnap))
sd->parent_fromsnap_guid = guid;
if (!sd->recursive) {
/*
* To allow a doall stream to work properly
* with a NULL fromsnap
*/
if (sd->doall && from == NULL && !sd->seenfrom)
sd->seenfrom = B_TRUE;
if (!sd->seenfrom && isfromsnap) {
sd->seenfrom = B_TRUE;
zfs_close(zhp);
return (0);
}
if ((sd->seento || !sd->seenfrom) && !istosnapwithnofrom) {
zfs_close(zhp);
return (0);
}
if (istosnap)
sd->seento = B_TRUE;
}
nvlist_t *nv = fnvlist_alloc();
send_iterate_prop(zhp, sd->backup, nv);
fnvlist_add_nvlist(sd->snapprops, snapname, nv);
fnvlist_free(nv);
if (sd->holds) {
nvlist_t *holds;
if (lzc_get_holds(zhp->zfs_name, &holds) == 0) {
fnvlist_add_nvlist(sd->snapholds, snapname, holds);
fnvlist_free(holds);
}
}
zfs_close(zhp);
return (0);
}
/*
* Collect all valid props from the handle snap into an nvlist.
*/
static void
send_iterate_prop(zfs_handle_t *zhp, boolean_t received_only, nvlist_t *nv)
{
nvlist_t *props;
if (received_only)
props = zfs_get_recvd_props(zhp);
else
props = zhp->zfs_props;
nvpair_t *elem = NULL;
while ((elem = nvlist_next_nvpair(props, elem)) != NULL) {
const char *propname = nvpair_name(elem);
zfs_prop_t prop = zfs_name_to_prop(propname);
if (!zfs_prop_user(propname)) {
/*
* Realistically, this should never happen. However,
* we want the ability to add DSL properties without
* needing to make incompatible version changes. We
* need to ignore unknown properties to allow older
* software to still send datasets containing these
* properties, with the unknown properties elided.
*/
if (prop == ZPROP_INVAL)
continue;
if (zfs_prop_readonly(prop))
continue;
}
nvlist_t *propnv = fnvpair_value_nvlist(elem);
boolean_t isspacelimit = (prop == ZFS_PROP_QUOTA ||
prop == ZFS_PROP_RESERVATION ||
prop == ZFS_PROP_REFQUOTA ||
prop == ZFS_PROP_REFRESERVATION);
if (isspacelimit && zhp->zfs_type == ZFS_TYPE_SNAPSHOT)
continue;
const char *source;
if (nvlist_lookup_string(propnv, ZPROP_SOURCE, &source) == 0) {
if (strcmp(source, zhp->zfs_name) != 0 &&
strcmp(source, ZPROP_SOURCE_VAL_RECVD) != 0)
continue;
} else {
/*
* May have no source before SPA_VERSION_RECVD_PROPS,
* but is still modifiable.
*/
if (!isspacelimit)
continue;
}
if (zfs_prop_user(propname) ||
zfs_prop_get_type(prop) == PROP_TYPE_STRING) {
const char *value;
value = fnvlist_lookup_string(propnv, ZPROP_VALUE);
fnvlist_add_string(nv, propname, value);
} else {
uint64_t value;
value = fnvlist_lookup_uint64(propnv, ZPROP_VALUE);
fnvlist_add_uint64(nv, propname, value);
}
}
}
/*
* returns snapshot guid
* and returns 0 if the snapshot does not exist
*/
static uint64_t
get_snap_guid(libzfs_handle_t *hdl, const char *fs, const char *snap)
{
char name[MAXPATHLEN + 1];
uint64_t guid = 0;
if (fs == NULL || fs[0] == '\0' || snap == NULL || snap[0] == '\0')
return (guid);
(void) snprintf(name, sizeof (name), "%s@%s", fs, snap);
zfs_handle_t *zhp = zfs_open(hdl, name, ZFS_TYPE_SNAPSHOT);
if (zhp != NULL) {
guid = zfs_prop_get_int(zhp, ZFS_PROP_GUID);
zfs_close(zhp);
}
return (guid);
}
/*
* returns snapshot creation txg
* and returns 0 if the snapshot does not exist
*/
static uint64_t
get_snap_txg(libzfs_handle_t *hdl, const char *fs, const char *snap)
{
char name[ZFS_MAX_DATASET_NAME_LEN];
uint64_t txg = 0;
if (fs == NULL || fs[0] == '\0' || snap == NULL || snap[0] == '\0')
return (txg);
(void) snprintf(name, sizeof (name), "%s@%s", fs, snap);
if (zfs_dataset_exists(hdl, name, ZFS_TYPE_SNAPSHOT)) {
zfs_handle_t *zhp = zfs_open(hdl, name, ZFS_TYPE_SNAPSHOT);
if (zhp != NULL) {
txg = zfs_prop_get_int(zhp, ZFS_PROP_CREATETXG);
zfs_close(zhp);
}
}
return (txg);
}
/*
* Recursively generate nvlists describing datasets. See comment
* for the data structure send_data_t above for description of contents
* of the nvlist.
*/
static int
send_iterate_fs(zfs_handle_t *zhp, void *arg)
{
send_data_t *sd = arg;
nvlist_t *nvfs = NULL, *nv = NULL;
int rv = 0;
uint64_t min_txg = 0, max_txg = 0;
uint64_t txg = zhp->zfs_dmustats.dds_creation_txg;
uint64_t guid = zhp->zfs_dmustats.dds_guid;
uint64_t fromsnap_txg, tosnap_txg;
char guidstring[64];
/* These fields are restored on return from a recursive call. */
uint64_t parent_fromsnap_guid_save = sd->parent_fromsnap_guid;
uint64_t fromsnap_txg_save = sd->fromsnap_txg;
uint64_t tosnap_txg_save = sd->tosnap_txg;
fromsnap_txg = get_snap_txg(zhp->zfs_hdl, zhp->zfs_name, sd->fromsnap);
if (fromsnap_txg != 0)
sd->fromsnap_txg = fromsnap_txg;
tosnap_txg = get_snap_txg(zhp->zfs_hdl, zhp->zfs_name, sd->tosnap);
if (tosnap_txg != 0)
sd->tosnap_txg = tosnap_txg;
/*
* On the send side, if the current dataset does not have tosnap,
* perform two additional checks:
*
* - Skip sending the current dataset if it was created later than
* the parent tosnap.
* - Return error if the current dataset was created earlier than
* the parent tosnap, unless --skip-missing specified. Then
* just print a warning.
*/
if (sd->tosnap != NULL && tosnap_txg == 0) {
if (sd->tosnap_txg != 0 && txg > sd->tosnap_txg) {
if (sd->verbose) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"skipping dataset %s: snapshot %s does "
"not exist\n"), zhp->zfs_name, sd->tosnap);
}
} else if (sd->skipmissing) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"WARNING: skipping dataset %s and its children:"
" snapshot %s does not exist\n"),
zhp->zfs_name, sd->tosnap);
} else {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"cannot send %s@%s%s: snapshot %s@%s does not "
"exist\n"), sd->fsname, sd->tosnap, sd->recursive ?
dgettext(TEXT_DOMAIN, " recursively") : "",
zhp->zfs_name, sd->tosnap);
rv = EZFS_NOENT;
}
goto out;
}
nvfs = fnvlist_alloc();
fnvlist_add_string(nvfs, "name", zhp->zfs_name);
fnvlist_add_uint64(nvfs, "parentfromsnap", sd->parent_fromsnap_guid);
if (zhp->zfs_dmustats.dds_origin[0] != '\0') {
zfs_handle_t *origin = zfs_open(zhp->zfs_hdl,
zhp->zfs_dmustats.dds_origin, ZFS_TYPE_SNAPSHOT);
if (origin == NULL) {
rv = -1;
goto out;
}
fnvlist_add_uint64(nvfs, "origin",
origin->zfs_dmustats.dds_guid);
zfs_close(origin);
}
/* Iterate over props. */
if (sd->props || sd->backup || sd->recursive) {
nv = fnvlist_alloc();
send_iterate_prop(zhp, sd->backup, nv);
fnvlist_add_nvlist(nvfs, "props", nv);
}
if (zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION) != ZIO_CRYPT_OFF) {
boolean_t encroot;
/* Determine if this dataset is an encryption root. */
if (zfs_crypto_get_encryption_root(zhp, &encroot, NULL) != 0) {
rv = -1;
goto out;
}
if (encroot)
fnvlist_add_boolean(nvfs, "is_encroot");
/*
* Encrypted datasets can only be sent with properties if
* the raw flag is specified because the receive side doesn't
* currently have a mechanism for recursively asking the user
* for new encryption parameters.
*/
if (!sd->raw) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"cannot send %s@%s: encrypted dataset %s may not "
"be sent with properties without the raw flag\n"),
sd->fsname, sd->tosnap, zhp->zfs_name);
rv = -1;
goto out;
}
}
/*
* Iterate over snaps, and set sd->parent_fromsnap_guid.
*
* If this is a "doall" send, a replicate send or we're just trying
* to gather a list of previous snapshots, iterate through all the
* snaps in the txg range. Otherwise just look at the one we're
* interested in.
*/
sd->parent_fromsnap_guid = 0;
sd->parent_snaps = fnvlist_alloc();
sd->snapprops = fnvlist_alloc();
if (sd->holds)
sd->snapholds = fnvlist_alloc();
if (sd->doall || sd->replicate || sd->tosnap == NULL) {
if (!sd->replicate && fromsnap_txg != 0)
min_txg = fromsnap_txg;
if (!sd->replicate && tosnap_txg != 0)
max_txg = tosnap_txg;
- (void) zfs_iter_snapshots_sorted(zhp, 0, send_iterate_snap, sd,
- min_txg, max_txg);
+ (void) zfs_iter_snapshots_sorted_v2(zhp, 0, send_iterate_snap,
+ sd, min_txg, max_txg);
} else {
char snapname[MAXPATHLEN] = { 0 };
zfs_handle_t *snap;
(void) snprintf(snapname, sizeof (snapname), "%s@%s",
zhp->zfs_name, sd->tosnap);
if (sd->fromsnap != NULL)
sd->seenfrom = B_TRUE;
snap = zfs_open(zhp->zfs_hdl, snapname, ZFS_TYPE_SNAPSHOT);
if (snap != NULL)
(void) send_iterate_snap(snap, sd);
}
fnvlist_add_nvlist(nvfs, "snaps", sd->parent_snaps);
fnvlist_free(sd->parent_snaps);
fnvlist_add_nvlist(nvfs, "snapprops", sd->snapprops);
fnvlist_free(sd->snapprops);
if (sd->holds) {
fnvlist_add_nvlist(nvfs, "snapholds", sd->snapholds);
fnvlist_free(sd->snapholds);
}
/* Do not allow the size of the properties list to exceed the limit */
if ((fnvlist_size(nvfs) + fnvlist_size(sd->fss)) >
zhp->zfs_hdl->libzfs_max_nvlist) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"warning: cannot send %s@%s: the size of the list of "
"snapshots and properties is too large to be received "
"successfully.\n"
"Select a smaller number of snapshots to send.\n"),
zhp->zfs_name, sd->tosnap);
rv = EZFS_NOSPC;
goto out;
}
/* Add this fs to nvlist. */
(void) snprintf(guidstring, sizeof (guidstring),
"0x%llx", (longlong_t)guid);
fnvlist_add_nvlist(sd->fss, guidstring, nvfs);
/* Iterate over children. */
if (sd->recursive)
- rv = zfs_iter_filesystems(zhp, 0, send_iterate_fs, sd);
+ rv = zfs_iter_filesystems_v2(zhp, 0, send_iterate_fs, sd);
out:
/* Restore saved fields. */
sd->parent_fromsnap_guid = parent_fromsnap_guid_save;
sd->fromsnap_txg = fromsnap_txg_save;
sd->tosnap_txg = tosnap_txg_save;
fnvlist_free(nv);
fnvlist_free(nvfs);
zfs_close(zhp);
return (rv);
}
static int
gather_nvlist(libzfs_handle_t *hdl, const char *fsname, const char *fromsnap,
const char *tosnap, boolean_t recursive, boolean_t raw, boolean_t doall,
boolean_t replicate, boolean_t skipmissing, boolean_t verbose,
boolean_t backup, boolean_t holds, boolean_t props, nvlist_t **nvlp,
avl_tree_t **avlp)
{
zfs_handle_t *zhp;
send_data_t sd = { 0 };
int error;
zhp = zfs_open(hdl, fsname, ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL)
return (EZFS_BADTYPE);
sd.fss = fnvlist_alloc();
sd.fsname = fsname;
sd.fromsnap = fromsnap;
sd.tosnap = tosnap;
sd.recursive = recursive;
sd.raw = raw;
sd.doall = doall;
sd.replicate = replicate;
sd.skipmissing = skipmissing;
sd.verbose = verbose;
sd.backup = backup;
sd.holds = holds;
sd.props = props;
if ((error = send_iterate_fs(zhp, &sd)) != 0) {
fnvlist_free(sd.fss);
if (avlp != NULL)
*avlp = NULL;
*nvlp = NULL;
return (error);
}
if (avlp != NULL && (*avlp = fsavl_create(sd.fss)) == NULL) {
fnvlist_free(sd.fss);
*nvlp = NULL;
return (EZFS_NOMEM);
}
*nvlp = sd.fss;
return (0);
}
/*
* Routines specific to "zfs send"
*/
typedef struct send_dump_data {
/* these are all just the short snapname (the part after the @) */
const char *fromsnap;
const char *tosnap;
char prevsnap[ZFS_MAX_DATASET_NAME_LEN];
uint64_t prevsnap_obj;
boolean_t seenfrom, seento, replicate, doall, fromorigin;
boolean_t dryrun, parsable, progress, embed_data, std_out;
boolean_t large_block, compress, raw, holds;
boolean_t progressastitle;
int outfd;
boolean_t err;
nvlist_t *fss;
nvlist_t *snapholds;
avl_tree_t *fsavl;
snapfilter_cb_t *filter_cb;
void *filter_cb_arg;
nvlist_t *debugnv;
char holdtag[ZFS_MAX_DATASET_NAME_LEN];
int cleanup_fd;
int verbosity;
uint64_t size;
} send_dump_data_t;
static int
zfs_send_space(zfs_handle_t *zhp, const char *snapname, const char *from,
enum lzc_send_flags flags, uint64_t *spacep)
{
assert(snapname != NULL);
int error = lzc_send_space(snapname, from, flags, spacep);
if (error == 0)
return (0);
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"warning: cannot estimate space for '%s'"), snapname);
libzfs_handle_t *hdl = zhp->zfs_hdl;
switch (error) {
case EXDEV:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"not an earlier snapshot from the same fs"));
return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf));
case ENOENT:
if (zfs_dataset_exists(hdl, snapname,
ZFS_TYPE_SNAPSHOT)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incremental source (%s) does not exist"),
snapname);
}
return (zfs_error(hdl, EZFS_NOENT, errbuf));
case EDQUOT:
case EFBIG:
case EIO:
case ENOLINK:
case ENOSPC:
case ENOSTR:
case ENXIO:
case EPIPE:
case ERANGE:
case EFAULT:
case EROFS:
case EINVAL:
zfs_error_aux(hdl, "%s", strerror(error));
return (zfs_error(hdl, EZFS_BADBACKUP, errbuf));
default:
return (zfs_standard_error(hdl, error, errbuf));
}
}
/*
* Dumps a backup of the given snapshot (incremental from fromsnap if it's not
* NULL) to the file descriptor specified by outfd.
*/
static int
dump_ioctl(zfs_handle_t *zhp, const char *fromsnap, uint64_t fromsnap_obj,
boolean_t fromorigin, int outfd, enum lzc_send_flags flags,
nvlist_t *debugnv)
{
zfs_cmd_t zc = {"\0"};
libzfs_handle_t *hdl = zhp->zfs_hdl;
nvlist_t *thisdbg;
assert(zhp->zfs_type == ZFS_TYPE_SNAPSHOT);
assert(fromsnap_obj == 0 || !fromorigin);
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
zc.zc_cookie = outfd;
zc.zc_obj = fromorigin;
zc.zc_sendobj = zfs_prop_get_int(zhp, ZFS_PROP_OBJSETID);
zc.zc_fromobj = fromsnap_obj;
zc.zc_flags = flags;
if (debugnv != NULL) {
thisdbg = fnvlist_alloc();
if (fromsnap != NULL && fromsnap[0] != '\0')
fnvlist_add_string(thisdbg, "fromsnap", fromsnap);
}
if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_SEND, &zc) != 0) {
char errbuf[ERRBUFLEN];
int error = errno;
(void) snprintf(errbuf, sizeof (errbuf), "%s '%s'",
dgettext(TEXT_DOMAIN, "warning: cannot send"),
zhp->zfs_name);
if (debugnv != NULL) {
fnvlist_add_uint64(thisdbg, "error", error);
fnvlist_add_nvlist(debugnv, zhp->zfs_name, thisdbg);
fnvlist_free(thisdbg);
}
switch (error) {
case EXDEV:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"not an earlier snapshot from the same fs"));
return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf));
case EACCES:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"source key must be loaded"));
return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf));
case ENOENT:
if (zfs_dataset_exists(hdl, zc.zc_name,
ZFS_TYPE_SNAPSHOT)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incremental source (@%s) does not exist"),
zc.zc_value);
}
return (zfs_error(hdl, EZFS_NOENT, errbuf));
case EDQUOT:
case EFBIG:
case EIO:
case ENOLINK:
case ENOSPC:
case ENOSTR:
case ENXIO:
case EPIPE:
case ERANGE:
case EFAULT:
case EROFS:
case EINVAL:
zfs_error_aux(hdl, "%s", strerror(errno));
return (zfs_error(hdl, EZFS_BADBACKUP, errbuf));
default:
return (zfs_standard_error(hdl, errno, errbuf));
}
}
if (debugnv != NULL) {
fnvlist_add_nvlist(debugnv, zhp->zfs_name, thisdbg);
fnvlist_free(thisdbg);
}
return (0);
}
static void
gather_holds(zfs_handle_t *zhp, send_dump_data_t *sdd)
{
assert(zhp->zfs_type == ZFS_TYPE_SNAPSHOT);
/*
* zfs_send() only sets snapholds for sends that need them,
* e.g. replication and doall.
*/
if (sdd->snapholds == NULL)
return;
fnvlist_add_string(sdd->snapholds, zhp->zfs_name, sdd->holdtag);
}
int
zfs_send_progress(zfs_handle_t *zhp, int fd, uint64_t *bytes_written,
uint64_t *blocks_visited)
{
zfs_cmd_t zc = {"\0"};
if (bytes_written != NULL)
*bytes_written = 0;
if (blocks_visited != NULL)
*blocks_visited = 0;
(void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name));
zc.zc_cookie = fd;
if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_SEND_PROGRESS, &zc) != 0)
return (errno);
if (bytes_written != NULL)
*bytes_written = zc.zc_cookie;
if (blocks_visited != NULL)
*blocks_visited = zc.zc_objset_type;
return (0);
}
static void *
send_progress_thread(void *arg)
{
progress_arg_t *pa = arg;
zfs_handle_t *zhp = pa->pa_zhp;
uint64_t bytes;
uint64_t blocks;
uint64_t total = pa->pa_size / 100;
char buf[16];
time_t t;
struct tm tm;
int err;
if (!pa->pa_parsable && pa->pa_progress) {
(void) fprintf(stderr,
"TIME %s %sSNAPSHOT %s\n",
pa->pa_estimate ? "BYTES" : " SENT",
pa->pa_verbosity >= 2 ? " BLOCKS " : "",
zhp->zfs_name);
}
/*
* Print the progress from ZFS_IOC_SEND_PROGRESS every second.
*/
for (;;) {
(void) sleep(1);
if ((err = zfs_send_progress(zhp, pa->pa_fd, &bytes,
&blocks)) != 0) {
if (err == EINTR || err == ENOENT)
return ((void *)0);
return ((void *)(uintptr_t)err);
}
(void) time(&t);
localtime_r(&t, &tm);
if (pa->pa_astitle) {
char buf_bytes[16];
char buf_size[16];
int pct;
zfs_nicenum(bytes, buf_bytes, sizeof (buf_bytes));
zfs_nicenum(pa->pa_size, buf_size, sizeof (buf_size));
pct = (total > 0) ? bytes / total : 100;
zfs_setproctitle("sending %s (%d%%: %s/%s)",
zhp->zfs_name, MIN(pct, 100), buf_bytes, buf_size);
}
if (pa->pa_verbosity >= 2 && pa->pa_parsable) {
(void) fprintf(stderr,
"%02d:%02d:%02d\t%llu\t%llu\t%s\n",
tm.tm_hour, tm.tm_min, tm.tm_sec,
(u_longlong_t)bytes, (u_longlong_t)blocks,
zhp->zfs_name);
} else if (pa->pa_verbosity >= 2) {
zfs_nicenum(bytes, buf, sizeof (buf));
(void) fprintf(stderr,
"%02d:%02d:%02d %5s %8llu %s\n",
tm.tm_hour, tm.tm_min, tm.tm_sec,
buf, (u_longlong_t)blocks, zhp->zfs_name);
} else if (pa->pa_parsable) {
(void) fprintf(stderr, "%02d:%02d:%02d\t%llu\t%s\n",
tm.tm_hour, tm.tm_min, tm.tm_sec,
(u_longlong_t)bytes, zhp->zfs_name);
} else if (pa->pa_progress) {
zfs_nicebytes(bytes, buf, sizeof (buf));
(void) fprintf(stderr, "%02d:%02d:%02d %5s %s\n",
tm.tm_hour, tm.tm_min, tm.tm_sec,
buf, zhp->zfs_name);
}
}
}
static boolean_t
send_progress_thread_exit(libzfs_handle_t *hdl, pthread_t ptid)
{
void *status = NULL;
(void) pthread_cancel(ptid);
(void) pthread_join(ptid, &status);
int error = (int)(uintptr_t)status;
if (error != 0 && status != PTHREAD_CANCELED)
return (zfs_standard_error(hdl, error,
dgettext(TEXT_DOMAIN, "progress thread exited nonzero")));
else
return (B_FALSE);
}
static void
send_print_verbose(FILE *fout, const char *tosnap, const char *fromsnap,
uint64_t size, boolean_t parsable)
{
if (parsable) {
if (fromsnap != NULL) {
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
"incremental\t%s\t%s"), fromsnap, tosnap);
} else {
/*
* Workaround for GCC 12+ with UBSan enabled deficencies.
*
* GCC 12+ invoked with -fsanitize=undefined incorrectly reports the code
* below as violating -Wformat-overflow.
*/
#if defined(__GNUC__) && !defined(__clang__) && \
defined(ZFS_UBSAN_ENABLED) && defined(HAVE_FORMAT_OVERFLOW)
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wformat-overflow"
#endif
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
"full\t%s"), tosnap);
#if defined(__GNUC__) && !defined(__clang__) && \
defined(ZFS_UBSAN_ENABLED) && defined(HAVE_FORMAT_OVERFLOW)
#pragma GCC diagnostic pop
#endif
}
(void) fprintf(fout, "\t%llu", (longlong_t)size);
} else {
if (fromsnap != NULL) {
if (strchr(fromsnap, '@') == NULL &&
strchr(fromsnap, '#') == NULL) {
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
"send from @%s to %s"), fromsnap, tosnap);
} else {
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
"send from %s to %s"), fromsnap, tosnap);
}
} else {
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
"full send of %s"), tosnap);
}
if (size != 0) {
char buf[16];
zfs_nicebytes(size, buf, sizeof (buf));
/*
* Workaround for GCC 12+ with UBSan enabled deficencies.
*
* GCC 12+ invoked with -fsanitize=undefined incorrectly reports the code
* below as violating -Wformat-overflow.
*/
#if defined(__GNUC__) && !defined(__clang__) && \
defined(ZFS_UBSAN_ENABLED) && defined(HAVE_FORMAT_OVERFLOW)
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wformat-overflow"
#endif
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
" estimated size is %s"), buf);
#if defined(__GNUC__) && !defined(__clang__) && \
defined(ZFS_UBSAN_ENABLED) && defined(HAVE_FORMAT_OVERFLOW)
#pragma GCC diagnostic pop
#endif
}
}
(void) fprintf(fout, "\n");
}
/*
* Send a single filesystem snapshot, updating the send dump data.
- * This interface is intended for use as a zfs_iter_snapshots_sorted visitor.
+ * This interface is intended for use as a zfs_iter_snapshots_v2_sorted visitor.
*/
static int
dump_snapshot(zfs_handle_t *zhp, void *arg)
{
send_dump_data_t *sdd = arg;
progress_arg_t pa = { 0 };
pthread_t tid;
char *thissnap;
enum lzc_send_flags flags = 0;
int err;
boolean_t isfromsnap, istosnap, fromorigin;
boolean_t exclude = B_FALSE;
FILE *fout = sdd->std_out ? stdout : stderr;
err = 0;
thissnap = strchr(zhp->zfs_name, '@') + 1;
isfromsnap = (sdd->fromsnap != NULL &&
strcmp(sdd->fromsnap, thissnap) == 0);
if (!sdd->seenfrom && isfromsnap) {
gather_holds(zhp, sdd);
sdd->seenfrom = B_TRUE;
(void) strlcpy(sdd->prevsnap, thissnap, sizeof (sdd->prevsnap));
sdd->prevsnap_obj = zfs_prop_get_int(zhp, ZFS_PROP_OBJSETID);
zfs_close(zhp);
return (0);
}
if (sdd->seento || !sdd->seenfrom) {
zfs_close(zhp);
return (0);
}
istosnap = (strcmp(sdd->tosnap, thissnap) == 0);
if (istosnap)
sdd->seento = B_TRUE;
if (sdd->large_block)
flags |= LZC_SEND_FLAG_LARGE_BLOCK;
if (sdd->embed_data)
flags |= LZC_SEND_FLAG_EMBED_DATA;
if (sdd->compress)
flags |= LZC_SEND_FLAG_COMPRESS;
if (sdd->raw)
flags |= LZC_SEND_FLAG_RAW;
if (!sdd->doall && !isfromsnap && !istosnap) {
if (sdd->replicate) {
const char *snapname;
nvlist_t *snapprops;
/*
* Filter out all intermediate snapshots except origin
* snapshots needed to replicate clones.
*/
nvlist_t *nvfs = fsavl_find(sdd->fsavl,
zhp->zfs_dmustats.dds_guid, &snapname);
if (nvfs != NULL) {
snapprops = fnvlist_lookup_nvlist(nvfs,
"snapprops");
snapprops = fnvlist_lookup_nvlist(snapprops,
thissnap);
exclude = !nvlist_exists(snapprops,
"is_clone_origin");
}
} else {
exclude = B_TRUE;
}
}
/*
* If a filter function exists, call it to determine whether
* this snapshot will be sent.
*/
if (exclude || (sdd->filter_cb != NULL &&
sdd->filter_cb(zhp, sdd->filter_cb_arg) == B_FALSE)) {
/*
* This snapshot is filtered out. Don't send it, and don't
* set prevsnap_obj, so it will be as if this snapshot didn't
* exist, and the next accepted snapshot will be sent as
* an incremental from the last accepted one, or as the
* first (and full) snapshot in the case of a replication,
* non-incremental send.
*/
zfs_close(zhp);
return (0);
}
gather_holds(zhp, sdd);
fromorigin = sdd->prevsnap[0] == '\0' &&
(sdd->fromorigin || sdd->replicate);
if (sdd->verbosity != 0) {
uint64_t size = 0;
char fromds[ZFS_MAX_DATASET_NAME_LEN];
if (sdd->prevsnap[0] != '\0') {
(void) strlcpy(fromds, zhp->zfs_name, sizeof (fromds));
*(strchr(fromds, '@') + 1) = '\0';
(void) strlcat(fromds, sdd->prevsnap, sizeof (fromds));
}
if (zfs_send_space(zhp, zhp->zfs_name,
sdd->prevsnap[0] ? fromds : NULL, flags, &size) == 0) {
send_print_verbose(fout, zhp->zfs_name,
sdd->prevsnap[0] ? sdd->prevsnap : NULL,
size, sdd->parsable);
sdd->size += size;
}
}
if (!sdd->dryrun) {
/*
* If progress reporting is requested, spawn a new thread to
* poll ZFS_IOC_SEND_PROGRESS at a regular interval.
*/
if (sdd->progress || sdd->progressastitle) {
pa.pa_zhp = zhp;
pa.pa_fd = sdd->outfd;
pa.pa_parsable = sdd->parsable;
pa.pa_estimate = B_FALSE;
pa.pa_verbosity = sdd->verbosity;
pa.pa_size = sdd->size;
pa.pa_astitle = sdd->progressastitle;
pa.pa_progress = sdd->progress;
if ((err = pthread_create(&tid, NULL,
send_progress_thread, &pa)) != 0) {
zfs_close(zhp);
return (err);
}
}
err = dump_ioctl(zhp, sdd->prevsnap, sdd->prevsnap_obj,
fromorigin, sdd->outfd, flags, sdd->debugnv);
if ((sdd->progress || sdd->progressastitle) &&
send_progress_thread_exit(zhp->zfs_hdl, tid))
return (-1);
}
(void) strlcpy(sdd->prevsnap, thissnap, sizeof (sdd->prevsnap));
sdd->prevsnap_obj = zfs_prop_get_int(zhp, ZFS_PROP_OBJSETID);
zfs_close(zhp);
return (err);
}
/*
* Send all snapshots for a filesystem, updating the send dump data.
*/
static int
dump_filesystem(zfs_handle_t *zhp, send_dump_data_t *sdd)
{
int rv = 0;
boolean_t missingfrom = B_FALSE;
zfs_cmd_t zc = {"\0"};
uint64_t min_txg = 0, max_txg = 0;
/*
* Make sure the tosnap exists.
*/
(void) snprintf(zc.zc_name, sizeof (zc.zc_name), "%s@%s",
zhp->zfs_name, sdd->tosnap);
if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_OBJSET_STATS, &zc) != 0) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"WARNING: could not send %s@%s: does not exist\n"),
zhp->zfs_name, sdd->tosnap);
sdd->err = B_TRUE;
return (0);
}
/*
* If this fs does not have fromsnap, and we're doing
* recursive, we need to send a full stream from the
* beginning (or an incremental from the origin if this
* is a clone). If we're doing non-recursive, then let
* them get the error.
*/
if (sdd->replicate && sdd->fromsnap) {
/*
* Make sure the fromsnap exists.
*/
(void) snprintf(zc.zc_name, sizeof (zc.zc_name), "%s@%s",
zhp->zfs_name, sdd->fromsnap);
if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_OBJSET_STATS, &zc) != 0)
missingfrom = B_TRUE;
}
sdd->seenfrom = sdd->seento = B_FALSE;
sdd->prevsnap[0] = '\0';
sdd->prevsnap_obj = 0;
if (sdd->fromsnap == NULL || missingfrom)
sdd->seenfrom = B_TRUE;
/*
* Iterate through all snapshots and process the ones we will be
* sending. If we only have a "from" and "to" snapshot to deal
* with, we can avoid iterating through all the other snapshots.
*/
if (sdd->doall || sdd->replicate || sdd->tosnap == NULL) {
if (!sdd->replicate) {
if (sdd->fromsnap != NULL) {
min_txg = get_snap_txg(zhp->zfs_hdl,
zhp->zfs_name, sdd->fromsnap);
}
if (sdd->tosnap != NULL) {
max_txg = get_snap_txg(zhp->zfs_hdl,
zhp->zfs_name, sdd->tosnap);
}
}
- rv = zfs_iter_snapshots_sorted(zhp, 0, dump_snapshot, sdd,
+ rv = zfs_iter_snapshots_sorted_v2(zhp, 0, dump_snapshot, sdd,
min_txg, max_txg);
} else {
char snapname[MAXPATHLEN] = { 0 };
zfs_handle_t *snap;
/* Dump fromsnap. */
if (!sdd->seenfrom) {
(void) snprintf(snapname, sizeof (snapname),
"%s@%s", zhp->zfs_name, sdd->fromsnap);
snap = zfs_open(zhp->zfs_hdl, snapname,
ZFS_TYPE_SNAPSHOT);
if (snap != NULL)
rv = dump_snapshot(snap, sdd);
else
rv = errno;
}
/* Dump tosnap. */
if (rv == 0) {
(void) snprintf(snapname, sizeof (snapname),
"%s@%s", zhp->zfs_name, sdd->tosnap);
snap = zfs_open(zhp->zfs_hdl, snapname,
ZFS_TYPE_SNAPSHOT);
if (snap != NULL)
rv = dump_snapshot(snap, sdd);
else
rv = errno;
}
}
if (!sdd->seenfrom) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"WARNING: could not send %s@%s:\n"
"incremental source (%s@%s) does not exist\n"),
zhp->zfs_name, sdd->tosnap,
zhp->zfs_name, sdd->fromsnap);
sdd->err = B_TRUE;
} else if (!sdd->seento) {
if (sdd->fromsnap) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"WARNING: could not send %s@%s:\n"
"incremental source (%s@%s) "
"is not earlier than it\n"),
zhp->zfs_name, sdd->tosnap,
zhp->zfs_name, sdd->fromsnap);
} else {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"WARNING: "
"could not send %s@%s: does not exist\n"),
zhp->zfs_name, sdd->tosnap);
}
sdd->err = B_TRUE;
}
return (rv);
}
/*
* Send all snapshots for all filesystems in sdd.
*/
static int
dump_filesystems(zfs_handle_t *rzhp, send_dump_data_t *sdd)
{
nvpair_t *fspair;
boolean_t needagain, progress;
if (!sdd->replicate)
return (dump_filesystem(rzhp, sdd));
/* Mark the clone origin snapshots. */
for (fspair = nvlist_next_nvpair(sdd->fss, NULL); fspair;
fspair = nvlist_next_nvpair(sdd->fss, fspair)) {
nvlist_t *nvfs;
uint64_t origin_guid = 0;
nvfs = fnvpair_value_nvlist(fspair);
(void) nvlist_lookup_uint64(nvfs, "origin", &origin_guid);
if (origin_guid != 0) {
const char *snapname;
nvlist_t *origin_nv = fsavl_find(sdd->fsavl,
origin_guid, &snapname);
if (origin_nv != NULL) {
nvlist_t *snapprops;
snapprops = fnvlist_lookup_nvlist(origin_nv,
"snapprops");
snapprops = fnvlist_lookup_nvlist(snapprops,
snapname);
fnvlist_add_boolean(snapprops,
"is_clone_origin");
}
}
}
again:
needagain = progress = B_FALSE;
for (fspair = nvlist_next_nvpair(sdd->fss, NULL); fspair;
fspair = nvlist_next_nvpair(sdd->fss, fspair)) {
nvlist_t *fslist, *parent_nv;
const char *fsname;
zfs_handle_t *zhp;
int err;
uint64_t origin_guid = 0;
uint64_t parent_guid = 0;
fslist = fnvpair_value_nvlist(fspair);
if (nvlist_lookup_boolean(fslist, "sent") == 0)
continue;
fsname = fnvlist_lookup_string(fslist, "name");
(void) nvlist_lookup_uint64(fslist, "origin", &origin_guid);
(void) nvlist_lookup_uint64(fslist, "parentfromsnap",
&parent_guid);
if (parent_guid != 0) {
parent_nv = fsavl_find(sdd->fsavl, parent_guid, NULL);
if (!nvlist_exists(parent_nv, "sent")) {
/* Parent has not been sent; skip this one. */
needagain = B_TRUE;
continue;
}
}
if (origin_guid != 0) {
nvlist_t *origin_nv = fsavl_find(sdd->fsavl,
origin_guid, NULL);
if (origin_nv != NULL &&
!nvlist_exists(origin_nv, "sent")) {
/*
* Origin has not been sent yet;
* skip this clone.
*/
needagain = B_TRUE;
continue;
}
}
zhp = zfs_open(rzhp->zfs_hdl, fsname, ZFS_TYPE_DATASET);
if (zhp == NULL)
return (-1);
err = dump_filesystem(zhp, sdd);
fnvlist_add_boolean(fslist, "sent");
progress = B_TRUE;
zfs_close(zhp);
if (err)
return (err);
}
if (needagain) {
assert(progress);
goto again;
}
/* Clean out the sent flags in case we reuse this fss. */
for (fspair = nvlist_next_nvpair(sdd->fss, NULL); fspair;
fspair = nvlist_next_nvpair(sdd->fss, fspair)) {
nvlist_t *fslist;
fslist = fnvpair_value_nvlist(fspair);
(void) nvlist_remove_all(fslist, "sent");
}
return (0);
}
nvlist_t *
zfs_send_resume_token_to_nvlist(libzfs_handle_t *hdl, const char *token)
{
unsigned int version;
int nread, i;
unsigned long long checksum, packed_len;
/*
* Decode token header, which is:
* <token version>-<checksum of payload>-<uncompressed payload length>
* Note that the only supported token version is 1.
*/
nread = sscanf(token, "%u-%llx-%llx-",
&version, &checksum, &packed_len);
if (nread != 3) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"resume token is corrupt (invalid format)"));
return (NULL);
}
if (version != ZFS_SEND_RESUME_TOKEN_VERSION) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"resume token is corrupt (invalid version %u)"),
version);
return (NULL);
}
/* Convert hexadecimal representation to binary. */
token = strrchr(token, '-') + 1;
int len = strlen(token) / 2;
unsigned char *compressed = zfs_alloc(hdl, len);
for (i = 0; i < len; i++) {
nread = sscanf(token + i * 2, "%2hhx", compressed + i);
if (nread != 1) {
free(compressed);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"resume token is corrupt "
"(payload is not hex-encoded)"));
return (NULL);
}
}
/* Verify checksum. */
zio_cksum_t cksum;
fletcher_4_native_varsize(compressed, len, &cksum);
if (cksum.zc_word[0] != checksum) {
free(compressed);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"resume token is corrupt (incorrect checksum)"));
return (NULL);
}
/* Uncompress. */
void *packed = zfs_alloc(hdl, packed_len);
uLongf packed_len_long = packed_len;
if (uncompress(packed, &packed_len_long, compressed, len) != Z_OK ||
packed_len_long != packed_len) {
free(packed);
free(compressed);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"resume token is corrupt (decompression failed)"));
return (NULL);
}
/* Unpack nvlist. */
nvlist_t *nv;
int error = nvlist_unpack(packed, packed_len, &nv, KM_SLEEP);
free(packed);
free(compressed);
if (error != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"resume token is corrupt (nvlist_unpack failed)"));
return (NULL);
}
return (nv);
}
static enum lzc_send_flags
lzc_flags_from_sendflags(const sendflags_t *flags)
{
enum lzc_send_flags lzc_flags = 0;
if (flags->largeblock)
lzc_flags |= LZC_SEND_FLAG_LARGE_BLOCK;
if (flags->embed_data)
lzc_flags |= LZC_SEND_FLAG_EMBED_DATA;
if (flags->compress)
lzc_flags |= LZC_SEND_FLAG_COMPRESS;
if (flags->raw)
lzc_flags |= LZC_SEND_FLAG_RAW;
if (flags->saved)
lzc_flags |= LZC_SEND_FLAG_SAVED;
return (lzc_flags);
}
static int
estimate_size(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
uint64_t resumeobj, uint64_t resumeoff, uint64_t bytes,
const char *redactbook, char *errbuf, uint64_t *sizep)
{
uint64_t size;
FILE *fout = flags->dryrun ? stdout : stderr;
progress_arg_t pa = { 0 };
int err = 0;
pthread_t ptid;
if (flags->progress || flags->progressastitle) {
pa.pa_zhp = zhp;
pa.pa_fd = fd;
pa.pa_parsable = flags->parsable;
pa.pa_estimate = B_TRUE;
pa.pa_verbosity = flags->verbosity;
err = pthread_create(&ptid, NULL,
send_progress_thread, &pa);
if (err != 0) {
zfs_error_aux(zhp->zfs_hdl, "%s", strerror(errno));
return (zfs_error(zhp->zfs_hdl,
EZFS_THREADCREATEFAILED, errbuf));
}
}
err = lzc_send_space_resume_redacted(zhp->zfs_name, from,
lzc_flags_from_sendflags(flags), resumeobj, resumeoff, bytes,
redactbook, fd, &size);
*sizep = size;
if ((flags->progress || flags->progressastitle) &&
send_progress_thread_exit(zhp->zfs_hdl, ptid))
return (-1);
if (!flags->progress && !flags->parsable)
return (err);
if (err != 0) {
zfs_error_aux(zhp->zfs_hdl, "%s", strerror(err));
return (zfs_error(zhp->zfs_hdl, EZFS_BADBACKUP,
errbuf));
}
send_print_verbose(fout, zhp->zfs_name, from, size,
flags->parsable);
if (flags->parsable) {
(void) fprintf(fout, "size\t%llu\n", (longlong_t)size);
} else {
char buf[16];
zfs_nicenum(size, buf, sizeof (buf));
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
"total estimated size is %s\n"), buf);
}
return (0);
}
static boolean_t
redact_snaps_contains(const uint64_t *snaps, uint64_t num_snaps, uint64_t guid)
{
for (int i = 0; i < num_snaps; i++) {
if (snaps[i] == guid)
return (B_TRUE);
}
return (B_FALSE);
}
static boolean_t
redact_snaps_equal(const uint64_t *snaps1, uint64_t num_snaps1,
const uint64_t *snaps2, uint64_t num_snaps2)
{
if (num_snaps1 != num_snaps2)
return (B_FALSE);
for (int i = 0; i < num_snaps1; i++) {
if (!redact_snaps_contains(snaps2, num_snaps2, snaps1[i]))
return (B_FALSE);
}
return (B_TRUE);
}
static int
get_bookmarks(const char *path, nvlist_t **bmarksp)
{
nvlist_t *props = fnvlist_alloc();
int error;
fnvlist_add_boolean(props, "redact_complete");
fnvlist_add_boolean(props, zfs_prop_to_name(ZFS_PROP_REDACT_SNAPS));
error = lzc_get_bookmarks(path, props, bmarksp);
fnvlist_free(props);
return (error);
}
static nvpair_t *
find_redact_pair(nvlist_t *bmarks, const uint64_t *redact_snap_guids,
int num_redact_snaps)
{
nvpair_t *pair;
for (pair = nvlist_next_nvpair(bmarks, NULL); pair;
pair = nvlist_next_nvpair(bmarks, pair)) {
nvlist_t *bmark = fnvpair_value_nvlist(pair);
nvlist_t *vallist = fnvlist_lookup_nvlist(bmark,
zfs_prop_to_name(ZFS_PROP_REDACT_SNAPS));
uint_t len = 0;
uint64_t *bmarksnaps = fnvlist_lookup_uint64_array(vallist,
ZPROP_VALUE, &len);
if (redact_snaps_equal(redact_snap_guids,
num_redact_snaps, bmarksnaps, len)) {
break;
}
}
return (pair);
}
static boolean_t
get_redact_complete(nvpair_t *pair)
{
nvlist_t *bmark = fnvpair_value_nvlist(pair);
nvlist_t *vallist = fnvlist_lookup_nvlist(bmark, "redact_complete");
boolean_t complete = fnvlist_lookup_boolean_value(vallist,
ZPROP_VALUE);
return (complete);
}
/*
* Check that the list of redaction snapshots in the bookmark matches the send
* we're resuming, and return whether or not it's complete.
*
* Note that the caller needs to free the contents of *bookname with free() if
* this function returns successfully.
*/
static int
find_redact_book(libzfs_handle_t *hdl, const char *path,
const uint64_t *redact_snap_guids, int num_redact_snaps,
char **bookname)
{
char errbuf[ERRBUFLEN];
nvlist_t *bmarks;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot resume send"));
int error = get_bookmarks(path, &bmarks);
if (error != 0) {
if (error == ESRCH) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"nonexistent redaction bookmark provided"));
} else if (error == ENOENT) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset to be sent no longer exists"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"unknown error: %s"), strerror(error));
}
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
}
nvpair_t *pair = find_redact_pair(bmarks, redact_snap_guids,
num_redact_snaps);
if (pair == NULL) {
fnvlist_free(bmarks);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"no appropriate redaction bookmark exists"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
}
boolean_t complete = get_redact_complete(pair);
if (!complete) {
fnvlist_free(bmarks);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incomplete redaction bookmark provided"));
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
}
*bookname = strndup(nvpair_name(pair), ZFS_MAX_DATASET_NAME_LEN);
ASSERT3P(*bookname, !=, NULL);
fnvlist_free(bmarks);
return (0);
}
static enum lzc_send_flags
lzc_flags_from_resume_nvl(nvlist_t *resume_nvl)
{
enum lzc_send_flags lzc_flags = 0;
if (nvlist_exists(resume_nvl, "largeblockok"))
lzc_flags |= LZC_SEND_FLAG_LARGE_BLOCK;
if (nvlist_exists(resume_nvl, "embedok"))
lzc_flags |= LZC_SEND_FLAG_EMBED_DATA;
if (nvlist_exists(resume_nvl, "compressok"))
lzc_flags |= LZC_SEND_FLAG_COMPRESS;
if (nvlist_exists(resume_nvl, "rawok"))
lzc_flags |= LZC_SEND_FLAG_RAW;
if (nvlist_exists(resume_nvl, "savedok"))
lzc_flags |= LZC_SEND_FLAG_SAVED;
return (lzc_flags);
}
static int
zfs_send_resume_impl_cb_impl(libzfs_handle_t *hdl, sendflags_t *flags,
int outfd, nvlist_t *resume_nvl)
{
char errbuf[ERRBUFLEN];
const char *toname;
const char *fromname = NULL;
uint64_t resumeobj, resumeoff, toguid, fromguid, bytes;
zfs_handle_t *zhp;
int error = 0;
char name[ZFS_MAX_DATASET_NAME_LEN];
FILE *fout = (flags->verbosity > 0 && flags->dryrun) ? stdout : stderr;
uint64_t *redact_snap_guids = NULL;
int num_redact_snaps = 0;
char *redact_book = NULL;
uint64_t size = 0;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot resume send"));
if (flags->verbosity != 0) {
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
"resume token contents:\n"));
nvlist_print(fout, resume_nvl);
}
if (nvlist_lookup_string(resume_nvl, "toname", &toname) != 0 ||
nvlist_lookup_uint64(resume_nvl, "object", &resumeobj) != 0 ||
nvlist_lookup_uint64(resume_nvl, "offset", &resumeoff) != 0 ||
nvlist_lookup_uint64(resume_nvl, "bytes", &bytes) != 0 ||
nvlist_lookup_uint64(resume_nvl, "toguid", &toguid) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"resume token is corrupt"));
return (zfs_error(hdl, EZFS_FAULT, errbuf));
}
fromguid = 0;
(void) nvlist_lookup_uint64(resume_nvl, "fromguid", &fromguid);
if (flags->saved) {
(void) strlcpy(name, toname, sizeof (name));
} else {
error = guid_to_name(hdl, toname, toguid, B_FALSE, name);
if (error != 0) {
if (zfs_dataset_exists(hdl, toname, ZFS_TYPE_DATASET)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is no longer the same snapshot "
"used in the initial send"), toname);
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' used in the initial send no "
"longer exists"), toname);
}
return (zfs_error(hdl, EZFS_BADPATH, errbuf));
}
}
zhp = zfs_open(hdl, name, ZFS_TYPE_DATASET);
if (zhp == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"unable to access '%s'"), name);
return (zfs_error(hdl, EZFS_BADPATH, errbuf));
}
if (nvlist_lookup_uint64_array(resume_nvl, "book_redact_snaps",
&redact_snap_guids, (uint_t *)&num_redact_snaps) != 0) {
num_redact_snaps = -1;
}
if (fromguid != 0) {
if (guid_to_name_redact_snaps(hdl, toname, fromguid, B_TRUE,
redact_snap_guids, num_redact_snaps, name) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incremental source %#llx no longer exists"),
(longlong_t)fromguid);
return (zfs_error(hdl, EZFS_BADPATH, errbuf));
}
fromname = name;
}
redact_snap_guids = NULL;
if (nvlist_lookup_uint64_array(resume_nvl,
zfs_prop_to_name(ZFS_PROP_REDACT_SNAPS), &redact_snap_guids,
(uint_t *)&num_redact_snaps) == 0) {
char path[ZFS_MAX_DATASET_NAME_LEN];
(void) strlcpy(path, toname, sizeof (path));
char *at = strchr(path, '@');
ASSERT3P(at, !=, NULL);
*at = '\0';
if ((error = find_redact_book(hdl, path, redact_snap_guids,
num_redact_snaps, &redact_book)) != 0) {
return (error);
}
}
enum lzc_send_flags lzc_flags = lzc_flags_from_sendflags(flags) |
lzc_flags_from_resume_nvl(resume_nvl);
if (flags->verbosity != 0 || flags->progressastitle) {
/*
* Some of these may have come from the resume token, set them
* here for size estimate purposes.
*/
sendflags_t tmpflags = *flags;
if (lzc_flags & LZC_SEND_FLAG_LARGE_BLOCK)
tmpflags.largeblock = B_TRUE;
if (lzc_flags & LZC_SEND_FLAG_COMPRESS)
tmpflags.compress = B_TRUE;
if (lzc_flags & LZC_SEND_FLAG_EMBED_DATA)
tmpflags.embed_data = B_TRUE;
if (lzc_flags & LZC_SEND_FLAG_RAW)
tmpflags.raw = B_TRUE;
if (lzc_flags & LZC_SEND_FLAG_SAVED)
tmpflags.saved = B_TRUE;
error = estimate_size(zhp, fromname, outfd, &tmpflags,
resumeobj, resumeoff, bytes, redact_book, errbuf, &size);
}
if (!flags->dryrun) {
progress_arg_t pa = { 0 };
pthread_t tid;
/*
* If progress reporting is requested, spawn a new thread to
* poll ZFS_IOC_SEND_PROGRESS at a regular interval.
*/
if (flags->progress || flags->progressastitle) {
pa.pa_zhp = zhp;
pa.pa_fd = outfd;
pa.pa_parsable = flags->parsable;
pa.pa_estimate = B_FALSE;
pa.pa_verbosity = flags->verbosity;
pa.pa_size = size;
pa.pa_astitle = flags->progressastitle;
pa.pa_progress = flags->progress;
error = pthread_create(&tid, NULL,
send_progress_thread, &pa);
if (error != 0) {
if (redact_book != NULL)
free(redact_book);
zfs_close(zhp);
return (error);
}
}
error = lzc_send_resume_redacted(zhp->zfs_name, fromname, outfd,
lzc_flags, resumeobj, resumeoff, redact_book);
if (redact_book != NULL)
free(redact_book);
if ((flags->progressastitle || flags->progress) &&
send_progress_thread_exit(hdl, tid)) {
zfs_close(zhp);
return (-1);
}
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"warning: cannot send '%s'"), zhp->zfs_name);
zfs_close(zhp);
switch (error) {
case 0:
return (0);
case EACCES:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"source key must be loaded"));
return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf));
case ESRCH:
if (lzc_exists(zhp->zfs_name)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incremental source could not be found"));
}
return (zfs_error(hdl, EZFS_NOENT, errbuf));
case EXDEV:
case ENOENT:
case EDQUOT:
case EFBIG:
case EIO:
case ENOLINK:
case ENOSPC:
case ENOSTR:
case ENXIO:
case EPIPE:
case ERANGE:
case EFAULT:
case EROFS:
zfs_error_aux(hdl, "%s", strerror(errno));
return (zfs_error(hdl, EZFS_BADBACKUP, errbuf));
default:
return (zfs_standard_error(hdl, errno, errbuf));
}
} else {
if (redact_book != NULL)
free(redact_book);
}
zfs_close(zhp);
return (error);
}
struct zfs_send_resume_impl {
libzfs_handle_t *hdl;
sendflags_t *flags;
nvlist_t *resume_nvl;
};
static int
zfs_send_resume_impl_cb(int outfd, void *arg)
{
struct zfs_send_resume_impl *zsri = arg;
return (zfs_send_resume_impl_cb_impl(zsri->hdl, zsri->flags, outfd,
zsri->resume_nvl));
}
static int
zfs_send_resume_impl(libzfs_handle_t *hdl, sendflags_t *flags, int outfd,
nvlist_t *resume_nvl)
{
struct zfs_send_resume_impl zsri = {
.hdl = hdl,
.flags = flags,
.resume_nvl = resume_nvl,
};
return (lzc_send_wrapper(zfs_send_resume_impl_cb, outfd, &zsri));
}
int
zfs_send_resume(libzfs_handle_t *hdl, sendflags_t *flags, int outfd,
const char *resume_token)
{
int ret;
char errbuf[ERRBUFLEN];
nvlist_t *resume_nvl;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot resume send"));
resume_nvl = zfs_send_resume_token_to_nvlist(hdl, resume_token);
if (resume_nvl == NULL) {
/*
* zfs_error_aux has already been set by
* zfs_send_resume_token_to_nvlist()
*/
return (zfs_error(hdl, EZFS_FAULT, errbuf));
}
ret = zfs_send_resume_impl(hdl, flags, outfd, resume_nvl);
fnvlist_free(resume_nvl);
return (ret);
}
int
zfs_send_saved(zfs_handle_t *zhp, sendflags_t *flags, int outfd,
const char *resume_token)
{
int ret;
libzfs_handle_t *hdl = zhp->zfs_hdl;
nvlist_t *saved_nvl = NULL, *resume_nvl = NULL;
uint64_t saved_guid = 0, resume_guid = 0;
uint64_t obj = 0, off = 0, bytes = 0;
char token_buf[ZFS_MAXPROPLEN];
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"saved send failed"));
ret = zfs_prop_get(zhp, ZFS_PROP_RECEIVE_RESUME_TOKEN,
token_buf, sizeof (token_buf), NULL, NULL, 0, B_TRUE);
if (ret != 0)
goto out;
saved_nvl = zfs_send_resume_token_to_nvlist(hdl, token_buf);
if (saved_nvl == NULL) {
/*
* zfs_error_aux has already been set by
* zfs_send_resume_token_to_nvlist()
*/
ret = zfs_error(hdl, EZFS_FAULT, errbuf);
goto out;
}
/*
* If a resume token is provided we use the object and offset
* from that instead of the default, which starts from the
* beginning.
*/
if (resume_token != NULL) {
resume_nvl = zfs_send_resume_token_to_nvlist(hdl,
resume_token);
if (resume_nvl == NULL) {
ret = zfs_error(hdl, EZFS_FAULT, errbuf);
goto out;
}
if (nvlist_lookup_uint64(resume_nvl, "object", &obj) != 0 ||
nvlist_lookup_uint64(resume_nvl, "offset", &off) != 0 ||
nvlist_lookup_uint64(resume_nvl, "bytes", &bytes) != 0 ||
nvlist_lookup_uint64(resume_nvl, "toguid",
&resume_guid) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"provided resume token is corrupt"));
ret = zfs_error(hdl, EZFS_FAULT, errbuf);
goto out;
}
if (nvlist_lookup_uint64(saved_nvl, "toguid",
&saved_guid)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset's resume token is corrupt"));
ret = zfs_error(hdl, EZFS_FAULT, errbuf);
goto out;
}
if (resume_guid != saved_guid) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"provided resume token does not match dataset"));
ret = zfs_error(hdl, EZFS_BADBACKUP, errbuf);
goto out;
}
}
(void) nvlist_remove_all(saved_nvl, "object");
fnvlist_add_uint64(saved_nvl, "object", obj);
(void) nvlist_remove_all(saved_nvl, "offset");
fnvlist_add_uint64(saved_nvl, "offset", off);
(void) nvlist_remove_all(saved_nvl, "bytes");
fnvlist_add_uint64(saved_nvl, "bytes", bytes);
(void) nvlist_remove_all(saved_nvl, "toname");
fnvlist_add_string(saved_nvl, "toname", zhp->zfs_name);
ret = zfs_send_resume_impl(hdl, flags, outfd, saved_nvl);
out:
fnvlist_free(saved_nvl);
fnvlist_free(resume_nvl);
return (ret);
}
/*
* This function informs the target system that the recursive send is complete.
* The record is also expected in the case of a send -p.
*/
static int
send_conclusion_record(int fd, zio_cksum_t *zc)
{
dmu_replay_record_t drr = { 0 };
drr.drr_type = DRR_END;
if (zc != NULL)
drr.drr_u.drr_end.drr_checksum = *zc;
if (write(fd, &drr, sizeof (drr)) == -1) {
return (errno);
}
return (0);
}
/*
* This function is responsible for sending the records that contain the
* necessary information for the target system's libzfs to be able to set the
* properties of the filesystem being received, or to be able to prepare for
* a recursive receive.
*
* The "zhp" argument is the handle of the snapshot we are sending
* (the "tosnap"). The "from" argument is the short snapshot name (the part
* after the @) of the incremental source.
*/
static int
send_prelim_records(zfs_handle_t *zhp, const char *from, int fd,
boolean_t gather_props, boolean_t recursive, boolean_t verbose,
boolean_t dryrun, boolean_t raw, boolean_t replicate, boolean_t skipmissing,
boolean_t backup, boolean_t holds, boolean_t props, boolean_t doall,
nvlist_t **fssp, avl_tree_t **fsavlp)
{
int err = 0;
char *packbuf = NULL;
size_t buflen = 0;
zio_cksum_t zc = { {0} };
int featureflags = 0;
/* name of filesystem/volume that contains snapshot we are sending */
char tofs[ZFS_MAX_DATASET_NAME_LEN];
/* short name of snap we are sending */
const char *tosnap = "";
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"warning: cannot send '%s'"), zhp->zfs_name);
if (zhp->zfs_type == ZFS_TYPE_FILESYSTEM && zfs_prop_get_int(zhp,
ZFS_PROP_VERSION) >= ZPL_VERSION_SA) {
featureflags |= DMU_BACKUP_FEATURE_SA_SPILL;
}
if (holds)
featureflags |= DMU_BACKUP_FEATURE_HOLDS;
(void) strlcpy(tofs, zhp->zfs_name, ZFS_MAX_DATASET_NAME_LEN);
char *at = strchr(tofs, '@');
if (at != NULL) {
*at = '\0';
tosnap = at + 1;
}
if (gather_props) {
nvlist_t *hdrnv = fnvlist_alloc();
nvlist_t *fss = NULL;
if (from != NULL)
fnvlist_add_string(hdrnv, "fromsnap", from);
fnvlist_add_string(hdrnv, "tosnap", tosnap);
if (!recursive)
fnvlist_add_boolean(hdrnv, "not_recursive");
if (raw) {
fnvlist_add_boolean(hdrnv, "raw");
}
if (gather_nvlist(zhp->zfs_hdl, tofs,
from, tosnap, recursive, raw, doall, replicate, skipmissing,
verbose, backup, holds, props, &fss, fsavlp) != 0) {
return (zfs_error(zhp->zfs_hdl, EZFS_BADBACKUP,
errbuf));
}
/*
* Do not allow the size of the properties list to exceed
* the limit
*/
if ((fnvlist_size(fss) + fnvlist_size(hdrnv)) >
zhp->zfs_hdl->libzfs_max_nvlist) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN, "warning: cannot send '%s': "
"the size of the list of snapshots and properties "
"is too large to be received successfully.\n"
"Select a smaller number of snapshots to send.\n"),
zhp->zfs_name);
return (zfs_error(zhp->zfs_hdl, EZFS_NOSPC,
errbuf));
}
fnvlist_add_nvlist(hdrnv, "fss", fss);
VERIFY0(nvlist_pack(hdrnv, &packbuf, &buflen, NV_ENCODE_XDR,
0));
if (fssp != NULL) {
*fssp = fss;
} else {
fnvlist_free(fss);
}
fnvlist_free(hdrnv);
}
if (!dryrun) {
dmu_replay_record_t drr = { 0 };
/* write first begin record */
drr.drr_type = DRR_BEGIN;
drr.drr_u.drr_begin.drr_magic = DMU_BACKUP_MAGIC;
DMU_SET_STREAM_HDRTYPE(drr.drr_u.drr_begin.
drr_versioninfo, DMU_COMPOUNDSTREAM);
DMU_SET_FEATUREFLAGS(drr.drr_u.drr_begin.
drr_versioninfo, featureflags);
if (snprintf(drr.drr_u.drr_begin.drr_toname,
sizeof (drr.drr_u.drr_begin.drr_toname), "%s@%s", tofs,
tosnap) >= sizeof (drr.drr_u.drr_begin.drr_toname)) {
return (zfs_error(zhp->zfs_hdl, EZFS_BADBACKUP,
errbuf));
}
drr.drr_payloadlen = buflen;
err = dump_record(&drr, packbuf, buflen, &zc, fd);
free(packbuf);
if (err != 0) {
zfs_error_aux(zhp->zfs_hdl, "%s", strerror(err));
return (zfs_error(zhp->zfs_hdl, EZFS_BADBACKUP,
errbuf));
}
err = send_conclusion_record(fd, &zc);
if (err != 0) {
zfs_error_aux(zhp->zfs_hdl, "%s", strerror(err));
return (zfs_error(zhp->zfs_hdl, EZFS_BADBACKUP,
errbuf));
}
}
return (0);
}
/*
* Generate a send stream. The "zhp" argument is the filesystem/volume
* that contains the snapshot to send. The "fromsnap" argument is the
* short name (the part after the '@') of the snapshot that is the
* incremental source to send from (if non-NULL). The "tosnap" argument
* is the short name of the snapshot to send.
*
* The content of the send stream is the snapshot identified by
* 'tosnap'. Incremental streams are requested in two ways:
* - from the snapshot identified by "fromsnap" (if non-null) or
* - from the origin of the dataset identified by zhp, which must
* be a clone. In this case, "fromsnap" is null and "fromorigin"
* is TRUE.
*
* The send stream is recursive (i.e. dumps a hierarchy of snapshots) and
* uses a special header (with a hdrtype field of DMU_COMPOUNDSTREAM)
* if "replicate" is set. If "doall" is set, dump all the intermediate
* snapshots. The DMU_COMPOUNDSTREAM header is used in the "doall"
* case too. If "props" is set, send properties.
*
* Pre-wrapped (cf. lzc_send_wrapper()).
*/
static int
zfs_send_cb_impl(zfs_handle_t *zhp, const char *fromsnap, const char *tosnap,
sendflags_t *flags, int outfd, snapfilter_cb_t filter_func,
void *cb_arg, nvlist_t **debugnvp)
{
char errbuf[ERRBUFLEN];
send_dump_data_t sdd = { 0 };
int err = 0;
nvlist_t *fss = NULL;
avl_tree_t *fsavl = NULL;
static uint64_t holdseq;
int spa_version;
FILE *fout;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot send '%s'"), zhp->zfs_name);
if (fromsnap && fromsnap[0] == '\0') {
zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN,
"zero-length incremental source"));
return (zfs_error(zhp->zfs_hdl, EZFS_NOENT, errbuf));
}
if (fromsnap) {
char full_fromsnap_name[ZFS_MAX_DATASET_NAME_LEN];
if (snprintf(full_fromsnap_name, sizeof (full_fromsnap_name),
"%s@%s", zhp->zfs_name, fromsnap) >=
sizeof (full_fromsnap_name)) {
err = EINVAL;
goto stderr_out;
}
zfs_handle_t *fromsnapn = zfs_open(zhp->zfs_hdl,
full_fromsnap_name, ZFS_TYPE_SNAPSHOT);
if (fromsnapn == NULL) {
err = -1;
goto err_out;
}
zfs_close(fromsnapn);
}
if (flags->replicate || flags->doall || flags->props ||
flags->holds || flags->backup) {
char full_tosnap_name[ZFS_MAX_DATASET_NAME_LEN];
if (snprintf(full_tosnap_name, sizeof (full_tosnap_name),
"%s@%s", zhp->zfs_name, tosnap) >=
sizeof (full_tosnap_name)) {
err = EINVAL;
goto stderr_out;
}
zfs_handle_t *tosnap = zfs_open(zhp->zfs_hdl,
full_tosnap_name, ZFS_TYPE_SNAPSHOT);
if (tosnap == NULL) {
err = -1;
goto err_out;
}
err = send_prelim_records(tosnap, fromsnap, outfd,
flags->replicate || flags->props || flags->holds,
flags->replicate, flags->verbosity > 0, flags->dryrun,
flags->raw, flags->replicate, flags->skipmissing,
flags->backup, flags->holds, flags->props, flags->doall,
&fss, &fsavl);
zfs_close(tosnap);
if (err != 0)
goto err_out;
}
/* dump each stream */
sdd.fromsnap = fromsnap;
sdd.tosnap = tosnap;
sdd.outfd = outfd;
sdd.replicate = flags->replicate;
sdd.doall = flags->doall;
sdd.fromorigin = flags->fromorigin;
sdd.fss = fss;
sdd.fsavl = fsavl;
sdd.verbosity = flags->verbosity;
sdd.parsable = flags->parsable;
sdd.progress = flags->progress;
sdd.progressastitle = flags->progressastitle;
sdd.dryrun = flags->dryrun;
sdd.large_block = flags->largeblock;
sdd.embed_data = flags->embed_data;
sdd.compress = flags->compress;
sdd.raw = flags->raw;
sdd.holds = flags->holds;
sdd.filter_cb = filter_func;
sdd.filter_cb_arg = cb_arg;
if (debugnvp)
sdd.debugnv = *debugnvp;
if (sdd.verbosity != 0 && sdd.dryrun)
sdd.std_out = B_TRUE;
fout = sdd.std_out ? stdout : stderr;
/*
* Some flags require that we place user holds on the datasets that are
* being sent so they don't get destroyed during the send. We can skip
* this step if the pool is imported read-only since the datasets cannot
* be destroyed.
*/
if (!flags->dryrun && !zpool_get_prop_int(zfs_get_pool_handle(zhp),
ZPOOL_PROP_READONLY, NULL) &&
zfs_spa_version(zhp, &spa_version) == 0 &&
spa_version >= SPA_VERSION_USERREFS &&
(flags->doall || flags->replicate)) {
++holdseq;
(void) snprintf(sdd.holdtag, sizeof (sdd.holdtag),
".send-%d-%llu", getpid(), (u_longlong_t)holdseq);
sdd.cleanup_fd = open(ZFS_DEV, O_RDWR | O_CLOEXEC);
if (sdd.cleanup_fd < 0) {
err = errno;
goto stderr_out;
}
sdd.snapholds = fnvlist_alloc();
} else {
sdd.cleanup_fd = -1;
sdd.snapholds = NULL;
}
if (flags->verbosity != 0 || sdd.snapholds != NULL) {
/*
* Do a verbose no-op dry run to get all the verbose output
* or to gather snapshot hold's before generating any data,
* then do a non-verbose real run to generate the streams.
*/
sdd.dryrun = B_TRUE;
err = dump_filesystems(zhp, &sdd);
if (err != 0)
goto stderr_out;
if (flags->verbosity != 0) {
if (flags->parsable) {
(void) fprintf(fout, "size\t%llu\n",
(longlong_t)sdd.size);
} else {
char buf[16];
zfs_nicebytes(sdd.size, buf, sizeof (buf));
(void) fprintf(fout, dgettext(TEXT_DOMAIN,
"total estimated size is %s\n"), buf);
}
}
/* Ensure no snaps found is treated as an error. */
if (!sdd.seento) {
err = ENOENT;
goto err_out;
}
/* Skip the second run if dryrun was requested. */
if (flags->dryrun)
goto err_out;
if (sdd.snapholds != NULL) {
err = zfs_hold_nvl(zhp, sdd.cleanup_fd, sdd.snapholds);
if (err != 0)
goto stderr_out;
fnvlist_free(sdd.snapholds);
sdd.snapholds = NULL;
}
sdd.dryrun = B_FALSE;
sdd.verbosity = 0;
}
err = dump_filesystems(zhp, &sdd);
fsavl_destroy(fsavl);
fnvlist_free(fss);
/* Ensure no snaps found is treated as an error. */
if (err == 0 && !sdd.seento)
err = ENOENT;
if (sdd.cleanup_fd != -1) {
VERIFY(0 == close(sdd.cleanup_fd));
sdd.cleanup_fd = -1;
}
if (!flags->dryrun && (flags->replicate || flags->doall ||
flags->props || flags->backup || flags->holds)) {
/*
* write final end record. NB: want to do this even if
* there was some error, because it might not be totally
* failed.
*/
int err2 = send_conclusion_record(outfd, NULL);
if (err2 != 0)
return (zfs_standard_error(zhp->zfs_hdl, err2, errbuf));
}
return (err || sdd.err);
stderr_out:
err = zfs_standard_error(zhp->zfs_hdl, err, errbuf);
err_out:
fsavl_destroy(fsavl);
fnvlist_free(fss);
fnvlist_free(sdd.snapholds);
if (sdd.cleanup_fd != -1)
VERIFY(0 == close(sdd.cleanup_fd));
return (err);
}
struct zfs_send {
zfs_handle_t *zhp;
const char *fromsnap;
const char *tosnap;
sendflags_t *flags;
snapfilter_cb_t *filter_func;
void *cb_arg;
nvlist_t **debugnvp;
};
static int
zfs_send_cb(int outfd, void *arg)
{
struct zfs_send *zs = arg;
return (zfs_send_cb_impl(zs->zhp, zs->fromsnap, zs->tosnap, zs->flags,
outfd, zs->filter_func, zs->cb_arg, zs->debugnvp));
}
int
zfs_send(zfs_handle_t *zhp, const char *fromsnap, const char *tosnap,
sendflags_t *flags, int outfd, snapfilter_cb_t filter_func,
void *cb_arg, nvlist_t **debugnvp)
{
struct zfs_send arg = {
.zhp = zhp,
.fromsnap = fromsnap,
.tosnap = tosnap,
.flags = flags,
.filter_func = filter_func,
.cb_arg = cb_arg,
.debugnvp = debugnvp,
};
return (lzc_send_wrapper(zfs_send_cb, outfd, &arg));
}
static zfs_handle_t *
name_to_dir_handle(libzfs_handle_t *hdl, const char *snapname)
{
char dirname[ZFS_MAX_DATASET_NAME_LEN];
(void) strlcpy(dirname, snapname, ZFS_MAX_DATASET_NAME_LEN);
char *c = strchr(dirname, '@');
if (c != NULL)
*c = '\0';
return (zfs_open(hdl, dirname, ZFS_TYPE_DATASET));
}
/*
* Returns B_TRUE if earlier is an earlier snapshot in later's timeline; either
* an earlier snapshot in the same filesystem, or a snapshot before later's
* origin, or it's origin's origin, etc.
*/
static boolean_t
snapshot_is_before(zfs_handle_t *earlier, zfs_handle_t *later)
{
boolean_t ret;
uint64_t later_txg =
(later->zfs_type == ZFS_TYPE_FILESYSTEM ||
later->zfs_type == ZFS_TYPE_VOLUME ?
UINT64_MAX : zfs_prop_get_int(later, ZFS_PROP_CREATETXG));
uint64_t earlier_txg = zfs_prop_get_int(earlier, ZFS_PROP_CREATETXG);
if (earlier_txg >= later_txg)
return (B_FALSE);
zfs_handle_t *earlier_dir = name_to_dir_handle(earlier->zfs_hdl,
earlier->zfs_name);
zfs_handle_t *later_dir = name_to_dir_handle(later->zfs_hdl,
later->zfs_name);
if (strcmp(earlier_dir->zfs_name, later_dir->zfs_name) == 0) {
zfs_close(earlier_dir);
zfs_close(later_dir);
return (B_TRUE);
}
char clonename[ZFS_MAX_DATASET_NAME_LEN];
if (zfs_prop_get(later_dir, ZFS_PROP_ORIGIN, clonename,
ZFS_MAX_DATASET_NAME_LEN, NULL, NULL, 0, B_TRUE) != 0) {
zfs_close(earlier_dir);
zfs_close(later_dir);
return (B_FALSE);
}
zfs_handle_t *origin = zfs_open(earlier->zfs_hdl, clonename,
ZFS_TYPE_DATASET);
uint64_t origin_txg = zfs_prop_get_int(origin, ZFS_PROP_CREATETXG);
/*
* If "earlier" is exactly the origin, then
* snapshot_is_before(earlier, origin) will return false (because
* they're the same).
*/
if (origin_txg == earlier_txg &&
strcmp(origin->zfs_name, earlier->zfs_name) == 0) {
zfs_close(earlier_dir);
zfs_close(later_dir);
zfs_close(origin);
return (B_TRUE);
}
zfs_close(earlier_dir);
zfs_close(later_dir);
ret = snapshot_is_before(earlier, origin);
zfs_close(origin);
return (ret);
}
/*
* The "zhp" argument is the handle of the dataset to send (typically a
* snapshot). The "from" argument is the full name of the snapshot or
* bookmark that is the incremental source.
*
* Pre-wrapped (cf. lzc_send_wrapper()).
*/
static int
zfs_send_one_cb_impl(zfs_handle_t *zhp, const char *from, int fd,
sendflags_t *flags, const char *redactbook)
{
int err;
libzfs_handle_t *hdl = zhp->zfs_hdl;
char *name = zhp->zfs_name;
pthread_t ptid;
progress_arg_t pa = { 0 };
uint64_t size = 0;
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"warning: cannot send '%s'"), name);
if (from != NULL && strchr(from, '@')) {
zfs_handle_t *from_zhp = zfs_open(hdl, from,
ZFS_TYPE_DATASET);
if (from_zhp == NULL)
return (-1);
if (!snapshot_is_before(from_zhp, zhp)) {
zfs_close(from_zhp);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"not an earlier snapshot from the same fs"));
return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf));
}
zfs_close(from_zhp);
}
if (redactbook != NULL) {
char bookname[ZFS_MAX_DATASET_NAME_LEN];
nvlist_t *redact_snaps;
zfs_handle_t *book_zhp;
char *at, *pound;
int dsnamelen;
pound = strchr(redactbook, '#');
if (pound != NULL)
redactbook = pound + 1;
at = strchr(name, '@');
if (at == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot do a redacted send to a filesystem"));
return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
}
dsnamelen = at - name;
if (snprintf(bookname, sizeof (bookname), "%.*s#%s",
dsnamelen, name, redactbook)
>= sizeof (bookname)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid bookmark name"));
return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf));
}
book_zhp = zfs_open(hdl, bookname, ZFS_TYPE_BOOKMARK);
if (book_zhp == NULL)
return (-1);
if (nvlist_lookup_nvlist(book_zhp->zfs_props,
zfs_prop_to_name(ZFS_PROP_REDACT_SNAPS),
&redact_snaps) != 0 || redact_snaps == NULL) {
zfs_close(book_zhp);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"not a redaction bookmark"));
return (zfs_error(hdl, EZFS_BADTYPE, errbuf));
}
zfs_close(book_zhp);
}
/*
* Send fs properties
*/
if (flags->props || flags->holds || flags->backup) {
/*
* Note: the header generated by send_prelim_records()
* assumes that the incremental source is in the same
* filesystem/volume as the target (which is a requirement
* when doing "zfs send -R"). But that isn't always the
* case here (e.g. send from snap in origin, or send from
* bookmark). We pass from=NULL, which will omit this
* information from the prelim records; it isn't used
* when receiving this type of stream.
*/
err = send_prelim_records(zhp, NULL, fd, B_TRUE, B_FALSE,
flags->verbosity > 0, flags->dryrun, flags->raw,
flags->replicate, B_FALSE, flags->backup, flags->holds,
flags->props, flags->doall, NULL, NULL);
if (err != 0)
return (err);
}
/*
* Perform size estimate if verbose was specified.
*/
if (flags->verbosity != 0 || flags->progressastitle) {
err = estimate_size(zhp, from, fd, flags, 0, 0, 0, redactbook,
errbuf, &size);
if (err != 0)
return (err);
}
if (flags->dryrun)
return (0);
/*
* If progress reporting is requested, spawn a new thread to poll
* ZFS_IOC_SEND_PROGRESS at a regular interval.
*/
if (flags->progress || flags->progressastitle) {
pa.pa_zhp = zhp;
pa.pa_fd = fd;
pa.pa_parsable = flags->parsable;
pa.pa_estimate = B_FALSE;
pa.pa_verbosity = flags->verbosity;
pa.pa_size = size;
pa.pa_astitle = flags->progressastitle;
pa.pa_progress = flags->progress;
err = pthread_create(&ptid, NULL,
send_progress_thread, &pa);
if (err != 0) {
zfs_error_aux(zhp->zfs_hdl, "%s", strerror(errno));
return (zfs_error(zhp->zfs_hdl,
EZFS_THREADCREATEFAILED, errbuf));
}
}
err = lzc_send_redacted(name, from, fd,
lzc_flags_from_sendflags(flags), redactbook);
if ((flags->progress || flags->progressastitle) &&
send_progress_thread_exit(hdl, ptid))
return (-1);
if (err == 0 && (flags->props || flags->holds || flags->backup)) {
/* Write the final end record. */
err = send_conclusion_record(fd, NULL);
if (err != 0)
return (zfs_standard_error(hdl, err, errbuf));
}
if (err != 0) {
switch (errno) {
case EXDEV:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"not an earlier snapshot from the same fs"));
return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf));
case ENOENT:
case ESRCH:
if (lzc_exists(name)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incremental source (%s) does not exist"),
from);
}
return (zfs_error(hdl, EZFS_NOENT, errbuf));
case EACCES:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset key must be loaded"));
return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf));
case EBUSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"target is busy; if a filesystem, "
"it must not be mounted"));
return (zfs_error(hdl, EZFS_BUSY, errbuf));
case EDQUOT:
case EFAULT:
case EFBIG:
case EINVAL:
case EIO:
case ENOLINK:
case ENOSPC:
case ENOSTR:
case ENXIO:
case EPIPE:
case ERANGE:
case EROFS:
zfs_error_aux(hdl, "%s", strerror(errno));
return (zfs_error(hdl, EZFS_BADBACKUP, errbuf));
default:
return (zfs_standard_error(hdl, errno, errbuf));
}
}
return (err != 0);
}
struct zfs_send_one {
zfs_handle_t *zhp;
const char *from;
sendflags_t *flags;
const char *redactbook;
};
static int
zfs_send_one_cb(int fd, void *arg)
{
struct zfs_send_one *zso = arg;
return (zfs_send_one_cb_impl(zso->zhp, zso->from, fd, zso->flags,
zso->redactbook));
}
int
zfs_send_one(zfs_handle_t *zhp, const char *from, int fd, sendflags_t *flags,
const char *redactbook)
{
struct zfs_send_one zso = {
.zhp = zhp,
.from = from,
.flags = flags,
.redactbook = redactbook,
};
return (lzc_send_wrapper(zfs_send_one_cb, fd, &zso));
}
/*
* Routines specific to "zfs recv"
*/
static int
recv_read(libzfs_handle_t *hdl, int fd, void *buf, int ilen,
boolean_t byteswap, zio_cksum_t *zc)
{
char *cp = buf;
int rv;
int len = ilen;
do {
rv = read(fd, cp, len);
cp += rv;
len -= rv;
} while (rv > 0);
if (rv < 0 || len != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"failed to read from stream"));
return (zfs_error(hdl, EZFS_BADSTREAM, dgettext(TEXT_DOMAIN,
"cannot receive")));
}
if (zc) {
if (byteswap)
fletcher_4_incremental_byteswap(buf, ilen, zc);
else
fletcher_4_incremental_native(buf, ilen, zc);
}
return (0);
}
static int
recv_read_nvlist(libzfs_handle_t *hdl, int fd, int len, nvlist_t **nvp,
boolean_t byteswap, zio_cksum_t *zc)
{
char *buf;
int err;
buf = zfs_alloc(hdl, len);
if (len > hdl->libzfs_max_nvlist) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "nvlist too large"));
free(buf);
return (ENOMEM);
}
err = recv_read(hdl, fd, buf, len, byteswap, zc);
if (err != 0) {
free(buf);
return (err);
}
err = nvlist_unpack(buf, len, nvp, 0);
free(buf);
if (err != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
"stream (malformed nvlist)"));
return (EINVAL);
}
return (0);
}
/*
* Returns the grand origin (origin of origin of origin...) of a given handle.
* If this dataset is not a clone, it simply returns a copy of the original
* handle.
*/
static zfs_handle_t *
recv_open_grand_origin(zfs_handle_t *zhp)
{
char origin[ZFS_MAX_DATASET_NAME_LEN];
zprop_source_t src;
zfs_handle_t *ozhp = zfs_handle_dup(zhp);
while (ozhp != NULL) {
if (zfs_prop_get(ozhp, ZFS_PROP_ORIGIN, origin,
sizeof (origin), &src, NULL, 0, B_FALSE) != 0)
break;
(void) zfs_close(ozhp);
ozhp = zfs_open(zhp->zfs_hdl, origin, ZFS_TYPE_FILESYSTEM);
}
return (ozhp);
}
static int
recv_rename_impl(zfs_handle_t *zhp, const char *name, const char *newname)
{
int err;
zfs_handle_t *ozhp = NULL;
/*
* Attempt to rename the dataset. If it fails with EACCES we have
* attempted to rename the dataset outside of its encryption root.
* Force the dataset to become an encryption root and try again.
*/
err = lzc_rename(name, newname);
if (err == EACCES) {
ozhp = recv_open_grand_origin(zhp);
if (ozhp == NULL) {
err = ENOENT;
goto out;
}
err = lzc_change_key(ozhp->zfs_name, DCP_CMD_FORCE_NEW_KEY,
NULL, NULL, 0);
if (err != 0)
goto out;
err = lzc_rename(name, newname);
}
out:
if (ozhp != NULL)
zfs_close(ozhp);
return (err);
}
static int
recv_rename(libzfs_handle_t *hdl, const char *name, const char *tryname,
int baselen, char *newname, recvflags_t *flags)
{
static int seq;
int err;
prop_changelist_t *clp = NULL;
zfs_handle_t *zhp = NULL;
zhp = zfs_open(hdl, name, ZFS_TYPE_DATASET);
if (zhp == NULL) {
err = -1;
goto out;
}
clp = changelist_gather(zhp, ZFS_PROP_NAME, 0,
flags->force ? MS_FORCE : 0);
if (clp == NULL) {
err = -1;
goto out;
}
err = changelist_prefix(clp);
if (err)
goto out;
if (tryname) {
(void) strlcpy(newname, tryname, ZFS_MAX_DATASET_NAME_LEN);
if (flags->verbose) {
(void) printf("attempting rename %s to %s\n",
name, newname);
}
err = recv_rename_impl(zhp, name, newname);
if (err == 0)
changelist_rename(clp, name, tryname);
} else {
err = ENOENT;
}
if (err != 0 && strncmp(name + baselen, "recv-", 5) != 0) {
seq++;
(void) snprintf(newname, ZFS_MAX_DATASET_NAME_LEN,
"%.*srecv-%u-%u", baselen, name, getpid(), seq);
if (flags->verbose) {
(void) printf("failed - trying rename %s to %s\n",
name, newname);
}
err = recv_rename_impl(zhp, name, newname);
if (err == 0)
changelist_rename(clp, name, newname);
if (err && flags->verbose) {
(void) printf("failed (%u) - "
"will try again on next pass\n", errno);
}
err = EAGAIN;
} else if (flags->verbose) {
if (err == 0)
(void) printf("success\n");
else
(void) printf("failed (%u)\n", errno);
}
(void) changelist_postfix(clp);
out:
if (clp != NULL)
changelist_free(clp);
if (zhp != NULL)
zfs_close(zhp);
return (err);
}
static int
recv_promote(libzfs_handle_t *hdl, const char *fsname,
const char *origin_fsname, recvflags_t *flags)
{
int err;
zfs_cmd_t zc = {"\0"};
zfs_handle_t *zhp = NULL, *ozhp = NULL;
if (flags->verbose)
(void) printf("promoting %s\n", fsname);
(void) strlcpy(zc.zc_value, origin_fsname, sizeof (zc.zc_value));
(void) strlcpy(zc.zc_name, fsname, sizeof (zc.zc_name));
/*
* Attempt to promote the dataset. If it fails with EACCES the
* promotion would cause this dataset to leave its encryption root.
* Force the origin to become an encryption root and try again.
*/
err = zfs_ioctl(hdl, ZFS_IOC_PROMOTE, &zc);
if (err == EACCES) {
zhp = zfs_open(hdl, fsname, ZFS_TYPE_DATASET);
if (zhp == NULL) {
err = -1;
goto out;
}
ozhp = recv_open_grand_origin(zhp);
if (ozhp == NULL) {
err = -1;
goto out;
}
err = lzc_change_key(ozhp->zfs_name, DCP_CMD_FORCE_NEW_KEY,
NULL, NULL, 0);
if (err != 0)
goto out;
err = zfs_ioctl(hdl, ZFS_IOC_PROMOTE, &zc);
}
out:
if (zhp != NULL)
zfs_close(zhp);
if (ozhp != NULL)
zfs_close(ozhp);
return (err);
}
static int
recv_destroy(libzfs_handle_t *hdl, const char *name, int baselen,
char *newname, recvflags_t *flags)
{
int err = 0;
prop_changelist_t *clp;
zfs_handle_t *zhp;
boolean_t defer = B_FALSE;
int spa_version;
zhp = zfs_open(hdl, name, ZFS_TYPE_DATASET);
if (zhp == NULL)
return (-1);
zfs_type_t type = zfs_get_type(zhp);
if (type == ZFS_TYPE_SNAPSHOT &&
zfs_spa_version(zhp, &spa_version) == 0 &&
spa_version >= SPA_VERSION_USERREFS)
defer = B_TRUE;
clp = changelist_gather(zhp, ZFS_PROP_NAME, 0,
flags->force ? MS_FORCE : 0);
zfs_close(zhp);
if (clp == NULL)
return (-1);
err = changelist_prefix(clp);
if (err)
return (err);
if (flags->verbose)
(void) printf("attempting destroy %s\n", name);
if (type == ZFS_TYPE_SNAPSHOT) {
nvlist_t *nv = fnvlist_alloc();
fnvlist_add_boolean(nv, name);
err = lzc_destroy_snaps(nv, defer, NULL);
fnvlist_free(nv);
} else {
err = lzc_destroy(name);
}
if (err == 0) {
if (flags->verbose)
(void) printf("success\n");
changelist_remove(clp, name);
}
(void) changelist_postfix(clp);
changelist_free(clp);
/*
* Deferred destroy might destroy the snapshot or only mark it to be
* destroyed later, and it returns success in either case.
*/
if (err != 0 || (defer && zfs_dataset_exists(hdl, name,
ZFS_TYPE_SNAPSHOT))) {
err = recv_rename(hdl, name, NULL, baselen, newname, flags);
}
return (err);
}
typedef struct guid_to_name_data {
uint64_t guid;
boolean_t bookmark_ok;
char *name;
char *skip;
uint64_t *redact_snap_guids;
uint64_t num_redact_snaps;
} guid_to_name_data_t;
static boolean_t
redact_snaps_match(zfs_handle_t *zhp, guid_to_name_data_t *gtnd)
{
uint64_t *bmark_snaps;
uint_t bmark_num_snaps;
nvlist_t *nvl;
if (zhp->zfs_type != ZFS_TYPE_BOOKMARK)
return (B_FALSE);
nvl = fnvlist_lookup_nvlist(zhp->zfs_props,
zfs_prop_to_name(ZFS_PROP_REDACT_SNAPS));
bmark_snaps = fnvlist_lookup_uint64_array(nvl, ZPROP_VALUE,
&bmark_num_snaps);
if (bmark_num_snaps != gtnd->num_redact_snaps)
return (B_FALSE);
int i = 0;
for (; i < bmark_num_snaps; i++) {
int j = 0;
for (; j < bmark_num_snaps; j++) {
if (bmark_snaps[i] == gtnd->redact_snap_guids[j])
break;
}
if (j == bmark_num_snaps)
break;
}
return (i == bmark_num_snaps);
}
static int
guid_to_name_cb(zfs_handle_t *zhp, void *arg)
{
guid_to_name_data_t *gtnd = arg;
const char *slash;
int err;
if (gtnd->skip != NULL &&
(slash = strrchr(zhp->zfs_name, '/')) != NULL &&
strcmp(slash + 1, gtnd->skip) == 0) {
zfs_close(zhp);
return (0);
}
if (zfs_prop_get_int(zhp, ZFS_PROP_GUID) == gtnd->guid &&
(gtnd->num_redact_snaps == -1 || redact_snaps_match(zhp, gtnd))) {
(void) strcpy(gtnd->name, zhp->zfs_name);
zfs_close(zhp);
return (EEXIST);
}
- err = zfs_iter_children(zhp, 0, guid_to_name_cb, gtnd);
+ err = zfs_iter_children_v2(zhp, 0, guid_to_name_cb, gtnd);
if (err != EEXIST && gtnd->bookmark_ok)
- err = zfs_iter_bookmarks(zhp, 0, guid_to_name_cb, gtnd);
+ err = zfs_iter_bookmarks_v2(zhp, 0, guid_to_name_cb, gtnd);
zfs_close(zhp);
return (err);
}
/*
* Attempt to find the local dataset associated with this guid. In the case of
* multiple matches, we attempt to find the "best" match by searching
* progressively larger portions of the hierarchy. This allows one to send a
* tree of datasets individually and guarantee that we will find the source
* guid within that hierarchy, even if there are multiple matches elsewhere.
*
* If num_redact_snaps is not -1, we attempt to find a redaction bookmark with
* the specified number of redaction snapshots. If num_redact_snaps isn't 0 or
* -1, then redact_snap_guids will be an array of the guids of the snapshots the
* redaction bookmark was created with. If num_redact_snaps is -1, then we will
* attempt to find a snapshot or bookmark (if bookmark_ok is passed) with the
* given guid. Note that a redaction bookmark can be returned if
* num_redact_snaps == -1.
*/
static int
guid_to_name_redact_snaps(libzfs_handle_t *hdl, const char *parent,
uint64_t guid, boolean_t bookmark_ok, uint64_t *redact_snap_guids,
uint64_t num_redact_snaps, char *name)
{
char pname[ZFS_MAX_DATASET_NAME_LEN];
guid_to_name_data_t gtnd;
gtnd.guid = guid;
gtnd.bookmark_ok = bookmark_ok;
gtnd.name = name;
gtnd.skip = NULL;
gtnd.redact_snap_guids = redact_snap_guids;
gtnd.num_redact_snaps = num_redact_snaps;
/*
* Search progressively larger portions of the hierarchy, starting
* with the filesystem specified by 'parent'. This will
* select the "most local" version of the origin snapshot in the case
* that there are multiple matching snapshots in the system.
*/
(void) strlcpy(pname, parent, sizeof (pname));
char *cp = strrchr(pname, '@');
if (cp == NULL)
cp = strchr(pname, '\0');
for (; cp != NULL; cp = strrchr(pname, '/')) {
/* Chop off the last component and open the parent */
*cp = '\0';
zfs_handle_t *zhp = make_dataset_handle(hdl, pname);
if (zhp == NULL)
continue;
int err = guid_to_name_cb(zfs_handle_dup(zhp), &gtnd);
if (err != EEXIST)
- err = zfs_iter_children(zhp, 0, guid_to_name_cb, &gtnd);
+ err = zfs_iter_children_v2(zhp, 0, guid_to_name_cb,
+ &gtnd);
if (err != EEXIST && bookmark_ok)
- err = zfs_iter_bookmarks(zhp, 0, guid_to_name_cb,
+ err = zfs_iter_bookmarks_v2(zhp, 0, guid_to_name_cb,
&gtnd);
zfs_close(zhp);
if (err == EEXIST)
return (0);
/*
* Remember the last portion of the dataset so we skip it next
* time through (as we've already searched that portion of the
* hierarchy).
*/
gtnd.skip = strrchr(pname, '/') + 1;
}
return (ENOENT);
}
static int
guid_to_name(libzfs_handle_t *hdl, const char *parent, uint64_t guid,
boolean_t bookmark_ok, char *name)
{
return (guid_to_name_redact_snaps(hdl, parent, guid, bookmark_ok, NULL,
-1, name));
}
/*
* Return +1 if guid1 is before guid2, 0 if they are the same, and -1 if
* guid1 is after guid2.
*/
static int
created_before(libzfs_handle_t *hdl, avl_tree_t *avl,
uint64_t guid1, uint64_t guid2)
{
nvlist_t *nvfs;
const char *fsname = NULL, *snapname = NULL;
char buf[ZFS_MAX_DATASET_NAME_LEN];
int rv;
zfs_handle_t *guid1hdl, *guid2hdl;
uint64_t create1, create2;
if (guid2 == 0)
return (0);
if (guid1 == 0)
return (1);
nvfs = fsavl_find(avl, guid1, &snapname);
fsname = fnvlist_lookup_string(nvfs, "name");
(void) snprintf(buf, sizeof (buf), "%s@%s", fsname, snapname);
guid1hdl = zfs_open(hdl, buf, ZFS_TYPE_SNAPSHOT);
if (guid1hdl == NULL)
return (-1);
nvfs = fsavl_find(avl, guid2, &snapname);
fsname = fnvlist_lookup_string(nvfs, "name");
(void) snprintf(buf, sizeof (buf), "%s@%s", fsname, snapname);
guid2hdl = zfs_open(hdl, buf, ZFS_TYPE_SNAPSHOT);
if (guid2hdl == NULL) {
zfs_close(guid1hdl);
return (-1);
}
create1 = zfs_prop_get_int(guid1hdl, ZFS_PROP_CREATETXG);
create2 = zfs_prop_get_int(guid2hdl, ZFS_PROP_CREATETXG);
if (create1 < create2)
rv = -1;
else if (create1 > create2)
rv = +1;
else
rv = 0;
zfs_close(guid1hdl);
zfs_close(guid2hdl);
return (rv);
}
/*
* This function reestablishes the hierarchy of encryption roots after a
* recursive incremental receive has completed. This must be done after the
* second call to recv_incremental_replication() has renamed and promoted all
* sent datasets to their final locations in the dataset hierarchy.
*/
static int
recv_fix_encryption_hierarchy(libzfs_handle_t *hdl, const char *top_zfs,
nvlist_t *stream_nv)
{
int err;
nvpair_t *fselem = NULL;
nvlist_t *stream_fss;
stream_fss = fnvlist_lookup_nvlist(stream_nv, "fss");
while ((fselem = nvlist_next_nvpair(stream_fss, fselem)) != NULL) {
zfs_handle_t *zhp = NULL;
uint64_t crypt;
nvlist_t *snaps, *props, *stream_nvfs = NULL;
nvpair_t *snapel = NULL;
boolean_t is_encroot, is_clone, stream_encroot;
char *cp;
const char *stream_keylocation = NULL;
char keylocation[MAXNAMELEN];
char fsname[ZFS_MAX_DATASET_NAME_LEN];
keylocation[0] = '\0';
stream_nvfs = fnvpair_value_nvlist(fselem);
snaps = fnvlist_lookup_nvlist(stream_nvfs, "snaps");
props = fnvlist_lookup_nvlist(stream_nvfs, "props");
stream_encroot = nvlist_exists(stream_nvfs, "is_encroot");
/* find a snapshot from the stream that exists locally */
err = ENOENT;
while ((snapel = nvlist_next_nvpair(snaps, snapel)) != NULL) {
uint64_t guid;
guid = fnvpair_value_uint64(snapel);
err = guid_to_name(hdl, top_zfs, guid, B_FALSE,
fsname);
if (err == 0)
break;
}
if (err != 0)
continue;
cp = strchr(fsname, '@');
if (cp != NULL)
*cp = '\0';
zhp = zfs_open(hdl, fsname, ZFS_TYPE_DATASET);
if (zhp == NULL) {
err = ENOENT;
goto error;
}
crypt = zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION);
is_clone = zhp->zfs_dmustats.dds_origin[0] != '\0';
(void) zfs_crypto_get_encryption_root(zhp, &is_encroot, NULL);
/* we don't need to do anything for unencrypted datasets */
if (crypt == ZIO_CRYPT_OFF) {
zfs_close(zhp);
continue;
}
/*
* If the dataset is flagged as an encryption root, was not
* received as a clone and is not currently an encryption root,
* force it to become one. Fixup the keylocation if necessary.
*/
if (stream_encroot) {
if (!is_clone && !is_encroot) {
err = lzc_change_key(fsname,
DCP_CMD_FORCE_NEW_KEY, NULL, NULL, 0);
if (err != 0) {
zfs_close(zhp);
goto error;
}
}
stream_keylocation = fnvlist_lookup_string(props,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION));
/*
* Refresh the properties in case the call to
* lzc_change_key() changed the value.
*/
zfs_refresh_properties(zhp);
err = zfs_prop_get(zhp, ZFS_PROP_KEYLOCATION,
keylocation, sizeof (keylocation), NULL, NULL,
0, B_TRUE);
if (err != 0) {
zfs_close(zhp);
goto error;
}
if (strcmp(keylocation, stream_keylocation) != 0) {
err = zfs_prop_set(zhp,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION),
stream_keylocation);
if (err != 0) {
zfs_close(zhp);
goto error;
}
}
}
/*
* If the dataset is not flagged as an encryption root and is
* currently an encryption root, force it to inherit from its
* parent. The root of a raw send should never be
* force-inherited.
*/
if (!stream_encroot && is_encroot &&
strcmp(top_zfs, fsname) != 0) {
err = lzc_change_key(fsname, DCP_CMD_FORCE_INHERIT,
NULL, NULL, 0);
if (err != 0) {
zfs_close(zhp);
goto error;
}
}
zfs_close(zhp);
}
return (0);
error:
return (err);
}
static int
recv_incremental_replication(libzfs_handle_t *hdl, const char *tofs,
recvflags_t *flags, nvlist_t *stream_nv, avl_tree_t *stream_avl,
nvlist_t *renamed)
{
nvlist_t *local_nv, *deleted = NULL;
avl_tree_t *local_avl;
nvpair_t *fselem, *nextfselem;
const char *fromsnap;
char newname[ZFS_MAX_DATASET_NAME_LEN];
char guidname[32];
int error;
boolean_t needagain, progress, recursive;
const char *s1, *s2;
fromsnap = fnvlist_lookup_string(stream_nv, "fromsnap");
recursive = (nvlist_lookup_boolean(stream_nv, "not_recursive") ==
ENOENT);
if (flags->dryrun)
return (0);
again:
needagain = progress = B_FALSE;
deleted = fnvlist_alloc();
if ((error = gather_nvlist(hdl, tofs, fromsnap, NULL,
recursive, B_TRUE, B_FALSE, recursive, B_FALSE, B_FALSE, B_FALSE,
B_FALSE, B_TRUE, &local_nv, &local_avl)) != 0)
return (error);
/*
* Process deletes and renames
*/
for (fselem = nvlist_next_nvpair(local_nv, NULL);
fselem; fselem = nextfselem) {
nvlist_t *nvfs, *snaps;
nvlist_t *stream_nvfs = NULL;
nvpair_t *snapelem, *nextsnapelem;
uint64_t fromguid = 0;
uint64_t originguid = 0;
uint64_t stream_originguid = 0;
uint64_t parent_fromsnap_guid, stream_parent_fromsnap_guid;
const char *fsname, *stream_fsname;
nextfselem = nvlist_next_nvpair(local_nv, fselem);
nvfs = fnvpair_value_nvlist(fselem);
snaps = fnvlist_lookup_nvlist(nvfs, "snaps");
fsname = fnvlist_lookup_string(nvfs, "name");
parent_fromsnap_guid = fnvlist_lookup_uint64(nvfs,
"parentfromsnap");
(void) nvlist_lookup_uint64(nvfs, "origin", &originguid);
/*
* First find the stream's fs, so we can check for
* a different origin (due to "zfs promote")
*/
for (snapelem = nvlist_next_nvpair(snaps, NULL);
snapelem; snapelem = nvlist_next_nvpair(snaps, snapelem)) {
uint64_t thisguid;
thisguid = fnvpair_value_uint64(snapelem);
stream_nvfs = fsavl_find(stream_avl, thisguid, NULL);
if (stream_nvfs != NULL)
break;
}
/* check for promote */
(void) nvlist_lookup_uint64(stream_nvfs, "origin",
&stream_originguid);
if (stream_nvfs && originguid != stream_originguid) {
switch (created_before(hdl, local_avl,
stream_originguid, originguid)) {
case 1: {
/* promote it! */
nvlist_t *origin_nvfs;
const char *origin_fsname;
origin_nvfs = fsavl_find(local_avl, originguid,
NULL);
origin_fsname = fnvlist_lookup_string(
origin_nvfs, "name");
error = recv_promote(hdl, fsname, origin_fsname,
flags);
if (error == 0)
progress = B_TRUE;
break;
}
default:
break;
case -1:
fsavl_destroy(local_avl);
fnvlist_free(local_nv);
return (-1);
}
/*
* We had/have the wrong origin, therefore our
* list of snapshots is wrong. Need to handle
* them on the next pass.
*/
needagain = B_TRUE;
continue;
}
for (snapelem = nvlist_next_nvpair(snaps, NULL);
snapelem; snapelem = nextsnapelem) {
uint64_t thisguid;
const char *stream_snapname;
nvlist_t *found, *props;
nextsnapelem = nvlist_next_nvpair(snaps, snapelem);
thisguid = fnvpair_value_uint64(snapelem);
found = fsavl_find(stream_avl, thisguid,
&stream_snapname);
/* check for delete */
if (found == NULL) {
char name[ZFS_MAX_DATASET_NAME_LEN];
if (!flags->force)
continue;
(void) snprintf(name, sizeof (name), "%s@%s",
fsname, nvpair_name(snapelem));
error = recv_destroy(hdl, name,
strlen(fsname)+1, newname, flags);
if (error)
needagain = B_TRUE;
else
progress = B_TRUE;
sprintf(guidname, "%llu",
(u_longlong_t)thisguid);
nvlist_add_boolean(deleted, guidname);
continue;
}
stream_nvfs = found;
if (0 == nvlist_lookup_nvlist(stream_nvfs, "snapprops",
&props) && 0 == nvlist_lookup_nvlist(props,
stream_snapname, &props)) {
zfs_cmd_t zc = {"\0"};
zc.zc_cookie = B_TRUE; /* received */
(void) snprintf(zc.zc_name, sizeof (zc.zc_name),
"%s@%s", fsname, nvpair_name(snapelem));
zcmd_write_src_nvlist(hdl, &zc, props);
(void) zfs_ioctl(hdl,
ZFS_IOC_SET_PROP, &zc);
zcmd_free_nvlists(&zc);
}
/* check for different snapname */
if (strcmp(nvpair_name(snapelem),
stream_snapname) != 0) {
char name[ZFS_MAX_DATASET_NAME_LEN];
char tryname[ZFS_MAX_DATASET_NAME_LEN];
(void) snprintf(name, sizeof (name), "%s@%s",
fsname, nvpair_name(snapelem));
(void) snprintf(tryname, sizeof (name), "%s@%s",
fsname, stream_snapname);
error = recv_rename(hdl, name, tryname,
strlen(fsname)+1, newname, flags);
if (error)
needagain = B_TRUE;
else
progress = B_TRUE;
}
if (strcmp(stream_snapname, fromsnap) == 0)
fromguid = thisguid;
}
/* check for delete */
if (stream_nvfs == NULL) {
if (!flags->force)
continue;
error = recv_destroy(hdl, fsname, strlen(tofs)+1,
newname, flags);
if (error)
needagain = B_TRUE;
else
progress = B_TRUE;
sprintf(guidname, "%llu",
(u_longlong_t)parent_fromsnap_guid);
nvlist_add_boolean(deleted, guidname);
continue;
}
if (fromguid == 0) {
if (flags->verbose) {
(void) printf("local fs %s does not have "
"fromsnap (%s in stream); must have "
"been deleted locally; ignoring\n",
fsname, fromsnap);
}
continue;
}
stream_fsname = fnvlist_lookup_string(stream_nvfs, "name");
stream_parent_fromsnap_guid = fnvlist_lookup_uint64(
stream_nvfs, "parentfromsnap");
s1 = strrchr(fsname, '/');
s2 = strrchr(stream_fsname, '/');
/*
* Check if we're going to rename based on parent guid change
* and the current parent guid was also deleted. If it was then
* rename will fail and is likely unneeded, so avoid this and
* force an early retry to determine the new
* parent_fromsnap_guid.
*/
if (stream_parent_fromsnap_guid != 0 &&
parent_fromsnap_guid != 0 &&
stream_parent_fromsnap_guid != parent_fromsnap_guid) {
sprintf(guidname, "%llu",
(u_longlong_t)parent_fromsnap_guid);
if (nvlist_exists(deleted, guidname)) {
progress = B_TRUE;
needagain = B_TRUE;
goto doagain;
}
}
/*
* Check for rename. If the exact receive path is specified, it
* does not count as a rename, but we still need to check the
* datasets beneath it.
*/
if ((stream_parent_fromsnap_guid != 0 &&
parent_fromsnap_guid != 0 &&
stream_parent_fromsnap_guid != parent_fromsnap_guid) ||
((flags->isprefix || strcmp(tofs, fsname) != 0) &&
(s1 != NULL) && (s2 != NULL) && strcmp(s1, s2) != 0)) {
nvlist_t *parent;
char tryname[ZFS_MAX_DATASET_NAME_LEN];
parent = fsavl_find(local_avl,
stream_parent_fromsnap_guid, NULL);
/*
* NB: parent might not be found if we used the
* tosnap for stream_parent_fromsnap_guid,
* because the parent is a newly-created fs;
* we'll be able to rename it after we recv the
* new fs.
*/
if (parent != NULL) {
const char *pname;
pname = fnvlist_lookup_string(parent, "name");
(void) snprintf(tryname, sizeof (tryname),
"%s%s", pname, strrchr(stream_fsname, '/'));
} else {
tryname[0] = '\0';
if (flags->verbose) {
(void) printf("local fs %s new parent "
"not found\n", fsname);
}
}
newname[0] = '\0';
error = recv_rename(hdl, fsname, tryname,
strlen(tofs)+1, newname, flags);
if (renamed != NULL && newname[0] != '\0') {
fnvlist_add_boolean(renamed, newname);
}
if (error)
needagain = B_TRUE;
else
progress = B_TRUE;
}
}
doagain:
fsavl_destroy(local_avl);
fnvlist_free(local_nv);
fnvlist_free(deleted);
if (needagain && progress) {
/* do another pass to fix up temporary names */
if (flags->verbose)
(void) printf("another pass:\n");
goto again;
}
return (needagain || error != 0);
}
static int
zfs_receive_package(libzfs_handle_t *hdl, int fd, const char *destname,
recvflags_t *flags, dmu_replay_record_t *drr, zio_cksum_t *zc,
char **top_zfs, nvlist_t *cmdprops)
{
nvlist_t *stream_nv = NULL;
avl_tree_t *stream_avl = NULL;
const char *fromsnap = NULL;
const char *sendsnap = NULL;
char *cp;
char tofs[ZFS_MAX_DATASET_NAME_LEN];
char sendfs[ZFS_MAX_DATASET_NAME_LEN];
char errbuf[ERRBUFLEN];
dmu_replay_record_t drre;
int error;
boolean_t anyerr = B_FALSE;
boolean_t softerr = B_FALSE;
boolean_t recursive, raw;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot receive"));
assert(drr->drr_type == DRR_BEGIN);
assert(drr->drr_u.drr_begin.drr_magic == DMU_BACKUP_MAGIC);
assert(DMU_GET_STREAM_HDRTYPE(drr->drr_u.drr_begin.drr_versioninfo) ==
DMU_COMPOUNDSTREAM);
/*
* Read in the nvlist from the stream.
*/
if (drr->drr_payloadlen != 0) {
error = recv_read_nvlist(hdl, fd, drr->drr_payloadlen,
&stream_nv, flags->byteswap, zc);
if (error) {
error = zfs_error(hdl, EZFS_BADSTREAM, errbuf);
goto out;
}
}
recursive = (nvlist_lookup_boolean(stream_nv, "not_recursive") ==
ENOENT);
raw = (nvlist_lookup_boolean(stream_nv, "raw") == 0);
if (recursive && strchr(destname, '@')) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot specify snapshot name for multi-snapshot stream"));
error = zfs_error(hdl, EZFS_BADSTREAM, errbuf);
goto out;
}
/*
* Read in the end record and verify checksum.
*/
if (0 != (error = recv_read(hdl, fd, &drre, sizeof (drre),
flags->byteswap, NULL)))
goto out;
if (flags->byteswap) {
drre.drr_type = BSWAP_32(drre.drr_type);
drre.drr_u.drr_end.drr_checksum.zc_word[0] =
BSWAP_64(drre.drr_u.drr_end.drr_checksum.zc_word[0]);
drre.drr_u.drr_end.drr_checksum.zc_word[1] =
BSWAP_64(drre.drr_u.drr_end.drr_checksum.zc_word[1]);
drre.drr_u.drr_end.drr_checksum.zc_word[2] =
BSWAP_64(drre.drr_u.drr_end.drr_checksum.zc_word[2]);
drre.drr_u.drr_end.drr_checksum.zc_word[3] =
BSWAP_64(drre.drr_u.drr_end.drr_checksum.zc_word[3]);
}
if (drre.drr_type != DRR_END) {
error = zfs_error(hdl, EZFS_BADSTREAM, errbuf);
goto out;
}
if (!ZIO_CHECKSUM_EQUAL(drre.drr_u.drr_end.drr_checksum, *zc)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incorrect header checksum"));
error = zfs_error(hdl, EZFS_BADSTREAM, errbuf);
goto out;
}
(void) nvlist_lookup_string(stream_nv, "fromsnap", &fromsnap);
if (drr->drr_payloadlen != 0) {
nvlist_t *stream_fss;
stream_fss = fnvlist_lookup_nvlist(stream_nv, "fss");
if ((stream_avl = fsavl_create(stream_fss)) == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"couldn't allocate avl tree"));
error = zfs_error(hdl, EZFS_NOMEM, errbuf);
goto out;
}
if (fromsnap != NULL && recursive) {
nvlist_t *renamed = NULL;
nvpair_t *pair = NULL;
(void) strlcpy(tofs, destname, sizeof (tofs));
if (flags->isprefix) {
struct drr_begin *drrb = &drr->drr_u.drr_begin;
int i;
if (flags->istail) {
cp = strrchr(drrb->drr_toname, '/');
if (cp == NULL) {
(void) strlcat(tofs, "/",
sizeof (tofs));
i = 0;
} else {
i = (cp - drrb->drr_toname);
}
} else {
i = strcspn(drrb->drr_toname, "/@");
}
/* zfs_receive_one() will create_parents() */
(void) strlcat(tofs, &drrb->drr_toname[i],
sizeof (tofs));
*strchr(tofs, '@') = '\0';
}
if (!flags->dryrun && !flags->nomount) {
renamed = fnvlist_alloc();
}
softerr = recv_incremental_replication(hdl, tofs, flags,
stream_nv, stream_avl, renamed);
/* Unmount renamed filesystems before receiving. */
while ((pair = nvlist_next_nvpair(renamed,
pair)) != NULL) {
zfs_handle_t *zhp;
prop_changelist_t *clp = NULL;
zhp = zfs_open(hdl, nvpair_name(pair),
ZFS_TYPE_FILESYSTEM);
if (zhp != NULL) {
clp = changelist_gather(zhp,
ZFS_PROP_MOUNTPOINT, 0,
flags->forceunmount ? MS_FORCE : 0);
zfs_close(zhp);
if (clp != NULL) {
softerr |=
changelist_prefix(clp);
changelist_free(clp);
}
}
}
fnvlist_free(renamed);
}
}
/*
* Get the fs specified by the first path in the stream (the top level
* specified by 'zfs send') and pass it to each invocation of
* zfs_receive_one().
*/
(void) strlcpy(sendfs, drr->drr_u.drr_begin.drr_toname,
sizeof (sendfs));
if ((cp = strchr(sendfs, '@')) != NULL) {
*cp = '\0';
/*
* Find the "sendsnap", the final snapshot in a replication
* stream. zfs_receive_one() handles certain errors
* differently, depending on if the contained stream is the
* last one or not.
*/
sendsnap = (cp + 1);
}
/* Finally, receive each contained stream */
do {
/*
* we should figure out if it has a recoverable
* error, in which case do a recv_skip() and drive on.
* Note, if we fail due to already having this guid,
* zfs_receive_one() will take care of it (ie,
* recv_skip() and return 0).
*/
error = zfs_receive_impl(hdl, destname, NULL, flags, fd,
sendfs, stream_nv, stream_avl, top_zfs, sendsnap, cmdprops);
if (error == ENODATA) {
error = 0;
break;
}
anyerr |= error;
} while (error == 0);
if (drr->drr_payloadlen != 0 && recursive && fromsnap != NULL) {
/*
* Now that we have the fs's they sent us, try the
* renames again.
*/
softerr = recv_incremental_replication(hdl, tofs, flags,
stream_nv, stream_avl, NULL);
}
if (raw && softerr == 0 && *top_zfs != NULL) {
softerr = recv_fix_encryption_hierarchy(hdl, *top_zfs,
stream_nv);
}
out:
fsavl_destroy(stream_avl);
fnvlist_free(stream_nv);
if (softerr)
error = -2;
if (anyerr)
error = -1;
return (error);
}
static void
trunc_prop_errs(int truncated)
{
ASSERT(truncated != 0);
if (truncated == 1)
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"1 more property could not be set\n"));
else
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"%d more properties could not be set\n"), truncated);
}
static int
recv_skip(libzfs_handle_t *hdl, int fd, boolean_t byteswap)
{
dmu_replay_record_t *drr;
void *buf = zfs_alloc(hdl, SPA_MAXBLOCKSIZE);
uint64_t payload_size;
char errbuf[ERRBUFLEN];
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot receive"));
/* XXX would be great to use lseek if possible... */
drr = buf;
while (recv_read(hdl, fd, drr, sizeof (dmu_replay_record_t),
byteswap, NULL) == 0) {
if (byteswap)
drr->drr_type = BSWAP_32(drr->drr_type);
switch (drr->drr_type) {
case DRR_BEGIN:
if (drr->drr_payloadlen != 0) {
(void) recv_read(hdl, fd, buf,
drr->drr_payloadlen, B_FALSE, NULL);
}
break;
case DRR_END:
free(buf);
return (0);
case DRR_OBJECT:
if (byteswap) {
drr->drr_u.drr_object.drr_bonuslen =
BSWAP_32(drr->drr_u.drr_object.
drr_bonuslen);
drr->drr_u.drr_object.drr_raw_bonuslen =
BSWAP_32(drr->drr_u.drr_object.
drr_raw_bonuslen);
}
payload_size =
DRR_OBJECT_PAYLOAD_SIZE(&drr->drr_u.drr_object);
(void) recv_read(hdl, fd, buf, payload_size,
B_FALSE, NULL);
break;
case DRR_WRITE:
if (byteswap) {
drr->drr_u.drr_write.drr_logical_size =
BSWAP_64(
drr->drr_u.drr_write.drr_logical_size);
drr->drr_u.drr_write.drr_compressed_size =
BSWAP_64(
drr->drr_u.drr_write.drr_compressed_size);
}
payload_size =
DRR_WRITE_PAYLOAD_SIZE(&drr->drr_u.drr_write);
assert(payload_size <= SPA_MAXBLOCKSIZE);
(void) recv_read(hdl, fd, buf,
payload_size, B_FALSE, NULL);
break;
case DRR_SPILL:
if (byteswap) {
drr->drr_u.drr_spill.drr_length =
BSWAP_64(drr->drr_u.drr_spill.drr_length);
drr->drr_u.drr_spill.drr_compressed_size =
BSWAP_64(drr->drr_u.drr_spill.
drr_compressed_size);
}
payload_size =
DRR_SPILL_PAYLOAD_SIZE(&drr->drr_u.drr_spill);
(void) recv_read(hdl, fd, buf, payload_size,
B_FALSE, NULL);
break;
case DRR_WRITE_EMBEDDED:
if (byteswap) {
drr->drr_u.drr_write_embedded.drr_psize =
BSWAP_32(drr->drr_u.drr_write_embedded.
drr_psize);
}
(void) recv_read(hdl, fd, buf,
P2ROUNDUP(drr->drr_u.drr_write_embedded.drr_psize,
8), B_FALSE, NULL);
break;
case DRR_OBJECT_RANGE:
case DRR_WRITE_BYREF:
case DRR_FREEOBJECTS:
case DRR_FREE:
break;
default:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid record type"));
free(buf);
return (zfs_error(hdl, EZFS_BADSTREAM, errbuf));
}
}
free(buf);
return (-1);
}
static void
recv_ecksum_set_aux(libzfs_handle_t *hdl, const char *target_snap,
boolean_t resumable, boolean_t checksum)
{
char target_fs[ZFS_MAX_DATASET_NAME_LEN];
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, (checksum ?
"checksum mismatch" : "incomplete stream")));
if (!resumable)
return;
(void) strlcpy(target_fs, target_snap, sizeof (target_fs));
*strchr(target_fs, '@') = '\0';
zfs_handle_t *zhp = zfs_open(hdl, target_fs,
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL)
return;
char token_buf[ZFS_MAXPROPLEN];
int error = zfs_prop_get(zhp, ZFS_PROP_RECEIVE_RESUME_TOKEN,
token_buf, sizeof (token_buf),
NULL, NULL, 0, B_TRUE);
if (error == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"checksum mismatch or incomplete stream.\n"
"Partially received snapshot is saved.\n"
"A resuming stream can be generated on the sending "
"system by running:\n"
" zfs send -t %s"),
token_buf);
}
zfs_close(zhp);
}
/*
* Prepare a new nvlist of properties that are to override (-o) or be excluded
* (-x) from the received dataset
* recvprops: received properties from the send stream
* cmdprops: raw input properties from command line
* origprops: properties, both locally-set and received, currently set on the
* target dataset if it exists, NULL otherwise.
* oxprops: valid output override (-o) and excluded (-x) properties
*/
static int
zfs_setup_cmdline_props(libzfs_handle_t *hdl, zfs_type_t type,
char *fsname, boolean_t zoned, boolean_t recursive, boolean_t newfs,
boolean_t raw, boolean_t toplevel, nvlist_t *recvprops, nvlist_t *cmdprops,
nvlist_t *origprops, nvlist_t **oxprops, uint8_t **wkeydata_out,
uint_t *wkeylen_out, const char *errbuf)
{
nvpair_t *nvp;
nvlist_t *oprops, *voprops;
zfs_handle_t *zhp = NULL;
zpool_handle_t *zpool_hdl = NULL;
char *cp;
int ret = 0;
char namebuf[ZFS_MAX_DATASET_NAME_LEN];
if (nvlist_empty(cmdprops))
return (0); /* No properties to override or exclude */
*oxprops = fnvlist_alloc();
oprops = fnvlist_alloc();
strlcpy(namebuf, fsname, ZFS_MAX_DATASET_NAME_LEN);
/*
* Get our dataset handle. The target dataset may not exist yet.
*/
if (zfs_dataset_exists(hdl, namebuf, ZFS_TYPE_DATASET)) {
zhp = zfs_open(hdl, namebuf, ZFS_TYPE_DATASET);
if (zhp == NULL) {
ret = -1;
goto error;
}
}
/* open the zpool handle */
cp = strchr(namebuf, '/');
if (cp != NULL)
*cp = '\0';
zpool_hdl = zpool_open(hdl, namebuf);
if (zpool_hdl == NULL) {
ret = -1;
goto error;
}
/* restore namebuf to match fsname for later use */
if (cp != NULL)
*cp = '/';
/*
* first iteration: process excluded (-x) properties now and gather
* added (-o) properties to be later processed by zfs_valid_proplist()
*/
nvp = NULL;
while ((nvp = nvlist_next_nvpair(cmdprops, nvp)) != NULL) {
const char *name = nvpair_name(nvp);
zfs_prop_t prop = zfs_name_to_prop(name);
/*
* It turns out, if we don't normalize "aliased" names
* e.g. compress= against the "real" names (e.g. compression)
* here, then setting/excluding them does not work as
* intended.
*
* But since user-defined properties wouldn't have a valid
* mapping here, we do this conditional dance.
*/
const char *newname = name;
if (prop >= ZFS_PROP_TYPE)
newname = zfs_prop_to_name(prop);
/* "origin" is processed separately, don't handle it here */
if (prop == ZFS_PROP_ORIGIN)
continue;
/* raw streams can't override encryption properties */
if ((zfs_prop_encryption_key_param(prop) ||
prop == ZFS_PROP_ENCRYPTION) && raw) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"encryption property '%s' cannot "
"be set or excluded for raw streams."), name);
ret = zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
/*
* For plain replicated send, we can ignore encryption
* properties other than first stream
*/
if ((zfs_prop_encryption_key_param(prop) || prop ==
ZFS_PROP_ENCRYPTION) && !newfs && recursive && !raw) {
continue;
}
/* incremental streams can only exclude encryption properties */
if ((zfs_prop_encryption_key_param(prop) ||
prop == ZFS_PROP_ENCRYPTION) && !newfs &&
nvpair_type(nvp) != DATA_TYPE_BOOLEAN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"encryption property '%s' cannot "
"be set for incremental streams."), name);
ret = zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
switch (nvpair_type(nvp)) {
case DATA_TYPE_BOOLEAN: /* -x property */
/*
* DATA_TYPE_BOOLEAN is the way we're asked to "exclude"
* a property: this is done by forcing an explicit
* inherit on the destination so the effective value is
* not the one we received from the send stream.
*/
if (!zfs_prop_valid_for_type(prop, type, B_FALSE) &&
!zfs_prop_user(name)) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN,
"Warning: %s: property '%s' does not "
"apply to datasets of this type\n"),
fsname, name);
continue;
}
/*
* We do this only if the property is not already
* locally-set, in which case its value will take
* priority over the received anyway.
*/
if (nvlist_exists(origprops, newname)) {
nvlist_t *attrs;
const char *source = NULL;
attrs = fnvlist_lookup_nvlist(origprops,
newname);
if (nvlist_lookup_string(attrs,
ZPROP_SOURCE, &source) == 0 &&
strcmp(source, ZPROP_SOURCE_VAL_RECVD) != 0)
continue;
}
/*
* We can't force an explicit inherit on non-inheritable
* properties: if we're asked to exclude this kind of
* values we remove them from "recvprops" input nvlist.
*/
if (!zfs_prop_user(name) && /* can be inherited too */
!zfs_prop_inheritable(prop) &&
nvlist_exists(recvprops, newname))
fnvlist_remove(recvprops, newname);
else
fnvlist_add_boolean(*oxprops, newname);
break;
case DATA_TYPE_STRING: /* -o property=value */
/*
* we're trying to override a property that does not
* make sense for this type of dataset, but we don't
* want to fail if the receive is recursive: this comes
* in handy when the send stream contains, for
* instance, a child ZVOL and we're trying to receive
* it with "-o atime=on"
*/
if (!zfs_prop_valid_for_type(prop, type, B_FALSE) &&
!zfs_prop_user(name)) {
if (recursive)
continue;
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' does not apply to datasets "
"of this type"), name);
ret = zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
fnvlist_add_string(oprops, newname,
fnvpair_value_string(nvp));
break;
default:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property '%s' must be a string or boolean"), name);
ret = zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
}
if (toplevel) {
/* convert override strings properties to native */
if ((voprops = zfs_valid_proplist(hdl, ZFS_TYPE_DATASET,
oprops, zoned, zhp, zpool_hdl, B_FALSE, errbuf)) == NULL) {
ret = zfs_error(hdl, EZFS_BADPROP, errbuf);
goto error;
}
/*
* zfs_crypto_create() requires the parent name. Get it
* by truncating the fsname copy stored in namebuf.
*/
cp = strrchr(namebuf, '/');
if (cp != NULL)
*cp = '\0';
if (!raw && !(!newfs && recursive) &&
zfs_crypto_create(hdl, namebuf, voprops, NULL,
B_FALSE, wkeydata_out, wkeylen_out) != 0) {
fnvlist_free(voprops);
ret = zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf);
goto error;
}
/* second pass: process "-o" properties */
fnvlist_merge(*oxprops, voprops);
fnvlist_free(voprops);
} else {
/* override props on child dataset are inherited */
nvp = NULL;
while ((nvp = nvlist_next_nvpair(oprops, nvp)) != NULL) {
const char *name = nvpair_name(nvp);
fnvlist_add_boolean(*oxprops, name);
}
}
error:
if (zhp != NULL)
zfs_close(zhp);
if (zpool_hdl != NULL)
zpool_close(zpool_hdl);
fnvlist_free(oprops);
return (ret);
}
/*
* Restores a backup of tosnap from the file descriptor specified by infd.
*/
static int
zfs_receive_one(libzfs_handle_t *hdl, int infd, const char *tosnap,
const char *originsnap, recvflags_t *flags, dmu_replay_record_t *drr,
dmu_replay_record_t *drr_noswap, const char *sendfs, nvlist_t *stream_nv,
avl_tree_t *stream_avl, char **top_zfs,
const char *finalsnap, nvlist_t *cmdprops)
{
struct timespec begin_time;
int ioctl_err, ioctl_errno, err;
char *cp;
struct drr_begin *drrb = &drr->drr_u.drr_begin;
char errbuf[ERRBUFLEN];
const char *chopprefix;
boolean_t newfs = B_FALSE;
boolean_t stream_wantsnewfs, stream_resumingnewfs;
boolean_t newprops = B_FALSE;
uint64_t read_bytes = 0;
uint64_t errflags = 0;
uint64_t parent_snapguid = 0;
prop_changelist_t *clp = NULL;
nvlist_t *snapprops_nvlist = NULL;
nvlist_t *snapholds_nvlist = NULL;
zprop_errflags_t prop_errflags;
nvlist_t *prop_errors = NULL;
boolean_t recursive;
const char *snapname = NULL;
char destsnap[MAXPATHLEN * 2];
char origin[MAXNAMELEN] = {0};
char name[MAXPATHLEN];
char tmp_keylocation[MAXNAMELEN] = {0};
nvlist_t *rcvprops = NULL; /* props received from the send stream */
nvlist_t *oxprops = NULL; /* override (-o) and exclude (-x) props */
nvlist_t *origprops = NULL; /* original props (if destination exists) */
zfs_type_t type = ZFS_TYPE_INVALID;
boolean_t toplevel = B_FALSE;
boolean_t zoned = B_FALSE;
boolean_t hastoken = B_FALSE;
boolean_t redacted;
uint8_t *wkeydata = NULL;
uint_t wkeylen = 0;
#ifndef CLOCK_MONOTONIC_RAW
#define CLOCK_MONOTONIC_RAW CLOCK_MONOTONIC
#endif
clock_gettime(CLOCK_MONOTONIC_RAW, &begin_time);
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot receive"));
recursive = (nvlist_lookup_boolean(stream_nv, "not_recursive") ==
ENOENT);
/* Did the user request holds be skipped via zfs recv -k? */
boolean_t holds = flags->holds && !flags->skipholds;
if (stream_avl != NULL) {
const char *keylocation = NULL;
nvlist_t *lookup = NULL;
nvlist_t *fs = fsavl_find(stream_avl, drrb->drr_toguid,
&snapname);
(void) nvlist_lookup_uint64(fs, "parentfromsnap",
&parent_snapguid);
err = nvlist_lookup_nvlist(fs, "props", &rcvprops);
if (err) {
rcvprops = fnvlist_alloc();
newprops = B_TRUE;
}
/*
* The keylocation property may only be set on encryption roots,
* but this dataset might not become an encryption root until
* recv_fix_encryption_hierarchy() is called. That function
* will fixup the keylocation anyway, so we temporarily unset
* the keylocation for now to avoid any errors from the receive
* ioctl.
*/
err = nvlist_lookup_string(rcvprops,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION), &keylocation);
if (err == 0) {
strlcpy(tmp_keylocation, keylocation, MAXNAMELEN);
(void) nvlist_remove_all(rcvprops,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION));
}
if (flags->canmountoff) {
fnvlist_add_uint64(rcvprops,
zfs_prop_to_name(ZFS_PROP_CANMOUNT), 0);
} else if (newprops) { /* nothing in rcvprops, eliminate it */
fnvlist_free(rcvprops);
rcvprops = NULL;
newprops = B_FALSE;
}
if (0 == nvlist_lookup_nvlist(fs, "snapprops", &lookup)) {
snapprops_nvlist = fnvlist_lookup_nvlist(lookup,
snapname);
}
if (holds) {
if (0 == nvlist_lookup_nvlist(fs, "snapholds",
&lookup)) {
snapholds_nvlist = fnvlist_lookup_nvlist(
lookup, snapname);
}
}
}
cp = NULL;
/*
* Determine how much of the snapshot name stored in the stream
* we are going to tack on to the name they specified on the
* command line, and how much we are going to chop off.
*
* If they specified a snapshot, chop the entire name stored in
* the stream.
*/
if (flags->istail) {
/*
* A filesystem was specified with -e. We want to tack on only
* the tail of the sent snapshot path.
*/
if (strchr(tosnap, '@')) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
"argument - snapshot not allowed with -e"));
err = zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
goto out;
}
chopprefix = strrchr(sendfs, '/');
if (chopprefix == NULL) {
/*
* The tail is the poolname, so we need to
* prepend a path separator.
*/
int len = strlen(drrb->drr_toname);
cp = umem_alloc(len + 2, UMEM_NOFAIL);
cp[0] = '/';
(void) strcpy(&cp[1], drrb->drr_toname);
chopprefix = cp;
} else {
chopprefix = drrb->drr_toname + (chopprefix - sendfs);
}
} else if (flags->isprefix) {
/*
* A filesystem was specified with -d. We want to tack on
* everything but the first element of the sent snapshot path
* (all but the pool name).
*/
if (strchr(tosnap, '@')) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
"argument - snapshot not allowed with -d"));
err = zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
goto out;
}
chopprefix = strchr(drrb->drr_toname, '/');
if (chopprefix == NULL)
chopprefix = strchr(drrb->drr_toname, '@');
} else if (strchr(tosnap, '@') == NULL) {
/*
* If a filesystem was specified without -d or -e, we want to
* tack on everything after the fs specified by 'zfs send'.
*/
chopprefix = drrb->drr_toname + strlen(sendfs);
} else {
/* A snapshot was specified as an exact path (no -d or -e). */
if (recursive) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot specify snapshot name for multi-snapshot "
"stream"));
err = zfs_error(hdl, EZFS_BADSTREAM, errbuf);
goto out;
}
chopprefix = drrb->drr_toname + strlen(drrb->drr_toname);
}
ASSERT(strstr(drrb->drr_toname, sendfs) == drrb->drr_toname);
ASSERT(chopprefix > drrb->drr_toname || strchr(sendfs, '/') == NULL);
ASSERT(chopprefix <= drrb->drr_toname + strlen(drrb->drr_toname) ||
strchr(sendfs, '/') == NULL);
ASSERT(chopprefix[0] == '/' || chopprefix[0] == '@' ||
chopprefix[0] == '\0');
/*
* Determine name of destination snapshot.
*/
(void) strlcpy(destsnap, tosnap, sizeof (destsnap));
(void) strlcat(destsnap, chopprefix, sizeof (destsnap));
if (cp != NULL)
umem_free(cp, strlen(cp) + 1);
if (!zfs_name_valid(destsnap, ZFS_TYPE_SNAPSHOT)) {
err = zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
goto out;
}
/*
* Determine the name of the origin snapshot.
*/
if (originsnap) {
(void) strlcpy(origin, originsnap, sizeof (origin));
if (flags->verbose)
(void) printf("using provided clone origin %s\n",
origin);
} else if (drrb->drr_flags & DRR_FLAG_CLONE) {
if (guid_to_name(hdl, destsnap,
drrb->drr_fromguid, B_FALSE, origin) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"local origin for clone %s does not exist"),
destsnap);
err = zfs_error(hdl, EZFS_NOENT, errbuf);
goto out;
}
if (flags->verbose)
(void) printf("found clone origin %s\n", origin);
}
if ((DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo) &
DMU_BACKUP_FEATURE_DEDUP)) {
(void) fprintf(stderr,
gettext("ERROR: \"zfs receive\" no longer supports "
"deduplicated send streams. Use\n"
"the \"zstream redup\" command to convert this stream "
"to a regular,\n"
"non-deduplicated stream.\n"));
err = zfs_error(hdl, EZFS_NOTSUP, errbuf);
goto out;
}
boolean_t resuming = DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo) &
DMU_BACKUP_FEATURE_RESUMING;
boolean_t raw = DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo) &
DMU_BACKUP_FEATURE_RAW;
boolean_t embedded = DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo) &
DMU_BACKUP_FEATURE_EMBED_DATA;
stream_wantsnewfs = (drrb->drr_fromguid == 0 ||
(drrb->drr_flags & DRR_FLAG_CLONE) || originsnap) && !resuming;
stream_resumingnewfs = (drrb->drr_fromguid == 0 ||
(drrb->drr_flags & DRR_FLAG_CLONE) || originsnap) && resuming;
if (stream_wantsnewfs) {
/*
* if the parent fs does not exist, look for it based on
* the parent snap GUID
*/
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot receive new filesystem stream"));
(void) strlcpy(name, destsnap, sizeof (name));
cp = strrchr(name, '/');
if (cp)
*cp = '\0';
if (cp &&
!zfs_dataset_exists(hdl, name, ZFS_TYPE_DATASET)) {
char suffix[ZFS_MAX_DATASET_NAME_LEN];
(void) strlcpy(suffix, strrchr(destsnap, '/'),
sizeof (suffix));
if (guid_to_name(hdl, name, parent_snapguid,
B_FALSE, destsnap) == 0) {
*strchr(destsnap, '@') = '\0';
(void) strlcat(destsnap, suffix,
sizeof (destsnap));
}
}
} else {
/*
* If the fs does not exist, look for it based on the
* fromsnap GUID.
*/
if (resuming) {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot receive resume stream"));
} else {
(void) snprintf(errbuf, sizeof (errbuf),
dgettext(TEXT_DOMAIN,
"cannot receive incremental stream"));
}
(void) strlcpy(name, destsnap, sizeof (name));
*strchr(name, '@') = '\0';
/*
* If the exact receive path was specified and this is the
* topmost path in the stream, then if the fs does not exist we
* should look no further.
*/
if ((flags->isprefix || (*(chopprefix = drrb->drr_toname +
strlen(sendfs)) != '\0' && *chopprefix != '@')) &&
!zfs_dataset_exists(hdl, name, ZFS_TYPE_DATASET)) {
char snap[ZFS_MAX_DATASET_NAME_LEN];
(void) strlcpy(snap, strchr(destsnap, '@'),
sizeof (snap));
if (guid_to_name(hdl, name, drrb->drr_fromguid,
B_FALSE, destsnap) == 0) {
*strchr(destsnap, '@') = '\0';
(void) strlcat(destsnap, snap,
sizeof (destsnap));
}
}
}
(void) strlcpy(name, destsnap, sizeof (name));
*strchr(name, '@') = '\0';
redacted = DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo) &
DMU_BACKUP_FEATURE_REDACTED;
if (flags->heal) {
if (flags->isprefix || flags->istail || flags->force ||
flags->canmountoff || flags->resumable || flags->nomount ||
flags->skipholds) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"corrective recv can not be used when combined with"
" this flag"));
err = zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
goto out;
}
uint64_t guid =
get_snap_guid(hdl, name, strchr(destsnap, '@') + 1);
if (guid == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"corrective recv must specify an existing snapshot"
" to heal"));
err = zfs_error(hdl, EZFS_INVALIDNAME, errbuf);
goto out;
} else if (guid != drrb->drr_toguid) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"local snapshot doesn't match the snapshot"
" in the provided stream"));
err = zfs_error(hdl, EZFS_WRONG_PARENT, errbuf);
goto out;
}
} else if (zfs_dataset_exists(hdl, name, ZFS_TYPE_DATASET)) {
zfs_cmd_t zc = {"\0"};
zfs_handle_t *zhp = NULL;
boolean_t encrypted;
(void) strcpy(zc.zc_name, name);
/*
* Destination fs exists. It must be one of these cases:
* - an incremental send stream
* - the stream specifies a new fs (full stream or clone)
* and they want us to blow away the existing fs (and
* have therefore specified -F and removed any snapshots)
* - we are resuming a failed receive.
*/
if (stream_wantsnewfs) {
boolean_t is_volume = drrb->drr_type == DMU_OST_ZVOL;
if (!flags->force) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination '%s' exists\n"
"must specify -F to overwrite it"), name);
err = zfs_error(hdl, EZFS_EXISTS, errbuf);
goto out;
}
if (zfs_ioctl(hdl, ZFS_IOC_SNAPSHOT_LIST_NEXT,
&zc) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination has snapshots (eg. %s)\n"
"must destroy them to overwrite it"),
zc.zc_name);
err = zfs_error(hdl, EZFS_EXISTS, errbuf);
goto out;
}
if (is_volume && strrchr(name, '/') == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination %s is the root dataset\n"
"cannot overwrite with a ZVOL"),
name);
err = zfs_error(hdl, EZFS_EXISTS, errbuf);
goto out;
}
if (is_volume &&
zfs_ioctl(hdl, ZFS_IOC_DATASET_LIST_NEXT,
&zc) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination has children (eg. %s)\n"
"cannot overwrite with a ZVOL"),
zc.zc_name);
err = zfs_error(hdl, EZFS_WRONG_PARENT, errbuf);
goto out;
}
}
if ((zhp = zfs_open(hdl, name,
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME)) == NULL) {
err = -1;
goto out;
}
/*
* When receiving full/newfs on existing dataset, then it
* should be done with "-F" flag. Its enforced for initial
* receive in previous checks in this function.
* Similarly, on resuming full/newfs recv on existing dataset,
* it should be done with "-F" flag.
*
* When dataset doesn't exist, then full/newfs recv is done on
* newly created dataset and it's marked INCONSISTENT. But
* When receiving on existing dataset, recv is first done on
* %recv and its marked INCONSISTENT. Existing dataset is not
* marked INCONSISTENT.
* Resume of full/newfs receive with dataset not INCONSISTENT
* indicates that its resuming newfs on existing dataset. So,
* enforce "-F" flag in this case.
*/
if (stream_resumingnewfs &&
!zfs_prop_get_int(zhp, ZFS_PROP_INCONSISTENT) &&
!flags->force) {
zfs_close(zhp);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Resuming recv on existing destination '%s'\n"
"must specify -F to overwrite it"), name);
err = zfs_error(hdl, EZFS_RESUME_EXISTS, errbuf);
goto out;
}
if (stream_wantsnewfs &&
zhp->zfs_dmustats.dds_origin[0]) {
zfs_close(zhp);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination '%s' is a clone\n"
"must destroy it to overwrite it"), name);
err = zfs_error(hdl, EZFS_EXISTS, errbuf);
goto out;
}
/*
* Raw sends can not be performed as an incremental on top
* of existing unencrypted datasets. zfs recv -F can't be
* used to blow away an existing encrypted filesystem. This
* is because it would require the dsl dir to point to the
* new key (or lack of a key) and the old key at the same
* time. The -F flag may still be used for deleting
* intermediate snapshots that would otherwise prevent the
* receive from working.
*/
encrypted = zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION) !=
ZIO_CRYPT_OFF;
if (!stream_wantsnewfs && !encrypted && raw) {
zfs_close(zhp);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"cannot perform raw receive on top of "
"existing unencrypted dataset"));
err = zfs_error(hdl, EZFS_BADRESTORE, errbuf);
goto out;
}
if (stream_wantsnewfs && flags->force &&
((raw && !encrypted) || encrypted)) {
zfs_close(zhp);
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"zfs receive -F cannot be used to destroy an "
"encrypted filesystem or overwrite an "
"unencrypted one with an encrypted one"));
err = zfs_error(hdl, EZFS_BADRESTORE, errbuf);
goto out;
}
if (!flags->dryrun && zhp->zfs_type == ZFS_TYPE_FILESYSTEM &&
(stream_wantsnewfs || stream_resumingnewfs)) {
/* We can't do online recv in this case */
clp = changelist_gather(zhp, ZFS_PROP_NAME, 0,
flags->forceunmount ? MS_FORCE : 0);
if (clp == NULL) {
zfs_close(zhp);
err = -1;
goto out;
}
if (changelist_prefix(clp) != 0) {
changelist_free(clp);
zfs_close(zhp);
err = -1;
goto out;
}
}
/*
* If we are resuming a newfs, set newfs here so that we will
* mount it if the recv succeeds this time. We can tell
* that it was a newfs on the first recv because the fs
* itself will be inconsistent (if the fs existed when we
* did the first recv, we would have received it into
* .../%recv).
*/
if (resuming && zfs_prop_get_int(zhp, ZFS_PROP_INCONSISTENT))
newfs = B_TRUE;
/* we want to know if we're zoned when validating -o|-x props */
zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED);
/* may need this info later, get it now we have zhp around */
if (zfs_prop_get(zhp, ZFS_PROP_RECEIVE_RESUME_TOKEN, NULL, 0,
NULL, NULL, 0, B_TRUE) == 0)
hastoken = B_TRUE;
/* gather existing properties on destination */
origprops = fnvlist_alloc();
fnvlist_merge(origprops, zhp->zfs_props);
fnvlist_merge(origprops, zhp->zfs_user_props);
zfs_close(zhp);
} else {
zfs_handle_t *zhp;
/*
* Destination filesystem does not exist. Therefore we better
* be creating a new filesystem (either from a full backup, or
* a clone). It would therefore be invalid if the user
* specified only the pool name (i.e. if the destination name
* contained no slash character).
*/
cp = strrchr(name, '/');
if (!stream_wantsnewfs || cp == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination '%s' does not exist"), name);
err = zfs_error(hdl, EZFS_NOENT, errbuf);
goto out;
}
/*
* Trim off the final dataset component so we perform the
* recvbackup ioctl to the filesystems's parent.
*/
*cp = '\0';
if (flags->isprefix && !flags->istail && !flags->dryrun &&
create_parents(hdl, destsnap, strlen(tosnap)) != 0) {
err = zfs_error(hdl, EZFS_BADRESTORE, errbuf);
goto out;
}
/* validate parent */
zhp = zfs_open(hdl, name, ZFS_TYPE_DATASET);
if (zhp == NULL) {
err = zfs_error(hdl, EZFS_BADRESTORE, errbuf);
goto out;
}
if (zfs_get_type(zhp) != ZFS_TYPE_FILESYSTEM) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"parent '%s' is not a filesystem"), name);
err = zfs_error(hdl, EZFS_WRONG_PARENT, errbuf);
zfs_close(zhp);
goto out;
}
zfs_close(zhp);
newfs = B_TRUE;
*cp = '/';
}
if (flags->verbose) {
(void) printf("%s %s%s stream of %s into %s\n",
flags->dryrun ? "would receive" : "receiving",
flags->heal ? " corrective" : "",
drrb->drr_fromguid ? "incremental" : "full",
drrb->drr_toname, destsnap);
(void) fflush(stdout);
}
/*
* If this is the top-level dataset, record it so we can use it
* for recursive operations later.
*/
if (top_zfs != NULL &&
(*top_zfs == NULL || strcmp(*top_zfs, name) == 0)) {
toplevel = B_TRUE;
if (*top_zfs == NULL)
*top_zfs = zfs_strdup(hdl, name);
}
if (drrb->drr_type == DMU_OST_ZVOL) {
type = ZFS_TYPE_VOLUME;
} else if (drrb->drr_type == DMU_OST_ZFS) {
type = ZFS_TYPE_FILESYSTEM;
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid record type: 0x%d"), drrb->drr_type);
err = zfs_error(hdl, EZFS_BADSTREAM, errbuf);
goto out;
}
if ((err = zfs_setup_cmdline_props(hdl, type, name, zoned, recursive,
stream_wantsnewfs, raw, toplevel, rcvprops, cmdprops, origprops,
&oxprops, &wkeydata, &wkeylen, errbuf)) != 0)
goto out;
/*
* When sending with properties (zfs send -p), the encryption property
* is not included because it is a SETONCE property and therefore
* treated as read only. However, we are always able to determine its
* value because raw sends will include it in the DRR_BDEGIN payload
* and non-raw sends with properties are not allowed for encrypted
* datasets. Therefore, if this is a non-raw properties stream, we can
* infer that the value should be ZIO_CRYPT_OFF and manually add that
* to the received properties.
*/
if (stream_wantsnewfs && !raw && rcvprops != NULL &&
!nvlist_exists(cmdprops, zfs_prop_to_name(ZFS_PROP_ENCRYPTION))) {
if (oxprops == NULL)
oxprops = fnvlist_alloc();
fnvlist_add_uint64(oxprops,
zfs_prop_to_name(ZFS_PROP_ENCRYPTION), ZIO_CRYPT_OFF);
}
if (flags->dryrun) {
void *buf = zfs_alloc(hdl, SPA_MAXBLOCKSIZE);
/*
* We have read the DRR_BEGIN record, but we have
* not yet read the payload. For non-dryrun sends
* this will be done by the kernel, so we must
* emulate that here, before attempting to read
* more records.
*/
err = recv_read(hdl, infd, buf, drr->drr_payloadlen,
flags->byteswap, NULL);
free(buf);
if (err != 0)
goto out;
err = recv_skip(hdl, infd, flags->byteswap);
goto out;
}
if (flags->heal) {
err = ioctl_err = lzc_receive_with_heal(destsnap, rcvprops,
oxprops, wkeydata, wkeylen, origin, flags->force,
flags->heal, flags->resumable, raw, infd, drr_noswap, -1,
&read_bytes, &errflags, NULL, &prop_errors);
} else {
err = ioctl_err = lzc_receive_with_cmdprops(destsnap, rcvprops,
oxprops, wkeydata, wkeylen, origin, flags->force,
flags->resumable, raw, infd, drr_noswap, -1, &read_bytes,
&errflags, NULL, &prop_errors);
}
ioctl_errno = ioctl_err;
prop_errflags = errflags;
if (err == 0) {
nvpair_t *prop_err = NULL;
while ((prop_err = nvlist_next_nvpair(prop_errors,
prop_err)) != NULL) {
char tbuf[1024];
zfs_prop_t prop;
int intval;
prop = zfs_name_to_prop(nvpair_name(prop_err));
(void) nvpair_value_int32(prop_err, &intval);
if (strcmp(nvpair_name(prop_err),
ZPROP_N_MORE_ERRORS) == 0) {
trunc_prop_errs(intval);
break;
} else if (snapname == NULL || finalsnap == NULL ||
strcmp(finalsnap, snapname) == 0 ||
strcmp(nvpair_name(prop_err),
zfs_prop_to_name(ZFS_PROP_REFQUOTA)) != 0) {
/*
* Skip the special case of, for example,
* "refquota", errors on intermediate
* snapshots leading up to a final one.
* That's why we have all of the checks above.
*
* See zfs_ioctl.c's extract_delay_props() for
* a list of props which can fail on
* intermediate snapshots, but shouldn't
* affect the overall receive.
*/
(void) snprintf(tbuf, sizeof (tbuf),
dgettext(TEXT_DOMAIN,
"cannot receive %s property on %s"),
nvpair_name(prop_err), name);
zfs_setprop_error(hdl, prop, intval, tbuf);
}
}
}
if (err == 0 && snapprops_nvlist) {
zfs_cmd_t zc = {"\0"};
(void) strlcpy(zc.zc_name, destsnap, sizeof (zc.zc_name));
zc.zc_cookie = B_TRUE; /* received */
zcmd_write_src_nvlist(hdl, &zc, snapprops_nvlist);
(void) zfs_ioctl(hdl, ZFS_IOC_SET_PROP, &zc);
zcmd_free_nvlists(&zc);
}
if (err == 0 && snapholds_nvlist) {
nvpair_t *pair;
nvlist_t *holds, *errors = NULL;
int cleanup_fd = -1;
VERIFY(0 == nvlist_alloc(&holds, 0, KM_SLEEP));
for (pair = nvlist_next_nvpair(snapholds_nvlist, NULL);
pair != NULL;
pair = nvlist_next_nvpair(snapholds_nvlist, pair)) {
fnvlist_add_string(holds, destsnap, nvpair_name(pair));
}
(void) lzc_hold(holds, cleanup_fd, &errors);
fnvlist_free(snapholds_nvlist);
fnvlist_free(holds);
}
if (err && (ioctl_errno == ENOENT || ioctl_errno == EEXIST)) {
/*
* It may be that this snapshot already exists,
* in which case we want to consume & ignore it
* rather than failing.
*/
avl_tree_t *local_avl;
nvlist_t *local_nv, *fs;
cp = strchr(destsnap, '@');
/*
* XXX Do this faster by just iterating over snaps in
* this fs. Also if zc_value does not exist, we will
* get a strange "does not exist" error message.
*/
*cp = '\0';
if (gather_nvlist(hdl, destsnap, NULL, NULL, B_FALSE, B_TRUE,
B_FALSE, B_FALSE, B_FALSE, B_FALSE, B_FALSE, B_FALSE,
B_TRUE, &local_nv, &local_avl) == 0) {
*cp = '@';
fs = fsavl_find(local_avl, drrb->drr_toguid, NULL);
fsavl_destroy(local_avl);
fnvlist_free(local_nv);
if (fs != NULL) {
if (flags->verbose) {
(void) printf("snap %s already exists; "
"ignoring\n", destsnap);
}
err = ioctl_err = recv_skip(hdl, infd,
flags->byteswap);
}
}
*cp = '@';
}
if (ioctl_err != 0) {
switch (ioctl_errno) {
case ENODEV:
cp = strchr(destsnap, '@');
*cp = '\0';
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"most recent snapshot of %s does not\n"
"match incremental source"), destsnap);
(void) zfs_error(hdl, EZFS_BADRESTORE, errbuf);
*cp = '@';
break;
case ETXTBSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination %s has been modified\n"
"since most recent snapshot"), name);
(void) zfs_error(hdl, EZFS_BADRESTORE, errbuf);
break;
case EACCES:
if (flags->heal) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"key must be loaded to do a non-raw "
"corrective recv on an encrypted "
"dataset."));
} else if (raw && stream_wantsnewfs) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"failed to create encryption key"));
} else if (raw && !stream_wantsnewfs) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"encryption key does not match "
"existing key"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"inherited key must be loaded"));
}
(void) zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf);
break;
case EEXIST:
cp = strchr(destsnap, '@');
if (newfs) {
/* it's the containing fs that exists */
*cp = '\0';
}
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination already exists"));
(void) zfs_error_fmt(hdl, EZFS_EXISTS,
dgettext(TEXT_DOMAIN, "cannot restore to %s"),
destsnap);
*cp = '@';
break;
case EINVAL:
if (embedded && !raw) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incompatible embedded data stream "
"feature with encrypted receive."));
} else if (flags->resumable) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"kernel modules must be upgraded to "
"receive this stream."));
}
(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
break;
case ECKSUM:
case ZFS_ERR_STREAM_TRUNCATED:
if (flags->heal)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"corrective receive was not able to "
"reconstruct the data needed for "
"healing."));
else
recv_ecksum_set_aux(hdl, destsnap,
flags->resumable, ioctl_err == ECKSUM);
(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
break;
case ZFS_ERR_STREAM_LARGE_BLOCK_MISMATCH:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"incremental send stream requires -L "
"(--large-block), to match previous receive."));
(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
break;
case ENOTSUP:
if (flags->heal)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"stream is not compatible with the "
"data in the pool."));
else
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool must be upgraded to receive this "
"stream."));
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
case ZFS_ERR_CRYPTO_NOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"stream uses crypto parameters not compatible with "
"this pool"));
(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
break;
case EDQUOT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination %s space quota exceeded."), name);
(void) zfs_error(hdl, EZFS_NOSPC, errbuf);
break;
case ZFS_ERR_FROM_IVSET_GUID_MISSING:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"IV set guid missing. See errata %u at "
"https://openzfs.github.io/openzfs-docs/msg/"
"ZFS-8000-ER."),
ZPOOL_ERRATA_ZOL_8308_ENCRYPTION);
(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
break;
case ZFS_ERR_FROM_IVSET_GUID_MISMATCH:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"IV set guid mismatch. See the 'zfs receive' "
"man page section\n discussing the limitations "
"of raw encrypted send streams."));
(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
break;
case ZFS_ERR_SPILL_BLOCK_FLAG_MISSING:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Spill block flag missing for raw send.\n"
"The zfs software on the sending system must "
"be updated."));
(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
break;
case ZFS_ERR_RESUME_EXISTS:
cp = strchr(destsnap, '@');
if (newfs) {
/* it's the containing fs that exists */
*cp = '\0';
}
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"Resuming recv on existing dataset without force"));
(void) zfs_error_fmt(hdl, EZFS_RESUME_EXISTS,
dgettext(TEXT_DOMAIN, "cannot resume recv %s"),
destsnap);
*cp = '@';
break;
case E2BIG:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"zfs receive required kernel memory allocation "
"larger than the system can support. Please file "
"an issue at the OpenZFS issue tracker:\n"
"https://github.com/openzfs/zfs/issues/new"));
(void) zfs_error(hdl, EZFS_BADSTREAM, errbuf);
break;
case EBUSY:
if (hastoken) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"destination %s contains "
"partially-complete state from "
"\"zfs receive -s\"."), name);
(void) zfs_error(hdl, EZFS_BUSY, errbuf);
break;
}
zfs_fallthrough;
default:
(void) zfs_standard_error(hdl, ioctl_errno, errbuf);
}
}
/*
* Mount the target filesystem (if created). Also mount any
* children of the target filesystem if we did a replication
* receive (indicated by stream_avl being non-NULL).
*/
if (clp) {
if (!flags->nomount)
err |= changelist_postfix(clp);
changelist_free(clp);
}
if ((newfs || stream_avl) && type == ZFS_TYPE_FILESYSTEM && !redacted)
flags->domount = B_TRUE;
if (prop_errflags & ZPROP_ERR_NOCLEAR) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN, "Warning: "
"failed to clear unreceived properties on %s"), name);
(void) fprintf(stderr, "\n");
}
if (prop_errflags & ZPROP_ERR_NORESTORE) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN, "Warning: "
"failed to restore original properties on %s"), name);
(void) fprintf(stderr, "\n");
}
if (err || ioctl_err) {
err = -1;
goto out;
}
if (flags->verbose) {
char buf1[64];
char buf2[64];
uint64_t bytes = read_bytes;
struct timespec delta;
clock_gettime(CLOCK_MONOTONIC_RAW, &delta);
if (begin_time.tv_nsec > delta.tv_nsec) {
delta.tv_nsec =
1000000000 + delta.tv_nsec - begin_time.tv_nsec;
delta.tv_sec -= 1;
} else
delta.tv_nsec -= begin_time.tv_nsec;
delta.tv_sec -= begin_time.tv_sec;
if (delta.tv_sec == 0 && delta.tv_nsec == 0)
delta.tv_nsec = 1;
double delta_f = delta.tv_sec + (delta.tv_nsec / 1e9);
zfs_nicebytes(bytes, buf1, sizeof (buf1));
zfs_nicebytes(bytes / delta_f, buf2, sizeof (buf2));
(void) printf("received %s stream in %.2f seconds (%s/sec)\n",
buf1, delta_f, buf2);
}
err = 0;
out:
if (prop_errors != NULL)
fnvlist_free(prop_errors);
if (tmp_keylocation[0] != '\0') {
fnvlist_add_string(rcvprops,
zfs_prop_to_name(ZFS_PROP_KEYLOCATION), tmp_keylocation);
}
if (newprops)
fnvlist_free(rcvprops);
fnvlist_free(oxprops);
fnvlist_free(origprops);
return (err);
}
/*
* Check properties we were asked to override (both -o|-x)
*/
static boolean_t
zfs_receive_checkprops(libzfs_handle_t *hdl, nvlist_t *props,
const char *errbuf)
{
nvpair_t *nvp = NULL;
zfs_prop_t prop;
const char *name;
while ((nvp = nvlist_next_nvpair(props, nvp)) != NULL) {
name = nvpair_name(nvp);
prop = zfs_name_to_prop(name);
if (prop == ZPROP_USERPROP) {
if (!zfs_prop_user(name)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"%s: invalid property '%s'"), errbuf, name);
return (B_FALSE);
}
continue;
}
/*
* "origin" is readonly but is used to receive datasets as
* clones so we don't raise an error here
*/
if (prop == ZFS_PROP_ORIGIN)
continue;
/* encryption params have their own verification later */
if (prop == ZFS_PROP_ENCRYPTION ||
zfs_prop_encryption_key_param(prop))
continue;
/*
* cannot override readonly, set-once and other specific
* settable properties
*/
if (zfs_prop_readonly(prop) || prop == ZFS_PROP_VERSION ||
prop == ZFS_PROP_VOLSIZE) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"%s: invalid property '%s'"), errbuf, name);
return (B_FALSE);
}
}
return (B_TRUE);
}
static int
zfs_receive_impl(libzfs_handle_t *hdl, const char *tosnap,
const char *originsnap, recvflags_t *flags, int infd, const char *sendfs,
nvlist_t *stream_nv, avl_tree_t *stream_avl, char **top_zfs,
const char *finalsnap, nvlist_t *cmdprops)
{
int err;
dmu_replay_record_t drr, drr_noswap;
struct drr_begin *drrb = &drr.drr_u.drr_begin;
char errbuf[ERRBUFLEN];
zio_cksum_t zcksum = { { 0 } };
uint64_t featureflags;
int hdrtype;
(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,
"cannot receive"));
/* check cmdline props, raise an error if they cannot be received */
if (!zfs_receive_checkprops(hdl, cmdprops, errbuf))
return (zfs_error(hdl, EZFS_BADPROP, errbuf));
if (flags->isprefix &&
!zfs_dataset_exists(hdl, tosnap, ZFS_TYPE_DATASET)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "specified fs "
"(%s) does not exist"), tosnap);
return (zfs_error(hdl, EZFS_NOENT, errbuf));
}
if (originsnap &&
!zfs_dataset_exists(hdl, originsnap, ZFS_TYPE_DATASET)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "specified origin fs "
"(%s) does not exist"), originsnap);
return (zfs_error(hdl, EZFS_NOENT, errbuf));
}
/* read in the BEGIN record */
if (0 != (err = recv_read(hdl, infd, &drr, sizeof (drr), B_FALSE,
&zcksum)))
return (err);
if (drr.drr_type == DRR_END || drr.drr_type == BSWAP_32(DRR_END)) {
/* It's the double end record at the end of a package */
return (ENODATA);
}
/* the kernel needs the non-byteswapped begin record */
drr_noswap = drr;
flags->byteswap = B_FALSE;
if (drrb->drr_magic == BSWAP_64(DMU_BACKUP_MAGIC)) {
/*
* We computed the checksum in the wrong byteorder in
* recv_read() above; do it again correctly.
*/
memset(&zcksum, 0, sizeof (zio_cksum_t));
fletcher_4_incremental_byteswap(&drr, sizeof (drr), &zcksum);
flags->byteswap = B_TRUE;
drr.drr_type = BSWAP_32(drr.drr_type);
drr.drr_payloadlen = BSWAP_32(drr.drr_payloadlen);
drrb->drr_magic = BSWAP_64(drrb->drr_magic);
drrb->drr_versioninfo = BSWAP_64(drrb->drr_versioninfo);
drrb->drr_creation_time = BSWAP_64(drrb->drr_creation_time);
drrb->drr_type = BSWAP_32(drrb->drr_type);
drrb->drr_flags = BSWAP_32(drrb->drr_flags);
drrb->drr_toguid = BSWAP_64(drrb->drr_toguid);
drrb->drr_fromguid = BSWAP_64(drrb->drr_fromguid);
}
if (drrb->drr_magic != DMU_BACKUP_MAGIC || drr.drr_type != DRR_BEGIN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
"stream (bad magic number)"));
return (zfs_error(hdl, EZFS_BADSTREAM, errbuf));
}
featureflags = DMU_GET_FEATUREFLAGS(drrb->drr_versioninfo);
hdrtype = DMU_GET_STREAM_HDRTYPE(drrb->drr_versioninfo);
if (!DMU_STREAM_SUPPORTED(featureflags) ||
(hdrtype != DMU_SUBSTREAM && hdrtype != DMU_COMPOUNDSTREAM)) {
/*
* Let's be explicit about this one, since rather than
* being a new feature we can't know, it's an old
* feature we dropped.
*/
if (featureflags & DMU_BACKUP_FEATURE_DEDUP) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"stream has deprecated feature: dedup, try "
"'zstream redup [send in a file] | zfs recv "
"[...]'"));
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"stream has unsupported feature, feature flags = "
"%llx (unknown flags = %llx)"),
(u_longlong_t)featureflags,
(u_longlong_t)((featureflags) &
~DMU_BACKUP_FEATURE_MASK));
}
return (zfs_error(hdl, EZFS_BADSTREAM, errbuf));
}
/* Holds feature is set once in the compound stream header. */
if (featureflags & DMU_BACKUP_FEATURE_HOLDS)
flags->holds = B_TRUE;
if (strchr(drrb->drr_toname, '@') == NULL) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid "
"stream (bad snapshot name)"));
return (zfs_error(hdl, EZFS_BADSTREAM, errbuf));
}
if (DMU_GET_STREAM_HDRTYPE(drrb->drr_versioninfo) == DMU_SUBSTREAM) {
char nonpackage_sendfs[ZFS_MAX_DATASET_NAME_LEN];
if (sendfs == NULL) {
/*
* We were not called from zfs_receive_package(). Get
* the fs specified by 'zfs send'.
*/
char *cp;
(void) strlcpy(nonpackage_sendfs,
drr.drr_u.drr_begin.drr_toname,
sizeof (nonpackage_sendfs));
if ((cp = strchr(nonpackage_sendfs, '@')) != NULL)
*cp = '\0';
sendfs = nonpackage_sendfs;
VERIFY(finalsnap == NULL);
}
return (zfs_receive_one(hdl, infd, tosnap, originsnap, flags,
&drr, &drr_noswap, sendfs, stream_nv, stream_avl, top_zfs,
finalsnap, cmdprops));
} else {
assert(DMU_GET_STREAM_HDRTYPE(drrb->drr_versioninfo) ==
DMU_COMPOUNDSTREAM);
return (zfs_receive_package(hdl, infd, tosnap, flags, &drr,
&zcksum, top_zfs, cmdprops));
}
}
/*
* Restores a backup of tosnap from the file descriptor specified by infd.
* Return 0 on total success, -2 if some things couldn't be
* destroyed/renamed/promoted, -1 if some things couldn't be received.
* (-1 will override -2, if -1 and the resumable flag was specified the
* transfer can be resumed if the sending side supports it).
*/
int
zfs_receive(libzfs_handle_t *hdl, const char *tosnap, nvlist_t *props,
recvflags_t *flags, int infd, avl_tree_t *stream_avl)
{
char *top_zfs = NULL;
int err;
struct stat sb;
const char *originsnap = NULL;
/*
* The only way fstat can fail is if we do not have a valid file
* descriptor.
*/
if (fstat(infd, &sb) == -1) {
perror("fstat");
return (-2);
}
if (props) {
err = nvlist_lookup_string(props, "origin", &originsnap);
if (err && err != ENOENT)
return (err);
}
err = zfs_receive_impl(hdl, tosnap, originsnap, flags, infd, NULL, NULL,
stream_avl, &top_zfs, NULL, props);
if (err == 0 && !flags->nomount && flags->domount && top_zfs) {
zfs_handle_t *zhp = NULL;
prop_changelist_t *clp = NULL;
zhp = zfs_open(hdl, top_zfs,
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME);
if (zhp == NULL) {
err = -1;
goto out;
} else {
if (zhp->zfs_type == ZFS_TYPE_VOLUME) {
zfs_close(zhp);
goto out;
}
clp = changelist_gather(zhp, ZFS_PROP_MOUNTPOINT,
CL_GATHER_MOUNT_ALWAYS,
flags->forceunmount ? MS_FORCE : 0);
zfs_close(zhp);
if (clp == NULL) {
err = -1;
goto out;
}
/* mount and share received datasets */
err = changelist_postfix(clp);
changelist_free(clp);
if (err != 0)
err = -1;
}
}
out:
if (top_zfs)
free(top_zfs);
return (err);
}
diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_util.c b/sys/contrib/openzfs/lib/libzfs/libzfs_util.c
index 5592332067f9..4b8a20160e02 100644
--- a/sys/contrib/openzfs/lib/libzfs/libzfs_util.c
+++ b/sys/contrib/openzfs/lib/libzfs/libzfs_util.c
@@ -1,2063 +1,2063 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2020 Joyent, Inc. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright 2016 Igor Kozhukhov <ikozhukhov@gmail.com>
* Copyright (c) 2017 Datto Inc.
* Copyright (c) 2020 The FreeBSD Foundation
*
* Portions of this software were developed by Allan Jude
* under sponsorship from the FreeBSD Foundation.
*/
/*
* Internal utility routines for the ZFS library.
*/
#include <errno.h>
#include <fcntl.h>
#include <libintl.h>
#include <stdarg.h>
#include <stdio.h>
#include <stdlib.h>
#include <strings.h>
#include <unistd.h>
#include <math.h>
#if LIBFETCH_DYNAMIC
#include <dlfcn.h>
#endif
#include <sys/stat.h>
#include <sys/mnttab.h>
#include <sys/mntent.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <libzfs.h>
#include <libzfs_core.h>
#include "libzfs_impl.h"
#include "zfs_prop.h"
#include "zfeature_common.h"
#include <zfs_fletcher.h>
#include <libzutil.h>
/*
* We only care about the scheme in order to match the scheme
* with the handler. Each handler should validate the full URI
* as necessary.
*/
#define URI_REGEX "^\\([A-Za-z][A-Za-z0-9+.\\-]*\\):"
int
libzfs_errno(libzfs_handle_t *hdl)
{
return (hdl->libzfs_error);
}
const char *
libzfs_error_action(libzfs_handle_t *hdl)
{
return (hdl->libzfs_action);
}
const char *
libzfs_error_description(libzfs_handle_t *hdl)
{
if (hdl->libzfs_desc[0] != '\0')
return (hdl->libzfs_desc);
switch (hdl->libzfs_error) {
case EZFS_NOMEM:
return (dgettext(TEXT_DOMAIN, "out of memory"));
case EZFS_BADPROP:
return (dgettext(TEXT_DOMAIN, "invalid property value"));
case EZFS_PROPREADONLY:
return (dgettext(TEXT_DOMAIN, "read-only property"));
case EZFS_PROPTYPE:
return (dgettext(TEXT_DOMAIN, "property doesn't apply to "
"datasets of this type"));
case EZFS_PROPNONINHERIT:
return (dgettext(TEXT_DOMAIN, "property cannot be inherited"));
case EZFS_PROPSPACE:
return (dgettext(TEXT_DOMAIN, "invalid quota or reservation"));
case EZFS_BADTYPE:
return (dgettext(TEXT_DOMAIN, "operation not applicable to "
"datasets of this type"));
case EZFS_BUSY:
return (dgettext(TEXT_DOMAIN, "pool or dataset is busy"));
case EZFS_EXISTS:
return (dgettext(TEXT_DOMAIN, "pool or dataset exists"));
case EZFS_NOENT:
return (dgettext(TEXT_DOMAIN, "no such pool or dataset"));
case EZFS_BADSTREAM:
return (dgettext(TEXT_DOMAIN, "invalid backup stream"));
case EZFS_DSREADONLY:
return (dgettext(TEXT_DOMAIN, "dataset is read-only"));
case EZFS_VOLTOOBIG:
return (dgettext(TEXT_DOMAIN, "volume size exceeds limit for "
"this system"));
case EZFS_INVALIDNAME:
return (dgettext(TEXT_DOMAIN, "invalid name"));
case EZFS_BADRESTORE:
return (dgettext(TEXT_DOMAIN, "unable to restore to "
"destination"));
case EZFS_BADBACKUP:
return (dgettext(TEXT_DOMAIN, "backup failed"));
case EZFS_BADTARGET:
return (dgettext(TEXT_DOMAIN, "invalid target vdev"));
case EZFS_NODEVICE:
return (dgettext(TEXT_DOMAIN, "no such device in pool"));
case EZFS_BADDEV:
return (dgettext(TEXT_DOMAIN, "invalid device"));
case EZFS_NOREPLICAS:
return (dgettext(TEXT_DOMAIN, "no valid replicas"));
case EZFS_RESILVERING:
return (dgettext(TEXT_DOMAIN, "currently resilvering"));
case EZFS_BADVERSION:
return (dgettext(TEXT_DOMAIN, "unsupported version or "
"feature"));
case EZFS_POOLUNAVAIL:
return (dgettext(TEXT_DOMAIN, "pool is unavailable"));
case EZFS_DEVOVERFLOW:
return (dgettext(TEXT_DOMAIN, "too many devices in one vdev"));
case EZFS_BADPATH:
return (dgettext(TEXT_DOMAIN, "must be an absolute path"));
case EZFS_CROSSTARGET:
return (dgettext(TEXT_DOMAIN, "operation crosses datasets or "
"pools"));
case EZFS_ZONED:
return (dgettext(TEXT_DOMAIN, "dataset in use by local zone"));
case EZFS_MOUNTFAILED:
return (dgettext(TEXT_DOMAIN, "mount failed"));
case EZFS_UMOUNTFAILED:
return (dgettext(TEXT_DOMAIN, "unmount failed"));
case EZFS_UNSHARENFSFAILED:
return (dgettext(TEXT_DOMAIN, "NFS share removal failed"));
case EZFS_SHARENFSFAILED:
return (dgettext(TEXT_DOMAIN, "NFS share creation failed"));
case EZFS_UNSHARESMBFAILED:
return (dgettext(TEXT_DOMAIN, "SMB share removal failed"));
case EZFS_SHARESMBFAILED:
return (dgettext(TEXT_DOMAIN, "SMB share creation failed"));
case EZFS_PERM:
return (dgettext(TEXT_DOMAIN, "permission denied"));
case EZFS_NOSPC:
return (dgettext(TEXT_DOMAIN, "out of space"));
case EZFS_FAULT:
return (dgettext(TEXT_DOMAIN, "bad address"));
case EZFS_IO:
return (dgettext(TEXT_DOMAIN, "I/O error"));
case EZFS_INTR:
return (dgettext(TEXT_DOMAIN, "signal received"));
case EZFS_CKSUM:
return (dgettext(TEXT_DOMAIN, "insufficient replicas"));
case EZFS_ISSPARE:
return (dgettext(TEXT_DOMAIN, "device is reserved as a hot "
"spare"));
case EZFS_INVALCONFIG:
return (dgettext(TEXT_DOMAIN, "invalid vdev configuration"));
case EZFS_RECURSIVE:
return (dgettext(TEXT_DOMAIN, "recursive dataset dependency"));
case EZFS_NOHISTORY:
return (dgettext(TEXT_DOMAIN, "no history available"));
case EZFS_POOLPROPS:
return (dgettext(TEXT_DOMAIN, "failed to retrieve "
"pool properties"));
case EZFS_POOL_NOTSUP:
return (dgettext(TEXT_DOMAIN, "operation not supported "
"on this type of pool"));
case EZFS_POOL_INVALARG:
return (dgettext(TEXT_DOMAIN, "invalid argument for "
"this pool operation"));
case EZFS_NAMETOOLONG:
return (dgettext(TEXT_DOMAIN, "dataset name is too long"));
case EZFS_OPENFAILED:
return (dgettext(TEXT_DOMAIN, "open failed"));
case EZFS_NOCAP:
return (dgettext(TEXT_DOMAIN,
"disk capacity information could not be retrieved"));
case EZFS_LABELFAILED:
return (dgettext(TEXT_DOMAIN, "write of label failed"));
case EZFS_BADWHO:
return (dgettext(TEXT_DOMAIN, "invalid user/group"));
case EZFS_BADPERM:
return (dgettext(TEXT_DOMAIN, "invalid permission"));
case EZFS_BADPERMSET:
return (dgettext(TEXT_DOMAIN, "invalid permission set name"));
case EZFS_NODELEGATION:
return (dgettext(TEXT_DOMAIN, "delegated administration is "
"disabled on pool"));
case EZFS_BADCACHE:
return (dgettext(TEXT_DOMAIN, "invalid or missing cache file"));
case EZFS_ISL2CACHE:
return (dgettext(TEXT_DOMAIN, "device is in use as a cache"));
case EZFS_VDEVNOTSUP:
return (dgettext(TEXT_DOMAIN, "vdev specification is not "
"supported"));
case EZFS_NOTSUP:
return (dgettext(TEXT_DOMAIN, "operation not supported "
"on this dataset"));
case EZFS_IOC_NOTSUPPORTED:
return (dgettext(TEXT_DOMAIN, "operation not supported by "
"zfs kernel module"));
case EZFS_ACTIVE_SPARE:
return (dgettext(TEXT_DOMAIN, "pool has active shared spare "
"device"));
case EZFS_UNPLAYED_LOGS:
return (dgettext(TEXT_DOMAIN, "log device has unplayed intent "
"logs"));
case EZFS_REFTAG_RELE:
return (dgettext(TEXT_DOMAIN, "no such tag on this dataset"));
case EZFS_REFTAG_HOLD:
return (dgettext(TEXT_DOMAIN, "tag already exists on this "
"dataset"));
case EZFS_TAGTOOLONG:
return (dgettext(TEXT_DOMAIN, "tag too long"));
case EZFS_PIPEFAILED:
return (dgettext(TEXT_DOMAIN, "pipe create failed"));
case EZFS_THREADCREATEFAILED:
return (dgettext(TEXT_DOMAIN, "thread create failed"));
case EZFS_POSTSPLIT_ONLINE:
return (dgettext(TEXT_DOMAIN, "disk was split from this pool "
"into a new one"));
case EZFS_SCRUB_PAUSED:
return (dgettext(TEXT_DOMAIN, "scrub is paused; "
"use 'zpool scrub' to resume"));
case EZFS_SCRUBBING:
return (dgettext(TEXT_DOMAIN, "currently scrubbing; "
"use 'zpool scrub -s' to cancel current scrub"));
case EZFS_NO_SCRUB:
return (dgettext(TEXT_DOMAIN, "there is no active scrub"));
case EZFS_DIFF:
return (dgettext(TEXT_DOMAIN, "unable to generate diffs"));
case EZFS_DIFFDATA:
return (dgettext(TEXT_DOMAIN, "invalid diff data"));
case EZFS_POOLREADONLY:
return (dgettext(TEXT_DOMAIN, "pool is read-only"));
case EZFS_NO_PENDING:
return (dgettext(TEXT_DOMAIN, "operation is not "
"in progress"));
case EZFS_CHECKPOINT_EXISTS:
return (dgettext(TEXT_DOMAIN, "checkpoint exists"));
case EZFS_DISCARDING_CHECKPOINT:
return (dgettext(TEXT_DOMAIN, "currently discarding "
"checkpoint"));
case EZFS_NO_CHECKPOINT:
return (dgettext(TEXT_DOMAIN, "checkpoint does not exist"));
case EZFS_DEVRM_IN_PROGRESS:
return (dgettext(TEXT_DOMAIN, "device removal in progress"));
case EZFS_VDEV_TOO_BIG:
return (dgettext(TEXT_DOMAIN, "device exceeds supported size"));
case EZFS_ACTIVE_POOL:
return (dgettext(TEXT_DOMAIN, "pool is imported on a "
"different host"));
case EZFS_CRYPTOFAILED:
return (dgettext(TEXT_DOMAIN, "encryption failure"));
case EZFS_TOOMANY:
return (dgettext(TEXT_DOMAIN, "argument list too long"));
case EZFS_INITIALIZING:
return (dgettext(TEXT_DOMAIN, "currently initializing"));
case EZFS_NO_INITIALIZE:
return (dgettext(TEXT_DOMAIN, "there is no active "
"initialization"));
case EZFS_WRONG_PARENT:
return (dgettext(TEXT_DOMAIN, "invalid parent dataset"));
case EZFS_TRIMMING:
return (dgettext(TEXT_DOMAIN, "currently trimming"));
case EZFS_NO_TRIM:
return (dgettext(TEXT_DOMAIN, "there is no active trim"));
case EZFS_TRIM_NOTSUP:
return (dgettext(TEXT_DOMAIN, "trim operations are not "
"supported by this device"));
case EZFS_NO_RESILVER_DEFER:
return (dgettext(TEXT_DOMAIN, "this action requires the "
"resilver_defer feature"));
case EZFS_EXPORT_IN_PROGRESS:
return (dgettext(TEXT_DOMAIN, "pool export in progress"));
case EZFS_REBUILDING:
return (dgettext(TEXT_DOMAIN, "currently sequentially "
"resilvering"));
case EZFS_VDEV_NOTSUP:
return (dgettext(TEXT_DOMAIN, "operation not supported "
"on this type of vdev"));
case EZFS_NOT_USER_NAMESPACE:
return (dgettext(TEXT_DOMAIN, "the provided file "
"was not a user namespace file"));
case EZFS_RESUME_EXISTS:
return (dgettext(TEXT_DOMAIN, "Resuming recv on existing "
"dataset without force"));
case EZFS_UNKNOWN:
return (dgettext(TEXT_DOMAIN, "unknown error"));
default:
assert(hdl->libzfs_error == 0);
return (dgettext(TEXT_DOMAIN, "no error"));
}
}
void
zfs_error_aux(libzfs_handle_t *hdl, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
(void) vsnprintf(hdl->libzfs_desc, sizeof (hdl->libzfs_desc),
fmt, ap);
hdl->libzfs_desc_active = 1;
va_end(ap);
}
static void
zfs_verror(libzfs_handle_t *hdl, int error, const char *fmt, va_list ap)
{
(void) vsnprintf(hdl->libzfs_action, sizeof (hdl->libzfs_action),
fmt, ap);
hdl->libzfs_error = error;
if (hdl->libzfs_desc_active)
hdl->libzfs_desc_active = 0;
else
hdl->libzfs_desc[0] = '\0';
if (hdl->libzfs_printerr) {
if (error == EZFS_UNKNOWN) {
(void) fprintf(stderr, dgettext(TEXT_DOMAIN, "internal "
"error: %s: %s\n"), hdl->libzfs_action,
libzfs_error_description(hdl));
abort();
}
(void) fprintf(stderr, "%s: %s\n", hdl->libzfs_action,
libzfs_error_description(hdl));
if (error == EZFS_NOMEM)
exit(1);
}
}
int
zfs_error(libzfs_handle_t *hdl, int error, const char *msg)
{
return (zfs_error_fmt(hdl, error, "%s", msg));
}
int
zfs_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
zfs_verror(hdl, error, fmt, ap);
va_end(ap);
return (-1);
}
static int
zfs_common_error(libzfs_handle_t *hdl, int error, const char *fmt,
va_list ap)
{
switch (error) {
case EPERM:
case EACCES:
zfs_verror(hdl, EZFS_PERM, fmt, ap);
return (-1);
case ECANCELED:
zfs_verror(hdl, EZFS_NODELEGATION, fmt, ap);
return (-1);
case EIO:
zfs_verror(hdl, EZFS_IO, fmt, ap);
return (-1);
case EFAULT:
zfs_verror(hdl, EZFS_FAULT, fmt, ap);
return (-1);
case EINTR:
zfs_verror(hdl, EZFS_INTR, fmt, ap);
return (-1);
case ECKSUM:
zfs_verror(hdl, EZFS_CKSUM, fmt, ap);
return (-1);
}
return (0);
}
int
zfs_standard_error(libzfs_handle_t *hdl, int error, const char *msg)
{
return (zfs_standard_error_fmt(hdl, error, "%s", msg));
}
int
zfs_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
if (zfs_common_error(hdl, error, fmt, ap) != 0) {
va_end(ap);
return (-1);
}
switch (error) {
case ENXIO:
case ENODEV:
case EPIPE:
zfs_verror(hdl, EZFS_IO, fmt, ap);
break;
case ENOENT:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset does not exist"));
zfs_verror(hdl, EZFS_NOENT, fmt, ap);
break;
case ENOSPC:
case EDQUOT:
zfs_verror(hdl, EZFS_NOSPC, fmt, ap);
break;
case EEXIST:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset already exists"));
zfs_verror(hdl, EZFS_EXISTS, fmt, ap);
break;
case EBUSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"dataset is busy"));
zfs_verror(hdl, EZFS_BUSY, fmt, ap);
break;
case EROFS:
zfs_verror(hdl, EZFS_POOLREADONLY, fmt, ap);
break;
case ENAMETOOLONG:
zfs_verror(hdl, EZFS_NAMETOOLONG, fmt, ap);
break;
case ENOTSUP:
zfs_verror(hdl, EZFS_BADVERSION, fmt, ap);
break;
case EAGAIN:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool I/O is currently suspended"));
zfs_verror(hdl, EZFS_POOLUNAVAIL, fmt, ap);
break;
case EREMOTEIO:
zfs_verror(hdl, EZFS_ACTIVE_POOL, fmt, ap);
break;
case ZFS_ERR_UNKNOWN_SEND_STREAM_FEATURE:
case ZFS_ERR_IOC_CMD_UNAVAIL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
"module does not support this operation. A reboot may "
"be required to enable this operation."));
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_IOC_ARG_UNAVAIL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
"module does not support an option for this operation. "
"A reboot may be required to enable this option."));
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_IOC_ARG_REQUIRED:
case ZFS_ERR_IOC_ARG_BADTYPE:
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_WRONG_PARENT:
zfs_verror(hdl, EZFS_WRONG_PARENT, fmt, ap);
break;
case ZFS_ERR_BADPROP:
zfs_verror(hdl, EZFS_BADPROP, fmt, ap);
break;
case ZFS_ERR_NOT_USER_NAMESPACE:
zfs_verror(hdl, EZFS_NOT_USER_NAMESPACE, fmt, ap);
break;
default:
zfs_error_aux(hdl, "%s", strerror(error));
zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap);
break;
}
va_end(ap);
return (-1);
}
void
zfs_setprop_error(libzfs_handle_t *hdl, zfs_prop_t prop, int err,
char *errbuf)
{
switch (err) {
case ENOSPC:
/*
* For quotas and reservations, ENOSPC indicates
* something different; setting a quota or reservation
* doesn't use any disk space.
*/
switch (prop) {
case ZFS_PROP_QUOTA:
case ZFS_PROP_REFQUOTA:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"size is less than current used or "
"reserved space"));
(void) zfs_error(hdl, EZFS_PROPSPACE, errbuf);
break;
case ZFS_PROP_RESERVATION:
case ZFS_PROP_REFRESERVATION:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"size is greater than available space"));
(void) zfs_error(hdl, EZFS_PROPSPACE, errbuf);
break;
default:
(void) zfs_standard_error(hdl, err, errbuf);
break;
}
break;
case EBUSY:
(void) zfs_standard_error(hdl, EBUSY, errbuf);
break;
case EROFS:
(void) zfs_error(hdl, EZFS_DSREADONLY, errbuf);
break;
case E2BIG:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property value too long"));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
break;
case ENOTSUP:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool and or dataset must be upgraded to set this "
"property or value"));
(void) zfs_error(hdl, EZFS_BADVERSION, errbuf);
break;
case ERANGE:
if (prop == ZFS_PROP_COMPRESSION ||
prop == ZFS_PROP_DNODESIZE ||
prop == ZFS_PROP_RECORDSIZE) {
(void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property setting is not allowed on "
"bootable datasets"));
(void) zfs_error(hdl, EZFS_NOTSUP, errbuf);
} else if (prop == ZFS_PROP_CHECKSUM ||
prop == ZFS_PROP_DEDUP) {
(void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"property setting is not allowed on "
"root pools"));
(void) zfs_error(hdl, EZFS_NOTSUP, errbuf);
} else {
(void) zfs_standard_error(hdl, err, errbuf);
}
break;
case EINVAL:
if (prop == ZPROP_INVAL) {
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
} else {
(void) zfs_standard_error(hdl, err, errbuf);
}
break;
case ZFS_ERR_BADPROP:
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
break;
case EACCES:
if (prop == ZFS_PROP_KEYLOCATION) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"keylocation may only be set on encryption roots"));
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
} else {
(void) zfs_standard_error(hdl, err, errbuf);
}
break;
case EOVERFLOW:
/*
* This platform can't address a volume this big.
*/
#ifdef _ILP32
if (prop == ZFS_PROP_VOLSIZE) {
(void) zfs_error(hdl, EZFS_VOLTOOBIG, errbuf);
break;
}
zfs_fallthrough;
#endif
default:
(void) zfs_standard_error(hdl, err, errbuf);
}
}
int
zpool_standard_error(libzfs_handle_t *hdl, int error, const char *msg)
{
return (zpool_standard_error_fmt(hdl, error, "%s", msg));
}
int
zpool_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
if (zfs_common_error(hdl, error, fmt, ap) != 0) {
va_end(ap);
return (-1);
}
switch (error) {
case ENODEV:
zfs_verror(hdl, EZFS_NODEVICE, fmt, ap);
break;
case ENOENT:
zfs_error_aux(hdl,
dgettext(TEXT_DOMAIN, "no such pool or dataset"));
zfs_verror(hdl, EZFS_NOENT, fmt, ap);
break;
case EEXIST:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool already exists"));
zfs_verror(hdl, EZFS_EXISTS, fmt, ap);
break;
case EBUSY:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool is busy"));
zfs_verror(hdl, EZFS_BUSY, fmt, ap);
break;
/* There is no pending operation to cancel */
case ENOTACTIVE:
zfs_verror(hdl, EZFS_NO_PENDING, fmt, ap);
break;
case ENXIO:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"one or more devices is currently unavailable"));
zfs_verror(hdl, EZFS_BADDEV, fmt, ap);
break;
case ENAMETOOLONG:
zfs_verror(hdl, EZFS_DEVOVERFLOW, fmt, ap);
break;
case ENOTSUP:
zfs_verror(hdl, EZFS_POOL_NOTSUP, fmt, ap);
break;
case EINVAL:
zfs_verror(hdl, EZFS_POOL_INVALARG, fmt, ap);
break;
case ENOSPC:
case EDQUOT:
zfs_verror(hdl, EZFS_NOSPC, fmt, ap);
break;
case EAGAIN:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"pool I/O is currently suspended"));
zfs_verror(hdl, EZFS_POOLUNAVAIL, fmt, ap);
break;
case EROFS:
zfs_verror(hdl, EZFS_POOLREADONLY, fmt, ap);
break;
case EDOM:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"block size out of range or does not match"));
zfs_verror(hdl, EZFS_BADPROP, fmt, ap);
break;
case EREMOTEIO:
zfs_verror(hdl, EZFS_ACTIVE_POOL, fmt, ap);
break;
case ZFS_ERR_CHECKPOINT_EXISTS:
zfs_verror(hdl, EZFS_CHECKPOINT_EXISTS, fmt, ap);
break;
case ZFS_ERR_DISCARDING_CHECKPOINT:
zfs_verror(hdl, EZFS_DISCARDING_CHECKPOINT, fmt, ap);
break;
case ZFS_ERR_NO_CHECKPOINT:
zfs_verror(hdl, EZFS_NO_CHECKPOINT, fmt, ap);
break;
case ZFS_ERR_DEVRM_IN_PROGRESS:
zfs_verror(hdl, EZFS_DEVRM_IN_PROGRESS, fmt, ap);
break;
case ZFS_ERR_VDEV_TOO_BIG:
zfs_verror(hdl, EZFS_VDEV_TOO_BIG, fmt, ap);
break;
case ZFS_ERR_EXPORT_IN_PROGRESS:
zfs_verror(hdl, EZFS_EXPORT_IN_PROGRESS, fmt, ap);
break;
case ZFS_ERR_RESILVER_IN_PROGRESS:
zfs_verror(hdl, EZFS_RESILVERING, fmt, ap);
break;
case ZFS_ERR_REBUILD_IN_PROGRESS:
zfs_verror(hdl, EZFS_REBUILDING, fmt, ap);
break;
case ZFS_ERR_BADPROP:
zfs_verror(hdl, EZFS_BADPROP, fmt, ap);
break;
case ZFS_ERR_VDEV_NOTSUP:
zfs_verror(hdl, EZFS_VDEV_NOTSUP, fmt, ap);
break;
case ZFS_ERR_IOC_CMD_UNAVAIL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
"module does not support this operation. A reboot may "
"be required to enable this operation."));
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_IOC_ARG_UNAVAIL:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs "
"module does not support an option for this operation. "
"A reboot may be required to enable this option."));
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
case ZFS_ERR_IOC_ARG_REQUIRED:
case ZFS_ERR_IOC_ARG_BADTYPE:
zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap);
break;
default:
zfs_error_aux(hdl, "%s", strerror(error));
zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap);
}
va_end(ap);
return (-1);
}
/*
* Display an out of memory error message and abort the current program.
*/
int
no_memory(libzfs_handle_t *hdl)
{
return (zfs_error(hdl, EZFS_NOMEM, "internal error"));
}
/*
* A safe form of malloc() which will die if the allocation fails.
*/
void *
zfs_alloc(libzfs_handle_t *hdl, size_t size)
{
void *data;
if ((data = calloc(1, size)) == NULL)
(void) no_memory(hdl);
return (data);
}
/*
* A safe form of asprintf() which will die if the allocation fails.
*/
char *
zfs_asprintf(libzfs_handle_t *hdl, const char *fmt, ...)
{
va_list ap;
char *ret;
int err;
va_start(ap, fmt);
err = vasprintf(&ret, fmt, ap);
va_end(ap);
if (err < 0) {
(void) no_memory(hdl);
ret = NULL;
}
return (ret);
}
/*
* A safe form of realloc(), which also zeroes newly allocated space.
*/
void *
zfs_realloc(libzfs_handle_t *hdl, void *ptr, size_t oldsize, size_t newsize)
{
void *ret;
if ((ret = realloc(ptr, newsize)) == NULL) {
(void) no_memory(hdl);
return (NULL);
}
memset((char *)ret + oldsize, 0, newsize - oldsize);
return (ret);
}
/*
* A safe form of strdup() which will die if the allocation fails.
*/
char *
zfs_strdup(libzfs_handle_t *hdl, const char *str)
{
char *ret;
if ((ret = strdup(str)) == NULL)
(void) no_memory(hdl);
return (ret);
}
void
libzfs_print_on_error(libzfs_handle_t *hdl, boolean_t printerr)
{
hdl->libzfs_printerr = printerr;
}
/*
* Read lines from an open file descriptor and store them in an array of
* strings until EOF. lines[] will be allocated and populated with all the
* lines read. All newlines are replaced with NULL terminators for
* convenience. lines[] must be freed after use with libzfs_free_str_array().
*
* Returns the number of lines read.
*/
static int
libzfs_read_stdout_from_fd(int fd, char **lines[])
{
FILE *fp;
int lines_cnt = 0;
size_t len = 0;
char *line = NULL;
char **tmp_lines = NULL, **tmp;
fp = fdopen(fd, "r");
if (fp == NULL) {
close(fd);
return (0);
}
while (getline(&line, &len, fp) != -1) {
tmp = realloc(tmp_lines, sizeof (*tmp_lines) * (lines_cnt + 1));
if (tmp == NULL) {
/* Return the lines we were able to process */
break;
}
tmp_lines = tmp;
/* Remove newline if not EOF */
if (line[strlen(line) - 1] == '\n')
line[strlen(line) - 1] = '\0';
tmp_lines[lines_cnt] = strdup(line);
if (tmp_lines[lines_cnt] == NULL)
break;
++lines_cnt;
}
free(line);
fclose(fp);
*lines = tmp_lines;
return (lines_cnt);
}
static int
libzfs_run_process_impl(const char *path, char *argv[], char *env[], int flags,
char **lines[], int *lines_cnt)
{
pid_t pid;
int error, devnull_fd;
int link[2];
/*
* Setup a pipe between our child and parent process if we're
* reading stdout.
*/
if (lines != NULL && pipe2(link, O_NONBLOCK | O_CLOEXEC) == -1)
return (-EPIPE);
pid = fork();
if (pid == 0) {
/* Child process */
devnull_fd = open("/dev/null", O_WRONLY | O_CLOEXEC);
if (devnull_fd < 0)
_exit(-1);
if (!(flags & STDOUT_VERBOSE) && (lines == NULL))
(void) dup2(devnull_fd, STDOUT_FILENO);
else if (lines != NULL) {
/* Save the output to lines[] */
dup2(link[1], STDOUT_FILENO);
}
if (!(flags & STDERR_VERBOSE))
(void) dup2(devnull_fd, STDERR_FILENO);
if (flags & NO_DEFAULT_PATH) {
if (env == NULL)
execv(path, argv);
else
execve(path, argv, env);
} else {
if (env == NULL)
execvp(path, argv);
else
execvpe(path, argv, env);
}
_exit(-1);
} else if (pid > 0) {
/* Parent process */
int status;
while ((error = waitpid(pid, &status, 0)) == -1 &&
errno == EINTR)
;
if (error < 0 || !WIFEXITED(status))
return (-1);
if (lines != NULL) {
close(link[1]);
*lines_cnt = libzfs_read_stdout_from_fd(link[0], lines);
}
return (WEXITSTATUS(status));
}
return (-1);
}
int
libzfs_run_process(const char *path, char *argv[], int flags)
{
return (libzfs_run_process_impl(path, argv, NULL, flags, NULL, NULL));
}
/*
* Run a command and store its stdout lines in an array of strings (lines[]).
* lines[] is allocated and populated for you, and the number of lines is set in
* lines_cnt. lines[] must be freed after use with libzfs_free_str_array().
* All newlines (\n) in lines[] are terminated for convenience.
*/
int
libzfs_run_process_get_stdout(const char *path, char *argv[], char *env[],
char **lines[], int *lines_cnt)
{
return (libzfs_run_process_impl(path, argv, env, 0, lines, lines_cnt));
}
/*
* Same as libzfs_run_process_get_stdout(), but run without $PATH set. This
* means that *path needs to be the full path to the executable.
*/
int
libzfs_run_process_get_stdout_nopath(const char *path, char *argv[],
char *env[], char **lines[], int *lines_cnt)
{
return (libzfs_run_process_impl(path, argv, env, NO_DEFAULT_PATH,
lines, lines_cnt));
}
/*
* Free an array of strings. Free both the strings contained in the array and
* the array itself.
*/
void
libzfs_free_str_array(char **strs, int count)
{
while (--count >= 0)
free(strs[count]);
free(strs);
}
/*
* Returns 1 if environment variable is set to "YES", "yes", "ON", "on", or
* a non-zero number.
*
* Returns 0 otherwise.
*/
boolean_t
libzfs_envvar_is_set(const char *envvar)
{
char *env = getenv(envvar);
return (env && (strtoul(env, NULL, 0) > 0 ||
(!strncasecmp(env, "YES", 3) && strnlen(env, 4) == 3) ||
(!strncasecmp(env, "ON", 2) && strnlen(env, 3) == 2)));
}
libzfs_handle_t *
libzfs_init(void)
{
libzfs_handle_t *hdl;
int error;
char *env;
if ((error = libzfs_load_module()) != 0) {
errno = error;
return (NULL);
}
if ((hdl = calloc(1, sizeof (libzfs_handle_t))) == NULL) {
return (NULL);
}
if (regcomp(&hdl->libzfs_urire, URI_REGEX, 0) != 0) {
free(hdl);
return (NULL);
}
if ((hdl->libzfs_fd = open(ZFS_DEV, O_RDWR|O_EXCL|O_CLOEXEC)) < 0) {
free(hdl);
return (NULL);
}
if (libzfs_core_init() != 0) {
(void) close(hdl->libzfs_fd);
free(hdl);
return (NULL);
}
zfs_prop_init();
zpool_prop_init();
zpool_feature_init();
vdev_prop_init();
libzfs_mnttab_init(hdl);
fletcher_4_init();
if (getenv("ZFS_PROP_DEBUG") != NULL) {
hdl->libzfs_prop_debug = B_TRUE;
}
if ((env = getenv("ZFS_SENDRECV_MAX_NVLIST")) != NULL) {
if ((error = zfs_nicestrtonum(hdl, env,
&hdl->libzfs_max_nvlist))) {
errno = error;
(void) close(hdl->libzfs_fd);
free(hdl);
return (NULL);
}
} else {
hdl->libzfs_max_nvlist = (SPA_MAXBLOCKSIZE * 4);
}
/*
* For testing, remove some settable properties and features
*/
if (libzfs_envvar_is_set("ZFS_SYSFS_PROP_SUPPORT_TEST")) {
zprop_desc_t *proptbl;
proptbl = zpool_prop_get_table();
proptbl[ZPOOL_PROP_COMMENT].pd_zfs_mod_supported = B_FALSE;
proptbl = zfs_prop_get_table();
proptbl[ZFS_PROP_DNODESIZE].pd_zfs_mod_supported = B_FALSE;
zfeature_info_t *ftbl = spa_feature_table;
ftbl[SPA_FEATURE_LARGE_BLOCKS].fi_zfs_mod_supported = B_FALSE;
}
return (hdl);
}
void
libzfs_fini(libzfs_handle_t *hdl)
{
(void) close(hdl->libzfs_fd);
zpool_free_handles(hdl);
namespace_clear(hdl);
libzfs_mnttab_fini(hdl);
libzfs_core_fini();
regfree(&hdl->libzfs_urire);
fletcher_4_fini();
#if LIBFETCH_DYNAMIC
if (hdl->libfetch != (void *)-1 && hdl->libfetch != NULL)
(void) dlclose(hdl->libfetch);
free(hdl->libfetch_load_error);
#endif
free(hdl);
}
libzfs_handle_t *
zpool_get_handle(zpool_handle_t *zhp)
{
return (zhp->zpool_hdl);
}
libzfs_handle_t *
zfs_get_handle(zfs_handle_t *zhp)
{
return (zhp->zfs_hdl);
}
zpool_handle_t *
zfs_get_pool_handle(const zfs_handle_t *zhp)
{
return (zhp->zpool_hdl);
}
/*
* Given a name, determine whether or not it's a valid path
* (starts with '/' or "./"). If so, walk the mnttab trying
* to match the device number. If not, treat the path as an
* fs/vol/snap/bkmark name.
*/
zfs_handle_t *
zfs_path_to_zhandle(libzfs_handle_t *hdl, const char *path, zfs_type_t argtype)
{
struct stat64 statbuf;
struct extmnttab entry;
if (path[0] != '/' && strncmp(path, "./", strlen("./")) != 0) {
/*
* It's not a valid path, assume it's a name of type 'argtype'.
*/
return (zfs_open(hdl, path, argtype));
}
if (getextmntent(path, &entry, &statbuf) != 0)
return (NULL);
if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0) {
(void) fprintf(stderr, gettext("'%s': not a ZFS filesystem\n"),
path);
return (NULL);
}
return (zfs_open(hdl, entry.mnt_special, ZFS_TYPE_FILESYSTEM));
}
/*
* Initialize the zc_nvlist_dst member to prepare for receiving an nvlist from
* an ioctl().
*/
void
zcmd_alloc_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, size_t len)
{
if (len == 0)
len = 256 * 1024;
zc->zc_nvlist_dst_size = len;
zc->zc_nvlist_dst =
(uint64_t)(uintptr_t)zfs_alloc(hdl, zc->zc_nvlist_dst_size);
}
/*
* Called when an ioctl() which returns an nvlist fails with ENOMEM. This will
* expand the nvlist to the size specified in 'zc_nvlist_dst_size', which was
* filled in by the kernel to indicate the actual required size.
*/
void
zcmd_expand_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc)
{
free((void *)(uintptr_t)zc->zc_nvlist_dst);
zc->zc_nvlist_dst =
(uint64_t)(uintptr_t)zfs_alloc(hdl, zc->zc_nvlist_dst_size);
}
/*
* Called to free the src and dst nvlists stored in the command structure.
*/
void
zcmd_free_nvlists(zfs_cmd_t *zc)
{
free((void *)(uintptr_t)zc->zc_nvlist_conf);
free((void *)(uintptr_t)zc->zc_nvlist_src);
free((void *)(uintptr_t)zc->zc_nvlist_dst);
zc->zc_nvlist_conf = 0;
zc->zc_nvlist_src = 0;
zc->zc_nvlist_dst = 0;
}
static void
zcmd_write_nvlist_com(libzfs_handle_t *hdl, uint64_t *outnv, uint64_t *outlen,
nvlist_t *nvl)
{
char *packed;
size_t len = fnvlist_size(nvl);
packed = zfs_alloc(hdl, len);
verify(nvlist_pack(nvl, &packed, &len, NV_ENCODE_NATIVE, 0) == 0);
*outnv = (uint64_t)(uintptr_t)packed;
*outlen = len;
}
void
zcmd_write_conf_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl)
{
zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_conf,
&zc->zc_nvlist_conf_size, nvl);
}
void
zcmd_write_src_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl)
{
zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_src,
&zc->zc_nvlist_src_size, nvl);
}
/*
* Unpacks an nvlist from the ZFS ioctl command structure.
*/
int
zcmd_read_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t **nvlp)
{
if (nvlist_unpack((void *)(uintptr_t)zc->zc_nvlist_dst,
zc->zc_nvlist_dst_size, nvlp, 0) != 0)
return (no_memory(hdl));
return (0);
}
/*
* ================================================================
* API shared by zfs and zpool property management
* ================================================================
*/
static void
zprop_print_headers(zprop_get_cbdata_t *cbp, zfs_type_t type)
{
zprop_list_t *pl;
int i;
char *title;
size_t len;
cbp->cb_first = B_FALSE;
if (cbp->cb_scripted)
return;
/*
* Start with the length of the column headers.
*/
cbp->cb_colwidths[GET_COL_NAME] = strlen(dgettext(TEXT_DOMAIN, "NAME"));
cbp->cb_colwidths[GET_COL_PROPERTY] = strlen(dgettext(TEXT_DOMAIN,
"PROPERTY"));
cbp->cb_colwidths[GET_COL_VALUE] = strlen(dgettext(TEXT_DOMAIN,
"VALUE"));
cbp->cb_colwidths[GET_COL_RECVD] = strlen(dgettext(TEXT_DOMAIN,
"RECEIVED"));
cbp->cb_colwidths[GET_COL_SOURCE] = strlen(dgettext(TEXT_DOMAIN,
"SOURCE"));
/* first property is always NAME */
assert(cbp->cb_proplist->pl_prop ==
((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME :
((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME : ZFS_PROP_NAME)));
/*
* Go through and calculate the widths for each column. For the
* 'source' column, we kludge it up by taking the worst-case scenario of
* inheriting from the longest name. This is acceptable because in the
* majority of cases 'SOURCE' is the last column displayed, and we don't
* use the width anyway. Note that the 'VALUE' column can be oversized,
* if the name of the property is much longer than any values we find.
*/
for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) {
/*
* 'PROPERTY' column
*/
if (pl->pl_prop != ZPROP_USERPROP) {
const char *propname = (type == ZFS_TYPE_POOL) ?
zpool_prop_to_name(pl->pl_prop) :
((type == ZFS_TYPE_VDEV) ?
vdev_prop_to_name(pl->pl_prop) :
zfs_prop_to_name(pl->pl_prop));
assert(propname != NULL);
len = strlen(propname);
if (len > cbp->cb_colwidths[GET_COL_PROPERTY])
cbp->cb_colwidths[GET_COL_PROPERTY] = len;
} else {
assert(pl->pl_user_prop != NULL);
len = strlen(pl->pl_user_prop);
if (len > cbp->cb_colwidths[GET_COL_PROPERTY])
cbp->cb_colwidths[GET_COL_PROPERTY] = len;
}
/*
* 'VALUE' column. The first property is always the 'name'
* property that was tacked on either by /sbin/zfs's
* zfs_do_get() or when calling zprop_expand_list(), so we
* ignore its width. If the user specified the name property
* to display, then it will be later in the list in any case.
*/
if (pl != cbp->cb_proplist &&
pl->pl_width > cbp->cb_colwidths[GET_COL_VALUE])
cbp->cb_colwidths[GET_COL_VALUE] = pl->pl_width;
/* 'RECEIVED' column. */
if (pl != cbp->cb_proplist &&
pl->pl_recvd_width > cbp->cb_colwidths[GET_COL_RECVD])
cbp->cb_colwidths[GET_COL_RECVD] = pl->pl_recvd_width;
/*
* 'NAME' and 'SOURCE' columns
*/
if (pl->pl_prop == ((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME :
((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME :
ZFS_PROP_NAME)) && pl->pl_width >
cbp->cb_colwidths[GET_COL_NAME]) {
cbp->cb_colwidths[GET_COL_NAME] = pl->pl_width;
cbp->cb_colwidths[GET_COL_SOURCE] = pl->pl_width +
strlen(dgettext(TEXT_DOMAIN, "inherited from"));
}
}
/*
* Now go through and print the headers.
*/
for (i = 0; i < ZFS_GET_NCOLS; i++) {
switch (cbp->cb_columns[i]) {
case GET_COL_NAME:
title = dgettext(TEXT_DOMAIN, "NAME");
break;
case GET_COL_PROPERTY:
title = dgettext(TEXT_DOMAIN, "PROPERTY");
break;
case GET_COL_VALUE:
title = dgettext(TEXT_DOMAIN, "VALUE");
break;
case GET_COL_RECVD:
title = dgettext(TEXT_DOMAIN, "RECEIVED");
break;
case GET_COL_SOURCE:
title = dgettext(TEXT_DOMAIN, "SOURCE");
break;
default:
title = NULL;
}
if (title != NULL) {
if (i == (ZFS_GET_NCOLS - 1) ||
cbp->cb_columns[i + 1] == GET_COL_NONE)
(void) printf("%s", title);
else
(void) printf("%-*s ",
cbp->cb_colwidths[cbp->cb_columns[i]],
title);
}
}
(void) printf("\n");
}
/*
* Display a single line of output, according to the settings in the callback
* structure.
*/
void
zprop_print_one_property(const char *name, zprop_get_cbdata_t *cbp,
const char *propname, const char *value, zprop_source_t sourcetype,
const char *source, const char *recvd_value)
{
int i;
const char *str = NULL;
char buf[128];
/*
* Ignore those source types that the user has chosen to ignore.
*/
if ((sourcetype & cbp->cb_sources) == 0)
return;
if (cbp->cb_first)
zprop_print_headers(cbp, cbp->cb_type);
for (i = 0; i < ZFS_GET_NCOLS; i++) {
switch (cbp->cb_columns[i]) {
case GET_COL_NAME:
str = name;
break;
case GET_COL_PROPERTY:
str = propname;
break;
case GET_COL_VALUE:
str = value;
break;
case GET_COL_SOURCE:
switch (sourcetype) {
case ZPROP_SRC_NONE:
str = "-";
break;
case ZPROP_SRC_DEFAULT:
str = "default";
break;
case ZPROP_SRC_LOCAL:
str = "local";
break;
case ZPROP_SRC_TEMPORARY:
str = "temporary";
break;
case ZPROP_SRC_INHERITED:
(void) snprintf(buf, sizeof (buf),
"inherited from %s", source);
str = buf;
break;
case ZPROP_SRC_RECEIVED:
str = "received";
break;
default:
str = NULL;
assert(!"unhandled zprop_source_t");
}
break;
case GET_COL_RECVD:
str = (recvd_value == NULL ? "-" : recvd_value);
break;
default:
continue;
}
if (i == (ZFS_GET_NCOLS - 1) ||
cbp->cb_columns[i + 1] == GET_COL_NONE)
(void) printf("%s", str);
else if (cbp->cb_scripted)
(void) printf("%s\t", str);
else
(void) printf("%-*s ",
cbp->cb_colwidths[cbp->cb_columns[i]],
str);
}
(void) printf("\n");
}
/*
* Given a numeric suffix, convert the value into a number of bits that the
* resulting value must be shifted.
*/
static int
str2shift(libzfs_handle_t *hdl, const char *buf)
{
const char *ends = "BKMGTPEZ";
int i;
if (buf[0] == '\0')
return (0);
for (i = 0; i < strlen(ends); i++) {
if (toupper(buf[0]) == ends[i])
break;
}
if (i == strlen(ends)) {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid numeric suffix '%s'"), buf);
return (-1);
}
/*
* Allow 'G' = 'GB' = 'GiB', case-insensitively.
* However, 'BB' and 'BiB' are disallowed.
*/
if (buf[1] == '\0' ||
(toupper(buf[0]) != 'B' &&
((toupper(buf[1]) == 'B' && buf[2] == '\0') ||
(toupper(buf[1]) == 'I' && toupper(buf[2]) == 'B' &&
buf[3] == '\0'))))
return (10 * i);
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid numeric suffix '%s'"), buf);
return (-1);
}
/*
* Convert a string of the form '100G' into a real number. Used when setting
* properties or creating a volume. 'buf' is used to place an extended error
* message for the caller to use.
*/
int
zfs_nicestrtonum(libzfs_handle_t *hdl, const char *value, uint64_t *num)
{
char *end;
int shift;
*num = 0;
/* Check to see if this looks like a number. */
if ((value[0] < '0' || value[0] > '9') && value[0] != '.') {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"bad numeric value '%s'"), value);
return (-1);
}
/* Rely on strtoull() to process the numeric portion. */
errno = 0;
*num = strtoull(value, &end, 10);
/*
* Check for ERANGE, which indicates that the value is too large to fit
* in a 64-bit value.
*/
if (errno == ERANGE) {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"numeric value is too large"));
return (-1);
}
/*
* If we have a decimal value, then do the computation with floating
* point arithmetic. Otherwise, use standard arithmetic.
*/
if (*end == '.') {
double fval = strtod(value, &end);
if ((shift = str2shift(hdl, end)) == -1)
return (-1);
fval *= pow(2, shift);
/*
* UINT64_MAX is not exactly representable as a double.
* The closest representation is UINT64_MAX + 1, so we
* use a >= comparison instead of > for the bounds check.
*/
if (fval >= (double)UINT64_MAX) {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"numeric value is too large"));
return (-1);
}
*num = (uint64_t)fval;
} else {
if ((shift = str2shift(hdl, end)) == -1)
return (-1);
/* Check for overflow */
if (shift >= 64 || (*num << shift) >> shift != *num) {
if (hdl)
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"numeric value is too large"));
return (-1);
}
*num <<= shift;
}
return (0);
}
/*
* Given a propname=value nvpair to set, parse any numeric properties
* (index, boolean, etc) if they are specified as strings and add the
* resulting nvpair to the returned nvlist.
*
* At the DSL layer, all properties are either 64-bit numbers or strings.
* We want the user to be able to ignore this fact and specify properties
* as native values (numbers, for example) or as strings (to simplify
* command line utilities). This also handles converting index types
* (compression, checksum, etc) from strings to their on-disk index.
*/
int
zprop_parse_value(libzfs_handle_t *hdl, nvpair_t *elem, int prop,
zfs_type_t type, nvlist_t *ret, const char **svalp, uint64_t *ivalp,
const char *errbuf)
{
data_type_t datatype = nvpair_type(elem);
zprop_type_t proptype;
const char *propname;
const char *value;
boolean_t isnone = B_FALSE;
boolean_t isauto = B_FALSE;
int err = 0;
if (type == ZFS_TYPE_POOL) {
proptype = zpool_prop_get_type(prop);
propname = zpool_prop_to_name(prop);
} else if (type == ZFS_TYPE_VDEV) {
proptype = vdev_prop_get_type(prop);
propname = vdev_prop_to_name(prop);
} else {
proptype = zfs_prop_get_type(prop);
propname = zfs_prop_to_name(prop);
}
/*
* Convert any properties to the internal DSL value types.
*/
*svalp = NULL;
*ivalp = 0;
switch (proptype) {
case PROP_TYPE_STRING:
if (datatype != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), nvpair_name(elem));
goto error;
}
err = nvpair_value_string(elem, svalp);
if (err != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is invalid"), nvpair_name(elem));
goto error;
}
if (strlen(*svalp) >= ZFS_MAXPROPLEN) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' is too long"), nvpair_name(elem));
goto error;
}
break;
case PROP_TYPE_NUMBER:
if (datatype == DATA_TYPE_STRING) {
(void) nvpair_value_string(elem, &value);
if (strcmp(value, "none") == 0) {
isnone = B_TRUE;
} else if (strcmp(value, "auto") == 0) {
isauto = B_TRUE;
} else if (zfs_nicestrtonum(hdl, value, ivalp) != 0) {
goto error;
}
} else if (datatype == DATA_TYPE_UINT64) {
(void) nvpair_value_uint64(elem, ivalp);
} else {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a number"), nvpair_name(elem));
goto error;
}
/*
* Quota special: force 'none' and don't allow 0.
*/
if ((type & ZFS_TYPE_DATASET) && *ivalp == 0 && !isnone &&
(prop == ZFS_PROP_QUOTA || prop == ZFS_PROP_REFQUOTA)) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"use 'none' to disable quota/refquota"));
goto error;
}
/*
* Special handling for "*_limit=none". In this case it's not
* 0 but UINT64_MAX.
*/
if ((type & ZFS_TYPE_DATASET) && isnone &&
(prop == ZFS_PROP_FILESYSTEM_LIMIT ||
prop == ZFS_PROP_SNAPSHOT_LIMIT)) {
*ivalp = UINT64_MAX;
}
/*
* Special handling for "checksum_*=none". In this case it's not
* 0 but UINT64_MAX.
*/
if ((type & ZFS_TYPE_VDEV) && isnone &&
(prop == VDEV_PROP_CHECKSUM_N ||
prop == VDEV_PROP_CHECKSUM_T ||
prop == VDEV_PROP_IO_N ||
prop == VDEV_PROP_IO_T)) {
*ivalp = UINT64_MAX;
}
/*
* Special handling for setting 'refreservation' to 'auto'. Use
* UINT64_MAX to tell the caller to use zfs_fix_auto_resv().
* 'auto' is only allowed on volumes.
*/
if (isauto) {
switch (prop) {
case ZFS_PROP_REFRESERVATION:
if ((type & ZFS_TYPE_VOLUME) == 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s=auto' only allowed on "
"volumes"), nvpair_name(elem));
goto error;
}
*ivalp = UINT64_MAX;
break;
default:
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'auto' is invalid value for '%s'"),
nvpair_name(elem));
goto error;
}
}
break;
case PROP_TYPE_INDEX:
if (datatype != DATA_TYPE_STRING) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be a string"), nvpair_name(elem));
goto error;
}
(void) nvpair_value_string(elem, &value);
if (zprop_string_to_index(prop, value, ivalp, type) != 0) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"'%s' must be one of '%s'"), propname,
zprop_values(prop, type));
goto error;
}
break;
default:
abort();
}
/*
* Add the result to our return set of properties.
*/
if (*svalp != NULL) {
if (nvlist_add_string(ret, propname, *svalp) != 0) {
(void) no_memory(hdl);
return (-1);
}
} else {
if (nvlist_add_uint64(ret, propname, *ivalp) != 0) {
(void) no_memory(hdl);
return (-1);
}
}
return (0);
error:
(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
return (-1);
}
static int
addlist(libzfs_handle_t *hdl, const char *propname, zprop_list_t **listp,
zfs_type_t type)
{
int prop = zprop_name_to_prop(propname, type);
if (prop != ZPROP_INVAL && !zprop_valid_for_type(prop, type, B_FALSE))
prop = ZPROP_INVAL;
/*
* Return failure if no property table entry was found and this isn't
* a user-defined property.
*/
if (prop == ZPROP_USERPROP && ((type == ZFS_TYPE_POOL &&
!zfs_prop_user(propname) &&
!zpool_prop_feature(propname) &&
!zpool_prop_unsupported(propname)) ||
((type == ZFS_TYPE_DATASET) && !zfs_prop_user(propname) &&
!zfs_prop_userquota(propname) && !zfs_prop_written(propname)) ||
((type == ZFS_TYPE_VDEV) && !vdev_prop_user(propname)))) {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"invalid property '%s'"), propname);
return (zfs_error(hdl, EZFS_BADPROP,
dgettext(TEXT_DOMAIN, "bad property list")));
}
zprop_list_t *entry = zfs_alloc(hdl, sizeof (*entry));
entry->pl_prop = prop;
if (prop == ZPROP_USERPROP) {
entry->pl_user_prop = zfs_strdup(hdl, propname);
entry->pl_width = strlen(propname);
} else {
entry->pl_width = zprop_width(prop, &entry->pl_fixed,
type);
}
*listp = entry;
return (0);
}
/*
* Given a comma-separated list of properties, construct a property list
* containing both user-defined and native properties. This function will
* return a NULL list if 'all' is specified, which can later be expanded
* by zprop_expand_list().
*/
int
zprop_get_list(libzfs_handle_t *hdl, char *props, zprop_list_t **listp,
zfs_type_t type)
{
*listp = NULL;
/*
* If 'all' is specified, return a NULL list.
*/
if (strcmp(props, "all") == 0)
return (0);
/*
* If no props were specified, return an error.
*/
if (props[0] == '\0') {
zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
"no properties specified"));
return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN,
"bad property list")));
}
for (char *p; (p = strsep(&props, ",")); )
if (strcmp(p, "space") == 0) {
static const char *const spaceprops[] = {
"name", "avail", "used", "usedbysnapshots",
"usedbydataset", "usedbyrefreservation",
"usedbychildren"
};
for (int i = 0; i < ARRAY_SIZE(spaceprops); i++) {
if (addlist(hdl, spaceprops[i], listp, type))
return (-1);
listp = &(*listp)->pl_next;
}
} else {
if (addlist(hdl, p, listp, type))
return (-1);
listp = &(*listp)->pl_next;
}
return (0);
}
void
zprop_free_list(zprop_list_t *pl)
{
zprop_list_t *next;
while (pl != NULL) {
next = pl->pl_next;
free(pl->pl_user_prop);
free(pl);
pl = next;
}
}
typedef struct expand_data {
zprop_list_t **last;
libzfs_handle_t *hdl;
zfs_type_t type;
} expand_data_t;
static int
zprop_expand_list_cb(int prop, void *cb)
{
zprop_list_t *entry;
expand_data_t *edp = cb;
entry = zfs_alloc(edp->hdl, sizeof (zprop_list_t));
entry->pl_prop = prop;
entry->pl_width = zprop_width(prop, &entry->pl_fixed, edp->type);
entry->pl_all = B_TRUE;
*(edp->last) = entry;
edp->last = &entry->pl_next;
return (ZPROP_CONT);
}
int
zprop_expand_list(libzfs_handle_t *hdl, zprop_list_t **plp, zfs_type_t type)
{
zprop_list_t *entry;
zprop_list_t **last;
expand_data_t exp;
if (*plp == NULL) {
/*
* If this is the very first time we've been called for an 'all'
* specification, expand the list to include all native
* properties.
*/
last = plp;
exp.last = last;
exp.hdl = hdl;
exp.type = type;
if (zprop_iter_common(zprop_expand_list_cb, &exp, B_FALSE,
B_FALSE, type) == ZPROP_INVAL)
return (-1);
/*
* Add 'name' to the beginning of the list, which is handled
* specially.
*/
entry = zfs_alloc(hdl, sizeof (zprop_list_t));
entry->pl_prop = ((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME :
((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME : ZFS_PROP_NAME));
entry->pl_width = zprop_width(entry->pl_prop,
&entry->pl_fixed, type);
entry->pl_all = B_TRUE;
entry->pl_next = *plp;
*plp = entry;
}
return (0);
}
int
zprop_iter(zprop_func func, void *cb, boolean_t show_all, boolean_t ordered,
zfs_type_t type)
{
return (zprop_iter_common(func, cb, show_all, ordered, type));
}
const char *
zfs_version_userland(void)
{
return (ZFS_META_ALIAS);
}
/*
* Prints both zfs userland and kernel versions
* Returns 0 on success, and -1 on error
*/
int
zfs_version_print(void)
{
(void) puts(ZFS_META_ALIAS);
char *kver = zfs_version_kernel();
if (kver == NULL) {
fprintf(stderr, "zfs_version_kernel() failed: %s\n",
strerror(errno));
return (-1);
}
(void) printf("zfs-kmod-%s\n", kver);
free(kver);
return (0);
}
/*
* Return 1 if the user requested ANSI color output, and our terminal supports
* it. Return 0 for no color.
*/
-static int
+int
use_color(void)
{
static int use_color = -1;
char *term;
/*
* Optimization:
*
* For each zpool invocation, we do a single check to see if we should
* be using color or not, and cache that value for the lifetime of the
* the zpool command. That makes it cheap to call use_color() when
* we're printing with color. We assume that the settings are not going
* to change during the invocation of a zpool command (the user isn't
* going to change the ZFS_COLOR value while zpool is running, for
* example).
*/
if (use_color != -1) {
/*
* We've already figured out if we should be using color or
* not. Return the cached value.
*/
return (use_color);
}
term = getenv("TERM");
/*
* The user sets the ZFS_COLOR env var set to enable zpool ANSI color
* output. However if NO_COLOR is set (https://no-color.org/) then
* don't use it. Also, don't use color if terminal doesn't support
* it.
*/
if (libzfs_envvar_is_set("ZFS_COLOR") &&
!libzfs_envvar_is_set("NO_COLOR") &&
isatty(STDOUT_FILENO) && term && strcmp("dumb", term) != 0 &&
strcmp("unknown", term) != 0) {
/* Color supported */
use_color = 1;
} else {
use_color = 0;
}
return (use_color);
}
/*
* The functions color_start() and color_end() are used for when you want
* to colorize a block of text.
*
* For example:
* color_start(ANSI_RED)
* printf("hello");
* printf("world");
* color_end();
*/
void
color_start(const char *color)
{
if (color && use_color()) {
fputs(color, stdout);
fflush(stdout);
}
}
void
color_end(void)
{
if (use_color()) {
fputs(ANSI_RESET, stdout);
fflush(stdout);
}
}
/*
* printf() with a color. If color is NULL, then do a normal printf.
*/
int
printf_color(const char *color, const char *format, ...)
{
va_list aptr;
int rc;
if (color)
color_start(color);
va_start(aptr, format);
rc = vprintf(format, aptr);
va_end(aptr);
if (color)
color_end();
return (rc);
}
diff --git a/sys/contrib/openzfs/lib/libzutil/zutil_import.c b/sys/contrib/openzfs/lib/libzutil/zutil_import.c
index 65f462e42cd0..19d8a4742813 100644
--- a/sys/contrib/openzfs/lib/libzutil/zutil_import.c
+++ b/sys/contrib/openzfs/lib/libzutil/zutil_import.c
@@ -1,1966 +1,1965 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2015 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
* Copyright 2015 RackTop Systems.
* Copyright (c) 2016, Intel Corporation.
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
*/
/*
* Pool import support functions.
*
* Used by zpool, ztest, zdb, and zhack to locate importable configs. Since
* these commands are expected to run in the global zone, we can assume
* that the devices are all readable when called.
*
* To import a pool, we rely on reading the configuration information from the
* ZFS label of each device. If we successfully read the label, then we
* organize the configuration information in the following hierarchy:
*
* pool guid -> toplevel vdev guid -> label txg
*
* Duplicate entries matching this same tuple will be discarded. Once we have
* examined every device, we pick the best label txg config for each toplevel
* vdev. We then arrange these toplevel vdevs into a complete pool config, and
* update any paths that have changed. Finally, we attempt to import the pool
* using our derived config, and record the results.
*/
#ifdef HAVE_AIO_H
#include <aio.h>
#endif
#include <ctype.h>
#include <dirent.h>
#include <errno.h>
#include <libintl.h>
#include <libgen.h>
#include <stddef.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/dktp/fdisk.h>
#include <sys/vdev_impl.h>
#include <sys/fs/zfs.h>
#include <thread_pool.h>
#include <libzutil.h>
#include <libnvpair.h>
#include "zutil_import.h"
const char *
libpc_error_description(libpc_handle_t *hdl)
{
if (hdl->lpc_desc[0] != '\0')
return (hdl->lpc_desc);
switch (hdl->lpc_error) {
case LPC_BADCACHE:
return (dgettext(TEXT_DOMAIN, "invalid or missing cache file"));
case LPC_BADPATH:
return (dgettext(TEXT_DOMAIN, "must be an absolute path"));
case LPC_NOMEM:
return (dgettext(TEXT_DOMAIN, "out of memory"));
case LPC_EACCESS:
return (dgettext(TEXT_DOMAIN, "some devices require root "
"privileges"));
case LPC_UNKNOWN:
return (dgettext(TEXT_DOMAIN, "unknown error"));
default:
assert(hdl->lpc_error == 0);
return (dgettext(TEXT_DOMAIN, "no error"));
}
}
static __attribute__((format(printf, 2, 3))) void
zutil_error_aux(libpc_handle_t *hdl, const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
(void) vsnprintf(hdl->lpc_desc, sizeof (hdl->lpc_desc), fmt, ap);
hdl->lpc_desc_active = B_TRUE;
va_end(ap);
}
static void
zutil_verror(libpc_handle_t *hdl, lpc_error_t error, const char *fmt,
va_list ap)
{
char action[1024];
(void) vsnprintf(action, sizeof (action), fmt, ap);
hdl->lpc_error = error;
if (hdl->lpc_desc_active)
hdl->lpc_desc_active = B_FALSE;
else
hdl->lpc_desc[0] = '\0';
if (hdl->lpc_printerr)
(void) fprintf(stderr, "%s: %s\n", action,
libpc_error_description(hdl));
}
static __attribute__((format(printf, 3, 4))) int
zutil_error_fmt(libpc_handle_t *hdl, lpc_error_t error,
const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
zutil_verror(hdl, error, fmt, ap);
va_end(ap);
return (-1);
}
static int
zutil_error(libpc_handle_t *hdl, lpc_error_t error, const char *msg)
{
return (zutil_error_fmt(hdl, error, "%s", msg));
}
static int
zutil_no_memory(libpc_handle_t *hdl)
{
zutil_error(hdl, LPC_NOMEM, "internal error");
exit(1);
}
void *
zutil_alloc(libpc_handle_t *hdl, size_t size)
{
void *data;
if ((data = calloc(1, size)) == NULL)
(void) zutil_no_memory(hdl);
return (data);
}
char *
zutil_strdup(libpc_handle_t *hdl, const char *str)
{
char *ret;
if ((ret = strdup(str)) == NULL)
(void) zutil_no_memory(hdl);
return (ret);
}
static char *
zutil_strndup(libpc_handle_t *hdl, const char *str, size_t n)
{
char *ret;
if ((ret = strndup(str, n)) == NULL)
(void) zutil_no_memory(hdl);
return (ret);
}
/*
* Intermediate structures used to gather configuration information.
*/
typedef struct config_entry {
uint64_t ce_txg;
nvlist_t *ce_config;
struct config_entry *ce_next;
} config_entry_t;
typedef struct vdev_entry {
uint64_t ve_guid;
config_entry_t *ve_configs;
struct vdev_entry *ve_next;
} vdev_entry_t;
typedef struct pool_entry {
uint64_t pe_guid;
vdev_entry_t *pe_vdevs;
struct pool_entry *pe_next;
} pool_entry_t;
typedef struct name_entry {
char *ne_name;
uint64_t ne_guid;
uint64_t ne_order;
uint64_t ne_num_labels;
struct name_entry *ne_next;
} name_entry_t;
typedef struct pool_list {
pool_entry_t *pools;
name_entry_t *names;
} pool_list_t;
/*
* Go through and fix up any path and/or devid information for the given vdev
* configuration.
*/
static int
fix_paths(libpc_handle_t *hdl, nvlist_t *nv, name_entry_t *names)
{
nvlist_t **child;
uint_t c, children;
uint64_t guid;
name_entry_t *ne, *best;
const char *path;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0) {
for (c = 0; c < children; c++)
if (fix_paths(hdl, child[c], names) != 0)
return (-1);
return (0);
}
/*
* This is a leaf (file or disk) vdev. In either case, go through
* the name list and see if we find a matching guid. If so, replace
* the path and see if we can calculate a new devid.
*
* There may be multiple names associated with a particular guid, in
* which case we have overlapping partitions or multiple paths to the
* same disk. In this case we prefer to use the path name which
* matches the ZPOOL_CONFIG_PATH. If no matching entry is found we
* use the lowest order device which corresponds to the first match
* while traversing the ZPOOL_IMPORT_PATH search path.
*/
verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) != 0)
path = NULL;
best = NULL;
for (ne = names; ne != NULL; ne = ne->ne_next) {
if (ne->ne_guid == guid) {
if (path == NULL) {
best = ne;
break;
}
if ((strlen(path) == strlen(ne->ne_name)) &&
strncmp(path, ne->ne_name, strlen(path)) == 0) {
best = ne;
break;
}
if (best == NULL) {
best = ne;
continue;
}
/* Prefer paths with move vdev labels. */
if (ne->ne_num_labels > best->ne_num_labels) {
best = ne;
continue;
}
/* Prefer paths earlier in the search order. */
if (ne->ne_num_labels == best->ne_num_labels &&
ne->ne_order < best->ne_order) {
best = ne;
continue;
}
}
}
if (best == NULL)
return (0);
if (nvlist_add_string(nv, ZPOOL_CONFIG_PATH, best->ne_name) != 0)
return (-1);
update_vdev_config_dev_strs(nv);
return (0);
}
/*
* Add the given configuration to the list of known devices.
*/
static int
add_config(libpc_handle_t *hdl, pool_list_t *pl, const char *path,
int order, int num_labels, nvlist_t *config)
{
uint64_t pool_guid, vdev_guid, top_guid, txg, state;
pool_entry_t *pe;
vdev_entry_t *ve;
config_entry_t *ce;
name_entry_t *ne;
/*
* If this is a hot spare not currently in use or level 2 cache
* device, add it to the list of names to translate, but don't do
* anything else.
*/
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
&state) == 0 &&
(state == POOL_STATE_SPARE || state == POOL_STATE_L2CACHE) &&
nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID, &vdev_guid) == 0) {
if ((ne = zutil_alloc(hdl, sizeof (name_entry_t))) == NULL)
return (-1);
if ((ne->ne_name = zutil_strdup(hdl, path)) == NULL) {
free(ne);
return (-1);
}
ne->ne_guid = vdev_guid;
ne->ne_order = order;
ne->ne_num_labels = num_labels;
ne->ne_next = pl->names;
pl->names = ne;
return (0);
}
/*
* If we have a valid config but cannot read any of these fields, then
* it means we have a half-initialized label. In vdev_label_init()
* we write a label with txg == 0 so that we can identify the device
* in case the user refers to the same disk later on. If we fail to
* create the pool, we'll be left with a label in this state
* which should not be considered part of a valid pool.
*/
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
&pool_guid) != 0 ||
nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID,
&vdev_guid) != 0 ||
nvlist_lookup_uint64(config, ZPOOL_CONFIG_TOP_GUID,
&top_guid) != 0 ||
nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_TXG,
&txg) != 0 || txg == 0) {
return (0);
}
/*
* First, see if we know about this pool. If not, then add it to the
* list of known pools.
*/
for (pe = pl->pools; pe != NULL; pe = pe->pe_next) {
if (pe->pe_guid == pool_guid)
break;
}
if (pe == NULL) {
if ((pe = zutil_alloc(hdl, sizeof (pool_entry_t))) == NULL) {
return (-1);
}
pe->pe_guid = pool_guid;
pe->pe_next = pl->pools;
pl->pools = pe;
}
/*
* Second, see if we know about this toplevel vdev. Add it if its
* missing.
*/
for (ve = pe->pe_vdevs; ve != NULL; ve = ve->ve_next) {
if (ve->ve_guid == top_guid)
break;
}
if (ve == NULL) {
if ((ve = zutil_alloc(hdl, sizeof (vdev_entry_t))) == NULL) {
return (-1);
}
ve->ve_guid = top_guid;
ve->ve_next = pe->pe_vdevs;
pe->pe_vdevs = ve;
}
/*
* Third, see if we have a config with a matching transaction group. If
* so, then we do nothing. Otherwise, add it to the list of known
* configs.
*/
for (ce = ve->ve_configs; ce != NULL; ce = ce->ce_next) {
if (ce->ce_txg == txg)
break;
}
if (ce == NULL) {
if ((ce = zutil_alloc(hdl, sizeof (config_entry_t))) == NULL) {
return (-1);
}
ce->ce_txg = txg;
ce->ce_config = fnvlist_dup(config);
ce->ce_next = ve->ve_configs;
ve->ve_configs = ce;
}
/*
* At this point we've successfully added our config to the list of
* known configs. The last thing to do is add the vdev guid -> path
* mappings so that we can fix up the configuration as necessary before
* doing the import.
*/
if ((ne = zutil_alloc(hdl, sizeof (name_entry_t))) == NULL)
return (-1);
if ((ne->ne_name = zutil_strdup(hdl, path)) == NULL) {
free(ne);
return (-1);
}
ne->ne_guid = vdev_guid;
ne->ne_order = order;
ne->ne_num_labels = num_labels;
ne->ne_next = pl->names;
pl->names = ne;
return (0);
}
static int
zutil_pool_active(libpc_handle_t *hdl, const char *name, uint64_t guid,
boolean_t *isactive)
{
ASSERT(hdl->lpc_ops->pco_pool_active != NULL);
int error = hdl->lpc_ops->pco_pool_active(hdl->lpc_lib_handle, name,
guid, isactive);
return (error);
}
static nvlist_t *
zutil_refresh_config(libpc_handle_t *hdl, nvlist_t *tryconfig)
{
ASSERT(hdl->lpc_ops->pco_refresh_config != NULL);
return (hdl->lpc_ops->pco_refresh_config(hdl->lpc_lib_handle,
tryconfig));
}
/*
* Determine if the vdev id is a hole in the namespace.
*/
static boolean_t
vdev_is_hole(uint64_t *hole_array, uint_t holes, uint_t id)
{
int c;
for (c = 0; c < holes; c++) {
/* Top-level is a hole */
if (hole_array[c] == id)
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Convert our list of pools into the definitive set of configurations. We
* start by picking the best config for each toplevel vdev. Once that's done,
* we assemble the toplevel vdevs into a full config for the pool. We make a
* pass to fix up any incorrect paths, and then add it to the main list to
* return to the user.
*/
static nvlist_t *
get_configs(libpc_handle_t *hdl, pool_list_t *pl, boolean_t active_ok,
nvlist_t *policy)
{
pool_entry_t *pe;
vdev_entry_t *ve;
config_entry_t *ce;
nvlist_t *ret = NULL, *config = NULL, *tmp = NULL, *nvtop, *nvroot;
nvlist_t **spares, **l2cache;
uint_t i, nspares, nl2cache;
boolean_t config_seen;
uint64_t best_txg;
const char *name, *hostname = NULL;
uint64_t guid;
uint_t children = 0;
nvlist_t **child = NULL;
uint64_t *hole_array, max_id;
uint_t c;
boolean_t isactive;
nvlist_t *nvl;
boolean_t valid_top_config = B_FALSE;
if (nvlist_alloc(&ret, 0, 0) != 0)
goto nomem;
for (pe = pl->pools; pe != NULL; pe = pe->pe_next) {
uint64_t id, max_txg = 0, hostid = 0;
uint_t holes = 0;
if (nvlist_alloc(&config, NV_UNIQUE_NAME, 0) != 0)
goto nomem;
config_seen = B_FALSE;
/*
* Iterate over all toplevel vdevs. Grab the pool configuration
* from the first one we find, and then go through the rest and
* add them as necessary to the 'vdevs' member of the config.
*/
for (ve = pe->pe_vdevs; ve != NULL; ve = ve->ve_next) {
/*
* Determine the best configuration for this vdev by
* selecting the config with the latest transaction
* group.
*/
best_txg = 0;
for (ce = ve->ve_configs; ce != NULL;
ce = ce->ce_next) {
if (ce->ce_txg > best_txg) {
tmp = ce->ce_config;
best_txg = ce->ce_txg;
}
}
/*
* We rely on the fact that the max txg for the
* pool will contain the most up-to-date information
* about the valid top-levels in the vdev namespace.
*/
if (best_txg > max_txg) {
(void) nvlist_remove(config,
ZPOOL_CONFIG_VDEV_CHILDREN,
DATA_TYPE_UINT64);
(void) nvlist_remove(config,
ZPOOL_CONFIG_HOLE_ARRAY,
DATA_TYPE_UINT64_ARRAY);
max_txg = best_txg;
hole_array = NULL;
holes = 0;
max_id = 0;
valid_top_config = B_FALSE;
if (nvlist_lookup_uint64(tmp,
ZPOOL_CONFIG_VDEV_CHILDREN, &max_id) == 0) {
verify(nvlist_add_uint64(config,
ZPOOL_CONFIG_VDEV_CHILDREN,
max_id) == 0);
valid_top_config = B_TRUE;
}
if (nvlist_lookup_uint64_array(tmp,
ZPOOL_CONFIG_HOLE_ARRAY, &hole_array,
&holes) == 0) {
verify(nvlist_add_uint64_array(config,
ZPOOL_CONFIG_HOLE_ARRAY,
hole_array, holes) == 0);
}
}
if (!config_seen) {
/*
* Copy the relevant pieces of data to the pool
* configuration:
*
* version
* pool guid
* name
* comment (if available)
* compatibility features (if available)
* pool state
* hostid (if available)
* hostname (if available)
*/
uint64_t state, version;
const char *comment = NULL;
const char *compatibility = NULL;
version = fnvlist_lookup_uint64(tmp,
ZPOOL_CONFIG_VERSION);
fnvlist_add_uint64(config,
ZPOOL_CONFIG_VERSION, version);
guid = fnvlist_lookup_uint64(tmp,
ZPOOL_CONFIG_POOL_GUID);
fnvlist_add_uint64(config,
ZPOOL_CONFIG_POOL_GUID, guid);
name = fnvlist_lookup_string(tmp,
ZPOOL_CONFIG_POOL_NAME);
fnvlist_add_string(config,
ZPOOL_CONFIG_POOL_NAME, name);
if (nvlist_lookup_string(tmp,
ZPOOL_CONFIG_COMMENT, &comment) == 0)
fnvlist_add_string(config,
ZPOOL_CONFIG_COMMENT, comment);
if (nvlist_lookup_string(tmp,
ZPOOL_CONFIG_COMPATIBILITY,
&compatibility) == 0)
fnvlist_add_string(config,
ZPOOL_CONFIG_COMPATIBILITY,
compatibility);
state = fnvlist_lookup_uint64(tmp,
ZPOOL_CONFIG_POOL_STATE);
fnvlist_add_uint64(config,
ZPOOL_CONFIG_POOL_STATE, state);
hostid = 0;
if (nvlist_lookup_uint64(tmp,
ZPOOL_CONFIG_HOSTID, &hostid) == 0) {
fnvlist_add_uint64(config,
ZPOOL_CONFIG_HOSTID, hostid);
hostname = fnvlist_lookup_string(tmp,
ZPOOL_CONFIG_HOSTNAME);
fnvlist_add_string(config,
ZPOOL_CONFIG_HOSTNAME, hostname);
}
config_seen = B_TRUE;
}
/*
* Add this top-level vdev to the child array.
*/
verify(nvlist_lookup_nvlist(tmp,
ZPOOL_CONFIG_VDEV_TREE, &nvtop) == 0);
verify(nvlist_lookup_uint64(nvtop, ZPOOL_CONFIG_ID,
&id) == 0);
if (id >= children) {
nvlist_t **newchild;
newchild = zutil_alloc(hdl, (id + 1) *
sizeof (nvlist_t *));
if (newchild == NULL)
goto nomem;
for (c = 0; c < children; c++)
newchild[c] = child[c];
free(child);
child = newchild;
children = id + 1;
}
if (nvlist_dup(nvtop, &child[id], 0) != 0)
goto nomem;
}
/*
* If we have information about all the top-levels then
* clean up the nvlist which we've constructed. This
* means removing any extraneous devices that are
* beyond the valid range or adding devices to the end
* of our array which appear to be missing.
*/
if (valid_top_config) {
if (max_id < children) {
for (c = max_id; c < children; c++)
nvlist_free(child[c]);
children = max_id;
} else if (max_id > children) {
nvlist_t **newchild;
newchild = zutil_alloc(hdl, (max_id) *
sizeof (nvlist_t *));
if (newchild == NULL)
goto nomem;
for (c = 0; c < children; c++)
newchild[c] = child[c];
free(child);
child = newchild;
children = max_id;
}
}
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
&guid) == 0);
/*
* The vdev namespace may contain holes as a result of
* device removal. We must add them back into the vdev
* tree before we process any missing devices.
*/
if (holes > 0) {
ASSERT(valid_top_config);
for (c = 0; c < children; c++) {
nvlist_t *holey;
if (child[c] != NULL ||
!vdev_is_hole(hole_array, holes, c))
continue;
if (nvlist_alloc(&holey, NV_UNIQUE_NAME,
0) != 0)
goto nomem;
/*
* Holes in the namespace are treated as
* "hole" top-level vdevs and have a
* special flag set on them.
*/
if (nvlist_add_string(holey,
ZPOOL_CONFIG_TYPE,
VDEV_TYPE_HOLE) != 0 ||
nvlist_add_uint64(holey,
ZPOOL_CONFIG_ID, c) != 0 ||
nvlist_add_uint64(holey,
ZPOOL_CONFIG_GUID, 0ULL) != 0) {
nvlist_free(holey);
goto nomem;
}
child[c] = holey;
}
}
/*
* Look for any missing top-level vdevs. If this is the case,
* create a faked up 'missing' vdev as a placeholder. We cannot
* simply compress the child array, because the kernel performs
* certain checks to make sure the vdev IDs match their location
* in the configuration.
*/
for (c = 0; c < children; c++) {
if (child[c] == NULL) {
nvlist_t *missing;
if (nvlist_alloc(&missing, NV_UNIQUE_NAME,
0) != 0)
goto nomem;
if (nvlist_add_string(missing,
ZPOOL_CONFIG_TYPE,
VDEV_TYPE_MISSING) != 0 ||
nvlist_add_uint64(missing,
ZPOOL_CONFIG_ID, c) != 0 ||
nvlist_add_uint64(missing,
ZPOOL_CONFIG_GUID, 0ULL) != 0) {
nvlist_free(missing);
goto nomem;
}
child[c] = missing;
}
}
/*
* Put all of this pool's top-level vdevs into a root vdev.
*/
if (nvlist_alloc(&nvroot, NV_UNIQUE_NAME, 0) != 0)
goto nomem;
if (nvlist_add_string(nvroot, ZPOOL_CONFIG_TYPE,
VDEV_TYPE_ROOT) != 0 ||
nvlist_add_uint64(nvroot, ZPOOL_CONFIG_ID, 0ULL) != 0 ||
nvlist_add_uint64(nvroot, ZPOOL_CONFIG_GUID, guid) != 0 ||
nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t **)child, children) != 0) {
nvlist_free(nvroot);
goto nomem;
}
for (c = 0; c < children; c++)
nvlist_free(child[c]);
free(child);
children = 0;
child = NULL;
/*
* Go through and fix up any paths and/or devids based on our
* known list of vdev GUID -> path mappings.
*/
if (fix_paths(hdl, nvroot, pl->names) != 0) {
nvlist_free(nvroot);
goto nomem;
}
/*
* Add the root vdev to this pool's configuration.
*/
if (nvlist_add_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
nvroot) != 0) {
nvlist_free(nvroot);
goto nomem;
}
nvlist_free(nvroot);
/*
* zdb uses this path to report on active pools that were
* imported or created using -R.
*/
if (active_ok)
goto add_pool;
/*
* Determine if this pool is currently active, in which case we
* can't actually import it.
*/
verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
&name) == 0);
verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
&guid) == 0);
if (zutil_pool_active(hdl, name, guid, &isactive) != 0)
goto error;
if (isactive) {
nvlist_free(config);
config = NULL;
continue;
}
if (policy != NULL) {
if (nvlist_add_nvlist(config, ZPOOL_LOAD_POLICY,
policy) != 0)
goto nomem;
}
if ((nvl = zutil_refresh_config(hdl, config)) == NULL) {
nvlist_free(config);
config = NULL;
continue;
}
nvlist_free(config);
config = nvl;
/*
* Go through and update the paths for spares, now that we have
* them.
*/
verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
&nvroot) == 0);
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0) {
for (i = 0; i < nspares; i++) {
if (fix_paths(hdl, spares[i], pl->names) != 0)
goto nomem;
}
}
/*
* Update the paths for l2cache devices.
*/
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0) {
for (i = 0; i < nl2cache; i++) {
if (fix_paths(hdl, l2cache[i], pl->names) != 0)
goto nomem;
}
}
/*
* Restore the original information read from the actual label.
*/
(void) nvlist_remove(config, ZPOOL_CONFIG_HOSTID,
DATA_TYPE_UINT64);
(void) nvlist_remove(config, ZPOOL_CONFIG_HOSTNAME,
DATA_TYPE_STRING);
if (hostid != 0) {
verify(nvlist_add_uint64(config, ZPOOL_CONFIG_HOSTID,
hostid) == 0);
verify(nvlist_add_string(config, ZPOOL_CONFIG_HOSTNAME,
hostname) == 0);
}
add_pool:
/*
* Add this pool to the list of configs.
*/
verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
&name) == 0);
if (nvlist_add_nvlist(ret, name, config) != 0)
goto nomem;
nvlist_free(config);
config = NULL;
}
return (ret);
nomem:
(void) zutil_no_memory(hdl);
error:
nvlist_free(config);
nvlist_free(ret);
for (c = 0; c < children; c++)
nvlist_free(child[c]);
free(child);
return (NULL);
}
/*
* Return the offset of the given label.
*/
static uint64_t
label_offset(uint64_t size, int l)
{
ASSERT(P2PHASE_TYPED(size, sizeof (vdev_label_t), uint64_t) == 0);
return (l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
0 : size - VDEV_LABELS * sizeof (vdev_label_t)));
}
/*
* The same description applies as to zpool_read_label below,
* except here we do it without aio, presumably because an aio call
* errored out in a way we think not using it could circumvent.
*/
static int
zpool_read_label_slow(int fd, nvlist_t **config, int *num_labels)
{
struct stat64 statbuf;
int l, count = 0;
vdev_phys_t *label;
nvlist_t *expected_config = NULL;
uint64_t expected_guid = 0, size;
*config = NULL;
if (fstat64_blk(fd, &statbuf) == -1)
return (0);
size = P2ALIGN_TYPED(statbuf.st_size, sizeof (vdev_label_t), uint64_t);
label = (vdev_phys_t *)umem_alloc_aligned(sizeof (*label), PAGESIZE,
UMEM_DEFAULT);
if (label == NULL)
return (-1);
for (l = 0; l < VDEV_LABELS; l++) {
uint64_t state, guid, txg;
off_t offset = label_offset(size, l) + VDEV_SKIP_SIZE;
if (pread64(fd, label, sizeof (vdev_phys_t),
offset) != sizeof (vdev_phys_t))
continue;
if (nvlist_unpack(label->vp_nvlist,
sizeof (label->vp_nvlist), config, 0) != 0)
continue;
if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_GUID,
&guid) != 0 || guid == 0) {
nvlist_free(*config);
continue;
}
if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_STATE,
&state) != 0 || state > POOL_STATE_L2CACHE) {
nvlist_free(*config);
continue;
}
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
(nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_TXG,
&txg) != 0 || txg == 0)) {
nvlist_free(*config);
continue;
}
if (expected_guid) {
if (expected_guid == guid)
count++;
nvlist_free(*config);
} else {
expected_config = *config;
expected_guid = guid;
count++;
}
}
if (num_labels != NULL)
*num_labels = count;
umem_free_aligned(label, sizeof (*label));
*config = expected_config;
return (0);
}
/*
* Given a file descriptor, read the label information and return an nvlist
* describing the configuration, if there is one. The number of valid
* labels found will be returned in num_labels when non-NULL.
*/
int
zpool_read_label(int fd, nvlist_t **config, int *num_labels)
{
#ifndef HAVE_AIO_H
return (zpool_read_label_slow(fd, config, num_labels));
#else
struct stat64 statbuf;
struct aiocb aiocbs[VDEV_LABELS];
struct aiocb *aiocbps[VDEV_LABELS];
vdev_phys_t *labels;
nvlist_t *expected_config = NULL;
uint64_t expected_guid = 0, size;
int error, l, count = 0;
*config = NULL;
if (fstat64_blk(fd, &statbuf) == -1)
return (0);
size = P2ALIGN_TYPED(statbuf.st_size, sizeof (vdev_label_t), uint64_t);
labels = (vdev_phys_t *)umem_alloc_aligned(
VDEV_LABELS * sizeof (*labels), PAGESIZE, UMEM_DEFAULT);
if (labels == NULL)
return (-1);
memset(aiocbs, 0, sizeof (aiocbs));
for (l = 0; l < VDEV_LABELS; l++) {
off_t offset = label_offset(size, l) + VDEV_SKIP_SIZE;
aiocbs[l].aio_fildes = fd;
aiocbs[l].aio_offset = offset;
aiocbs[l].aio_buf = &labels[l];
aiocbs[l].aio_nbytes = sizeof (vdev_phys_t);
aiocbs[l].aio_lio_opcode = LIO_READ;
aiocbps[l] = &aiocbs[l];
}
if (lio_listio(LIO_WAIT, aiocbps, VDEV_LABELS, NULL) != 0) {
int saved_errno = errno;
boolean_t do_slow = B_FALSE;
error = -1;
if (errno == EAGAIN || errno == EINTR || errno == EIO) {
/*
* A portion of the requests may have been submitted.
* Clean them up.
*/
for (l = 0; l < VDEV_LABELS; l++) {
errno = 0;
switch (aio_error(&aiocbs[l])) {
case EINVAL:
break;
case EINPROGRESS:
// This shouldn't be possible to
// encounter, die if we do.
ASSERT(B_FALSE);
zfs_fallthrough;
case EOPNOTSUPP:
case ENOSYS:
do_slow = B_TRUE;
zfs_fallthrough;
case 0:
default:
(void) aio_return(&aiocbs[l]);
}
}
}
if (do_slow) {
/*
* At least some IO involved access unsafe-for-AIO
* files. Let's try again, without AIO this time.
*/
error = zpool_read_label_slow(fd, config, num_labels);
saved_errno = errno;
}
umem_free_aligned(labels, VDEV_LABELS * sizeof (*labels));
errno = saved_errno;
return (error);
}
for (l = 0; l < VDEV_LABELS; l++) {
uint64_t state, guid, txg;
if (aio_return(&aiocbs[l]) != sizeof (vdev_phys_t))
continue;
if (nvlist_unpack(labels[l].vp_nvlist,
sizeof (labels[l].vp_nvlist), config, 0) != 0)
continue;
if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_GUID,
&guid) != 0 || guid == 0) {
nvlist_free(*config);
continue;
}
if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_STATE,
&state) != 0 || state > POOL_STATE_L2CACHE) {
nvlist_free(*config);
continue;
}
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
(nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_TXG,
&txg) != 0 || txg == 0)) {
nvlist_free(*config);
continue;
}
if (expected_guid) {
if (expected_guid == guid)
count++;
nvlist_free(*config);
} else {
expected_config = *config;
expected_guid = guid;
count++;
}
}
if (num_labels != NULL)
*num_labels = count;
umem_free_aligned(labels, VDEV_LABELS * sizeof (*labels));
*config = expected_config;
return (0);
#endif
}
/*
* Sorted by full path and then vdev guid to allow for multiple entries with
* the same full path name. This is required because it's possible to
* have multiple block devices with labels that refer to the same
* ZPOOL_CONFIG_PATH yet have different vdev guids. In this case both
* entries need to be added to the cache. Scenarios where this can occur
* include overwritten pool labels, devices which are visible from multiple
* hosts and multipath devices.
*/
int
slice_cache_compare(const void *arg1, const void *arg2)
{
const char *nm1 = ((rdsk_node_t *)arg1)->rn_name;
const char *nm2 = ((rdsk_node_t *)arg2)->rn_name;
uint64_t guid1 = ((rdsk_node_t *)arg1)->rn_vdev_guid;
uint64_t guid2 = ((rdsk_node_t *)arg2)->rn_vdev_guid;
int rv;
rv = TREE_ISIGN(strcmp(nm1, nm2));
if (rv)
return (rv);
return (TREE_CMP(guid1, guid2));
}
static int
label_paths_impl(libpc_handle_t *hdl, nvlist_t *nvroot, uint64_t pool_guid,
uint64_t vdev_guid, const char **path, const char **devid)
{
nvlist_t **child;
uint_t c, children;
uint64_t guid;
const char *val;
int error;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0) {
for (c = 0; c < children; c++) {
error = label_paths_impl(hdl, child[c],
pool_guid, vdev_guid, path, devid);
if (error)
return (error);
}
return (0);
}
if (nvroot == NULL)
return (0);
error = nvlist_lookup_uint64(nvroot, ZPOOL_CONFIG_GUID, &guid);
if ((error != 0) || (guid != vdev_guid))
return (0);
error = nvlist_lookup_string(nvroot, ZPOOL_CONFIG_PATH, &val);
if (error == 0)
*path = val;
error = nvlist_lookup_string(nvroot, ZPOOL_CONFIG_DEVID, &val);
if (error == 0)
*devid = val;
return (0);
}
/*
* Given a disk label fetch the ZPOOL_CONFIG_PATH and ZPOOL_CONFIG_DEVID
* and store these strings as config_path and devid_path respectively.
* The returned pointers are only valid as long as label remains valid.
*/
int
label_paths(libpc_handle_t *hdl, nvlist_t *label, const char **path,
const char **devid)
{
nvlist_t *nvroot;
uint64_t pool_guid;
uint64_t vdev_guid;
*path = NULL;
*devid = NULL;
if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_VDEV_TREE, &nvroot) ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID, &pool_guid) ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &vdev_guid))
return (ENOENT);
return (label_paths_impl(hdl, nvroot, pool_guid, vdev_guid, path,
devid));
}
static void
zpool_find_import_scan_add_slice(libpc_handle_t *hdl, pthread_mutex_t *lock,
avl_tree_t *cache, const char *path, const char *name, int order)
{
avl_index_t where;
rdsk_node_t *slice;
slice = zutil_alloc(hdl, sizeof (rdsk_node_t));
if (asprintf(&slice->rn_name, "%s/%s", path, name) == -1) {
free(slice);
return;
}
slice->rn_vdev_guid = 0;
slice->rn_lock = lock;
slice->rn_avl = cache;
slice->rn_hdl = hdl;
slice->rn_order = order + IMPORT_ORDER_SCAN_OFFSET;
slice->rn_labelpaths = B_FALSE;
pthread_mutex_lock(lock);
if (avl_find(cache, slice, &where)) {
free(slice->rn_name);
free(slice);
} else {
avl_insert(cache, slice, where);
}
pthread_mutex_unlock(lock);
}
static int
zpool_find_import_scan_dir(libpc_handle_t *hdl, pthread_mutex_t *lock,
avl_tree_t *cache, const char *dir, int order)
{
int error;
char path[MAXPATHLEN];
struct dirent64 *dp;
DIR *dirp;
if (realpath(dir, path) == NULL) {
error = errno;
if (error == ENOENT)
return (0);
zutil_error_aux(hdl, "%s", strerror(error));
(void) zutil_error_fmt(hdl, LPC_BADPATH, dgettext(TEXT_DOMAIN,
"cannot resolve path '%s'"), dir);
return (error);
}
dirp = opendir(path);
if (dirp == NULL) {
error = errno;
zutil_error_aux(hdl, "%s", strerror(error));
(void) zutil_error_fmt(hdl, LPC_BADPATH, dgettext(TEXT_DOMAIN,
"cannot open '%s'"), path);
return (error);
}
while ((dp = readdir64(dirp)) != NULL) {
const char *name = dp->d_name;
if (strcmp(name, ".") == 0 || strcmp(name, "..") == 0)
continue;
switch (dp->d_type) {
case DT_UNKNOWN:
case DT_BLK:
case DT_LNK:
#ifdef __FreeBSD__
case DT_CHR:
#endif
case DT_REG:
break;
default:
continue;
}
zpool_find_import_scan_add_slice(hdl, lock, cache, path, name,
order);
}
(void) closedir(dirp);
return (0);
}
static int
zpool_find_import_scan_path(libpc_handle_t *hdl, pthread_mutex_t *lock,
avl_tree_t *cache, const char *dir, int order)
{
int error = 0;
char path[MAXPATHLEN];
char *d = NULL;
ssize_t dl;
const char *dpath, *name;
/*
* Separate the directory and the basename.
* We do this so that we can get the realpath of
* the directory. We don't get the realpath on the
* whole path because if it's a symlink, we want the
* path of the symlink not where it points to.
*/
name = zfs_basename(dir);
if ((dl = zfs_dirnamelen(dir)) == -1)
dpath = ".";
else
dpath = d = zutil_strndup(hdl, dir, dl);
if (realpath(dpath, path) == NULL) {
error = errno;
if (error == ENOENT) {
error = 0;
goto out;
}
zutil_error_aux(hdl, "%s", strerror(error));
(void) zutil_error_fmt(hdl, LPC_BADPATH, dgettext(TEXT_DOMAIN,
"cannot resolve path '%s'"), dir);
goto out;
}
zpool_find_import_scan_add_slice(hdl, lock, cache, path, name, order);
out:
free(d);
return (error);
}
/*
* Scan a list of directories for zfs devices.
*/
static int
zpool_find_import_scan(libpc_handle_t *hdl, pthread_mutex_t *lock,
avl_tree_t **slice_cache, const char * const *dir, size_t dirs)
{
avl_tree_t *cache;
rdsk_node_t *slice;
void *cookie;
int i, error;
*slice_cache = NULL;
cache = zutil_alloc(hdl, sizeof (avl_tree_t));
avl_create(cache, slice_cache_compare, sizeof (rdsk_node_t),
offsetof(rdsk_node_t, rn_node));
for (i = 0; i < dirs; i++) {
struct stat sbuf;
if (stat(dir[i], &sbuf) != 0) {
error = errno;
if (error == ENOENT)
continue;
zutil_error_aux(hdl, "%s", strerror(error));
(void) zutil_error_fmt(hdl, LPC_BADPATH, dgettext(
TEXT_DOMAIN, "cannot resolve path '%s'"), dir[i]);
goto error;
}
/*
* If dir[i] is a directory, we walk through it and add all
* the entries to the cache. If it's not a directory, we just
* add it to the cache.
*/
if (S_ISDIR(sbuf.st_mode)) {
if ((error = zpool_find_import_scan_dir(hdl, lock,
cache, dir[i], i)) != 0)
goto error;
} else {
if ((error = zpool_find_import_scan_path(hdl, lock,
cache, dir[i], i)) != 0)
goto error;
}
}
*slice_cache = cache;
return (0);
error:
cookie = NULL;
while ((slice = avl_destroy_nodes(cache, &cookie)) != NULL) {
free(slice->rn_name);
free(slice);
}
free(cache);
return (error);
}
/*
* Given a list of directories to search, find all pools stored on disk. This
* includes partial pools which are not available to import. If no args are
* given (argc is 0), then the default directory (/dev/dsk) is searched.
* poolname or guid (but not both) are provided by the caller when trying
* to import a specific pool.
*/
static nvlist_t *
zpool_find_import_impl(libpc_handle_t *hdl, importargs_t *iarg,
pthread_mutex_t *lock, avl_tree_t *cache)
{
(void) lock;
nvlist_t *ret = NULL;
pool_list_t pools = { 0 };
pool_entry_t *pe, *penext;
vdev_entry_t *ve, *venext;
config_entry_t *ce, *cenext;
name_entry_t *ne, *nenext;
rdsk_node_t *slice;
void *cookie;
tpool_t *t;
verify(iarg->poolname == NULL || iarg->guid == 0);
/*
* Create a thread pool to parallelize the process of reading and
* validating labels, a large number of threads can be used due to
* minimal contention.
*/
t = tpool_create(1, 2 * sysconf(_SC_NPROCESSORS_ONLN), 0, NULL);
for (slice = avl_first(cache); slice;
(slice = avl_walk(cache, slice, AVL_AFTER)))
(void) tpool_dispatch(t, zpool_open_func, slice);
tpool_wait(t);
tpool_destroy(t);
/*
* Process the cache, filtering out any entries which are not
* for the specified pool then adding matching label configs.
*/
cookie = NULL;
while ((slice = avl_destroy_nodes(cache, &cookie)) != NULL) {
if (slice->rn_config != NULL) {
nvlist_t *config = slice->rn_config;
boolean_t matched = B_TRUE;
boolean_t aux = B_FALSE;
int fd;
/*
* Check if it's a spare or l2cache device. If it is,
* we need to skip the name and guid check since they
* don't exist on aux device label.
*/
if (iarg->poolname != NULL || iarg->guid != 0) {
uint64_t state;
aux = nvlist_lookup_uint64(config,
ZPOOL_CONFIG_POOL_STATE, &state) == 0 &&
(state == POOL_STATE_SPARE ||
state == POOL_STATE_L2CACHE);
}
if (iarg->poolname != NULL && !aux) {
const char *pname;
matched = nvlist_lookup_string(config,
ZPOOL_CONFIG_POOL_NAME, &pname) == 0 &&
strcmp(iarg->poolname, pname) == 0;
} else if (iarg->guid != 0 && !aux) {
uint64_t this_guid;
matched = nvlist_lookup_uint64(config,
ZPOOL_CONFIG_POOL_GUID, &this_guid) == 0 &&
iarg->guid == this_guid;
}
if (matched) {
/*
* Verify all remaining entries can be opened
* exclusively. This will prune all underlying
* multipath devices which otherwise could
* result in the vdev appearing as UNAVAIL.
*
* Under zdb, this step isn't required and
* would prevent a zdb -e of active pools with
* no cachefile.
*/
fd = open(slice->rn_name,
O_RDONLY | O_EXCL | O_CLOEXEC);
if (fd >= 0 || iarg->can_be_active) {
if (fd >= 0)
close(fd);
add_config(hdl, &pools,
slice->rn_name, slice->rn_order,
slice->rn_num_labels, config);
}
}
nvlist_free(config);
}
free(slice->rn_name);
free(slice);
}
avl_destroy(cache);
free(cache);
ret = get_configs(hdl, &pools, iarg->can_be_active, iarg->policy);
for (pe = pools.pools; pe != NULL; pe = penext) {
penext = pe->pe_next;
for (ve = pe->pe_vdevs; ve != NULL; ve = venext) {
venext = ve->ve_next;
for (ce = ve->ve_configs; ce != NULL; ce = cenext) {
cenext = ce->ce_next;
nvlist_free(ce->ce_config);
free(ce);
}
free(ve);
}
free(pe);
}
for (ne = pools.names; ne != NULL; ne = nenext) {
nenext = ne->ne_next;
free(ne->ne_name);
free(ne);
}
return (ret);
}
/*
* Given a config, discover the paths for the devices which
* exist in the config.
*/
static int
discover_cached_paths(libpc_handle_t *hdl, nvlist_t *nv,
avl_tree_t *cache, pthread_mutex_t *lock)
{
const char *path = NULL;
ssize_t dl;
uint_t children;
nvlist_t **child;
if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children) == 0) {
for (int c = 0; c < children; c++) {
discover_cached_paths(hdl, child[c], cache, lock);
}
}
/*
* Once we have the path, we need to add the directory to
* our directory cache.
*/
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0) {
int ret;
char c = '\0';
if ((dl = zfs_dirnamelen(path)) == -1) {
path = ".";
} else {
c = path[dl];
((char *)path)[dl] = '\0';
}
ret = zpool_find_import_scan_dir(hdl, lock, cache,
path, 0);
if (c != '\0')
((char *)path)[dl] = c;
return (ret);
}
return (0);
}
/*
* Given a cache file, return the contents as a list of importable pools.
* poolname or guid (but not both) are provided by the caller when trying
* to import a specific pool.
*/
static nvlist_t *
zpool_find_import_cached(libpc_handle_t *hdl, importargs_t *iarg)
{
char *buf;
int fd;
struct stat64 statbuf;
nvlist_t *raw, *src, *dst;
nvlist_t *pools;
nvpair_t *elem;
const char *name;
uint64_t this_guid;
boolean_t active;
verify(iarg->poolname == NULL || iarg->guid == 0);
if ((fd = open(iarg->cachefile, O_RDONLY | O_CLOEXEC)) < 0) {
zutil_error_aux(hdl, "%s", strerror(errno));
(void) zutil_error(hdl, LPC_BADCACHE, dgettext(TEXT_DOMAIN,
"failed to open cache file"));
return (NULL);
}
if (fstat64(fd, &statbuf) != 0) {
zutil_error_aux(hdl, "%s", strerror(errno));
(void) close(fd);
(void) zutil_error(hdl, LPC_BADCACHE, dgettext(TEXT_DOMAIN,
"failed to get size of cache file"));
return (NULL);
}
if ((buf = zutil_alloc(hdl, statbuf.st_size)) == NULL) {
(void) close(fd);
return (NULL);
}
if (read(fd, buf, statbuf.st_size) != statbuf.st_size) {
(void) close(fd);
free(buf);
(void) zutil_error(hdl, LPC_BADCACHE, dgettext(TEXT_DOMAIN,
"failed to read cache file contents"));
return (NULL);
}
(void) close(fd);
if (nvlist_unpack(buf, statbuf.st_size, &raw, 0) != 0) {
free(buf);
(void) zutil_error(hdl, LPC_BADCACHE, dgettext(TEXT_DOMAIN,
"invalid or corrupt cache file contents"));
return (NULL);
}
free(buf);
/*
* Go through and get the current state of the pools and refresh their
* state.
*/
if (nvlist_alloc(&pools, 0, 0) != 0) {
(void) zutil_no_memory(hdl);
nvlist_free(raw);
return (NULL);
}
elem = NULL;
while ((elem = nvlist_next_nvpair(raw, elem)) != NULL) {
src = fnvpair_value_nvlist(elem);
name = fnvlist_lookup_string(src, ZPOOL_CONFIG_POOL_NAME);
if (iarg->poolname != NULL && strcmp(iarg->poolname, name) != 0)
continue;
this_guid = fnvlist_lookup_uint64(src, ZPOOL_CONFIG_POOL_GUID);
if (iarg->guid != 0 && iarg->guid != this_guid)
continue;
if (zutil_pool_active(hdl, name, this_guid, &active) != 0) {
nvlist_free(raw);
nvlist_free(pools);
return (NULL);
}
if (active)
continue;
if (iarg->scan) {
uint64_t saved_guid = iarg->guid;
const char *saved_poolname = iarg->poolname;
pthread_mutex_t lock;
/*
* Create the device cache that will hold the
* devices we will scan based on the cachefile.
* This will get destroyed and freed by
* zpool_find_import_impl.
*/
avl_tree_t *cache = zutil_alloc(hdl,
sizeof (avl_tree_t));
avl_create(cache, slice_cache_compare,
sizeof (rdsk_node_t),
offsetof(rdsk_node_t, rn_node));
nvlist_t *nvroot = fnvlist_lookup_nvlist(src,
ZPOOL_CONFIG_VDEV_TREE);
/*
* We only want to find the pool with this_guid.
* We will reset these values back later.
*/
iarg->guid = this_guid;
iarg->poolname = NULL;
/*
* We need to build up a cache of devices that exists
* in the paths pointed to by the cachefile. This allows
* us to preserve the device namespace that was
* originally specified by the user but also lets us
* scan devices in those directories in case they had
* been renamed.
*/
pthread_mutex_init(&lock, NULL);
discover_cached_paths(hdl, nvroot, cache, &lock);
nvlist_t *nv = zpool_find_import_impl(hdl, iarg,
&lock, cache);
pthread_mutex_destroy(&lock);
/*
* zpool_find_import_impl will return back
* a list of pools that it found based on the
* device cache. There should only be one pool
* since we're looking for a specific guid.
* We will use that pool to build up the final
* pool nvlist which is returned back to the
* caller.
*/
nvpair_t *pair = nvlist_next_nvpair(nv, NULL);
if (pair == NULL)
continue;
fnvlist_add_nvlist(pools, nvpair_name(pair),
fnvpair_value_nvlist(pair));
VERIFY3P(nvlist_next_nvpair(nv, pair), ==, NULL);
iarg->guid = saved_guid;
iarg->poolname = saved_poolname;
continue;
}
if (nvlist_add_string(src, ZPOOL_CONFIG_CACHEFILE,
iarg->cachefile) != 0) {
(void) zutil_no_memory(hdl);
nvlist_free(raw);
nvlist_free(pools);
return (NULL);
}
update_vdevs_config_dev_sysfs_path(src);
if ((dst = zutil_refresh_config(hdl, src)) == NULL) {
nvlist_free(raw);
nvlist_free(pools);
return (NULL);
}
if (nvlist_add_nvlist(pools, nvpair_name(elem), dst) != 0) {
(void) zutil_no_memory(hdl);
nvlist_free(dst);
nvlist_free(raw);
nvlist_free(pools);
return (NULL);
}
nvlist_free(dst);
}
nvlist_free(raw);
return (pools);
}
static nvlist_t *
zpool_find_import(libpc_handle_t *hdl, importargs_t *iarg)
{
pthread_mutex_t lock;
avl_tree_t *cache;
nvlist_t *pools = NULL;
verify(iarg->poolname == NULL || iarg->guid == 0);
pthread_mutex_init(&lock, NULL);
/*
* Locate pool member vdevs by blkid or by directory scanning.
* On success a newly allocated AVL tree which is populated with an
* entry for each discovered vdev will be returned in the cache.
* It's the caller's responsibility to consume and destroy this tree.
*/
if (iarg->scan || iarg->paths != 0) {
size_t dirs = iarg->paths;
const char * const *dir = (const char * const *)iarg->path;
if (dirs == 0)
dir = zpool_default_search_paths(&dirs);
if (zpool_find_import_scan(hdl, &lock, &cache,
dir, dirs) != 0) {
pthread_mutex_destroy(&lock);
return (NULL);
}
} else {
if (zpool_find_import_blkid(hdl, &lock, &cache) != 0) {
pthread_mutex_destroy(&lock);
return (NULL);
}
}
pools = zpool_find_import_impl(hdl, iarg, &lock, cache);
pthread_mutex_destroy(&lock);
return (pools);
}
nvlist_t *
zpool_search_import(libpc_handle_t *hdl, importargs_t *import)
{
nvlist_t *pools = NULL;
verify(import->poolname == NULL || import->guid == 0);
if (import->cachefile != NULL)
pools = zpool_find_import_cached(hdl, import);
else
pools = zpool_find_import(hdl, import);
if ((pools == NULL || nvlist_empty(pools)) &&
hdl->lpc_open_access_error && geteuid() != 0) {
(void) zutil_error(hdl, LPC_EACCESS, dgettext(TEXT_DOMAIN,
"no pools found"));
}
return (pools);
}
static boolean_t
pool_match(nvlist_t *cfg, const char *tgt)
{
uint64_t v, guid = strtoull(tgt, NULL, 0);
const char *s;
if (guid != 0) {
if (nvlist_lookup_uint64(cfg, ZPOOL_CONFIG_POOL_GUID, &v) == 0)
return (v == guid);
} else {
if (nvlist_lookup_string(cfg, ZPOOL_CONFIG_POOL_NAME, &s) == 0)
return (strcmp(s, tgt) == 0);
}
return (B_FALSE);
}
int
zpool_find_config(libpc_handle_t *hdl, const char *target, nvlist_t **configp,
importargs_t *args)
{
nvlist_t *pools;
nvlist_t *match = NULL;
nvlist_t *config = NULL;
char *sepp = NULL;
int count = 0;
char *targetdup = strdup(target);
if (targetdup == NULL)
return (ENOMEM);
*configp = NULL;
if ((sepp = strpbrk(targetdup, "/@")) != NULL)
*sepp = '\0';
pools = zpool_search_import(hdl, args);
if (pools != NULL) {
nvpair_t *elem = NULL;
while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) {
VERIFY0(nvpair_value_nvlist(elem, &config));
if (pool_match(config, targetdup)) {
count++;
if (match != NULL) {
/* multiple matches found */
continue;
} else {
match = fnvlist_dup(config);
}
}
}
fnvlist_free(pools);
}
if (count == 0) {
free(targetdup);
return (ENOENT);
}
if (count > 1) {
free(targetdup);
fnvlist_free(match);
return (EINVAL);
}
*configp = match;
free(targetdup);
return (0);
}
/*
* Internal function for iterating over the vdevs.
*
* For each vdev, func() will be called and will be passed 'zhp' (which is
* typically the zpool_handle_t cast as a void pointer), the vdev's nvlist, and
* a user-defined data pointer).
*
* The return values from all the func() calls will be OR'd together and
* returned.
*/
int
for_each_vdev_cb(void *zhp, nvlist_t *nv, pool_vdev_iter_f func,
void *data)
{
nvlist_t **child;
uint_t c, children;
int ret = 0;
int i;
const char *type;
const char *list[] = {
ZPOOL_CONFIG_SPARES,
ZPOOL_CONFIG_L2CACHE,
ZPOOL_CONFIG_CHILDREN
};
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) != 0)
return (ret);
- /* Don't run our function on root or indirect vdevs */
- if ((strcmp(type, VDEV_TYPE_ROOT) != 0) &&
- (strcmp(type, VDEV_TYPE_INDIRECT) != 0)) {
+ /* Don't run our function on indirect vdevs */
+ if (strcmp(type, VDEV_TYPE_INDIRECT) != 0) {
ret |= func(zhp, nv, data);
}
for (i = 0; i < ARRAY_SIZE(list); i++) {
if (nvlist_lookup_nvlist_array(nv, list[i], &child,
&children) == 0) {
for (c = 0; c < children; c++) {
uint64_t ishole = 0;
(void) nvlist_lookup_uint64(child[c],
ZPOOL_CONFIG_IS_HOLE, &ishole);
if (ishole)
continue;
ret |= for_each_vdev_cb(zhp, child[c],
func, data);
}
}
}
return (ret);
}
/*
* Given an ZPOOL_CONFIG_VDEV_TREE nvpair, iterate over all the vdevs, calling
* func() for each one. func() is passed the vdev's nvlist and an optional
* user-defined 'data' pointer.
*/
int
for_each_vdev_in_nvlist(nvlist_t *nvroot, pool_vdev_iter_f func, void *data)
{
return (for_each_vdev_cb(NULL, nvroot, func, data));
}
diff --git a/sys/contrib/openzfs/man/man4/zfs.4 b/sys/contrib/openzfs/man/man4/zfs.4
index e8e2cfec61e8..d529147464fe 100644
--- a/sys/contrib/openzfs/man/man4/zfs.4
+++ b/sys/contrib/openzfs/man/man4/zfs.4
@@ -1,2585 +1,2593 @@
.\"
.\" Copyright (c) 2013 by Turbo Fredriksson <turbo@bayour.com>. All rights reserved.
.\" Copyright (c) 2019, 2021 by Delphix. All rights reserved.
.\" Copyright (c) 2019 Datto Inc.
.\" The contents of this file are subject to the terms of the Common Development
.\" and Distribution License (the "License"). You may not use this file except
.\" in compliance with the License. You can obtain a copy of the license at
.\" usr/src/OPENSOLARIS.LICENSE or https://opensource.org/licenses/CDDL-1.0.
.\"
.\" See the License for the specific language governing permissions and
.\" limitations under the License. When distributing Covered Code, include this
.\" CDDL HEADER in each file and include the License file at
.\" usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this
.\" CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your
.\" own identifying information:
.\" Portions Copyright [yyyy] [name of copyright owner]
.\"
.Dd January 10, 2023
.Dt ZFS 4
.Os
.
.Sh NAME
.Nm zfs
.Nd tuning of the ZFS kernel module
.
.Sh DESCRIPTION
The ZFS module supports these parameters:
.Bl -tag -width Ds
.It Sy dbuf_cache_max_bytes Ns = Ns Sy UINT64_MAX Ns B Pq u64
Maximum size in bytes of the dbuf cache.
The target size is determined by the MIN versus
.No 1/2^ Ns Sy dbuf_cache_shift Pq 1/32nd
of the target ARC size.
The behavior of the dbuf cache and its associated settings
can be observed via the
.Pa /proc/spl/kstat/zfs/dbufstats
kstat.
.
.It Sy dbuf_metadata_cache_max_bytes Ns = Ns Sy UINT64_MAX Ns B Pq u64
Maximum size in bytes of the metadata dbuf cache.
The target size is determined by the MIN versus
.No 1/2^ Ns Sy dbuf_metadata_cache_shift Pq 1/64th
of the target ARC size.
The behavior of the metadata dbuf cache and its associated settings
can be observed via the
.Pa /proc/spl/kstat/zfs/dbufstats
kstat.
.
.It Sy dbuf_cache_hiwater_pct Ns = Ns Sy 10 Ns % Pq uint
The percentage over
.Sy dbuf_cache_max_bytes
when dbufs must be evicted directly.
.
.It Sy dbuf_cache_lowater_pct Ns = Ns Sy 10 Ns % Pq uint
The percentage below
.Sy dbuf_cache_max_bytes
when the evict thread stops evicting dbufs.
.
.It Sy dbuf_cache_shift Ns = Ns Sy 5 Pq uint
Set the size of the dbuf cache
.Pq Sy dbuf_cache_max_bytes
to a log2 fraction of the target ARC size.
.
.It Sy dbuf_metadata_cache_shift Ns = Ns Sy 6 Pq uint
Set the size of the dbuf metadata cache
.Pq Sy dbuf_metadata_cache_max_bytes
to a log2 fraction of the target ARC size.
.
.It Sy dbuf_mutex_cache_shift Ns = Ns Sy 0 Pq uint
Set the size of the mutex array for the dbuf cache.
When set to
.Sy 0
the array is dynamically sized based on total system memory.
.
.It Sy dmu_object_alloc_chunk_shift Ns = Ns Sy 7 Po 128 Pc Pq uint
dnode slots allocated in a single operation as a power of 2.
The default value minimizes lock contention for the bulk operation performed.
.
.It Sy dmu_prefetch_max Ns = Ns Sy 134217728 Ns B Po 128 MiB Pc Pq uint
Limit the amount we can prefetch with one call to this amount in bytes.
This helps to limit the amount of memory that can be used by prefetching.
.
.It Sy ignore_hole_birth Pq int
Alias for
.Sy send_holes_without_birth_time .
.
.It Sy l2arc_feed_again Ns = Ns Sy 1 Ns | Ns 0 Pq int
Turbo L2ARC warm-up.
When the L2ARC is cold the fill interval will be set as fast as possible.
.
.It Sy l2arc_feed_min_ms Ns = Ns Sy 200 Pq u64
Min feed interval in milliseconds.
Requires
.Sy l2arc_feed_again Ns = Ns Ar 1
and only applicable in related situations.
.
.It Sy l2arc_feed_secs Ns = Ns Sy 1 Pq u64
Seconds between L2ARC writing.
.
.It Sy l2arc_headroom Ns = Ns Sy 2 Pq u64
How far through the ARC lists to search for L2ARC cacheable content,
expressed as a multiplier of
.Sy l2arc_write_max .
ARC persistence across reboots can be achieved with persistent L2ARC
by setting this parameter to
.Sy 0 ,
allowing the full length of ARC lists to be searched for cacheable content.
.
.It Sy l2arc_headroom_boost Ns = Ns Sy 200 Ns % Pq u64
Scales
.Sy l2arc_headroom
by this percentage when L2ARC contents are being successfully compressed
before writing.
A value of
.Sy 100
disables this feature.
.
.It Sy l2arc_exclude_special Ns = Ns Sy 0 Ns | Ns 1 Pq int
Controls whether buffers present on special vdevs are eligible for caching
into L2ARC.
If set to 1, exclude dbufs on special vdevs from being cached to L2ARC.
.
.It Sy l2arc_mfuonly Ns = Ns Sy 0 Ns | Ns 1 Pq int
Controls whether only MFU metadata and data are cached from ARC into L2ARC.
This may be desired to avoid wasting space on L2ARC when reading/writing large
amounts of data that are not expected to be accessed more than once.
.Pp
The default is off,
meaning both MRU and MFU data and metadata are cached.
When turning off this feature, some MRU buffers will still be present
in ARC and eventually cached on L2ARC.
.No If Sy l2arc_noprefetch Ns = Ns Sy 0 ,
some prefetched buffers will be cached to L2ARC, and those might later
transition to MRU, in which case the
.Sy l2arc_mru_asize No arcstat will not be Sy 0 .
.Pp
Regardless of
.Sy l2arc_noprefetch ,
some MFU buffers might be evicted from ARC,
accessed later on as prefetches and transition to MRU as prefetches.
If accessed again they are counted as MRU and the
.Sy l2arc_mru_asize No arcstat will not be Sy 0 .
.Pp
The ARC status of L2ARC buffers when they were first cached in
L2ARC can be seen in the
.Sy l2arc_mru_asize , Sy l2arc_mfu_asize , No and Sy l2arc_prefetch_asize
arcstats when importing the pool or onlining a cache
device if persistent L2ARC is enabled.
.Pp
The
.Sy evict_l2_eligible_mru
arcstat does not take into account if this option is enabled as the information
provided by the
.Sy evict_l2_eligible_m[rf]u
arcstats can be used to decide if toggling this option is appropriate
for the current workload.
.
.It Sy l2arc_meta_percent Ns = Ns Sy 33 Ns % Pq uint
Percent of ARC size allowed for L2ARC-only headers.
Since L2ARC buffers are not evicted on memory pressure,
too many headers on a system with an irrationally large L2ARC
can render it slow or unusable.
This parameter limits L2ARC writes and rebuilds to achieve the target.
.
.It Sy l2arc_trim_ahead Ns = Ns Sy 0 Ns % Pq u64
Trims ahead of the current write size
.Pq Sy l2arc_write_max
on L2ARC devices by this percentage of write size if we have filled the device.
If set to
.Sy 100
we TRIM twice the space required to accommodate upcoming writes.
A minimum of
.Sy 64 MiB
will be trimmed.
It also enables TRIM of the whole L2ARC device upon creation
or addition to an existing pool or if the header of the device is
invalid upon importing a pool or onlining a cache device.
A value of
.Sy 0
disables TRIM on L2ARC altogether and is the default as it can put significant
stress on the underlying storage devices.
This will vary depending of how well the specific device handles these commands.
.
.It Sy l2arc_noprefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int
Do not write buffers to L2ARC if they were prefetched but not used by
applications.
In case there are prefetched buffers in L2ARC and this option
is later set, we do not read the prefetched buffers from L2ARC.
Unsetting this option is useful for caching sequential reads from the
disks to L2ARC and serve those reads from L2ARC later on.
This may be beneficial in case the L2ARC device is significantly faster
in sequential reads than the disks of the pool.
.Pp
Use
.Sy 1
to disable and
.Sy 0
to enable caching/reading prefetches to/from L2ARC.
.
.It Sy l2arc_norw Ns = Ns Sy 0 Ns | Ns 1 Pq int
No reads during writes.
.
.It Sy l2arc_write_boost Ns = Ns Sy 8388608 Ns B Po 8 MiB Pc Pq u64
Cold L2ARC devices will have
.Sy l2arc_write_max
increased by this amount while they remain cold.
.
.It Sy l2arc_write_max Ns = Ns Sy 8388608 Ns B Po 8 MiB Pc Pq u64
Max write bytes per interval.
.
.It Sy l2arc_rebuild_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Rebuild the L2ARC when importing a pool (persistent L2ARC).
This can be disabled if there are problems importing a pool
or attaching an L2ARC device (e.g. the L2ARC device is slow
in reading stored log metadata, or the metadata
has become somehow fragmented/unusable).
.
.It Sy l2arc_rebuild_blocks_min_l2size Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64
Mininum size of an L2ARC device required in order to write log blocks in it.
The log blocks are used upon importing the pool to rebuild the persistent L2ARC.
.Pp
For L2ARC devices less than 1 GiB, the amount of data
.Fn l2arc_evict
evicts is significant compared to the amount of restored L2ARC data.
In this case, do not write log blocks in L2ARC in order not to waste space.
.
.It Sy metaslab_aliquot Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
Metaslab granularity, in bytes.
This is roughly similar to what would be referred to as the "stripe size"
in traditional RAID arrays.
In normal operation, ZFS will try to write this amount of data to each disk
before moving on to the next top-level vdev.
.
.It Sy metaslab_bias_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable metaslab group biasing based on their vdevs' over- or under-utilization
relative to the pool.
.
.It Sy metaslab_force_ganging Ns = Ns Sy 16777217 Ns B Po 16 MiB + 1 B Pc Pq u64
Make some blocks above a certain size be gang blocks.
This option is used by the test suite to facilitate testing.
.
.It Sy zfs_default_bs Ns = Ns Sy 9 Po 512 B Pc Pq int
Default dnode block size as a power of 2.
.
.It Sy zfs_default_ibs Ns = Ns Sy 17 Po 128 KiB Pc Pq int
Default dnode indirect block size as a power of 2.
.
.It Sy zfs_history_output_max Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
When attempting to log an output nvlist of an ioctl in the on-disk history,
the output will not be stored if it is larger than this size (in bytes).
This must be less than
.Sy DMU_MAX_ACCESS Pq 64 MiB .
This applies primarily to
.Fn zfs_ioc_channel_program Pq cf. Xr zfs-program 8 .
.
.It Sy zfs_keep_log_spacemaps_at_export Ns = Ns Sy 0 Ns | Ns 1 Pq int
Prevent log spacemaps from being destroyed during pool exports and destroys.
.
.It Sy zfs_metaslab_segment_weight_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable/disable segment-based metaslab selection.
.
.It Sy zfs_metaslab_switch_threshold Ns = Ns Sy 2 Pq int
When using segment-based metaslab selection, continue allocating
from the active metaslab until this option's
worth of buckets have been exhausted.
.
.It Sy metaslab_debug_load Ns = Ns Sy 0 Ns | Ns 1 Pq int
Load all metaslabs during pool import.
.
.It Sy metaslab_debug_unload Ns = Ns Sy 0 Ns | Ns 1 Pq int
Prevent metaslabs from being unloaded.
.
.It Sy metaslab_fragmentation_factor_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable use of the fragmentation metric in computing metaslab weights.
.
.It Sy metaslab_df_max_search Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint
Maximum distance to search forward from the last offset.
Without this limit, fragmented pools can see
.Em >100`000
iterations and
.Fn metaslab_block_picker
becomes the performance limiting factor on high-performance storage.
.Pp
With the default setting of
.Sy 16 MiB ,
we typically see less than
.Em 500
iterations, even with very fragmented
.Sy ashift Ns = Ns Sy 9
pools.
The maximum number of iterations possible is
.Sy metaslab_df_max_search / 2^(ashift+1) .
With the default setting of
.Sy 16 MiB
this is
.Em 16*1024 Pq with Sy ashift Ns = Ns Sy 9
or
.Em 2*1024 Pq with Sy ashift Ns = Ns Sy 12 .
.
.It Sy metaslab_df_use_largest_segment Ns = Ns Sy 0 Ns | Ns 1 Pq int
If not searching forward (due to
.Sy metaslab_df_max_search , metaslab_df_free_pct ,
.No or Sy metaslab_df_alloc_threshold ) ,
this tunable controls which segment is used.
If set, we will use the largest free segment.
If unset, we will use a segment of at least the requested size.
.
.It Sy zfs_metaslab_max_size_cache_sec Ns = Ns Sy 3600 Ns s Po 1 hour Pc Pq u64
When we unload a metaslab, we cache the size of the largest free chunk.
We use that cached size to determine whether or not to load a metaslab
for a given allocation.
As more frees accumulate in that metaslab while it's unloaded,
the cached max size becomes less and less accurate.
After a number of seconds controlled by this tunable,
we stop considering the cached max size and start
considering only the histogram instead.
.
.It Sy zfs_metaslab_mem_limit Ns = Ns Sy 25 Ns % Pq uint
When we are loading a new metaslab, we check the amount of memory being used
to store metaslab range trees.
If it is over a threshold, we attempt to unload the least recently used metaslab
to prevent the system from clogging all of its memory with range trees.
This tunable sets the percentage of total system memory that is the threshold.
.
.It Sy zfs_metaslab_try_hard_before_gang Ns = Ns Sy 0 Ns | Ns 1 Pq int
.Bl -item -compact
.It
If unset, we will first try normal allocation.
.It
If that fails then we will do a gang allocation.
.It
If that fails then we will do a "try hard" gang allocation.
.It
If that fails then we will have a multi-layer gang block.
.El
.Pp
.Bl -item -compact
.It
If set, we will first try normal allocation.
.It
If that fails then we will do a "try hard" allocation.
.It
If that fails we will do a gang allocation.
.It
If that fails we will do a "try hard" gang allocation.
.It
If that fails then we will have a multi-layer gang block.
.El
.
.It Sy zfs_metaslab_find_max_tries Ns = Ns Sy 100 Pq uint
When not trying hard, we only consider this number of the best metaslabs.
This improves performance, especially when there are many metaslabs per vdev
and the allocation can't actually be satisfied
(so we would otherwise iterate all metaslabs).
.
.It Sy zfs_vdev_default_ms_count Ns = Ns Sy 200 Pq uint
When a vdev is added, target this number of metaslabs per top-level vdev.
.
.It Sy zfs_vdev_default_ms_shift Ns = Ns Sy 29 Po 512 MiB Pc Pq uint
-Default limit for metaslab size.
+Default lower limit for metaslab size.
+.
+.It Sy zfs_vdev_max_ms_shift Ns = Ns Sy 34 Po 16 GiB Pc Pq uint
+Default upper limit for metaslab size.
.
.It Sy zfs_vdev_max_auto_ashift Ns = Ns Sy 14 Pq uint
Maximum ashift used when optimizing for logical \[->] physical sector size on
new
top-level vdevs.
May be increased up to
.Sy ASHIFT_MAX Po 16 Pc ,
but this may negatively impact pool space efficiency.
.
.It Sy zfs_vdev_min_auto_ashift Ns = Ns Sy ASHIFT_MIN Po 9 Pc Pq uint
Minimum ashift used when creating new top-level vdevs.
.
.It Sy zfs_vdev_min_ms_count Ns = Ns Sy 16 Pq uint
Minimum number of metaslabs to create in a top-level vdev.
.
.It Sy vdev_validate_skip Ns = Ns Sy 0 Ns | Ns 1 Pq int
Skip label validation steps during pool import.
Changing is not recommended unless you know what you're doing
and are recovering a damaged label.
.
.It Sy zfs_vdev_ms_count_limit Ns = Ns Sy 131072 Po 128k Pc Pq uint
Practical upper limit of total metaslabs per top-level vdev.
.
.It Sy metaslab_preload_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable metaslab group preloading.
.
.It Sy metaslab_lba_weighting_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Give more weight to metaslabs with lower LBAs,
assuming they have greater bandwidth,
as is typically the case on a modern constant angular velocity disk drive.
.
.It Sy metaslab_unload_delay Ns = Ns Sy 32 Pq uint
After a metaslab is used, we keep it loaded for this many TXGs, to attempt to
reduce unnecessary reloading.
Note that both this many TXGs and
.Sy metaslab_unload_delay_ms
milliseconds must pass before unloading will occur.
.
.It Sy metaslab_unload_delay_ms Ns = Ns Sy 600000 Ns ms Po 10 min Pc Pq uint
After a metaslab is used, we keep it loaded for this many milliseconds,
to attempt to reduce unnecessary reloading.
Note, that both this many milliseconds and
.Sy metaslab_unload_delay
TXGs must pass before unloading will occur.
.
.It Sy reference_history Ns = Ns Sy 3 Pq uint
Maximum reference holders being tracked when reference_tracking_enable is
active.
.
.It Sy reference_tracking_enable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Track reference holders to
.Sy refcount_t
objects (debug builds only).
.
.It Sy send_holes_without_birth_time Ns = Ns Sy 1 Ns | Ns 0 Pq int
When set, the
.Sy hole_birth
optimization will not be used, and all holes will always be sent during a
.Nm zfs Cm send .
This is useful if you suspect your datasets are affected by a bug in
.Sy hole_birth .
.
.It Sy spa_config_path Ns = Ns Pa /etc/zfs/zpool.cache Pq charp
SPA config file.
.
.It Sy spa_asize_inflation Ns = Ns Sy 24 Pq uint
Multiplication factor used to estimate actual disk consumption from the
size of data being written.
The default value is a worst case estimate,
but lower values may be valid for a given pool depending on its configuration.
Pool administrators who understand the factors involved
may wish to specify a more realistic inflation factor,
particularly if they operate close to quota or capacity limits.
.
.It Sy spa_load_print_vdev_tree Ns = Ns Sy 0 Ns | Ns 1 Pq int
Whether to print the vdev tree in the debugging message buffer during pool
import.
.
.It Sy spa_load_verify_data Ns = Ns Sy 1 Ns | Ns 0 Pq int
Whether to traverse data blocks during an "extreme rewind"
.Pq Fl X
import.
.Pp
An extreme rewind import normally performs a full traversal of all
blocks in the pool for verification.
If this parameter is unset, the traversal skips non-metadata blocks.
It can be toggled once the
import has started to stop or start the traversal of non-metadata blocks.
.
.It Sy spa_load_verify_metadata Ns = Ns Sy 1 Ns | Ns 0 Pq int
Whether to traverse blocks during an "extreme rewind"
.Pq Fl X
pool import.
.Pp
An extreme rewind import normally performs a full traversal of all
blocks in the pool for verification.
If this parameter is unset, the traversal is not performed.
It can be toggled once the import has started to stop or start the traversal.
.
.It Sy spa_load_verify_shift Ns = Ns Sy 4 Po 1/16th Pc Pq uint
Sets the maximum number of bytes to consume during pool import to the log2
fraction of the target ARC size.
.
.It Sy spa_slop_shift Ns = Ns Sy 5 Po 1/32nd Pc Pq int
Normally, we don't allow the last
.Sy 3.2% Pq Sy 1/2^spa_slop_shift
of space in the pool to be consumed.
This ensures that we don't run the pool completely out of space,
due to unaccounted changes (e.g. to the MOS).
It also limits the worst-case time to allocate space.
If we have less than this amount of free space,
most ZPL operations (e.g. write, create) will return
.Sy ENOSPC .
.
.It Sy spa_upgrade_errlog_limit Ns = Ns Sy 0 Pq uint
Limits the number of on-disk error log entries that will be converted to the
new format when enabling the
.Sy head_errlog
feature.
The default is to convert all log entries.
.
.It Sy vdev_removal_max_span Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint
During top-level vdev removal, chunks of data are copied from the vdev
which may include free space in order to trade bandwidth for IOPS.
This parameter determines the maximum span of free space, in bytes,
which will be included as "unnecessary" data in a chunk of copied data.
.Pp
The default value here was chosen to align with
.Sy zfs_vdev_read_gap_limit ,
which is a similar concept when doing
regular reads (but there's no reason it has to be the same).
.
.It Sy vdev_file_logical_ashift Ns = Ns Sy 9 Po 512 B Pc Pq u64
Logical ashift for file-based devices.
.
.It Sy vdev_file_physical_ashift Ns = Ns Sy 9 Po 512 B Pc Pq u64
Physical ashift for file-based devices.
.
.It Sy zap_iterate_prefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int
If set, when we start iterating over a ZAP object,
prefetch the entire object (all leaf blocks).
However, this is limited by
.Sy dmu_prefetch_max .
.
.It Sy zap_micro_max_size Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq int
Maximum micro ZAP size.
A micro ZAP is upgraded to a fat ZAP, once it grows beyond the specified size.
.
.It Sy zfetch_array_rd_sz Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
If prefetching is enabled, disable prefetching for reads larger than this size.
.
.It Sy zfetch_min_distance Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint
Min bytes to prefetch per stream.
Prefetch distance starts from the demand access size and quickly grows to
this value, doubling on each hit.
After that it may grow further by 1/8 per hit, but only if some prefetch
since last time haven't completed in time to satisfy demand request, i.e.
prefetch depth didn't cover the read latency or the pool got saturated.
.
.It Sy zfetch_max_distance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint
Max bytes to prefetch per stream.
.
.It Sy zfetch_max_idistance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint
Max bytes to prefetch indirects for per stream.
.
.It Sy zfetch_max_streams Ns = Ns Sy 8 Pq uint
Max number of streams per zfetch (prefetch streams per file).
.
.It Sy zfetch_min_sec_reap Ns = Ns Sy 1 Pq uint
Min time before inactive prefetch stream can be reclaimed
.
.It Sy zfetch_max_sec_reap Ns = Ns Sy 2 Pq uint
Max time before inactive prefetch stream can be deleted
.
.It Sy zfs_abd_scatter_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enables ARC from using scatter/gather lists and forces all allocations to be
linear in kernel memory.
Disabling can improve performance in some code paths
at the expense of fragmented kernel memory.
.
.It Sy zfs_abd_scatter_max_order Ns = Ns Sy MAX_ORDER\-1 Pq uint
Maximum number of consecutive memory pages allocated in a single block for
scatter/gather lists.
.Pp
The value of
.Sy MAX_ORDER
depends on kernel configuration.
.
.It Sy zfs_abd_scatter_min_size Ns = Ns Sy 1536 Ns B Po 1.5 KiB Pc Pq uint
This is the minimum allocation size that will use scatter (page-based) ABDs.
Smaller allocations will use linear ABDs.
.
.It Sy zfs_arc_dnode_limit Ns = Ns Sy 0 Ns B Pq u64
When the number of bytes consumed by dnodes in the ARC exceeds this number of
bytes, try to unpin some of it in response to demand for non-metadata.
This value acts as a ceiling to the amount of dnode metadata, and defaults to
.Sy 0 ,
which indicates that a percent which is based on
.Sy zfs_arc_dnode_limit_percent
of the ARC meta buffers that may be used for dnodes.
.It Sy zfs_arc_dnode_limit_percent Ns = Ns Sy 10 Ns % Pq u64
Percentage that can be consumed by dnodes of ARC meta buffers.
.Pp
See also
.Sy zfs_arc_dnode_limit ,
which serves a similar purpose but has a higher priority if nonzero.
.
.It Sy zfs_arc_dnode_reduce_percent Ns = Ns Sy 10 Ns % Pq u64
Percentage of ARC dnodes to try to scan in response to demand for non-metadata
when the number of bytes consumed by dnodes exceeds
.Sy zfs_arc_dnode_limit .
.
.It Sy zfs_arc_average_blocksize Ns = Ns Sy 8192 Ns B Po 8 KiB Pc Pq uint
The ARC's buffer hash table is sized based on the assumption of an average
block size of this value.
This works out to roughly 1 MiB of hash table per 1 GiB of physical memory
with 8-byte pointers.
For configurations with a known larger average block size,
this value can be increased to reduce the memory footprint.
.
.It Sy zfs_arc_eviction_pct Ns = Ns Sy 200 Ns % Pq uint
When
.Fn arc_is_overflowing ,
.Fn arc_get_data_impl
waits for this percent of the requested amount of data to be evicted.
For example, by default, for every
.Em 2 KiB
that's evicted,
.Em 1 KiB
of it may be "reused" by a new allocation.
Since this is above
.Sy 100 Ns % ,
it ensures that progress is made towards getting
.Sy arc_size No under Sy arc_c .
Since this is finite, it ensures that allocations can still happen,
even during the potentially long time that
.Sy arc_size No is more than Sy arc_c .
.
.It Sy zfs_arc_evict_batch_limit Ns = Ns Sy 10 Pq uint
Number ARC headers to evict per sub-list before proceeding to another sub-list.
This batch-style operation prevents entire sub-lists from being evicted at once
but comes at a cost of additional unlocking and locking.
.
.It Sy zfs_arc_grow_retry Ns = Ns Sy 0 Ns s Pq uint
If set to a non zero value, it will replace the
.Sy arc_grow_retry
value with this value.
The
.Sy arc_grow_retry
.No value Pq default Sy 5 Ns s
is the number of seconds the ARC will wait before
trying to resume growth after a memory pressure event.
.
.It Sy zfs_arc_lotsfree_percent Ns = Ns Sy 10 Ns % Pq int
Throttle I/O when free system memory drops below this percentage of total
system memory.
Setting this value to
.Sy 0
will disable the throttle.
.
.It Sy zfs_arc_max Ns = Ns Sy 0 Ns B Pq u64
Max size of ARC in bytes.
If
.Sy 0 ,
then the max size of ARC is determined by the amount of system memory installed.
Under Linux, half of system memory will be used as the limit.
Under
.Fx ,
the larger of
.Sy all_system_memory No \- Sy 1 GiB
and
.Sy 5/8 No \(mu Sy all_system_memory
will be used as the limit.
This value must be at least
.Sy 67108864 Ns B Pq 64 MiB .
.Pp
This value can be changed dynamically, with some caveats.
It cannot be set back to
.Sy 0
while running, and reducing it below the current ARC size will not cause
the ARC to shrink without memory pressure to induce shrinking.
.
.It Sy zfs_arc_meta_balance Ns = Ns Sy 500 Pq uint
Balance between metadata and data on ghost hits.
Values above 100 increase metadata caching by proportionally reducing effect
of ghost data hits on target data/metadata rate.
.
.It Sy zfs_arc_min Ns = Ns Sy 0 Ns B Pq u64
Min size of ARC in bytes.
.No If set to Sy 0 , arc_c_min
will default to consuming the larger of
.Sy 32 MiB
and
.Sy all_system_memory No / Sy 32 .
.
.It Sy zfs_arc_min_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 1s Pc Pq uint
Minimum time prefetched blocks are locked in the ARC.
.
.It Sy zfs_arc_min_prescient_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 6s Pc Pq uint
Minimum time "prescient prefetched" blocks are locked in the ARC.
These blocks are meant to be prefetched fairly aggressively ahead of
the code that may use them.
.
.It Sy zfs_arc_prune_task_threads Ns = Ns Sy 1 Pq int
Number of arc_prune threads.
.Fx
does not need more than one.
Linux may theoretically use one per mount point up to number of CPUs,
but that was not proven to be useful.
.
.It Sy zfs_max_missing_tvds Ns = Ns Sy 0 Pq int
Number of missing top-level vdevs which will be allowed during
pool import (only in read-only mode).
.
.It Sy zfs_max_nvlist_src_size Ns = Sy 0 Pq u64
Maximum size in bytes allowed to be passed as
.Sy zc_nvlist_src_size
for ioctls on
.Pa /dev/zfs .
This prevents a user from causing the kernel to allocate
an excessive amount of memory.
When the limit is exceeded, the ioctl fails with
.Sy EINVAL
and a description of the error is sent to the
.Pa zfs-dbgmsg
log.
This parameter should not need to be touched under normal circumstances.
If
.Sy 0 ,
equivalent to a quarter of the user-wired memory limit under
.Fx
and to
.Sy 134217728 Ns B Pq 128 MiB
under Linux.
.
.It Sy zfs_multilist_num_sublists Ns = Ns Sy 0 Pq uint
To allow more fine-grained locking, each ARC state contains a series
of lists for both data and metadata objects.
Locking is performed at the level of these "sub-lists".
This parameters controls the number of sub-lists per ARC state,
and also applies to other uses of the multilist data structure.
.Pp
If
.Sy 0 ,
equivalent to the greater of the number of online CPUs and
.Sy 4 .
.
.It Sy zfs_arc_overflow_shift Ns = Ns Sy 8 Pq int
The ARC size is considered to be overflowing if it exceeds the current
ARC target size
.Pq Sy arc_c
by thresholds determined by this parameter.
Exceeding by
.Sy ( arc_c No >> Sy zfs_arc_overflow_shift ) No / Sy 2
starts ARC reclamation process.
If that appears insufficient, exceeding by
.Sy ( arc_c No >> Sy zfs_arc_overflow_shift ) No \(mu Sy 1.5
blocks new buffer allocation until the reclaim thread catches up.
Started reclamation process continues till ARC size returns below the
target size.
.Pp
The default value of
.Sy 8
causes the ARC to start reclamation if it exceeds the target size by
.Em 0.2%
of the target size, and block allocations by
.Em 0.6% .
.
.It Sy zfs_arc_shrink_shift Ns = Ns Sy 0 Pq uint
If nonzero, this will update
.Sy arc_shrink_shift Pq default Sy 7
with the new value.
.
.It Sy zfs_arc_pc_percent Ns = Ns Sy 0 Ns % Po off Pc Pq uint
Percent of pagecache to reclaim ARC to.
.Pp
This tunable allows the ZFS ARC to play more nicely
with the kernel's LRU pagecache.
It can guarantee that the ARC size won't collapse under scanning
pressure on the pagecache, yet still allows the ARC to be reclaimed down to
.Sy zfs_arc_min
if necessary.
This value is specified as percent of pagecache size (as measured by
.Sy NR_FILE_PAGES ) ,
where that percent may exceed
.Sy 100 .
This
only operates during memory pressure/reclaim.
.
.It Sy zfs_arc_shrinker_limit Ns = Ns Sy 10000 Pq int
This is a limit on how many pages the ARC shrinker makes available for
eviction in response to one page allocation attempt.
Note that in practice, the kernel's shrinker can ask us to evict
up to about four times this for one allocation attempt.
.Pp
The default limit of
.Sy 10000 Pq in practice, Em 160 MiB No per allocation attempt with 4 KiB pages
limits the amount of time spent attempting to reclaim ARC memory to
less than 100 ms per allocation attempt,
even with a small average compressed block size of ~8 KiB.
.Pp
The parameter can be set to 0 (zero) to disable the limit,
and only applies on Linux.
.
.It Sy zfs_arc_sys_free Ns = Ns Sy 0 Ns B Pq u64
The target number of bytes the ARC should leave as free memory on the system.
If zero, equivalent to the bigger of
.Sy 512 KiB No and Sy all_system_memory/64 .
.
.It Sy zfs_autoimport_disable Ns = Ns Sy 1 Ns | Ns 0 Pq int
Disable pool import at module load by ignoring the cache file
.Pq Sy spa_config_path .
.
.It Sy zfs_checksum_events_per_second Ns = Ns Sy 20 Ns /s Pq uint
Rate limit checksum events to this many per second.
Note that this should not be set below the ZED thresholds
(currently 10 checksums over 10 seconds)
or else the daemon may not trigger any action.
.
.It Sy zfs_commit_timeout_pct Ns = Ns Sy 5 Ns % Pq uint
This controls the amount of time that a ZIL block (lwb) will remain "open"
when it isn't "full", and it has a thread waiting for it to be committed to
stable storage.
The timeout is scaled based on a percentage of the last lwb
latency to avoid significantly impacting the latency of each individual
transaction record (itx).
.
.It Sy zfs_condense_indirect_commit_entry_delay_ms Ns = Ns Sy 0 Ns ms Pq int
Vdev indirection layer (used for device removal) sleeps for this many
milliseconds during mapping generation.
Intended for use with the test suite to throttle vdev removal speed.
.
.It Sy zfs_condense_indirect_obsolete_pct Ns = Ns Sy 25 Ns % Pq uint
Minimum percent of obsolete bytes in vdev mapping required to attempt to
condense
.Pq see Sy zfs_condense_indirect_vdevs_enable .
Intended for use with the test suite
to facilitate triggering condensing as needed.
.
.It Sy zfs_condense_indirect_vdevs_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable condensing indirect vdev mappings.
When set, attempt to condense indirect vdev mappings
if the mapping uses more than
.Sy zfs_condense_min_mapping_bytes
bytes of memory and if the obsolete space map object uses more than
.Sy zfs_condense_max_obsolete_bytes
bytes on-disk.
The condensing process is an attempt to save memory by removing obsolete
mappings.
.
.It Sy zfs_condense_max_obsolete_bytes Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64
Only attempt to condense indirect vdev mappings if the on-disk size
of the obsolete space map object is greater than this number of bytes
.Pq see Sy zfs_condense_indirect_vdevs_enable .
.
.It Sy zfs_condense_min_mapping_bytes Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq u64
Minimum size vdev mapping to attempt to condense
.Pq see Sy zfs_condense_indirect_vdevs_enable .
.
.It Sy zfs_dbgmsg_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int
Internally ZFS keeps a small log to facilitate debugging.
The log is enabled by default, and can be disabled by unsetting this option.
The contents of the log can be accessed by reading
.Pa /proc/spl/kstat/zfs/dbgmsg .
Writing
.Sy 0
to the file clears the log.
.Pp
This setting does not influence debug prints due to
.Sy zfs_flags .
.
.It Sy zfs_dbgmsg_maxsize Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint
Maximum size of the internal ZFS debug log.
.
.It Sy zfs_dbuf_state_index Ns = Ns Sy 0 Pq int
Historically used for controlling what reporting was available under
.Pa /proc/spl/kstat/zfs .
No effect.
.
.It Sy zfs_deadman_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
When a pool sync operation takes longer than
.Sy zfs_deadman_synctime_ms ,
or when an individual I/O operation takes longer than
.Sy zfs_deadman_ziotime_ms ,
then the operation is considered to be "hung".
If
.Sy zfs_deadman_enabled
is set, then the deadman behavior is invoked as described by
.Sy zfs_deadman_failmode .
By default, the deadman is enabled and set to
.Sy wait
which results in "hung" I/O operations only being logged.
The deadman is automatically disabled when a pool gets suspended.
.
.It Sy zfs_deadman_failmode Ns = Ns Sy wait Pq charp
Controls the failure behavior when the deadman detects a "hung" I/O operation.
Valid values are:
.Bl -tag -compact -offset 4n -width "continue"
.It Sy wait
Wait for a "hung" operation to complete.
For each "hung" operation a "deadman" event will be posted
describing that operation.
.It Sy continue
Attempt to recover from a "hung" operation by re-dispatching it
to the I/O pipeline if possible.
.It Sy panic
Panic the system.
This can be used to facilitate automatic fail-over
to a properly configured fail-over partner.
.El
.
.It Sy zfs_deadman_checktime_ms Ns = Ns Sy 60000 Ns ms Po 1 min Pc Pq u64
Check time in milliseconds.
This defines the frequency at which we check for hung I/O requests
and potentially invoke the
.Sy zfs_deadman_failmode
behavior.
.
.It Sy zfs_deadman_synctime_ms Ns = Ns Sy 600000 Ns ms Po 10 min Pc Pq u64
Interval in milliseconds after which the deadman is triggered and also
the interval after which a pool sync operation is considered to be "hung".
Once this limit is exceeded the deadman will be invoked every
.Sy zfs_deadman_checktime_ms
milliseconds until the pool sync completes.
.
.It Sy zfs_deadman_ziotime_ms Ns = Ns Sy 300000 Ns ms Po 5 min Pc Pq u64
Interval in milliseconds after which the deadman is triggered and an
individual I/O operation is considered to be "hung".
As long as the operation remains "hung",
the deadman will be invoked every
.Sy zfs_deadman_checktime_ms
milliseconds until the operation completes.
.
.It Sy zfs_dedup_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int
Enable prefetching dedup-ed blocks which are going to be freed.
.
.It Sy zfs_delay_min_dirty_percent Ns = Ns Sy 60 Ns % Pq uint
Start to delay each transaction once there is this amount of dirty data,
expressed as a percentage of
.Sy zfs_dirty_data_max .
This value should be at least
.Sy zfs_vdev_async_write_active_max_dirty_percent .
.No See Sx ZFS TRANSACTION DELAY .
.
.It Sy zfs_delay_scale Ns = Ns Sy 500000 Pq int
This controls how quickly the transaction delay approaches infinity.
Larger values cause longer delays for a given amount of dirty data.
.Pp
For the smoothest delay, this value should be about 1 billion divided
by the maximum number of operations per second.
This will smoothly handle between ten times and a tenth of this number.
.No See Sx ZFS TRANSACTION DELAY .
.Pp
.Sy zfs_delay_scale No \(mu Sy zfs_dirty_data_max Em must No be smaller than Sy 2^64 .
.
.It Sy zfs_disable_ivset_guid_check Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disables requirement for IVset GUIDs to be present and match when doing a raw
receive of encrypted datasets.
Intended for users whose pools were created with
OpenZFS pre-release versions and now have compatibility issues.
.
.It Sy zfs_key_max_salt_uses Ns = Ns Sy 400000000 Po 4*10^8 Pc Pq ulong
Maximum number of uses of a single salt value before generating a new one for
encrypted datasets.
The default value is also the maximum.
.
.It Sy zfs_object_mutex_size Ns = Ns Sy 64 Pq uint
Size of the znode hashtable used for holds.
.Pp
Due to the need to hold locks on objects that may not exist yet, kernel mutexes
are not created per-object and instead a hashtable is used where collisions
will result in objects waiting when there is not actually contention on the
same object.
.
.It Sy zfs_slow_io_events_per_second Ns = Ns Sy 20 Ns /s Pq int
Rate limit delay and deadman zevents (which report slow I/O operations) to this
many per
second.
.
.It Sy zfs_unflushed_max_mem_amt Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64
Upper-bound limit for unflushed metadata changes to be held by the
log spacemap in memory, in bytes.
.
.It Sy zfs_unflushed_max_mem_ppm Ns = Ns Sy 1000 Ns ppm Po 0.1% Pc Pq u64
Part of overall system memory that ZFS allows to be used
for unflushed metadata changes by the log spacemap, in millionths.
.
.It Sy zfs_unflushed_log_block_max Ns = Ns Sy 131072 Po 128k Pc Pq u64
Describes the maximum number of log spacemap blocks allowed for each pool.
The default value means that the space in all the log spacemaps
can add up to no more than
.Sy 131072
blocks (which means
.Em 16 GiB
of logical space before compression and ditto blocks,
assuming that blocksize is
.Em 128 KiB ) .
.Pp
This tunable is important because it involves a trade-off between import
time after an unclean export and the frequency of flushing metaslabs.
The higher this number is, the more log blocks we allow when the pool is
active which means that we flush metaslabs less often and thus decrease
the number of I/O operations for spacemap updates per TXG.
At the same time though, that means that in the event of an unclean export,
there will be more log spacemap blocks for us to read, inducing overhead
in the import time of the pool.
The lower the number, the amount of flushing increases, destroying log
blocks quicker as they become obsolete faster, which leaves less blocks
to be read during import time after a crash.
.Pp
Each log spacemap block existing during pool import leads to approximately
one extra logical I/O issued.
This is the reason why this tunable is exposed in terms of blocks rather
than space used.
.
.It Sy zfs_unflushed_log_block_min Ns = Ns Sy 1000 Pq u64
If the number of metaslabs is small and our incoming rate is high,
we could get into a situation that we are flushing all our metaslabs every TXG.
Thus we always allow at least this many log blocks.
.
.It Sy zfs_unflushed_log_block_pct Ns = Ns Sy 400 Ns % Pq u64
Tunable used to determine the number of blocks that can be used for
the spacemap log, expressed as a percentage of the total number of
unflushed metaslabs in the pool.
.
.It Sy zfs_unflushed_log_txg_max Ns = Ns Sy 1000 Pq u64
Tunable limiting maximum time in TXGs any metaslab may remain unflushed.
It effectively limits maximum number of unflushed per-TXG spacemap logs
that need to be read after unclean pool export.
.
.It Sy zfs_unlink_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint
When enabled, files will not be asynchronously removed from the list of pending
unlinks and the space they consume will be leaked.
Once this option has been disabled and the dataset is remounted,
the pending unlinks will be processed and the freed space returned to the pool.
This option is used by the test suite.
.
.It Sy zfs_delete_blocks Ns = Ns Sy 20480 Pq ulong
This is the used to define a large file for the purposes of deletion.
Files containing more than
.Sy zfs_delete_blocks
will be deleted asynchronously, while smaller files are deleted synchronously.
Decreasing this value will reduce the time spent in an
.Xr unlink 2
system call, at the expense of a longer delay before the freed space is
available.
This only applies on Linux.
.
.It Sy zfs_dirty_data_max Ns = Pq int
Determines the dirty space limit in bytes.
Once this limit is exceeded, new writes are halted until space frees up.
This parameter takes precedence over
.Sy zfs_dirty_data_max_percent .
.No See Sx ZFS TRANSACTION DELAY .
.Pp
Defaults to
.Sy physical_ram/10 ,
capped at
.Sy zfs_dirty_data_max_max .
.
.It Sy zfs_dirty_data_max_max Ns = Pq int
Maximum allowable value of
.Sy zfs_dirty_data_max ,
expressed in bytes.
This limit is only enforced at module load time, and will be ignored if
.Sy zfs_dirty_data_max
is later changed.
This parameter takes precedence over
.Sy zfs_dirty_data_max_max_percent .
.No See Sx ZFS TRANSACTION DELAY .
.Pp
Defaults to
.Sy min(physical_ram/4, 4GiB) ,
or
.Sy min(physical_ram/4, 1GiB)
for 32-bit systems.
.
.It Sy zfs_dirty_data_max_max_percent Ns = Ns Sy 25 Ns % Pq uint
Maximum allowable value of
.Sy zfs_dirty_data_max ,
expressed as a percentage of physical RAM.
This limit is only enforced at module load time, and will be ignored if
.Sy zfs_dirty_data_max
is later changed.
The parameter
.Sy zfs_dirty_data_max_max
takes precedence over this one.
.No See Sx ZFS TRANSACTION DELAY .
.
.It Sy zfs_dirty_data_max_percent Ns = Ns Sy 10 Ns % Pq uint
Determines the dirty space limit, expressed as a percentage of all memory.
Once this limit is exceeded, new writes are halted until space frees up.
The parameter
.Sy zfs_dirty_data_max
takes precedence over this one.
.No See Sx ZFS TRANSACTION DELAY .
.Pp
Subject to
.Sy zfs_dirty_data_max_max .
.
.It Sy zfs_dirty_data_sync_percent Ns = Ns Sy 20 Ns % Pq uint
Start syncing out a transaction group if there's at least this much dirty data
.Pq as a percentage of Sy zfs_dirty_data_max .
This should be less than
.Sy zfs_vdev_async_write_active_min_dirty_percent .
.
.It Sy zfs_wrlog_data_max Ns = Pq int
The upper limit of write-transaction zil log data size in bytes.
Write operations are throttled when approaching the limit until log data is
cleared out after transaction group sync.
Because of some overhead, it should be set at least 2 times the size of
.Sy zfs_dirty_data_max
.No to prevent harming normal write throughput .
It also should be smaller than the size of the slog device if slog is present.
.Pp
Defaults to
.Sy zfs_dirty_data_max*2
.
.It Sy zfs_fallocate_reserve_percent Ns = Ns Sy 110 Ns % Pq uint
Since ZFS is a copy-on-write filesystem with snapshots, blocks cannot be
preallocated for a file in order to guarantee that later writes will not
run out of space.
Instead,
.Xr fallocate 2
space preallocation only checks that sufficient space is currently available
in the pool or the user's project quota allocation,
and then creates a sparse file of the requested size.
The requested space is multiplied by
.Sy zfs_fallocate_reserve_percent
to allow additional space for indirect blocks and other internal metadata.
Setting this to
.Sy 0
disables support for
.Xr fallocate 2
and causes it to return
.Sy EOPNOTSUPP .
.
.It Sy zfs_fletcher_4_impl Ns = Ns Sy fastest Pq string
Select a fletcher 4 implementation.
.Pp
Supported selectors are:
.Sy fastest , scalar , sse2 , ssse3 , avx2 , avx512f , avx512bw ,
.No and Sy aarch64_neon .
All except
.Sy fastest No and Sy scalar
require instruction set extensions to be available,
and will only appear if ZFS detects that they are present at runtime.
If multiple implementations of fletcher 4 are available, the
.Sy fastest
will be chosen using a micro benchmark.
Selecting
.Sy scalar
results in the original CPU-based calculation being used.
Selecting any option other than
.Sy fastest No or Sy scalar
results in vector instructions
from the respective CPU instruction set being used.
.
.It Sy zfs_blake3_impl Ns = Ns Sy fastest Pq string
Select a BLAKE3 implementation.
.Pp
Supported selectors are:
.Sy cycle , fastest , generic , sse2 , sse41 , avx2 , avx512 .
All except
.Sy cycle , fastest No and Sy generic
require instruction set extensions to be available,
and will only appear if ZFS detects that they are present at runtime.
If multiple implementations of BLAKE3 are available, the
.Sy fastest will be chosen using a micro benchmark. You can see the
benchmark results by reading this kstat file:
.Pa /proc/spl/kstat/zfs/chksum_bench .
.
.It Sy zfs_free_bpobj_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable/disable the processing of the free_bpobj object.
.
.It Sy zfs_async_block_max_blocks Ns = Ns Sy UINT64_MAX Po unlimited Pc Pq u64
Maximum number of blocks freed in a single TXG.
.
.It Sy zfs_max_async_dedup_frees Ns = Ns Sy 100000 Po 10^5 Pc Pq u64
Maximum number of dedup blocks freed in a single TXG.
.
.It Sy zfs_vdev_async_read_max_active Ns = Ns Sy 3 Pq uint
Maximum asynchronous read I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_read_min_active Ns = Ns Sy 1 Pq uint
Minimum asynchronous read I/O operation active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_write_active_max_dirty_percent Ns = Ns Sy 60 Ns % Pq uint
When the pool has more than this much dirty data, use
.Sy zfs_vdev_async_write_max_active
to limit active async writes.
If the dirty data is between the minimum and maximum,
the active I/O limit is linearly interpolated.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_write_active_min_dirty_percent Ns = Ns Sy 30 Ns % Pq uint
When the pool has less than this much dirty data, use
.Sy zfs_vdev_async_write_min_active
to limit active async writes.
If the dirty data is between the minimum and maximum,
the active I/O limit is linearly
interpolated.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_write_max_active Ns = Ns Sy 10 Pq uint
Maximum asynchronous write I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_async_write_min_active Ns = Ns Sy 2 Pq uint
Minimum asynchronous write I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.Pp
Lower values are associated with better latency on rotational media but poorer
resilver performance.
The default value of
.Sy 2
was chosen as a compromise.
A value of
.Sy 3
has been shown to improve resilver performance further at a cost of
further increasing latency.
.
.It Sy zfs_vdev_initializing_max_active Ns = Ns Sy 1 Pq uint
Maximum initializing I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_initializing_min_active Ns = Ns Sy 1 Pq uint
Minimum initializing I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_max_active Ns = Ns Sy 1000 Pq uint
The maximum number of I/O operations active to each device.
Ideally, this will be at least the sum of each queue's
.Sy max_active .
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_open_timeout_ms Ns = Ns Sy 1000 Pq uint
Timeout value to wait before determining a device is missing
during import.
This is helpful for transient missing paths due
to links being briefly removed and recreated in response to
udev events.
.
.It Sy zfs_vdev_rebuild_max_active Ns = Ns Sy 3 Pq uint
Maximum sequential resilver I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_rebuild_min_active Ns = Ns Sy 1 Pq uint
Minimum sequential resilver I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_removal_max_active Ns = Ns Sy 2 Pq uint
Maximum removal I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_removal_min_active Ns = Ns Sy 1 Pq uint
Minimum removal I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_scrub_max_active Ns = Ns Sy 2 Pq uint
Maximum scrub I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_scrub_min_active Ns = Ns Sy 1 Pq uint
Minimum scrub I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_sync_read_max_active Ns = Ns Sy 10 Pq uint
Maximum synchronous read I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_sync_read_min_active Ns = Ns Sy 10 Pq uint
Minimum synchronous read I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_sync_write_max_active Ns = Ns Sy 10 Pq uint
Maximum synchronous write I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_sync_write_min_active Ns = Ns Sy 10 Pq uint
Minimum synchronous write I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_trim_max_active Ns = Ns Sy 2 Pq uint
Maximum trim/discard I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_trim_min_active Ns = Ns Sy 1 Pq uint
Minimum trim/discard I/O operations active to each device.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_nia_delay Ns = Ns Sy 5 Pq uint
For non-interactive I/O (scrub, resilver, removal, initialize and rebuild),
the number of concurrently-active I/O operations is limited to
.Sy zfs_*_min_active ,
unless the vdev is "idle".
When there are no interactive I/O operations active (synchronous or otherwise),
and
.Sy zfs_vdev_nia_delay
operations have completed since the last interactive operation,
then the vdev is considered to be "idle",
and the number of concurrently-active non-interactive operations is increased to
.Sy zfs_*_max_active .
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_nia_credit Ns = Ns Sy 5 Pq uint
Some HDDs tend to prioritize sequential I/O so strongly, that concurrent
random I/O latency reaches several seconds.
On some HDDs this happens even if sequential I/O operations
are submitted one at a time, and so setting
.Sy zfs_*_max_active Ns = Sy 1
does not help.
To prevent non-interactive I/O, like scrub,
from monopolizing the device, no more than
.Sy zfs_vdev_nia_credit operations can be sent
while there are outstanding incomplete interactive operations.
This enforced wait ensures the HDD services the interactive I/O
within a reasonable amount of time.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_queue_depth_pct Ns = Ns Sy 1000 Ns % Pq uint
Maximum number of queued allocations per top-level vdev expressed as
a percentage of
.Sy zfs_vdev_async_write_max_active ,
which allows the system to detect devices that are more capable
of handling allocations and to allocate more blocks to those devices.
This allows for dynamic allocation distribution when devices are imbalanced,
as fuller devices will tend to be slower than empty devices.
.Pp
Also see
.Sy zio_dva_throttle_enabled .
.
+.It Sy zfs_vdev_def_queue_depth Ns = Ns Sy 32 Pq uint
+Default queue depth for each vdev IO allocator.
+Higher values allow for better coalescing of sequential writes before sending
+them to the disk, but can increase transaction commit times.
+.
.It Sy zfs_vdev_failfast_mask Ns = Ns Sy 1 Pq uint
Defines if the driver should retire on a given error type.
The following options may be bitwise-ored together:
.TS
box;
lbz r l l .
Value Name Description
_
1 Device No driver retries on device errors
2 Transport No driver retries on transport errors.
4 Driver No driver retries on driver errors.
.TE
.
.It Sy zfs_expire_snapshot Ns = Ns Sy 300 Ns s Pq int
Time before expiring
.Pa .zfs/snapshot .
.
.It Sy zfs_admin_snapshot Ns = Ns Sy 0 Ns | Ns 1 Pq int
Allow the creation, removal, or renaming of entries in the
.Sy .zfs/snapshot
directory to cause the creation, destruction, or renaming of snapshots.
When enabled, this functionality works both locally and over NFS exports
which have the
.Em no_root_squash
option set.
.
.It Sy zfs_flags Ns = Ns Sy 0 Pq int
Set additional debugging flags.
The following flags may be bitwise-ored together:
.TS
box;
lbz r l l .
Value Name Description
_
1 ZFS_DEBUG_DPRINTF Enable dprintf entries in the debug log.
* 2 ZFS_DEBUG_DBUF_VERIFY Enable extra dbuf verifications.
* 4 ZFS_DEBUG_DNODE_VERIFY Enable extra dnode verifications.
8 ZFS_DEBUG_SNAPNAMES Enable snapshot name verification.
* 16 ZFS_DEBUG_MODIFY Check for illegally modified ARC buffers.
64 ZFS_DEBUG_ZIO_FREE Enable verification of block frees.
128 ZFS_DEBUG_HISTOGRAM_VERIFY Enable extra spacemap histogram verifications.
256 ZFS_DEBUG_METASLAB_VERIFY Verify space accounting on disk matches in-memory \fBrange_trees\fP.
512 ZFS_DEBUG_SET_ERROR Enable \fBSET_ERROR\fP and dprintf entries in the debug log.
1024 ZFS_DEBUG_INDIRECT_REMAP Verify split blocks created by device removal.
2048 ZFS_DEBUG_TRIM Verify TRIM ranges are always within the allocatable range tree.
4096 ZFS_DEBUG_LOG_SPACEMAP Verify that the log summary is consistent with the spacemap log
and enable \fBzfs_dbgmsgs\fP for metaslab loading and flushing.
.TE
.Sy \& * No Requires debug build .
.
.It Sy zfs_btree_verify_intensity Ns = Ns Sy 0 Pq uint
Enables btree verification.
The following settings are culminative:
.TS
box;
lbz r l l .
Value Description
1 Verify height.
2 Verify pointers from children to parent.
3 Verify element counts.
4 Verify element order. (expensive)
* 5 Verify unused memory is poisoned. (expensive)
.TE
.Sy \& * No Requires debug build .
.
.It Sy zfs_free_leak_on_eio Ns = Ns Sy 0 Ns | Ns 1 Pq int
If destroy encounters an
.Sy EIO
while reading metadata (e.g. indirect blocks),
space referenced by the missing metadata can not be freed.
Normally this causes the background destroy to become "stalled",
as it is unable to make forward progress.
While in this stalled state, all remaining space to free
from the error-encountering filesystem is "temporarily leaked".
Set this flag to cause it to ignore the
.Sy EIO ,
permanently leak the space from indirect blocks that can not be read,
and continue to free everything else that it can.
.Pp
The default "stalling" behavior is useful if the storage partially
fails (i.e. some but not all I/O operations fail), and then later recovers.
In this case, we will be able to continue pool operations while it is
partially failed, and when it recovers, we can continue to free the
space, with no leaks.
Note, however, that this case is actually fairly rare.
.Pp
Typically pools either
.Bl -enum -compact -offset 4n -width "1."
.It
fail completely (but perhaps temporarily,
e.g. due to a top-level vdev going offline), or
.It
have localized, permanent errors (e.g. disk returns the wrong data
due to bit flip or firmware bug).
.El
In the former case, this setting does not matter because the
pool will be suspended and the sync thread will not be able to make
forward progress regardless.
In the latter, because the error is permanent, the best we can do
is leak the minimum amount of space,
which is what setting this flag will do.
It is therefore reasonable for this flag to normally be set,
but we chose the more conservative approach of not setting it,
so that there is no possibility of
leaking space in the "partial temporary" failure case.
.
.It Sy zfs_free_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1s Pc Pq uint
During a
.Nm zfs Cm destroy
operation using the
.Sy async_destroy
feature,
a minimum of this much time will be spent working on freeing blocks per TXG.
.
.It Sy zfs_obsolete_min_time_ms Ns = Ns Sy 500 Ns ms Pq uint
Similar to
.Sy zfs_free_min_time_ms ,
but for cleanup of old indirection records for removed vdevs.
.
.It Sy zfs_immediate_write_sz Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq s64
Largest data block to write to the ZIL.
Larger blocks will be treated as if the dataset being written to had the
.Sy logbias Ns = Ns Sy throughput
property set.
.
.It Sy zfs_initialize_value Ns = Ns Sy 16045690984833335022 Po 0xDEADBEEFDEADBEEE Pc Pq u64
Pattern written to vdev free space by
.Xr zpool-initialize 8 .
.
.It Sy zfs_initialize_chunk_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
Size of writes used by
.Xr zpool-initialize 8 .
This option is used by the test suite.
.
.It Sy zfs_livelist_max_entries Ns = Ns Sy 500000 Po 5*10^5 Pc Pq u64
The threshold size (in block pointers) at which we create a new sub-livelist.
Larger sublists are more costly from a memory perspective but the fewer
sublists there are, the lower the cost of insertion.
.
.It Sy zfs_livelist_min_percent_shared Ns = Ns Sy 75 Ns % Pq int
If the amount of shared space between a snapshot and its clone drops below
this threshold, the clone turns off the livelist and reverts to the old
deletion method.
This is in place because livelists no long give us a benefit
once a clone has been overwritten enough.
.
.It Sy zfs_livelist_condense_new_alloc Ns = Ns Sy 0 Pq int
Incremented each time an extra ALLOC blkptr is added to a livelist entry while
it is being condensed.
This option is used by the test suite to track race conditions.
.
.It Sy zfs_livelist_condense_sync_cancel Ns = Ns Sy 0 Pq int
Incremented each time livelist condensing is canceled while in
.Fn spa_livelist_condense_sync .
This option is used by the test suite to track race conditions.
.
.It Sy zfs_livelist_condense_sync_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int
When set, the livelist condense process pauses indefinitely before
executing the synctask \(em
.Fn spa_livelist_condense_sync .
This option is used by the test suite to trigger race conditions.
.
.It Sy zfs_livelist_condense_zthr_cancel Ns = Ns Sy 0 Pq int
Incremented each time livelist condensing is canceled while in
.Fn spa_livelist_condense_cb .
This option is used by the test suite to track race conditions.
.
.It Sy zfs_livelist_condense_zthr_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int
When set, the livelist condense process pauses indefinitely before
executing the open context condensing work in
.Fn spa_livelist_condense_cb .
This option is used by the test suite to trigger race conditions.
.
.It Sy zfs_lua_max_instrlimit Ns = Ns Sy 100000000 Po 10^8 Pc Pq u64
The maximum execution time limit that can be set for a ZFS channel program,
specified as a number of Lua instructions.
.
.It Sy zfs_lua_max_memlimit Ns = Ns Sy 104857600 Po 100 MiB Pc Pq u64
The maximum memory limit that can be set for a ZFS channel program, specified
in bytes.
.
.It Sy zfs_max_dataset_nesting Ns = Ns Sy 50 Pq int
The maximum depth of nested datasets.
This value can be tuned temporarily to
fix existing datasets that exceed the predefined limit.
.
.It Sy zfs_max_log_walking Ns = Ns Sy 5 Pq u64
The number of past TXGs that the flushing algorithm of the log spacemap
feature uses to estimate incoming log blocks.
.
.It Sy zfs_max_logsm_summary_length Ns = Ns Sy 10 Pq u64
Maximum number of rows allowed in the summary of the spacemap log.
.
.It Sy zfs_max_recordsize Ns = Ns Sy 16777216 Po 16 MiB Pc Pq uint
We currently support block sizes from
.Em 512 Po 512 B Pc No to Em 16777216 Po 16 MiB Pc .
The benefits of larger blocks, and thus larger I/O,
need to be weighed against the cost of COWing a giant block to modify one byte.
Additionally, very large blocks can have an impact on I/O latency,
and also potentially on the memory allocator.
Therefore, we formerly forbade creating blocks larger than 1M.
Larger blocks could be created by changing it,
and pools with larger blocks can always be imported and used,
regardless of this setting.
.
.It Sy zfs_allow_redacted_dataset_mount Ns = Ns Sy 0 Ns | Ns 1 Pq int
Allow datasets received with redacted send/receive to be mounted.
Normally disabled because these datasets may be missing key data.
.
.It Sy zfs_min_metaslabs_to_flush Ns = Ns Sy 1 Pq u64
Minimum number of metaslabs to flush per dirty TXG.
.
.It Sy zfs_metaslab_fragmentation_threshold Ns = Ns Sy 70 Ns % Pq uint
Allow metaslabs to keep their active state as long as their fragmentation
percentage is no more than this value.
An active metaslab that exceeds this threshold
will no longer keep its active status allowing better metaslabs to be selected.
.
.It Sy zfs_mg_fragmentation_threshold Ns = Ns Sy 95 Ns % Pq uint
Metaslab groups are considered eligible for allocations if their
fragmentation metric (measured as a percentage) is less than or equal to
this value.
If a metaslab group exceeds this threshold then it will be
skipped unless all metaslab groups within the metaslab class have also
crossed this threshold.
.
.It Sy zfs_mg_noalloc_threshold Ns = Ns Sy 0 Ns % Pq uint
Defines a threshold at which metaslab groups should be eligible for allocations.
The value is expressed as a percentage of free space
beyond which a metaslab group is always eligible for allocations.
If a metaslab group's free space is less than or equal to the
threshold, the allocator will avoid allocating to that group
unless all groups in the pool have reached the threshold.
Once all groups have reached the threshold, all groups are allowed to accept
allocations.
The default value of
.Sy 0
disables the feature and causes all metaslab groups to be eligible for
allocations.
.Pp
This parameter allows one to deal with pools having heavily imbalanced
vdevs such as would be the case when a new vdev has been added.
Setting the threshold to a non-zero percentage will stop allocations
from being made to vdevs that aren't filled to the specified percentage
and allow lesser filled vdevs to acquire more allocations than they
otherwise would under the old
.Sy zfs_mg_alloc_failures
facility.
.
.It Sy zfs_ddt_data_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int
If enabled, ZFS will place DDT data into the special allocation class.
.
.It Sy zfs_user_indirect_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int
If enabled, ZFS will place user data indirect blocks
into the special allocation class.
.
.It Sy zfs_multihost_history Ns = Ns Sy 0 Pq uint
Historical statistics for this many latest multihost updates will be available
in
.Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /multihost .
.
.It Sy zfs_multihost_interval Ns = Ns Sy 1000 Ns ms Po 1 s Pc Pq u64
Used to control the frequency of multihost writes which are performed when the
.Sy multihost
pool property is on.
This is one of the factors used to determine the
length of the activity check during import.
.Pp
The multihost write period is
.Sy zfs_multihost_interval No / Sy leaf-vdevs .
On average a multihost write will be issued for each leaf vdev
every
.Sy zfs_multihost_interval
milliseconds.
In practice, the observed period can vary with the I/O load
and this observed value is the delay which is stored in the uberblock.
.
.It Sy zfs_multihost_import_intervals Ns = Ns Sy 20 Pq uint
Used to control the duration of the activity test on import.
Smaller values of
.Sy zfs_multihost_import_intervals
will reduce the import time but increase
the risk of failing to detect an active pool.
The total activity check time is never allowed to drop below one second.
.Pp
On import the activity check waits a minimum amount of time determined by
.Sy zfs_multihost_interval No \(mu Sy zfs_multihost_import_intervals ,
or the same product computed on the host which last had the pool imported,
whichever is greater.
The activity check time may be further extended if the value of MMP
delay found in the best uberblock indicates actual multihost updates happened
at longer intervals than
.Sy zfs_multihost_interval .
A minimum of
.Em 100 ms
is enforced.
.Pp
.Sy 0 No is equivalent to Sy 1 .
.
.It Sy zfs_multihost_fail_intervals Ns = Ns Sy 10 Pq uint
Controls the behavior of the pool when multihost write failures or delays are
detected.
.Pp
When
.Sy 0 ,
multihost write failures or delays are ignored.
The failures will still be reported to the ZED which depending on
its configuration may take action such as suspending the pool or offlining a
device.
.Pp
Otherwise, the pool will be suspended if
.Sy zfs_multihost_fail_intervals No \(mu Sy zfs_multihost_interval
milliseconds pass without a successful MMP write.
This guarantees the activity test will see MMP writes if the pool is imported.
.Sy 1 No is equivalent to Sy 2 ;
this is necessary to prevent the pool from being suspended
due to normal, small I/O latency variations.
.
.It Sy zfs_no_scrub_io Ns = Ns Sy 0 Ns | Ns 1 Pq int
Set to disable scrub I/O.
This results in scrubs not actually scrubbing data and
simply doing a metadata crawl of the pool instead.
.
.It Sy zfs_no_scrub_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int
Set to disable block prefetching for scrubs.
.
.It Sy zfs_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable cache flush operations on disks when writing.
Setting this will cause pool corruption on power loss
if a volatile out-of-order write cache is enabled.
.
.It Sy zfs_nopwrite_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Allow no-operation writes.
The occurrence of nopwrites will further depend on other pool properties
.Pq i.a. the checksumming and compression algorithms .
.
.It Sy zfs_dmu_offset_next_sync Ns = Ns Sy 1 Ns | Ns 0 Pq int
Enable forcing TXG sync to find holes.
When enabled forces ZFS to sync data when
.Sy SEEK_HOLE No or Sy SEEK_DATA
flags are used allowing holes in a file to be accurately reported.
When disabled holes will not be reported in recently dirtied files.
.
.It Sy zfs_pd_bytes_max Ns = Ns Sy 52428800 Ns B Po 50 MiB Pc Pq int
The number of bytes which should be prefetched during a pool traversal, like
.Nm zfs Cm send
or other data crawling operations.
.
.It Sy zfs_traverse_indirect_prefetch_limit Ns = Ns Sy 32 Pq uint
The number of blocks pointed by indirect (non-L0) block which should be
prefetched during a pool traversal, like
.Nm zfs Cm send
or other data crawling operations.
.
.It Sy zfs_per_txg_dirty_frees_percent Ns = Ns Sy 30 Ns % Pq u64
Control percentage of dirtied indirect blocks from frees allowed into one TXG.
After this threshold is crossed, additional frees will wait until the next TXG.
.Sy 0 No disables this throttle .
.
.It Sy zfs_prefetch_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable predictive prefetch.
Note that it leaves "prescient" prefetch
.Pq for, e.g., Nm zfs Cm send
intact.
Unlike predictive prefetch, prescient prefetch never issues I/O
that ends up not being needed, so it can't hurt performance.
.
.It Sy zfs_qat_checksum_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable QAT hardware acceleration for SHA256 checksums.
May be unset after the ZFS modules have been loaded to initialize the QAT
hardware as long as support is compiled in and the QAT driver is present.
.
.It Sy zfs_qat_compress_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable QAT hardware acceleration for gzip compression.
May be unset after the ZFS modules have been loaded to initialize the QAT
hardware as long as support is compiled in and the QAT driver is present.
.
.It Sy zfs_qat_encrypt_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable QAT hardware acceleration for AES-GCM encryption.
May be unset after the ZFS modules have been loaded to initialize the QAT
hardware as long as support is compiled in and the QAT driver is present.
.
.It Sy zfs_vnops_read_chunk_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
Bytes to read per chunk.
.
.It Sy zfs_read_history Ns = Ns Sy 0 Pq uint
Historical statistics for this many latest reads will be available in
.Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /reads .
.
.It Sy zfs_read_history_hits Ns = Ns Sy 0 Ns | Ns 1 Pq int
Include cache hits in read history
.
.It Sy zfs_rebuild_max_segment Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64
Maximum read segment size to issue when sequentially resilvering a
top-level vdev.
.
.It Sy zfs_rebuild_scrub_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Automatically start a pool scrub when the last active sequential resilver
completes in order to verify the checksums of all blocks which have been
resilvered.
This is enabled by default and strongly recommended.
.
.It Sy zfs_rebuild_vdev_limit Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq u64
Maximum amount of I/O that can be concurrently issued for a sequential
resilver per leaf device, given in bytes.
.
.It Sy zfs_reconstruct_indirect_combinations_max Ns = Ns Sy 4096 Pq int
If an indirect split block contains more than this many possible unique
combinations when being reconstructed, consider it too computationally
expensive to check them all.
Instead, try at most this many randomly selected
combinations each time the block is accessed.
This allows all segment copies to participate fairly
in the reconstruction when all combinations
cannot be checked and prevents repeated use of one bad copy.
.
.It Sy zfs_recover Ns = Ns Sy 0 Ns | Ns 1 Pq int
Set to attempt to recover from fatal errors.
This should only be used as a last resort,
as it typically results in leaked space, or worse.
.
.It Sy zfs_removal_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int
Ignore hard I/O errors during device removal.
When set, if a device encounters a hard I/O error during the removal process
the removal will not be cancelled.
This can result in a normally recoverable block becoming permanently damaged
and is hence not recommended.
This should only be used as a last resort when the
pool cannot be returned to a healthy state prior to removing the device.
.
.It Sy zfs_removal_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint
This is used by the test suite so that it can ensure that certain actions
happen while in the middle of a removal.
.
.It Sy zfs_remove_max_segment Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint
The largest contiguous segment that we will attempt to allocate when removing
a device.
If there is a performance problem with attempting to allocate large blocks,
consider decreasing this.
The default value is also the maximum.
.
.It Sy zfs_resilver_disable_defer Ns = Ns Sy 0 Ns | Ns 1 Pq int
Ignore the
.Sy resilver_defer
feature, causing an operation that would start a resilver to
immediately restart the one in progress.
.
.It Sy zfs_resilver_min_time_ms Ns = Ns Sy 3000 Ns ms Po 3 s Pc Pq uint
Resilvers are processed by the sync thread.
While resilvering, it will spend at least this much time
working on a resilver between TXG flushes.
.
.It Sy zfs_scan_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int
If set, remove the DTL (dirty time list) upon completion of a pool scan (scrub),
even if there were unrepairable errors.
Intended to be used during pool repair or recovery to
stop resilvering when the pool is next imported.
.
.It Sy zfs_scrub_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1 s Pc Pq uint
Scrubs are processed by the sync thread.
While scrubbing, it will spend at least this much time
working on a scrub between TXG flushes.
.
.It Sy zfs_scan_checkpoint_intval Ns = Ns Sy 7200 Ns s Po 2 hour Pc Pq uint
To preserve progress across reboots, the sequential scan algorithm periodically
needs to stop metadata scanning and issue all the verification I/O to disk.
The frequency of this flushing is determined by this tunable.
.
.It Sy zfs_scan_fill_weight Ns = Ns Sy 3 Pq uint
This tunable affects how scrub and resilver I/O segments are ordered.
A higher number indicates that we care more about how filled in a segment is,
while a lower number indicates we care more about the size of the extent without
considering the gaps within a segment.
This value is only tunable upon module insertion.
Changing the value afterwards will have no effect on scrub or resilver
performance.
.
.It Sy zfs_scan_issue_strategy Ns = Ns Sy 0 Pq uint
Determines the order that data will be verified while scrubbing or resilvering:
.Bl -tag -compact -offset 4n -width "a"
.It Sy 1
Data will be verified as sequentially as possible, given the
amount of memory reserved for scrubbing
.Pq see Sy zfs_scan_mem_lim_fact .
This may improve scrub performance if the pool's data is very fragmented.
.It Sy 2
The largest mostly-contiguous chunk of found data will be verified first.
By deferring scrubbing of small segments, we may later find adjacent data
to coalesce and increase the segment size.
.It Sy 0
.No Use strategy Sy 1 No during normal verification
.No and strategy Sy 2 No while taking a checkpoint .
.El
.
.It Sy zfs_scan_legacy Ns = Ns Sy 0 Ns | Ns 1 Pq int
If unset, indicates that scrubs and resilvers will gather metadata in
memory before issuing sequential I/O.
Otherwise indicates that the legacy algorithm will be used,
where I/O is initiated as soon as it is discovered.
Unsetting will not affect scrubs or resilvers that are already in progress.
.
.It Sy zfs_scan_max_ext_gap Ns = Ns Sy 2097152 Ns B Po 2 MiB Pc Pq int
Sets the largest gap in bytes between scrub/resilver I/O operations
that will still be considered sequential for sorting purposes.
Changing this value will not
affect scrubs or resilvers that are already in progress.
.
.It Sy zfs_scan_mem_lim_fact Ns = Ns Sy 20 Ns ^-1 Pq uint
Maximum fraction of RAM used for I/O sorting by sequential scan algorithm.
This tunable determines the hard limit for I/O sorting memory usage.
When the hard limit is reached we stop scanning metadata and start issuing
data verification I/O.
This is done until we get below the soft limit.
.
.It Sy zfs_scan_mem_lim_soft_fact Ns = Ns Sy 20 Ns ^-1 Pq uint
The fraction of the hard limit used to determined the soft limit for I/O sorting
by the sequential scan algorithm.
When we cross this limit from below no action is taken.
When we cross this limit from above it is because we are issuing verification
I/O.
In this case (unless the metadata scan is done) we stop issuing verification I/O
and start scanning metadata again until we get to the hard limit.
.
.It Sy zfs_scan_report_txgs Ns = Ns Sy 0 Ns | Ns 1 Pq uint
When reporting resilver throughput and estimated completion time use the
performance observed over roughly the last
.Sy zfs_scan_report_txgs
TXGs.
When set to zero performance is calculated over the time between checkpoints.
.
.It Sy zfs_scan_strict_mem_lim Ns = Ns Sy 0 Ns | Ns 1 Pq int
Enforce tight memory limits on pool scans when a sequential scan is in progress.
When disabled, the memory limit may be exceeded by fast disks.
.
.It Sy zfs_scan_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq int
Freezes a scrub/resilver in progress without actually pausing it.
Intended for testing/debugging.
.
.It Sy zfs_scan_vdev_limit Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq int
Maximum amount of data that can be concurrently issued at once for scrubs and
resilvers per leaf device, given in bytes.
.
.It Sy zfs_send_corrupt_data Ns = Ns Sy 0 Ns | Ns 1 Pq int
Allow sending of corrupt data (ignore read/checksum errors when sending).
.
.It Sy zfs_send_unmodified_spill_blocks Ns = Ns Sy 1 Ns | Ns 0 Pq int
Include unmodified spill blocks in the send stream.
Under certain circumstances, previous versions of ZFS could incorrectly
remove the spill block from an existing object.
Including unmodified copies of the spill blocks creates a backwards-compatible
stream which will recreate a spill block if it was incorrectly removed.
.
.It Sy zfs_send_no_prefetch_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint
The fill fraction of the
.Nm zfs Cm send
internal queues.
The fill fraction controls the timing with which internal threads are woken up.
.
.It Sy zfs_send_no_prefetch_queue_length Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint
The maximum number of bytes allowed in
.Nm zfs Cm send Ns 's
internal queues.
.
.It Sy zfs_send_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint
The fill fraction of the
.Nm zfs Cm send
prefetch queue.
The fill fraction controls the timing with which internal threads are woken up.
.
.It Sy zfs_send_queue_length Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint
The maximum number of bytes allowed that will be prefetched by
.Nm zfs Cm send .
This value must be at least twice the maximum block size in use.
.
.It Sy zfs_recv_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint
The fill fraction of the
.Nm zfs Cm receive
queue.
The fill fraction controls the timing with which internal threads are woken up.
.
.It Sy zfs_recv_queue_length Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint
The maximum number of bytes allowed in the
.Nm zfs Cm receive
queue.
This value must be at least twice the maximum block size in use.
.
.It Sy zfs_recv_write_batch_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint
The maximum amount of data, in bytes, that
.Nm zfs Cm receive
will write in one DMU transaction.
This is the uncompressed size, even when receiving a compressed send stream.
This setting will not reduce the write size below a single block.
Capped at a maximum of
.Sy 32 MiB .
.
.It Sy zfs_recv_best_effort_corrective Ns = Ns Sy 0 Pq int
When this variable is set to non-zero a corrective receive:
.Bl -enum -compact -offset 4n -width "1."
.It
Does not enforce the restriction of source & destination snapshot GUIDs
matching.
.It
If there is an error during healing, the healing receive is not
terminated instead it moves on to the next record.
.El
.
.It Sy zfs_override_estimate_recordsize Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Setting this variable overrides the default logic for estimating block
sizes when doing a
.Nm zfs Cm send .
The default heuristic is that the average block size
will be the current recordsize.
Override this value if most data in your dataset is not of that size
and you require accurate zfs send size estimates.
.
.It Sy zfs_sync_pass_deferred_free Ns = Ns Sy 2 Pq uint
Flushing of data to disk is done in passes.
Defer frees starting in this pass.
.
.It Sy zfs_spa_discard_memory_limit Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq int
Maximum memory used for prefetching a checkpoint's space map on each
vdev while discarding the checkpoint.
.
.It Sy zfs_special_class_metadata_reserve_pct Ns = Ns Sy 25 Ns % Pq uint
Only allow small data blocks to be allocated on the special and dedup vdev
types when the available free space percentage on these vdevs exceeds this
value.
This ensures reserved space is available for pool metadata as the
special vdevs approach capacity.
.
.It Sy zfs_sync_pass_dont_compress Ns = Ns Sy 8 Pq uint
Starting in this sync pass, disable compression (including of metadata).
With the default setting, in practice, we don't have this many sync passes,
so this has no effect.
.Pp
The original intent was that disabling compression would help the sync passes
to converge.
However, in practice, disabling compression increases
the average number of sync passes; because when we turn compression off,
many blocks' size will change, and thus we have to re-allocate
(not overwrite) them.
It also increases the number of
.Em 128 KiB
allocations (e.g. for indirect blocks and spacemaps)
because these will not be compressed.
The
.Em 128 KiB
allocations are especially detrimental to performance
on highly fragmented systems, which may have very few free segments of this
size,
and may need to load new metaslabs to satisfy these allocations.
.
.It Sy zfs_sync_pass_rewrite Ns = Ns Sy 2 Pq uint
Rewrite new block pointers starting in this pass.
.
.It Sy zfs_sync_taskq_batch_pct Ns = Ns Sy 75 Ns % Pq int
This controls the number of threads used by
.Sy dp_sync_taskq .
The default value of
.Sy 75%
will create a maximum of one thread per CPU.
.
.It Sy zfs_trim_extent_bytes_max Ns = Ns Sy 134217728 Ns B Po 128 MiB Pc Pq uint
Maximum size of TRIM command.
Larger ranges will be split into chunks no larger than this value before
issuing.
.
.It Sy zfs_trim_extent_bytes_min Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint
Minimum size of TRIM commands.
TRIM ranges smaller than this will be skipped,
unless they're part of a larger range which was chunked.
This is done because it's common for these small TRIMs
to negatively impact overall performance.
.
.It Sy zfs_trim_metaslab_skip Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Skip uninitialized metaslabs during the TRIM process.
This option is useful for pools constructed from large thinly-provisioned
devices
where TRIM operations are slow.
As a pool ages, an increasing fraction of the pool's metaslabs
will be initialized, progressively degrading the usefulness of this option.
This setting is stored when starting a manual TRIM and will
persist for the duration of the requested TRIM.
.
.It Sy zfs_trim_queue_limit Ns = Ns Sy 10 Pq uint
Maximum number of queued TRIMs outstanding per leaf vdev.
The number of concurrent TRIM commands issued to the device is controlled by
.Sy zfs_vdev_trim_min_active No and Sy zfs_vdev_trim_max_active .
.
.It Sy zfs_trim_txg_batch Ns = Ns Sy 32 Pq uint
The number of transaction groups' worth of frees which should be aggregated
before TRIM operations are issued to the device.
This setting represents a trade-off between issuing larger,
more efficient TRIM operations and the delay
before the recently trimmed space is available for use by the device.
.Pp
Increasing this value will allow frees to be aggregated for a longer time.
This will result is larger TRIM operations and potentially increased memory
usage.
Decreasing this value will have the opposite effect.
The default of
.Sy 32
was determined to be a reasonable compromise.
.
.It Sy zfs_txg_history Ns = Ns Sy 0 Pq uint
Historical statistics for this many latest TXGs will be available in
.Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /TXGs .
.
.It Sy zfs_txg_timeout Ns = Ns Sy 5 Ns s Pq uint
Flush dirty data to disk at least every this many seconds (maximum TXG
duration).
.
.It Sy zfs_vdev_aggregate_trim Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Allow TRIM I/O operations to be aggregated.
This is normally not helpful because the extents to be trimmed
will have been already been aggregated by the metaslab.
This option is provided for debugging and performance analysis.
.
.It Sy zfs_vdev_aggregation_limit Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint
Max vdev I/O aggregation size.
.
.It Sy zfs_vdev_aggregation_limit_non_rotating Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint
Max vdev I/O aggregation size for non-rotating media.
.
.It Sy zfs_vdev_cache_bshift Ns = Ns Sy 16 Po 64 KiB Pc Pq uint
Shift size to inflate reads to.
.
.It Sy zfs_vdev_cache_max Ns = Ns Sy 16384 Ns B Po 16 KiB Pc Pq uint
Inflate reads smaller than this value to meet the
.Sy zfs_vdev_cache_bshift
size
.Pq default Sy 64 KiB .
.
.It Sy zfs_vdev_cache_size Ns = Ns Sy 0 Pq uint
Total size of the per-disk cache in bytes.
.Pp
Currently this feature is disabled, as it has been found to not be helpful
for performance and in some cases harmful.
.
.It Sy zfs_vdev_mirror_rotating_inc Ns = Ns Sy 0 Pq int
A number by which the balancing algorithm increments the load calculation for
the purpose of selecting the least busy mirror member when an I/O operation
immediately follows its predecessor on rotational vdevs
for the purpose of making decisions based on load.
.
.It Sy zfs_vdev_mirror_rotating_seek_inc Ns = Ns Sy 5 Pq int
A number by which the balancing algorithm increments the load calculation for
the purpose of selecting the least busy mirror member when an I/O operation
lacks locality as defined by
.Sy zfs_vdev_mirror_rotating_seek_offset .
Operations within this that are not immediately following the previous operation
are incremented by half.
.
.It Sy zfs_vdev_mirror_rotating_seek_offset Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq int
The maximum distance for the last queued I/O operation in which
the balancing algorithm considers an operation to have locality.
.No See Sx ZFS I/O SCHEDULER .
.
.It Sy zfs_vdev_mirror_non_rotating_inc Ns = Ns Sy 0 Pq int
A number by which the balancing algorithm increments the load calculation for
the purpose of selecting the least busy mirror member on non-rotational vdevs
when I/O operations do not immediately follow one another.
.
.It Sy zfs_vdev_mirror_non_rotating_seek_inc Ns = Ns Sy 1 Pq int
A number by which the balancing algorithm increments the load calculation for
the purpose of selecting the least busy mirror member when an I/O operation
lacks
locality as defined by the
.Sy zfs_vdev_mirror_rotating_seek_offset .
Operations within this that are not immediately following the previous operation
are incremented by half.
.
.It Sy zfs_vdev_read_gap_limit Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint
Aggregate read I/O operations if the on-disk gap between them is within this
threshold.
.
.It Sy zfs_vdev_write_gap_limit Ns = Ns Sy 4096 Ns B Po 4 KiB Pc Pq uint
Aggregate write I/O operations if the on-disk gap between them is within this
threshold.
.
.It Sy zfs_vdev_raidz_impl Ns = Ns Sy fastest Pq string
Select the raidz parity implementation to use.
.Pp
Variants that don't depend on CPU-specific features
may be selected on module load, as they are supported on all systems.
The remaining options may only be set after the module is loaded,
as they are available only if the implementations are compiled in
and supported on the running system.
.Pp
Once the module is loaded,
.Pa /sys/module/zfs/parameters/zfs_vdev_raidz_impl
will show the available options,
with the currently selected one enclosed in square brackets.
.Pp
.TS
lb l l .
fastest selected by built-in benchmark
original original implementation
scalar scalar implementation
sse2 SSE2 instruction set 64-bit x86
ssse3 SSSE3 instruction set 64-bit x86
avx2 AVX2 instruction set 64-bit x86
avx512f AVX512F instruction set 64-bit x86
avx512bw AVX512F & AVX512BW instruction sets 64-bit x86
aarch64_neon NEON Aarch64/64-bit ARMv8
aarch64_neonx2 NEON with more unrolling Aarch64/64-bit ARMv8
powerpc_altivec Altivec PowerPC
.TE
.
.It Sy zfs_vdev_scheduler Pq charp
.Sy DEPRECATED .
Prints warning to kernel log for compatibility.
.
.It Sy zfs_zevent_len_max Ns = Ns Sy 512 Pq uint
Max event queue length.
Events in the queue can be viewed with
.Xr zpool-events 8 .
.
.It Sy zfs_zevent_retain_max Ns = Ns Sy 2000 Pq int
Maximum recent zevent records to retain for duplicate checking.
Setting this to
.Sy 0
disables duplicate detection.
.
.It Sy zfs_zevent_retain_expire_secs Ns = Ns Sy 900 Ns s Po 15 min Pc Pq int
Lifespan for a recent ereport that was retained for duplicate checking.
.
.It Sy zfs_zil_clean_taskq_maxalloc Ns = Ns Sy 1048576 Pq int
The maximum number of taskq entries that are allowed to be cached.
When this limit is exceeded transaction records (itxs)
will be cleaned synchronously.
.
.It Sy zfs_zil_clean_taskq_minalloc Ns = Ns Sy 1024 Pq int
The number of taskq entries that are pre-populated when the taskq is first
created and are immediately available for use.
.
.It Sy zfs_zil_clean_taskq_nthr_pct Ns = Ns Sy 100 Ns % Pq int
This controls the number of threads used by
.Sy dp_zil_clean_taskq .
The default value of
.Sy 100%
will create a maximum of one thread per cpu.
.
.It Sy zil_maxblocksize Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint
This sets the maximum block size used by the ZIL.
On very fragmented pools, lowering this
.Pq typically to Sy 36 KiB
can improve performance.
.
.It Sy zil_min_commit_timeout Ns = Ns Sy 5000 Pq u64
This sets the minimum delay in nanoseconds ZIL care to delay block commit,
waiting for more records.
If ZIL writes are too fast, kernel may not be able sleep for so short interval,
increasing log latency above allowed by
.Sy zfs_commit_timeout_pct .
.
.It Sy zil_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable the cache flush commands that are normally sent to disk by
the ZIL after an LWB write has completed.
Setting this will cause ZIL corruption on power loss
if a volatile out-of-order write cache is enabled.
.
.It Sy zil_replay_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int
Disable intent logging replay.
Can be disabled for recovery from corrupted ZIL.
.
.It Sy zil_slog_bulk Ns = Ns Sy 786432 Ns B Po 768 KiB Pc Pq u64
Limit SLOG write size per commit executed with synchronous priority.
Any writes above that will be executed with lower (asynchronous) priority
to limit potential SLOG device abuse by single active ZIL writer.
.
.It Sy zfs_zil_saxattr Ns = Ns Sy 1 Ns | Ns 0 Pq int
Setting this tunable to zero disables ZIL logging of new
.Sy xattr Ns = Ns Sy sa
records if the
.Sy org.openzfs:zilsaxattr
feature is enabled on the pool.
This would only be necessary to work around bugs in the ZIL logging or replay
code for this record type.
The tunable has no effect if the feature is disabled.
.
.It Sy zfs_embedded_slog_min_ms Ns = Ns Sy 64 Pq uint
Usually, one metaslab from each normal-class vdev is dedicated for use by
the ZIL to log synchronous writes.
However, if there are fewer than
.Sy zfs_embedded_slog_min_ms
metaslabs in the vdev, this functionality is disabled.
This ensures that we don't set aside an unreasonable amount of space for the
ZIL.
.
.It Sy zstd_earlyabort_pass Ns = Ns Sy 1 Pq uint
Whether heuristic for detection of incompressible data with zstd levels >= 3
using LZ4 and zstd-1 passes is enabled.
.
.It Sy zstd_abort_size Ns = Ns Sy 131072 Pq uint
Minimal uncompressed size (inclusive) of a record before the early abort
heuristic will be attempted.
.
.It Sy zio_deadman_log_all Ns = Ns Sy 0 Ns | Ns 1 Pq int
If non-zero, the zio deadman will produce debugging messages
.Pq see Sy zfs_dbgmsg_enable
for all zios, rather than only for leaf zios possessing a vdev.
This is meant to be used by developers to gain
diagnostic information for hang conditions which don't involve a mutex
or other locking primitive: typically conditions in which a thread in
the zio pipeline is looping indefinitely.
.
.It Sy zio_slow_io_ms Ns = Ns Sy 30000 Ns ms Po 30 s Pc Pq int
When an I/O operation takes more than this much time to complete,
it's marked as slow.
Each slow operation causes a delay zevent.
Slow I/O counters can be seen with
.Nm zpool Cm status Fl s .
.
.It Sy zio_dva_throttle_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int
Throttle block allocations in the I/O pipeline.
This allows for dynamic allocation distribution when devices are imbalanced.
When enabled, the maximum number of pending allocations per top-level vdev
is limited by
.Sy zfs_vdev_queue_depth_pct .
.
.It Sy zfs_xattr_compat Ns = Ns 0 Ns | Ns 1 Pq int
Control the naming scheme used when setting new xattrs in the user namespace.
If
.Sy 0
.Pq the default on Linux ,
user namespace xattr names are prefixed with the namespace, to be backwards
compatible with previous versions of ZFS on Linux.
If
.Sy 1
.Pq the default on Fx ,
user namespace xattr names are not prefixed, to be backwards compatible with
previous versions of ZFS on illumos and
.Fx .
.Pp
Either naming scheme can be read on this and future versions of ZFS, regardless
of this tunable, but legacy ZFS on illumos or
.Fx
are unable to read user namespace xattrs written in the Linux format, and
legacy versions of ZFS on Linux are unable to read user namespace xattrs written
in the legacy ZFS format.
.Pp
An existing xattr with the alternate naming scheme is removed when overwriting
the xattr so as to not accumulate duplicates.
.
.It Sy zio_requeue_io_start_cut_in_line Ns = Ns Sy 0 Ns | Ns 1 Pq int
Prioritize requeued I/O.
.
.It Sy zio_taskq_batch_pct Ns = Ns Sy 80 Ns % Pq uint
Percentage of online CPUs which will run a worker thread for I/O.
These workers are responsible for I/O work such as compression and
checksum calculations.
Fractional number of CPUs will be rounded down.
.Pp
The default value of
.Sy 80%
was chosen to avoid using all CPUs which can result in
latency issues and inconsistent application performance,
especially when slower compression and/or checksumming is enabled.
.
.It Sy zio_taskq_batch_tpq Ns = Ns Sy 0 Pq uint
Number of worker threads per taskq.
Lower values improve I/O ordering and CPU utilization,
while higher reduces lock contention.
.Pp
If
.Sy 0 ,
generate a system-dependent value close to 6 threads per taskq.
.
.It Sy zvol_inhibit_dev Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Do not create zvol device nodes.
This may slightly improve startup time on
systems with a very large number of zvols.
.
.It Sy zvol_major Ns = Ns Sy 230 Pq uint
Major number for zvol block devices.
.
.It Sy zvol_max_discard_blocks Ns = Ns Sy 16384 Pq long
Discard (TRIM) operations done on zvols will be done in batches of this
many blocks, where block size is determined by the
.Sy volblocksize
property of a zvol.
.
.It Sy zvol_prefetch_bytes Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint
When adding a zvol to the system, prefetch this many bytes
from the start and end of the volume.
Prefetching these regions of the volume is desirable,
because they are likely to be accessed immediately by
.Xr blkid 8
or the kernel partitioner.
.
.It Sy zvol_request_sync Ns = Ns Sy 0 Ns | Ns 1 Pq uint
When processing I/O requests for a zvol, submit them synchronously.
This effectively limits the queue depth to
.Em 1
for each I/O submitter.
When unset, requests are handled asynchronously by a thread pool.
The number of requests which can be handled concurrently is controlled by
.Sy zvol_threads .
.Sy zvol_request_sync
is ignored when running on a kernel that supports block multiqueue
.Pq Li blk-mq .
.
.It Sy zvol_threads Ns = Ns Sy 0 Pq uint
The number of system wide threads to use for processing zvol block IOs.
If
.Sy 0
(the default) then internally set
.Sy zvol_threads
to the number of CPUs present or 32 (whichever is greater).
.
.It Sy zvol_blk_mq_threads Ns = Ns Sy 0 Pq uint
The number of threads per zvol to use for queuing IO requests.
This parameter will only appear if your kernel supports
.Li blk-mq
and is only read and assigned to a zvol at zvol load time.
If
.Sy 0
(the default) then internally set
.Sy zvol_blk_mq_threads
to the number of CPUs present.
.
.It Sy zvol_use_blk_mq Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Set to
.Sy 1
to use the
.Li blk-mq
API for zvols.
Set to
.Sy 0
(the default) to use the legacy zvol APIs.
This setting can give better or worse zvol performance depending on
the workload.
This parameter will only appear if your kernel supports
.Li blk-mq
and is only read and assigned to a zvol at zvol load time.
.
.It Sy zvol_blk_mq_blocks_per_thread Ns = Ns Sy 8 Pq uint
If
.Sy zvol_use_blk_mq
is enabled, then process this number of
.Sy volblocksize Ns -sized blocks per zvol thread.
This tunable can be use to favor better performance for zvol reads (lower
values) or writes (higher values).
If set to
.Sy 0 ,
then the zvol layer will process the maximum number of blocks
per thread that it can.
This parameter will only appear if your kernel supports
.Li blk-mq
and is only applied at each zvol's load time.
.
.It Sy zvol_blk_mq_queue_depth Ns = Ns Sy 0 Pq uint
The queue_depth value for the zvol
.Li blk-mq
interface.
This parameter will only appear if your kernel supports
.Li blk-mq
and is only applied at each zvol's load time.
If
.Sy 0
(the default) then use the kernel's default queue depth.
Values are clamped to the kernel's
.Dv BLKDEV_MIN_RQ
and
.Dv BLKDEV_MAX_RQ Ns / Ns Dv BLKDEV_DEFAULT_RQ
limits.
.
.It Sy zvol_volmode Ns = Ns Sy 1 Pq uint
Defines zvol block devices behaviour when
.Sy volmode Ns = Ns Sy default :
.Bl -tag -compact -offset 4n -width "a"
.It Sy 1
.No equivalent to Sy full
.It Sy 2
.No equivalent to Sy dev
.It Sy 3
.No equivalent to Sy none
.El
.
.It Sy zvol_enforce_quotas Ns = Ns Sy 0 Ns | Ns 1 Pq uint
Enable strict ZVOL quota enforcement.
The strict quota enforcement may have a performance impact.
.El
.
.Sh ZFS I/O SCHEDULER
ZFS issues I/O operations to leaf vdevs to satisfy and complete I/O operations.
The scheduler determines when and in what order those operations are issued.
The scheduler divides operations into five I/O classes,
prioritized in the following order: sync read, sync write, async read,
async write, and scrub/resilver.
Each queue defines the minimum and maximum number of concurrent operations
that may be issued to the device.
In addition, the device has an aggregate maximum,
.Sy zfs_vdev_max_active .
Note that the sum of the per-queue minima must not exceed the aggregate maximum.
If the sum of the per-queue maxima exceeds the aggregate maximum,
then the number of active operations may reach
.Sy zfs_vdev_max_active ,
in which case no further operations will be issued,
regardless of whether all per-queue minima have been met.
.Pp
For many physical devices, throughput increases with the number of
concurrent operations, but latency typically suffers.
Furthermore, physical devices typically have a limit
at which more concurrent operations have no
effect on throughput or can actually cause it to decrease.
.Pp
The scheduler selects the next operation to issue by first looking for an
I/O class whose minimum has not been satisfied.
Once all are satisfied and the aggregate maximum has not been hit,
the scheduler looks for classes whose maximum has not been satisfied.
Iteration through the I/O classes is done in the order specified above.
No further operations are issued
if the aggregate maximum number of concurrent operations has been hit,
or if there are no operations queued for an I/O class that has not hit its
maximum.
Every time an I/O operation is queued or an operation completes,
the scheduler looks for new operations to issue.
.Pp
In general, smaller
.Sy max_active Ns s
will lead to lower latency of synchronous operations.
Larger
.Sy max_active Ns s
may lead to higher overall throughput, depending on underlying storage.
.Pp
The ratio of the queues'
.Sy max_active Ns s
determines the balance of performance between reads, writes, and scrubs.
For example, increasing
.Sy zfs_vdev_scrub_max_active
will cause the scrub or resilver to complete more quickly,
but reads and writes to have higher latency and lower throughput.
.Pp
All I/O classes have a fixed maximum number of outstanding operations,
except for the async write class.
Asynchronous writes represent the data that is committed to stable storage
during the syncing stage for transaction groups.
Transaction groups enter the syncing state periodically,
so the number of queued async writes will quickly burst up
and then bleed down to zero.
Rather than servicing them as quickly as possible,
the I/O scheduler changes the maximum number of active async write operations
according to the amount of dirty data in the pool.
Since both throughput and latency typically increase with the number of
concurrent operations issued to physical devices, reducing the
burstiness in the number of simultaneous operations also stabilizes the
response time of operations from other queues, in particular synchronous ones.
In broad strokes, the I/O scheduler will issue more concurrent operations
from the async write queue as there is more dirty data in the pool.
.
.Ss Async Writes
The number of concurrent operations issued for the async write I/O class
follows a piece-wise linear function defined by a few adjustable points:
.Bd -literal
| o---------| <-- \fBzfs_vdev_async_write_max_active\fP
^ | /^ |
| | / | |
active | / | |
I/O | / | |
count | / | |
| / | |
|-------o | | <-- \fBzfs_vdev_async_write_min_active\fP
0|_______^______|_________|
0% | | 100% of \fBzfs_dirty_data_max\fP
| |
| `-- \fBzfs_vdev_async_write_active_max_dirty_percent\fP
`--------- \fBzfs_vdev_async_write_active_min_dirty_percent\fP
.Ed
.Pp
Until the amount of dirty data exceeds a minimum percentage of the dirty
data allowed in the pool, the I/O scheduler will limit the number of
concurrent operations to the minimum.
As that threshold is crossed, the number of concurrent operations issued
increases linearly to the maximum at the specified maximum percentage
of the dirty data allowed in the pool.
.Pp
Ideally, the amount of dirty data on a busy pool will stay in the sloped
part of the function between
.Sy zfs_vdev_async_write_active_min_dirty_percent
and
.Sy zfs_vdev_async_write_active_max_dirty_percent .
If it exceeds the maximum percentage,
this indicates that the rate of incoming data is
greater than the rate that the backend storage can handle.
In this case, we must further throttle incoming writes,
as described in the next section.
.
.Sh ZFS TRANSACTION DELAY
We delay transactions when we've determined that the backend storage
isn't able to accommodate the rate of incoming writes.
.Pp
If there is already a transaction waiting, we delay relative to when
that transaction will finish waiting.
This way the calculated delay time
is independent of the number of threads concurrently executing transactions.
.Pp
If we are the only waiter, wait relative to when the transaction started,
rather than the current time.
This credits the transaction for "time already served",
e.g. reading indirect blocks.
.Pp
The minimum time for a transaction to take is calculated as
.D1 min_time = min( Ns Sy zfs_delay_scale No \(mu Po Sy dirty No \- Sy min Pc / Po Sy max No \- Sy dirty Pc , 100ms)
.Pp
The delay has two degrees of freedom that can be adjusted via tunables.
The percentage of dirty data at which we start to delay is defined by
.Sy zfs_delay_min_dirty_percent .
This should typically be at or above
.Sy zfs_vdev_async_write_active_max_dirty_percent ,
so that we only start to delay after writing at full speed
has failed to keep up with the incoming write rate.
The scale of the curve is defined by
.Sy zfs_delay_scale .
Roughly speaking, this variable determines the amount of delay at the midpoint
of the curve.
.Bd -literal
delay
10ms +-------------------------------------------------------------*+
| *|
9ms + *+
| *|
8ms + *+
| * |
7ms + * +
| * |
6ms + * +
| * |
5ms + * +
| * |
4ms + * +
| * |
3ms + * +
| * |
2ms + (midpoint) * +
| | ** |
1ms + v *** +
| \fBzfs_delay_scale\fP ----------> ******** |
0 +-------------------------------------*********----------------+
0% <- \fBzfs_dirty_data_max\fP -> 100%
.Ed
.Pp
Note, that since the delay is added to the outstanding time remaining on the
most recent transaction it's effectively the inverse of IOPS.
Here, the midpoint of
.Em 500 us
translates to
.Em 2000 IOPS .
The shape of the curve
was chosen such that small changes in the amount of accumulated dirty data
in the first three quarters of the curve yield relatively small differences
in the amount of delay.
.Pp
The effects can be easier to understand when the amount of delay is
represented on a logarithmic scale:
.Bd -literal
delay
100ms +-------------------------------------------------------------++
+ +
| |
+ *+
10ms + *+
+ ** +
| (midpoint) ** |
+ | ** +
1ms + v **** +
+ \fBzfs_delay_scale\fP ----------> ***** +
| **** |
+ **** +
100us + ** +
+ * +
| * |
+ * +
10us + * +
+ +
| |
+ +
+--------------------------------------------------------------+
0% <- \fBzfs_dirty_data_max\fP -> 100%
.Ed
.Pp
Note here that only as the amount of dirty data approaches its limit does
the delay start to increase rapidly.
The goal of a properly tuned system should be to keep the amount of dirty data
out of that range by first ensuring that the appropriate limits are set
for the I/O scheduler to reach optimal throughput on the back-end storage,
and then by changing the value of
.Sy zfs_delay_scale
to increase the steepness of the curve.
diff --git a/sys/contrib/openzfs/man/man7/dracut.zfs.7 b/sys/contrib/openzfs/man/man7/dracut.zfs.7
index 2db2f6639eaf..c1475c695e83 100644
--- a/sys/contrib/openzfs/man/man7/dracut.zfs.7
+++ b/sys/contrib/openzfs/man/man7/dracut.zfs.7
@@ -1,282 +1,282 @@
.\" SPDX-License-Identifier: 0BSD
.\"
-.Dd April 4, 2022
+.Dd March 28, 2023
.Dt DRACUT.ZFS 7
.Os
.
.Sh NAME
.Nm dracut.zfs
.Nd overview of ZFS dracut hooks
.
.Sh SYNOPSIS
.Bd -literal -compact
parse-zfs.sh \(-> dracut-cmdline.service
| \(da
| …
| \(da
\e\(em\(em\(em\(em\(em\(em\(em\(em\(-> dracut-initqueue.service
| zfs-import-opts.sh
zfs-load-module.service \(da | |
| | sysinit.target \(da |
\(da | | zfs-import-scan.service \(da
zfs-import-scan.service \(da \(da | zfs-import-cache.service
| zfs-import-cache.service basic.target | |
\e__________________| | \(da \(da
\(da | zfs-load-key.sh
zfs-env-bootfs.service | |
\(da \(da \(da
zfs-import.target \(-> dracut-pre-mount.service
| \(ua |
| dracut-zfs-generator |
- | ____________________/|
+ | _____________________/|
|/ \(da
- | sysroot.mount \(<-\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em dracut-zfs-generator
- | | \(da |
- | \(da sysroot-{usr,etc,lib,&c.}.mount |
- | initrd-root-fs.target \(<-\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em or \(da
- | | zfs-nonroot-necessities.service
+ | sysroot.mount \(<-\(em\(em\(em dracut-zfs-generator
+ | |
+ | \(da
+ | initrd-root-fs.target \(<-\(em zfs-nonroot-necessities.service
+ | | |
| \(da |
\(da dracut-mount.service |
zfs-snapshot-bootfs.service | |
| \(da |
\(da … |
zfs-rollback-bootfs.service | |
| \(da |
- | sysroot-usr.mount \(<-\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em/
+ | /sysroot/{usr,etc,lib,&c.} \(<-\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em\(em/
| |
| \(da
| initrd-fs.target
\e______________________ |
\e|
\(da
export-zfs.sh initrd.target
| |
\(da \(da
dracut-shutdown.service …
|
\(da
zfs-needshutdown.sh \(-> initrd-cleanup.service
.Ed
.Pp
Compare
.Xr dracut.bootup 7
for the full flowchart.
.
.Sh DESCRIPTION
Under dracut, booting with
.No ZFS-on- Ns Pa /
is facilitated by a number of hooks in the
.Nm 90zfs
module.
.Pp
Booting into a ZFS dataset requires
.Sy mountpoint Ns = Ns Pa /
to be set on the dataset containing the root filesystem (henceforth "the boot
dataset") and at the very least either the
.Sy bootfs
property to be set to that dataset, or the
.Sy root=
kernel cmdline (or dracut drop-in) argument to specify it.
.Pp
All children of the boot dataset with
.Sy canmount Ns = Ns Sy on
with
.Sy mountpoint Ns s
matching
.Pa /etc , /bin , /lib , /lib?? , /libx32 , No and Pa /usr
globs are deemed essential and will be mounted as well.
.Pp
.Xr zfs-mount-generator 8
is recommended for proper functioning of the system afterward (correct mount
properties, remounting, &c.).
.
.Sh CMDLINE
.Ss Standard
.Bl -tag -compact -width ".Sy root=zfs:AUTO , root=zfs: , root=zfs , Op Sy root="
.It Sy root=zfs:\& Ns Ar dataset , Sy root=ZFS= Ns Ar dataset
Use
.Ar dataset
as the boot dataset.
All pluses
.Pq Sq +
are replaced with spaces
.Pq Sq \ .
.
.It Sy root=zfs:AUTO , root=zfs:\& , root=zfs , Op Sy root=
After import, search for the first pool with the
.Sy bootfs
property set, use its value as-if specified as the
.Ar dataset
above.
.
.It Sy rootfstype=zfs root= Ns Ar dataset
Equivalent to
.Sy root=zfs:\& Ns Ar dataset .
.
.It Sy rootfstype=zfs Op Sy root=
Equivalent to
.Sy root=zfs:AUTO .
.
.It Sy rootflags= Ns Ar flags
Mount the boot dataset with
.Fl o Ar flags ;
cf.\&
.Sx Temporary Mount Point Properties
in
.Xr zfsprops 7 .
These properties will not last, since all filesystems will be re-mounted from
the real root.
.
.It Sy debug
If specified,
.Nm dracut-zfs-generator
logs to the journal.
.El
.Pp
Be careful about setting neither
.Sy rootfstype=zfs
nor
.Sy root=zfs:\& Ns Ar dataset
\(em other automatic boot selection methods, like
.Nm systemd-gpt-auto-generator
and
.Nm systemd-fstab-generator
might take precedent.
.
.Ss ZFS-specific
.Bl -tag -compact -width ".Sy bootfs.snapshot Ns Op Sy = Ns Ar snapshot-name"
.It Sy bootfs.snapshot Ns Op Sy = Ns Ar snapshot-name
Execute
.Nm zfs Cm snapshot Ar boot-dataset Ns Sy @ Ns Ar snapshot-name
before pivoting to the real root.
.Ar snapshot-name
defaults to the current kernel release.
.
.It Sy bootfs.rollback Ns Op Sy = Ns Ar snapshot-name
Execute
.Nm zfs Cm snapshot Fl Rf Ar boot-dataset Ns Sy @ Ns Ar snapshot-name
before pivoting to the real root.
.Ar snapshot-name
defaults to the current kernel release.
.
.It Sy spl_hostid= Ns Ar host-id
Use
.Xr zgenhostid 8
to set the host ID to
.Ar host-id ;
otherwise,
.Pa /etc/hostid
inherited from the real root is used.
.
.It Sy zfs_force , zfs.force , zfsforce
Appends
.Fl f
to all
.Nm zpool Cm import
invocations; primarily useful in conjunction with
.Sy spl_hostid= ,
or if no host ID was inherited.
.El
.
.Sh FILES
.Bl -tag -width 0
.It Pa parse-zfs.sh Pq Sy cmdline
Processes
.Sy spl_hostid= .
If
.Sy root=
matches a known pattern, above, provides
.Pa /dev/root
and delays the initqueue until
.Xr zfs 4
is loaded,
.
.It Pa zfs-import-opts.sh Pq Nm systemd No environment generator
Turns
.Sy zfs_force , zfs.force , No or Sy zfsforce
into
.Ev ZPOOL_IMPORT_OPTS Ns = Ns Fl f
for
.Pa zfs-import-scan.service
or
.Pa zfs-import-cache.service .
.
.It Pa zfs-load-key.sh Pq Sy pre-mount
Loads encryption keys for the boot dataset and its essential descendants.
.Bl -tag -compact -offset 4n -width ".Sy keylocation Ns = Ns Sy https:// Ns Ar URL , Sy keylocation Ns = Ns Sy http:// Ns Ar URL"
.It Sy keylocation Ns = Ns Sy prompt
Is prompted for via
.Nm systemd-ask-password
thrice.
.
.It Sy keylocation Ns = Ns Sy https:// Ns Ar URL , Sy keylocation Ns = Ns Sy http:// Ns Ar URL
.Pa network-online.target
is started before loading.
.
.It Sy keylocation Ns = Ns Sy file:// Ns Ar path
If
.Ar path
doesn't exist,
.Nm udevadm No is Cm settle Ns d .
If it still doesn't, it's waited for for up to
.Sy 10 Ns s .
.El
.
.It Pa zfs-env-bootfs.service Pq Nm systemd No service
After pool import, sets
.Ev BOOTFS Ns =
in the systemd environment to the first non-null
.Sy bootfs
value in iteration order.
.
.It Pa dracut-zfs-generator Pq Nm systemd No generator
Generates
.Pa sysroot.mount Pq using Sy rootflags= , No if any .
If an explicit boot dataset was specified, also generates essential mountpoints
.Pq Pa sysroot-etc.mount , sysroot-bin.mount , No &c.\& ,
otherwise generates
.Pa zfs-nonroot-necessities.service
which mounts them explicitly after
.Pa /sysroot
using
.Ev BOOTFS Ns = .
.
.It Pa zfs-snapshot-bootfs.service , zfs-rollback-bootfs.service Pq Nm systemd No services
Consume
.Sy bootfs.snapshot
and
.Sy bootfs.rollback
as described in
.Sx CMDLINE .
Use
.Ev BOOTFS Ns =
if no explicit boot dataset was specified.
.
.It Pa zfs-needshutdown.sh Pq Sy cleanup
If any pools were imported, signals that shutdown hooks are required.
.
.It Pa export-zfs.sh Pq Sy shutdown
Forcibly exports all pools.
.
.It Pa /etc/hostid , /etc/zfs/zpool.cache , /etc/zfs/vdev_id.conf Pq regular files
Included verbatim, hostonly.
.
.It Pa mount-zfs.sh Pq Sy mount
Does nothing on
.Nm systemd
systems
.Pq if Pa dracut-zfs-generator No succeeded .
Otherwise, loads encryption key for the boot dataset from the console or via
plymouth.
It may not work at all!
.El
.
.Sh SEE ALSO
.Xr dracut.bootup 7 ,
.Xr zfsprops 7 ,
.Xr zpoolprops 7 ,
.Xr dracut-shutdown.service 8 ,
.Xr systemd-fstab-generator 8 ,
.Xr systemd-gpt-auto-generator 8 ,
.Xr zfs-mount-generator 8 ,
.Xr zgenhostid 8
diff --git a/sys/contrib/openzfs/man/man7/zfsprops.7 b/sys/contrib/openzfs/man/man7/zfsprops.7
index 1f6ffb59e32f..8f6b919cfc0b 100644
--- a/sys/contrib/openzfs/man/man7/zfsprops.7
+++ b/sys/contrib/openzfs/man/man7/zfsprops.7
@@ -1,2108 +1,2112 @@
.\"
.\" CDDL HEADER START
.\"
.\" The contents of this file are subject to the terms of the
.\" Common Development and Distribution License (the "License").
.\" You may not use this file except in compliance with the License.
.\"
.\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
.\" or https://opensource.org/licenses/CDDL-1.0.
.\" See the License for the specific language governing permissions
.\" and limitations under the License.
.\"
.\" When distributing Covered Code, include this CDDL HEADER in each
.\" file and include the License file at usr/src/OPENSOLARIS.LICENSE.
.\" If applicable, add the following below this CDDL HEADER, with the
.\" fields enclosed by brackets "[]" replaced with your own identifying
.\" information: Portions Copyright [yyyy] [name of copyright owner]
.\"
.\" CDDL HEADER END
.\"
.\" Copyright (c) 2009 Sun Microsystems, Inc. All Rights Reserved.
.\" Copyright 2011 Joshua M. Clulow <josh@sysmgr.org>
.\" Copyright (c) 2011, 2019 by Delphix. All rights reserved.
.\" Copyright (c) 2011, Pawel Jakub Dawidek <pjd@FreeBSD.org>
.\" Copyright (c) 2012, Glen Barber <gjb@FreeBSD.org>
.\" Copyright (c) 2012, Bryan Drewery <bdrewery@FreeBSD.org>
.\" Copyright (c) 2013, Steven Hartland <smh@FreeBSD.org>
.\" Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
.\" Copyright (c) 2014, Joyent, Inc. All rights reserved.
.\" Copyright (c) 2014 by Adam Stevko. All rights reserved.
.\" Copyright (c) 2014 Integros [integros.com]
.\" Copyright (c) 2016 Nexenta Systems, Inc. All Rights Reserved.
.\" Copyright (c) 2014, Xin LI <delphij@FreeBSD.org>
.\" Copyright (c) 2014-2015, The FreeBSD Foundation, All Rights Reserved.
.\" Copyright 2019 Richard Laager. All rights reserved.
.\" Copyright 2018 Nexenta Systems, Inc.
.\" Copyright 2019 Joyent, Inc.
.\" Copyright (c) 2019, Kjeld Schouten-Lebbing
.\" Copyright (c) 2022 Hewlett Packard Enterprise Development LP.
.\"
-.Dd July 21, 2022
+.Dd April 18, 2023
.Dt ZFSPROPS 7
.Os
.
.Sh NAME
.Nm zfsprops
.Nd native and user-defined properties of ZFS datasets
.
.Sh DESCRIPTION
Properties are divided into two types, native properties and user-defined
.Po or
.Qq user
.Pc
properties.
Native properties either export internal statistics or control ZFS behavior.
In addition, native properties are either editable or read-only.
User properties have no effect on ZFS behavior, but you can use them to annotate
datasets in a way that is meaningful in your environment.
For more information about user properties, see the
.Sx User Properties
section, below.
.
.Ss Native Properties
Every dataset has a set of properties that export statistics about the dataset
as well as control various behaviors.
Properties are inherited from the parent unless overridden by the child.
Some properties apply only to certain types of datasets
.Pq file systems, volumes, or snapshots .
.Pp
The values of numeric properties can be specified using human-readable suffixes
.Po for example,
.Sy k ,
.Sy KB ,
.Sy M ,
.Sy Gb ,
and so forth, up to
.Sy Z
for zettabyte
.Pc .
The following are all valid
.Pq and equal
specifications:
-.Li 1536M, 1.5g, 1.50GB .
+.Li 1536M ,
+.Li 1.5g ,
+.Li 1.50GB .
.Pp
The values of non-numeric properties are case sensitive and must be lowercase,
except for
.Sy mountpoint ,
.Sy sharenfs ,
and
.Sy sharesmb .
.Pp
The following native properties consist of read-only statistics about the
dataset.
These properties can be neither set, nor inherited.
Native properties apply to all dataset types unless otherwise noted.
.Bl -tag -width "usedbyrefreservation"
.It Sy available
The amount of space available to the dataset and all its children, assuming that
there is no other activity in the pool.
Because space is shared within a pool, availability can be limited by any number
of factors, including physical pool size, quotas, reservations, or other
datasets within the pool.
.Pp
This property can also be referred to by its shortened column name,
.Sy avail .
.It Sy compressratio
For non-snapshots, the compression ratio achieved for the
.Sy used
space of this dataset, expressed as a multiplier.
The
.Sy used
property includes descendant datasets, and, for clones, does not include the
space shared with the origin snapshot.
For snapshots, the
.Sy compressratio
is the same as the
.Sy refcompressratio
property.
Compression can be turned on by running:
.Nm zfs Cm set Sy compression Ns = Ns Sy on Ar dataset .
The default value is
.Sy off .
.It Sy createtxg
The transaction group (txg) in which the dataset was created.
Bookmarks have the same
.Sy createtxg
as the snapshot they are initially tied to.
This property is suitable for ordering a list of snapshots,
e.g. for incremental send and receive.
.It Sy creation
The time this dataset was created.
.It Sy clones
For snapshots, this property is a comma-separated list of filesystems or volumes
which are clones of this snapshot.
The clones'
.Sy origin
property is this snapshot.
If the
.Sy clones
property is not empty, then this snapshot can not be destroyed
.Po even with the
.Fl r
or
.Fl f
options
.Pc .
The roles of origin and clone can be swapped by promoting the clone with the
.Nm zfs Cm promote
command.
.It Sy defer_destroy
This property is
.Sy on
if the snapshot has been marked for deferred destroy by using the
.Nm zfs Cm destroy Fl d
command.
Otherwise, the property is
.Sy off .
.It Sy encryptionroot
For encrypted datasets, indicates where the dataset is currently inheriting its
encryption key from.
Loading or unloading a key for the
.Sy encryptionroot
will implicitly load / unload the key for any inheriting datasets (see
.Nm zfs Cm load-key
and
.Nm zfs Cm unload-key
for details).
Clones will always share an
encryption key with their origin.
See the
.Sx Encryption
section of
.Xr zfs-load-key 8
for details.
.It Sy filesystem_count
The total number of filesystems and volumes that exist under this location in
the dataset tree.
This value is only available when a
.Sy filesystem_limit
has been set somewhere in the tree under which the dataset resides.
.It Sy keystatus
Indicates if an encryption key is currently loaded into ZFS.
The possible values are
.Sy none ,
.Sy available ,
and
.Sy unavailable .
See
.Nm zfs Cm load-key
and
.Nm zfs Cm unload-key .
.It Sy guid
The 64 bit GUID of this dataset or bookmark which does not change over its
entire lifetime.
When a snapshot is sent to another pool, the received snapshot has the same
GUID.
Thus, the
.Sy guid
is suitable to identify a snapshot across pools.
.It Sy logicalreferenced
The amount of space that is
.Qq logically
accessible by this dataset.
See the
.Sy referenced
property.
The logical space ignores the effect of the
.Sy compression
and
.Sy copies
properties, giving a quantity closer to the amount of data that applications
see.
However, it does include space consumed by metadata.
.Pp
This property can also be referred to by its shortened column name,
.Sy lrefer .
.It Sy logicalused
The amount of space that is
.Qq logically
consumed by this dataset and all its descendents.
See the
.Sy used
property.
The logical space ignores the effect of the
.Sy compression
and
.Sy copies
properties, giving a quantity closer to the amount of data that applications
see.
However, it does include space consumed by metadata.
.Pp
This property can also be referred to by its shortened column name,
.Sy lused .
.It Sy mounted
For file systems, indicates whether the file system is currently mounted.
This property can be either
.Sy yes
or
.Sy no .
.It Sy objsetid
A unique identifier for this dataset within the pool.
Unlike the dataset's
.Sy guid , No the Sy objsetid
of a dataset is not transferred to other pools when the snapshot is copied
with a send/receive operation.
The
.Sy objsetid
can be reused (for a new dataset) after the dataset is deleted.
.It Sy origin
For cloned file systems or volumes, the snapshot from which the clone was
created.
See also the
.Sy clones
property.
.It Sy receive_resume_token
For filesystems or volumes which have saved partially-completed state from
.Nm zfs Cm receive Fl s ,
this opaque token can be provided to
.Nm zfs Cm send Fl t
to resume and complete the
.Nm zfs Cm receive .
.It Sy redact_snaps
For bookmarks, this is the list of snapshot guids the bookmark contains a
redaction
list for.
For snapshots, this is the list of snapshot guids the snapshot is redacted with
respect to.
.It Sy referenced
The amount of data that is accessible by this dataset, which may or may not be
shared with other datasets in the pool.
When a snapshot or clone is created, it initially references the same amount of
space as the file system or snapshot it was created from, since its contents are
identical.
.Pp
This property can also be referred to by its shortened column name,
.Sy refer .
.It Sy refcompressratio
The compression ratio achieved for the
.Sy referenced
space of this dataset, expressed as a multiplier.
See also the
.Sy compressratio
property.
.It Sy snapshot_count
The total number of snapshots that exist under this location in the dataset
tree.
This value is only available when a
.Sy snapshot_limit
has been set somewhere in the tree under which the dataset resides.
.It Sy type
The type of dataset:
.Sy filesystem ,
.Sy volume ,
.Sy snapshot ,
or
.Sy bookmark .
.It Sy used
The amount of space consumed by this dataset and all its descendents.
This is the value that is checked against this dataset's quota and reservation.
The space used does not include this dataset's reservation, but does take into
account the reservations of any descendent datasets.
The amount of space that a dataset consumes from its parent, as well as the
amount of space that is freed if this dataset is recursively destroyed, is the
greater of its space used and its reservation.
.Pp
The used space of a snapshot
.Po see the
.Sx Snapshots
section of
.Xr zfsconcepts 7
.Pc
is space that is referenced exclusively by this snapshot.
If this snapshot is destroyed, the amount of
.Sy used
space will be freed.
Space that is shared by multiple snapshots isn't accounted for in this metric.
When a snapshot is destroyed, space that was previously shared with this
snapshot can become unique to snapshots adjacent to it, thus changing the used
space of those snapshots.
The used space of the latest snapshot can also be affected by changes in the
file system.
Note that the
.Sy used
space of a snapshot is a subset of the
.Sy written
space of the snapshot.
.Pp
The amount of space used, available, or referenced does not take into account
pending changes.
Pending changes are generally accounted for within a few seconds.
Committing a change to a disk using
.Xr fsync 2
or
.Sy O_SYNC
does not necessarily guarantee that the space usage information is updated
immediately.
.It Sy usedby*
The
.Sy usedby*
properties decompose the
.Sy used
properties into the various reasons that space is used.
Specifically,
.Sy used No =
.Sy usedbychildren No +
.Sy usedbydataset No +
.Sy usedbyrefreservation No +
.Sy usedbysnapshots .
These properties are only available for datasets created on
.Nm zpool
.Qo version 13 Qc
pools.
.It Sy usedbychildren
The amount of space used by children of this dataset, which would be freed if
all the dataset's children were destroyed.
.It Sy usedbydataset
The amount of space used by this dataset itself, which would be freed if the
dataset were destroyed
.Po after first removing any
.Sy refreservation
and destroying any necessary snapshots or descendents
.Pc .
.It Sy usedbyrefreservation
The amount of space used by a
.Sy refreservation
set on this dataset, which would be freed if the
.Sy refreservation
was removed.
.It Sy usedbysnapshots
The amount of space consumed by snapshots of this dataset.
In particular, it is the amount of space that would be freed if all of this
dataset's snapshots were destroyed.
Note that this is not simply the sum of the snapshots'
.Sy used
properties because space can be shared by multiple snapshots.
.It Sy userused Ns @ Ns Ar user
The amount of space consumed by the specified user in this dataset.
Space is charged to the owner of each file, as displayed by
.Nm ls Fl l .
The amount of space charged is displayed by
.Nm du No and Nm ls Fl s .
See the
.Nm zfs Cm userspace
command for more information.
.Pp
Unprivileged users can access only their own space usage.
The root user, or a user who has been granted the
.Sy userused
privilege with
.Nm zfs Cm allow ,
can access everyone's usage.
.Pp
The
.Sy userused Ns @ Ns Ar …
properties are not displayed by
.Nm zfs Cm get Sy all .
The user's name must be appended after the
.Sy @
symbol, using one of the following forms:
.Bl -bullet -compact -offset 4n
.It
POSIX name
.Pq Qq joe
.It
POSIX numeric ID
.Pq Qq 789
.It
SID name
.Pq Qq joe.smith@mydomain
.It
SID numeric ID
.Pq Qq S-1-123-456-789
.El
.Pp
Files created on Linux always have POSIX owners.
.It Sy userobjused Ns @ Ns Ar user
The
.Sy userobjused
property is similar to
.Sy userused
but instead it counts the number of objects consumed by a user.
This property counts all objects allocated on behalf of the user,
it may differ from the results of system tools such as
.Nm df Fl i .
.Pp
When the property
.Sy xattr Ns = Ns Sy on
is set on a file system additional objects will be created per-file to store
extended attributes.
These additional objects are reflected in the
.Sy userobjused
value and are counted against the user's
.Sy userobjquota .
When a file system is configured to use
.Sy xattr Ns = Ns Sy sa
no additional internal objects are normally required.
.It Sy userrefs
This property is set to the number of user holds on this snapshot.
User holds are set by using the
.Nm zfs Cm hold
command.
.It Sy groupused Ns @ Ns Ar group
The amount of space consumed by the specified group in this dataset.
Space is charged to the group of each file, as displayed by
.Nm ls Fl l .
See the
.Sy userused Ns @ Ns Ar user
property for more information.
.Pp
Unprivileged users can only access their own groups' space usage.
The root user, or a user who has been granted the
.Sy groupused
privilege with
.Nm zfs Cm allow ,
can access all groups' usage.
.It Sy groupobjused Ns @ Ns Ar group
The number of objects consumed by the specified group in this dataset.
Multiple objects may be charged to the group for each file when extended
attributes are in use.
See the
.Sy userobjused Ns @ Ns Ar user
property for more information.
.Pp
Unprivileged users can only access their own groups' space usage.
The root user, or a user who has been granted the
.Sy groupobjused
privilege with
.Nm zfs Cm allow ,
can access all groups' usage.
.It Sy projectused Ns @ Ns Ar project
The amount of space consumed by the specified project in this dataset.
Project is identified via the project identifier (ID) that is object-based
numeral attribute.
An object can inherit the project ID from its parent object (if the
parent has the flag of inherit project ID that can be set and changed via
.Nm chattr Fl /+P
or
.Nm zfs project Fl s )
when being created.
The privileged user can set and change object's project
ID via
.Nm chattr Fl p
or
.Nm zfs project Fl s
anytime.
Space is charged to the project of each file, as displayed by
.Nm lsattr Fl p
or
.Nm zfs project .
See the
.Sy userused Ns @ Ns Ar user
property for more information.
.Pp
The root user, or a user who has been granted the
.Sy projectused
privilege with
.Nm zfs allow ,
can access all projects' usage.
.It Sy projectobjused Ns @ Ns Ar project
The
.Sy projectobjused
is similar to
.Sy projectused
but instead it counts the number of objects consumed by project.
When the property
.Sy xattr Ns = Ns Sy on
is set on a fileset, ZFS will create additional objects per-file to store
extended attributes.
These additional objects are reflected in the
.Sy projectobjused
value and are counted against the project's
.Sy projectobjquota .
When a filesystem is configured to use
.Sy xattr Ns = Ns Sy sa
no additional internal objects are required.
See the
.Sy userobjused Ns @ Ns Ar user
property for more information.
.Pp
The root user, or a user who has been granted the
.Sy projectobjused
privilege with
.Nm zfs allow ,
can access all projects' objects usage.
.It Sy snapshots_changed
Provides a mechanism to quickly determine whether snapshot list has
changed without having to mount a dataset or iterate the snapshot list.
Specifies the time at which a snapshot for a dataset was last
created or deleted.
.Pp
This allows us to be more efficient how often we query snapshots.
The property is persistent across mount and unmount operations only if the
.Sy extensible_dataset
feature is enabled.
.It Sy volblocksize
For volumes, specifies the block size of the volume.
The
.Sy blocksize
cannot be changed once the volume has been written, so it should be set at
volume creation time.
The default
.Sy blocksize
for volumes is 16 Kbytes.
Any power of 2 from 512 bytes to 128 Kbytes is valid.
.Pp
This property can also be referred to by its shortened column name,
.Sy volblock .
.It Sy written
The amount of space
.Sy referenced
by this dataset, that was written since the previous snapshot
.Pq i.e. that is not referenced by the previous snapshot .
.It Sy written Ns @ Ns Ar snapshot
The amount of
.Sy referenced
space written to this dataset since the specified snapshot.
This is the space that is referenced by this dataset but was not referenced by
the specified snapshot.
.Pp
The
.Ar snapshot
may be specified as a short snapshot name
.Pq just the part after the Sy @ ,
in which case it will be interpreted as a snapshot in the same filesystem as
this dataset.
The
.Ar snapshot
may be a full snapshot name
.Pq Ar filesystem Ns @ Ns Ar snapshot ,
which for clones may be a snapshot in the origin's filesystem
.Pq or the origin of the origin's filesystem, etc.
.El
.Pp
The following native properties can be used to change the behavior of a ZFS
dataset.
.Bl -tag -width ""
.It Xo
.Sy aclinherit Ns = Ns Sy discard Ns | Ns Sy noallow Ns | Ns
.Sy restricted Ns | Ns Sy passthrough Ns | Ns Sy passthrough-x
.Xc
Controls how ACEs are inherited when files and directories are created.
.Bl -tag -compact -offset 4n -width "passthrough-x"
.It Sy discard
does not inherit any ACEs.
.It Sy noallow
only inherits inheritable ACEs that specify
.Qq deny
permissions.
.It Sy restricted
default, removes the
.Sy write_acl
and
.Sy write_owner
permissions when the ACE is inherited.
.It Sy passthrough
inherits all inheritable ACEs without any modifications.
.It Sy passthrough-x
same meaning as
.Sy passthrough ,
except that the
.Sy owner@ , group@ , No and Sy everyone@
ACEs inherit the execute permission only if the file creation mode also requests
the execute bit.
.El
.Pp
When the property value is set to
.Sy passthrough ,
files are created with a mode determined by the inheritable ACEs.
If no inheritable ACEs exist that affect the mode, then the mode is set in
accordance to the requested mode from the application.
.Pp
The
.Sy aclinherit
property does not apply to POSIX ACLs.
.It Xo
.Sy aclmode Ns = Ns Sy discard Ns | Ns Sy groupmask Ns | Ns
.Sy passthrough Ns | Ns Sy restricted Ns
.Xc
Controls how an ACL is modified during chmod(2) and how inherited ACEs
are modified by the file creation mode:
.Bl -tag -compact -offset 4n -width "passthrough"
.It Sy discard
default, deletes all
.Sy ACEs
except for those representing
the mode of the file or directory requested by
.Xr chmod 2 .
.It Sy groupmask
reduces permissions granted in all
.Sy ALLOW
entries found in the
.Sy ACL
such that they are no greater than the group permissions specified by
.Xr chmod 2 .
.It Sy passthrough
indicates that no changes are made to the ACL other than creating or updating
the necessary ACL entries to represent the new mode of the file or directory.
.It Sy restricted
will cause the
.Xr chmod 2
operation to return an error when used on any file or directory which has
a non-trivial ACL whose entries can not be represented by a mode.
.Xr chmod 2
is required to change the set user ID, set group ID, or sticky bits on a file
or directory, as they do not have equivalent ACL entries.
In order to use
.Xr chmod 2
on a file or directory with a non-trivial ACL when
.Sy aclmode
is set to
.Sy restricted ,
you must first remove all ACL entries which do not represent the current mode.
.El
.It Sy acltype Ns = Ns Sy off Ns | Ns Sy nfsv4 Ns | Ns Sy posix
Controls whether ACLs are enabled and if so what type of ACL to use.
When this property is set to a type of ACL not supported by the current
platform, the behavior is the same as if it were set to
.Sy off .
.Bl -tag -compact -offset 4n -width "posixacl"
.It Sy off
default on Linux, when a file system has the
.Sy acltype
property set to off then ACLs are disabled.
.It Sy noacl
an alias for
.Sy off
.It Sy nfsv4
default on
.Fx ,
indicates that NFSv4-style ZFS ACLs should be used.
These ACLs can be managed with the
.Xr getfacl 1
and
.Xr setfacl 1 .
The
.Sy nfsv4
ZFS ACL type is not yet supported on Linux.
.It Sy posix
indicates POSIX ACLs should be used.
POSIX ACLs are specific to Linux and are not functional on other platforms.
POSIX ACLs are stored as an extended
attribute and therefore will not overwrite any existing NFSv4 ACLs which
may be set.
.It Sy posixacl
an alias for
.Sy posix
.El
.Pp
To obtain the best performance when setting
.Sy posix
users are strongly encouraged to set the
.Sy xattr Ns = Ns Sy sa
property.
This will result in the POSIX ACL being stored more efficiently on disk.
But as a consequence, all new extended attributes will only be
accessible from OpenZFS implementations which support the
.Sy xattr Ns = Ns Sy sa
property.
See the
.Sy xattr
property for more details.
.It Sy atime Ns = Ns Sy on Ns | Ns Sy off
Controls whether the access time for files is updated when they are read.
Turning this property off avoids producing write traffic when reading files and
can result in significant performance gains, though it might confuse mailers
and other similar utilities.
The values
.Sy on
and
.Sy off
are equivalent to the
.Sy atime
and
.Sy noatime
mount options.
The default value is
.Sy on .
See also
.Sy relatime
below.
.It Sy canmount Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy noauto
If this property is set to
.Sy off ,
the file system cannot be mounted, and is ignored by
.Nm zfs Cm mount Fl a .
Setting this property to
.Sy off
is similar to setting the
.Sy mountpoint
property to
.Sy none ,
except that the dataset still has a normal
.Sy mountpoint
property, which can be inherited.
Setting this property to
.Sy off
allows datasets to be used solely as a mechanism to inherit properties.
One example of setting
.Sy canmount Ns = Ns Sy off
is to have two datasets with the same
.Sy mountpoint ,
so that the children of both datasets appear in the same directory, but might
have different inherited characteristics.
.Pp
When set to
.Sy noauto ,
a dataset can only be mounted and unmounted explicitly.
The dataset is not mounted automatically when the dataset is created or
imported, nor is it mounted by the
.Nm zfs Cm mount Fl a
command or unmounted by the
.Nm zfs Cm unmount Fl a
command.
.Pp
This property is not inherited.
.It Xo
.Sy checksum Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy fletcher2 Ns | Ns
.Sy fletcher4 Ns | Ns Sy sha256 Ns | Ns Sy noparity Ns | Ns
.Sy sha512 Ns | Ns Sy skein Ns | Ns Sy edonr Ns | Ns Sy blake3
.Xc
Controls the checksum used to verify data integrity.
The default value is
.Sy on ,
which automatically selects an appropriate algorithm
.Po currently,
.Sy fletcher4 ,
but this may change in future releases
.Pc .
The value
.Sy off
disables integrity checking on user data.
The value
.Sy noparity
not only disables integrity but also disables maintaining parity for user data.
This setting is used internally by a dump device residing on a RAID-Z pool and
should not be used by any other dataset.
Disabling checksums is
.Em NOT
a recommended practice.
.Pp
The
.Sy sha512 ,
.Sy skein ,
.Sy edonr ,
and
.Sy blake3
checksum algorithms require enabling the appropriate features on the pool.
.Pp
Please see
.Xr zpool-features 7
for more information on these algorithms.
.Pp
Changing this property affects only newly-written data.
.It Xo
.Sy compression Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy gzip Ns | Ns
.Sy gzip- Ns Ar N Ns | Ns Sy lz4 Ns | Ns Sy lzjb Ns | Ns Sy zle Ns | Ns Sy zstd Ns | Ns
.Sy zstd- Ns Ar N Ns | Ns Sy zstd-fast Ns | Ns Sy zstd-fast- Ns Ar N
.Xc
Controls the compression algorithm used for this dataset.
.Pp
When set to
.Sy on
(the default), indicates that the current default compression algorithm should
be used.
The default balances compression and decompression speed, with compression ratio
and is expected to work well on a wide variety of workloads.
Unlike all other settings for this property,
.Sy on
does not select a fixed compression type.
As new compression algorithms are added to ZFS and enabled on a pool, the
default compression algorithm may change.
The current default compression algorithm is either
.Sy lzjb
or, if the
.Sy lz4_compress
feature is enabled,
.Sy lz4 .
.Pp
The
.Sy lz4
compression algorithm is a high-performance replacement for the
.Sy lzjb
algorithm.
It features significantly faster compression and decompression, as well as a
moderately higher compression ratio than
.Sy lzjb ,
but can only be used on pools with the
.Sy lz4_compress
feature set to
.Sy enabled .
See
.Xr zpool-features 7
for details on ZFS feature flags and the
.Sy lz4_compress
feature.
.Pp
The
.Sy lzjb
compression algorithm is optimized for performance while providing decent data
compression.
.Pp
The
.Sy gzip
compression algorithm uses the same compression as the
.Xr gzip 1
command.
You can specify the
.Sy gzip
level by using the value
.Sy gzip- Ns Ar N ,
where
.Ar N
is an integer from 1
.Pq fastest
to 9
.Pq best compression ratio .
Currently,
.Sy gzip
is equivalent to
.Sy gzip-6
.Po which is also the default for
.Xr gzip 1
.Pc .
.Pp
The
.Sy zstd
compression algorithm provides both high compression ratios and good
performance.
You can specify the
.Sy zstd
level by using the value
.Sy zstd- Ns Ar N ,
where
.Ar N
is an integer from 1
.Pq fastest
to 19
.Pq best compression ratio .
.Sy zstd
is equivalent to
.Sy zstd-3 .
.Pp
Faster speeds at the cost of the compression ratio can be requested by
setting a negative
.Sy zstd
level.
This is done using
.Sy zstd-fast- Ns Ar N ,
where
.Ar N
is an integer in
.Bq Sy 1 Ns - Ns Sy 10 , 20 , 30 , No … , Sy 100 , 500 , 1000
which maps to a negative
.Sy zstd
level.
The lower the level the faster the compression \(em
.Sy 1000
provides the fastest compression and lowest compression ratio.
.Sy zstd-fast
is equivalent to
.Sy zstd-fast- Ns Ar 1 .
.Pp
The
.Sy zle
compression algorithm compresses runs of zeros.
.Pp
This property can also be referred to by its shortened column name
.Sy compress .
Changing this property affects only newly-written data.
.Pp
When any setting except
.Sy off
is selected, compression will explicitly check for blocks consisting of only
zeroes (the NUL byte).
When a zero-filled block is detected, it is stored as
a hole and not compressed using the indicated compression algorithm.
.Pp
Any block being compressed must be no larger than 7/8 of its original size
after compression, otherwise the compression will not be considered worthwhile
and the block saved uncompressed.
Note that when the logical block is less than
8 times the disk sector size this effectively reduces the necessary compression
ratio; for example, 8 KiB blocks on disks with 4 KiB disk sectors must compress
to 1/2
or less of their original size.
.It Xo
.Sy context Ns = Ns Sy none Ns | Ns
.Ar SELinux-User : Ns Ar SELinux-Role : Ns Ar SELinux-Type : Ns Ar Sensitivity-Level
.Xc
This flag sets the SELinux context for all files in the file system under
a mount point for that file system.
See
.Xr selinux 8
for more information.
.It Xo
.Sy fscontext Ns = Ns Sy none Ns | Ns
.Ar SELinux-User : Ns Ar SELinux-Role : Ns Ar SELinux-Type : Ns Ar Sensitivity-Level
.Xc
This flag sets the SELinux context for the file system file system being
mounted.
See
.Xr selinux 8
for more information.
.It Xo
.Sy defcontext Ns = Ns Sy none Ns | Ns
.Ar SELinux-User : Ns Ar SELinux-Role : Ns Ar SELinux-Type : Ns Ar Sensitivity-Level
.Xc
This flag sets the SELinux default context for unlabeled files.
See
.Xr selinux 8
for more information.
.It Xo
.Sy rootcontext Ns = Ns Sy none Ns | Ns
.Ar SELinux-User : Ns Ar SELinux-Role : Ns Ar SELinux-Type : Ns Ar Sensitivity-Level
.Xc
This flag sets the SELinux context for the root inode of the file system.
See
.Xr selinux 8
for more information.
.It Sy copies Ns = Ns Sy 1 Ns | Ns Sy 2 Ns | Ns Sy 3
Controls the number of copies of data stored for this dataset.
These copies are in addition to any redundancy provided by the pool, for
example, mirroring or RAID-Z.
The copies are stored on different disks, if possible.
The space used by multiple copies is charged to the associated file and dataset,
changing the
.Sy used
property and counting against quotas and reservations.
.Pp
Changing this property only affects newly-written data.
Therefore, set this property at file system creation time by using the
.Fl o Sy copies Ns = Ns Ar N
option.
.Pp
Remember that ZFS will not import a pool with a missing top-level vdev.
Do
.Em NOT
create, for example a two-disk striped pool and set
.Sy copies Ns = Ns Ar 2
on some datasets thinking you have setup redundancy for them.
When a disk fails you will not be able to import the pool
and will have lost all of your data.
.Pp
Encrypted datasets may not have
.Sy copies Ns = Ns Ar 3
since the implementation stores some encryption metadata where the third copy
would normally be.
.It Sy devices Ns = Ns Sy on Ns | Ns Sy off
Controls whether device nodes can be opened on this file system.
The default value is
.Sy on .
The values
.Sy on
and
.Sy off
are equivalent to the
.Sy dev
and
.Sy nodev
mount options.
.It Xo
.Sy dedup Ns = Ns Sy off Ns | Ns Sy on Ns | Ns Sy verify Ns | Ns
.Sy sha256 Ns Oo , Ns Sy verify Oc Ns | Ns Sy sha512 Ns Oo , Ns Sy verify Oc Ns | Ns Sy skein Ns Oo , Ns Sy verify Oc Ns | Ns
.Sy edonr , Ns Sy verify Ns | Ns Sy blake3 Ns Oo , Ns Sy verify Oc Ns
.Xc
Configures deduplication for a dataset.
The default value is
.Sy off .
The default deduplication checksum is
.Sy sha256
(this may change in the future).
When
.Sy dedup
is enabled, the checksum defined here overrides the
.Sy checksum
property.
Setting the value to
.Sy verify
has the same effect as the setting
.Sy sha256 , Ns Sy verify .
.Pp
If set to
.Sy verify ,
ZFS will do a byte-to-byte comparison in case of two blocks having the same
signature to make sure the block contents are identical.
Specifying
.Sy verify
is mandatory for the
.Sy edonr
algorithm.
.Pp
Unless necessary, deduplication should
.Em not
be enabled on a system.
See the
.Sx Deduplication
section of
.Xr zfsconcepts 7 .
.It Xo
.Sy dnodesize Ns = Ns Sy legacy Ns | Ns Sy auto Ns | Ns Sy 1k Ns | Ns
.Sy 2k Ns | Ns Sy 4k Ns | Ns Sy 8k Ns | Ns Sy 16k
.Xc
Specifies a compatibility mode or literal value for the size of dnodes in the
file system.
The default value is
.Sy legacy .
Setting this property to a value other than
.Sy legacy No requires the Sy large_dnode No pool feature to be enabled .
.Pp
Consider setting
.Sy dnodesize
to
.Sy auto
if the dataset uses the
.Sy xattr Ns = Ns Sy sa
property setting and the workload makes heavy use of extended attributes.
This
may be applicable to SELinux-enabled systems, Lustre servers, and Samba
servers, for example.
Literal values are supported for cases where the optimal
size is known in advance and for performance testing.
.Pp
Leave
.Sy dnodesize
set to
.Sy legacy
if you need to receive a send stream of this dataset on a pool that doesn't
enable the
.Sy large_dnode
feature, or if you need to import this pool on a system that doesn't support the
.Sy large_dnode No feature .
.Pp
This property can also be referred to by its shortened column name,
.Sy dnsize .
.It Xo
.Sy encryption Ns = Ns Sy off Ns | Ns Sy on Ns | Ns Sy aes-128-ccm Ns | Ns
.Sy aes-192-ccm Ns | Ns Sy aes-256-ccm Ns | Ns Sy aes-128-gcm Ns | Ns
.Sy aes-192-gcm Ns | Ns Sy aes-256-gcm
.Xc
Controls the encryption cipher suite (block cipher, key length, and mode) used
for this dataset.
Requires the
.Sy encryption
feature to be enabled on the pool.
Requires a
.Sy keyformat
to be set at dataset creation time.
.Pp
Selecting
.Sy encryption Ns = Ns Sy on
when creating a dataset indicates that the default encryption suite will be
selected, which is currently
.Sy aes-256-gcm .
In order to provide consistent data protection, encryption must be specified at
dataset creation time and it cannot be changed afterwards.
.Pp
For more details and caveats about encryption see the
.Sx Encryption
section of
.Xr zfs-load-key 8 .
.It Sy keyformat Ns = Ns Sy raw Ns | Ns Sy hex Ns | Ns Sy passphrase
Controls what format the user's encryption key will be provided as.
This property is only set when the dataset is encrypted.
.Pp
Raw keys and hex keys must be 32 bytes long (regardless of the chosen
encryption suite) and must be randomly generated.
A raw key can be generated with the following command:
.Dl # Nm dd Sy if=/dev/urandom bs=32 count=1 Sy of= Ns Pa /path/to/output/key
.Pp
Passphrases must be between 8 and 512 bytes long and will be processed through
PBKDF2 before being used (see the
.Sy pbkdf2iters
property).
Even though the encryption suite cannot be changed after dataset creation,
the keyformat can be with
.Nm zfs Cm change-key .
.It Xo
.Sy keylocation Ns = Ns Sy prompt Ns | Ns Sy file:// Ns Ar /absolute/file/path Ns | Ns Sy https:// Ns Ar address Ns | Ns Sy http:// Ns Ar address
.Xc
Controls where the user's encryption key will be loaded from by default for
commands such as
.Nm zfs Cm load-key
and
.Nm zfs Cm mount Fl l .
This property is only set for encrypted datasets which are encryption roots.
If unspecified, the default is
.Sy prompt .
.Pp
Even though the encryption suite cannot be changed after dataset creation, the
keylocation can be with either
.Nm zfs Cm set
or
.Nm zfs Cm change-key .
If
.Sy prompt
is selected ZFS will ask for the key at the command prompt when it is required
to access the encrypted data (see
.Nm zfs Cm load-key
for details).
This setting will also allow the key to be passed in via the standard input
stream,
but users should be careful not to place keys which should be kept secret on
the command line.
If a file URI is selected, the key will be loaded from the
specified absolute file path.
If an HTTPS or HTTP URL is selected, it will be GETted using
.Xr fetch 3 ,
libcurl, or nothing, depending on compile-time configuration and run-time
availability.
The
.Sy SSL_CA_CERT_FILE
environment variable can be set to set the location
of the concatenated certificate store.
The
.Sy SSL_CA_CERT_PATH
environment variable can be set to override the location
of the directory containing the certificate authority bundle.
The
.Sy SSL_CLIENT_CERT_FILE
and
.Sy SSL_CLIENT_KEY_FILE
environment variables can be set to configure the path
to the client certificate and its key.
.It Sy pbkdf2iters Ns = Ns Ar iterations
Controls the number of PBKDF2 iterations that a
.Sy passphrase
encryption key should be run through when processing it into an encryption key.
This property is only defined when encryption is enabled and a keyformat of
.Sy passphrase
is selected.
The goal of PBKDF2 is to significantly increase the
computational difficulty needed to brute force a user's passphrase.
This is accomplished by forcing the attacker to run each passphrase through a
computationally expensive hashing function many times before they arrive at the
resulting key.
A user who actually knows the passphrase will only have to pay this cost once.
As CPUs become better at processing, this number should be
raised to ensure that a brute force attack is still not possible.
The current default is
.Sy 350000
and the minimum is
.Sy 100000 .
This property may be changed with
.Nm zfs Cm change-key .
.It Sy exec Ns = Ns Sy on Ns | Ns Sy off
Controls whether processes can be executed from within this file system.
The default value is
.Sy on .
The values
.Sy on
and
.Sy off
are equivalent to the
.Sy exec
and
.Sy noexec
mount options.
.It Sy filesystem_limit Ns = Ns Ar count Ns | Ns Sy none
Limits the number of filesystems and volumes that can exist under this point in
the dataset tree.
The limit is not enforced if the user is allowed to change the limit.
Setting a
.Sy filesystem_limit
to
.Sy on
a descendent of a filesystem that already has a
.Sy filesystem_limit
does not override the ancestor's
.Sy filesystem_limit ,
but rather imposes an additional limit.
This feature must be enabled to be used
.Po see
.Xr zpool-features 7
.Pc .
.It Sy special_small_blocks Ns = Ns Ar size
This value represents the threshold block size for including small file
blocks into the special allocation class.
Blocks smaller than or equal to this
value will be assigned to the special allocation class while greater blocks
will be assigned to the regular class.
Valid values are zero or a power of two from 512 up to 1048576 (1 MiB).
The default size is 0 which means no small file blocks
will be allocated in the special class.
.Pp
Before setting this property, a special class vdev must be added to the
pool.
See
.Xr zpoolconcepts 7
for more details on the special allocation class.
.It Sy mountpoint Ns = Ns Pa path Ns | Ns Sy none Ns | Ns Sy legacy
Controls the mount point used for this file system.
See the
.Sx Mount Points
section of
.Xr zfsconcepts 7
for more information on how this property is used.
.Pp
When the
.Sy mountpoint
property is changed for a file system, the file system and any children that
inherit the mount point are unmounted.
If the new value is
.Sy legacy ,
then they remain unmounted.
Otherwise, they are automatically remounted in the new location if the property
was previously
.Sy legacy
or
.Sy none ,
or if they were mounted before the property was changed.
In addition, any shared file systems are unshared and shared in the new
location.
.It Sy nbmand Ns = Ns Sy on Ns | Ns Sy off
Controls whether the file system should be mounted with
.Sy nbmand
.Pq Non-blocking mandatory locks .
-This is used for SMB clients.
Changes to this property only take effect when the file system is umounted and
remounted.
-Support for these locks is scarce and not described by POSIX.
+This was only supported by Linux prior to 5.15, and was buggy there,
+and is not supported by
+.Fx .
+On Solaris it's used for SMB clients.
.It Sy overlay Ns = Ns Sy on Ns | Ns Sy off
Allow mounting on a busy directory or a directory which already contains
files or directories.
This is the default mount behavior for Linux and
.Fx
file systems.
On these platforms the property is
.Sy on
by default.
Set to
.Sy off
to disable overlay mounts for consistency with OpenZFS on other platforms.
.It Sy primarycache Ns = Ns Sy all Ns | Ns Sy none Ns | Ns Sy metadata
Controls what is cached in the primary cache
.Pq ARC .
If this property is set to
.Sy all ,
then both user data and metadata is cached.
If this property is set to
.Sy none ,
then neither user data nor metadata is cached.
If this property is set to
.Sy metadata ,
then only metadata is cached.
The default value is
.Sy all .
.It Sy quota Ns = Ns Ar size Ns | Ns Sy none
Limits the amount of space a dataset and its descendents can consume.
This property enforces a hard limit on the amount of space used.
This includes all space consumed by descendents, including file systems and
snapshots.
Setting a quota on a descendent of a dataset that already has a quota does not
override the ancestor's quota, but rather imposes an additional limit.
.Pp
Quotas cannot be set on volumes, as the
.Sy volsize
property acts as an implicit quota.
.It Sy snapshot_limit Ns = Ns Ar count Ns | Ns Sy none
Limits the number of snapshots that can be created on a dataset and its
descendents.
Setting a
.Sy snapshot_limit
on a descendent of a dataset that already has a
.Sy snapshot_limit
does not override the ancestor's
.Sy snapshot_limit ,
but rather imposes an additional limit.
The limit is not enforced if the user is allowed to change the limit.
For example, this means that recursive snapshots taken from the global zone are
counted against each delegated dataset within a zone.
This feature must be enabled to be used
.Po see
.Xr zpool-features 7
.Pc .
.It Sy userquota@ Ns Ar user Ns = Ns Ar size Ns | Ns Sy none
Limits the amount of space consumed by the specified user.
User space consumption is identified by the
.Sy userspace@ Ns Ar user
property.
.Pp
Enforcement of user quotas may be delayed by several seconds.
This delay means that a user might exceed their quota before the system notices
that they are over quota and begins to refuse additional writes with the
.Er EDQUOT
error message.
See the
.Nm zfs Cm userspace
command for more information.
.Pp
Unprivileged users can only access their own groups' space usage.
The root user, or a user who has been granted the
.Sy userquota
privilege with
.Nm zfs Cm allow ,
can get and set everyone's quota.
.Pp
This property is not available on volumes, on file systems before version 4, or
on pools before version 15.
The
.Sy userquota@ Ns Ar …
properties are not displayed by
.Nm zfs Cm get Sy all .
The user's name must be appended after the
.Sy @
symbol, using one of the following forms:
.Bl -bullet -compact -offset 4n
.It
POSIX name
.Pq Qq joe
.It
POSIX numeric ID
.Pq Qq 789
.It
SID name
.Pq Qq joe.smith@mydomain
.It
SID numeric ID
.Pq Qq S-1-123-456-789
.El
.Pp
Files created on Linux always have POSIX owners.
.It Sy userobjquota@ Ns Ar user Ns = Ns Ar size Ns | Ns Sy none
The
.Sy userobjquota
is similar to
.Sy userquota
but it limits the number of objects a user can create.
Please refer to
.Sy userobjused
for more information about how objects are counted.
.It Sy groupquota@ Ns Ar group Ns = Ns Ar size Ns | Ns Sy none
Limits the amount of space consumed by the specified group.
Group space consumption is identified by the
.Sy groupused@ Ns Ar group
property.
.Pp
Unprivileged users can access only their own groups' space usage.
The root user, or a user who has been granted the
.Sy groupquota
privilege with
.Nm zfs Cm allow ,
can get and set all groups' quotas.
.It Sy groupobjquota@ Ns Ar group Ns = Ns Ar size Ns | Ns Sy none
The
.Sy groupobjquota
is similar to
.Sy groupquota
but it limits number of objects a group can consume.
Please refer to
.Sy userobjused
for more information about how objects are counted.
.It Sy projectquota@ Ns Ar project Ns = Ns Ar size Ns | Ns Sy none
Limits the amount of space consumed by the specified project.
Project space consumption is identified by the
.Sy projectused@ Ns Ar project
property.
Please refer to
.Sy projectused
for more information about how project is identified and set/changed.
.Pp
The root user, or a user who has been granted the
.Sy projectquota
privilege with
.Nm zfs allow ,
can access all projects' quota.
.It Sy projectobjquota@ Ns Ar project Ns = Ns Ar size Ns | Ns Sy none
The
.Sy projectobjquota
is similar to
.Sy projectquota
but it limits number of objects a project can consume.
Please refer to
.Sy userobjused
for more information about how objects are counted.
.It Sy readonly Ns = Ns Sy on Ns | Ns Sy off
Controls whether this dataset can be modified.
The default value is
.Sy off .
The values
.Sy on
and
.Sy off
are equivalent to the
.Sy ro
and
.Sy rw
mount options.
.Pp
This property can also be referred to by its shortened column name,
.Sy rdonly .
.It Sy recordsize Ns = Ns Ar size
Specifies a suggested block size for files in the file system.
This property is designed solely for use with database workloads that access
files in fixed-size records.
ZFS automatically tunes block sizes according to internal algorithms optimized
for typical access patterns.
.Pp
For databases that create very large files but access them in small random
chunks, these algorithms may be suboptimal.
Specifying a
.Sy recordsize
greater than or equal to the record size of the database can result in
significant performance gains.
Use of this property for general purpose file systems is strongly discouraged,
and may adversely affect performance.
.Pp
The size specified must be a power of two greater than or equal to
.Ar 512 B
and less than or equal to
.Ar 128 KiB .
If the
.Sy large_blocks
feature is enabled on the pool, the size may be up to
.Ar 1 MiB .
See
.Xr zpool-features 7
for details on ZFS feature flags.
.Pp
Changing the file system's
.Sy recordsize
affects only files created afterward; existing files are unaffected.
.Pp
This property can also be referred to by its shortened column name,
.Sy recsize .
.It Sy redundant_metadata Ns = Ns Sy all Ns | Ns Sy most Ns | Ns Sy some Ns | Ns Sy none
Controls what types of metadata are stored redundantly.
ZFS stores an extra copy of metadata, so that if a single block is corrupted,
the amount of user data lost is limited.
This extra copy is in addition to any redundancy provided at the pool level
.Pq e.g. by mirroring or RAID-Z ,
and is in addition to an extra copy specified by the
.Sy copies
property
.Pq up to a total of 3 copies .
For example if the pool is mirrored,
.Sy copies Ns = Ns 2 ,
and
.Sy redundant_metadata Ns = Ns Sy most ,
then ZFS stores 6 copies of most metadata, and 4 copies of data and some
metadata.
.Pp
When set to
.Sy all ,
ZFS stores an extra copy of all metadata.
If a single on-disk block is corrupt, at worst a single block of user data
.Po which is
.Sy recordsize
bytes long
.Pc
can be lost.
.Pp
When set to
.Sy most ,
ZFS stores an extra copy of most types of metadata.
This can improve performance of random writes, because less metadata must be
written.
In practice, at worst about 1000 blocks
.Po of
.Sy recordsize
bytes each
.Pc
of user data can be lost if a single on-disk block is corrupt.
The exact behavior of which metadata blocks are stored redundantly may change in
future releases.
.Pp
When set to
.Sy some ,
ZFS stores an extra copy of only critical metadata.
This can improve file create performance since less metadata
needs to be written.
If a single on-disk block is corrupt, at worst a single user file can be lost.
.Pp
When set to
.Sy none ,
ZFS does not store any copies of metadata redundantly.
If a single on-disk block is corrupt, an entire dataset can be lost.
.Pp
The default value is
.Sy all .
.It Sy refquota Ns = Ns Ar size Ns | Ns Sy none
Limits the amount of space a dataset can consume.
This property enforces a hard limit on the amount of space used.
This hard limit does not include space used by descendents, including file
systems and snapshots.
.It Sy refreservation Ns = Ns Ar size Ns | Ns Sy none Ns | Ns Sy auto
The minimum amount of space guaranteed to a dataset, not including its
descendents.
When the amount of space used is below this value, the dataset is treated as if
it were taking up the amount of space specified by
.Sy refreservation .
The
.Sy refreservation
reservation is accounted for in the parent datasets' space used, and counts
against the parent datasets' quotas and reservations.
.Pp
If
.Sy refreservation
is set, a snapshot is only allowed if there is enough free pool space outside of
this reservation to accommodate the current number of
.Qq referenced
bytes in the dataset.
.Pp
If
.Sy refreservation
is set to
.Sy auto ,
a volume is thick provisioned
.Po or
.Qq not sparse
.Pc .
.Sy refreservation Ns = Ns Sy auto
is only supported on volumes.
See
.Sy volsize
in the
.Sx Native Properties
section for more information about sparse volumes.
.Pp
This property can also be referred to by its shortened column name,
.Sy refreserv .
.It Sy relatime Ns = Ns Sy on Ns | Ns Sy off
Controls the manner in which the access time is updated when
.Sy atime Ns = Ns Sy on
is set.
Turning this property on causes the access time to be updated relative
to the modify or change time.
Access time is only updated if the previous
access time was earlier than the current modify or change time or if the
existing access time hasn't been updated within the past 24 hours.
The default value is
.Sy on .
The values
.Sy on
and
.Sy off
are equivalent to the
.Sy relatime
and
.Sy norelatime
mount options.
.It Sy reservation Ns = Ns Ar size Ns | Ns Sy none
The minimum amount of space guaranteed to a dataset and its descendants.
When the amount of space used is below this value, the dataset is treated as if
it were taking up the amount of space specified by its reservation.
Reservations are accounted for in the parent datasets' space used, and count
against the parent datasets' quotas and reservations.
.Pp
This property can also be referred to by its shortened column name,
.Sy reserv .
.It Sy secondarycache Ns = Ns Sy all Ns | Ns Sy none Ns | Ns Sy metadata
Controls what is cached in the secondary cache
.Pq L2ARC .
If this property is set to
.Sy all ,
then both user data and metadata is cached.
If this property is set to
.Sy none ,
then neither user data nor metadata is cached.
If this property is set to
.Sy metadata ,
then only metadata is cached.
The default value is
.Sy all .
.It Sy setuid Ns = Ns Sy on Ns | Ns Sy off
Controls whether the setuid bit is respected for the file system.
The default value is
.Sy on .
The values
.Sy on
and
.Sy off
are equivalent to the
.Sy suid
and
.Sy nosuid
mount options.
.It Sy sharesmb Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Ar opts
Controls whether the file system is shared by using
.Sy Samba USERSHARES
and what options are to be used.
Otherwise, the file system is automatically shared and unshared with the
.Nm zfs Cm share
and
.Nm zfs Cm unshare
commands.
If the property is set to on, the
.Xr net 8
command is invoked to create a
.Sy USERSHARE .
.Pp
Because SMB shares requires a resource name, a unique resource name is
constructed from the dataset name.
The constructed name is a copy of the
dataset name except that the characters in the dataset name, which would be
invalid in the resource name, are replaced with underscore (_) characters.
Linux does not currently support additional options which might be available
on Solaris.
.Pp
If the
.Sy sharesmb
property is set to
.Sy off ,
the file systems are unshared.
.Pp
The share is created with the ACL (Access Control List) "Everyone:F" ("F"
stands for "full permissions", i.e. read and write permissions) and no guest
access (which means Samba must be able to authenticate a real user \(em
.Xr passwd 5 Ns / Ns Xr shadow 5 Ns - ,
LDAP- or
.Xr smbpasswd 5 Ns -based )
by default.
This means that any additional access control
(disallow specific user specific access etc) must be done on the underlying file
system.
.It Sy sharenfs Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Ar opts
Controls whether the file system is shared via NFS, and what options are to be
used.
A file system with a
.Sy sharenfs
property of
.Sy off
is managed with the
.Xr exportfs 8
command and entries in the
.Pa /etc/exports
file.
Otherwise, the file system is automatically shared and unshared with the
.Nm zfs Cm share
and
.Nm zfs Cm unshare
commands.
If the property is set to
.Sy on ,
the dataset is shared using the default options:
.Dl sec=sys,rw,crossmnt,no_subtree_check
.Pp
Please note that the options are comma-separated, unlike those found in
.Xr exports 5 .
This is done to negate the need for quoting, as well as to make parsing
with scripts easier.
.Pp
See
.Xr exports 5
for the meaning of the default options.
Otherwise, the
.Xr exportfs 8
command is invoked with options equivalent to the contents of this property.
.Pp
When the
.Sy sharenfs
property is changed for a dataset, the dataset and any children inheriting the
property are re-shared with the new options, only if the property was previously
.Sy off ,
or if they were shared before the property was changed.
If the new property is
.Sy off ,
the file systems are unshared.
.It Sy logbias Ns = Ns Sy latency Ns | Ns Sy throughput
Provide a hint to ZFS about handling of synchronous requests in this dataset.
If
.Sy logbias
is set to
.Sy latency
.Pq the default ,
ZFS will use pool log devices
.Pq if configured
to handle the requests at low latency.
If
.Sy logbias
is set to
.Sy throughput ,
ZFS will not use configured pool log devices.
ZFS will instead optimize synchronous operations for global pool throughput and
efficient use of resources.
.It Sy snapdev Ns = Ns Sy hidden Ns | Ns Sy visible
Controls whether the volume snapshot devices under
.Pa /dev/zvol/ Ns Aq Ar pool
are hidden or visible.
The default value is
.Sy hidden .
.It Sy snapdir Ns = Ns Sy hidden Ns | Ns Sy visible
Controls whether the
.Pa .zfs
directory is hidden or visible in the root of the file system as discussed in
the
.Sx Snapshots
section of
.Xr zfsconcepts 7 .
The default value is
.Sy hidden .
.It Sy sync Ns = Ns Sy standard Ns | Ns Sy always Ns | Ns Sy disabled
Controls the behavior of synchronous requests
.Pq e.g. fsync, O_DSYNC .
.Sy standard
is the POSIX-specified behavior of ensuring all synchronous requests
are written to stable storage and all devices are flushed to ensure
data is not cached by device controllers
.Pq this is the default .
.Sy always
causes every file system transaction to be written and flushed before its
system call returns.
This has a large performance penalty.
.Sy disabled
disables synchronous requests.
File system transactions are only committed to stable storage periodically.
This option will give the highest performance.
However, it is very dangerous as ZFS would be ignoring the synchronous
transaction demands of applications such as databases or NFS.
Administrators should only use this option when the risks are understood.
.It Sy version Ns = Ns Ar N Ns | Ns Sy current
The on-disk version of this file system, which is independent of the pool
version.
This property can only be set to later supported versions.
See the
.Nm zfs Cm upgrade
command.
.It Sy volsize Ns = Ns Ar size
For volumes, specifies the logical size of the volume.
By default, creating a volume establishes a reservation of equal size.
For storage pools with a version number of 9 or higher, a
.Sy refreservation
is set instead.
Any changes to
.Sy volsize
are reflected in an equivalent change to the reservation
.Pq or Sy refreservation .
The
.Sy volsize
can only be set to a multiple of
.Sy volblocksize ,
and cannot be zero.
.Pp
The reservation is kept equal to the volume's logical size to prevent unexpected
behavior for consumers.
Without the reservation, the volume could run out of space, resulting in
undefined behavior or data corruption, depending on how the volume is used.
These effects can also occur when the volume size is changed while it is in use
.Pq particularly when shrinking the size .
Extreme care should be used when adjusting the volume size.
.Pp
Though not recommended, a
.Qq sparse volume
.Po also known as
.Qq thin provisioned
.Pc
can be created by specifying the
.Fl s
option to the
.Nm zfs Cm create Fl V
command, or by changing the value of the
.Sy refreservation
property
.Po or
.Sy reservation
property on pool version 8 or earlier
.Pc
after the volume has been created.
A
.Qq sparse volume
is a volume where the value of
.Sy refreservation
is less than the size of the volume plus the space required to store its
metadata.
Consequently, writes to a sparse volume can fail with
.Er ENOSPC
when the pool is low on space.
For a sparse volume, changes to
.Sy volsize
are not reflected in the
.Sy refreservation .
A volume that is not sparse is said to be
.Qq thick provisioned .
A sparse volume can become thick provisioned by setting
.Sy refreservation
to
.Sy auto .
.It Sy volmode Ns = Ns Sy default Ns | Ns Sy full Ns | Ns Sy geom Ns | Ns Sy dev Ns | Ns Sy none
This property specifies how volumes should be exposed to the OS.
Setting it to
.Sy full
exposes volumes as fully fledged block devices, providing maximal
functionality.
The value
.Sy geom
is just an alias for
.Sy full
and is kept for compatibility.
Setting it to
.Sy dev
hides its partitions.
Volumes with property set to
.Sy none
are not exposed outside ZFS, but can be snapshotted, cloned, replicated, etc,
that can be suitable for backup purposes.
Value
.Sy default
means that volumes exposition is controlled by system-wide tunable
.Sy zvol_volmode ,
where
.Sy full ,
.Sy dev
and
.Sy none
are encoded as 1, 2 and 3 respectively.
The default value is
.Sy full .
.It Sy vscan Ns = Ns Sy on Ns | Ns Sy off
Controls whether regular files should be scanned for viruses when a file is
opened and closed.
In addition to enabling this property, the virus scan service must also be
enabled for virus scanning to occur.
The default value is
.Sy off .
This property is not used by OpenZFS.
.It Sy xattr Ns = Ns Sy on Ns | Ns Sy off Ns | Ns Sy sa
Controls whether extended attributes are enabled for this file system.
Two styles of extended attributes are supported: either directory-based
or system-attribute-based.
.Pp
The default value of
.Sy on
enables directory-based extended attributes.
This style of extended attribute imposes no practical limit
on either the size or number of attributes which can be set on a file.
Although under Linux the
.Xr getxattr 2
and
.Xr setxattr 2
system calls limit the maximum size to
.Sy 64K .
This is the most compatible
style of extended attribute and is supported by all ZFS implementations.
.Pp
System-attribute-based xattrs can be enabled by setting the value to
.Sy sa .
The key advantage of this type of xattr is improved performance.
Storing extended attributes as system attributes
significantly decreases the amount of disk I/O required.
Up to
.Sy 64K
of data may be stored per-file in the space reserved for system attributes.
If there is not enough space available for an extended attribute
then it will be automatically written as a directory-based xattr.
System-attribute-based extended attributes are not accessible
on platforms which do not support the
.Sy xattr Ns = Ns Sy sa
feature.
OpenZFS supports
.Sy xattr Ns = Ns Sy sa
on both
.Fx
and Linux.
.Pp
The use of system-attribute-based xattrs is strongly encouraged for users of
SELinux or POSIX ACLs.
Both of these features heavily rely on extended
attributes and benefit significantly from the reduced access time.
.Pp
The values
.Sy on
and
.Sy off
are equivalent to the
.Sy xattr
and
.Sy noxattr
mount options.
.It Sy jailed Ns = Ns Sy off Ns | Ns Sy on
Controls whether the dataset is managed from a jail.
See
.Xr zfs-jail 8
for more information.
Jails are a
.Fx
feature and are not relevant on other platforms.
The default value is
.Sy off .
.It Sy zoned Ns = Ns Sy on Ns | Ns Sy off
Controls whether the dataset is managed from a non-global zone or namespace.
The default value is
.Sy off .
.El
.Pp
The following three properties cannot be changed after the file system is
created, and therefore, should be set when the file system is created.
If the properties are not set with the
.Nm zfs Cm create
or
.Nm zpool Cm create
commands, these properties are inherited from the parent dataset.
If the parent dataset lacks these properties due to having been created prior to
these features being supported, the new file system will have the default values
for these properties.
.Bl -tag -width ""
.It Xo
.Sy casesensitivity Ns = Ns Sy sensitive Ns | Ns
.Sy insensitive Ns | Ns Sy mixed
.Xc
Indicates whether the file name matching algorithm used by the file system
should be case-sensitive, case-insensitive, or allow a combination of both
styles of matching.
The default value for the
.Sy casesensitivity
property is
.Sy sensitive .
Traditionally,
.Ux
and POSIX file systems have case-sensitive file names.
.Pp
The
.Sy mixed
value for the
.Sy casesensitivity
property indicates that the file system can support requests for both
case-sensitive and case-insensitive matching behavior.
Currently, case-insensitive matching behavior on a file system that supports
mixed behavior is limited to the SMB server product.
For more information about the
.Sy mixed
value behavior, see the "ZFS Administration Guide".
.It Xo
.Sy normalization Ns = Ns Sy none Ns | Ns Sy formC Ns | Ns
.Sy formD Ns | Ns Sy formKC Ns | Ns Sy formKD
.Xc
Indicates whether the file system should perform a
.Sy unicode
normalization of file names whenever two file names are compared, and which
normalization algorithm should be used.
File names are always stored unmodified, names are normalized as part of any
comparison process.
If this property is set to a legal value other than
.Sy none ,
and the
.Sy utf8only
property was left unspecified, the
.Sy utf8only
property is automatically set to
.Sy on .
The default value of the
.Sy normalization
property is
.Sy none .
This property cannot be changed after the file system is created.
.It Sy utf8only Ns = Ns Sy on Ns | Ns Sy off
Indicates whether the file system should reject file names that include
characters that are not present in the
.Sy UTF-8
character code set.
If this property is explicitly set to
.Sy off ,
the normalization property must either not be explicitly set or be set to
.Sy none .
The default value for the
.Sy utf8only
property is
.Sy off .
This property cannot be changed after the file system is created.
.El
.Pp
The
.Sy casesensitivity ,
.Sy normalization ,
and
.Sy utf8only
properties are also new permissions that can be assigned to non-privileged users
by using the ZFS delegated administration feature.
.
.Ss Temporary Mount Point Properties
When a file system is mounted, either through
.Xr mount 8
for legacy mounts or the
.Nm zfs Cm mount
command for normal file systems, its mount options are set according to its
properties.
The correlation between properties and mount options is as follows:
.Bl -tag -compact -offset Ds -width "rootcontext="
.It Sy atime
atime/noatime
.It Sy canmount
auto/noauto
.It Sy devices
dev/nodev
.It Sy exec
exec/noexec
.It Sy readonly
ro/rw
.It Sy relatime
relatime/norelatime
.It Sy setuid
suid/nosuid
.It Sy xattr
xattr/noxattr
.It Sy nbmand
mand/nomand
.It Sy context Ns =
context=
.It Sy fscontext Ns =
fscontext=
.It Sy defcontext Ns =
defcontext=
.It Sy rootcontext Ns =
rootcontext=
.El
.Pp
In addition, these options can be set on a per-mount basis using the
.Fl o
option, without affecting the property that is stored on disk.
The values specified on the command line override the values stored in the
dataset.
The
.Sy nosuid
option is an alias for
.Sy nodevices , Ns Sy nosetuid .
These properties are reported as
.Qq temporary
by the
.Nm zfs Cm get
command.
If the properties are changed while the dataset is mounted, the new setting
overrides any temporary settings.
.
.Ss User Properties
In addition to the standard native properties, ZFS supports arbitrary user
properties.
User properties have no effect on ZFS behavior, but applications or
administrators can use them to annotate datasets
.Pq file systems, volumes, and snapshots .
.Pp
User property names must contain a colon
.Pq Qq Sy \&:
character to distinguish them from native properties.
They may contain lowercase letters, numbers, and the following punctuation
characters: colon
.Pq Qq Sy \&: ,
dash
.Pq Qq Sy - ,
period
.Pq Qq Sy \&. ,
and underscore
.Pq Qq Sy _ .
The expected convention is that the property name is divided into two portions
such as
.Ar module : Ns Ar property ,
but this namespace is not enforced by ZFS.
User property names can be at most 256 characters, and cannot begin with a dash
.Pq Qq Sy - .
.Pp
When making programmatic use of user properties, it is strongly suggested to use
a reversed DNS domain name for the
.Ar module
component of property names to reduce the chance that two
independently-developed packages use the same property name for different
purposes.
.Pp
The values of user properties are arbitrary strings, are always inherited, and
are never validated.
All of the commands that operate on properties
.Po Nm zfs Cm list ,
.Nm zfs Cm get ,
.Nm zfs Cm set ,
and so forth
.Pc
can be used to manipulate both native properties and user properties.
Use the
.Nm zfs Cm inherit
command to clear a user property.
If the property is not defined in any parent dataset, it is removed entirely.
Property values are limited to 8192 bytes.
diff --git a/sys/contrib/openzfs/man/man7/zpool-features.7 b/sys/contrib/openzfs/man/man7/zpool-features.7
index 4cd7526858a3..efe9e833996a 100644
--- a/sys/contrib/openzfs/man/man7/zpool-features.7
+++ b/sys/contrib/openzfs/man/man7/zpool-features.7
@@ -1,937 +1,953 @@
.\"
.\" Copyright (c) 2012, 2018 by Delphix. All rights reserved.
.\" Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
.\" Copyright (c) 2014, Joyent, Inc. All rights reserved.
.\" The contents of this file are subject to the terms of the Common Development
.\" and Distribution License (the "License"). You may not use this file except
.\" in compliance with the License. You can obtain a copy of the license at
.\" usr/src/OPENSOLARIS.LICENSE or https://opensource.org/licenses/CDDL-1.0.
.\"
.\" See the License for the specific language governing permissions and
.\" limitations under the License. When distributing Covered Code, include this
.\" CDDL HEADER in each file and include the License file at
.\" usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this
.\" CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your
.\" own identifying information:
.\" Portions Copyright [yyyy] [name of copyright owner]
.\" Copyright (c) 2019, Klara Inc.
.\" Copyright (c) 2019, Allan Jude
.\" Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
.\"
.Dd June 23, 2022
.Dt ZPOOL-FEATURES 7
.Os
.
.Sh NAME
.Nm zpool-features
.Nd description of ZFS pool features
.
.Sh DESCRIPTION
ZFS pool on-disk format versions are specified via
.Dq features
which replace the old on-disk format numbers
.Pq the last supported on-disk format number is 28 .
To enable a feature on a pool use the
.Nm zpool Cm upgrade ,
or set the
.Sy feature Ns @ Ns Ar feature-name
property to
.Sy enabled .
Please also see the
.Sx Compatibility feature sets
section for information on how sets of features may be enabled together.
.Pp
The pool format does not affect file system version compatibility or the ability
to send file systems between pools.
.Pp
Since most features can be enabled independently of each other, the on-disk
format of the pool is specified by the set of all features marked as
.Sy active
on the pool.
If the pool was created by another software version
this set may include unsupported features.
.
.Ss Identifying features
Every feature has a GUID of the form
.Ar com.example : Ns Ar feature-name .
The reversed DNS name ensures that the feature's GUID is unique across all ZFS
implementations.
When unsupported features are encountered on a pool they will
be identified by their GUIDs.
Refer to the documentation for the ZFS
implementation that created the pool for information about those features.
.Pp
Each supported feature also has a short name.
By convention a feature's short name is the portion of its GUID which follows
the
.Sq \&:
.Po
i.e.
.Ar com.example : Ns Ar feature-name
would have the short name
.Ar feature-name
.Pc ,
however a feature's short name may differ across ZFS implementations if
following the convention would result in name conflicts.
.
.Ss Feature states
Features can be in one of three states:
.Bl -tag -width "disabled"
.It Sy active
This feature's on-disk format changes are in effect on the pool.
Support for this feature is required to import the pool in read-write mode.
If this feature is not read-only compatible,
support is also required to import the pool in read-only mode
.Pq see Sx Read-only compatibility .
.It Sy enabled
An administrator has marked this feature as enabled on the pool, but the
feature's on-disk format changes have not been made yet.
The pool can still be imported by software that does not support this feature,
but changes may be made to the on-disk format at any time
which will move the feature to the
.Sy active
state.
Some features may support returning to the
.Sy enabled
state after becoming
.Sy active .
See feature-specific documentation for details.
.It Sy disabled
This feature's on-disk format changes have not been made and will not be made
unless an administrator moves the feature to the
.Sy enabled
state.
Features cannot be disabled once they have been enabled.
.El
.Pp
The state of supported features is exposed through pool properties of the form
.Sy feature Ns @ Ns Ar short-name .
.
.Ss Read-only compatibility
Some features may make on-disk format changes that do not interfere with other
software's ability to read from the pool.
These features are referred to as
.Dq read-only compatible .
If all unsupported features on a pool are read-only compatible,
the pool can be imported in read-only mode by setting the
.Sy readonly
property during import
.Po see
.Xr zpool-import 8
for details on importing pools
.Pc .
.
.Ss Unsupported features
For each unsupported feature enabled on an imported pool, a pool property
named
.Sy unsupported Ns @ Ns Ar feature-name
will indicate why the import was allowed despite the unsupported feature.
Possible values for this property are:
.Bl -tag -width "readonly"
.It Sy inactive
The feature is in the
.Sy enabled
state and therefore the pool's on-disk
format is still compatible with software that does not support this feature.
.It Sy readonly
The feature is read-only compatible and the pool has been imported in
read-only mode.
.El
.
.Ss Feature dependencies
Some features depend on other features being enabled in order to function.
Enabling a feature will automatically enable any features it depends on.
.
.Ss Compatibility feature sets
It is sometimes necessary for a pool to maintain compatibility with a
specific on-disk format, by enabling and disabling particular features.
The
.Sy compatibility
feature facilitates this by allowing feature sets to be read from text files.
When set to
.Sy off
.Pq the default ,
compatibility feature sets are disabled
.Pq i.e. all features are enabled ;
when set to
.Sy legacy ,
no features are enabled.
When set to a comma-separated list of filenames
.Po
each filename may either be an absolute path, or relative to
.Pa /etc/zfs/compatibility.d
or
.Pa /usr/share/zfs/compatibility.d
.Pc ,
the lists of requested features are read from those files,
separated by whitespace and/or commas.
Only features present in all files are enabled.
.Pp
Simple sanity checks are applied to the files:
they must be between 1 B and 16 KiB in size, and must end with a newline
character.
.Pp
The requested features are applied when a pool is created using
.Nm zpool Cm create Fl o Sy compatibility Ns = Ns Ar …
and controls which features are enabled when using
.Nm zpool Cm upgrade .
.Nm zpool Cm status
will not show a warning about disabled features which are not part
of the requested feature set.
.Pp
The special value
.Sy legacy
prevents any features from being enabled, either via
.Nm zpool Cm upgrade
or
.Nm zpool Cm set Sy feature Ns @ Ns Ar feature-name Ns = Ns Sy enabled .
This setting also prevents pools from being upgraded to newer on-disk versions.
This is a safety measure to prevent new features from being
accidentally enabled, breaking compatibility.
.Pp
By convention, compatibility files in
.Pa /usr/share/zfs/compatibility.d
are provided by the distribution, and include feature sets
supported by important versions of popular distributions, and feature
sets commonly supported at the start of each year.
Compatibility files in
.Pa /etc/zfs/compatibility.d ,
if present, will take precedence over files with the same name in
.Pa /usr/share/zfs/compatibility.d .
.Pp
If an unrecognized feature is found in these files, an error message will
be shown.
If the unrecognized feature is in a file in
.Pa /etc/zfs/compatibility.d ,
this is treated as an error and processing will stop.
If the unrecognized feature is under
.Pa /usr/share/zfs/compatibility.d ,
this is treated as a warning and processing will continue.
This difference is to allow distributions to include features
which might not be recognized by the currently-installed binaries.
.Pp
Compatibility files may include comments:
any text from
.Sq #
to the end of the line is ignored.
.Pp
.Sy Example :
.Bd -literal -compact -offset 4n
.No example# Nm cat Pa /usr/share/zfs/compatibility.d/grub2
# Features which are supported by GRUB2
async_destroy
bookmarks
embedded_data
empty_bpobj
enabled_txg
extensible_dataset
filesystem_limits
hole_birth
large_blocks
lz4_compress
spacemap_histogram
.No example# Nm zpool Cm create Fl o Sy compatibility Ns = Ns Ar grub2 Ar bootpool Ar vdev
.Ed
.Pp
See
.Xr zpool-create 8
and
.Xr zpool-upgrade 8
for more information on how these commands are affected by feature sets.
.
.de feature
.It Sy \\$2
.Bl -tag -compact -width "READ-ONLY COMPATIBLE"
.It GUID
.Sy \\$1:\\$2
.if !"\\$4"" \{\
.It DEPENDENCIES
\fB\\$4\fP\c
.if !"\\$5"" , \fB\\$5\fP\c
.if !"\\$6"" , \fB\\$6\fP\c
.if !"\\$7"" , \fB\\$7\fP\c
.if !"\\$8"" , \fB\\$8\fP\c
.if !"\\$9"" , \fB\\$9\fP\c
.\}
.It READ-ONLY COMPATIBLE
\\$3
.El
.Pp
..
.
.ds instant-never \
.No This feature becomes Sy active No as soon as it is enabled \
and will never return to being Sy enabled .
.
.ds remount-upgrade \
.No Each filesystem will be upgraded automatically when remounted, \
or when a new file is created under that filesystem. \
The upgrade can also be triggered on filesystems via \
Nm zfs Cm set Sy version Ns = Ns Sy current Ar fs . \
No The upgrade process runs in the background and may take a while to complete \
for filesystems containing large amounts of files .
.
.de checksum-spiel
When the
.Sy \\$1
feature is set to
.Sy enabled ,
the administrator can turn on the
.Sy \\$1
checksum on any dataset using
.Nm zfs Cm set Sy checksum Ns = Ns Sy \\$1 Ar dset
.Po see Xr zfs-set 8 Pc .
This feature becomes
.Sy active
once a
.Sy checksum
property has been set to
.Sy \\$1 ,
and will return to being
.Sy enabled
once all filesystems that have ever had their checksum set to
.Sy \\$1
are destroyed.
..
.
.Sh FEATURES
The following features are supported on this system:
.Bl -tag -width Ds
.feature org.zfsonlinux allocation_classes yes
This feature enables support for separate allocation classes.
.Pp
This feature becomes
.Sy active
when a dedicated allocation class vdev
.Pq dedup or special
is created with the
.Nm zpool Cm create No or Nm zpool Cm add No commands .
With device removal, it can be returned to the
.Sy enabled
state if all the dedicated allocation class vdevs are removed.
.
.feature com.delphix async_destroy yes
Destroying a file system requires traversing all of its data in order to
return its used space to the pool.
Without
.Sy async_destroy ,
the file system is not fully removed until all space has been reclaimed.
If the destroy operation is interrupted by a reboot or power outage,
the next attempt to open the pool will need to complete the destroy
operation synchronously.
.Pp
When
.Sy async_destroy
is enabled, the file system's data will be reclaimed by a background process,
allowing the destroy operation to complete
without traversing the entire file system.
The background process is able to resume
interrupted destroys after the pool has been opened, eliminating the need
to finish interrupted destroys as part of the open operation.
The amount of space remaining to be reclaimed by the background process
is available through the
.Sy freeing
property.
.Pp
This feature is only
.Sy active
while
.Sy freeing
is non-zero.
.
.feature org.openzfs blake3 no extensible_dataset
This feature enables the use of the BLAKE3 hash algorithm for checksum and
dedup.
BLAKE3 is a secure hash algorithm focused on high performance.
.Pp
.checksum-spiel blake3
.
.feature com.fudosecurity block_cloning yes
When this feature is enabled ZFS will use block cloning for operations like
.Fn copy_file_range 2 .
Block cloning allows to create multiple references to a single block.
It is much faster than copying the data (as the actual data is neither read nor
written) and takes no additional space.
Blocks can be cloned across datasets under some conditions (like disabled
encryption and equal
.Nm recordsize ) .
.Pp
This feature becomes
.Sy active
when first block is cloned.
When the last cloned block is freed, it goes back to the enabled state.
.feature com.delphix bookmarks yes extensible_dataset
This feature enables use of the
.Nm zfs Cm bookmark
command.
.Pp
This feature is
.Sy active
while any bookmarks exist in the pool.
All bookmarks in the pool can be listed by running
.Nm zfs Cm list Fl t Sy bookmark Fl r Ar poolname .
.
.feature com.datto bookmark_v2 no bookmark extensible_dataset
This feature enables the creation and management of larger bookmarks which are
needed for other features in ZFS.
.Pp
This feature becomes
.Sy active
when a v2 bookmark is created and will be returned to the
.Sy enabled
state when all v2 bookmarks are destroyed.
.
.feature com.delphix bookmark_written no bookmark extensible_dataset bookmark_v2
This feature enables additional bookmark accounting fields, enabling the
.Sy written Ns # Ns Ar bookmark
property
.Pq space written since a bookmark
and estimates of send stream sizes for incrementals from bookmarks.
.Pp
This feature becomes
.Sy active
when a bookmark is created and will be
returned to the
.Sy enabled
state when all bookmarks with these fields are destroyed.
.
.feature org.openzfs device_rebuild yes
This feature enables the ability for the
.Nm zpool Cm attach
and
.Nm zpool Cm replace
commands to perform sequential reconstruction
.Pq instead of healing reconstruction
when resilvering.
.Pp
Sequential reconstruction resilvers a device in LBA order without immediately
verifying the checksums.
Once complete, a scrub is started, which then verifies the checksums.
This approach allows full redundancy to be restored to the pool
in the minimum amount of time.
This two-phase approach will take longer than a healing resilver
when the time to verify the checksums is included.
However, unless there is additional pool damage,
no checksum errors should be reported by the scrub.
This feature is incompatible with raidz configurations.
.
This feature becomes
.Sy active
while a sequential resilver is in progress, and returns to
.Sy enabled
when the resilver completes.
.
.feature com.delphix device_removal no
This feature enables the
.Nm zpool Cm remove
command to remove top-level vdevs,
evacuating them to reduce the total size of the pool.
.Pp
This feature becomes
.Sy active
when the
.Nm zpool Cm remove
command is used
on a top-level vdev, and will never return to being
.Sy enabled .
.
.feature org.openzfs draid no
This feature enables use of the
.Sy draid
vdev type.
dRAID is a variant of RAID-Z which provides integrated distributed
hot spares that allow faster resilvering while retaining the benefits of RAID-Z.
Data, parity, and spare space are organized in redundancy groups
and distributed evenly over all of the devices.
.Pp
This feature becomes
.Sy active
when creating a pool which uses the
.Sy draid
vdev type, or when adding a new
.Sy draid
vdev to an existing pool.
.
.feature org.illumos edonr no extensible_dataset
This feature enables the use of the Edon-R hash algorithm for checksum,
including for nopwrite
.Po if compression is also enabled, an overwrite of
a block whose checksum matches the data being written will be ignored
.Pc .
In an abundance of caution, Edon-R requires verification when used with
dedup:
.Nm zfs Cm set Sy dedup Ns = Ns Sy edonr , Ns Sy verify
.Po see Xr zfs-set 8 Pc .
.Pp
Edon-R is a very high-performance hash algorithm that was part
of the NIST SHA-3 competition.
It provides extremely high hash performance
.Pq over 350% faster than SHA-256 ,
but was not selected because of its unsuitability
as a general purpose secure hash algorithm.
This implementation utilizes the new salted checksumming functionality
in ZFS, which means that the checksum is pre-seeded with a secret
256-bit random key
.Pq stored on the pool
before being fed the data block to be checksummed.
Thus the produced checksums are unique to a given pool,
preventing hash collision attacks on systems with dedup.
.Pp
.checksum-spiel edonr
.
.feature com.delphix embedded_data no
This feature improves the performance and compression ratio of
highly-compressible blocks.
Blocks whose contents can compress to 112 bytes
or smaller can take advantage of this feature.
.Pp
When this feature is enabled, the contents of highly-compressible blocks are
stored in the block
.Dq pointer
itself
.Po a misnomer in this case, as it contains
the compressed data, rather than a pointer to its location on disk
.Pc .
Thus the space of the block
.Pq one sector, typically 512 B or 4 KiB
is saved, and no additional I/O is needed to read and write the data block.
.
\*[instant-never]
.
.feature com.delphix empty_bpobj yes
This feature increases the performance of creating and using a large
number of snapshots of a single filesystem or volume, and also reduces
the disk space required.
.Pp
When there are many snapshots, each snapshot uses many Block Pointer
Objects
.Pq bpobjs
to track blocks associated with that snapshot.
However, in common use cases, most of these bpobjs are empty.
This feature allows us to create each bpobj on-demand,
thus eliminating the empty bpobjs.
.Pp
This feature is
.Sy active
while there are any filesystems, volumes,
or snapshots which were created after enabling this feature.
.
.feature com.delphix enabled_txg yes
Once this feature is enabled, ZFS records the transaction group number
in which new features are enabled.
This has no user-visible impact, but other features may depend on this feature.
.Pp
This feature becomes
.Sy active
as soon as it is enabled and will never return to being
.Sy enabled .
.
.feature com.datto encryption no bookmark_v2 extensible_dataset
This feature enables the creation and management of natively encrypted datasets.
.Pp
This feature becomes
.Sy active
when an encrypted dataset is created and will be returned to the
.Sy enabled
state when all datasets that use this feature are destroyed.
.
.feature com.delphix extensible_dataset no
This feature allows more flexible use of internal ZFS data structures,
and exists for other features to depend on.
.Pp
This feature will be
.Sy active
when the first dependent feature uses it, and will be returned to the
.Sy enabled
state when all datasets that use this feature are destroyed.
.
.feature com.joyent filesystem_limits yes extensible_dataset
This feature enables filesystem and snapshot limits.
These limits can be used to control how many filesystems and/or snapshots
can be created at the point in the tree on which the limits are set.
.Pp
This feature is
.Sy active
once either of the limit properties has been set on a dataset
and will never return to being
.Sy enabled .
.
.feature com.delphix head_errlog no
This feature enables the upgraded version of errlog, which required an on-disk
error log format change.
Now the error log of each head dataset is stored separately in the zap object
and keyed by the head id.
In case of encrypted filesystems with unloaded keys or unmounted encrypted
filesystems we are unable to check their snapshots or clones for errors and
these will not be reported.
In this case no filenames will be reported either.
With this feature enabled, every dataset affected by an error block is listed
in the output of
.Nm zpool Cm status .
.Pp
\*[instant-never]
.
.feature com.delphix hole_birth no enabled_txg
This feature has/had bugs, the result of which is that, if you do a
.Nm zfs Cm send Fl i
.Pq or Fl R , No since it uses Fl i
from an affected dataset, the receiving party will not see any checksum
or other errors, but the resulting destination snapshot
will not match the source.
Its use by
.Nm zfs Cm send Fl i
has been disabled by default
.Po
see
.Sy send_holes_without_birth_time
in
.Xr zfs 4
.Pc .
.Pp
This feature improves performance of incremental sends
.Pq Nm zfs Cm send Fl i
and receives for objects with many holes.
The most common case of hole-filled objects is zvols.
.Pp
An incremental send stream from snapshot
.Sy A No to snapshot Sy B
contains information about every block that changed between
.Sy A No and Sy B .
Blocks which did not change between those snapshots can be
identified and omitted from the stream using a piece of metadata called
the
.Dq block birth time ,
but birth times are not recorded for holes
.Pq blocks filled only with zeroes .
Since holes created after
.Sy A No cannot be distinguished from holes created before Sy A ,
information about every hole in the entire filesystem or zvol
is included in the send stream.
.Pp
For workloads where holes are rare this is not a problem.
However, when incrementally replicating filesystems or zvols with many holes
.Pq for example a zvol formatted with another filesystem
a lot of time will be spent sending and receiving unnecessary information
about holes that already exist on the receiving side.
.Pp
Once the
.Sy hole_birth
feature has been enabled the block birth times
of all new holes will be recorded.
Incremental sends between snapshots created after this feature is enabled
will use this new metadata to avoid sending information about holes that
already exist on the receiving side.
.Pp
\*[instant-never]
.
.feature org.open-zfs large_blocks no extensible_dataset
This feature allows the record size on a dataset to be set larger than 128 KiB.
.Pp
This feature becomes
.Sy active
once a dataset contains a file with a block size larger than 128 KiB,
and will return to being
.Sy enabled
once all filesystems that have ever had their recordsize larger than 128 KiB
are destroyed.
.
.feature org.zfsonlinux large_dnode no extensible_dataset
This feature allows the size of dnodes in a dataset to be set larger than 512 B.
.
This feature becomes
.Sy active
once a dataset contains an object with a dnode larger than 512 B,
which occurs as a result of setting the
.Sy dnodesize
dataset property to a value other than
.Sy legacy .
The feature will return to being
.Sy enabled
once all filesystems that have ever contained a dnode larger than 512 B
are destroyed.
Large dnodes allow more data to be stored in the bonus buffer,
thus potentially improving performance by avoiding the use of spill blocks.
.
.feature com.delphix livelist yes
This feature allows clones to be deleted faster than the traditional method
when a large number of random/sparse writes have been made to the clone.
All blocks allocated and freed after a clone is created are tracked by the
the clone's livelist which is referenced during the deletion of the clone.
The feature is activated when a clone is created and remains
.Sy active
until all clones have been destroyed.
.
.feature com.delphix log_spacemap yes com.delphix:spacemap_v2
This feature improves performance for heavily-fragmented pools,
especially when workloads are heavy in random-writes.
It does so by logging all the metaslab changes on a single spacemap every TXG
instead of scattering multiple writes to all the metaslab spacemaps.
.Pp
\*[instant-never]
.
.feature org.illumos lz4_compress no
.Sy lz4
is a high-performance real-time compression algorithm that
features significantly faster compression and decompression as well as a
higher compression ratio than the older
.Sy lzjb
compression.
Typically,
.Sy lz4
compression is approximately 50% faster on compressible data and 200% faster
on incompressible data than
.Sy lzjb .
It is also approximately 80% faster on decompression,
while giving approximately a 10% better compression ratio.
.Pp
When the
.Sy lz4_compress
feature is set to
.Sy enabled ,
the administrator can turn on
.Sy lz4
compression on any dataset on the pool using the
.Xr zfs-set 8
command.
All newly written metadata will be compressed with the
.Sy lz4
algorithm.
.Pp
\*[instant-never]
.
.feature com.joyent multi_vdev_crash_dump no
This feature allows a dump device to be configured with a pool comprised
of multiple vdevs.
Those vdevs may be arranged in any mirrored or raidz configuration.
.Pp
When the
.Sy multi_vdev_crash_dump
feature is set to
.Sy enabled ,
the administrator can use
.Xr dumpadm 8
to configure a dump device on a pool comprised of multiple vdevs.
.Pp
Under
.Fx
and Linux this feature is unused, but registered for compatibility.
New pools created on these systems will have the feature
.Sy enabled
but will never transition to
.Sy active ,
as this functionality is not required for crash dump support.
Existing pools where this feature is
.Sy active
can be imported.
.
.feature com.delphix obsolete_counts yes device_removal
This feature is an enhancement of
.Sy device_removal ,
which will over time reduce the memory used to track removed devices.
When indirect blocks are freed or remapped,
we note that their part of the indirect mapping is
.Dq obsolete
– no longer needed.
.Pp
This feature becomes
.Sy active
when the
.Nm zpool Cm remove
command is used on a top-level vdev, and will never return to being
.Sy enabled .
.
.feature org.zfsonlinux project_quota yes extensible_dataset
This feature allows administrators to account the spaces and objects usage
information against the project identifier
.Pq ID .
.Pp
The project ID is an object-based attribute.
When upgrading an existing filesystem,
objects without a project ID will be assigned a zero project ID.
When this feature is enabled, newly created objects inherit
their parent directories' project ID if the parent's inherit flag is set
.Pq via Nm chattr Sy [+-]P No or Nm zfs Cm project Fl s Ns | Ns Fl C .
Otherwise, the new object's project ID will be zero.
An object's project ID can be changed at any time by the owner
.Pq or privileged user
via
.Nm chattr Fl p Ar prjid
or
.Nm zfs Cm project Fl p Ar prjid .
.Pp
This feature will become
.Sy active
as soon as it is enabled and will never return to being
.Sy disabled .
\*[remount-upgrade]
.
.feature com.delphix redaction_bookmarks no bookmarks extensible_dataset
This feature enables the use of redacted
.Nm zfs Cm send Ns s ,
which create redaction bookmarks storing the list of blocks
redacted by the send that created them.
For more information about redacted sends, see
.Xr zfs-send 8 .
.
.feature com.delphix redacted_datasets no extensible_dataset
This feature enables the receiving of redacted
.Nm zfs Cm send
streams, which create redacted datasets when received.
These datasets are missing some of their blocks,
and so cannot be safely mounted, and their contents cannot be safely read.
For more information about redacted receives, see
.Xr zfs-send 8 .
.
.feature com.datto resilver_defer yes
This feature allows ZFS to postpone new resilvers if an existing one is already
in progress.
Without this feature, any new resilvers will cause the currently
running one to be immediately restarted from the beginning.
.Pp
This feature becomes
.Sy active
once a resilver has been deferred, and returns to being
.Sy enabled
when the deferred resilver begins.
.
.feature org.illumos sha512 no extensible_dataset
This feature enables the use of the SHA-512/256 truncated hash algorithm
.Pq FIPS 180-4
for checksum and dedup.
The native 64-bit arithmetic of SHA-512 provides an approximate 50%
performance boost over SHA-256 on 64-bit hardware
and is thus a good minimum-change replacement candidate
for systems where hash performance is important,
but these systems cannot for whatever reason utilize the faster
.Sy skein No and Sy edonr
algorithms.
.Pp
.checksum-spiel sha512
.
.feature org.illumos skein no extensible_dataset
This feature enables the use of the Skein hash algorithm for checksum and dedup.
Skein is a high-performance secure hash algorithm that was a
finalist in the NIST SHA-3 competition.
It provides a very high security margin and high performance on 64-bit hardware
.Pq 80% faster than SHA-256 .
This implementation also utilizes the new salted checksumming
functionality in ZFS, which means that the checksum is pre-seeded with a
secret 256-bit random key
.Pq stored on the pool
before being fed the data block to be checksummed.
Thus the produced checksums are unique to a given pool,
preventing hash collision attacks on systems with dedup.
.Pp
.checksum-spiel skein
.
.feature com.delphix spacemap_histogram yes
This features allows ZFS to maintain more information about how free space
is organized within the pool.
If this feature is
.Sy enabled ,
it will be activated when a new space map object is created, or
an existing space map is upgraded to the new format,
and never returns back to being
.Sy enabled .
.
.feature com.delphix spacemap_v2 yes
This feature enables the use of the new space map encoding which
consists of two words
.Pq instead of one
whenever it is advantageous.
The new encoding allows space maps to represent large regions of
space more efficiently on-disk while also increasing their maximum
addressable offset.
.Pp
This feature becomes
.Sy active
once it is
.Sy enabled ,
and never returns back to being
.Sy enabled .
.
.feature org.zfsonlinux userobj_accounting yes extensible_dataset
This feature allows administrators to account the object usage information
by user and group.
.Pp
\*[instant-never]
\*[remount-upgrade]
.
+.feature com.klarasystems vdev_zaps_v2 no
+This feature creates a ZAP object for the root vdev.
+.Pp
+This feature becomes active after the next
+.Nm zpool Cm import
+or
+.Nm zpool reguid .
+.
+Properties can be retrieved or set on the root vdev using
+.Nm zpool Cm get
+and
+.Nm zpool Cm set
+with
+.Sy root
+as the vdev name which is an alias for
+.Sy root-0 .
.feature org.openzfs zilsaxattr yes extensible_dataset
This feature enables
.Sy xattr Ns = Ns Sy sa
extended attribute logging in the ZIL.
If enabled, extended attribute changes
.Pq both Sy xattrdir Ns = Ns Sy dir No and Sy xattr Ns = Ns Sy sa
are guaranteed to be durable if either the dataset had
.Sy sync Ns = Ns Sy always
set at the time the changes were made, or
.Xr sync 2
is called on the dataset after the changes were made.
.Pp
This feature becomes
.Sy active
when a ZIL is created for at least one dataset and will be returned to the
.Sy enabled
state when it is destroyed for all datasets that use this feature.
.
.feature com.delphix zpool_checkpoint yes
This feature enables the
.Nm zpool Cm checkpoint
command that can checkpoint the state of the pool
at the time it was issued and later rewind back to it or discard it.
.Pp
This feature becomes
.Sy active
when the
.Nm zpool Cm checkpoint
command is used to checkpoint the pool.
The feature will only return back to being
.Sy enabled
when the pool is rewound or the checkpoint has been discarded.
.
.feature org.freebsd zstd_compress no extensible_dataset
.Sy zstd
is a high-performance compression algorithm that features a
combination of high compression ratios and high speed.
Compared to
.Sy gzip ,
.Sy zstd
offers slightly better compression at much higher speeds.
Compared to
.Sy lz4 ,
.Sy zstd
offers much better compression while being only modestly slower.
Typically,
.Sy zstd
compression speed ranges from 250 to 500 MB/s per thread
and decompression speed is over 1 GB/s per thread.
.Pp
When the
.Sy zstd
feature is set to
.Sy enabled ,
the administrator can turn on
.Sy zstd
compression of any dataset using
.Nm zfs Cm set Sy compress Ns = Ns Sy zstd Ar dset
.Po see Xr zfs-set 8 Pc .
This feature becomes
.Sy active
once a
.Sy compress
property has been set to
.Sy zstd ,
and will return to being
.Sy enabled
once all filesystems that have ever had their
.Sy compress
property set to
.Sy zstd
are destroyed.
.El
.
.Sh SEE ALSO
.Xr zfs 8 ,
.Xr zpool 8
diff --git a/sys/contrib/openzfs/man/man7/zpoolconcepts.7 b/sys/contrib/openzfs/man/man7/zpoolconcepts.7
index 4ef96b157564..db3fd4926236 100644
--- a/sys/contrib/openzfs/man/man7/zpoolconcepts.7
+++ b/sys/contrib/openzfs/man/man7/zpoolconcepts.7
@@ -1,512 +1,514 @@
.\"
.\" CDDL HEADER START
.\"
.\" The contents of this file are subject to the terms of the
.\" Common Development and Distribution License (the "License").
.\" You may not use this file except in compliance with the License.
.\"
.\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
.\" or https://opensource.org/licenses/CDDL-1.0.
.\" See the License for the specific language governing permissions
.\" and limitations under the License.
.\"
.\" When distributing Covered Code, include this CDDL HEADER in each
.\" file and include the License file at usr/src/OPENSOLARIS.LICENSE.
.\" If applicable, add the following below this CDDL HEADER, with the
.\" fields enclosed by brackets "[]" replaced with your own identifying
.\" information: Portions Copyright [yyyy] [name of copyright owner]
.\"
.\" CDDL HEADER END
.\"
.\" Copyright (c) 2007, Sun Microsystems, Inc. All Rights Reserved.
.\" Copyright (c) 2012, 2018 by Delphix. All rights reserved.
.\" Copyright (c) 2012 Cyril Plisko. All Rights Reserved.
.\" Copyright (c) 2017 Datto Inc.
.\" Copyright (c) 2018 George Melikov. All Rights Reserved.
.\" Copyright 2017 Nexenta Systems, Inc.
.\" Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
.\"
-.Dd June 2, 2021
+.Dd April 7, 2023
.Dt ZPOOLCONCEPTS 7
.Os
.
.Sh NAME
.Nm zpoolconcepts
.Nd overview of ZFS storage pools
.
.Sh DESCRIPTION
.Ss Virtual Devices (vdevs)
-A "virtual device" describes a single device or a collection of devices
+A "virtual device" describes a single device or a collection of devices,
organized according to certain performance and fault characteristics.
The following virtual devices are supported:
.Bl -tag -width "special"
.It Sy disk
A block device, typically located under
.Pa /dev .
ZFS can use individual slices or partitions, though the recommended mode of
operation is to use whole disks.
A disk can be specified by a full path, or it can be a shorthand name
.Po the relative portion of the path under
.Pa /dev
.Pc .
A whole disk can be specified by omitting the slice or partition designation.
For example,
.Pa sda
is equivalent to
.Pa /dev/sda .
When given a whole disk, ZFS automatically labels the disk, if necessary.
.It Sy file
A regular file.
The use of files as a backing store is strongly discouraged.
It is designed primarily for experimental purposes, as the fault tolerance of a
file is only as good as the file system on which it resides.
A file must be specified by a full path.
.It Sy mirror
A mirror of two or more devices.
Data is replicated in an identical fashion across all components of a mirror.
A mirror with
.Em N No disks of size Em X No can hold Em X No bytes and can withstand Em N-1
-devices failing without losing data.
+devices failing, without losing data.
.It Sy raidz , raidz1 , raidz2 , raidz3
-A variation on RAID-5 that allows for better distribution of parity and
-eliminates the RAID-5
-.Qq write hole
+A distributed-parity layout, similar to RAID-5/6, with improved distribution of
+parity, and which does not suffer from the RAID-5/6
+.Qq write hole ,
.Pq in which data and parity become inconsistent after a power loss .
-Data and parity is striped across all disks within a raidz group.
+Data and parity is striped across all disks within a raidz group, though not
+necessarily in a consistent stripe width.
.Pp
A raidz group can have single, double, or triple parity, meaning that the
raidz group can sustain one, two, or three failures, respectively, without
losing any data.
The
.Sy raidz1
vdev type specifies a single-parity raidz group; the
.Sy raidz2
vdev type specifies a double-parity raidz group; and the
.Sy raidz3
vdev type specifies a triple-parity raidz group.
The
.Sy raidz
vdev type is an alias for
.Sy raidz1 .
.Pp
A raidz group with
.Em N No disks of size Em X No with Em P No parity disks can hold approximately
.Em (N-P)*X No bytes and can withstand Em P No devices failing without losing data .
The minimum number of devices in a raidz group is one more than the number of
parity disks.
The recommended number is between 3 and 9 to help increase performance.
.It Sy draid , draid1 , draid2 , draid3
-A variant of raidz that provides integrated distributed hot spares which
-allows for faster resilvering while retaining the benefits of raidz.
+A variant of raidz that provides integrated distributed hot spares, allowing
+for faster resilvering, while retaining the benefits of raidz.
A dRAID vdev is constructed from multiple internal raidz groups, each with
.Em D No data devices and Em P No parity devices .
These groups are distributed over all of the children in order to fully
utilize the available disk performance.
.Pp
Unlike raidz, dRAID uses a fixed stripe width (padding as necessary with
zeros) to allow fully sequential resilvering.
-This fixed stripe width significantly effects both usable capacity and IOPS.
+This fixed stripe width significantly affects both usable capacity and IOPS.
For example, with the default
.Em D=8 No and Em 4 KiB No disk sectors the minimum allocation size is Em 32 KiB .
If using compression, this relatively large allocation size can reduce the
effective compression ratio.
-When using ZFS volumes and dRAID, the default of the
+When using ZFS volumes (zvols) and dRAID, the default of the
.Sy volblocksize
property is increased to account for the allocation size.
If a dRAID pool will hold a significant amount of small blocks, it is
recommended to also add a mirrored
.Sy special
vdev to store those blocks.
.Pp
-In regards to I/O, performance is similar to raidz since for any read all
+In regards to I/O, performance is similar to raidz since, for any read, all
.Em D No data disks must be accessed .
Delivered random IOPS can be reasonably approximated as
.Sy floor((N-S)/(D+P))*single_drive_IOPS .
.Pp
Like raidz, a dRAID can have single-, double-, or triple-parity.
The
.Sy draid1 ,
.Sy draid2 ,
and
.Sy draid3
types can be used to specify the parity level.
The
.Sy draid
vdev type is an alias for
.Sy draid1 .
.Pp
A dRAID with
.Em N No disks of size Em X , D No data disks per redundancy group , Em P
.No parity level, and Em S No distributed hot spares can hold approximately
.Em (N-S)*(D/(D+P))*X No bytes and can withstand Em P
devices failing without losing data.
.It Sy draid Ns Oo Ar parity Oc Ns Oo Sy \&: Ns Ar data Ns Sy d Oc Ns Oo Sy \&: Ns Ar children Ns Sy c Oc Ns Oo Sy \&: Ns Ar spares Ns Sy s Oc
A non-default dRAID configuration can be specified by appending one or more
of the following optional arguments to the
.Sy draid
keyword:
.Bl -tag -compact -width "children"
.It Ar parity
The parity level (1-3).
.It Ar data
The number of data devices per redundancy group.
In general, a smaller value of
.Em D No will increase IOPS, improve the compression ratio ,
and speed up resilvering at the expense of total usable capacity.
Defaults to
.Em 8 , No unless Em N-P-S No is less than Em 8 .
.It Ar children
The expected number of children.
Useful as a cross-check when listing a large number of devices.
An error is returned when the provided number of children differs.
.It Ar spares
The number of distributed hot spares.
Defaults to zero.
.El
.It Sy spare
A pseudo-vdev which keeps track of available hot spares for a pool.
For more information, see the
.Sx Hot Spares
section.
.It Sy log
A separate intent log device.
If more than one log device is specified, then writes are load-balanced between
devices.
Log devices can be mirrored.
However, raidz vdev types are not supported for the intent log.
For more information, see the
.Sx Intent Log
section.
.It Sy dedup
-A device dedicated solely for deduplication tables.
+A device solely dedicated for deduplication tables.
The redundancy of this device should match the redundancy of the other normal
devices in the pool.
If more than one dedup device is specified, then
allocations are load-balanced between those devices.
.It Sy special
A device dedicated solely for allocating various kinds of internal metadata,
and optionally small file blocks.
The redundancy of this device should match the redundancy of the other normal
devices in the pool.
If more than one special device is specified, then
allocations are load-balanced between those devices.
.Pp
For more information on special allocations, see the
.Sx Special Allocation Class
section.
.It Sy cache
A device used to cache storage pool data.
A cache device cannot be configured as a mirror or raidz group.
For more information, see the
.Sx Cache Devices
section.
.El
.Pp
Virtual devices cannot be nested, so a mirror or raidz virtual device can only
contain files or disks.
Mirrors of mirrors
.Pq or other combinations
are not allowed.
.Pp
A pool can have any number of virtual devices at the top of the configuration
.Po known as
.Qq root vdevs
.Pc .
Data is dynamically distributed across all top-level devices to balance data
among devices.
As new virtual devices are added, ZFS automatically places data on the newly
available devices.
.Pp
Virtual devices are specified one at a time on the command line,
separated by whitespace.
Keywords like
.Sy mirror No and Sy raidz
are used to distinguish where a group ends and another begins.
For example, the following creates a pool with two root vdevs,
each a mirror of two disks:
.Dl # Nm zpool Cm create Ar mypool Sy mirror Ar sda sdb Sy mirror Ar sdc sdd
.
.Ss Device Failure and Recovery
ZFS supports a rich set of mechanisms for handling device failure and data
corruption.
All metadata and data is checksummed, and ZFS automatically repairs bad data
-from a good copy when corruption is detected.
+from a good copy, when corruption is detected.
.Pp
In order to take advantage of these features, a pool must make use of some form
of redundancy, using either mirrored or raidz groups.
While ZFS supports running in a non-redundant configuration, where each root
vdev is simply a disk or file, this is strongly discouraged.
A single case of bit corruption can render some or all of your data unavailable.
.Pp
A pool's health status is described by one of three states:
.Sy online , degraded , No or Sy faulted .
An online pool has all devices operating normally.
A degraded pool is one in which one or more devices have failed, but the data is
still available due to a redundant configuration.
A faulted pool has corrupted metadata, or one or more faulted devices, and
insufficient replicas to continue functioning.
.Pp
The health of the top-level vdev, such as a mirror or raidz device,
-is potentially impacted by the state of its associated vdevs,
+is potentially impacted by the state of its associated vdevs
or component devices.
A top-level vdev or component device is in one of the following states:
.Bl -tag -width "DEGRADED"
.It Sy DEGRADED
One or more top-level vdevs is in the degraded state because one or more
component devices are offline.
Sufficient replicas exist to continue functioning.
.Pp
One or more component devices is in the degraded or faulted state, but
sufficient replicas exist to continue functioning.
The underlying conditions are as follows:
.Bl -bullet -compact
.It
The number of checksum errors exceeds acceptable levels and the device is
degraded as an indication that something may be wrong.
ZFS continues to use the device as necessary.
.It
The number of I/O errors exceeds acceptable levels.
The device could not be marked as faulted because there are insufficient
replicas to continue functioning.
.El
.It Sy FAULTED
One or more top-level vdevs is in the faulted state because one or more
component devices are offline.
Insufficient replicas exist to continue functioning.
.Pp
One or more component devices is in the faulted state, and insufficient
replicas exist to continue functioning.
The underlying conditions are as follows:
.Bl -bullet -compact
.It
The device could be opened, but the contents did not match expected values.
.It
The number of I/O errors exceeds acceptable levels and the device is faulted to
prevent further use of the device.
.El
.It Sy OFFLINE
The device was explicitly taken offline by the
.Nm zpool Cm offline
command.
.It Sy ONLINE
The device is online and functioning.
.It Sy REMOVED
The device was physically removed while the system was running.
Device removal detection is hardware-dependent and may not be supported on all
platforms.
.It Sy UNAVAIL
The device could not be opened.
If a pool is imported when a device was unavailable, then the device will be
identified by a unique identifier instead of its path since the path was never
correct in the first place.
.El
.Pp
Checksum errors represent events where a disk returned data that was expected
to be correct, but was not.
In other words, these are instances of silent data corruption.
The checksum errors are reported in
.Nm zpool Cm status
and
.Nm zpool Cm events .
When a block is stored redundantly, a damaged block may be reconstructed
(e.g. from raidz parity or a mirrored copy).
In this case, ZFS reports the checksum error against the disks that contained
damaged data.
If a block is unable to be reconstructed (e.g. due to 3 disks being damaged
in a raidz2 group), it is not possible to determine which disks were silently
corrupted.
In this case, checksum errors are reported for all disks on which the block
is stored.
.Pp
If a device is removed and later re-attached to the system,
-ZFS attempts online the device automatically.
+ZFS attempts to bring the device online automatically.
Device attachment detection is hardware-dependent
and might not be supported on all platforms.
.
.Ss Hot Spares
ZFS allows devices to be associated with pools as
.Qq hot spares .
-These devices are not actively used in the pool, but when an active device
+These devices are not actively used in the pool.
+But, when an active device
fails, it is automatically replaced by a hot spare.
To create a pool with hot spares, specify a
.Sy spare
vdev with any number of devices.
For example,
.Dl # Nm zpool Cm create Ar pool Sy mirror Ar sda sdb Sy spare Ar sdc sdd
.Pp
Spares can be shared across multiple pools, and can be added with the
.Nm zpool Cm add
command and removed with the
.Nm zpool Cm remove
command.
Once a spare replacement is initiated, a new
.Sy spare
vdev is created within the configuration that will remain there until the
original device is replaced.
-At this point, the hot spare becomes available again if another device fails.
+At this point, the hot spare becomes available again, if another device fails.
.Pp
-If a pool has a shared spare that is currently being used, the pool can not be
-exported since other pools may use this shared spare, which may lead to
+If a pool has a shared spare that is currently being used, the pool cannot be
+exported, since other pools may use this shared spare, which may lead to
potential data corruption.
.Pp
Shared spares add some risk.
If the pools are imported on different hosts,
and both pools suffer a device failure at the same time,
both could attempt to use the spare at the same time.
This may not be detected, resulting in data corruption.
.Pp
An in-progress spare replacement can be cancelled by detaching the hot spare.
If the original faulted device is detached, then the hot spare assumes its
place in the configuration, and is removed from the spare list of all active
pools.
.Pp
The
.Sy draid
vdev type provides distributed hot spares.
These hot spares are named after the dRAID vdev they're a part of
.Po Sy draid1 Ns - Ns Ar 2 Ns - Ns Ar 3 No specifies spare Ar 3 No of vdev Ar 2 ,
.No which is a single parity dRAID Pc
and may only be used by that dRAID vdev.
Otherwise, they behave the same as normal hot spares.
.Pp
Spares cannot replace log devices.
.
.Ss Intent Log
The ZFS Intent Log (ZIL) satisfies POSIX requirements for synchronous
transactions.
For instance, databases often require their transactions to be on stable storage
devices when returning from a system call.
NFS and other applications can also use
.Xr fsync 2
to ensure data stability.
By default, the intent log is allocated from blocks within the main pool.
However, it might be possible to get better performance using separate intent
log devices such as NVRAM or a dedicated disk.
For example:
.Dl # Nm zpool Cm create Ar pool sda sdb Sy log Ar sdc
.Pp
Multiple log devices can also be specified, and they can be mirrored.
See the
.Sx EXAMPLES
section for an example of mirroring multiple log devices.
.Pp
-Log devices can be added, replaced, attached, detached and removed.
+Log devices can be added, replaced, attached, detached, and removed.
In addition, log devices are imported and exported as part of the pool
that contains them.
Mirrored devices can be removed by specifying the top-level mirror vdev.
.
.Ss Cache Devices
Devices can be added to a storage pool as
.Qq cache devices .
These devices provide an additional layer of caching between main memory and
disk.
For read-heavy workloads, where the working set size is much larger than what
can be cached in main memory, using cache devices allows much more of this
working set to be served from low latency media.
Using cache devices provides the greatest performance improvement for random
read-workloads of mostly static content.
.Pp
To create a pool with cache devices, specify a
.Sy cache
vdev with any number of devices.
For example:
.Dl # Nm zpool Cm create Ar pool sda sdb Sy cache Ar sdc sdd
.Pp
Cache devices cannot be mirrored or part of a raidz configuration.
If a read error is encountered on a cache device, that read I/O is reissued to
the original storage pool device, which might be part of a mirrored or raidz
configuration.
.Pp
The content of the cache devices is persistent across reboots and restored
asynchronously when importing the pool in L2ARC (persistent L2ARC).
This can be disabled by setting
.Sy l2arc_rebuild_enabled Ns = Ns Sy 0 .
For cache devices smaller than
.Em 1 GiB ,
-we do not write the metadata structures
-required for rebuilding the L2ARC in order not to waste space.
+ZFS does not write the metadata structures
+required for rebuilding the L2ARC, to conserve space.
This can be changed with
.Sy l2arc_rebuild_blocks_min_l2size .
The cache device header
.Pq Em 512 B
is updated even if no metadata structures are written.
Setting
.Sy l2arc_headroom Ns = Ns Sy 0
will result in scanning the full-length ARC lists for cacheable content to be
written in L2ARC (persistent ARC).
If a cache device is added with
-.Nm zpool Cm add
-its label and header will be overwritten and its contents are not going to be
+.Nm zpool Cm add ,
+its label and header will be overwritten and its contents will not be
restored in L2ARC, even if the device was previously part of the pool.
If a cache device is onlined with
-.Nm zpool Cm online
+.Nm zpool Cm online ,
its contents will be restored in L2ARC.
-This is useful in case of memory pressure
+This is useful in case of memory pressure,
where the contents of the cache device are not fully restored in L2ARC.
-The user can off- and online the cache device when there is less memory pressure
-in order to fully restore its contents to L2ARC.
+The user can off- and online the cache device when there is less memory
+pressure, to fully restore its contents to L2ARC.
.
.Ss Pool checkpoint
Before starting critical procedures that include destructive actions
.Pq like Nm zfs Cm destroy ,
-an administrator can checkpoint the pool's state and in the case of a
+an administrator can checkpoint the pool's state and, in the case of a
mistake or failure, rewind the entire pool back to the checkpoint.
Otherwise, the checkpoint can be discarded when the procedure has completed
successfully.
.Pp
A pool checkpoint can be thought of as a pool-wide snapshot and should be used
with care as it contains every part of the pool's state, from properties to vdev
configuration.
Thus, certain operations are not allowed while a pool has a checkpoint.
Specifically, vdev removal/attach/detach, mirror splitting, and
changing the pool's GUID.
Adding a new vdev is supported, but in the case of a rewind it will have to be
added again.
Finally, users of this feature should keep in mind that scrubs in a pool that
has a checkpoint do not repair checkpointed data.
.Pp
To create a checkpoint for a pool:
.Dl # Nm zpool Cm checkpoint Ar pool
.Pp
To later rewind to its checkpointed state, you need to first export it and
then rewind it during import:
.Dl # Nm zpool Cm export Ar pool
.Dl # Nm zpool Cm import Fl -rewind-to-checkpoint Ar pool
.Pp
To discard the checkpoint from a pool:
.Dl # Nm zpool Cm checkpoint Fl d Ar pool
.Pp
Dataset reservations (controlled by the
.Sy reservation No and Sy refreservation
properties) may be unenforceable while a checkpoint exists, because the
checkpoint is allowed to consume the dataset's reservation.
Finally, data that is part of the checkpoint but has been freed in the
current state of the pool won't be scanned during a scrub.
.
.Ss Special Allocation Class
Allocations in the special class are dedicated to specific block types.
-By default this includes all metadata, the indirect blocks of user data, and
+By default, this includes all metadata, the indirect blocks of user data, and
any deduplication tables.
The class can also be provisioned to accept small file blocks.
.Pp
A pool must always have at least one normal
.Pq non- Ns Sy dedup Ns /- Ns Sy special
vdev before
other devices can be assigned to the special class.
If the
.Sy special
class becomes full, then allocations intended for it
will spill back into the normal class.
.Pp
Deduplication tables can be excluded from the special class by unsetting the
.Sy zfs_ddt_data_is_special
ZFS module parameter.
.Pp
Inclusion of small file blocks in the special class is opt-in.
Each dataset can control the size of small file blocks allowed
in the special class by setting the
.Sy special_small_blocks
property to nonzero.
See
.Xr zfsprops 7
for more info on this property.
diff --git a/sys/contrib/openzfs/module/Makefile.bsd b/sys/contrib/openzfs/module/Makefile.bsd
index 8ec094d4ad1c..365609fb8585 100644
--- a/sys/contrib/openzfs/module/Makefile.bsd
+++ b/sys/contrib/openzfs/module/Makefile.bsd
@@ -1,520 +1,530 @@
.if !defined(WITH_CTF)
WITH_CTF=1
.endif
.include <bsd.sys.mk>
SRCDIR=${.CURDIR}
INCDIR=${.CURDIR:H}/include
KMOD= openzfs
.PATH: ${SRCDIR}/avl \
${SRCDIR}/lua \
${SRCDIR}/nvpair \
${SRCDIR}/icp/algs/blake3 \
${SRCDIR}/icp/algs/edonr \
${SRCDIR}/icp/algs/sha2 \
${SRCDIR}/icp/asm-aarch64/blake3 \
${SRCDIR}/icp/asm-aarch64/sha2 \
${SRCDIR}/icp/asm-arm/sha2 \
${SRCDIR}/icp/asm-ppc64/sha2 \
${SRCDIR}/icp/asm-ppc64/blake3 \
${SRCDIR}/icp/asm-x86_64/blake3 \
${SRCDIR}/icp/asm-x86_64/sha2 \
${SRCDIR}/os/freebsd/spl \
${SRCDIR}/os/freebsd/zfs \
${SRCDIR}/unicode \
${SRCDIR}/zcommon \
${SRCDIR}/zfs \
${SRCDIR}/zstd \
${SRCDIR}/zstd/lib/common \
${SRCDIR}/zstd/lib/compress \
${SRCDIR}/zstd/lib/decompress
CFLAGS+= -I${INCDIR}
CFLAGS+= -I${SRCDIR}/icp/include
CFLAGS+= -I${INCDIR}/os/freebsd
CFLAGS+= -I${INCDIR}/os/freebsd/spl
CFLAGS+= -I${INCDIR}/os/freebsd/zfs
CFLAGS+= -I${SRCDIR}/zstd/include
CFLAGS+= -include ${INCDIR}/os/freebsd/spl/sys/ccompile.h
CFLAGS+= -I${.CURDIR}
CFLAGS+= -D__KERNEL__ -DFREEBSD_NAMECACHE -DBUILDING_ZFS -D__BSD_VISIBLE=1 \
-DHAVE_UIO_ZEROCOPY -DWITHOUT_NETDUMP -D__KERNEL -D_SYS_CONDVAR_H_ \
-D_SYS_VMEM_H_ -DKDTRACE_HOOKS -DCOMPAT_FREEBSD11
.if ${MACHINE_ARCH} == "amd64"
CFLAGS+= -D__x86_64 -DHAVE_SSE2 -DHAVE_SSSE3 -DHAVE_SSE4_1 -DHAVE_SSE4_2 \
-DHAVE_AVX -DHAVE_AVX2 -DHAVE_AVX512F -DHAVE_AVX512VL -DHAVE_AVX512BW
.endif
.if defined(WITH_DEBUG) && ${WITH_DEBUG} == "true"
CFLAGS+= -DZFS_DEBUG -g
.if defined(WITH_INVARIANTS) && ${WITH_INVARIANTS} == "true"
CFLAGS+= -DINVARIANTS -DWITNESS -DOPENSOLARIS_WITNESS
.endif
.if defined(WITH_O0) && ${WITH_O0} == "true"
CFLAGS+= -O0
.endif
.else
CFLAGS += -DNDEBUG
.endif
.if defined(WITH_VFS_DEBUG) && ${WITH_VFS_DEBUG} == "true"
# kernel must also be built with this option for this to work
CFLAGS+= -DDEBUG_VFS_LOCKS
.endif
.if defined(WITH_GCOV) && ${WITH_GCOV} == "true"
CFLAGS+= -fprofile-arcs -ftest-coverage
.endif
DEBUG_FLAGS=-g
.if ${MACHINE_ARCH} == "i386" || ${MACHINE_ARCH} == "powerpc" || \
${MACHINE_ARCH} == "powerpcspe" || ${MACHINE_ARCH} == "arm"
CFLAGS+= -DBITS_PER_LONG=32
.else
CFLAGS+= -DBITS_PER_LONG=64
.endif
SRCS= vnode_if.h device_if.h bus_if.h
# avl
SRCS+= avl.c
# icp
SRCS+= edonr.c
#icp/algs/blake3
SRCS+= blake3.c \
blake3_generic.c \
blake3_impl.c
#icp/asm-aarch64/blake3
SRCS+= b3_aarch64_sse2.S \
b3_aarch64_sse41.S
#icp/asm-ppc64/blake3
SRCS+= b3_ppc64le_sse2.S \
b3_ppc64le_sse41.S
#icp/asm-x86_64/blake3
SRCS+= blake3_avx2.S \
blake3_avx512.S \
blake3_sse2.S \
blake3_sse41.S
#icp/algs/sha2
SRCS+= sha2_generic.c \
sha256_impl.c \
sha512_impl.c
#icp/asm-arm/sha2
SRCS+= sha256-armv7.S \
sha512-armv7.S
#icp/asm-aarch64/sha2
SRCS+= sha256-armv8.S \
sha512-armv8.S
#icp/asm-ppc64/sha2
SRCS+= sha256-p8.S \
sha512-p8.S \
sha256-ppc.S \
sha512-ppc.S
#icp/asm-x86_64/sha2
SRCS+= sha256-x86_64.S \
sha512-x86_64.S
#lua
SRCS+= lapi.c \
lauxlib.c \
lbaselib.c \
lcode.c \
lcompat.c \
lcorolib.c \
lctype.c \
ldebug.c \
ldo.c \
lfunc.c \
lgc.c \
llex.c \
lmem.c \
lobject.c \
lopcodes.c \
lparser.c \
lstate.c \
lstring.c \
lstrlib.c \
ltable.c \
ltablib.c \
ltm.c \
lvm.c \
lzio.c
#nvpair
SRCS+= nvpair.c \
fnvpair.c \
nvpair_alloc_spl.c \
nvpair_alloc_fixed.c
#os/freebsd/spl
SRCS+= acl_common.c \
callb.c \
list.c \
spl_acl.c \
spl_cmn_err.c \
spl_dtrace.c \
spl_kmem.c \
spl_kstat.c \
spl_misc.c \
spl_policy.c \
spl_procfs_list.c \
spl_string.c \
spl_sunddi.c \
spl_sysevent.c \
spl_taskq.c \
spl_uio.c \
spl_vfs.c \
spl_vm.c \
spl_zlib.c \
spl_zone.c
.if ${MACHINE_ARCH} == "i386" || ${MACHINE_ARCH} == "powerpc" || \
${MACHINE_ARCH} == "powerpcspe" || ${MACHINE_ARCH} == "arm"
SRCS+= spl_atomic.c
.endif
#os/freebsd/zfs
SRCS+= abd_os.c \
arc_os.c \
crypto_os.c \
dmu_os.c \
event_os.c \
hkdf.c \
kmod_core.c \
spa_os.c \
sysctl_os.c \
vdev_file.c \
vdev_geom.c \
vdev_label_os.c \
zfs_acl.c \
zfs_ctldir.c \
zfs_debug.c \
zfs_dir.c \
zfs_ioctl_compat.c \
zfs_ioctl_os.c \
zfs_racct.c \
zfs_vfsops.c \
zfs_vnops_os.c \
zfs_znode.c \
zio_crypt.c \
zvol_os.c
#unicode
SRCS+= uconv.c \
u8_textprep.c
#zcommon
SRCS+= zfeature_common.c \
zfs_comutil.c \
zfs_deleg.c \
zfs_fletcher.c \
zfs_fletcher_avx512.c \
zfs_fletcher_intel.c \
zfs_fletcher_sse.c \
zfs_fletcher_superscalar.c \
zfs_fletcher_superscalar4.c \
zfs_namecheck.c \
zfs_prop.c \
zpool_prop.c \
zprop_common.c
#zfs
SRCS+= abd.c \
aggsum.c \
arc.c \
blake3_zfs.c \
blkptr.c \
bplist.c \
bpobj.c \
brt.c \
btree.c \
cityhash.c \
dbuf.c \
dbuf_stats.c \
bptree.c \
bqueue.c \
dataset_kstats.c \
ddt.c \
ddt_zap.c \
dmu.c \
dmu_diff.c \
dmu_object.c \
dmu_objset.c \
dmu_recv.c \
dmu_redact.c \
dmu_send.c \
dmu_traverse.c \
dmu_tx.c \
dmu_zfetch.c \
dnode.c \
dnode_sync.c \
dsl_dataset.c \
dsl_deadlist.c \
dsl_deleg.c \
dsl_bookmark.c \
dsl_dir.c \
dsl_crypt.c \
dsl_destroy.c \
dsl_pool.c \
dsl_prop.c \
dsl_scan.c \
dsl_synctask.c \
dsl_userhold.c \
edonr_zfs.c \
fm.c \
gzip.c \
lzjb.c \
lz4.c \
lz4_zfs.c \
metaslab.c \
mmp.c \
multilist.c \
objlist.c \
pathname.c \
range_tree.c \
refcount.c \
rrwlock.c \
sa.c \
sha2_zfs.c \
skein_zfs.c \
spa.c \
spa_checkpoint.c \
spa_config.c \
spa_errlog.c \
spa_history.c \
spa_log_spacemap.c \
spa_misc.c \
spa_stats.c \
space_map.c \
space_reftree.c \
txg.c \
uberblock.c \
unique.c \
vdev.c \
vdev_cache.c \
vdev_draid.c \
vdev_draid_rand.c \
vdev_indirect.c \
vdev_indirect_births.c \
vdev_indirect_mapping.c \
vdev_initialize.c \
vdev_label.c \
vdev_mirror.c \
vdev_missing.c \
vdev_queue.c \
vdev_raidz.c \
vdev_raidz_math.c \
vdev_raidz_math_scalar.c \
vdev_raidz_math_avx2.c \
vdev_raidz_math_avx512bw.c \
vdev_raidz_math_avx512f.c \
vdev_raidz_math_sse2.c \
vdev_raidz_math_ssse3.c \
vdev_rebuild.c \
vdev_removal.c \
vdev_root.c \
vdev_trim.c \
zap.c \
zap_leaf.c \
zap_micro.c \
zcp.c \
zcp_get.c \
zcp_global.c \
zcp_iter.c \
zcp_set.c \
zcp_synctask.c \
zfeature.c \
zfs_byteswap.c \
zfs_chksum.c \
zfs_file_os.c \
zfs_fm.c \
zfs_fuid.c \
zfs_impl.c \
zfs_ioctl.c \
zfs_log.c \
zfs_onexit.c \
zfs_quota.c \
zfs_ratelimit.c \
zfs_replay.c \
zfs_rlock.c \
zfs_sa.c \
zfs_vnops.c \
zil.c \
zio.c \
zio_checksum.c \
zio_compress.c \
zio_inject.c \
zle.c \
zrlock.c \
zthr.c \
zvol.c
#zstd
SRCS+= zfs_zstd.c \
entropy_common.c \
error_private.c \
fse_compress.c \
fse_decompress.c \
hist.c \
huf_compress.c \
huf_decompress.c \
pool.c \
xxhash.c \
zstd_common.c \
zstd_compress.c \
zstd_compress_literals.c \
zstd_compress_sequences.c \
zstd_compress_superblock.c \
zstd_ddict.c \
zstd_decompress.c \
zstd_decompress_block.c \
zstd_double_fast.c \
zstd_fast.c \
zstd_lazy.c \
zstd_ldm.c \
zstd_opt.c
beforeinstall:
.if ${MK_DEBUG_FILES} != "no"
mtree -eu \
-f /etc/mtree/BSD.debug.dist \
-p ${DESTDIR}/usr/lib
.endif
.include <bsd.kmod.mk>
CFLAGS.sysctl_os.c= -include ../zfs_config.h
CFLAGS.xxhash.c+= -include ${SYSDIR}/sys/_null.h
CFLAGS.gcc+= -Wno-pointer-to-int-cast
CFLAGS.abd.c= -Wno-cast-qual
CFLAGS.ddt.c= -Wno-cast-qual
CFLAGS.dmu.c= -Wno-cast-qual
CFLAGS.dmu_traverse.c= -Wno-cast-qual
CFLAGS.dnode.c= ${NO_WUNUSED_BUT_SET_VARIABLE}
CFLAGS.dsl_deadlist.c= -Wno-cast-qual
CFLAGS.dsl_dir.c= -Wno-cast-qual
CFLAGS.dsl_prop.c= -Wno-cast-qual
CFLAGS.edonr.c= -Wno-cast-qual
CFLAGS.fm.c= -Wno-cast-qual
CFLAGS.hist.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.lapi.c= -Wno-cast-qual
CFLAGS.lcompat.c= -Wno-cast-qual
CFLAGS.ldo.c= ${NO_WINFINITE_RECURSION}
CFLAGS.lobject.c= -Wno-cast-qual
CFLAGS.ltable.c= -Wno-cast-qual
CFLAGS.lvm.c= -Wno-cast-qual
CFLAGS.lz4.c= -Wno-cast-qual
CFLAGS.lz4_zfs.c= -Wno-cast-qual
CFLAGS.nvpair.c= -Wno-cast-qual -DHAVE_RPC_TYPES ${NO_WSTRINGOP_OVERREAD}
CFLAGS.pool.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.pool.c= -U__BMI__ -fno-tree-vectorize
CFLAGS.spa.c= -Wno-cast-qual
CFLAGS.spa_misc.c= -Wno-cast-qual
CFLAGS.spl_string.c= -Wno-cast-qual
CFLAGS.spl_vm.c= -Wno-cast-qual
CFLAGS.spl_zlib.c= -Wno-cast-qual
CFLAGS.u8_textprep.c= -Wno-cast-qual
CFLAGS.vdev_draid.c= -Wno-cast-qual
CFLAGS.vdev_raidz.c= -Wno-cast-qual
CFLAGS.vdev_raidz_math.c= -Wno-cast-qual
CFLAGS.vdev_raidz_math_avx2.c= -Wno-cast-qual -Wno-duplicate-decl-specifier
CFLAGS.vdev_raidz_math_avx512f.c= -Wno-cast-qual -Wno-duplicate-decl-specifier
CFLAGS.vdev_raidz_math_scalar.c= -Wno-cast-qual
CFLAGS.vdev_raidz_math_sse2.c= -Wno-cast-qual -Wno-duplicate-decl-specifier
CFLAGS.zap_leaf.c= -Wno-cast-qual
CFLAGS.zap_micro.c= -Wno-cast-qual
CFLAGS.zcp.c= -Wno-cast-qual
CFLAGS.zfs_fletcher.c= -Wno-cast-qual -Wno-pointer-arith
CFLAGS.zfs_fletcher_avx512.c= -Wno-cast-qual -Wno-pointer-arith
CFLAGS.zfs_fletcher_intel.c= -Wno-cast-qual -Wno-pointer-arith
CFLAGS.zfs_fletcher_sse.c= -Wno-cast-qual -Wno-pointer-arith
CFLAGS.zfs_fm.c= -Wno-cast-qual ${NO_WUNUSED_BUT_SET_VARIABLE}
CFLAGS.zfs_ioctl.c= -Wno-cast-qual
CFLAGS.zfs_log.c= -Wno-cast-qual
CFLAGS.zfs_vnops_os.c= -Wno-pointer-arith
CFLAGS.zfs_zstd.c= -Wno-cast-qual -Wno-pointer-arith
CFLAGS.zil.c= -Wno-cast-qual
CFLAGS.zio.c= -Wno-cast-qual
CFLAGS.zprop_common.c= -Wno-cast-qual
CFLAGS.zrlock.c= -Wno-cast-qual
#zstd
CFLAGS.entropy_common.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.error_private.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.fse_compress.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL} ${NO_WUNUSED_BUT_SET_VARIABLE}
CFLAGS.fse_decompress.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.huf_compress.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.huf_decompress.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.xxhash.c+= -U__BMI__ -fno-tree-vectorize
CFLAGS.xxhash.c+= ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_common.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_compress.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_compress_literals.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_compress_sequences.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_compress_superblock.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL} ${NO_WUNUSED_BUT_SET_VARIABLE}
CFLAGS.zstd_ddict.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_decompress.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_decompress_block.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_double_fast.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_fast.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_lazy.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_ldm.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
CFLAGS.zstd_opt.c= -U__BMI__ -fno-tree-vectorize ${NO_WBITWISE_INSTEAD_OF_LOGICAL}
.if ${MACHINE_CPUARCH} == "aarch64"
__ZFS_ZSTD_AARCH64_FLAGS= -include ${SRCDIR}/zstd/include/aarch64_compat.h
CFLAGS.zstd.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.entropy_common.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.error_private.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.fse_compress.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.fse_decompress.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.hist.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.huf_compress.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.huf_decompress.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.pool.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.xxhash.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_common.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_compress.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_compress_literals.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_compress_sequences.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_compress_superblock.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_ddict.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_decompress.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_decompress_block.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_double_fast.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_fast.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_lazy.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_ldm.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
CFLAGS.zstd_opt.c+= ${__ZFS_ZSTD_AARCH64_FLAGS}
+sha256-armv8.o: sha256-armv8.S
+ ${CC} -c ${CFLAGS:N-mgeneral-regs-only} ${WERROR} ${.IMPSRC} \
+ -o ${.TARGET}
+ ${CTFCONVERT_CMD}
+
+sha512-armv8.o: sha512-armv8.S
+ ${CC} -c ${CFLAGS:N-mgeneral-regs-only} ${WERROR} ${.IMPSRC} \
+ -o ${.TARGET}
+ ${CTFCONVERT_CMD}
+
b3_aarch64_sse2.o: b3_aarch64_sse2.S
${CC} -c ${CFLAGS:N-mgeneral-regs-only} ${WERROR} ${.IMPSRC} \
-o ${.TARGET}
${CTFCONVERT_CMD}
b3_aarch64_sse41.o: b3_aarch64_sse41.S
${CC} -c ${CFLAGS:N-mgeneral-regs-only} ${WERROR} ${.IMPSRC} \
-o ${.TARGET}
${CTFCONVERT_CMD}
.endif
diff --git a/sys/contrib/openzfs/module/lua/ldo.c b/sys/contrib/openzfs/module/lua/ldo.c
index bf525588e260..38bd4e08a73d 100644
--- a/sys/contrib/openzfs/module/lua/ldo.c
+++ b/sys/contrib/openzfs/module/lua/ldo.c
@@ -1,760 +1,762 @@
/*
** $Id: ldo.c,v 2.108.1.3 2013/11/08 18:22:50 roberto Exp $
** Stack and Call structure of Lua
** See Copyright Notice in lua.h
*/
#define ldo_c
#define LUA_CORE
#include <sys/lua/lua.h>
#include <sys/asm_linkage.h>
#include "lapi.h"
#include "ldebug.h"
#include "ldo.h"
#include "lfunc.h"
#include "lgc.h"
#include "lmem.h"
#include "lobject.h"
#include "lopcodes.h"
#include "lparser.h"
#include "lstate.h"
#include "lstring.h"
#include "ltable.h"
#include "ltm.h"
#include "lvm.h"
#include "lzio.h"
/* Return the number of bytes available on the stack. */
#if defined (_KERNEL) && defined(__linux__)
#include <asm/current.h>
static intptr_t stack_remaining(void) {
intptr_t local;
local = (intptr_t)&local - (intptr_t)current->stack;
return local;
}
#elif defined (_KERNEL) && defined(__FreeBSD__)
#include <sys/pcpu.h>
static intptr_t stack_remaining(void) {
intptr_t local;
local = (intptr_t)&local - (intptr_t)curthread->td_kstack;
return local;
}
#else
static intptr_t stack_remaining(void) {
return INTPTR_MAX;
}
#endif
/*
** {======================================================
** Error-recovery functions
** =======================================================
*/
/*
** LUAI_THROW/LUAI_TRY define how Lua does exception handling. By
** default, Lua handles errors with exceptions when compiling as
** C++ code, with _longjmp/_setjmp when asked to use them, and with
** longjmp/setjmp otherwise.
*/
#if !defined(LUAI_THROW)
#ifdef _KERNEL
#ifdef __linux__
#if defined(__i386__)
#define JMP_BUF_CNT 6
#elif defined(__x86_64__)
#define JMP_BUF_CNT 8
#elif defined(__sparc__) && defined(__arch64__)
#define JMP_BUF_CNT 6
#elif defined(__powerpc__)
#define JMP_BUF_CNT 26
#elif defined(__aarch64__)
#define JMP_BUF_CNT 64
#elif defined(__arm__)
#define JMP_BUF_CNT 65
#elif defined(__mips__)
#define JMP_BUF_CNT 12
#elif defined(__s390x__)
#define JMP_BUF_CNT 18
#elif defined(__riscv)
#define JMP_BUF_CNT 64
+#elif defined(__loongarch_lp64)
+#define JMP_BUF_CNT 64
#else
#define JMP_BUF_CNT 1
#endif
typedef struct _label_t { long long unsigned val[JMP_BUF_CNT]; } label_t;
int ASMABI setjmp(label_t *) __attribute__ ((__nothrow__));
extern __attribute__((noreturn)) void ASMABI longjmp(label_t *);
#define LUAI_THROW(L,c) longjmp(&(c)->b)
#define LUAI_TRY(L,c,a) if (setjmp(&(c)->b) == 0) { a }
#define luai_jmpbuf label_t
/* unsupported arches will build but not be able to run lua programs */
#if JMP_BUF_CNT == 1
int setjmp (label_t *buf) {
return 1;
}
void longjmp (label_t * buf) {
for (;;);
}
#endif
#else
#define LUAI_THROW(L,c) longjmp((c)->b, 1)
#define LUAI_TRY(L,c,a) if (setjmp((c)->b) == 0) { a }
#define luai_jmpbuf jmp_buf
#endif
#else /* _KERNEL */
#if defined(__cplusplus) && !defined(LUA_USE_LONGJMP)
/* C++ exceptions */
#define LUAI_THROW(L,c) throw(c)
#define LUAI_TRY(L,c,a) \
try { a } catch(...) { if ((c)->status == 0) (c)->status = -1; }
#define luai_jmpbuf int /* dummy variable */
#elif defined(LUA_USE_ULONGJMP)
/* in Unix, try _longjmp/_setjmp (more efficient) */
#define LUAI_THROW(L,c) _longjmp((c)->b, 1)
#define LUAI_TRY(L,c,a) if (_setjmp((c)->b) == 0) { a }
#define luai_jmpbuf jmp_buf
#else
/* default handling with long jumps */
#define LUAI_THROW(L,c) longjmp((c)->b, 1)
#define LUAI_TRY(L,c,a) if (setjmp((c)->b) == 0) { a }
#define luai_jmpbuf jmp_buf
#endif
#endif /* _KERNEL */
#endif /* LUAI_THROW */
/* chain list of long jump buffers */
struct lua_longjmp {
struct lua_longjmp *previous;
luai_jmpbuf b;
volatile int status; /* error code */
};
static void seterrorobj (lua_State *L, int errcode, StkId oldtop) {
switch (errcode) {
case LUA_ERRMEM: { /* memory error? */
setsvalue2s(L, oldtop, G(L)->memerrmsg); /* reuse preregistered msg. */
break;
}
case LUA_ERRERR: {
setsvalue2s(L, oldtop, luaS_newliteral(L, "error in error handling"));
break;
}
default: {
setobjs2s(L, oldtop, L->top - 1); /* error message on current top */
break;
}
}
L->top = oldtop + 1;
}
/*
* Silence infinite recursion warning which was added to -Wall in gcc 12.1
*/
#if defined(__GNUC__) && !defined(__clang__) && \
defined(HAVE_KERNEL_INFINITE_RECURSION)
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Winfinite-recursion"
#endif
l_noret luaD_throw (lua_State *L, int errcode) {
if (L->errorJmp) { /* thread has an error handler? */
L->errorJmp->status = errcode; /* set status */
LUAI_THROW(L, L->errorJmp); /* jump to it */
}
else { /* thread has no error handler */
L->status = cast_byte(errcode); /* mark it as dead */
if (G(L)->mainthread->errorJmp) { /* main thread has a handler? */
setobjs2s(L, G(L)->mainthread->top++, L->top - 1); /* copy error obj. */
luaD_throw(G(L)->mainthread, errcode); /* re-throw in main thread */
}
else { /* no handler at all; abort */
if (G(L)->panic) { /* panic function? */
lua_unlock(L);
G(L)->panic(L); /* call it (last chance to jump out) */
}
panic("no error handler");
}
}
}
#if defined(__GNUC__) && !defined(__clang__) && \
defined(HAVE_INFINITE_RECURSION)
#pragma GCC diagnostic pop
#endif
int luaD_rawrunprotected (lua_State *L, Pfunc f, void *ud) {
unsigned short oldnCcalls = L->nCcalls;
struct lua_longjmp lj;
lj.status = LUA_OK;
lj.previous = L->errorJmp; /* chain new error handler */
L->errorJmp = &lj;
LUAI_TRY(L, &lj,
(*f)(L, ud);
);
L->errorJmp = lj.previous; /* restore old error handler */
L->nCcalls = oldnCcalls;
return lj.status;
}
/* }====================================================== */
static void correctstack (lua_State *L, TValue *oldstack) {
CallInfo *ci;
GCObject *up;
L->top = (L->top - oldstack) + L->stack;
for (up = L->openupval; up != NULL; up = up->gch.next)
gco2uv(up)->v = (gco2uv(up)->v - oldstack) + L->stack;
for (ci = L->ci; ci != NULL; ci = ci->previous) {
ci->top = (ci->top - oldstack) + L->stack;
ci->func = (ci->func - oldstack) + L->stack;
if (isLua(ci))
ci->u.l.base = (ci->u.l.base - oldstack) + L->stack;
}
}
/* some space for error handling */
#define ERRORSTACKSIZE (LUAI_MAXSTACK + 200)
void luaD_reallocstack (lua_State *L, int newsize) {
TValue *oldstack = L->stack;
int lim = L->stacksize;
lua_assert(newsize <= LUAI_MAXSTACK || newsize == ERRORSTACKSIZE);
lua_assert(L->stack_last - L->stack == L->stacksize - EXTRA_STACK);
luaM_reallocvector(L, L->stack, L->stacksize, newsize, TValue);
for (; lim < newsize; lim++)
setnilvalue(L->stack + lim); /* erase new segment */
L->stacksize = newsize;
L->stack_last = L->stack + newsize - EXTRA_STACK;
correctstack(L, oldstack);
}
void luaD_growstack (lua_State *L, int n) {
int size = L->stacksize;
if (size > LUAI_MAXSTACK) /* error after extra size? */
luaD_throw(L, LUA_ERRERR);
else {
int needed = cast_int(L->top - L->stack) + n + EXTRA_STACK;
int newsize = 2 * size;
if (newsize > LUAI_MAXSTACK) newsize = LUAI_MAXSTACK;
if (newsize < needed) newsize = needed;
if (newsize > LUAI_MAXSTACK) { /* stack overflow? */
luaD_reallocstack(L, ERRORSTACKSIZE);
luaG_runerror(L, "stack overflow");
}
else
luaD_reallocstack(L, newsize);
}
}
static int stackinuse (lua_State *L) {
CallInfo *ci;
StkId lim = L->top;
for (ci = L->ci; ci != NULL; ci = ci->previous) {
lua_assert(ci->top <= L->stack_last);
if (lim < ci->top) lim = ci->top;
}
return cast_int(lim - L->stack) + 1; /* part of stack in use */
}
void luaD_shrinkstack (lua_State *L) {
int inuse = stackinuse(L);
int goodsize = inuse + (inuse / 8) + 2*EXTRA_STACK;
if (goodsize > LUAI_MAXSTACK) goodsize = LUAI_MAXSTACK;
if (inuse > LUAI_MAXSTACK || /* handling stack overflow? */
goodsize >= L->stacksize) /* would grow instead of shrink? */
condmovestack(L); /* don't change stack (change only for debugging) */
else
luaD_reallocstack(L, goodsize); /* shrink it */
}
void luaD_hook (lua_State *L, int event, int line) {
lua_Hook hook = L->hook;
if (hook && L->allowhook) {
CallInfo *ci = L->ci;
ptrdiff_t top = savestack(L, L->top);
ptrdiff_t ci_top = savestack(L, ci->top);
lua_Debug ar;
ar.event = event;
ar.currentline = line;
ar.i_ci = ci;
luaD_checkstack(L, LUA_MINSTACK); /* ensure minimum stack size */
ci->top = L->top + LUA_MINSTACK;
lua_assert(ci->top <= L->stack_last);
L->allowhook = 0; /* cannot call hooks inside a hook */
ci->callstatus |= CIST_HOOKED;
lua_unlock(L);
(*hook)(L, &ar);
lua_lock(L);
lua_assert(!L->allowhook);
L->allowhook = 1;
ci->top = restorestack(L, ci_top);
L->top = restorestack(L, top);
ci->callstatus &= ~CIST_HOOKED;
}
}
static void callhook (lua_State *L, CallInfo *ci) {
int hook = LUA_HOOKCALL;
ci->u.l.savedpc++; /* hooks assume 'pc' is already incremented */
if (isLua(ci->previous) &&
GET_OPCODE(*(ci->previous->u.l.savedpc - 1)) == OP_TAILCALL) {
ci->callstatus |= CIST_TAIL;
hook = LUA_HOOKTAILCALL;
}
luaD_hook(L, hook, -1);
ci->u.l.savedpc--; /* correct 'pc' */
}
static StkId adjust_varargs (lua_State *L, Proto *p, int actual) {
int i;
int nfixargs = p->numparams;
StkId base, fixed;
lua_assert(actual >= nfixargs);
/* move fixed parameters to final position */
luaD_checkstack(L, p->maxstacksize); /* check again for new 'base' */
fixed = L->top - actual; /* first fixed argument */
base = L->top; /* final position of first argument */
for (i=0; i<nfixargs; i++) {
setobjs2s(L, L->top++, fixed + i);
setnilvalue(fixed + i);
}
return base;
}
static StkId tryfuncTM (lua_State *L, StkId func) {
const TValue *tm = luaT_gettmbyobj(L, func, TM_CALL);
StkId p;
ptrdiff_t funcr = savestack(L, func);
if (!ttisfunction(tm))
luaG_typeerror(L, func, "call");
/* Open a hole inside the stack at `func' */
for (p = L->top; p > func; p--) setobjs2s(L, p, p-1);
incr_top(L);
func = restorestack(L, funcr); /* previous call may change stack */
setobj2s(L, func, tm); /* tag method is the new function to be called */
return func;
}
#define next_ci(L) (L->ci = (L->ci->next ? L->ci->next : luaE_extendCI(L)))
/*
** returns true if function has been executed (C function)
*/
int luaD_precall (lua_State *L, StkId func, int nresults) {
lua_CFunction f;
CallInfo *ci;
int n; /* number of arguments (Lua) or returns (C) */
ptrdiff_t funcr = savestack(L, func);
switch (ttype(func)) {
case LUA_TLCF: /* light C function */
f = fvalue(func);
goto Cfunc;
case LUA_TCCL: { /* C closure */
f = clCvalue(func)->f;
Cfunc:
luaD_checkstack(L, LUA_MINSTACK); /* ensure minimum stack size */
ci = next_ci(L); /* now 'enter' new function */
ci->nresults = nresults;
ci->func = restorestack(L, funcr);
ci->top = L->top + LUA_MINSTACK;
lua_assert(ci->top <= L->stack_last);
ci->callstatus = 0;
luaC_checkGC(L); /* stack grow uses memory */
if (L->hookmask & LUA_MASKCALL)
luaD_hook(L, LUA_HOOKCALL, -1);
lua_unlock(L);
n = (*f)(L); /* do the actual call */
lua_lock(L);
api_checknelems(L, n);
luaD_poscall(L, L->top - n);
return 1;
}
case LUA_TLCL: { /* Lua function: prepare its call */
StkId base;
Proto *p = clLvalue(func)->p;
n = cast_int(L->top - func) - 1; /* number of real arguments */
luaD_checkstack(L, p->maxstacksize + p->numparams);
for (; n < p->numparams; n++)
setnilvalue(L->top++); /* complete missing arguments */
if (!p->is_vararg) {
func = restorestack(L, funcr);
base = func + 1;
}
else {
base = adjust_varargs(L, p, n);
func = restorestack(L, funcr); /* previous call can change stack */
}
ci = next_ci(L); /* now 'enter' new function */
ci->nresults = nresults;
ci->func = func;
ci->u.l.base = base;
ci->top = base + p->maxstacksize;
lua_assert(ci->top <= L->stack_last);
ci->u.l.savedpc = p->code; /* starting point */
ci->callstatus = CIST_LUA;
L->top = ci->top;
luaC_checkGC(L); /* stack grow uses memory */
if (L->hookmask & LUA_MASKCALL)
callhook(L, ci);
return 0;
}
default: { /* not a function */
func = tryfuncTM(L, func); /* retry with 'function' tag method */
return luaD_precall(L, func, nresults); /* now it must be a function */
}
}
}
int luaD_poscall (lua_State *L, StkId firstResult) {
StkId res;
int wanted, i;
CallInfo *ci = L->ci;
if (L->hookmask & (LUA_MASKRET | LUA_MASKLINE)) {
if (L->hookmask & LUA_MASKRET) {
ptrdiff_t fr = savestack(L, firstResult); /* hook may change stack */
luaD_hook(L, LUA_HOOKRET, -1);
firstResult = restorestack(L, fr);
}
L->oldpc = ci->previous->u.l.savedpc; /* 'oldpc' for caller function */
}
res = ci->func; /* res == final position of 1st result */
wanted = ci->nresults;
L->ci = ci->previous; /* back to caller */
/* move results to correct place */
for (i = wanted; i != 0 && firstResult < L->top; i--)
setobjs2s(L, res++, firstResult++);
while (i-- > 0)
setnilvalue(res++);
L->top = res;
return (wanted - LUA_MULTRET); /* 0 iff wanted == LUA_MULTRET */
}
/*
** Call a function (C or Lua). The function to be called is at *func.
** The arguments are on the stack, right after the function.
** When returns, all the results are on the stack, starting at the original
** function position.
*/
void luaD_call (lua_State *L, StkId func, int nResults, int allowyield) {
if (++L->nCcalls >= LUAI_MAXCCALLS) {
if (L->nCcalls == LUAI_MAXCCALLS)
luaG_runerror(L, "C stack overflow");
else if (L->nCcalls >= (LUAI_MAXCCALLS + (LUAI_MAXCCALLS>>3)))
luaD_throw(L, LUA_ERRERR); /* error while handling stack error */
}
intptr_t remaining = stack_remaining();
if (L->runerror == 0 && remaining < LUAI_MINCSTACK)
luaG_runerror(L, "C stack overflow");
if (L->runerror != 0 && remaining < LUAI_MINCSTACK / 2)
luaD_throw(L, LUA_ERRERR); /* error while handling stack error */
if (!allowyield) L->nny++;
if (!luaD_precall(L, func, nResults)) /* is a Lua function? */
luaV_execute(L); /* call it */
if (!allowyield) L->nny--;
L->nCcalls--;
}
static void finishCcall (lua_State *L) {
CallInfo *ci = L->ci;
int n;
lua_assert(ci->u.c.k != NULL); /* must have a continuation */
lua_assert(L->nny == 0);
if (ci->callstatus & CIST_YPCALL) { /* was inside a pcall? */
ci->callstatus &= ~CIST_YPCALL; /* finish 'lua_pcall' */
L->errfunc = ci->u.c.old_errfunc;
}
/* finish 'lua_callk'/'lua_pcall' */
adjustresults(L, ci->nresults);
/* call continuation function */
if (!(ci->callstatus & CIST_STAT)) /* no call status? */
ci->u.c.status = LUA_YIELD; /* 'default' status */
lua_assert(ci->u.c.status != LUA_OK);
ci->callstatus = (ci->callstatus & ~(CIST_YPCALL | CIST_STAT)) | CIST_YIELDED;
lua_unlock(L);
n = (*ci->u.c.k)(L);
lua_lock(L);
api_checknelems(L, n);
/* finish 'luaD_precall' */
luaD_poscall(L, L->top - n);
}
static void unroll (lua_State *L, void *ud) {
UNUSED(ud);
for (;;) {
if (L->ci == &L->base_ci) /* stack is empty? */
return; /* coroutine finished normally */
if (!isLua(L->ci)) /* C function? */
finishCcall(L);
else { /* Lua function */
luaV_finishOp(L); /* finish interrupted instruction */
luaV_execute(L); /* execute down to higher C 'boundary' */
}
}
}
/*
** check whether thread has a suspended protected call
*/
static CallInfo *findpcall (lua_State *L) {
CallInfo *ci;
for (ci = L->ci; ci != NULL; ci = ci->previous) { /* search for a pcall */
if (ci->callstatus & CIST_YPCALL)
return ci;
}
return NULL; /* no pending pcall */
}
static int recover (lua_State *L, int status) {
StkId oldtop;
CallInfo *ci = findpcall(L);
if (ci == NULL) return 0; /* no recovery point */
/* "finish" luaD_pcall */
oldtop = restorestack(L, ci->extra);
luaF_close(L, oldtop);
seterrorobj(L, status, oldtop);
L->ci = ci;
L->allowhook = ci->u.c.old_allowhook;
L->nny = 0; /* should be zero to be yieldable */
luaD_shrinkstack(L);
L->errfunc = ci->u.c.old_errfunc;
ci->callstatus |= CIST_STAT; /* call has error status */
ci->u.c.status = status; /* (here it is) */
return 1; /* continue running the coroutine */
}
/*
** signal an error in the call to 'resume', not in the execution of the
** coroutine itself. (Such errors should not be handled by any coroutine
** error handler and should not kill the coroutine.)
*/
static l_noret resume_error (lua_State *L, const char *msg, StkId firstArg) {
L->top = firstArg; /* remove args from the stack */
setsvalue2s(L, L->top, luaS_new(L, msg)); /* push error message */
api_incr_top(L);
luaD_throw(L, -1); /* jump back to 'lua_resume' */
}
/*
** do the work for 'lua_resume' in protected mode
*/
static void resume_cb (lua_State *L, void *ud) {
int nCcalls = L->nCcalls;
StkId firstArg = cast(StkId, ud);
CallInfo *ci = L->ci;
if (nCcalls >= LUAI_MAXCCALLS)
resume_error(L, "C stack overflow", firstArg);
if (L->status == LUA_OK) { /* may be starting a coroutine */
if (ci != &L->base_ci) /* not in base level? */
resume_error(L, "cannot resume non-suspended coroutine", firstArg);
/* coroutine is in base level; start running it */
if (!luaD_precall(L, firstArg - 1, LUA_MULTRET)) /* Lua function? */
luaV_execute(L); /* call it */
}
else if (L->status != LUA_YIELD)
resume_error(L, "cannot resume dead coroutine", firstArg);
else { /* resuming from previous yield */
L->status = LUA_OK;
ci->func = restorestack(L, ci->extra);
if (isLua(ci)) /* yielded inside a hook? */
luaV_execute(L); /* just continue running Lua code */
else { /* 'common' yield */
if (ci->u.c.k != NULL) { /* does it have a continuation? */
int n;
ci->u.c.status = LUA_YIELD; /* 'default' status */
ci->callstatus |= CIST_YIELDED;
lua_unlock(L);
n = (*ci->u.c.k)(L); /* call continuation */
lua_lock(L);
api_checknelems(L, n);
firstArg = L->top - n; /* yield results come from continuation */
}
luaD_poscall(L, firstArg); /* finish 'luaD_precall' */
}
unroll(L, NULL);
}
lua_assert(nCcalls == L->nCcalls);
}
LUA_API int lua_resume (lua_State *L, lua_State *from, int nargs) {
int status;
int oldnny = L->nny; /* save 'nny' */
lua_lock(L);
luai_userstateresume(L, nargs);
L->nCcalls = (from) ? from->nCcalls + 1 : 1;
L->nny = 0; /* allow yields */
api_checknelems(L, (L->status == LUA_OK) ? nargs + 1 : nargs);
status = luaD_rawrunprotected(L, resume_cb, L->top - nargs);
if (status == -1) /* error calling 'lua_resume'? */
status = LUA_ERRRUN;
else { /* yield or regular error */
while (status != LUA_OK && status != LUA_YIELD) { /* error? */
if (recover(L, status)) /* recover point? */
status = luaD_rawrunprotected(L, unroll, NULL); /* run continuation */
else { /* unrecoverable error */
L->status = cast_byte(status); /* mark thread as `dead' */
seterrorobj(L, status, L->top);
L->ci->top = L->top;
break;
}
}
lua_assert(status == L->status);
}
L->nny = oldnny; /* restore 'nny' */
L->nCcalls--;
lua_assert(L->nCcalls == ((from) ? from->nCcalls : 0));
lua_unlock(L);
return status;
}
LUA_API int lua_yieldk (lua_State *L, int nresults, int ctx, lua_CFunction k) {
CallInfo *ci = L->ci;
luai_userstateyield(L, nresults);
lua_lock(L);
api_checknelems(L, nresults);
if (L->nny > 0) {
if (L != G(L)->mainthread)
luaG_runerror(L, "attempt to yield across a C-call boundary");
else
luaG_runerror(L, "attempt to yield from outside a coroutine");
}
L->status = LUA_YIELD;
ci->extra = savestack(L, ci->func); /* save current 'func' */
if (isLua(ci)) { /* inside a hook? */
api_check(L, k == NULL, "hooks cannot continue after yielding");
}
else {
if ((ci->u.c.k = k) != NULL) /* is there a continuation? */
ci->u.c.ctx = ctx; /* save context */
ci->func = L->top - nresults - 1; /* protect stack below results */
luaD_throw(L, LUA_YIELD);
}
lua_assert(ci->callstatus & CIST_HOOKED); /* must be inside a hook */
lua_unlock(L);
return 0; /* return to 'luaD_hook' */
}
int luaD_pcall (lua_State *L, Pfunc func, void *u,
ptrdiff_t old_top, ptrdiff_t ef) {
int status;
CallInfo *old_ci = L->ci;
lu_byte old_allowhooks = L->allowhook;
unsigned short old_nny = L->nny;
ptrdiff_t old_errfunc = L->errfunc;
L->errfunc = ef;
status = luaD_rawrunprotected(L, func, u);
if (status != LUA_OK) { /* an error occurred? */
StkId oldtop = restorestack(L, old_top);
luaF_close(L, oldtop); /* close possible pending closures */
seterrorobj(L, status, oldtop);
L->ci = old_ci;
L->allowhook = old_allowhooks;
L->nny = old_nny;
luaD_shrinkstack(L);
}
L->errfunc = old_errfunc;
return status;
}
/*
** Execute a protected parser.
*/
struct SParser { /* data to `f_parser' */
ZIO *z;
Mbuffer buff; /* dynamic structure used by the scanner */
Dyndata dyd; /* dynamic structures used by the parser */
const char *mode;
const char *name;
};
static void checkmode (lua_State *L, const char *mode, const char *x) {
if (mode && strchr(mode, x[0]) == NULL) {
luaO_pushfstring(L,
"attempt to load a %s chunk (mode is " LUA_QS ")", x, mode);
luaD_throw(L, LUA_ERRSYNTAX);
}
}
static void f_parser (lua_State *L, void *ud) {
int i;
Closure *cl;
struct SParser *p = cast(struct SParser *, ud);
int c = zgetc(p->z); /* read first character */
lua_assert(c != LUA_SIGNATURE[0]); /* binary not supported */
checkmode(L, p->mode, "text");
cl = luaY_parser(L, p->z, &p->buff, &p->dyd, p->name, c);
lua_assert(cl->l.nupvalues == cl->l.p->sizeupvalues);
for (i = 0; i < cl->l.nupvalues; i++) { /* initialize upvalues */
UpVal *up = luaF_newupval(L);
cl->l.upvals[i] = up;
luaC_objbarrier(L, cl, up);
}
}
int luaD_protectedparser (lua_State *L, ZIO *z, const char *name,
const char *mode) {
struct SParser p;
int status;
L->nny++; /* cannot yield during parsing */
p.z = z; p.name = name; p.mode = mode;
p.dyd.actvar.arr = NULL; p.dyd.actvar.size = 0;
p.dyd.gt.arr = NULL; p.dyd.gt.size = 0;
p.dyd.label.arr = NULL; p.dyd.label.size = 0;
luaZ_initbuffer(L, &p.buff);
status = luaD_pcall(L, f_parser, &p, savestack(L, L->top), L->errfunc);
luaZ_freebuffer(L, &p.buff);
luaM_freearray(L, p.dyd.actvar.arr, p.dyd.actvar.size);
luaM_freearray(L, p.dyd.gt.arr, p.dyd.gt.size);
luaM_freearray(L, p.dyd.label.arr, p.dyd.label.size);
L->nny--;
return status;
}
diff --git a/sys/contrib/openzfs/module/lua/setjmp/setjmp.S b/sys/contrib/openzfs/module/lua/setjmp/setjmp.S
index 1f461a0a4ef3..6f03eea92711 100644
--- a/sys/contrib/openzfs/module/lua/setjmp/setjmp.S
+++ b/sys/contrib/openzfs/module/lua/setjmp/setjmp.S
@@ -1,19 +1,21 @@
#if defined(__x86_64__)
#include "setjmp_x86_64.S"
#elif defined(__i386__)
#include "setjmp_i386.S"
#elif defined(__aarch64__)
#include "setjmp_aarch64.S"
#elif defined(__arm__)
#include "setjmp_arm.S"
#elif defined(__sparc__) && defined(__arch64__)
#include "setjmp_sparc64.S"
#elif defined(__powerpc__)
#include "setjmp_ppc.S"
#elif defined(__mips__)
#include "setjmp_mips.S"
#elif defined(__s390x__)
#include "setjmp_s390x.S"
#elif defined(__riscv)
#include "setjmp_rv64g.S"
+#elif defined(__loongarch_lp64)
+#include "setjmp_loongarch64.S"
#endif
diff --git a/sys/contrib/openzfs/module/lua/setjmp/setjmp_loongarch64.S b/sys/contrib/openzfs/module/lua/setjmp/setjmp_loongarch64.S
new file mode 100644
index 000000000000..216b829ff236
--- /dev/null
+++ b/sys/contrib/openzfs/module/lua/setjmp/setjmp_loongarch64.S
@@ -0,0 +1,82 @@
+/*-
+ * Copyright 2022 Han Gao <gaohan@uniontech.com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#if __loongarch_lp64
+
+#define ENTRY(symbol) \
+ .text; \
+ .globl symbol; \
+ .align 3; \
+ .type symbol, @function; \
+ symbol:
+
+#define END(function) \
+ .size function, .- function;
+
+ENTRY(setjmp)
+ st.d $ra, $a0, 0*8
+ st.d $sp, $a0, 1*8
+ st.d $r21, $a0, 2*8
+ st.d $fp, $a0, 3*8
+ st.d $s0, $a0, 4*8
+ st.d $s1, $a0, 5*8
+ st.d $s2, $a0, 6*8
+ st.d $s3, $a0, 7*8
+ st.d $s4, $a0, 8*8
+ st.d $s5, $a0, 9*8
+ st.d $s6, $a0, 10*8
+ st.d $s7, $a0, 11*8
+ st.d $s8, $a0, 12*8
+
+ li.w $a0, 0
+ jr $ra
+END(setjmp)
+
+ENTRY(longjmp)
+ ld.d $ra, $a0, 0*8
+ ld.d $sp, $a0, 1*8
+ ld.d $r21, $a0, 2*8
+ ld.d $fp, $a0, 3*8
+ ld.d $s0, $a0, 4*8
+ ld.d $s1, $a0, 5*8
+ ld.d $s2, $a0, 6*8
+ ld.d $s3, $a0, 7*8
+ ld.d $s4, $a0, 8*8
+ ld.d $s5, $a0, 9*8
+ ld.d $s6, $a0, 10*8
+ ld.d $s7, $a0, 11*8
+ ld.d $s8, $a0, 12*8
+
+ sltui $a0, $a1, 1
+ add.d $a0, $a0, $a1 // a0 = (a1 == 0) ? 1 : a1
+ jr $ra
+END(longjmp)
+
+#ifdef __ELF__
+.section .note.GNU-stack,"",%progbits
+#endif
+
+#endif
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/sysctl_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/sysctl_os.c
index eccb91deff4f..cc616f33db96 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/sysctl_os.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/sysctl_os.c
@@ -1,904 +1,896 @@
/*
* Copyright (c) 2020 iXsystems, Inc.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/types.h>
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/conf.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/errno.h>
#include <sys/uio.h>
#include <sys/buf.h>
#include <sys/file.h>
#include <sys/kmem.h>
#include <sys/conf.h>
#include <sys/cmn_err.h>
#include <sys/stat.h>
#include <sys/zfs_ioctl.h>
#include <sys/zfs_vfsops.h>
#include <sys/zfs_znode.h>
#include <sys/zap.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
#include <sys/arc_os.h>
#include <sys/dmu.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_deleg.h>
#include <sys/dmu_objset.h>
#include <sys/dmu_impl.h>
#include <sys/dmu_tx.h>
#include <sys/sunddi.h>
#include <sys/policy.h>
#include <sys/zone.h>
#include <sys/nvpair.h>
#include <sys/mount.h>
#include <sys/taskqueue.h>
#include <sys/sdt.h>
#include <sys/fs/zfs.h>
#include <sys/zfs_ctldir.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_onexit.h>
#include <sys/zvol.h>
#include <sys/dsl_scan.h>
#include <sys/dmu_objset.h>
#include <sys/dmu_send.h>
#include <sys/dsl_destroy.h>
#include <sys/dsl_bookmark.h>
#include <sys/dsl_userhold.h>
#include <sys/zfeature.h>
#include <sys/zcp.h>
#include <sys/zio_checksum.h>
#include <sys/vdev_removal.h>
#include <sys/dsl_crypt.h>
#include <sys/zfs_ioctl_compat.h>
#include <sys/zfs_context.h>
#include <sys/arc_impl.h>
#include <sys/dsl_pool.h>
#include <sys/vmmeter.h>
SYSCTL_DECL(_vfs_zfs);
SYSCTL_NODE(_vfs_zfs, OID_AUTO, arc, CTLFLAG_RW, 0,
"ZFS adaptive replacement cache");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, brt, CTLFLAG_RW, 0,
"ZFS Block Reference Table");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, condense, CTLFLAG_RW, 0, "ZFS condense");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, dbuf, CTLFLAG_RW, 0, "ZFS disk buf cache");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, dbuf_cache, CTLFLAG_RW, 0,
"ZFS disk buf cache");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, deadman, CTLFLAG_RW, 0, "ZFS deadman");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, dedup, CTLFLAG_RW, 0, "ZFS dedup");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, l2arc, CTLFLAG_RW, 0, "ZFS l2arc");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, livelist, CTLFLAG_RW, 0, "ZFS livelist");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, lua, CTLFLAG_RW, 0, "ZFS lua");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, metaslab, CTLFLAG_RW, 0, "ZFS metaslab");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, mg, CTLFLAG_RW, 0, "ZFS metaslab group");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, multihost, CTLFLAG_RW, 0,
"ZFS multihost protection");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, prefetch, CTLFLAG_RW, 0, "ZFS prefetch");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, reconstruct, CTLFLAG_RW, 0, "ZFS reconstruct");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, recv, CTLFLAG_RW, 0, "ZFS receive");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, send, CTLFLAG_RW, 0, "ZFS send");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, spa, CTLFLAG_RW, 0, "ZFS space allocation");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, trim, CTLFLAG_RW, 0, "ZFS TRIM");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, txg, CTLFLAG_RW, 0, "ZFS transaction group");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, vdev, CTLFLAG_RW, 0, "ZFS VDEV");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, vnops, CTLFLAG_RW, 0, "ZFS VNOPS");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, zevent, CTLFLAG_RW, 0, "ZFS event");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, zil, CTLFLAG_RW, 0, "ZFS ZIL");
SYSCTL_NODE(_vfs_zfs, OID_AUTO, zio, CTLFLAG_RW, 0, "ZFS ZIO");
SYSCTL_NODE(_vfs_zfs_livelist, OID_AUTO, condense, CTLFLAG_RW, 0,
"ZFS livelist condense");
SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, cache, CTLFLAG_RW, 0, "ZFS VDEV Cache");
SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, file, CTLFLAG_RW, 0, "ZFS VDEV file");
SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, mirror, CTLFLAG_RD, 0,
"ZFS VDEV mirror");
SYSCTL_DECL(_vfs_zfs_version);
SYSCTL_CONST_STRING(_vfs_zfs_version, OID_AUTO, module, CTLFLAG_RD,
(ZFS_META_VERSION "-" ZFS_META_RELEASE), "OpenZFS module version");
/* arc.c */
int
param_set_arc_u64(SYSCTL_HANDLER_ARGS)
{
int err;
err = sysctl_handle_64(oidp, arg1, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
arc_tuning_update(B_TRUE);
return (0);
}
int
param_set_arc_int(SYSCTL_HANDLER_ARGS)
{
int err;
err = sysctl_handle_int(oidp, arg1, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
arc_tuning_update(B_TRUE);
return (0);
}
int
param_set_arc_max(SYSCTL_HANDLER_ARGS)
{
unsigned long val;
int err;
val = zfs_arc_max;
err = sysctl_handle_64(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (SET_ERROR(err));
if (val != 0 && (val < MIN_ARC_MAX || val <= arc_c_min ||
val >= arc_all_memory()))
return (SET_ERROR(EINVAL));
zfs_arc_max = val;
arc_tuning_update(B_TRUE);
/* Update the sysctl to the tuned value */
if (val != 0)
zfs_arc_max = arc_c_max;
return (0);
}
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_max,
CTLTYPE_ULONG | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
NULL, 0, param_set_arc_max, "LU",
"Maximum ARC size in bytes (LEGACY)");
/* END CSTYLED */
int
param_set_arc_min(SYSCTL_HANDLER_ARGS)
{
unsigned long val;
int err;
val = zfs_arc_min;
err = sysctl_handle_64(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (SET_ERROR(err));
if (val != 0 && (val < 2ULL << SPA_MAXBLOCKSHIFT || val > arc_c_max))
return (SET_ERROR(EINVAL));
zfs_arc_min = val;
arc_tuning_update(B_TRUE);
/* Update the sysctl to the tuned value */
if (val != 0)
zfs_arc_min = arc_c_min;
return (0);
}
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_min,
CTLTYPE_ULONG | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
NULL, 0, param_set_arc_min, "LU",
"Minimum ARC size in bytes (LEGACY)");
/* END CSTYLED */
extern uint_t zfs_arc_free_target;
int
param_set_arc_free_target(SYSCTL_HANDLER_ARGS)
{
uint_t val;
int err;
val = zfs_arc_free_target;
err = sysctl_handle_int(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
if (val < minfree)
return (EINVAL);
if (val > vm_cnt.v_page_count)
return (EINVAL);
zfs_arc_free_target = val;
return (0);
}
/*
* NOTE: This sysctl is CTLFLAG_RW not CTLFLAG_RWTUN due to its dependency on
* pagedaemon initialization.
*/
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_free_target,
CTLTYPE_UINT | CTLFLAG_RW | CTLFLAG_MPSAFE,
NULL, 0, param_set_arc_free_target, "IU",
"Desired number of free pages below which ARC triggers reclaim"
" (LEGACY)");
/* END CSTYLED */
int
param_set_arc_no_grow_shift(SYSCTL_HANDLER_ARGS)
{
int err, val;
val = arc_no_grow_shift;
err = sysctl_handle_int(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
if (val < 0 || val >= arc_shrink_shift)
return (EINVAL);
arc_no_grow_shift = val;
return (0);
}
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_no_grow_shift,
CTLTYPE_INT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
NULL, 0, param_set_arc_no_grow_shift, "I",
"log2(fraction of ARC which must be free to allow growing) (LEGACY)");
/* END CSTYLED */
extern uint64_t l2arc_write_max;
/* BEGIN CSTYLED */
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_max,
CTLFLAG_RWTUN, &l2arc_write_max, 0,
"Max write bytes per interval (LEGACY)");
/* END CSTYLED */
extern uint64_t l2arc_write_boost;
/* BEGIN CSTYLED */
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_boost,
CTLFLAG_RWTUN, &l2arc_write_boost, 0,
"Extra write bytes during device warmup (LEGACY)");
/* END CSTYLED */
extern uint64_t l2arc_headroom;
/* BEGIN CSTYLED */
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_headroom,
CTLFLAG_RWTUN, &l2arc_headroom, 0,
"Number of max device writes to precache (LEGACY)");
/* END CSTYLED */
extern uint64_t l2arc_headroom_boost;
/* BEGIN CSTYLED */
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_headroom_boost,
CTLFLAG_RWTUN, &l2arc_headroom_boost, 0,
"Compressed l2arc_headroom multiplier (LEGACY)");
/* END CSTYLED */
extern uint64_t l2arc_feed_secs;
/* BEGIN CSTYLED */
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_secs,
CTLFLAG_RWTUN, &l2arc_feed_secs, 0,
"Seconds between L2ARC writing (LEGACY)");
/* END CSTYLED */
extern uint64_t l2arc_feed_min_ms;
/* BEGIN CSTYLED */
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_min_ms,
CTLFLAG_RWTUN, &l2arc_feed_min_ms, 0,
"Min feed interval in milliseconds (LEGACY)");
/* END CSTYLED */
extern int l2arc_noprefetch;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_noprefetch,
CTLFLAG_RWTUN, &l2arc_noprefetch, 0,
"Skip caching prefetched buffers (LEGACY)");
/* END CSTYLED */
extern int l2arc_feed_again;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_feed_again,
CTLFLAG_RWTUN, &l2arc_feed_again, 0,
"Turbo L2ARC warmup (LEGACY)");
/* END CSTYLED */
extern int l2arc_norw;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_norw,
CTLFLAG_RWTUN, &l2arc_norw, 0,
"No reads during writes (LEGACY)");
/* END CSTYLED */
static int
param_get_arc_state_size(SYSCTL_HANDLER_ARGS)
{
arc_state_t *state = (arc_state_t *)arg1;
int64_t val;
val = zfs_refcount_count(&state->arcs_size[ARC_BUFC_DATA]) +
zfs_refcount_count(&state->arcs_size[ARC_BUFC_METADATA]);
return (sysctl_handle_64(oidp, &val, 0, req));
}
extern arc_state_t ARC_anon;
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, anon_size,
CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
&ARC_anon, 0, param_get_arc_state_size, "Q",
"size of anonymous state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_metadata_esize, CTLFLAG_RD,
&ARC_anon.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
"size of evictable metadata in anonymous state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_data_esize, CTLFLAG_RD,
&ARC_anon.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
"size of evictable data in anonymous state");
/* END CSTYLED */
extern arc_state_t ARC_mru;
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, mru_size,
CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
&ARC_mru, 0, param_get_arc_state_size, "Q",
"size of mru state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_metadata_esize, CTLFLAG_RD,
&ARC_mru.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
"size of evictable metadata in mru state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_data_esize, CTLFLAG_RD,
&ARC_mru.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
"size of evictable data in mru state");
/* END CSTYLED */
extern arc_state_t ARC_mru_ghost;
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, mru_ghost_size,
CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
&ARC_mru_ghost, 0, param_get_arc_state_size, "Q",
"size of mru ghost state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_metadata_esize, CTLFLAG_RD,
&ARC_mru_ghost.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
"size of evictable metadata in mru ghost state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_data_esize, CTLFLAG_RD,
&ARC_mru_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
"size of evictable data in mru ghost state");
/* END CSTYLED */
extern arc_state_t ARC_mfu;
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, mfu_size,
CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
&ARC_mfu, 0, param_get_arc_state_size, "Q",
"size of mfu state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_metadata_esize, CTLFLAG_RD,
&ARC_mfu.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
"size of evictable metadata in mfu state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_data_esize, CTLFLAG_RD,
&ARC_mfu.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
"size of evictable data in mfu state");
/* END CSTYLED */
extern arc_state_t ARC_mfu_ghost;
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, mfu_ghost_size,
CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
&ARC_mfu_ghost, 0, param_get_arc_state_size, "Q",
"size of mfu ghost state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_metadata_esize, CTLFLAG_RD,
&ARC_mfu_ghost.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
"size of evictable metadata in mfu ghost state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_data_esize, CTLFLAG_RD,
&ARC_mfu_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
"size of evictable data in mfu ghost state");
/* END CSTYLED */
extern arc_state_t ARC_uncached;
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, uncached_size,
CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
&ARC_uncached, 0, param_get_arc_state_size, "Q",
"size of uncached state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, uncached_metadata_esize, CTLFLAG_RD,
&ARC_uncached.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
"size of evictable metadata in uncached state");
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, uncached_data_esize, CTLFLAG_RD,
&ARC_uncached.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
"size of evictable data in uncached state");
/* END CSTYLED */
extern arc_state_t ARC_l2c_only;
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, l2c_only_size,
CTLTYPE_S64 | CTLFLAG_RD | CTLFLAG_MPSAFE,
&ARC_l2c_only, 0, param_get_arc_state_size, "Q",
"size of l2c_only state");
/* END CSTYLED */
/* dbuf.c */
/* dmu.c */
/* dmu_zfetch.c */
SYSCTL_NODE(_vfs_zfs, OID_AUTO, zfetch, CTLFLAG_RW, 0, "ZFS ZFETCH (LEGACY)");
extern uint32_t zfetch_max_distance;
/* BEGIN CSTYLED */
SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_distance,
CTLFLAG_RWTUN, &zfetch_max_distance, 0,
"Max bytes to prefetch per stream (LEGACY)");
/* END CSTYLED */
extern uint32_t zfetch_max_idistance;
/* BEGIN CSTYLED */
SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_idistance,
CTLFLAG_RWTUN, &zfetch_max_idistance, 0,
"Max bytes to prefetch indirects for per stream (LEGACY)");
/* END CSTYLED */
/* dsl_pool.c */
/* dnode.c */
/* dsl_scan.c */
/* metaslab.c */
/*
* In pools where the log space map feature is not enabled we touch
* multiple metaslabs (and their respective space maps) with each
* transaction group. Thus, we benefit from having a small space map
* block size since it allows us to issue more I/O operations scattered
* around the disk. So a sane default for the space map block size
* is 8~16K.
*/
extern int zfs_metaslab_sm_blksz_no_log;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_no_log,
CTLFLAG_RDTUN, &zfs_metaslab_sm_blksz_no_log, 0,
"Block size for space map in pools with log space map disabled. "
"Power of 2 greater than 4096.");
/* END CSTYLED */
/*
* When the log space map feature is enabled, we accumulate a lot of
* changes per metaslab that are flushed once in a while so we benefit
* from a bigger block size like 128K for the metaslab space maps.
*/
extern int zfs_metaslab_sm_blksz_with_log;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_with_log,
CTLFLAG_RDTUN, &zfs_metaslab_sm_blksz_with_log, 0,
"Block size for space map in pools with log space map enabled. "
"Power of 2 greater than 4096.");
/* END CSTYLED */
/*
* The in-core space map representation is more compact than its on-disk form.
* The zfs_condense_pct determines how much more compact the in-core
* space map representation must be before we compact it on-disk.
* Values should be greater than or equal to 100.
*/
extern uint_t zfs_condense_pct;
/* BEGIN CSTYLED */
SYSCTL_UINT(_vfs_zfs, OID_AUTO, condense_pct,
CTLFLAG_RWTUN, &zfs_condense_pct, 0,
"Condense on-disk spacemap when it is more than this many percents"
" of in-memory counterpart");
/* END CSTYLED */
extern uint_t zfs_remove_max_segment;
/* BEGIN CSTYLED */
SYSCTL_UINT(_vfs_zfs, OID_AUTO, remove_max_segment,
CTLFLAG_RWTUN, &zfs_remove_max_segment, 0,
"Largest contiguous segment ZFS will attempt to allocate when removing"
" a device");
/* END CSTYLED */
extern int zfs_removal_suspend_progress;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, removal_suspend_progress,
CTLFLAG_RWTUN, &zfs_removal_suspend_progress, 0,
"Ensures certain actions can happen while in the middle of a removal");
/* END CSTYLED */
/*
* Minimum size which forces the dynamic allocator to change
* it's allocation strategy. Once the space map cannot satisfy
* an allocation of this size then it switches to using more
* aggressive strategy (i.e search by size rather than offset).
*/
extern uint64_t metaslab_df_alloc_threshold;
/* BEGIN CSTYLED */
SYSCTL_QUAD(_vfs_zfs_metaslab, OID_AUTO, df_alloc_threshold,
CTLFLAG_RWTUN, &metaslab_df_alloc_threshold, 0,
"Minimum size which forces the dynamic allocator to change its"
" allocation strategy");
/* END CSTYLED */
/*
* The minimum free space, in percent, which must be available
* in a space map to continue allocations in a first-fit fashion.
* Once the space map's free space drops below this level we dynamically
* switch to using best-fit allocations.
*/
extern uint_t metaslab_df_free_pct;
/* BEGIN CSTYLED */
SYSCTL_UINT(_vfs_zfs_metaslab, OID_AUTO, df_free_pct,
CTLFLAG_RWTUN, &metaslab_df_free_pct, 0,
"The minimum free space, in percent, which must be available in a"
" space map to continue allocations in a first-fit fashion");
/* END CSTYLED */
/*
* Percentage of all cpus that can be used by the metaslab taskq.
*/
extern int metaslab_load_pct;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, load_pct,
CTLFLAG_RWTUN, &metaslab_load_pct, 0,
"Percentage of cpus that can be used by the metaslab taskq");
/* END CSTYLED */
/*
* Max number of metaslabs per group to preload.
*/
extern uint_t metaslab_preload_limit;
/* BEGIN CSTYLED */
SYSCTL_UINT(_vfs_zfs_metaslab, OID_AUTO, preload_limit,
CTLFLAG_RWTUN, &metaslab_preload_limit, 0,
"Max number of metaslabs per group to preload");
/* END CSTYLED */
/* mmp.c */
int
param_set_multihost_interval(SYSCTL_HANDLER_ARGS)
{
int err;
err = sysctl_handle_64(oidp, &zfs_multihost_interval, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
if (spa_mode_global != SPA_MODE_UNINIT)
mmp_signal_all_threads();
return (0);
}
/* spa.c */
extern int zfs_ccw_retry_interval;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, ccw_retry_interval,
CTLFLAG_RWTUN, &zfs_ccw_retry_interval, 0,
"Configuration cache file write, retry after failure, interval"
" (seconds)");
/* END CSTYLED */
extern uint64_t zfs_max_missing_tvds_cachefile;
/* BEGIN CSTYLED */
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, max_missing_tvds_cachefile,
CTLFLAG_RWTUN, &zfs_max_missing_tvds_cachefile, 0,
"Allow importing pools with missing top-level vdevs in cache file");
/* END CSTYLED */
extern uint64_t zfs_max_missing_tvds_scan;
/* BEGIN CSTYLED */
SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, max_missing_tvds_scan,
CTLFLAG_RWTUN, &zfs_max_missing_tvds_scan, 0,
"Allow importing pools with missing top-level vdevs during scan");
/* END CSTYLED */
/* spa_misc.c */
extern int zfs_flags;
static int
sysctl_vfs_zfs_debug_flags(SYSCTL_HANDLER_ARGS)
{
int err, val;
val = zfs_flags;
err = sysctl_handle_int(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
/*
* ZFS_DEBUG_MODIFY must be enabled prior to boot so all
* arc buffers in the system have the necessary additional
* checksum data. However, it is safe to disable at any
* time.
*/
if (!(zfs_flags & ZFS_DEBUG_MODIFY))
val &= ~ZFS_DEBUG_MODIFY;
zfs_flags = val;
return (0);
}
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, debugflags,
CTLTYPE_UINT | CTLFLAG_MPSAFE | CTLFLAG_RWTUN, NULL, 0,
sysctl_vfs_zfs_debug_flags, "IU", "Debug flags for ZFS testing.");
/* END CSTYLED */
int
param_set_deadman_synctime(SYSCTL_HANDLER_ARGS)
{
unsigned long val;
int err;
val = zfs_deadman_synctime_ms;
err = sysctl_handle_64(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
zfs_deadman_synctime_ms = val;
spa_set_deadman_synctime(MSEC2NSEC(zfs_deadman_synctime_ms));
return (0);
}
int
param_set_deadman_ziotime(SYSCTL_HANDLER_ARGS)
{
unsigned long val;
int err;
val = zfs_deadman_ziotime_ms;
err = sysctl_handle_64(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
zfs_deadman_ziotime_ms = val;
spa_set_deadman_ziotime(MSEC2NSEC(zfs_deadman_synctime_ms));
return (0);
}
int
param_set_deadman_failmode(SYSCTL_HANDLER_ARGS)
{
char buf[16];
int rc;
if (req->newptr == NULL)
strlcpy(buf, zfs_deadman_failmode, sizeof (buf));
rc = sysctl_handle_string(oidp, buf, sizeof (buf), req);
if (rc || req->newptr == NULL)
return (rc);
if (strcmp(buf, zfs_deadman_failmode) == 0)
return (0);
if (strcmp(buf, "wait") == 0)
zfs_deadman_failmode = "wait";
if (strcmp(buf, "continue") == 0)
zfs_deadman_failmode = "continue";
if (strcmp(buf, "panic") == 0)
zfs_deadman_failmode = "panic";
return (-param_set_deadman_failmode_common(buf));
}
int
param_set_slop_shift(SYSCTL_HANDLER_ARGS)
{
int val;
int err;
val = spa_slop_shift;
err = sysctl_handle_int(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (err);
if (val < 1 || val > 31)
return (EINVAL);
spa_slop_shift = val;
return (0);
}
/* spacemap.c */
extern int space_map_ibs;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, space_map_ibs, CTLFLAG_RWTUN,
&space_map_ibs, 0, "Space map indirect block shift");
/* END CSTYLED */
/* vdev.c */
int
param_set_min_auto_ashift(SYSCTL_HANDLER_ARGS)
{
int val;
int err;
val = zfs_vdev_min_auto_ashift;
err = sysctl_handle_int(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (SET_ERROR(err));
if (val < ASHIFT_MIN || val > zfs_vdev_max_auto_ashift)
return (SET_ERROR(EINVAL));
zfs_vdev_min_auto_ashift = val;
return (0);
}
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift,
CTLTYPE_UINT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
&zfs_vdev_min_auto_ashift, sizeof (zfs_vdev_min_auto_ashift),
param_set_min_auto_ashift, "IU",
"Min ashift used when creating new top-level vdev. (LEGACY)");
/* END CSTYLED */
int
param_set_max_auto_ashift(SYSCTL_HANDLER_ARGS)
{
int val;
int err;
val = zfs_vdev_max_auto_ashift;
err = sysctl_handle_int(oidp, &val, 0, req);
if (err != 0 || req->newptr == NULL)
return (SET_ERROR(err));
if (val > ASHIFT_MAX || val < zfs_vdev_min_auto_ashift)
return (SET_ERROR(EINVAL));
zfs_vdev_max_auto_ashift = val;
return (0);
}
/* BEGIN CSTYLED */
SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift,
CTLTYPE_UINT | CTLFLAG_RWTUN | CTLFLAG_MPSAFE,
&zfs_vdev_max_auto_ashift, sizeof (zfs_vdev_max_auto_ashift),
param_set_max_auto_ashift, "IU",
"Max ashift used when optimizing for logical -> physical sector size on"
" new top-level vdevs. (LEGACY)");
/* END CSTYLED */
/*
* Since the DTL space map of a vdev is not expected to have a lot of
* entries, we default its block size to 4K.
*/
extern int zfs_vdev_dtl_sm_blksz;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, dtl_sm_blksz,
CTLFLAG_RDTUN, &zfs_vdev_dtl_sm_blksz, 0,
"Block size for DTL space map. Power of 2 greater than 4096.");
/* END CSTYLED */
/*
* vdev-wide space maps that have lots of entries written to them at
* the end of each transaction can benefit from a higher I/O bandwidth
* (e.g. vdev_obsolete_sm), thus we default their block size to 128K.
*/
extern int zfs_vdev_standard_sm_blksz;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, standard_sm_blksz,
CTLFLAG_RDTUN, &zfs_vdev_standard_sm_blksz, 0,
"Block size for standard space map. Power of 2 greater than 4096.");
/* END CSTYLED */
extern int vdev_validate_skip;
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs, OID_AUTO, validate_skip,
CTLFLAG_RDTUN, &vdev_validate_skip, 0,
"Enable to bypass vdev_validate().");
/* END CSTYLED */
/* vdev_cache.c */
/* vdev_mirror.c */
/* vdev_queue.c */
extern uint_t zfs_vdev_max_active;
/* BEGIN CSTYLED */
SYSCTL_UINT(_vfs_zfs, OID_AUTO, top_maxinflight,
CTLFLAG_RWTUN, &zfs_vdev_max_active, 0,
"The maximum number of I/Os of all types active for each device."
" (LEGACY)");
/* END CSTYLED */
-extern uint_t zfs_vdev_def_queue_depth;
-
-/* BEGIN CSTYLED */
-SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, def_queue_depth,
- CTLFLAG_RWTUN, &zfs_vdev_def_queue_depth, 0,
- "Default queue depth for each allocator");
-/* END CSTYLED */
-
/* zio.c */
/* BEGIN CSTYLED */
SYSCTL_INT(_vfs_zfs_zio, OID_AUTO, exclude_metadata,
CTLFLAG_RDTUN, &zio_exclude_metadata, 0,
"Exclude metadata buffers from dumps as well");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c
index 9f735dbb558c..a077076927a1 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_acl.c
@@ -1,2677 +1,2677 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013 by Delphix. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc. All rights reserved.
*/
#include <sys/types.h>
#include <sys/param.h>
#include <sys/time.h>
#include <sys/systm.h>
#include <sys/sysmacros.h>
#include <sys/resource.h>
#include <sys/vfs.h>
#include <sys/vnode.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/kmem.h>
#include <sys/cmn_err.h>
#include <sys/errno.h>
#include <sys/unistd.h>
#include <sys/sdt.h>
#include <sys/fs/zfs.h>
#include <sys/policy.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_fuid.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_quota.h>
#include <sys/zfs_vfsops.h>
#include <sys/dmu.h>
#include <sys/dnode.h>
#include <sys/zap.h>
#include <sys/sa.h>
#include <acl/acl_common.h>
#define ALLOW ACE_ACCESS_ALLOWED_ACE_TYPE
#define DENY ACE_ACCESS_DENIED_ACE_TYPE
#define MAX_ACE_TYPE ACE_SYSTEM_ALARM_CALLBACK_OBJECT_ACE_TYPE
#define MIN_ACE_TYPE ALLOW
#define OWNING_GROUP (ACE_GROUP|ACE_IDENTIFIER_GROUP)
#define EVERYONE_ALLOW_MASK (ACE_READ_ACL|ACE_READ_ATTRIBUTES | \
ACE_READ_NAMED_ATTRS|ACE_SYNCHRONIZE)
#define EVERYONE_DENY_MASK (ACE_WRITE_ACL|ACE_WRITE_OWNER | \
ACE_WRITE_ATTRIBUTES|ACE_WRITE_NAMED_ATTRS)
#define OWNER_ALLOW_MASK (ACE_WRITE_ACL | ACE_WRITE_OWNER | \
ACE_WRITE_ATTRIBUTES|ACE_WRITE_NAMED_ATTRS)
#define ZFS_CHECKED_MASKS (ACE_READ_ACL|ACE_READ_ATTRIBUTES|ACE_READ_DATA| \
ACE_READ_NAMED_ATTRS|ACE_WRITE_DATA|ACE_WRITE_ATTRIBUTES| \
ACE_WRITE_NAMED_ATTRS|ACE_APPEND_DATA|ACE_EXECUTE|ACE_WRITE_OWNER| \
ACE_WRITE_ACL|ACE_DELETE|ACE_DELETE_CHILD|ACE_SYNCHRONIZE)
#define WRITE_MASK_DATA (ACE_WRITE_DATA|ACE_APPEND_DATA|ACE_WRITE_NAMED_ATTRS)
#define WRITE_MASK_ATTRS (ACE_WRITE_ACL|ACE_WRITE_OWNER|ACE_WRITE_ATTRIBUTES| \
ACE_DELETE|ACE_DELETE_CHILD)
#define WRITE_MASK (WRITE_MASK_DATA|WRITE_MASK_ATTRS)
#define OGE_CLEAR (ACE_READ_DATA|ACE_LIST_DIRECTORY|ACE_WRITE_DATA| \
ACE_ADD_FILE|ACE_APPEND_DATA|ACE_ADD_SUBDIRECTORY|ACE_EXECUTE)
#define OKAY_MASK_BITS (ACE_READ_DATA|ACE_LIST_DIRECTORY|ACE_WRITE_DATA| \
ACE_ADD_FILE|ACE_APPEND_DATA|ACE_ADD_SUBDIRECTORY|ACE_EXECUTE)
#define ALL_INHERIT (ACE_FILE_INHERIT_ACE|ACE_DIRECTORY_INHERIT_ACE | \
ACE_NO_PROPAGATE_INHERIT_ACE|ACE_INHERIT_ONLY_ACE|ACE_INHERITED_ACE)
#define RESTRICTED_CLEAR (ACE_WRITE_ACL|ACE_WRITE_OWNER)
#define V4_ACL_WIDE_FLAGS (ZFS_ACL_AUTO_INHERIT|ZFS_ACL_DEFAULTED|\
ZFS_ACL_PROTECTED)
#define ZFS_ACL_WIDE_FLAGS (V4_ACL_WIDE_FLAGS|ZFS_ACL_TRIVIAL|ZFS_INHERIT_ACE|\
ZFS_ACL_OBJ_ACE)
#define ALL_MODE_EXECS (S_IXUSR | S_IXGRP | S_IXOTH)
static uint16_t
zfs_ace_v0_get_type(void *acep)
{
return (((zfs_oldace_t *)acep)->z_type);
}
static uint16_t
zfs_ace_v0_get_flags(void *acep)
{
return (((zfs_oldace_t *)acep)->z_flags);
}
static uint32_t
zfs_ace_v0_get_mask(void *acep)
{
return (((zfs_oldace_t *)acep)->z_access_mask);
}
static uint64_t
zfs_ace_v0_get_who(void *acep)
{
return (((zfs_oldace_t *)acep)->z_fuid);
}
static void
zfs_ace_v0_set_type(void *acep, uint16_t type)
{
((zfs_oldace_t *)acep)->z_type = type;
}
static void
zfs_ace_v0_set_flags(void *acep, uint16_t flags)
{
((zfs_oldace_t *)acep)->z_flags = flags;
}
static void
zfs_ace_v0_set_mask(void *acep, uint32_t mask)
{
((zfs_oldace_t *)acep)->z_access_mask = mask;
}
static void
zfs_ace_v0_set_who(void *acep, uint64_t who)
{
((zfs_oldace_t *)acep)->z_fuid = who;
}
static size_t
zfs_ace_v0_size(void *acep)
{
(void) acep;
return (sizeof (zfs_oldace_t));
}
static size_t
zfs_ace_v0_abstract_size(void)
{
return (sizeof (zfs_oldace_t));
}
static int
zfs_ace_v0_mask_off(void)
{
return (offsetof(zfs_oldace_t, z_access_mask));
}
static int
zfs_ace_v0_data(void *acep, void **datap)
{
(void) acep;
*datap = NULL;
return (0);
}
static const acl_ops_t zfs_acl_v0_ops = {
zfs_ace_v0_get_mask,
zfs_ace_v0_set_mask,
zfs_ace_v0_get_flags,
zfs_ace_v0_set_flags,
zfs_ace_v0_get_type,
zfs_ace_v0_set_type,
zfs_ace_v0_get_who,
zfs_ace_v0_set_who,
zfs_ace_v0_size,
zfs_ace_v0_abstract_size,
zfs_ace_v0_mask_off,
zfs_ace_v0_data
};
static uint16_t
zfs_ace_fuid_get_type(void *acep)
{
return (((zfs_ace_hdr_t *)acep)->z_type);
}
static uint16_t
zfs_ace_fuid_get_flags(void *acep)
{
return (((zfs_ace_hdr_t *)acep)->z_flags);
}
static uint32_t
zfs_ace_fuid_get_mask(void *acep)
{
return (((zfs_ace_hdr_t *)acep)->z_access_mask);
}
static uint64_t
zfs_ace_fuid_get_who(void *args)
{
uint16_t entry_type;
zfs_ace_t *acep = args;
entry_type = acep->z_hdr.z_flags & ACE_TYPE_FLAGS;
if (entry_type == ACE_OWNER || entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE)
return (-1);
return (((zfs_ace_t *)acep)->z_fuid);
}
static void
zfs_ace_fuid_set_type(void *acep, uint16_t type)
{
((zfs_ace_hdr_t *)acep)->z_type = type;
}
static void
zfs_ace_fuid_set_flags(void *acep, uint16_t flags)
{
((zfs_ace_hdr_t *)acep)->z_flags = flags;
}
static void
zfs_ace_fuid_set_mask(void *acep, uint32_t mask)
{
((zfs_ace_hdr_t *)acep)->z_access_mask = mask;
}
static void
zfs_ace_fuid_set_who(void *arg, uint64_t who)
{
zfs_ace_t *acep = arg;
uint16_t entry_type = acep->z_hdr.z_flags & ACE_TYPE_FLAGS;
if (entry_type == ACE_OWNER || entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE)
return;
acep->z_fuid = who;
}
static size_t
zfs_ace_fuid_size(void *acep)
{
zfs_ace_hdr_t *zacep = acep;
uint16_t entry_type;
switch (zacep->z_type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
return (sizeof (zfs_object_ace_t));
case ALLOW:
case DENY:
entry_type =
(((zfs_ace_hdr_t *)acep)->z_flags & ACE_TYPE_FLAGS);
if (entry_type == ACE_OWNER ||
entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE)
return (sizeof (zfs_ace_hdr_t));
zfs_fallthrough;
default:
return (sizeof (zfs_ace_t));
}
}
static size_t
zfs_ace_fuid_abstract_size(void)
{
return (sizeof (zfs_ace_hdr_t));
}
static int
zfs_ace_fuid_mask_off(void)
{
return (offsetof(zfs_ace_hdr_t, z_access_mask));
}
static int
zfs_ace_fuid_data(void *acep, void **datap)
{
zfs_ace_t *zacep = acep;
zfs_object_ace_t *zobjp;
switch (zacep->z_hdr.z_type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
zobjp = acep;
*datap = (caddr_t)zobjp + sizeof (zfs_ace_t);
return (sizeof (zfs_object_ace_t) - sizeof (zfs_ace_t));
default:
*datap = NULL;
return (0);
}
}
static const acl_ops_t zfs_acl_fuid_ops = {
zfs_ace_fuid_get_mask,
zfs_ace_fuid_set_mask,
zfs_ace_fuid_get_flags,
zfs_ace_fuid_set_flags,
zfs_ace_fuid_get_type,
zfs_ace_fuid_set_type,
zfs_ace_fuid_get_who,
zfs_ace_fuid_set_who,
zfs_ace_fuid_size,
zfs_ace_fuid_abstract_size,
zfs_ace_fuid_mask_off,
zfs_ace_fuid_data
};
/*
* The following three functions are provided for compatibility with
* older ZPL version in order to determine if the file use to have
* an external ACL and what version of ACL previously existed on the
* file. Would really be nice to not need this, sigh.
*/
uint64_t
zfs_external_acl(znode_t *zp)
{
zfs_acl_phys_t acl_phys;
int error;
if (zp->z_is_sa)
return (0);
/*
* Need to deal with a potential
* race where zfs_sa_upgrade could cause
* z_isa_sa to change.
*
* If the lookup fails then the state of z_is_sa should have
* changed.
*/
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_ZNODE_ACL(zp->z_zfsvfs),
&acl_phys, sizeof (acl_phys))) == 0)
return (acl_phys.z_acl_extern_obj);
else {
/*
* after upgrade the SA_ZPL_ZNODE_ACL should have been
* removed
*/
VERIFY(zp->z_is_sa);
VERIFY3S(error, ==, ENOENT);
return (0);
}
}
/*
* Determine size of ACL in bytes
*
* This is more complicated than it should be since we have to deal
* with old external ACLs.
*/
static int
zfs_acl_znode_info(znode_t *zp, int *aclsize, int *aclcount,
zfs_acl_phys_t *aclphys)
{
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
uint64_t acl_count;
int size;
int error;
ASSERT(MUTEX_HELD(&zp->z_acl_lock));
if (zp->z_is_sa) {
if ((error = sa_size(zp->z_sa_hdl, SA_ZPL_DACL_ACES(zfsvfs),
&size)) != 0)
return (error);
*aclsize = size;
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_DACL_COUNT(zfsvfs),
&acl_count, sizeof (acl_count))) != 0)
return (error);
*aclcount = acl_count;
} else {
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_ZNODE_ACL(zfsvfs),
aclphys, sizeof (*aclphys))) != 0)
return (error);
if (aclphys->z_acl_version == ZFS_ACL_VERSION_INITIAL) {
*aclsize = ZFS_ACL_SIZE(aclphys->z_acl_size);
*aclcount = aclphys->z_acl_size;
} else {
*aclsize = aclphys->z_acl_size;
*aclcount = aclphys->z_acl_count;
}
}
return (0);
}
int
zfs_znode_acl_version(znode_t *zp)
{
zfs_acl_phys_t acl_phys;
if (zp->z_is_sa)
return (ZFS_ACL_VERSION_FUID);
else {
int error;
/*
* Need to deal with a potential
* race where zfs_sa_upgrade could cause
* z_isa_sa to change.
*
* If the lookup fails then the state of z_is_sa should have
* changed.
*/
if ((error = sa_lookup(zp->z_sa_hdl,
SA_ZPL_ZNODE_ACL(zp->z_zfsvfs),
&acl_phys, sizeof (acl_phys))) == 0)
return (acl_phys.z_acl_version);
else {
/*
* After upgrade SA_ZPL_ZNODE_ACL should have
* been removed.
*/
VERIFY(zp->z_is_sa);
VERIFY3S(error, ==, ENOENT);
return (ZFS_ACL_VERSION_FUID);
}
}
}
static int
zfs_acl_version(int version)
{
if (version < ZPL_VERSION_FUID)
return (ZFS_ACL_VERSION_INITIAL);
else
return (ZFS_ACL_VERSION_FUID);
}
static int
zfs_acl_version_zp(znode_t *zp)
{
return (zfs_acl_version(zp->z_zfsvfs->z_version));
}
zfs_acl_t *
zfs_acl_alloc(int vers)
{
zfs_acl_t *aclp;
aclp = kmem_zalloc(sizeof (zfs_acl_t), KM_SLEEP);
list_create(&aclp->z_acl, sizeof (zfs_acl_node_t),
offsetof(zfs_acl_node_t, z_next));
aclp->z_version = vers;
if (vers == ZFS_ACL_VERSION_FUID)
aclp->z_ops = &zfs_acl_fuid_ops;
else
aclp->z_ops = &zfs_acl_v0_ops;
return (aclp);
}
zfs_acl_node_t *
zfs_acl_node_alloc(size_t bytes)
{
zfs_acl_node_t *aclnode;
aclnode = kmem_zalloc(sizeof (zfs_acl_node_t), KM_SLEEP);
if (bytes) {
aclnode->z_acldata = kmem_alloc(bytes, KM_SLEEP);
aclnode->z_allocdata = aclnode->z_acldata;
aclnode->z_allocsize = bytes;
aclnode->z_size = bytes;
}
return (aclnode);
}
static void
zfs_acl_node_free(zfs_acl_node_t *aclnode)
{
if (aclnode->z_allocsize)
kmem_free(aclnode->z_allocdata, aclnode->z_allocsize);
kmem_free(aclnode, sizeof (zfs_acl_node_t));
}
static void
zfs_acl_release_nodes(zfs_acl_t *aclp)
{
zfs_acl_node_t *aclnode;
while ((aclnode = list_head(&aclp->z_acl))) {
list_remove(&aclp->z_acl, aclnode);
zfs_acl_node_free(aclnode);
}
aclp->z_acl_count = 0;
aclp->z_acl_bytes = 0;
}
void
zfs_acl_free(zfs_acl_t *aclp)
{
zfs_acl_release_nodes(aclp);
list_destroy(&aclp->z_acl);
kmem_free(aclp, sizeof (zfs_acl_t));
}
static boolean_t
zfs_acl_valid_ace_type(uint_t type, uint_t flags)
{
uint16_t entry_type;
switch (type) {
case ALLOW:
case DENY:
case ACE_SYSTEM_AUDIT_ACE_TYPE:
case ACE_SYSTEM_ALARM_ACE_TYPE:
entry_type = flags & ACE_TYPE_FLAGS;
return (entry_type == ACE_OWNER ||
entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE || entry_type == 0 ||
entry_type == ACE_IDENTIFIER_GROUP);
default:
if (type <= MAX_ACE_TYPE)
return (B_TRUE);
}
return (B_FALSE);
}
static boolean_t
zfs_ace_valid(vtype_t obj_type, zfs_acl_t *aclp, uint16_t type, uint16_t iflags)
{
/*
* first check type of entry
*/
if (!zfs_acl_valid_ace_type(type, iflags))
return (B_FALSE);
switch (type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
if (aclp->z_version < ZFS_ACL_VERSION_FUID)
return (B_FALSE);
aclp->z_hints |= ZFS_ACL_OBJ_ACE;
}
/*
* next check inheritance level flags
*/
if (obj_type == VDIR &&
(iflags & (ACE_FILE_INHERIT_ACE|ACE_DIRECTORY_INHERIT_ACE)))
aclp->z_hints |= ZFS_INHERIT_ACE;
if (iflags & (ACE_INHERIT_ONLY_ACE|ACE_NO_PROPAGATE_INHERIT_ACE)) {
if ((iflags & (ACE_FILE_INHERIT_ACE|
ACE_DIRECTORY_INHERIT_ACE)) == 0) {
return (B_FALSE);
}
}
return (B_TRUE);
}
static void *
zfs_acl_next_ace(zfs_acl_t *aclp, void *start, uint64_t *who,
uint32_t *access_mask, uint16_t *iflags, uint16_t *type)
{
zfs_acl_node_t *aclnode;
ASSERT3P(aclp, !=, NULL);
if (start == NULL) {
aclnode = list_head(&aclp->z_acl);
if (aclnode == NULL)
return (NULL);
aclp->z_next_ace = aclnode->z_acldata;
aclp->z_curr_node = aclnode;
aclnode->z_ace_idx = 0;
}
aclnode = aclp->z_curr_node;
if (aclnode == NULL)
return (NULL);
if (aclnode->z_ace_idx >= aclnode->z_ace_count) {
aclnode = list_next(&aclp->z_acl, aclnode);
if (aclnode == NULL)
return (NULL);
else {
aclp->z_curr_node = aclnode;
aclnode->z_ace_idx = 0;
aclp->z_next_ace = aclnode->z_acldata;
}
}
if (aclnode->z_ace_idx < aclnode->z_ace_count) {
void *acep = aclp->z_next_ace;
size_t ace_size;
/*
* Make sure we don't overstep our bounds
*/
ace_size = aclp->z_ops->ace_size(acep);
if (((caddr_t)acep + ace_size) >
((caddr_t)aclnode->z_acldata + aclnode->z_size)) {
return (NULL);
}
*iflags = aclp->z_ops->ace_flags_get(acep);
*type = aclp->z_ops->ace_type_get(acep);
*access_mask = aclp->z_ops->ace_mask_get(acep);
*who = aclp->z_ops->ace_who_get(acep);
aclp->z_next_ace = (caddr_t)aclp->z_next_ace + ace_size;
aclnode->z_ace_idx++;
return ((void *)acep);
}
return (NULL);
}
static uintptr_t
zfs_ace_walk(void *datap, uintptr_t cookie, int aclcnt,
uint16_t *flags, uint16_t *type, uint32_t *mask)
{
(void) aclcnt;
zfs_acl_t *aclp = datap;
zfs_ace_hdr_t *acep = (zfs_ace_hdr_t *)(uintptr_t)cookie;
uint64_t who;
acep = zfs_acl_next_ace(aclp, acep, &who, mask,
flags, type);
return ((uintptr_t)acep);
}
/*
* Copy ACE to internal ZFS format.
* While processing the ACL each ACE will be validated for correctness.
* ACE FUIDs will be created later.
*/
static int
zfs_copy_ace_2_fuid(zfsvfs_t *zfsvfs, vtype_t obj_type, zfs_acl_t *aclp,
void *datap, zfs_ace_t *z_acl, uint64_t aclcnt, size_t *size,
zfs_fuid_info_t **fuidp, cred_t *cr)
{
int i;
uint16_t entry_type;
zfs_ace_t *aceptr = z_acl;
ace_t *acep = datap;
zfs_object_ace_t *zobjacep;
ace_object_t *aceobjp;
for (i = 0; i != aclcnt; i++) {
aceptr->z_hdr.z_access_mask = acep->a_access_mask;
aceptr->z_hdr.z_flags = acep->a_flags;
aceptr->z_hdr.z_type = acep->a_type;
entry_type = aceptr->z_hdr.z_flags & ACE_TYPE_FLAGS;
if (entry_type != ACE_OWNER && entry_type != OWNING_GROUP &&
entry_type != ACE_EVERYONE) {
aceptr->z_fuid = zfs_fuid_create(zfsvfs, acep->a_who,
cr, (entry_type == 0) ?
ZFS_ACE_USER : ZFS_ACE_GROUP, fuidp);
}
/*
* Make sure ACE is valid
*/
if (zfs_ace_valid(obj_type, aclp, aceptr->z_hdr.z_type,
aceptr->z_hdr.z_flags) != B_TRUE)
return (SET_ERROR(EINVAL));
switch (acep->a_type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
zobjacep = (zfs_object_ace_t *)aceptr;
aceobjp = (ace_object_t *)acep;
memcpy(zobjacep->z_object_type, aceobjp->a_obj_type,
sizeof (aceobjp->a_obj_type));
memcpy(zobjacep->z_inherit_type,
aceobjp->a_inherit_obj_type,
sizeof (aceobjp->a_inherit_obj_type));
acep = (ace_t *)((caddr_t)acep + sizeof (ace_object_t));
break;
default:
acep = (ace_t *)((caddr_t)acep + sizeof (ace_t));
}
aceptr = (zfs_ace_t *)((caddr_t)aceptr +
aclp->z_ops->ace_size(aceptr));
}
*size = (caddr_t)aceptr - (caddr_t)z_acl;
return (0);
}
/*
* Copy ZFS ACEs to fixed size ace_t layout
*/
static void
zfs_copy_fuid_2_ace(zfsvfs_t *zfsvfs, zfs_acl_t *aclp, cred_t *cr,
void *datap, int filter)
{
uint64_t who;
uint32_t access_mask;
uint16_t iflags, type;
zfs_ace_hdr_t *zacep = NULL;
ace_t *acep = datap;
ace_object_t *objacep;
zfs_object_ace_t *zobjacep;
size_t ace_size;
uint16_t entry_type;
while ((zacep = zfs_acl_next_ace(aclp, zacep,
&who, &access_mask, &iflags, &type))) {
switch (type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
if (filter) {
continue;
}
zobjacep = (zfs_object_ace_t *)zacep;
objacep = (ace_object_t *)acep;
memcpy(objacep->a_obj_type,
zobjacep->z_object_type,
sizeof (zobjacep->z_object_type));
memcpy(objacep->a_inherit_obj_type,
zobjacep->z_inherit_type,
sizeof (zobjacep->z_inherit_type));
ace_size = sizeof (ace_object_t);
break;
default:
ace_size = sizeof (ace_t);
break;
}
entry_type = (iflags & ACE_TYPE_FLAGS);
if ((entry_type != ACE_OWNER &&
entry_type != OWNING_GROUP &&
entry_type != ACE_EVERYONE)) {
acep->a_who = zfs_fuid_map_id(zfsvfs, who,
cr, (entry_type & ACE_IDENTIFIER_GROUP) ?
ZFS_ACE_GROUP : ZFS_ACE_USER);
} else {
acep->a_who = (uid_t)(int64_t)who;
}
acep->a_access_mask = access_mask;
acep->a_flags = iflags;
acep->a_type = type;
acep = (ace_t *)((caddr_t)acep + ace_size);
}
}
static int
zfs_copy_ace_2_oldace(vtype_t obj_type, zfs_acl_t *aclp, ace_t *acep,
zfs_oldace_t *z_acl, int aclcnt, size_t *size)
{
int i;
zfs_oldace_t *aceptr = z_acl;
for (i = 0; i != aclcnt; i++, aceptr++) {
aceptr->z_access_mask = acep[i].a_access_mask;
aceptr->z_type = acep[i].a_type;
aceptr->z_flags = acep[i].a_flags;
aceptr->z_fuid = acep[i].a_who;
/*
* Make sure ACE is valid
*/
if (zfs_ace_valid(obj_type, aclp, aceptr->z_type,
aceptr->z_flags) != B_TRUE)
return (SET_ERROR(EINVAL));
}
*size = (caddr_t)aceptr - (caddr_t)z_acl;
return (0);
}
/*
* convert old ACL format to new
*/
void
zfs_acl_xform(znode_t *zp, zfs_acl_t *aclp, cred_t *cr)
{
zfs_oldace_t *oldaclp;
int i;
uint16_t type, iflags;
uint32_t access_mask;
uint64_t who;
void *cookie = NULL;
zfs_acl_node_t *newaclnode;
ASSERT3U(aclp->z_version, ==, ZFS_ACL_VERSION_INITIAL);
/*
* First create the ACE in a contiguous piece of memory
* for zfs_copy_ace_2_fuid().
*
* We only convert an ACL once, so this won't happen
* everytime.
*/
oldaclp = kmem_alloc(sizeof (zfs_oldace_t) * aclp->z_acl_count,
KM_SLEEP);
i = 0;
while ((cookie = zfs_acl_next_ace(aclp, cookie, &who,
&access_mask, &iflags, &type))) {
oldaclp[i].z_flags = iflags;
oldaclp[i].z_type = type;
oldaclp[i].z_fuid = who;
oldaclp[i++].z_access_mask = access_mask;
}
newaclnode = zfs_acl_node_alloc(aclp->z_acl_count *
sizeof (zfs_object_ace_t));
aclp->z_ops = &zfs_acl_fuid_ops;
VERIFY0(zfs_copy_ace_2_fuid(zp->z_zfsvfs, ZTOV(zp)->v_type, aclp,
oldaclp, newaclnode->z_acldata, aclp->z_acl_count,
&newaclnode->z_size, NULL, cr));
newaclnode->z_ace_count = aclp->z_acl_count;
aclp->z_version = ZFS_ACL_VERSION;
kmem_free(oldaclp, aclp->z_acl_count * sizeof (zfs_oldace_t));
/*
* Release all previous ACL nodes
*/
zfs_acl_release_nodes(aclp);
list_insert_head(&aclp->z_acl, newaclnode);
aclp->z_acl_bytes = newaclnode->z_size;
aclp->z_acl_count = newaclnode->z_ace_count;
}
/*
* Convert unix access mask to v4 access mask
*/
static uint32_t
zfs_unix_to_v4(uint32_t access_mask)
{
uint32_t new_mask = 0;
if (access_mask & S_IXOTH)
new_mask |= ACE_EXECUTE;
if (access_mask & S_IWOTH)
new_mask |= ACE_WRITE_DATA;
if (access_mask & S_IROTH)
new_mask |= ACE_READ_DATA;
return (new_mask);
}
static void
zfs_set_ace(zfs_acl_t *aclp, void *acep, uint32_t access_mask,
uint16_t access_type, uint64_t fuid, uint16_t entry_type)
{
uint16_t type = entry_type & ACE_TYPE_FLAGS;
aclp->z_ops->ace_mask_set(acep, access_mask);
aclp->z_ops->ace_type_set(acep, access_type);
aclp->z_ops->ace_flags_set(acep, entry_type);
if ((type != ACE_OWNER && type != OWNING_GROUP &&
type != ACE_EVERYONE))
aclp->z_ops->ace_who_set(acep, fuid);
}
/*
* Determine mode of file based on ACL.
*/
uint64_t
zfs_mode_compute(uint64_t fmode, zfs_acl_t *aclp,
uint64_t *pflags, uint64_t fuid, uint64_t fgid)
{
int entry_type;
mode_t mode;
mode_t seen = 0;
zfs_ace_hdr_t *acep = NULL;
uint64_t who;
uint16_t iflags, type;
uint32_t access_mask;
boolean_t an_exec_denied = B_FALSE;
mode = (fmode & (S_IFMT | S_ISUID | S_ISGID | S_ISVTX));
while ((acep = zfs_acl_next_ace(aclp, acep, &who,
&access_mask, &iflags, &type))) {
if (!zfs_acl_valid_ace_type(type, iflags))
continue;
entry_type = (iflags & ACE_TYPE_FLAGS);
/*
* Skip over any inherit_only ACEs
*/
if (iflags & ACE_INHERIT_ONLY_ACE)
continue;
if (entry_type == ACE_OWNER || (entry_type == 0 &&
who == fuid)) {
if ((access_mask & ACE_READ_DATA) &&
(!(seen & S_IRUSR))) {
seen |= S_IRUSR;
if (type == ALLOW) {
mode |= S_IRUSR;
}
}
if ((access_mask & ACE_WRITE_DATA) &&
(!(seen & S_IWUSR))) {
seen |= S_IWUSR;
if (type == ALLOW) {
mode |= S_IWUSR;
}
}
if ((access_mask & ACE_EXECUTE) &&
(!(seen & S_IXUSR))) {
seen |= S_IXUSR;
if (type == ALLOW) {
mode |= S_IXUSR;
}
}
} else if (entry_type == OWNING_GROUP ||
(entry_type == ACE_IDENTIFIER_GROUP && who == fgid)) {
if ((access_mask & ACE_READ_DATA) &&
(!(seen & S_IRGRP))) {
seen |= S_IRGRP;
if (type == ALLOW) {
mode |= S_IRGRP;
}
}
if ((access_mask & ACE_WRITE_DATA) &&
(!(seen & S_IWGRP))) {
seen |= S_IWGRP;
if (type == ALLOW) {
mode |= S_IWGRP;
}
}
if ((access_mask & ACE_EXECUTE) &&
(!(seen & S_IXGRP))) {
seen |= S_IXGRP;
if (type == ALLOW) {
mode |= S_IXGRP;
}
}
} else if (entry_type == ACE_EVERYONE) {
if ((access_mask & ACE_READ_DATA)) {
if (!(seen & S_IRUSR)) {
seen |= S_IRUSR;
if (type == ALLOW) {
mode |= S_IRUSR;
}
}
if (!(seen & S_IRGRP)) {
seen |= S_IRGRP;
if (type == ALLOW) {
mode |= S_IRGRP;
}
}
if (!(seen & S_IROTH)) {
seen |= S_IROTH;
if (type == ALLOW) {
mode |= S_IROTH;
}
}
}
if ((access_mask & ACE_WRITE_DATA)) {
if (!(seen & S_IWUSR)) {
seen |= S_IWUSR;
if (type == ALLOW) {
mode |= S_IWUSR;
}
}
if (!(seen & S_IWGRP)) {
seen |= S_IWGRP;
if (type == ALLOW) {
mode |= S_IWGRP;
}
}
if (!(seen & S_IWOTH)) {
seen |= S_IWOTH;
if (type == ALLOW) {
mode |= S_IWOTH;
}
}
}
if ((access_mask & ACE_EXECUTE)) {
if (!(seen & S_IXUSR)) {
seen |= S_IXUSR;
if (type == ALLOW) {
mode |= S_IXUSR;
}
}
if (!(seen & S_IXGRP)) {
seen |= S_IXGRP;
if (type == ALLOW) {
mode |= S_IXGRP;
}
}
if (!(seen & S_IXOTH)) {
seen |= S_IXOTH;
if (type == ALLOW) {
mode |= S_IXOTH;
}
}
}
} else {
/*
* Only care if this IDENTIFIER_GROUP or
* USER ACE denies execute access to someone,
* mode is not affected
*/
if ((access_mask & ACE_EXECUTE) && type == DENY)
an_exec_denied = B_TRUE;
}
}
/*
* Failure to allow is effectively a deny, so execute permission
* is denied if it was never mentioned or if we explicitly
* weren't allowed it.
*/
if (!an_exec_denied &&
((seen & ALL_MODE_EXECS) != ALL_MODE_EXECS ||
(mode & ALL_MODE_EXECS) != ALL_MODE_EXECS))
an_exec_denied = B_TRUE;
if (an_exec_denied)
*pflags &= ~ZFS_NO_EXECS_DENIED;
else
*pflags |= ZFS_NO_EXECS_DENIED;
return (mode);
}
/*
* Read an external acl object. If the intent is to modify, always
* create a new acl and leave any cached acl in place.
*/
int
zfs_acl_node_read(znode_t *zp, boolean_t have_lock, zfs_acl_t **aclpp,
boolean_t will_modify)
{
zfs_acl_t *aclp;
int aclsize;
int acl_count;
zfs_acl_node_t *aclnode;
zfs_acl_phys_t znode_acl;
int version;
int error;
ASSERT(MUTEX_HELD(&zp->z_acl_lock));
if (zp->z_zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_LOCKED(ZTOV(zp), __func__);
if (zp->z_acl_cached && !will_modify) {
*aclpp = zp->z_acl_cached;
return (0);
}
version = zfs_znode_acl_version(zp);
if ((error = zfs_acl_znode_info(zp, &aclsize,
&acl_count, &znode_acl)) != 0) {
goto done;
}
aclp = zfs_acl_alloc(version);
aclp->z_acl_count = acl_count;
aclp->z_acl_bytes = aclsize;
aclnode = zfs_acl_node_alloc(aclsize);
aclnode->z_ace_count = aclp->z_acl_count;
aclnode->z_size = aclsize;
if (!zp->z_is_sa) {
if (znode_acl.z_acl_extern_obj) {
error = dmu_read(zp->z_zfsvfs->z_os,
znode_acl.z_acl_extern_obj, 0, aclnode->z_size,
aclnode->z_acldata, DMU_READ_PREFETCH);
} else {
memcpy(aclnode->z_acldata, znode_acl.z_ace_data,
aclnode->z_size);
}
} else {
error = sa_lookup(zp->z_sa_hdl, SA_ZPL_DACL_ACES(zp->z_zfsvfs),
aclnode->z_acldata, aclnode->z_size);
}
if (error != 0) {
zfs_acl_free(aclp);
zfs_acl_node_free(aclnode);
/* convert checksum errors into IO errors */
if (error == ECKSUM)
error = SET_ERROR(EIO);
goto done;
}
list_insert_head(&aclp->z_acl, aclnode);
*aclpp = aclp;
if (!will_modify)
zp->z_acl_cached = aclp;
done:
return (error);
}
void
zfs_acl_data_locator(void **dataptr, uint32_t *length, uint32_t buflen,
boolean_t start, void *userdata)
{
(void) buflen;
zfs_acl_locator_cb_t *cb = (zfs_acl_locator_cb_t *)userdata;
if (start) {
cb->cb_acl_node = list_head(&cb->cb_aclp->z_acl);
} else {
cb->cb_acl_node = list_next(&cb->cb_aclp->z_acl,
cb->cb_acl_node);
}
ASSERT3P(cb->cb_acl_node, !=, NULL);
*dataptr = cb->cb_acl_node->z_acldata;
*length = cb->cb_acl_node->z_size;
}
int
zfs_acl_chown_setattr(znode_t *zp)
{
int error;
zfs_acl_t *aclp;
if (zp->z_zfsvfs->z_replay == B_FALSE) {
ASSERT_VOP_ELOCKED(ZTOV(zp), __func__);
ASSERT_VOP_IN_SEQC(ZTOV(zp));
}
ASSERT(MUTEX_HELD(&zp->z_acl_lock));
if ((error = zfs_acl_node_read(zp, B_TRUE, &aclp, B_FALSE)) == 0)
zp->z_mode = zfs_mode_compute(zp->z_mode, aclp,
&zp->z_pflags, zp->z_uid, zp->z_gid);
return (error);
}
/*
* common code for setting ACLs.
*
* This function is called from zfs_mode_update, zfs_perm_init, and zfs_setacl.
* zfs_setacl passes a non-NULL inherit pointer (ihp) to indicate that it's
* already checked the acl and knows whether to inherit.
*/
int
zfs_aclset_common(znode_t *zp, zfs_acl_t *aclp, cred_t *cr, dmu_tx_t *tx)
{
int error;
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
dmu_object_type_t otype;
zfs_acl_locator_cb_t locate = { 0 };
uint64_t mode;
sa_bulk_attr_t bulk[5];
uint64_t ctime[2];
int count = 0;
zfs_acl_phys_t acl_phys;
if (zp->z_zfsvfs->z_replay == B_FALSE) {
ASSERT_VOP_IN_SEQC(ZTOV(zp));
}
mode = zp->z_mode;
mode = zfs_mode_compute(mode, aclp, &zp->z_pflags,
zp->z_uid, zp->z_gid);
zp->z_mode = mode;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL,
&mode, sizeof (mode));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, sizeof (zp->z_pflags));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
&ctime, sizeof (ctime));
if (zp->z_acl_cached) {
zfs_acl_free(zp->z_acl_cached);
zp->z_acl_cached = NULL;
}
/*
* Upgrade needed?
*/
if (!zfsvfs->z_use_fuids) {
otype = DMU_OT_OLDACL;
} else {
if ((aclp->z_version == ZFS_ACL_VERSION_INITIAL) &&
(zfsvfs->z_version >= ZPL_VERSION_FUID))
zfs_acl_xform(zp, aclp, cr);
ASSERT3U(aclp->z_version, >=, ZFS_ACL_VERSION_FUID);
otype = DMU_OT_ACL;
}
/*
* Arrgh, we have to handle old on disk format
* as well as newer (preferred) SA format.
*/
if (zp->z_is_sa) { /* the easy case, just update the ACL attribute */
locate.cb_aclp = aclp;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_DACL_ACES(zfsvfs),
zfs_acl_data_locator, &locate, aclp->z_acl_bytes);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_DACL_COUNT(zfsvfs),
NULL, &aclp->z_acl_count, sizeof (uint64_t));
} else { /* Painful legacy way */
zfs_acl_node_t *aclnode;
uint64_t off = 0;
uint64_t aoid;
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_ZNODE_ACL(zfsvfs),
&acl_phys, sizeof (acl_phys))) != 0)
return (error);
aoid = acl_phys.z_acl_extern_obj;
if (aclp->z_acl_bytes > ZFS_ACE_SPACE) {
/*
* If ACL was previously external and we are now
* converting to new ACL format then release old
* ACL object and create a new one.
*/
if (aoid &&
aclp->z_version != acl_phys.z_acl_version) {
error = dmu_object_free(zfsvfs->z_os, aoid, tx);
if (error)
return (error);
aoid = 0;
}
if (aoid == 0) {
aoid = dmu_object_alloc(zfsvfs->z_os,
otype, aclp->z_acl_bytes,
otype == DMU_OT_ACL ?
DMU_OT_SYSACL : DMU_OT_NONE,
otype == DMU_OT_ACL ?
DN_OLD_MAX_BONUSLEN : 0, tx);
} else {
(void) dmu_object_set_blocksize(zfsvfs->z_os,
aoid, aclp->z_acl_bytes, 0, tx);
}
acl_phys.z_acl_extern_obj = aoid;
for (aclnode = list_head(&aclp->z_acl); aclnode;
aclnode = list_next(&aclp->z_acl, aclnode)) {
if (aclnode->z_ace_count == 0)
continue;
dmu_write(zfsvfs->z_os, aoid, off,
aclnode->z_size, aclnode->z_acldata, tx);
off += aclnode->z_size;
}
} else {
void *start = acl_phys.z_ace_data;
/*
* Migrating back embedded?
*/
if (acl_phys.z_acl_extern_obj) {
error = dmu_object_free(zfsvfs->z_os,
acl_phys.z_acl_extern_obj, tx);
if (error)
return (error);
acl_phys.z_acl_extern_obj = 0;
}
for (aclnode = list_head(&aclp->z_acl); aclnode;
aclnode = list_next(&aclp->z_acl, aclnode)) {
if (aclnode->z_ace_count == 0)
continue;
memcpy(start, aclnode->z_acldata,
aclnode->z_size);
start = (caddr_t)start + aclnode->z_size;
}
}
/*
* If Old version then swap count/bytes to match old
* layout of znode_acl_phys_t.
*/
if (aclp->z_version == ZFS_ACL_VERSION_INITIAL) {
acl_phys.z_acl_size = aclp->z_acl_count;
acl_phys.z_acl_count = aclp->z_acl_bytes;
} else {
acl_phys.z_acl_size = aclp->z_acl_bytes;
acl_phys.z_acl_count = aclp->z_acl_count;
}
acl_phys.z_acl_version = aclp->z_version;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ZNODE_ACL(zfsvfs), NULL,
&acl_phys, sizeof (acl_phys));
}
/*
* Replace ACL wide bits, but first clear them.
*/
zp->z_pflags &= ~ZFS_ACL_WIDE_FLAGS;
zp->z_pflags |= aclp->z_hints;
if (ace_trivial_common(aclp, 0, zfs_ace_walk) == 0)
zp->z_pflags |= ZFS_ACL_TRIVIAL;
zfs_tstamp_update_setup(zp, STATE_CHANGED, NULL, ctime);
return (sa_bulk_update(zp->z_sa_hdl, bulk, count, tx));
}
static void
zfs_acl_chmod(vtype_t vtype, uint64_t mode, boolean_t split, boolean_t trim,
zfs_acl_t *aclp)
{
void *acep = NULL;
uint64_t who;
int new_count, new_bytes;
int ace_size;
int entry_type;
uint16_t iflags, type;
uint32_t access_mask;
zfs_acl_node_t *newnode;
size_t abstract_size = aclp->z_ops->ace_abstract_size();
void *zacep;
boolean_t isdir;
trivial_acl_t masks;
new_count = new_bytes = 0;
isdir = (vtype == VDIR);
acl_trivial_access_masks((mode_t)mode, isdir, &masks);
newnode = zfs_acl_node_alloc((abstract_size * 6) + aclp->z_acl_bytes);
zacep = newnode->z_acldata;
if (masks.allow0) {
zfs_set_ace(aclp, zacep, masks.allow0, ALLOW, -1, ACE_OWNER);
zacep = (void *)((uintptr_t)zacep + abstract_size);
new_count++;
new_bytes += abstract_size;
}
if (masks.deny1) {
zfs_set_ace(aclp, zacep, masks.deny1, DENY, -1, ACE_OWNER);
zacep = (void *)((uintptr_t)zacep + abstract_size);
new_count++;
new_bytes += abstract_size;
}
if (masks.deny2) {
zfs_set_ace(aclp, zacep, masks.deny2, DENY, -1, OWNING_GROUP);
zacep = (void *)((uintptr_t)zacep + abstract_size);
new_count++;
new_bytes += abstract_size;
}
while ((acep = zfs_acl_next_ace(aclp, acep, &who, &access_mask,
&iflags, &type))) {
entry_type = (iflags & ACE_TYPE_FLAGS);
/*
* ACEs used to represent the file mode may be divided
* into an equivalent pair of inherit-only and regular
* ACEs, if they are inheritable.
* Skip regular ACEs, which are replaced by the new mode.
*/
if (split && (entry_type == ACE_OWNER ||
entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE)) {
if (!isdir || !(iflags &
(ACE_FILE_INHERIT_ACE|ACE_DIRECTORY_INHERIT_ACE)))
continue;
/*
* We preserve owner@, group@, or @everyone
* permissions, if they are inheritable, by
* copying them to inherit_only ACEs. This
* prevents inheritable permissions from being
* altered along with the file mode.
*/
iflags |= ACE_INHERIT_ONLY_ACE;
}
/*
* If this ACL has any inheritable ACEs, mark that in
* the hints (which are later masked into the pflags)
* so create knows to do inheritance.
*/
if (isdir && (iflags &
(ACE_FILE_INHERIT_ACE|ACE_DIRECTORY_INHERIT_ACE)))
aclp->z_hints |= ZFS_INHERIT_ACE;
if ((type != ALLOW && type != DENY) ||
(iflags & ACE_INHERIT_ONLY_ACE)) {
switch (type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
aclp->z_hints |= ZFS_ACL_OBJ_ACE;
break;
}
} else {
/*
* Limit permissions granted by ACEs to be no greater
* than permissions of the requested group mode.
* Applies when the "aclmode" property is set to
* "groupmask".
*/
if ((type == ALLOW) && trim)
access_mask &= masks.group;
}
zfs_set_ace(aclp, zacep, access_mask, type, who, iflags);
ace_size = aclp->z_ops->ace_size(acep);
zacep = (void *)((uintptr_t)zacep + ace_size);
new_count++;
new_bytes += ace_size;
}
zfs_set_ace(aclp, zacep, masks.owner, ALLOW, -1, ACE_OWNER);
zacep = (void *)((uintptr_t)zacep + abstract_size);
zfs_set_ace(aclp, zacep, masks.group, ALLOW, -1, OWNING_GROUP);
zacep = (void *)((uintptr_t)zacep + abstract_size);
zfs_set_ace(aclp, zacep, masks.everyone, ALLOW, -1, ACE_EVERYONE);
new_count += 3;
new_bytes += abstract_size * 3;
zfs_acl_release_nodes(aclp);
aclp->z_acl_count = new_count;
aclp->z_acl_bytes = new_bytes;
newnode->z_ace_count = new_count;
newnode->z_size = new_bytes;
list_insert_tail(&aclp->z_acl, newnode);
}
int
zfs_acl_chmod_setattr(znode_t *zp, zfs_acl_t **aclp, uint64_t mode)
{
int error = 0;
mutex_enter(&zp->z_acl_lock);
if (zp->z_zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_ELOCKED(ZTOV(zp), __func__);
if (zp->z_zfsvfs->z_acl_mode == ZFS_ACL_DISCARD)
*aclp = zfs_acl_alloc(zfs_acl_version_zp(zp));
else
error = zfs_acl_node_read(zp, B_TRUE, aclp, B_TRUE);
if (error == 0) {
(*aclp)->z_hints = zp->z_pflags & V4_ACL_WIDE_FLAGS;
zfs_acl_chmod(ZTOV(zp)->v_type, mode, B_TRUE,
(zp->z_zfsvfs->z_acl_mode == ZFS_ACL_GROUPMASK), *aclp);
}
mutex_exit(&zp->z_acl_lock);
return (error);
}
/*
* Should ACE be inherited?
*/
static int
zfs_ace_can_use(vtype_t vtype, uint16_t acep_flags)
{
int iflags = (acep_flags & 0xf);
if ((vtype == VDIR) && (iflags & ACE_DIRECTORY_INHERIT_ACE))
return (1);
else if (iflags & ACE_FILE_INHERIT_ACE)
return (!((vtype == VDIR) &&
(iflags & ACE_NO_PROPAGATE_INHERIT_ACE)));
return (0);
}
/*
* inherit inheritable ACEs from parent
*/
static zfs_acl_t *
zfs_acl_inherit(zfsvfs_t *zfsvfs, vtype_t vtype, zfs_acl_t *paclp,
uint64_t mode, boolean_t *need_chmod)
{
void *pacep = NULL;
void *acep;
zfs_acl_node_t *aclnode;
zfs_acl_t *aclp = NULL;
uint64_t who;
uint32_t access_mask;
uint16_t iflags, newflags, type;
size_t ace_size;
void *data1, *data2;
size_t data1sz, data2sz;
uint_t aclinherit;
boolean_t isdir = (vtype == VDIR);
boolean_t isreg = (vtype == VREG);
*need_chmod = B_TRUE;
aclp = zfs_acl_alloc(paclp->z_version);
aclinherit = zfsvfs->z_acl_inherit;
if (aclinherit == ZFS_ACL_DISCARD || vtype == VLNK)
return (aclp);
while ((pacep = zfs_acl_next_ace(paclp, pacep, &who,
&access_mask, &iflags, &type))) {
/*
* don't inherit bogus ACEs
*/
if (!zfs_acl_valid_ace_type(type, iflags))
continue;
/*
* Check if ACE is inheritable by this vnode
*/
if ((aclinherit == ZFS_ACL_NOALLOW && type == ALLOW) ||
!zfs_ace_can_use(vtype, iflags))
continue;
/*
* If owner@, group@, or everyone@ inheritable
* then zfs_acl_chmod() isn't needed.
*/
if ((aclinherit == ZFS_ACL_PASSTHROUGH ||
aclinherit == ZFS_ACL_PASSTHROUGH_X) &&
((iflags & (ACE_OWNER|ACE_EVERYONE)) ||
((iflags & OWNING_GROUP) == OWNING_GROUP)) &&
(isreg || (isdir && (iflags & ACE_DIRECTORY_INHERIT_ACE))))
*need_chmod = B_FALSE;
/*
* Strip inherited execute permission from file if
* not in mode
*/
if (aclinherit == ZFS_ACL_PASSTHROUGH_X && type == ALLOW &&
!isdir && ((mode & (S_IXUSR|S_IXGRP|S_IXOTH)) == 0)) {
access_mask &= ~ACE_EXECUTE;
}
/*
* Strip write_acl and write_owner from permissions
* when inheriting an ACE
*/
if (aclinherit == ZFS_ACL_RESTRICTED && type == ALLOW) {
access_mask &= ~RESTRICTED_CLEAR;
}
ace_size = aclp->z_ops->ace_size(pacep);
aclnode = zfs_acl_node_alloc(ace_size);
list_insert_tail(&aclp->z_acl, aclnode);
acep = aclnode->z_acldata;
zfs_set_ace(aclp, acep, access_mask, type,
who, iflags|ACE_INHERITED_ACE);
/*
* Copy special opaque data if any
*/
if ((data1sz = paclp->z_ops->ace_data(pacep, &data1)) != 0) {
data2sz = aclp->z_ops->ace_data(acep, &data2);
VERIFY3U(data2sz, ==, data1sz);
memcpy(data2, data1, data2sz);
}
aclp->z_acl_count++;
aclnode->z_ace_count++;
aclp->z_acl_bytes += aclnode->z_size;
newflags = aclp->z_ops->ace_flags_get(acep);
/*
* If ACE is not to be inherited further, or if the vnode is
* not a directory, remove all inheritance flags
*/
if (!isdir || (iflags & ACE_NO_PROPAGATE_INHERIT_ACE)) {
newflags &= ~ALL_INHERIT;
aclp->z_ops->ace_flags_set(acep,
newflags|ACE_INHERITED_ACE);
continue;
}
/*
* This directory has an inheritable ACE
*/
aclp->z_hints |= ZFS_INHERIT_ACE;
/*
* If only FILE_INHERIT is set then turn on
* inherit_only
*/
if ((iflags & (ACE_FILE_INHERIT_ACE |
ACE_DIRECTORY_INHERIT_ACE)) == ACE_FILE_INHERIT_ACE) {
newflags |= ACE_INHERIT_ONLY_ACE;
aclp->z_ops->ace_flags_set(acep,
newflags|ACE_INHERITED_ACE);
} else {
newflags &= ~ACE_INHERIT_ONLY_ACE;
aclp->z_ops->ace_flags_set(acep,
newflags|ACE_INHERITED_ACE);
}
}
if (zfsvfs->z_acl_mode == ZFS_ACL_RESTRICTED &&
aclp->z_acl_count != 0) {
*need_chmod = B_FALSE;
}
return (aclp);
}
/*
* Create file system object initial permissions
* including inheritable ACEs.
* Also, create FUIDs for owner and group.
*/
int
zfs_acl_ids_create(znode_t *dzp, int flag, vattr_t *vap, cred_t *cr,
- vsecattr_t *vsecp, zfs_acl_ids_t *acl_ids, zuserns_t *mnt_ns)
+ vsecattr_t *vsecp, zfs_acl_ids_t *acl_ids, zidmap_t *mnt_ns)
{
int error;
zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
zfs_acl_t *paclp;
gid_t gid;
boolean_t need_chmod = B_TRUE;
boolean_t trim = B_FALSE;
boolean_t inherited = B_FALSE;
if ((flag & IS_ROOT_NODE) == 0) {
if (zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_ELOCKED(ZTOV(dzp), __func__);
} else
ASSERT3P(dzp->z_vnode, ==, NULL);
memset(acl_ids, 0, sizeof (zfs_acl_ids_t));
acl_ids->z_mode = MAKEIMODE(vap->va_type, vap->va_mode);
if (vsecp)
if ((error = zfs_vsec_2_aclp(zfsvfs, vap->va_type, vsecp, cr,
&acl_ids->z_fuidp, &acl_ids->z_aclp)) != 0)
return (error);
/*
* Determine uid and gid.
*/
if ((flag & IS_ROOT_NODE) || zfsvfs->z_replay ||
((flag & IS_XATTR) && (vap->va_type == VDIR))) {
acl_ids->z_fuid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_uid, cr,
ZFS_OWNER, &acl_ids->z_fuidp);
acl_ids->z_fgid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_gid, cr,
ZFS_GROUP, &acl_ids->z_fuidp);
gid = vap->va_gid;
} else {
uid_t id = crgetuid(cr);
if (IS_EPHEMERAL(id))
id = UID_NOBODY;
acl_ids->z_fuid = (uint64_t)id;
acl_ids->z_fgid = 0;
if (vap->va_mask & AT_GID) {
acl_ids->z_fgid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_gid,
cr, ZFS_GROUP, &acl_ids->z_fuidp);
gid = vap->va_gid;
if (acl_ids->z_fgid != dzp->z_gid &&
!groupmember(vap->va_gid, cr) &&
secpolicy_vnode_create_gid(cr) != 0)
acl_ids->z_fgid = 0;
}
if (acl_ids->z_fgid == 0) {
const char *domain;
uint32_t rid;
acl_ids->z_fgid = dzp->z_gid;
gid = zfs_fuid_map_id(zfsvfs, acl_ids->z_fgid,
cr, ZFS_GROUP);
if (zfsvfs->z_use_fuids &&
IS_EPHEMERAL(acl_ids->z_fgid)) {
domain =
zfs_fuid_idx_domain(&zfsvfs->z_fuid_idx,
FUID_INDEX(acl_ids->z_fgid));
rid = FUID_RID(acl_ids->z_fgid);
zfs_fuid_node_add(&acl_ids->z_fuidp,
domain, rid, FUID_INDEX(acl_ids->z_fgid),
acl_ids->z_fgid, ZFS_GROUP);
}
}
}
/*
* If we're creating a directory, and the parent directory has the
* set-GID bit set, set in on the new directory.
* Otherwise, if the user is neither privileged nor a member of the
* file's new group, clear the file's set-GID bit.
*/
if (!(flag & IS_ROOT_NODE) && (dzp->z_mode & S_ISGID) &&
(vap->va_type == VDIR)) {
acl_ids->z_mode |= S_ISGID;
} else {
if ((acl_ids->z_mode & S_ISGID) &&
secpolicy_vnode_setids_setgids(ZTOV(dzp), cr, gid) != 0)
acl_ids->z_mode &= ~S_ISGID;
}
if (acl_ids->z_aclp == NULL) {
mutex_enter(&dzp->z_acl_lock);
if (!(flag & IS_ROOT_NODE) &&
(dzp->z_pflags & ZFS_INHERIT_ACE) &&
!(dzp->z_pflags & ZFS_XATTR)) {
VERIFY0(zfs_acl_node_read(dzp, B_TRUE,
&paclp, B_FALSE));
acl_ids->z_aclp = zfs_acl_inherit(zfsvfs,
vap->va_type, paclp, acl_ids->z_mode, &need_chmod);
inherited = B_TRUE;
} else {
acl_ids->z_aclp =
zfs_acl_alloc(zfs_acl_version_zp(dzp));
acl_ids->z_aclp->z_hints |= ZFS_ACL_TRIVIAL;
}
mutex_exit(&dzp->z_acl_lock);
if (need_chmod) {
if (vap->va_type == VDIR)
acl_ids->z_aclp->z_hints |=
ZFS_ACL_AUTO_INHERIT;
if (zfsvfs->z_acl_mode == ZFS_ACL_GROUPMASK &&
zfsvfs->z_acl_inherit != ZFS_ACL_PASSTHROUGH &&
zfsvfs->z_acl_inherit != ZFS_ACL_PASSTHROUGH_X)
trim = B_TRUE;
zfs_acl_chmod(vap->va_type, acl_ids->z_mode, B_FALSE,
trim, acl_ids->z_aclp);
}
}
if (inherited || vsecp) {
acl_ids->z_mode = zfs_mode_compute(acl_ids->z_mode,
acl_ids->z_aclp, &acl_ids->z_aclp->z_hints,
acl_ids->z_fuid, acl_ids->z_fgid);
if (ace_trivial_common(acl_ids->z_aclp, 0, zfs_ace_walk) == 0)
acl_ids->z_aclp->z_hints |= ZFS_ACL_TRIVIAL;
}
return (0);
}
/*
* Free ACL and fuid_infop, but not the acl_ids structure
*/
void
zfs_acl_ids_free(zfs_acl_ids_t *acl_ids)
{
if (acl_ids->z_aclp)
zfs_acl_free(acl_ids->z_aclp);
if (acl_ids->z_fuidp)
zfs_fuid_info_free(acl_ids->z_fuidp);
acl_ids->z_aclp = NULL;
acl_ids->z_fuidp = NULL;
}
boolean_t
zfs_acl_ids_overquota(zfsvfs_t *zv, zfs_acl_ids_t *acl_ids, uint64_t projid)
{
return (zfs_id_overquota(zv, DMU_USERUSED_OBJECT, acl_ids->z_fuid) ||
zfs_id_overquota(zv, DMU_GROUPUSED_OBJECT, acl_ids->z_fgid) ||
(projid != ZFS_DEFAULT_PROJID && projid != ZFS_INVALID_PROJID &&
zfs_id_overquota(zv, DMU_PROJECTUSED_OBJECT, projid)));
}
/*
* Retrieve a file's ACL
*/
int
zfs_getacl(znode_t *zp, vsecattr_t *vsecp, boolean_t skipaclchk, cred_t *cr)
{
zfs_acl_t *aclp;
ulong_t mask;
int error;
int count = 0;
int largeace = 0;
mask = vsecp->vsa_mask & (VSA_ACE | VSA_ACECNT |
VSA_ACE_ACLFLAGS | VSA_ACE_ALLTYPES);
if (mask == 0)
return (SET_ERROR(ENOSYS));
if ((error = zfs_zaccess(zp, ACE_READ_ACL, 0, skipaclchk, cr, NULL)))
return (error);
mutex_enter(&zp->z_acl_lock);
if (zp->z_zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_LOCKED(ZTOV(zp), __func__);
error = zfs_acl_node_read(zp, B_TRUE, &aclp, B_FALSE);
if (error != 0) {
mutex_exit(&zp->z_acl_lock);
return (error);
}
/*
* Scan ACL to determine number of ACEs
*/
if ((zp->z_pflags & ZFS_ACL_OBJ_ACE) && !(mask & VSA_ACE_ALLTYPES)) {
void *zacep = NULL;
uint64_t who;
uint32_t access_mask;
uint16_t type, iflags;
while ((zacep = zfs_acl_next_ace(aclp, zacep,
&who, &access_mask, &iflags, &type))) {
switch (type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
largeace++;
continue;
default:
count++;
}
}
vsecp->vsa_aclcnt = count;
} else
count = (int)aclp->z_acl_count;
if (mask & VSA_ACECNT) {
vsecp->vsa_aclcnt = count;
}
if (mask & VSA_ACE) {
size_t aclsz;
aclsz = count * sizeof (ace_t) +
sizeof (ace_object_t) * largeace;
vsecp->vsa_aclentp = kmem_alloc(aclsz, KM_SLEEP);
vsecp->vsa_aclentsz = aclsz;
if (aclp->z_version == ZFS_ACL_VERSION_FUID)
zfs_copy_fuid_2_ace(zp->z_zfsvfs, aclp, cr,
vsecp->vsa_aclentp, !(mask & VSA_ACE_ALLTYPES));
else {
zfs_acl_node_t *aclnode;
void *start = vsecp->vsa_aclentp;
for (aclnode = list_head(&aclp->z_acl); aclnode;
aclnode = list_next(&aclp->z_acl, aclnode)) {
memcpy(start, aclnode->z_acldata,
aclnode->z_size);
start = (caddr_t)start + aclnode->z_size;
}
ASSERT3U((caddr_t)start - (caddr_t)vsecp->vsa_aclentp,
==, aclp->z_acl_bytes);
}
}
if (mask & VSA_ACE_ACLFLAGS) {
vsecp->vsa_aclflags = 0;
if (zp->z_pflags & ZFS_ACL_DEFAULTED)
vsecp->vsa_aclflags |= ACL_DEFAULTED;
if (zp->z_pflags & ZFS_ACL_PROTECTED)
vsecp->vsa_aclflags |= ACL_PROTECTED;
if (zp->z_pflags & ZFS_ACL_AUTO_INHERIT)
vsecp->vsa_aclflags |= ACL_AUTO_INHERIT;
}
mutex_exit(&zp->z_acl_lock);
return (0);
}
int
zfs_vsec_2_aclp(zfsvfs_t *zfsvfs, umode_t obj_type,
vsecattr_t *vsecp, cred_t *cr, zfs_fuid_info_t **fuidp, zfs_acl_t **zaclp)
{
zfs_acl_t *aclp;
zfs_acl_node_t *aclnode;
int aclcnt = vsecp->vsa_aclcnt;
int error;
if (vsecp->vsa_aclcnt > MAX_ACL_ENTRIES || vsecp->vsa_aclcnt <= 0)
return (SET_ERROR(EINVAL));
aclp = zfs_acl_alloc(zfs_acl_version(zfsvfs->z_version));
aclp->z_hints = 0;
aclnode = zfs_acl_node_alloc(aclcnt * sizeof (zfs_object_ace_t));
if (aclp->z_version == ZFS_ACL_VERSION_INITIAL) {
if ((error = zfs_copy_ace_2_oldace(obj_type, aclp,
(ace_t *)vsecp->vsa_aclentp, aclnode->z_acldata,
aclcnt, &aclnode->z_size)) != 0) {
zfs_acl_free(aclp);
zfs_acl_node_free(aclnode);
return (error);
}
} else {
if ((error = zfs_copy_ace_2_fuid(zfsvfs, obj_type, aclp,
vsecp->vsa_aclentp, aclnode->z_acldata, aclcnt,
&aclnode->z_size, fuidp, cr)) != 0) {
zfs_acl_free(aclp);
zfs_acl_node_free(aclnode);
return (error);
}
}
aclp->z_acl_bytes = aclnode->z_size;
aclnode->z_ace_count = aclcnt;
aclp->z_acl_count = aclcnt;
list_insert_head(&aclp->z_acl, aclnode);
/*
* If flags are being set then add them to z_hints
*/
if (vsecp->vsa_mask & VSA_ACE_ACLFLAGS) {
if (vsecp->vsa_aclflags & ACL_PROTECTED)
aclp->z_hints |= ZFS_ACL_PROTECTED;
if (vsecp->vsa_aclflags & ACL_DEFAULTED)
aclp->z_hints |= ZFS_ACL_DEFAULTED;
if (vsecp->vsa_aclflags & ACL_AUTO_INHERIT)
aclp->z_hints |= ZFS_ACL_AUTO_INHERIT;
}
*zaclp = aclp;
return (0);
}
/*
* Set a file's ACL
*/
int
zfs_setacl(znode_t *zp, vsecattr_t *vsecp, boolean_t skipaclchk, cred_t *cr)
{
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
zilog_t *zilog = zfsvfs->z_log;
ulong_t mask = vsecp->vsa_mask & (VSA_ACE | VSA_ACECNT);
dmu_tx_t *tx;
int error;
zfs_acl_t *aclp;
zfs_fuid_info_t *fuidp = NULL;
boolean_t fuid_dirtied;
uint64_t acl_obj;
if (zp->z_zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_ELOCKED(ZTOV(zp), __func__);
if (mask == 0)
return (SET_ERROR(ENOSYS));
if (zp->z_pflags & ZFS_IMMUTABLE)
return (SET_ERROR(EPERM));
if ((error = zfs_zaccess(zp, ACE_WRITE_ACL, 0, skipaclchk, cr, NULL)))
return (error);
error = zfs_vsec_2_aclp(zfsvfs, ZTOV(zp)->v_type, vsecp, cr, &fuidp,
&aclp);
if (error)
return (error);
/*
* If ACL wide flags aren't being set then preserve any
* existing flags.
*/
if (!(vsecp->vsa_mask & VSA_ACE_ACLFLAGS)) {
aclp->z_hints |=
(zp->z_pflags & V4_ACL_WIDE_FLAGS);
}
top:
mutex_enter(&zp->z_acl_lock);
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
/*
* If old version and ACL won't fit in bonus and we aren't
* upgrading then take out necessary DMU holds
*/
if ((acl_obj = zfs_external_acl(zp)) != 0) {
if (zfsvfs->z_version >= ZPL_VERSION_FUID &&
zfs_znode_acl_version(zp) <= ZFS_ACL_VERSION_INITIAL) {
dmu_tx_hold_free(tx, acl_obj, 0,
DMU_OBJECT_END);
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0,
aclp->z_acl_bytes);
} else {
dmu_tx_hold_write(tx, acl_obj, 0, aclp->z_acl_bytes);
}
} else if (!zp->z_is_sa && aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, aclp->z_acl_bytes);
}
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_NOWAIT);
if (error) {
mutex_exit(&zp->z_acl_lock);
if (error == ERESTART) {
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
dmu_tx_abort(tx);
zfs_acl_free(aclp);
return (error);
}
error = zfs_aclset_common(zp, aclp, cr, tx);
ASSERT0(error);
ASSERT3P(zp->z_acl_cached, ==, NULL);
zp->z_acl_cached = aclp;
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
zfs_log_acl(zilog, tx, zp, vsecp, fuidp);
if (fuidp)
zfs_fuid_info_free(fuidp);
dmu_tx_commit(tx);
mutex_exit(&zp->z_acl_lock);
return (error);
}
/*
* Check accesses of interest (AoI) against attributes of the dataset
* such as read-only. Returns zero if no AoI conflict with dataset
* attributes, otherwise an appropriate errno is returned.
*/
static int
zfs_zaccess_dataset_check(znode_t *zp, uint32_t v4_mode)
{
if ((v4_mode & WRITE_MASK) &&
(zp->z_zfsvfs->z_vfs->vfs_flag & VFS_RDONLY) &&
(!IS_DEVVP(ZTOV(zp)) || (v4_mode & WRITE_MASK_ATTRS))) {
return (SET_ERROR(EROFS));
}
/*
* Intentionally allow ZFS_READONLY through here.
* See zfs_zaccess_common().
*/
if ((v4_mode & WRITE_MASK_DATA) &&
(zp->z_pflags & ZFS_IMMUTABLE)) {
return (SET_ERROR(EPERM));
}
/*
* In FreeBSD we allow to modify directory's content is ZFS_NOUNLINK
* (sunlnk) is set. We just don't allow directory removal, which is
* handled in zfs_zaccess_delete().
*/
if ((v4_mode & ACE_DELETE) &&
(zp->z_pflags & ZFS_NOUNLINK)) {
return (EPERM);
}
if (((v4_mode & (ACE_READ_DATA|ACE_EXECUTE)) &&
(zp->z_pflags & ZFS_AV_QUARANTINED))) {
return (SET_ERROR(EACCES));
}
return (0);
}
/*
* The primary usage of this function is to loop through all of the
* ACEs in the znode, determining what accesses of interest (AoI) to
* the caller are allowed or denied. The AoI are expressed as bits in
* the working_mode parameter. As each ACE is processed, bits covered
* by that ACE are removed from the working_mode. This removal
* facilitates two things. The first is that when the working mode is
* empty (= 0), we know we've looked at all the AoI. The second is
* that the ACE interpretation rules don't allow a later ACE to undo
* something granted or denied by an earlier ACE. Removing the
* discovered access or denial enforces this rule. At the end of
* processing the ACEs, all AoI that were found to be denied are
* placed into the working_mode, giving the caller a mask of denied
* accesses. Returns:
* 0 if all AoI granted
* EACCESS if the denied mask is non-zero
* other error if abnormal failure (e.g., IO error)
*
* A secondary usage of the function is to determine if any of the
* AoI are granted. If an ACE grants any access in
* the working_mode, we immediately short circuit out of the function.
* This mode is chosen by setting anyaccess to B_TRUE. The
* working_mode is not a denied access mask upon exit if the function
* is used in this manner.
*/
static int
zfs_zaccess_aces_check(znode_t *zp, uint32_t *working_mode,
boolean_t anyaccess, cred_t *cr)
{
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
zfs_acl_t *aclp;
int error;
uid_t uid = crgetuid(cr);
uint64_t who;
uint16_t type, iflags;
uint16_t entry_type;
uint32_t access_mask;
uint32_t deny_mask = 0;
zfs_ace_hdr_t *acep = NULL;
boolean_t checkit;
uid_t gowner;
uid_t fowner;
zfs_fuid_map_ids(zp, cr, &fowner, &gowner);
mutex_enter(&zp->z_acl_lock);
if (zp->z_zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_LOCKED(ZTOV(zp), __func__);
error = zfs_acl_node_read(zp, B_TRUE, &aclp, B_FALSE);
if (error != 0) {
mutex_exit(&zp->z_acl_lock);
return (error);
}
ASSERT3P(zp->z_acl_cached, !=, NULL);
while ((acep = zfs_acl_next_ace(aclp, acep, &who, &access_mask,
&iflags, &type))) {
uint32_t mask_matched;
if (!zfs_acl_valid_ace_type(type, iflags))
continue;
if (ZTOV(zp)->v_type == VDIR && (iflags & ACE_INHERIT_ONLY_ACE))
continue;
/* Skip ACE if it does not affect any AoI */
mask_matched = (access_mask & *working_mode);
if (!mask_matched)
continue;
entry_type = (iflags & ACE_TYPE_FLAGS);
checkit = B_FALSE;
switch (entry_type) {
case ACE_OWNER:
if (uid == fowner)
checkit = B_TRUE;
break;
case OWNING_GROUP:
who = gowner;
zfs_fallthrough;
case ACE_IDENTIFIER_GROUP:
checkit = zfs_groupmember(zfsvfs, who, cr);
break;
case ACE_EVERYONE:
checkit = B_TRUE;
break;
/* USER Entry */
default:
if (entry_type == 0) {
uid_t newid;
newid = zfs_fuid_map_id(zfsvfs, who, cr,
ZFS_ACE_USER);
if (newid != UID_NOBODY &&
uid == newid)
checkit = B_TRUE;
break;
} else {
mutex_exit(&zp->z_acl_lock);
return (SET_ERROR(EIO));
}
}
if (checkit) {
if (type == DENY) {
DTRACE_PROBE3(zfs__ace__denies,
znode_t *, zp,
zfs_ace_hdr_t *, acep,
uint32_t, mask_matched);
deny_mask |= mask_matched;
} else {
DTRACE_PROBE3(zfs__ace__allows,
znode_t *, zp,
zfs_ace_hdr_t *, acep,
uint32_t, mask_matched);
if (anyaccess) {
mutex_exit(&zp->z_acl_lock);
return (0);
}
}
*working_mode &= ~mask_matched;
}
/* Are we done? */
if (*working_mode == 0)
break;
}
mutex_exit(&zp->z_acl_lock);
/* Put the found 'denies' back on the working mode */
if (deny_mask) {
*working_mode |= deny_mask;
return (SET_ERROR(EACCES));
} else if (*working_mode) {
return (-1);
}
return (0);
}
/*
* Return true if any access whatsoever granted, we don't actually
* care what access is granted.
*/
boolean_t
zfs_has_access(znode_t *zp, cred_t *cr)
{
uint32_t have = ACE_ALL_PERMS;
if (zfs_zaccess_aces_check(zp, &have, B_TRUE, cr) != 0) {
uid_t owner;
owner = zfs_fuid_map_id(zp->z_zfsvfs, zp->z_uid, cr, ZFS_OWNER);
return (secpolicy_vnode_any_access(cr, ZTOV(zp), owner) == 0);
}
return (B_TRUE);
}
static int
zfs_zaccess_common(znode_t *zp, uint32_t v4_mode, uint32_t *working_mode,
boolean_t *check_privs, boolean_t skipaclchk, cred_t *cr)
{
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
int err;
*working_mode = v4_mode;
*check_privs = B_TRUE;
/*
* Short circuit empty requests
*/
if (v4_mode == 0 || zfsvfs->z_replay) {
*working_mode = 0;
return (0);
}
if ((err = zfs_zaccess_dataset_check(zp, v4_mode)) != 0) {
*check_privs = B_FALSE;
return (err);
}
/*
* The caller requested that the ACL check be skipped. This
* would only happen if the caller checked VOP_ACCESS() with a
* 32 bit ACE mask and already had the appropriate permissions.
*/
if (skipaclchk) {
*working_mode = 0;
return (0);
}
/*
* Note: ZFS_READONLY represents the "DOS R/O" attribute.
* When that flag is set, we should behave as if write access
* were not granted by anything in the ACL. In particular:
* We _must_ allow writes after opening the file r/w, then
* setting the DOS R/O attribute, and writing some more.
* (Similar to how you can write after fchmod(fd, 0444).)
*
* Therefore ZFS_READONLY is ignored in the dataset check
* above, and checked here as if part of the ACL check.
* Also note: DOS R/O is ignored for directories.
*/
if ((v4_mode & WRITE_MASK_DATA) &&
(ZTOV(zp)->v_type != VDIR) &&
(zp->z_pflags & ZFS_READONLY)) {
return (SET_ERROR(EPERM));
}
return (zfs_zaccess_aces_check(zp, working_mode, B_FALSE, cr));
}
static int
zfs_zaccess_append(znode_t *zp, uint32_t *working_mode, boolean_t *check_privs,
cred_t *cr)
{
if (*working_mode != ACE_WRITE_DATA)
return (SET_ERROR(EACCES));
return (zfs_zaccess_common(zp, ACE_APPEND_DATA, working_mode,
check_privs, B_FALSE, cr));
}
/*
* Check if VEXEC is allowed.
*
* This routine is based on zfs_fastaccesschk_execute which has slowpath
* calling zfs_zaccess. This would be incorrect on FreeBSD (see
* zfs_freebsd_access for the difference). Thus this variant let's the
* caller handle the slowpath (if necessary).
*
* On top of that we perform a lockless check for ZFS_NO_EXECS_DENIED.
*
* Safe access to znode_t is provided by the vnode lock.
*/
int
zfs_fastaccesschk_execute(znode_t *zdp, cred_t *cr)
{
boolean_t is_attr;
if (zdp->z_pflags & ZFS_AV_QUARANTINED)
return (1);
is_attr = ((zdp->z_pflags & ZFS_XATTR) &&
(ZTOV(zdp)->v_type == VDIR));
if (is_attr)
return (1);
if (zdp->z_pflags & ZFS_NO_EXECS_DENIED)
return (0);
return (1);
}
/*
* Determine whether Access should be granted/denied.
*
* The least priv subsystem is always consulted as a basic privilege
* can define any form of access.
*/
int
zfs_zaccess(znode_t *zp, int mode, int flags, boolean_t skipaclchk, cred_t *cr,
- zuserns_t *mnt_ns)
+ zidmap_t *mnt_ns)
{
uint32_t working_mode;
int error;
int is_attr;
boolean_t check_privs;
znode_t *xzp = NULL;
znode_t *check_zp = zp;
mode_t needed_bits;
uid_t owner;
is_attr = ((zp->z_pflags & ZFS_XATTR) && (ZTOV(zp)->v_type == VDIR));
/*
* In FreeBSD, we don't care about permissions of individual ADS.
* Note that not checking them is not just an optimization - without
* this shortcut, EA operations may bogusly fail with EACCES.
*/
if (zp->z_pflags & ZFS_XATTR)
return (0);
owner = zfs_fuid_map_id(zp->z_zfsvfs, zp->z_uid, cr, ZFS_OWNER);
/*
* Map the bits required to the standard vnode flags VREAD|VWRITE|VEXEC
* in needed_bits. Map the bits mapped by working_mode (currently
* missing) in missing_bits.
* Call secpolicy_vnode_access2() with (needed_bits & ~checkmode),
* needed_bits.
*/
needed_bits = 0;
working_mode = mode;
if ((working_mode & (ACE_READ_ACL|ACE_READ_ATTRIBUTES)) &&
owner == crgetuid(cr))
working_mode &= ~(ACE_READ_ACL|ACE_READ_ATTRIBUTES);
if (working_mode & (ACE_READ_DATA|ACE_READ_NAMED_ATTRS|
ACE_READ_ACL|ACE_READ_ATTRIBUTES|ACE_SYNCHRONIZE))
needed_bits |= VREAD;
if (working_mode & (ACE_WRITE_DATA|ACE_WRITE_NAMED_ATTRS|
ACE_APPEND_DATA|ACE_WRITE_ATTRIBUTES|ACE_SYNCHRONIZE))
needed_bits |= VWRITE;
if (working_mode & ACE_EXECUTE)
needed_bits |= VEXEC;
if ((error = zfs_zaccess_common(check_zp, mode, &working_mode,
&check_privs, skipaclchk, cr)) == 0) {
if (is_attr)
VN_RELE(ZTOV(xzp));
return (secpolicy_vnode_access2(cr, ZTOV(zp), owner,
needed_bits, needed_bits));
}
if (error && !check_privs) {
if (is_attr)
VN_RELE(ZTOV(xzp));
return (error);
}
if (error && (flags & V_APPEND)) {
error = zfs_zaccess_append(zp, &working_mode, &check_privs, cr);
}
if (error && check_privs) {
mode_t checkmode = 0;
vnode_t *check_vp = ZTOV(check_zp);
/*
* First check for implicit owner permission on
* read_acl/read_attributes
*/
ASSERT3U(working_mode, !=, 0);
if ((working_mode & (ACE_READ_ACL|ACE_READ_ATTRIBUTES) &&
owner == crgetuid(cr)))
working_mode &= ~(ACE_READ_ACL|ACE_READ_ATTRIBUTES);
if (working_mode & (ACE_READ_DATA|ACE_READ_NAMED_ATTRS|
ACE_READ_ACL|ACE_READ_ATTRIBUTES|ACE_SYNCHRONIZE))
checkmode |= VREAD;
if (working_mode & (ACE_WRITE_DATA|ACE_WRITE_NAMED_ATTRS|
ACE_APPEND_DATA|ACE_WRITE_ATTRIBUTES|ACE_SYNCHRONIZE))
checkmode |= VWRITE;
if (working_mode & ACE_EXECUTE)
checkmode |= VEXEC;
error = secpolicy_vnode_access2(cr, check_vp, owner,
needed_bits & ~checkmode, needed_bits);
if (error == 0 && (working_mode & ACE_WRITE_OWNER))
error = secpolicy_vnode_chown(check_vp, cr, owner);
if (error == 0 && (working_mode & ACE_WRITE_ACL))
error = secpolicy_vnode_setdac(check_vp, cr, owner);
if (error == 0 && (working_mode &
(ACE_DELETE|ACE_DELETE_CHILD)))
error = secpolicy_vnode_remove(check_vp, cr);
if (error == 0 && (working_mode & ACE_SYNCHRONIZE)) {
error = secpolicy_vnode_chown(check_vp, cr, owner);
}
if (error == 0) {
/*
* See if any bits other than those already checked
* for are still present. If so then return EACCES
*/
if (working_mode & ~(ZFS_CHECKED_MASKS)) {
error = SET_ERROR(EACCES);
}
}
} else if (error == 0) {
error = secpolicy_vnode_access2(cr, ZTOV(zp), owner,
needed_bits, needed_bits);
}
if (is_attr)
VN_RELE(ZTOV(xzp));
return (error);
}
/*
* Translate traditional unix VREAD/VWRITE/VEXEC mode into
* NFSv4-style ZFS ACL format and call zfs_zaccess()
*/
int
zfs_zaccess_rwx(znode_t *zp, mode_t mode, int flags, cred_t *cr,
- zuserns_t *mnt_ns)
+ zidmap_t *mnt_ns)
{
return (zfs_zaccess(zp, zfs_unix_to_v4(mode >> 6), flags, B_FALSE, cr,
mnt_ns));
}
/*
* Access function for secpolicy_vnode_setattr
*/
int
zfs_zaccess_unix(void *zp, int mode, cred_t *cr)
{
int v4_mode = zfs_unix_to_v4(mode >> 6);
return (zfs_zaccess(zp, v4_mode, 0, B_FALSE, cr, NULL));
}
static int
zfs_delete_final_check(znode_t *zp, znode_t *dzp,
mode_t available_perms, cred_t *cr)
{
int error;
uid_t downer;
downer = zfs_fuid_map_id(dzp->z_zfsvfs, dzp->z_uid, cr, ZFS_OWNER);
error = secpolicy_vnode_access2(cr, ZTOV(dzp),
downer, available_perms, VWRITE|VEXEC);
if (error == 0)
error = zfs_sticky_remove_access(dzp, zp, cr);
return (error);
}
/*
* Determine whether Access should be granted/deny, without
* consulting least priv subsystem.
*
* The following chart is the recommended NFSv4 enforcement for
* ability to delete an object.
*
* -------------------------------------------------------
* | Parent Dir | Target Object Permissions |
* | permissions | |
* -------------------------------------------------------
* | | ACL Allows | ACL Denies| Delete |
* | | Delete | Delete | unspecified|
* -------------------------------------------------------
* | ACL Allows | Permit | Permit | Permit |
* | DELETE_CHILD | |
* -------------------------------------------------------
* | ACL Denies | Permit | Deny | Deny |
* | DELETE_CHILD | | | |
* -------------------------------------------------------
* | ACL specifies | | | |
* | only allow | Permit | Permit | Permit |
* | write and | | | |
* | execute | | | |
* -------------------------------------------------------
* | ACL denies | | | |
* | write and | Permit | Deny | Deny |
* | execute | | | |
* -------------------------------------------------------
* ^
* |
* No search privilege, can't even look up file?
*
*/
int
-zfs_zaccess_delete(znode_t *dzp, znode_t *zp, cred_t *cr, zuserns_t *mnt_ns)
+zfs_zaccess_delete(znode_t *dzp, znode_t *zp, cred_t *cr, zidmap_t *mnt_ns)
{
uint32_t dzp_working_mode = 0;
uint32_t zp_working_mode = 0;
int dzp_error, zp_error;
mode_t available_perms;
boolean_t dzpcheck_privs = B_TRUE;
boolean_t zpcheck_privs = B_TRUE;
/*
* We want specific DELETE permissions to
* take precedence over WRITE/EXECUTE. We don't
* want an ACL such as this to mess us up.
* user:joe:write_data:deny,user:joe:delete:allow
*
* However, deny permissions may ultimately be overridden
* by secpolicy_vnode_access().
*
* We will ask for all of the necessary permissions and then
* look at the working modes from the directory and target object
* to determine what was found.
*/
if (zp->z_pflags & (ZFS_IMMUTABLE | ZFS_NOUNLINK))
return (SET_ERROR(EPERM));
/*
* First row
* If the directory permissions allow the delete, we are done.
*/
if ((dzp_error = zfs_zaccess_common(dzp, ACE_DELETE_CHILD,
&dzp_working_mode, &dzpcheck_privs, B_FALSE, cr)) == 0)
return (0);
/*
* If target object has delete permission then we are done
*/
if ((zp_error = zfs_zaccess_common(zp, ACE_DELETE, &zp_working_mode,
&zpcheck_privs, B_FALSE, cr)) == 0)
return (0);
ASSERT(dzp_error);
ASSERT(zp_error);
if (!dzpcheck_privs)
return (dzp_error);
if (!zpcheck_privs)
return (zp_error);
/*
* Second row
*
* If directory returns EACCES then delete_child was denied
* due to deny delete_child. In this case send the request through
* secpolicy_vnode_remove(). We don't use zfs_delete_final_check()
* since that *could* allow the delete based on write/execute permission
* and we want delete permissions to override write/execute.
*/
if (dzp_error == EACCES) {
/* XXXPJD: s/dzp/zp/ ? */
return (secpolicy_vnode_remove(ZTOV(dzp), cr));
}
/*
* Third Row
* only need to see if we have write/execute on directory.
*/
dzp_error = zfs_zaccess_common(dzp, ACE_EXECUTE|ACE_WRITE_DATA,
&dzp_working_mode, &dzpcheck_privs, B_FALSE, cr);
if (dzp_error != 0 && !dzpcheck_privs)
return (dzp_error);
/*
* Fourth row
*/
available_perms = (dzp_working_mode & ACE_WRITE_DATA) ? 0 : VWRITE;
available_perms |= (dzp_working_mode & ACE_EXECUTE) ? 0 : VEXEC;
return (zfs_delete_final_check(zp, dzp, available_perms, cr));
}
int
zfs_zaccess_rename(znode_t *sdzp, znode_t *szp, znode_t *tdzp,
- znode_t *tzp, cred_t *cr, zuserns_t *mnt_ns)
+ znode_t *tzp, cred_t *cr, zidmap_t *mnt_ns)
{
int add_perm;
int error;
if (szp->z_pflags & ZFS_AV_QUARANTINED)
return (SET_ERROR(EACCES));
add_perm = (ZTOV(szp)->v_type == VDIR) ?
ACE_ADD_SUBDIRECTORY : ACE_ADD_FILE;
/*
* Rename permissions are combination of delete permission +
* add file/subdir permission.
*
* BSD operating systems also require write permission
* on the directory being moved from one parent directory
* to another.
*/
if (ZTOV(szp)->v_type == VDIR && ZTOV(sdzp) != ZTOV(tdzp)) {
if ((error = zfs_zaccess(szp, ACE_WRITE_DATA, 0, B_FALSE, cr,
mnt_ns)))
return (error);
}
/*
* first make sure we do the delete portion.
*
* If that succeeds then check for add_file/add_subdir permissions
*/
if ((error = zfs_zaccess_delete(sdzp, szp, cr, mnt_ns)))
return (error);
/*
* If we have a tzp, see if we can delete it?
*/
if (tzp && (error = zfs_zaccess_delete(tdzp, tzp, cr, mnt_ns)))
return (error);
/*
* Now check for add permissions
*/
error = zfs_zaccess(tdzp, add_perm, 0, B_FALSE, cr, mnt_ns);
return (error);
}
diff --git a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c
index 5079d0afa9e4..0d6d006781d6 100644
--- a/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c
+++ b/sys/contrib/openzfs/module/os/freebsd/zfs/zfs_vnops_os.c
@@ -1,6435 +1,6438 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2015 by Delphix. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2017 Nexenta Systems, Inc.
*/
/* Portions Copyright 2007 Jeremy Teo */
/* Portions Copyright 2010 Robert Milkowski */
#include <sys/param.h>
#include <sys/time.h>
#include <sys/systm.h>
#include <sys/sysmacros.h>
#include <sys/resource.h>
#include <security/mac/mac_framework.h>
#include <sys/vfs.h>
#include <sys/endian.h>
#include <sys/vm.h>
#include <sys/vnode.h>
#if __FreeBSD_version >= 1300102
#include <sys/smr.h>
#endif
#include <sys/dirent.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/kmem.h>
#include <sys/taskq.h>
#include <sys/uio.h>
#include <sys/atomic.h>
#include <sys/namei.h>
#include <sys/mman.h>
#include <sys/cmn_err.h>
#include <sys/kdb.h>
#include <sys/sysproto.h>
#include <sys/errno.h>
#include <sys/unistd.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_ioctl.h>
#include <sys/fs/zfs.h>
#include <sys/dmu.h>
#include <sys/dmu_objset.h>
#include <sys/spa.h>
#include <sys/txg.h>
#include <sys/dbuf.h>
#include <sys/zap.h>
#include <sys/sa.h>
#include <sys/policy.h>
#include <sys/sunddi.h>
#include <sys/filio.h>
#include <sys/sid.h>
#include <sys/zfs_ctldir.h>
#include <sys/zfs_fuid.h>
#include <sys/zfs_quota.h>
#include <sys/zfs_sa.h>
#include <sys/zfs_rlock.h>
#include <sys/bio.h>
#include <sys/buf.h>
#include <sys/sched.h>
#include <sys/acl.h>
#include <sys/vmmeter.h>
#include <vm/vm_param.h>
#include <sys/zil.h>
#include <sys/zfs_vnops.h>
#include <sys/module.h>
#include <sys/sysent.h>
-
#include <sys/dmu_impl.h>
#include <sys/brt.h>
#include <sys/zfeature.h>
#include <vm/vm_object.h>
#include <sys/extattr.h>
#include <sys/priv.h>
#ifndef VN_OPEN_INVFS
#define VN_OPEN_INVFS 0x0
#endif
VFS_SMR_DECLARE;
#if __FreeBSD_version < 1300103
#define NDFREE_PNBUF(ndp) NDFREE((ndp), NDF_ONLY_PNBUF)
#endif
#if __FreeBSD_version >= 1300047
#define vm_page_wire_lock(pp)
#define vm_page_wire_unlock(pp)
#else
#define vm_page_wire_lock(pp) vm_page_lock(pp)
#define vm_page_wire_unlock(pp) vm_page_unlock(pp)
#endif
#ifdef DEBUG_VFS_LOCKS
#define VNCHECKREF(vp) \
VNASSERT((vp)->v_holdcnt > 0 && (vp)->v_usecount > 0, vp, \
("%s: wrong ref counts", __func__));
#else
#define VNCHECKREF(vp)
#endif
#if __FreeBSD_version >= 1400045
typedef uint64_t cookie_t;
#else
typedef ulong_t cookie_t;
#endif
/*
* Programming rules.
*
* Each vnode op performs some logical unit of work. To do this, the ZPL must
* properly lock its in-core state, create a DMU transaction, do the work,
* record this work in the intent log (ZIL), commit the DMU transaction,
* and wait for the intent log to commit if it is a synchronous operation.
* Moreover, the vnode ops must work in both normal and log replay context.
* The ordering of events is important to avoid deadlocks and references
* to freed memory. The example below illustrates the following Big Rules:
*
* (1) A check must be made in each zfs thread for a mounted file system.
* This is done avoiding races using zfs_enter(zfsvfs).
* A zfs_exit(zfsvfs) is needed before all returns. Any znodes
* must be checked with zfs_verify_zp(zp). Both of these macros
* can return EIO from the calling function.
*
* (2) VN_RELE() should always be the last thing except for zil_commit()
* (if necessary) and zfs_exit(). This is for 3 reasons:
* First, if it's the last reference, the vnode/znode
* can be freed, so the zp may point to freed memory. Second, the last
* reference will call zfs_zinactive(), which may induce a lot of work --
* pushing cached pages (which acquires range locks) and syncing out
* cached atime changes. Third, zfs_zinactive() may require a new tx,
* which could deadlock the system if you were already holding one.
* If you must call VN_RELE() within a tx then use VN_RELE_ASYNC().
*
* (3) All range locks must be grabbed before calling dmu_tx_assign(),
* as they can span dmu_tx_assign() calls.
*
* (4) If ZPL locks are held, pass TXG_NOWAIT as the second argument to
* dmu_tx_assign(). This is critical because we don't want to block
* while holding locks.
*
* If no ZPL locks are held (aside from zfs_enter()), use TXG_WAIT. This
* reduces lock contention and CPU usage when we must wait (note that if
* throughput is constrained by the storage, nearly every transaction
* must wait).
*
* Note, in particular, that if a lock is sometimes acquired before
* the tx assigns, and sometimes after (e.g. z_lock), then failing
* to use a non-blocking assign can deadlock the system. The scenario:
*
* Thread A has grabbed a lock before calling dmu_tx_assign().
* Thread B is in an already-assigned tx, and blocks for this lock.
* Thread A calls dmu_tx_assign(TXG_WAIT) and blocks in txg_wait_open()
* forever, because the previous txg can't quiesce until B's tx commits.
*
* If dmu_tx_assign() returns ERESTART and zfsvfs->z_assign is TXG_NOWAIT,
* then drop all locks, call dmu_tx_wait(), and try again. On subsequent
* calls to dmu_tx_assign(), pass TXG_NOTHROTTLE in addition to TXG_NOWAIT,
* to indicate that this operation has already called dmu_tx_wait().
* This will ensure that we don't retry forever, waiting a short bit
* each time.
*
* (5) If the operation succeeded, generate the intent log entry for it
* before dropping locks. This ensures that the ordering of events
* in the intent log matches the order in which they actually occurred.
* During ZIL replay the zfs_log_* functions will update the sequence
* number to indicate the zil transaction has replayed.
*
* (6) At the end of each vnode op, the DMU tx must always commit,
* regardless of whether there were any errors.
*
* (7) After dropping all locks, invoke zil_commit(zilog, foid)
* to ensure that synchronous semantics are provided when necessary.
*
* In general, this is how things should be ordered in each vnode op:
*
* zfs_enter(zfsvfs); // exit if unmounted
* top:
* zfs_dirent_lookup(&dl, ...) // lock directory entry (may VN_HOLD())
* rw_enter(...); // grab any other locks you need
* tx = dmu_tx_create(...); // get DMU tx
* dmu_tx_hold_*(); // hold each object you might modify
* error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
* if (error) {
* rw_exit(...); // drop locks
* zfs_dirent_unlock(dl); // unlock directory entry
* VN_RELE(...); // release held vnodes
* if (error == ERESTART) {
* waited = B_TRUE;
* dmu_tx_wait(tx);
* dmu_tx_abort(tx);
* goto top;
* }
* dmu_tx_abort(tx); // abort DMU tx
* zfs_exit(zfsvfs); // finished in zfs
* return (error); // really out of space
* }
* error = do_real_work(); // do whatever this VOP does
* if (error == 0)
* zfs_log_*(...); // on success, make ZIL entry
* dmu_tx_commit(tx); // commit DMU tx -- error or not
* rw_exit(...); // drop locks
* zfs_dirent_unlock(dl); // unlock directory entry
* VN_RELE(...); // release held vnodes
* zil_commit(zilog, foid); // synchronous when necessary
* zfs_exit(zfsvfs); // finished in zfs
* return (error); // done, report error
*/
static int
zfs_open(vnode_t **vpp, int flag, cred_t *cr)
{
(void) cr;
znode_t *zp = VTOZ(*vpp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if ((flag & FWRITE) && (zp->z_pflags & ZFS_APPENDONLY) &&
((flag & FAPPEND) == 0)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
/* Keep a count of the synchronous opens in the znode */
if (flag & O_SYNC)
atomic_inc_32(&zp->z_sync_cnt);
zfs_exit(zfsvfs, FTAG);
return (0);
}
static int
zfs_close(vnode_t *vp, int flag, int count, offset_t offset, cred_t *cr)
{
(void) offset, (void) cr;
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
/* Decrement the synchronous opens in the znode */
if ((flag & O_SYNC) && (count == 1))
atomic_dec_32(&zp->z_sync_cnt);
zfs_exit(zfsvfs, FTAG);
return (0);
}
static int
zfs_ioctl(vnode_t *vp, ulong_t com, intptr_t data, int flag, cred_t *cred,
int *rvalp)
{
(void) flag, (void) cred, (void) rvalp;
loff_t off;
int error;
switch (com) {
case _FIOFFS:
{
return (0);
/*
* The following two ioctls are used by bfu. Faking out,
* necessary to avoid bfu errors.
*/
}
case _FIOGDIO:
case _FIOSDIO:
{
return (0);
}
case F_SEEK_DATA:
case F_SEEK_HOLE:
{
off = *(offset_t *)data;
/* offset parameter is in/out */
error = zfs_holey(VTOZ(vp), com, &off);
if (error)
return (error);
*(offset_t *)data = off;
return (0);
}
}
return (SET_ERROR(ENOTTY));
}
static vm_page_t
page_busy(vnode_t *vp, int64_t start, int64_t off, int64_t nbytes)
{
vm_object_t obj;
vm_page_t pp;
int64_t end;
/*
* At present vm_page_clear_dirty extends the cleared range to DEV_BSIZE
* aligned boundaries, if the range is not aligned. As a result a
* DEV_BSIZE subrange with partially dirty data may get marked as clean.
* It may happen that all DEV_BSIZE subranges are marked clean and thus
* the whole page would be considered clean despite have some
* dirty data.
* For this reason we should shrink the range to DEV_BSIZE aligned
* boundaries before calling vm_page_clear_dirty.
*/
end = rounddown2(off + nbytes, DEV_BSIZE);
off = roundup2(off, DEV_BSIZE);
nbytes = end - off;
obj = vp->v_object;
zfs_vmobject_assert_wlocked_12(obj);
#if __FreeBSD_version < 1300050
for (;;) {
if ((pp = vm_page_lookup(obj, OFF_TO_IDX(start))) != NULL &&
pp->valid) {
if (vm_page_xbusied(pp)) {
/*
* Reference the page before unlocking and
* sleeping so that the page daemon is less
* likely to reclaim it.
*/
vm_page_reference(pp);
vm_page_lock(pp);
zfs_vmobject_wunlock(obj);
vm_page_busy_sleep(pp, "zfsmwb", true);
zfs_vmobject_wlock(obj);
continue;
}
vm_page_sbusy(pp);
} else if (pp != NULL) {
ASSERT(!pp->valid);
pp = NULL;
}
if (pp != NULL) {
ASSERT3U(pp->valid, ==, VM_PAGE_BITS_ALL);
vm_object_pip_add(obj, 1);
pmap_remove_write(pp);
if (nbytes != 0)
vm_page_clear_dirty(pp, off, nbytes);
}
break;
}
#else
vm_page_grab_valid_unlocked(&pp, obj, OFF_TO_IDX(start),
VM_ALLOC_NOCREAT | VM_ALLOC_SBUSY | VM_ALLOC_NORMAL |
VM_ALLOC_IGN_SBUSY);
if (pp != NULL) {
ASSERT3U(pp->valid, ==, VM_PAGE_BITS_ALL);
vm_object_pip_add(obj, 1);
pmap_remove_write(pp);
if (nbytes != 0)
vm_page_clear_dirty(pp, off, nbytes);
}
#endif
return (pp);
}
static void
page_unbusy(vm_page_t pp)
{
vm_page_sunbusy(pp);
#if __FreeBSD_version >= 1300041
vm_object_pip_wakeup(pp->object);
#else
vm_object_pip_subtract(pp->object, 1);
#endif
}
#if __FreeBSD_version > 1300051
static vm_page_t
page_hold(vnode_t *vp, int64_t start)
{
vm_object_t obj;
vm_page_t m;
obj = vp->v_object;
vm_page_grab_valid_unlocked(&m, obj, OFF_TO_IDX(start),
VM_ALLOC_NOCREAT | VM_ALLOC_WIRED | VM_ALLOC_IGN_SBUSY |
VM_ALLOC_NOBUSY);
return (m);
}
#else
static vm_page_t
page_hold(vnode_t *vp, int64_t start)
{
vm_object_t obj;
vm_page_t pp;
obj = vp->v_object;
zfs_vmobject_assert_wlocked(obj);
for (;;) {
if ((pp = vm_page_lookup(obj, OFF_TO_IDX(start))) != NULL &&
pp->valid) {
if (vm_page_xbusied(pp)) {
/*
* Reference the page before unlocking and
* sleeping so that the page daemon is less
* likely to reclaim it.
*/
vm_page_reference(pp);
vm_page_lock(pp);
zfs_vmobject_wunlock(obj);
vm_page_busy_sleep(pp, "zfsmwb", true);
zfs_vmobject_wlock(obj);
continue;
}
ASSERT3U(pp->valid, ==, VM_PAGE_BITS_ALL);
vm_page_wire_lock(pp);
vm_page_hold(pp);
vm_page_wire_unlock(pp);
} else
pp = NULL;
break;
}
return (pp);
}
#endif
static void
page_unhold(vm_page_t pp)
{
vm_page_wire_lock(pp);
#if __FreeBSD_version >= 1300035
vm_page_unwire(pp, PQ_ACTIVE);
#else
vm_page_unhold(pp);
#endif
vm_page_wire_unlock(pp);
}
/*
* When a file is memory mapped, we must keep the IO data synchronized
* between the DMU cache and the memory mapped pages. What this means:
*
* On Write: If we find a memory mapped page, we write to *both*
* the page and the dmu buffer.
*/
void
update_pages(znode_t *zp, int64_t start, int len, objset_t *os)
{
vm_object_t obj;
struct sf_buf *sf;
vnode_t *vp = ZTOV(zp);
caddr_t va;
int off;
ASSERT3P(vp->v_mount, !=, NULL);
obj = vp->v_object;
ASSERT3P(obj, !=, NULL);
off = start & PAGEOFFSET;
zfs_vmobject_wlock_12(obj);
#if __FreeBSD_version >= 1300041
vm_object_pip_add(obj, 1);
#endif
for (start &= PAGEMASK; len > 0; start += PAGESIZE) {
vm_page_t pp;
int nbytes = imin(PAGESIZE - off, len);
if ((pp = page_busy(vp, start, off, nbytes)) != NULL) {
zfs_vmobject_wunlock_12(obj);
va = zfs_map_page(pp, &sf);
(void) dmu_read(os, zp->z_id, start + off, nbytes,
va + off, DMU_READ_PREFETCH);
zfs_unmap_page(sf);
zfs_vmobject_wlock_12(obj);
page_unbusy(pp);
}
len -= nbytes;
off = 0;
}
#if __FreeBSD_version >= 1300041
vm_object_pip_wakeup(obj);
#else
vm_object_pip_wakeupn(obj, 0);
#endif
zfs_vmobject_wunlock_12(obj);
}
/*
* Read with UIO_NOCOPY flag means that sendfile(2) requests
* ZFS to populate a range of page cache pages with data.
*
* NOTE: this function could be optimized to pre-allocate
* all pages in advance, drain exclusive busy on all of them,
* map them into contiguous KVA region and populate them
* in one single dmu_read() call.
*/
int
mappedread_sf(znode_t *zp, int nbytes, zfs_uio_t *uio)
{
vnode_t *vp = ZTOV(zp);
objset_t *os = zp->z_zfsvfs->z_os;
struct sf_buf *sf;
vm_object_t obj;
vm_page_t pp;
int64_t start;
caddr_t va;
int len = nbytes;
int error = 0;
ASSERT3U(zfs_uio_segflg(uio), ==, UIO_NOCOPY);
ASSERT3P(vp->v_mount, !=, NULL);
obj = vp->v_object;
ASSERT3P(obj, !=, NULL);
ASSERT0(zfs_uio_offset(uio) & PAGEOFFSET);
zfs_vmobject_wlock_12(obj);
for (start = zfs_uio_offset(uio); len > 0; start += PAGESIZE) {
int bytes = MIN(PAGESIZE, len);
pp = vm_page_grab_unlocked(obj, OFF_TO_IDX(start),
VM_ALLOC_SBUSY | VM_ALLOC_NORMAL | VM_ALLOC_IGN_SBUSY);
if (vm_page_none_valid(pp)) {
zfs_vmobject_wunlock_12(obj);
va = zfs_map_page(pp, &sf);
error = dmu_read(os, zp->z_id, start, bytes, va,
DMU_READ_PREFETCH);
if (bytes != PAGESIZE && error == 0)
memset(va + bytes, 0, PAGESIZE - bytes);
zfs_unmap_page(sf);
zfs_vmobject_wlock_12(obj);
#if __FreeBSD_version >= 1300081
if (error == 0) {
vm_page_valid(pp);
vm_page_activate(pp);
vm_page_do_sunbusy(pp);
} else {
zfs_vmobject_wlock(obj);
if (!vm_page_wired(pp) && pp->valid == 0 &&
vm_page_busy_tryupgrade(pp))
vm_page_free(pp);
else
vm_page_sunbusy(pp);
zfs_vmobject_wunlock(obj);
}
#else
vm_page_do_sunbusy(pp);
vm_page_lock(pp);
if (error) {
if (pp->wire_count == 0 && pp->valid == 0 &&
!vm_page_busied(pp))
vm_page_free(pp);
} else {
pp->valid = VM_PAGE_BITS_ALL;
vm_page_activate(pp);
}
vm_page_unlock(pp);
#endif
} else {
ASSERT3U(pp->valid, ==, VM_PAGE_BITS_ALL);
vm_page_do_sunbusy(pp);
}
if (error)
break;
zfs_uio_advance(uio, bytes);
len -= bytes;
}
zfs_vmobject_wunlock_12(obj);
return (error);
}
/*
* When a file is memory mapped, we must keep the IO data synchronized
* between the DMU cache and the memory mapped pages. What this means:
*
* On Read: We "read" preferentially from memory mapped pages,
* else we default from the dmu buffer.
*
* NOTE: We will always "break up" the IO into PAGESIZE uiomoves when
* the file is memory mapped.
*/
int
mappedread(znode_t *zp, int nbytes, zfs_uio_t *uio)
{
vnode_t *vp = ZTOV(zp);
vm_object_t obj;
int64_t start;
int len = nbytes;
int off;
int error = 0;
ASSERT3P(vp->v_mount, !=, NULL);
obj = vp->v_object;
ASSERT3P(obj, !=, NULL);
start = zfs_uio_offset(uio);
off = start & PAGEOFFSET;
zfs_vmobject_wlock_12(obj);
for (start &= PAGEMASK; len > 0; start += PAGESIZE) {
vm_page_t pp;
uint64_t bytes = MIN(PAGESIZE - off, len);
if ((pp = page_hold(vp, start))) {
struct sf_buf *sf;
caddr_t va;
zfs_vmobject_wunlock_12(obj);
va = zfs_map_page(pp, &sf);
error = vn_io_fault_uiomove(va + off, bytes,
GET_UIO_STRUCT(uio));
zfs_unmap_page(sf);
zfs_vmobject_wlock_12(obj);
page_unhold(pp);
} else {
zfs_vmobject_wunlock_12(obj);
error = dmu_read_uio_dbuf(sa_get_db(zp->z_sa_hdl),
uio, bytes);
zfs_vmobject_wlock_12(obj);
}
len -= bytes;
off = 0;
if (error)
break;
}
zfs_vmobject_wunlock_12(obj);
return (error);
}
int
zfs_write_simple(znode_t *zp, const void *data, size_t len,
loff_t pos, size_t *presid)
{
int error = 0;
ssize_t resid;
error = vn_rdwr(UIO_WRITE, ZTOV(zp), __DECONST(void *, data), len, pos,
UIO_SYSSPACE, IO_SYNC, kcred, NOCRED, &resid, curthread);
if (error) {
return (SET_ERROR(error));
} else if (presid == NULL) {
if (resid != 0) {
error = SET_ERROR(EIO);
}
} else {
*presid = resid;
}
return (error);
}
void
zfs_zrele_async(znode_t *zp)
{
vnode_t *vp = ZTOV(zp);
objset_t *os = ITOZSB(vp)->z_os;
VN_RELE_ASYNC(vp, dsl_pool_zrele_taskq(dmu_objset_pool(os)));
}
static int
zfs_dd_callback(struct mount *mp, void *arg, int lkflags, struct vnode **vpp)
{
int error;
*vpp = arg;
error = vn_lock(*vpp, lkflags);
if (error != 0)
vrele(*vpp);
return (error);
}
static int
zfs_lookup_lock(vnode_t *dvp, vnode_t *vp, const char *name, int lkflags)
{
znode_t *zdp = VTOZ(dvp);
zfsvfs_t *zfsvfs __unused = zdp->z_zfsvfs;
int error;
int ltype;
if (zfsvfs->z_replay == B_FALSE)
ASSERT_VOP_LOCKED(dvp, __func__);
if (name[0] == 0 || (name[0] == '.' && name[1] == 0)) {
ASSERT3P(dvp, ==, vp);
vref(dvp);
ltype = lkflags & LK_TYPE_MASK;
if (ltype != VOP_ISLOCKED(dvp)) {
if (ltype == LK_EXCLUSIVE)
vn_lock(dvp, LK_UPGRADE | LK_RETRY);
else /* if (ltype == LK_SHARED) */
vn_lock(dvp, LK_DOWNGRADE | LK_RETRY);
/*
* Relock for the "." case could leave us with
* reclaimed vnode.
*/
if (VN_IS_DOOMED(dvp)) {
vrele(dvp);
return (SET_ERROR(ENOENT));
}
}
return (0);
} else if (name[0] == '.' && name[1] == '.' && name[2] == 0) {
/*
* Note that in this case, dvp is the child vnode, and we
* are looking up the parent vnode - exactly reverse from
* normal operation. Unlocking dvp requires some rather
* tricky unlock/relock dance to prevent mp from being freed;
* use vn_vget_ino_gen() which takes care of all that.
*
* XXX Note that there is a time window when both vnodes are
* unlocked. It is possible, although highly unlikely, that
* during that window the parent-child relationship between
* the vnodes may change, for example, get reversed.
* In that case we would have a wrong lock order for the vnodes.
* All other filesystems seem to ignore this problem, so we
* do the same here.
* A potential solution could be implemented as follows:
* - using LK_NOWAIT when locking the second vnode and retrying
* if necessary
* - checking that the parent-child relationship still holds
* after locking both vnodes and retrying if it doesn't
*/
error = vn_vget_ino_gen(dvp, zfs_dd_callback, vp, lkflags, &vp);
return (error);
} else {
error = vn_lock(vp, lkflags);
if (error != 0)
vrele(vp);
return (error);
}
}
/*
* Lookup an entry in a directory, or an extended attribute directory.
* If it exists, return a held vnode reference for it.
*
* IN: dvp - vnode of directory to search.
* nm - name of entry to lookup.
* pnp - full pathname to lookup [UNUSED].
* flags - LOOKUP_XATTR set if looking for an attribute.
* rdir - root directory vnode [UNUSED].
* cr - credentials of caller.
* ct - caller context
*
* OUT: vpp - vnode of located entry, NULL if not found.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* NA
*/
static int
zfs_lookup(vnode_t *dvp, const char *nm, vnode_t **vpp,
struct componentname *cnp, int nameiop, cred_t *cr, int flags,
boolean_t cached)
{
znode_t *zdp = VTOZ(dvp);
znode_t *zp;
zfsvfs_t *zfsvfs = zdp->z_zfsvfs;
#if __FreeBSD_version > 1300124
seqc_t dvp_seqc;
#endif
int error = 0;
/*
* Fast path lookup, however we must skip DNLC lookup
* for case folding or normalizing lookups because the
* DNLC code only stores the passed in name. This means
* creating 'a' and removing 'A' on a case insensitive
* file system would work, but DNLC still thinks 'a'
* exists and won't let you create it again on the next
* pass through fast path.
*/
if (!(flags & LOOKUP_XATTR)) {
if (dvp->v_type != VDIR) {
return (SET_ERROR(ENOTDIR));
} else if (zdp->z_sa_hdl == NULL) {
return (SET_ERROR(EIO));
}
}
DTRACE_PROBE2(zfs__fastpath__lookup__miss, vnode_t *, dvp,
const char *, nm);
if ((error = zfs_enter_verify_zp(zfsvfs, zdp, FTAG)) != 0)
return (error);
#if __FreeBSD_version > 1300124
dvp_seqc = vn_seqc_read_notmodify(dvp);
#endif
*vpp = NULL;
if (flags & LOOKUP_XATTR) {
/*
* If the xattr property is off, refuse the lookup request.
*/
if (!(zfsvfs->z_flags & ZSB_XATTR)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EOPNOTSUPP));
}
/*
* We don't allow recursive attributes..
* Maybe someday we will.
*/
if (zdp->z_pflags & ZFS_XATTR) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if ((error = zfs_get_xattrdir(VTOZ(dvp), &zp, cr, flags))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
*vpp = ZTOV(zp);
/*
* Do we have permission to get into attribute directory?
*/
error = zfs_zaccess(zp, ACE_EXECUTE, 0, B_FALSE, cr, NULL);
if (error) {
vrele(ZTOV(zp));
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Check accessibility of directory if we're not coming in via
* VOP_CACHEDLOOKUP.
*/
if (!cached) {
#ifdef NOEXECCHECK
if ((cnp->cn_flags & NOEXECCHECK) != 0) {
cnp->cn_flags &= ~NOEXECCHECK;
} else
#endif
if ((error = zfs_zaccess(zdp, ACE_EXECUTE, 0, B_FALSE, cr,
NULL))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
if (zfsvfs->z_utf8 && u8_validate(nm, strlen(nm),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
/*
* First handle the special cases.
*/
if ((cnp->cn_flags & ISDOTDOT) != 0) {
/*
* If we are a snapshot mounted under .zfs, return
* the vp for the snapshot directory.
*/
if (zdp->z_id == zfsvfs->z_root && zfsvfs->z_parent != zfsvfs) {
struct componentname cn;
vnode_t *zfsctl_vp;
int ltype;
zfs_exit(zfsvfs, FTAG);
ltype = VOP_ISLOCKED(dvp);
VOP_UNLOCK1(dvp);
error = zfsctl_root(zfsvfs->z_parent, LK_SHARED,
&zfsctl_vp);
if (error == 0) {
cn.cn_nameptr = "snapshot";
cn.cn_namelen = strlen(cn.cn_nameptr);
cn.cn_nameiop = cnp->cn_nameiop;
cn.cn_flags = cnp->cn_flags & ~ISDOTDOT;
cn.cn_lkflags = cnp->cn_lkflags;
error = VOP_LOOKUP(zfsctl_vp, vpp, &cn);
vput(zfsctl_vp);
}
vn_lock(dvp, ltype | LK_RETRY);
return (error);
}
}
if (zfs_has_ctldir(zdp) && strcmp(nm, ZFS_CTLDIR_NAME) == 0) {
zfs_exit(zfsvfs, FTAG);
if ((cnp->cn_flags & ISLASTCN) != 0 && nameiop != LOOKUP)
return (SET_ERROR(ENOTSUP));
error = zfsctl_root(zfsvfs, cnp->cn_lkflags, vpp);
return (error);
}
/*
* The loop is retry the lookup if the parent-child relationship
* changes during the dot-dot locking complexities.
*/
for (;;) {
uint64_t parent;
error = zfs_dirlook(zdp, nm, &zp);
if (error == 0)
*vpp = ZTOV(zp);
zfs_exit(zfsvfs, FTAG);
if (error != 0)
break;
error = zfs_lookup_lock(dvp, *vpp, nm, cnp->cn_lkflags);
if (error != 0) {
/*
* If we've got a locking error, then the vnode
* got reclaimed because of a force unmount.
* We never enter doomed vnodes into the name cache.
*/
*vpp = NULL;
return (error);
}
if ((cnp->cn_flags & ISDOTDOT) == 0)
break;
if ((error = zfs_enter(zfsvfs, FTAG)) != 0) {
vput(ZTOV(zp));
*vpp = NULL;
return (error);
}
if (zdp->z_sa_hdl == NULL) {
error = SET_ERROR(EIO);
} else {
error = sa_lookup(zdp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs),
&parent, sizeof (parent));
}
if (error != 0) {
zfs_exit(zfsvfs, FTAG);
vput(ZTOV(zp));
break;
}
if (zp->z_id == parent) {
zfs_exit(zfsvfs, FTAG);
break;
}
vput(ZTOV(zp));
}
if (error != 0)
*vpp = NULL;
/* Translate errors and add SAVENAME when needed. */
if (cnp->cn_flags & ISLASTCN) {
switch (nameiop) {
case CREATE:
case RENAME:
if (error == ENOENT) {
error = EJUSTRETURN;
#if __FreeBSD_version < 1400068
cnp->cn_flags |= SAVENAME;
#endif
break;
}
zfs_fallthrough;
case DELETE:
#if __FreeBSD_version < 1400068
if (error == 0)
cnp->cn_flags |= SAVENAME;
#endif
break;
}
}
#if __FreeBSD_version > 1300124
if ((cnp->cn_flags & ISDOTDOT) != 0) {
/*
* FIXME: zfs_lookup_lock relocks vnodes and does nothing to
* handle races. In particular different callers may end up
* with different vnodes and will try to add conflicting
* entries to the namecache.
*
* While finding different result may be acceptable in face
* of concurrent modification, adding conflicting entries
* trips over an assert in the namecache.
*
* Ultimately let an entry through once everything settles.
*/
if (!vn_seqc_consistent(dvp, dvp_seqc)) {
cnp->cn_flags &= ~MAKEENTRY;
}
}
#endif
/* Insert name into cache (as non-existent) if appropriate. */
if (zfsvfs->z_use_namecache && !zfsvfs->z_replay &&
error == ENOENT && (cnp->cn_flags & MAKEENTRY) != 0)
cache_enter(dvp, NULL, cnp);
/* Insert name into cache if appropriate. */
if (zfsvfs->z_use_namecache && !zfsvfs->z_replay &&
error == 0 && (cnp->cn_flags & MAKEENTRY)) {
if (!(cnp->cn_flags & ISLASTCN) ||
(nameiop != DELETE && nameiop != RENAME)) {
cache_enter(dvp, *vpp, cnp);
}
}
return (error);
}
/*
* Attempt to create a new entry in a directory. If the entry
* already exists, truncate the file if permissible, else return
* an error. Return the vp of the created or trunc'd file.
*
* IN: dvp - vnode of directory to put new file entry in.
* name - name of new file entry.
* vap - attributes of new file.
* excl - flag indicating exclusive or non-exclusive mode.
* mode - mode to open file with.
* cr - credentials of caller.
* flag - large file flag [UNUSED].
* ct - caller context
* vsecp - ACL to be set
* mnt_ns - Unused on FreeBSD
*
* OUT: vpp - vnode of created or trunc'd entry.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dvp - ctime|mtime updated if new entry created
* vp - ctime|mtime always, atime if new
*/
int
zfs_create(znode_t *dzp, const char *name, vattr_t *vap, int excl, int mode,
- znode_t **zpp, cred_t *cr, int flag, vsecattr_t *vsecp, zuserns_t *mnt_ns)
+ znode_t **zpp, cred_t *cr, int flag, vsecattr_t *vsecp, zidmap_t *mnt_ns)
{
(void) excl, (void) mode, (void) flag;
znode_t *zp;
zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
zilog_t *zilog;
objset_t *os;
dmu_tx_t *tx;
int error;
uid_t uid = crgetuid(cr);
gid_t gid = crgetgid(cr);
uint64_t projid = ZFS_DEFAULT_PROJID;
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
uint64_t txtype;
#ifdef DEBUG_VFS_LOCKS
vnode_t *dvp = ZTOV(dzp);
#endif
/*
* If we have an ephemeral id, ACL, or XVATTR then
* make sure file system is at proper version
*/
if (zfsvfs->z_use_fuids == B_FALSE &&
(vsecp || (vap->va_mask & AT_XVATTR) ||
IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid)))
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
os = zfsvfs->z_os;
zilog = zfsvfs->z_log;
if (zfsvfs->z_utf8 && u8_validate(name, strlen(name),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (vap->va_mask & AT_XVATTR) {
if ((error = secpolicy_xvattr(ZTOV(dzp), (xvattr_t *)vap,
crgetuid(cr), cr, vap->va_type)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
*zpp = NULL;
if ((vap->va_mode & S_ISVTX) && secpolicy_vnode_stky_modify(cr))
vap->va_mode &= ~S_ISVTX;
error = zfs_dirent_lookup(dzp, name, &zp, ZNEW);
if (error) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
ASSERT3P(zp, ==, NULL);
/*
* Create a new file object and update the directory
* to reference it.
*/
if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) {
goto out;
}
/*
* We only support the creation of regular files in
* extended attribute directories.
*/
if ((dzp->z_pflags & ZFS_XATTR) &&
(vap->va_type != VREG)) {
error = SET_ERROR(EINVAL);
goto out;
}
if ((error = zfs_acl_ids_create(dzp, 0, vap,
cr, vsecp, &acl_ids, NULL)) != 0)
goto out;
if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode))
projid = zfs_inherit_projid(dzp);
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, projid)) {
zfs_acl_ids_free(&acl_ids);
error = SET_ERROR(EDQUOT);
goto out;
}
getnewvnode_reserve_();
tx = dmu_tx_create(os);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name);
dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE);
if (!zfsvfs->z_use_sa &&
acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, acl_ids.z_aclp->z_acl_bytes);
}
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
getnewvnode_drop_reserve();
zfs_exit(zfsvfs, FTAG);
return (error);
}
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
(void) zfs_link_create(dzp, name, zp, tx, ZNEW);
txtype = zfs_log_create_txtype(Z_FILE, vsecp, vap);
zfs_log_create(zilog, tx, txtype, dzp, zp, name,
vsecp, acl_ids.z_fuidp, vap);
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
getnewvnode_drop_reserve();
out:
VNCHECKREF(dvp);
if (error == 0) {
*zpp = zp;
}
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Remove an entry from a directory.
*
* IN: dvp - vnode of directory to remove entry from.
* name - name of entry to remove.
* cr - credentials of caller.
* ct - caller context
* flags - case flags
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dvp - ctime|mtime
* vp - ctime (if nlink > 0)
*/
static int
zfs_remove_(vnode_t *dvp, vnode_t *vp, const char *name, cred_t *cr)
{
znode_t *dzp = VTOZ(dvp);
znode_t *zp;
znode_t *xzp;
zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
zilog_t *zilog;
uint64_t xattr_obj;
uint64_t obj = 0;
dmu_tx_t *tx;
boolean_t unlinked;
uint64_t txtype;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zp = VTOZ(vp);
if ((error = zfs_verify_zp(zp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
zilog = zfsvfs->z_log;
xattr_obj = 0;
xzp = NULL;
if ((error = zfs_zaccess_delete(dzp, zp, cr, NULL))) {
goto out;
}
/*
* Need to use rmdir for removing directories.
*/
if (vp->v_type == VDIR) {
error = SET_ERROR(EPERM);
goto out;
}
vnevent_remove(vp, dvp, name, ct);
obj = zp->z_id;
/* are there any extended attributes? */
error = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj, sizeof (xattr_obj));
if (error == 0 && xattr_obj) {
error = zfs_zget(zfsvfs, xattr_obj, &xzp);
ASSERT0(error);
}
/*
* We may delete the znode now, or we may put it in the unlinked set;
* it depends on whether we're the last link, and on whether there are
* other holds on the vnode. So we dmu_tx_hold() the right things to
* allow for either case.
*/
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
zfs_sa_upgrade_txholds(tx, dzp);
if (xzp) {
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
dmu_tx_hold_sa(tx, xzp->z_sa_hdl, B_FALSE);
}
/* charge as an update -- would be nice not to charge at all */
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
/*
* Mark this transaction as typically resulting in a net free of space
*/
dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Remove the directory entry.
*/
error = zfs_link_destroy(dzp, name, zp, tx, ZEXISTS, &unlinked);
if (error) {
dmu_tx_commit(tx);
goto out;
}
if (unlinked) {
zfs_unlinked_add(zp, tx);
vp->v_vflag |= VV_NOSYNC;
}
/* XXX check changes to linux vnops */
txtype = TX_REMOVE;
zfs_log_remove(zilog, tx, txtype, dzp, name, obj, unlinked);
dmu_tx_commit(tx);
out:
if (xzp)
vrele(ZTOV(xzp));
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
static int
zfs_lookup_internal(znode_t *dzp, const char *name, vnode_t **vpp,
struct componentname *cnp, int nameiop)
{
zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
int error;
cnp->cn_nameptr = __DECONST(char *, name);
cnp->cn_namelen = strlen(name);
cnp->cn_nameiop = nameiop;
cnp->cn_flags = ISLASTCN;
#if __FreeBSD_version < 1400068
cnp->cn_flags |= SAVENAME;
#endif
cnp->cn_lkflags = LK_EXCLUSIVE | LK_RETRY;
cnp->cn_cred = kcred;
#if __FreeBSD_version < 1400037
cnp->cn_thread = curthread;
#endif
if (zfsvfs->z_use_namecache && !zfsvfs->z_replay) {
struct vop_lookup_args a;
a.a_gen.a_desc = &vop_lookup_desc;
a.a_dvp = ZTOV(dzp);
a.a_vpp = vpp;
a.a_cnp = cnp;
error = vfs_cache_lookup(&a);
} else {
error = zfs_lookup(ZTOV(dzp), name, vpp, cnp, nameiop, kcred, 0,
B_FALSE);
}
#ifdef ZFS_DEBUG
if (error) {
printf("got error %d on name %s on op %d\n", error, name,
nameiop);
kdb_backtrace();
}
#endif
return (error);
}
int
zfs_remove(znode_t *dzp, const char *name, cred_t *cr, int flags)
{
vnode_t *vp;
int error;
struct componentname cn;
if ((error = zfs_lookup_internal(dzp, name, &vp, &cn, DELETE)))
return (error);
error = zfs_remove_(ZTOV(dzp), vp, name, cr);
vput(vp);
return (error);
}
/*
* Create a new directory and insert it into dvp using the name
* provided. Return a pointer to the inserted directory.
*
* IN: dvp - vnode of directory to add subdir to.
* dirname - name of new directory.
* vap - attributes of new directory.
* cr - credentials of caller.
* ct - caller context
* flags - case flags
* vsecp - ACL to be set
* mnt_ns - Unused on FreeBSD
*
* OUT: vpp - vnode of created directory.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dvp - ctime|mtime updated
* vp - ctime|mtime|atime updated
*/
int
zfs_mkdir(znode_t *dzp, const char *dirname, vattr_t *vap, znode_t **zpp,
- cred_t *cr, int flags, vsecattr_t *vsecp, zuserns_t *mnt_ns)
+ cred_t *cr, int flags, vsecattr_t *vsecp, zidmap_t *mnt_ns)
{
(void) flags, (void) vsecp;
znode_t *zp;
zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
zilog_t *zilog;
uint64_t txtype;
dmu_tx_t *tx;
int error;
uid_t uid = crgetuid(cr);
gid_t gid = crgetgid(cr);
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
ASSERT3U(vap->va_type, ==, VDIR);
/*
* If we have an ephemeral id, ACL, or XVATTR then
* make sure file system is at proper version
*/
if (zfsvfs->z_use_fuids == B_FALSE &&
((vap->va_mask & AT_XVATTR) ||
IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid)))
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (dzp->z_pflags & ZFS_XATTR) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (zfsvfs->z_utf8 && u8_validate(dirname,
strlen(dirname), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (vap->va_mask & AT_XVATTR) {
if ((error = secpolicy_xvattr(ZTOV(dzp), (xvattr_t *)vap,
crgetuid(cr), cr, vap->va_type)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
if ((error = zfs_acl_ids_create(dzp, 0, vap, cr,
NULL, &acl_ids, NULL)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* First make sure the new directory doesn't exist.
*
* Existence is checked first to make sure we don't return
* EACCES instead of EEXIST which can cause some applications
* to fail.
*/
*zpp = NULL;
if ((error = zfs_dirent_lookup(dzp, dirname, &zp, ZNEW))) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (error);
}
ASSERT3P(zp, ==, NULL);
if ((error = zfs_zaccess(dzp, ACE_ADD_SUBDIRECTORY, 0, B_FALSE, cr,
mnt_ns))) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, zfs_inherit_projid(dzp))) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EDQUOT));
}
/*
* Add a new entry to the directory.
*/
getnewvnode_reserve_();
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, dirname);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0,
acl_ids.z_aclp->z_acl_bytes);
}
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
getnewvnode_drop_reserve();
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Create new node.
*/
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
/*
* Now put new name in parent dir.
*/
(void) zfs_link_create(dzp, dirname, zp, tx, ZNEW);
*zpp = zp;
txtype = zfs_log_create_txtype(Z_DIR, NULL, vap);
zfs_log_create(zilog, tx, txtype, dzp, zp, dirname, NULL,
acl_ids.z_fuidp, vap);
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
getnewvnode_drop_reserve();
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (0);
}
#if __FreeBSD_version < 1300124
static void
cache_vop_rmdir(struct vnode *dvp, struct vnode *vp)
{
cache_purge(dvp);
cache_purge(vp);
}
#endif
/*
* Remove a directory subdir entry. If the current working
* directory is the same as the subdir to be removed, the
* remove will fail.
*
* IN: dvp - vnode of directory to remove from.
* name - name of directory to be removed.
* cwd - vnode of current working directory.
* cr - credentials of caller.
* ct - caller context
* flags - case flags
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dvp - ctime|mtime updated
*/
static int
zfs_rmdir_(vnode_t *dvp, vnode_t *vp, const char *name, cred_t *cr)
{
znode_t *dzp = VTOZ(dvp);
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
zilog_t *zilog;
dmu_tx_t *tx;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
if ((error = zfs_verify_zp(zp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
zilog = zfsvfs->z_log;
if ((error = zfs_zaccess_delete(dzp, zp, cr, NULL))) {
goto out;
}
if (vp->v_type != VDIR) {
error = SET_ERROR(ENOTDIR);
goto out;
}
vnevent_rmdir(vp, dvp, name, ct);
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
zfs_sa_upgrade_txholds(tx, zp);
zfs_sa_upgrade_txholds(tx, dzp);
dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
error = zfs_link_destroy(dzp, name, zp, tx, ZEXISTS, NULL);
if (error == 0) {
uint64_t txtype = TX_RMDIR;
zfs_log_remove(zilog, tx, txtype, dzp, name,
ZFS_NO_OBJECT, B_FALSE);
}
dmu_tx_commit(tx);
if (zfsvfs->z_use_namecache)
cache_vop_rmdir(dvp, vp);
out:
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
int
zfs_rmdir(znode_t *dzp, const char *name, znode_t *cwd, cred_t *cr, int flags)
{
struct componentname cn;
vnode_t *vp;
int error;
if ((error = zfs_lookup_internal(dzp, name, &vp, &cn, DELETE)))
return (error);
error = zfs_rmdir_(ZTOV(dzp), vp, name, cr);
vput(vp);
return (error);
}
/*
* Read as many directory entries as will fit into the provided
* buffer from the given directory cursor position (specified in
* the uio structure).
*
* IN: vp - vnode of directory to read.
* uio - structure supplying read location, range info,
* and return buffer.
* cr - credentials of caller.
* ct - caller context
*
* OUT: uio - updated offset and range, buffer filled.
* eofp - set to true if end-of-file detected.
* ncookies- number of entries in cookies
* cookies - offsets to directory entries
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* vp - atime updated
*
* Note that the low 4 bits of the cookie returned by zap is always zero.
* This allows us to use the low range for "special" directory entries:
* We use 0 for '.', and 1 for '..'. If this is the root of the filesystem,
* we use the offset 2 for the '.zfs' directory.
*/
static int
zfs_readdir(vnode_t *vp, zfs_uio_t *uio, cred_t *cr, int *eofp,
int *ncookies, cookie_t **cookies)
{
znode_t *zp = VTOZ(vp);
iovec_t *iovp;
dirent64_t *odp;
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
objset_t *os;
caddr_t outbuf;
size_t bufsize;
zap_cursor_t zc;
zap_attribute_t zap;
uint_t bytes_wanted;
uint64_t offset; /* must be unsigned; checks for < 1 */
uint64_t parent;
int local_eof;
int outcount;
int error;
uint8_t prefetch;
uint8_t type;
int ncooks;
cookie_t *cooks = NULL;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs),
&parent, sizeof (parent))) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* If we are not given an eof variable,
* use a local one.
*/
if (eofp == NULL)
eofp = &local_eof;
/*
* Check for valid iov_len.
*/
if (GET_UIO_STRUCT(uio)->uio_iov->iov_len <= 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Quit if directory has been removed (posix)
*/
if ((*eofp = zp->z_unlinked) != 0) {
zfs_exit(zfsvfs, FTAG);
return (0);
}
error = 0;
os = zfsvfs->z_os;
offset = zfs_uio_offset(uio);
prefetch = zp->z_zn_prefetch;
/*
* Initialize the iterator cursor.
*/
if (offset <= 3) {
/*
* Start iteration from the beginning of the directory.
*/
zap_cursor_init(&zc, os, zp->z_id);
} else {
/*
* The offset is a serialized cursor.
*/
zap_cursor_init_serialized(&zc, os, zp->z_id, offset);
}
/*
* Get space to change directory entries into fs independent format.
*/
iovp = GET_UIO_STRUCT(uio)->uio_iov;
bytes_wanted = iovp->iov_len;
if (zfs_uio_segflg(uio) != UIO_SYSSPACE || zfs_uio_iovcnt(uio) != 1) {
bufsize = bytes_wanted;
outbuf = kmem_alloc(bufsize, KM_SLEEP);
odp = (struct dirent64 *)outbuf;
} else {
bufsize = bytes_wanted;
outbuf = NULL;
odp = (struct dirent64 *)iovp->iov_base;
}
if (ncookies != NULL) {
/*
* Minimum entry size is dirent size and 1 byte for a file name.
*/
ncooks = zfs_uio_resid(uio) / (sizeof (struct dirent) -
sizeof (((struct dirent *)NULL)->d_name) + 1);
cooks = malloc(ncooks * sizeof (*cooks), M_TEMP, M_WAITOK);
*cookies = cooks;
*ncookies = ncooks;
}
/*
* Transform to file-system independent format
*/
outcount = 0;
while (outcount < bytes_wanted) {
ino64_t objnum;
ushort_t reclen;
off64_t *next = NULL;
/*
* Special case `.', `..', and `.zfs'.
*/
if (offset == 0) {
(void) strcpy(zap.za_name, ".");
zap.za_normalization_conflict = 0;
objnum = zp->z_id;
type = DT_DIR;
} else if (offset == 1) {
(void) strcpy(zap.za_name, "..");
zap.za_normalization_conflict = 0;
objnum = parent;
type = DT_DIR;
} else if (offset == 2 && zfs_show_ctldir(zp)) {
(void) strcpy(zap.za_name, ZFS_CTLDIR_NAME);
zap.za_normalization_conflict = 0;
objnum = ZFSCTL_INO_ROOT;
type = DT_DIR;
} else {
/*
* Grab next entry.
*/
if ((error = zap_cursor_retrieve(&zc, &zap))) {
if ((*eofp = (error == ENOENT)) != 0)
break;
else
goto update;
}
if (zap.za_integer_length != 8 ||
zap.za_num_integers != 1) {
cmn_err(CE_WARN, "zap_readdir: bad directory "
"entry, obj = %lld, offset = %lld\n",
(u_longlong_t)zp->z_id,
(u_longlong_t)offset);
error = SET_ERROR(ENXIO);
goto update;
}
objnum = ZFS_DIRENT_OBJ(zap.za_first_integer);
/*
* MacOS X can extract the object type here such as:
* uint8_t type = ZFS_DIRENT_TYPE(zap.za_first_integer);
*/
type = ZFS_DIRENT_TYPE(zap.za_first_integer);
}
reclen = DIRENT64_RECLEN(strlen(zap.za_name));
/*
* Will this entry fit in the buffer?
*/
if (outcount + reclen > bufsize) {
/*
* Did we manage to fit anything in the buffer?
*/
if (!outcount) {
error = SET_ERROR(EINVAL);
goto update;
}
break;
}
/*
* Add normal entry:
*/
odp->d_ino = objnum;
odp->d_reclen = reclen;
odp->d_namlen = strlen(zap.za_name);
/* NOTE: d_off is the offset for the *next* entry. */
next = &odp->d_off;
strlcpy(odp->d_name, zap.za_name, odp->d_namlen + 1);
odp->d_type = type;
dirent_terminate(odp);
odp = (dirent64_t *)((intptr_t)odp + reclen);
outcount += reclen;
ASSERT3S(outcount, <=, bufsize);
/* Prefetch znode */
if (prefetch)
dmu_prefetch(os, objnum, 0, 0, 0,
ZIO_PRIORITY_SYNC_READ);
/*
* Move to the next entry, fill in the previous offset.
*/
if (offset > 2 || (offset == 2 && !zfs_show_ctldir(zp))) {
zap_cursor_advance(&zc);
offset = zap_cursor_serialize(&zc);
} else {
offset += 1;
}
/* Fill the offset right after advancing the cursor. */
if (next != NULL)
*next = offset;
if (cooks != NULL) {
*cooks++ = offset;
ncooks--;
KASSERT(ncooks >= 0, ("ncookies=%d", ncooks));
}
}
zp->z_zn_prefetch = B_FALSE; /* a lookup will re-enable pre-fetching */
/* Subtract unused cookies */
if (ncookies != NULL)
*ncookies -= ncooks;
if (zfs_uio_segflg(uio) == UIO_SYSSPACE && zfs_uio_iovcnt(uio) == 1) {
iovp->iov_base += outcount;
iovp->iov_len -= outcount;
zfs_uio_resid(uio) -= outcount;
} else if ((error =
zfs_uiomove(outbuf, (long)outcount, UIO_READ, uio))) {
/*
* Reset the pointer.
*/
offset = zfs_uio_offset(uio);
}
update:
zap_cursor_fini(&zc);
if (zfs_uio_segflg(uio) != UIO_SYSSPACE || zfs_uio_iovcnt(uio) != 1)
kmem_free(outbuf, bufsize);
if (error == ENOENT)
error = 0;
ZFS_ACCESSTIME_STAMP(zfsvfs, zp);
zfs_uio_setoffset(uio, offset);
zfs_exit(zfsvfs, FTAG);
if (error != 0 && cookies != NULL) {
free(*cookies, M_TEMP);
*cookies = NULL;
*ncookies = 0;
}
return (error);
}
/*
* Get the requested file attributes and place them in the provided
* vattr structure.
*
* IN: vp - vnode of file.
* vap - va_mask identifies requested attributes.
* If AT_XVATTR set, then optional attrs are requested
* flags - ATTR_NOACLCHECK (CIFS server context)
* cr - credentials of caller.
*
* OUT: vap - attribute values.
*
* RETURN: 0 (always succeeds).
*/
static int
zfs_getattr(vnode_t *vp, vattr_t *vap, int flags, cred_t *cr)
{
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
int error = 0;
uint32_t blksize;
u_longlong_t nblocks;
uint64_t mtime[2], ctime[2], crtime[2], rdev;
xvattr_t *xvap = (xvattr_t *)vap; /* vap may be an xvattr_t * */
xoptattr_t *xoap = NULL;
boolean_t skipaclchk = (flags & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
sa_bulk_attr_t bulk[4];
int count = 0;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
zfs_fuid_map_ids(zp, cr, &vap->va_uid, &vap->va_gid);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CRTIME(zfsvfs), NULL, &crtime, 16);
if (vp->v_type == VBLK || vp->v_type == VCHR)
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_RDEV(zfsvfs), NULL,
&rdev, 8);
if ((error = sa_bulk_lookup(zp->z_sa_hdl, bulk, count)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* If ACL is trivial don't bother looking for ACE_READ_ATTRIBUTES.
* Also, if we are the owner don't bother, since owner should
* always be allowed to read basic attributes of file.
*/
if (!(zp->z_pflags & ZFS_ACL_TRIVIAL) &&
(vap->va_uid != crgetuid(cr))) {
if ((error = zfs_zaccess(zp, ACE_READ_ATTRIBUTES, 0,
skipaclchk, cr, NULL))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
/*
* Return all attributes. It's cheaper to provide the answer
* than to determine whether we were asked the question.
*/
vap->va_type = IFTOVT(zp->z_mode);
vap->va_mode = zp->z_mode & ~S_IFMT;
vn_fsid(vp, vap);
vap->va_nodeid = zp->z_id;
vap->va_nlink = zp->z_links;
if ((vp->v_flag & VROOT) && zfs_show_ctldir(zp) &&
zp->z_links < ZFS_LINK_MAX)
vap->va_nlink++;
vap->va_size = zp->z_size;
if (vp->v_type == VBLK || vp->v_type == VCHR)
vap->va_rdev = zfs_cmpldev(rdev);
vap->va_gen = zp->z_gen;
vap->va_flags = 0; /* FreeBSD: Reset chflags(2) flags. */
vap->va_filerev = zp->z_seq;
/*
* Add in any requested optional attributes and the create time.
* Also set the corresponding bits in the returned attribute bitmap.
*/
if ((xoap = xva_getxoptattr(xvap)) != NULL && zfsvfs->z_use_fuids) {
if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE)) {
xoap->xoa_archive =
((zp->z_pflags & ZFS_ARCHIVE) != 0);
XVA_SET_RTN(xvap, XAT_ARCHIVE);
}
if (XVA_ISSET_REQ(xvap, XAT_READONLY)) {
xoap->xoa_readonly =
((zp->z_pflags & ZFS_READONLY) != 0);
XVA_SET_RTN(xvap, XAT_READONLY);
}
if (XVA_ISSET_REQ(xvap, XAT_SYSTEM)) {
xoap->xoa_system =
((zp->z_pflags & ZFS_SYSTEM) != 0);
XVA_SET_RTN(xvap, XAT_SYSTEM);
}
if (XVA_ISSET_REQ(xvap, XAT_HIDDEN)) {
xoap->xoa_hidden =
((zp->z_pflags & ZFS_HIDDEN) != 0);
XVA_SET_RTN(xvap, XAT_HIDDEN);
}
if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) {
xoap->xoa_nounlink =
((zp->z_pflags & ZFS_NOUNLINK) != 0);
XVA_SET_RTN(xvap, XAT_NOUNLINK);
}
if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) {
xoap->xoa_immutable =
((zp->z_pflags & ZFS_IMMUTABLE) != 0);
XVA_SET_RTN(xvap, XAT_IMMUTABLE);
}
if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) {
xoap->xoa_appendonly =
((zp->z_pflags & ZFS_APPENDONLY) != 0);
XVA_SET_RTN(xvap, XAT_APPENDONLY);
}
if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) {
xoap->xoa_nodump =
((zp->z_pflags & ZFS_NODUMP) != 0);
XVA_SET_RTN(xvap, XAT_NODUMP);
}
if (XVA_ISSET_REQ(xvap, XAT_OPAQUE)) {
xoap->xoa_opaque =
((zp->z_pflags & ZFS_OPAQUE) != 0);
XVA_SET_RTN(xvap, XAT_OPAQUE);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) {
xoap->xoa_av_quarantined =
((zp->z_pflags & ZFS_AV_QUARANTINED) != 0);
XVA_SET_RTN(xvap, XAT_AV_QUARANTINED);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) {
xoap->xoa_av_modified =
((zp->z_pflags & ZFS_AV_MODIFIED) != 0);
XVA_SET_RTN(xvap, XAT_AV_MODIFIED);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP) &&
vp->v_type == VREG) {
zfs_sa_get_scanstamp(zp, xvap);
}
if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) {
xoap->xoa_reparse = ((zp->z_pflags & ZFS_REPARSE) != 0);
XVA_SET_RTN(xvap, XAT_REPARSE);
}
if (XVA_ISSET_REQ(xvap, XAT_GEN)) {
xoap->xoa_generation = zp->z_gen;
XVA_SET_RTN(xvap, XAT_GEN);
}
if (XVA_ISSET_REQ(xvap, XAT_OFFLINE)) {
xoap->xoa_offline =
((zp->z_pflags & ZFS_OFFLINE) != 0);
XVA_SET_RTN(xvap, XAT_OFFLINE);
}
if (XVA_ISSET_REQ(xvap, XAT_SPARSE)) {
xoap->xoa_sparse =
((zp->z_pflags & ZFS_SPARSE) != 0);
XVA_SET_RTN(xvap, XAT_SPARSE);
}
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT)) {
xoap->xoa_projinherit =
((zp->z_pflags & ZFS_PROJINHERIT) != 0);
XVA_SET_RTN(xvap, XAT_PROJINHERIT);
}
if (XVA_ISSET_REQ(xvap, XAT_PROJID)) {
xoap->xoa_projid = zp->z_projid;
XVA_SET_RTN(xvap, XAT_PROJID);
}
}
ZFS_TIME_DECODE(&vap->va_atime, zp->z_atime);
ZFS_TIME_DECODE(&vap->va_mtime, mtime);
ZFS_TIME_DECODE(&vap->va_ctime, ctime);
ZFS_TIME_DECODE(&vap->va_birthtime, crtime);
sa_object_size(zp->z_sa_hdl, &blksize, &nblocks);
vap->va_blksize = blksize;
vap->va_bytes = nblocks << 9; /* nblocks * 512 */
if (zp->z_blksz == 0) {
/*
* Block size hasn't been set; suggest maximal I/O transfers.
*/
vap->va_blksize = zfsvfs->z_max_blksz;
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* Set the file attributes to the values contained in the
* vattr structure.
*
* IN: zp - znode of file to be modified.
* vap - new attribute values.
* If AT_XVATTR set, then optional attrs are being set
* flags - ATTR_UTIME set if non-default time values provided.
* - ATTR_NOACLCHECK (CIFS context only).
* cr - credentials of caller.
* mnt_ns - Unused on FreeBSD
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* vp - ctime updated, mtime updated if size changed.
*/
int
-zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zuserns_t *mnt_ns)
+zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zidmap_t *mnt_ns)
{
vnode_t *vp = ZTOV(zp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
objset_t *os;
zilog_t *zilog;
dmu_tx_t *tx;
vattr_t oldva;
xvattr_t tmpxvattr;
uint_t mask = vap->va_mask;
uint_t saved_mask = 0;
uint64_t saved_mode;
int trim_mask = 0;
uint64_t new_mode;
uint64_t new_uid, new_gid;
uint64_t xattr_obj;
uint64_t mtime[2], ctime[2];
uint64_t projid = ZFS_INVALID_PROJID;
znode_t *attrzp;
int need_policy = FALSE;
int err, err2;
zfs_fuid_info_t *fuidp = NULL;
xvattr_t *xvap = (xvattr_t *)vap; /* vap may be an xvattr_t * */
xoptattr_t *xoap;
zfs_acl_t *aclp;
boolean_t skipaclchk = (flags & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
boolean_t fuid_dirtied = B_FALSE;
sa_bulk_attr_t bulk[7], xattr_bulk[7];
int count = 0, xattr_count = 0;
if (mask == 0)
return (0);
if (mask & AT_NOSET)
return (SET_ERROR(EINVAL));
if ((err = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (err);
os = zfsvfs->z_os;
zilog = zfsvfs->z_log;
/*
* Make sure that if we have ephemeral uid/gid or xvattr specified
* that file system is at proper version level
*/
if (zfsvfs->z_use_fuids == B_FALSE &&
(((mask & AT_UID) && IS_EPHEMERAL(vap->va_uid)) ||
((mask & AT_GID) && IS_EPHEMERAL(vap->va_gid)) ||
(mask & AT_XVATTR))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (mask & AT_SIZE && vp->v_type == VDIR) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EISDIR));
}
if (mask & AT_SIZE && vp->v_type != VREG && vp->v_type != VFIFO) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* If this is an xvattr_t, then get a pointer to the structure of
* optional attributes. If this is NULL, then we have a vattr_t.
*/
xoap = xva_getxoptattr(xvap);
xva_init(&tmpxvattr);
/*
* Immutable files can only alter immutable bit and atime
*/
if ((zp->z_pflags & ZFS_IMMUTABLE) &&
((mask & (AT_SIZE|AT_UID|AT_GID|AT_MTIME|AT_MODE)) ||
((mask & AT_XVATTR) && XVA_ISSET_REQ(xvap, XAT_CREATETIME)))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
/*
* Note: ZFS_READONLY is handled in zfs_zaccess_common.
*/
/*
* Verify timestamps doesn't overflow 32 bits.
* ZFS can handle large timestamps, but 32bit syscalls can't
* handle times greater than 2039. This check should be removed
* once large timestamps are fully supported.
*/
if (mask & (AT_ATIME | AT_MTIME)) {
if (((mask & AT_ATIME) && TIMESPEC_OVERFLOW(&vap->va_atime)) ||
((mask & AT_MTIME) && TIMESPEC_OVERFLOW(&vap->va_mtime))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EOVERFLOW));
}
}
if (xoap != NULL && (mask & AT_XVATTR)) {
if (XVA_ISSET_REQ(xvap, XAT_CREATETIME) &&
TIMESPEC_OVERFLOW(&vap->va_birthtime)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EOVERFLOW));
}
if (XVA_ISSET_REQ(xvap, XAT_PROJID)) {
if (!dmu_objset_projectquota_enabled(os) ||
(!S_ISREG(zp->z_mode) && !S_ISDIR(zp->z_mode))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EOPNOTSUPP));
}
projid = xoap->xoa_projid;
if (unlikely(projid == ZFS_INVALID_PROJID)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (projid == zp->z_projid && zp->z_pflags & ZFS_PROJID)
projid = ZFS_INVALID_PROJID;
else
need_policy = TRUE;
}
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT) &&
(xoap->xoa_projinherit !=
((zp->z_pflags & ZFS_PROJINHERIT) != 0)) &&
(!dmu_objset_projectquota_enabled(os) ||
(!S_ISREG(zp->z_mode) && !S_ISDIR(zp->z_mode)))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EOPNOTSUPP));
}
}
attrzp = NULL;
aclp = NULL;
if (zfsvfs->z_vfs->vfs_flag & VFS_RDONLY) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EROFS));
}
/*
* First validate permissions
*/
if (mask & AT_SIZE) {
/*
* XXX - Note, we are not providing any open
* mode flags here (like FNDELAY), so we may
* block if there are locks present... this
* should be addressed in openat().
*/
/* XXX - would it be OK to generate a log record here? */
err = zfs_freesp(zp, vap->va_size, 0, 0, FALSE);
if (err) {
zfs_exit(zfsvfs, FTAG);
return (err);
}
}
if (mask & (AT_ATIME|AT_MTIME) ||
((mask & AT_XVATTR) && (XVA_ISSET_REQ(xvap, XAT_HIDDEN) ||
XVA_ISSET_REQ(xvap, XAT_READONLY) ||
XVA_ISSET_REQ(xvap, XAT_ARCHIVE) ||
XVA_ISSET_REQ(xvap, XAT_OFFLINE) ||
XVA_ISSET_REQ(xvap, XAT_SPARSE) ||
XVA_ISSET_REQ(xvap, XAT_CREATETIME) ||
XVA_ISSET_REQ(xvap, XAT_SYSTEM)))) {
need_policy = zfs_zaccess(zp, ACE_WRITE_ATTRIBUTES, 0,
skipaclchk, cr, mnt_ns);
}
if (mask & (AT_UID|AT_GID)) {
int idmask = (mask & (AT_UID|AT_GID));
int take_owner;
int take_group;
/*
* NOTE: even if a new mode is being set,
* we may clear S_ISUID/S_ISGID bits.
*/
if (!(mask & AT_MODE))
vap->va_mode = zp->z_mode;
/*
* Take ownership or chgrp to group we are a member of
*/
take_owner = (mask & AT_UID) && (vap->va_uid == crgetuid(cr));
take_group = (mask & AT_GID) &&
zfs_groupmember(zfsvfs, vap->va_gid, cr);
/*
* If both AT_UID and AT_GID are set then take_owner and
* take_group must both be set in order to allow taking
* ownership.
*
* Otherwise, send the check through secpolicy_vnode_setattr()
*
*/
if (((idmask == (AT_UID|AT_GID)) && take_owner && take_group) ||
((idmask == AT_UID) && take_owner) ||
((idmask == AT_GID) && take_group)) {
if (zfs_zaccess(zp, ACE_WRITE_OWNER, 0,
skipaclchk, cr, mnt_ns) == 0) {
/*
* Remove setuid/setgid for non-privileged users
*/
secpolicy_setid_clear(vap, vp, cr);
trim_mask = (mask & (AT_UID|AT_GID));
} else {
need_policy = TRUE;
}
} else {
need_policy = TRUE;
}
}
oldva.va_mode = zp->z_mode;
zfs_fuid_map_ids(zp, cr, &oldva.va_uid, &oldva.va_gid);
if (mask & AT_XVATTR) {
/*
* Update xvattr mask to include only those attributes
* that are actually changing.
*
* the bits will be restored prior to actually setting
* the attributes so the caller thinks they were set.
*/
if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) {
if (xoap->xoa_appendonly !=
((zp->z_pflags & ZFS_APPENDONLY) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_APPENDONLY);
XVA_SET_REQ(&tmpxvattr, XAT_APPENDONLY);
}
}
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT)) {
if (xoap->xoa_projinherit !=
((zp->z_pflags & ZFS_PROJINHERIT) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_PROJINHERIT);
XVA_SET_REQ(&tmpxvattr, XAT_PROJINHERIT);
}
}
if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) {
if (xoap->xoa_nounlink !=
((zp->z_pflags & ZFS_NOUNLINK) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_NOUNLINK);
XVA_SET_REQ(&tmpxvattr, XAT_NOUNLINK);
}
}
if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) {
if (xoap->xoa_immutable !=
((zp->z_pflags & ZFS_IMMUTABLE) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_IMMUTABLE);
XVA_SET_REQ(&tmpxvattr, XAT_IMMUTABLE);
}
}
if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) {
if (xoap->xoa_nodump !=
((zp->z_pflags & ZFS_NODUMP) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_NODUMP);
XVA_SET_REQ(&tmpxvattr, XAT_NODUMP);
}
}
if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) {
if (xoap->xoa_av_modified !=
((zp->z_pflags & ZFS_AV_MODIFIED) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_AV_MODIFIED);
XVA_SET_REQ(&tmpxvattr, XAT_AV_MODIFIED);
}
}
if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) {
if ((vp->v_type != VREG &&
xoap->xoa_av_quarantined) ||
xoap->xoa_av_quarantined !=
((zp->z_pflags & ZFS_AV_QUARANTINED) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_AV_QUARANTINED);
XVA_SET_REQ(&tmpxvattr, XAT_AV_QUARANTINED);
}
}
if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if (need_policy == FALSE &&
(XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP) ||
XVA_ISSET_REQ(xvap, XAT_OPAQUE))) {
need_policy = TRUE;
}
}
if (mask & AT_MODE) {
if (zfs_zaccess(zp, ACE_WRITE_ACL, 0, skipaclchk, cr,
mnt_ns) == 0) {
err = secpolicy_setid_setsticky_clear(vp, vap,
&oldva, cr);
if (err) {
zfs_exit(zfsvfs, FTAG);
return (err);
}
trim_mask |= AT_MODE;
} else {
need_policy = TRUE;
}
}
if (need_policy) {
/*
* If trim_mask is set then take ownership
* has been granted or write_acl is present and user
* has the ability to modify mode. In that case remove
* UID|GID and or MODE from mask so that
* secpolicy_vnode_setattr() doesn't revoke it.
*/
if (trim_mask) {
saved_mask = vap->va_mask;
vap->va_mask &= ~trim_mask;
if (trim_mask & AT_MODE) {
/*
* Save the mode, as secpolicy_vnode_setattr()
* will overwrite it with ova.va_mode.
*/
saved_mode = vap->va_mode;
}
}
err = secpolicy_vnode_setattr(cr, vp, vap, &oldva, flags,
(int (*)(void *, int, cred_t *))zfs_zaccess_unix, zp);
if (err) {
zfs_exit(zfsvfs, FTAG);
return (err);
}
if (trim_mask) {
vap->va_mask |= saved_mask;
if (trim_mask & AT_MODE) {
/*
* Recover the mode after
* secpolicy_vnode_setattr().
*/
vap->va_mode = saved_mode;
}
}
}
/*
* secpolicy_vnode_setattr, or take ownership may have
* changed va_mask
*/
mask = vap->va_mask;
if ((mask & (AT_UID | AT_GID)) || projid != ZFS_INVALID_PROJID) {
err = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj, sizeof (xattr_obj));
if (err == 0 && xattr_obj) {
err = zfs_zget(zp->z_zfsvfs, xattr_obj, &attrzp);
if (err == 0) {
err = vn_lock(ZTOV(attrzp), LK_EXCLUSIVE);
if (err != 0)
vrele(ZTOV(attrzp));
}
if (err)
goto out2;
}
if (mask & AT_UID) {
new_uid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_uid, cr, ZFS_OWNER, &fuidp);
if (new_uid != zp->z_uid &&
zfs_id_overquota(zfsvfs, DMU_USERUSED_OBJECT,
new_uid)) {
if (attrzp)
vput(ZTOV(attrzp));
err = SET_ERROR(EDQUOT);
goto out2;
}
}
if (mask & AT_GID) {
new_gid = zfs_fuid_create(zfsvfs, (uint64_t)vap->va_gid,
cr, ZFS_GROUP, &fuidp);
if (new_gid != zp->z_gid &&
zfs_id_overquota(zfsvfs, DMU_GROUPUSED_OBJECT,
new_gid)) {
if (attrzp)
vput(ZTOV(attrzp));
err = SET_ERROR(EDQUOT);
goto out2;
}
}
if (projid != ZFS_INVALID_PROJID &&
zfs_id_overquota(zfsvfs, DMU_PROJECTUSED_OBJECT, projid)) {
if (attrzp)
vput(ZTOV(attrzp));
err = SET_ERROR(EDQUOT);
goto out2;
}
}
tx = dmu_tx_create(os);
if (mask & AT_MODE) {
uint64_t pmode = zp->z_mode;
uint64_t acl_obj;
new_mode = (pmode & S_IFMT) | (vap->va_mode & ~S_IFMT);
if (zp->z_zfsvfs->z_acl_mode == ZFS_ACL_RESTRICTED &&
!(zp->z_pflags & ZFS_ACL_TRIVIAL)) {
err = SET_ERROR(EPERM);
goto out;
}
if ((err = zfs_acl_chmod_setattr(zp, &aclp, new_mode)))
goto out;
if (!zp->z_is_sa && ((acl_obj = zfs_external_acl(zp)) != 0)) {
/*
* Are we upgrading ACL from old V0 format
* to V1 format?
*/
if (zfsvfs->z_version >= ZPL_VERSION_FUID &&
zfs_znode_acl_version(zp) ==
ZFS_ACL_VERSION_INITIAL) {
dmu_tx_hold_free(tx, acl_obj, 0,
DMU_OBJECT_END);
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, aclp->z_acl_bytes);
} else {
dmu_tx_hold_write(tx, acl_obj, 0,
aclp->z_acl_bytes);
}
} else if (!zp->z_is_sa && aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, aclp->z_acl_bytes);
}
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
} else {
if (((mask & AT_XVATTR) &&
XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) ||
(projid != ZFS_INVALID_PROJID &&
!(zp->z_pflags & ZFS_PROJID)))
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
else
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
}
if (attrzp) {
dmu_tx_hold_sa(tx, attrzp->z_sa_hdl, B_FALSE);
}
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
zfs_sa_upgrade_txholds(tx, zp);
err = dmu_tx_assign(tx, TXG_WAIT);
if (err)
goto out;
count = 0;
/*
* Set each attribute requested.
* We group settings according to the locks they need to acquire.
*
* Note: you cannot set ctime directly, although it will be
* updated as a side-effect of calling this function.
*/
if (projid != ZFS_INVALID_PROJID && !(zp->z_pflags & ZFS_PROJID)) {
/*
* For the existed object that is upgraded from old system,
* its on-disk layout has no slot for the project ID attribute.
* But quota accounting logic needs to access related slots by
* offset directly. So we need to adjust old objects' layout
* to make the project ID to some unified and fixed offset.
*/
if (attrzp)
err = sa_add_projid(attrzp->z_sa_hdl, tx, projid);
if (err == 0)
err = sa_add_projid(zp->z_sa_hdl, tx, projid);
if (unlikely(err == EEXIST))
err = 0;
else if (err != 0)
goto out;
else
projid = ZFS_INVALID_PROJID;
}
if (mask & (AT_UID|AT_GID|AT_MODE))
mutex_enter(&zp->z_acl_lock);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, sizeof (zp->z_pflags));
if (attrzp) {
if (mask & (AT_UID|AT_GID|AT_MODE))
mutex_enter(&attrzp->z_acl_lock);
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_FLAGS(zfsvfs), NULL, &attrzp->z_pflags,
sizeof (attrzp->z_pflags));
if (projid != ZFS_INVALID_PROJID) {
attrzp->z_projid = projid;
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_PROJID(zfsvfs), NULL, &attrzp->z_projid,
sizeof (attrzp->z_projid));
}
}
if (mask & (AT_UID|AT_GID)) {
if (mask & AT_UID) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL,
&new_uid, sizeof (new_uid));
zp->z_uid = new_uid;
if (attrzp) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_UID(zfsvfs), NULL, &new_uid,
sizeof (new_uid));
attrzp->z_uid = new_uid;
}
}
if (mask & AT_GID) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs),
NULL, &new_gid, sizeof (new_gid));
zp->z_gid = new_gid;
if (attrzp) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_GID(zfsvfs), NULL, &new_gid,
sizeof (new_gid));
attrzp->z_gid = new_gid;
}
}
if (!(mask & AT_MODE)) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs),
NULL, &new_mode, sizeof (new_mode));
new_mode = zp->z_mode;
}
err = zfs_acl_chown_setattr(zp);
ASSERT0(err);
if (attrzp) {
vn_seqc_write_begin(ZTOV(attrzp));
err = zfs_acl_chown_setattr(attrzp);
vn_seqc_write_end(ZTOV(attrzp));
ASSERT0(err);
}
}
if (mask & AT_MODE) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL,
&new_mode, sizeof (new_mode));
zp->z_mode = new_mode;
ASSERT3P(aclp, !=, NULL);
err = zfs_aclset_common(zp, aclp, cr, tx);
ASSERT0(err);
if (zp->z_acl_cached)
zfs_acl_free(zp->z_acl_cached);
zp->z_acl_cached = aclp;
aclp = NULL;
}
if (mask & AT_ATIME) {
ZFS_TIME_ENCODE(&vap->va_atime, zp->z_atime);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL,
&zp->z_atime, sizeof (zp->z_atime));
}
if (mask & AT_MTIME) {
ZFS_TIME_ENCODE(&vap->va_mtime, mtime);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL,
mtime, sizeof (mtime));
}
if (projid != ZFS_INVALID_PROJID) {
zp->z_projid = projid;
SA_ADD_BULK_ATTR(bulk, count,
SA_ZPL_PROJID(zfsvfs), NULL, &zp->z_projid,
sizeof (zp->z_projid));
}
/* XXX - shouldn't this be done *before* the ATIME/MTIME checks? */
if (mask & AT_SIZE && !(mask & AT_MTIME)) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs),
NULL, mtime, sizeof (mtime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
&ctime, sizeof (ctime));
zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime);
} else if (mask != 0) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
&ctime, sizeof (ctime));
zfs_tstamp_update_setup(zp, STATE_CHANGED, mtime, ctime);
if (attrzp) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_CTIME(zfsvfs), NULL,
&ctime, sizeof (ctime));
zfs_tstamp_update_setup(attrzp, STATE_CHANGED,
mtime, ctime);
}
}
/*
* Do this after setting timestamps to prevent timestamp
* update from toggling bit
*/
if (xoap && (mask & AT_XVATTR)) {
if (XVA_ISSET_REQ(xvap, XAT_CREATETIME))
xoap->xoa_createtime = vap->va_birthtime;
/*
* restore trimmed off masks
* so that return masks can be set for caller.
*/
if (XVA_ISSET_REQ(&tmpxvattr, XAT_APPENDONLY)) {
XVA_SET_REQ(xvap, XAT_APPENDONLY);
}
if (XVA_ISSET_REQ(&tmpxvattr, XAT_NOUNLINK)) {
XVA_SET_REQ(xvap, XAT_NOUNLINK);
}
if (XVA_ISSET_REQ(&tmpxvattr, XAT_IMMUTABLE)) {
XVA_SET_REQ(xvap, XAT_IMMUTABLE);
}
if (XVA_ISSET_REQ(&tmpxvattr, XAT_NODUMP)) {
XVA_SET_REQ(xvap, XAT_NODUMP);
}
if (XVA_ISSET_REQ(&tmpxvattr, XAT_AV_MODIFIED)) {
XVA_SET_REQ(xvap, XAT_AV_MODIFIED);
}
if (XVA_ISSET_REQ(&tmpxvattr, XAT_AV_QUARANTINED)) {
XVA_SET_REQ(xvap, XAT_AV_QUARANTINED);
}
if (XVA_ISSET_REQ(&tmpxvattr, XAT_PROJINHERIT)) {
XVA_SET_REQ(xvap, XAT_PROJINHERIT);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP))
ASSERT3S(vp->v_type, ==, VREG);
zfs_xvattr_set(zp, xvap, tx);
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
if (mask != 0)
zfs_log_setattr(zilog, tx, TX_SETATTR, zp, vap, mask, fuidp);
if (mask & (AT_UID|AT_GID|AT_MODE))
mutex_exit(&zp->z_acl_lock);
if (attrzp) {
if (mask & (AT_UID|AT_GID|AT_MODE))
mutex_exit(&attrzp->z_acl_lock);
}
out:
if (err == 0 && attrzp) {
err2 = sa_bulk_update(attrzp->z_sa_hdl, xattr_bulk,
xattr_count, tx);
ASSERT0(err2);
}
if (attrzp)
vput(ZTOV(attrzp));
if (aclp)
zfs_acl_free(aclp);
if (fuidp) {
zfs_fuid_info_free(fuidp);
fuidp = NULL;
}
if (err) {
dmu_tx_abort(tx);
} else {
err2 = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
dmu_tx_commit(tx);
}
out2:
if (os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (err);
}
/*
* Look up the directory entries corresponding to the source and target
* directory/name pairs.
*/
static int
zfs_rename_relock_lookup(znode_t *sdzp, const struct componentname *scnp,
znode_t **szpp, znode_t *tdzp, const struct componentname *tcnp,
znode_t **tzpp)
{
zfsvfs_t *zfsvfs;
znode_t *szp, *tzp;
int error;
/*
* Before using sdzp and tdzp we must ensure that they are live.
* As a porting legacy from illumos we have two things to worry
* about. One is typical for FreeBSD and it is that the vnode is
* not reclaimed (doomed). The other is that the znode is live.
* The current code can invalidate the znode without acquiring the
* corresponding vnode lock if the object represented by the znode
* and vnode is no longer valid after a rollback or receive operation.
* z_teardown_lock hidden behind zfs_enter and zfs_exit is the lock
* that protects the znodes from the invalidation.
*/
zfsvfs = sdzp->z_zfsvfs;
ASSERT3P(zfsvfs, ==, tdzp->z_zfsvfs);
if ((error = zfs_enter_verify_zp(zfsvfs, sdzp, FTAG)) != 0)
return (error);
if ((error = zfs_verify_zp(tdzp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Re-resolve svp to be certain it still exists and fetch the
* correct vnode.
*/
error = zfs_dirent_lookup(sdzp, scnp->cn_nameptr, &szp, ZEXISTS);
if (error != 0) {
/* Source entry invalid or not there. */
if ((scnp->cn_flags & ISDOTDOT) != 0 ||
(scnp->cn_namelen == 1 && scnp->cn_nameptr[0] == '.'))
error = SET_ERROR(EINVAL);
goto out;
}
*szpp = szp;
/*
* Re-resolve tvp, if it disappeared we just carry on.
*/
error = zfs_dirent_lookup(tdzp, tcnp->cn_nameptr, &tzp, 0);
if (error != 0) {
vrele(ZTOV(szp));
if ((tcnp->cn_flags & ISDOTDOT) != 0)
error = SET_ERROR(EINVAL);
goto out;
}
*tzpp = tzp;
out:
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* We acquire all but fdvp locks using non-blocking acquisitions. If we
* fail to acquire any lock in the path we will drop all held locks,
* acquire the new lock in a blocking fashion, and then release it and
* restart the rename. This acquire/release step ensures that we do not
* spin on a lock waiting for release. On error release all vnode locks
* and decrement references the way tmpfs_rename() would do.
*/
static int
zfs_rename_relock(struct vnode *sdvp, struct vnode **svpp,
struct vnode *tdvp, struct vnode **tvpp,
const struct componentname *scnp, const struct componentname *tcnp)
{
struct vnode *nvp, *svp, *tvp;
znode_t *sdzp, *tdzp, *szp, *tzp;
int error;
VOP_UNLOCK1(tdvp);
if (*tvpp != NULL && *tvpp != tdvp)
VOP_UNLOCK1(*tvpp);
relock:
error = vn_lock(sdvp, LK_EXCLUSIVE);
if (error)
goto out;
error = vn_lock(tdvp, LK_EXCLUSIVE | LK_NOWAIT);
if (error != 0) {
VOP_UNLOCK1(sdvp);
if (error != EBUSY)
goto out;
error = vn_lock(tdvp, LK_EXCLUSIVE);
if (error)
goto out;
VOP_UNLOCK1(tdvp);
goto relock;
}
tdzp = VTOZ(tdvp);
sdzp = VTOZ(sdvp);
error = zfs_rename_relock_lookup(sdzp, scnp, &szp, tdzp, tcnp, &tzp);
if (error != 0) {
VOP_UNLOCK1(sdvp);
VOP_UNLOCK1(tdvp);
goto out;
}
svp = ZTOV(szp);
tvp = tzp != NULL ? ZTOV(tzp) : NULL;
/*
* Now try acquire locks on svp and tvp.
*/
nvp = svp;
error = vn_lock(nvp, LK_EXCLUSIVE | LK_NOWAIT);
if (error != 0) {
VOP_UNLOCK1(sdvp);
VOP_UNLOCK1(tdvp);
if (tvp != NULL)
vrele(tvp);
if (error != EBUSY) {
vrele(nvp);
goto out;
}
error = vn_lock(nvp, LK_EXCLUSIVE);
if (error != 0) {
vrele(nvp);
goto out;
}
VOP_UNLOCK1(nvp);
/*
* Concurrent rename race.
* XXX ?
*/
if (nvp == tdvp) {
vrele(nvp);
error = SET_ERROR(EINVAL);
goto out;
}
vrele(*svpp);
*svpp = nvp;
goto relock;
}
vrele(*svpp);
*svpp = nvp;
if (*tvpp != NULL)
vrele(*tvpp);
*tvpp = NULL;
if (tvp != NULL) {
nvp = tvp;
error = vn_lock(nvp, LK_EXCLUSIVE | LK_NOWAIT);
if (error != 0) {
VOP_UNLOCK1(sdvp);
VOP_UNLOCK1(tdvp);
VOP_UNLOCK1(*svpp);
if (error != EBUSY) {
vrele(nvp);
goto out;
}
error = vn_lock(nvp, LK_EXCLUSIVE);
if (error != 0) {
vrele(nvp);
goto out;
}
vput(nvp);
goto relock;
}
*tvpp = nvp;
}
return (0);
out:
return (error);
}
/*
* Note that we must use VRELE_ASYNC in this function as it walks
* up the directory tree and vrele may need to acquire an exclusive
* lock if a last reference to a vnode is dropped.
*/
static int
zfs_rename_check(znode_t *szp, znode_t *sdzp, znode_t *tdzp)
{
zfsvfs_t *zfsvfs;
znode_t *zp, *zp1;
uint64_t parent;
int error;
zfsvfs = tdzp->z_zfsvfs;
if (tdzp == szp)
return (SET_ERROR(EINVAL));
if (tdzp == sdzp)
return (0);
if (tdzp->z_id == zfsvfs->z_root)
return (0);
zp = tdzp;
for (;;) {
ASSERT(!zp->z_unlinked);
if ((error = sa_lookup(zp->z_sa_hdl,
SA_ZPL_PARENT(zfsvfs), &parent, sizeof (parent))) != 0)
break;
if (parent == szp->z_id) {
error = SET_ERROR(EINVAL);
break;
}
if (parent == zfsvfs->z_root)
break;
if (parent == sdzp->z_id)
break;
error = zfs_zget(zfsvfs, parent, &zp1);
if (error != 0)
break;
if (zp != tdzp)
VN_RELE_ASYNC(ZTOV(zp),
dsl_pool_zrele_taskq(
dmu_objset_pool(zfsvfs->z_os)));
zp = zp1;
}
if (error == ENOTDIR)
panic("checkpath: .. not a directory\n");
if (zp != tdzp)
VN_RELE_ASYNC(ZTOV(zp),
dsl_pool_zrele_taskq(dmu_objset_pool(zfsvfs->z_os)));
return (error);
}
#if __FreeBSD_version < 1300124
static void
cache_vop_rename(struct vnode *fdvp, struct vnode *fvp, struct vnode *tdvp,
struct vnode *tvp, struct componentname *fcnp, struct componentname *tcnp)
{
cache_purge(fvp);
if (tvp != NULL)
cache_purge(tvp);
cache_purge_negative(tdvp);
}
#endif
static int
zfs_do_rename_impl(vnode_t *sdvp, vnode_t **svpp, struct componentname *scnp,
vnode_t *tdvp, vnode_t **tvpp, struct componentname *tcnp,
cred_t *cr);
/*
* Move an entry from the provided source directory to the target
* directory. Change the entry name as indicated.
*
* IN: sdvp - Source directory containing the "old entry".
* scnp - Old entry name.
* tdvp - Target directory to contain the "new entry".
* tcnp - New entry name.
* cr - credentials of caller.
* INOUT: svpp - Source file
* tvpp - Target file, may point to NULL initially
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* sdvp,tdvp - ctime|mtime updated
*/
static int
zfs_do_rename(vnode_t *sdvp, vnode_t **svpp, struct componentname *scnp,
vnode_t *tdvp, vnode_t **tvpp, struct componentname *tcnp,
cred_t *cr)
{
int error;
ASSERT_VOP_ELOCKED(tdvp, __func__);
if (*tvpp != NULL)
ASSERT_VOP_ELOCKED(*tvpp, __func__);
/* Reject renames across filesystems. */
if ((*svpp)->v_mount != tdvp->v_mount ||
((*tvpp) != NULL && (*svpp)->v_mount != (*tvpp)->v_mount)) {
error = SET_ERROR(EXDEV);
goto out;
}
if (zfsctl_is_node(tdvp)) {
error = SET_ERROR(EXDEV);
goto out;
}
/*
* Lock all four vnodes to ensure safety and semantics of renaming.
*/
error = zfs_rename_relock(sdvp, svpp, tdvp, tvpp, scnp, tcnp);
if (error != 0) {
/* no vnodes are locked in the case of error here */
return (error);
}
error = zfs_do_rename_impl(sdvp, svpp, scnp, tdvp, tvpp, tcnp, cr);
VOP_UNLOCK1(sdvp);
VOP_UNLOCK1(*svpp);
out:
if (*tvpp != NULL)
VOP_UNLOCK1(*tvpp);
if (tdvp != *tvpp)
VOP_UNLOCK1(tdvp);
return (error);
}
static int
zfs_do_rename_impl(vnode_t *sdvp, vnode_t **svpp, struct componentname *scnp,
vnode_t *tdvp, vnode_t **tvpp, struct componentname *tcnp,
cred_t *cr)
{
dmu_tx_t *tx;
zfsvfs_t *zfsvfs;
zilog_t *zilog;
znode_t *tdzp, *sdzp, *tzp, *szp;
const char *snm = scnp->cn_nameptr;
const char *tnm = tcnp->cn_nameptr;
int error;
tdzp = VTOZ(tdvp);
sdzp = VTOZ(sdvp);
zfsvfs = tdzp->z_zfsvfs;
if ((error = zfs_enter_verify_zp(zfsvfs, tdzp, FTAG)) != 0)
return (error);
if ((error = zfs_verify_zp(sdzp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
zilog = zfsvfs->z_log;
if (zfsvfs->z_utf8 && u8_validate(tnm,
strlen(tnm), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
error = SET_ERROR(EILSEQ);
goto out;
}
/* If source and target are the same file, there is nothing to do. */
if ((*svpp) == (*tvpp)) {
error = 0;
goto out;
}
if (((*svpp)->v_type == VDIR && (*svpp)->v_mountedhere != NULL) ||
((*tvpp) != NULL && (*tvpp)->v_type == VDIR &&
(*tvpp)->v_mountedhere != NULL)) {
error = SET_ERROR(EXDEV);
goto out;
}
szp = VTOZ(*svpp);
if ((error = zfs_verify_zp(szp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
tzp = *tvpp == NULL ? NULL : VTOZ(*tvpp);
if (tzp != NULL) {
if ((error = zfs_verify_zp(tzp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
/*
* This is to prevent the creation of links into attribute space
* by renaming a linked file into/outof an attribute directory.
* See the comment in zfs_link() for why this is considered bad.
*/
if ((tdzp->z_pflags & ZFS_XATTR) != (sdzp->z_pflags & ZFS_XATTR)) {
error = SET_ERROR(EINVAL);
goto out;
}
/*
* If we are using project inheritance, means if the directory has
* ZFS_PROJINHERIT set, then its descendant directories will inherit
* not only the project ID, but also the ZFS_PROJINHERIT flag. Under
* such case, we only allow renames into our tree when the project
* IDs are the same.
*/
if (tdzp->z_pflags & ZFS_PROJINHERIT &&
tdzp->z_projid != szp->z_projid) {
error = SET_ERROR(EXDEV);
goto out;
}
/*
* Must have write access at the source to remove the old entry
* and write access at the target to create the new entry.
* Note that if target and source are the same, this can be
* done in a single check.
*/
if ((error = zfs_zaccess_rename(sdzp, szp, tdzp, tzp, cr, NULL)))
goto out;
if ((*svpp)->v_type == VDIR) {
/*
* Avoid ".", "..", and aliases of "." for obvious reasons.
*/
if ((scnp->cn_namelen == 1 && scnp->cn_nameptr[0] == '.') ||
sdzp == szp ||
(scnp->cn_flags | tcnp->cn_flags) & ISDOTDOT) {
error = EINVAL;
goto out;
}
/*
* Check to make sure rename is valid.
* Can't do a move like this: /usr/a/b to /usr/a/b/c/d
*/
if ((error = zfs_rename_check(szp, sdzp, tdzp)))
goto out;
}
/*
* Does target exist?
*/
if (tzp) {
/*
* Source and target must be the same type.
*/
if ((*svpp)->v_type == VDIR) {
if ((*tvpp)->v_type != VDIR) {
error = SET_ERROR(ENOTDIR);
goto out;
} else {
cache_purge(tdvp);
if (sdvp != tdvp)
cache_purge(sdvp);
}
} else {
if ((*tvpp)->v_type == VDIR) {
error = SET_ERROR(EISDIR);
goto out;
}
}
}
vn_seqc_write_begin(*svpp);
vn_seqc_write_begin(sdvp);
if (*tvpp != NULL)
vn_seqc_write_begin(*tvpp);
if (tdvp != *tvpp)
vn_seqc_write_begin(tdvp);
vnevent_rename_src(*svpp, sdvp, scnp->cn_nameptr, ct);
if (tzp)
vnevent_rename_dest(*tvpp, tdvp, tnm, ct);
/*
* notify the target directory if it is not the same
* as source directory.
*/
if (tdvp != sdvp) {
vnevent_rename_dest_dir(tdvp, ct);
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE);
dmu_tx_hold_sa(tx, sdzp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, sdzp->z_id, FALSE, snm);
dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, tnm);
if (sdzp != tdzp) {
dmu_tx_hold_sa(tx, tdzp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, tdzp);
}
if (tzp) {
dmu_tx_hold_sa(tx, tzp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, tzp);
}
zfs_sa_upgrade_txholds(tx, szp);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
goto out_seq;
}
if (tzp) /* Attempt to remove the existing target */
error = zfs_link_destroy(tdzp, tnm, tzp, tx, 0, NULL);
if (error == 0) {
error = zfs_link_create(tdzp, tnm, szp, tx, ZRENAMING);
if (error == 0) {
szp->z_pflags |= ZFS_AV_MODIFIED;
error = sa_update(szp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs),
(void *)&szp->z_pflags, sizeof (uint64_t), tx);
ASSERT0(error);
error = zfs_link_destroy(sdzp, snm, szp, tx, ZRENAMING,
NULL);
if (error == 0) {
zfs_log_rename(zilog, tx, TX_RENAME, sdzp,
snm, tdzp, tnm, szp);
} else {
/*
* At this point, we have successfully created
* the target name, but have failed to remove
* the source name. Since the create was done
* with the ZRENAMING flag, there are
* complications; for one, the link count is
* wrong. The easiest way to deal with this
* is to remove the newly created target, and
* return the original error. This must
* succeed; fortunately, it is very unlikely to
* fail, since we just created it.
*/
VERIFY0(zfs_link_destroy(tdzp, tnm, szp, tx,
ZRENAMING, NULL));
}
}
if (error == 0) {
cache_vop_rename(sdvp, *svpp, tdvp, *tvpp, scnp, tcnp);
}
}
dmu_tx_commit(tx);
out_seq:
vn_seqc_write_end(*svpp);
vn_seqc_write_end(sdvp);
if (*tvpp != NULL)
vn_seqc_write_end(*tvpp);
if (tdvp != *tvpp)
vn_seqc_write_end(tdvp);
out:
if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
int
zfs_rename(znode_t *sdzp, const char *sname, znode_t *tdzp, const char *tname,
- cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap, zuserns_t *mnt_ns)
+ cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap, zidmap_t *mnt_ns)
{
struct componentname scn, tcn;
vnode_t *sdvp, *tdvp;
vnode_t *svp, *tvp;
int error;
svp = tvp = NULL;
if (rflags != 0 || wo_vap != NULL)
return (SET_ERROR(EINVAL));
sdvp = ZTOV(sdzp);
tdvp = ZTOV(tdzp);
error = zfs_lookup_internal(sdzp, sname, &svp, &scn, DELETE);
if (sdzp->z_zfsvfs->z_replay == B_FALSE)
VOP_UNLOCK1(sdvp);
if (error != 0)
goto fail;
VOP_UNLOCK1(svp);
vn_lock(tdvp, LK_EXCLUSIVE | LK_RETRY);
error = zfs_lookup_internal(tdzp, tname, &tvp, &tcn, RENAME);
if (error == EJUSTRETURN)
tvp = NULL;
else if (error != 0) {
VOP_UNLOCK1(tdvp);
goto fail;
}
error = zfs_do_rename(sdvp, &svp, &scn, tdvp, &tvp, &tcn, cr);
fail:
if (svp != NULL)
vrele(svp);
if (tvp != NULL)
vrele(tvp);
return (error);
}
/*
* Insert the indicated symbolic reference entry into the directory.
*
* IN: dvp - Directory to contain new symbolic link.
* link - Name for new symlink entry.
* vap - Attributes of new entry.
* cr - credentials of caller.
* ct - caller context
* flags - case flags
* mnt_ns - Unused on FreeBSD
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dvp - ctime|mtime updated
*/
int
zfs_symlink(znode_t *dzp, const char *name, vattr_t *vap,
- const char *link, znode_t **zpp, cred_t *cr, int flags, zuserns_t *mnt_ns)
+ const char *link, znode_t **zpp, cred_t *cr, int flags, zidmap_t *mnt_ns)
{
(void) flags;
znode_t *zp;
dmu_tx_t *tx;
zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
zilog_t *zilog;
uint64_t len = strlen(link);
int error;
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
uint64_t txtype = TX_SYMLINK;
ASSERT3S(vap->va_type, ==, VLNK);
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (zfsvfs->z_utf8 && u8_validate(name, strlen(name),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (len > MAXPATHLEN) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENAMETOOLONG));
}
if ((error = zfs_acl_ids_create(dzp, 0,
vap, cr, NULL, &acl_ids, NULL)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Attempt to lock directory; fail if entry already exists.
*/
error = zfs_dirent_lookup(dzp, name, &zp, ZNEW);
if (error) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids,
0 /* projid */)) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EDQUOT));
}
getnewvnode_reserve_();
tx = dmu_tx_create(zfsvfs->z_os);
fuid_dirtied = zfsvfs->z_fuid_dirty;
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, MAX(1, len));
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE + len);
dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE);
if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0,
acl_ids.z_aclp->z_acl_bytes);
}
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
getnewvnode_drop_reserve();
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Create a new object for the symlink.
* for version 4 ZPL datasets the symlink will be an SA attribute
*/
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
if (zp->z_is_sa)
error = sa_update(zp->z_sa_hdl, SA_ZPL_SYMLINK(zfsvfs),
__DECONST(void *, link), len, tx);
else
zfs_sa_symlink(zp, __DECONST(char *, link), len, tx);
zp->z_size = len;
(void) sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zfsvfs),
&zp->z_size, sizeof (zp->z_size), tx);
/*
* Insert the new object into the directory.
*/
(void) zfs_link_create(dzp, name, zp, tx, ZNEW);
zfs_log_symlink(zilog, tx, txtype, dzp, zp, name, link);
*zpp = zp;
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
getnewvnode_drop_reserve();
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Return, in the buffer contained in the provided uio structure,
* the symbolic path referred to by vp.
*
* IN: vp - vnode of symbolic link.
* uio - structure to contain the link path.
* cr - credentials of caller.
* ct - caller context
*
* OUT: uio - structure containing the link path.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* vp - atime updated
*/
static int
zfs_readlink(vnode_t *vp, zfs_uio_t *uio, cred_t *cr, caller_context_t *ct)
{
(void) cr, (void) ct;
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if (zp->z_is_sa)
error = sa_lookup_uio(zp->z_sa_hdl,
SA_ZPL_SYMLINK(zfsvfs), uio);
else
error = zfs_sa_readlink(zp, uio);
ZFS_ACCESSTIME_STAMP(zfsvfs, zp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Insert a new entry into directory tdvp referencing svp.
*
* IN: tdvp - Directory to contain new entry.
* svp - vnode of new entry.
* name - name of new entry.
* cr - credentials of caller.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* tdvp - ctime|mtime updated
* svp - ctime updated
*/
int
zfs_link(znode_t *tdzp, znode_t *szp, const char *name, cred_t *cr,
int flags)
{
(void) flags;
znode_t *tzp;
zfsvfs_t *zfsvfs = tdzp->z_zfsvfs;
zilog_t *zilog;
dmu_tx_t *tx;
int error;
uint64_t parent;
uid_t owner;
ASSERT3S(ZTOV(tdzp)->v_type, ==, VDIR);
if ((error = zfs_enter_verify_zp(zfsvfs, tdzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
/*
* POSIX dictates that we return EPERM here.
* Better choices include ENOTSUP or EISDIR.
*/
if (ZTOV(szp)->v_type == VDIR) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if ((error = zfs_verify_zp(szp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* If we are using project inheritance, means if the directory has
* ZFS_PROJINHERIT set, then its descendant directories will inherit
* not only the project ID, but also the ZFS_PROJINHERIT flag. Under
* such case, we only allow hard link creation in our tree when the
* project IDs are the same.
*/
if (tdzp->z_pflags & ZFS_PROJINHERIT &&
tdzp->z_projid != szp->z_projid) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
if (szp->z_pflags & (ZFS_APPENDONLY |
ZFS_IMMUTABLE | ZFS_READONLY)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
/* Prevent links to .zfs/shares files */
if ((error = sa_lookup(szp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs),
&parent, sizeof (uint64_t))) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (parent == zfsvfs->z_shares_dir) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if (zfsvfs->z_utf8 && u8_validate(name,
strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
/*
* We do not support links between attributes and non-attributes
* because of the potential security risk of creating links
* into "normal" file space in order to circumvent restrictions
* imposed in attribute space.
*/
if ((szp->z_pflags & ZFS_XATTR) != (tdzp->z_pflags & ZFS_XATTR)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
owner = zfs_fuid_map_id(zfsvfs, szp->z_uid, cr, ZFS_OWNER);
if (owner != crgetuid(cr) && secpolicy_basic_link(ZTOV(szp), cr) != 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if ((error = zfs_zaccess(tdzp, ACE_ADD_FILE, 0, B_FALSE, cr, NULL))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Attempt to lock directory; fail if entry already exists.
*/
error = zfs_dirent_lookup(tdzp, name, &tzp, ZNEW);
if (error) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, name);
zfs_sa_upgrade_txholds(tx, szp);
zfs_sa_upgrade_txholds(tx, tdzp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
error = zfs_link_create(tdzp, name, szp, tx, 0);
if (error == 0) {
uint64_t txtype = TX_LINK;
zfs_log_link(zilog, tx, txtype, tdzp, szp, name);
}
dmu_tx_commit(tx);
if (error == 0) {
vnevent_link(ZTOV(szp), ct);
}
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Free or allocate space in a file. Currently, this function only
* supports the `F_FREESP' command. However, this command is somewhat
* misnamed, as its functionality includes the ability to allocate as
* well as free space.
*
* IN: ip - inode of file to free data in.
* cmd - action to take (only F_FREESP supported).
* bfp - section of file to free/alloc.
* flag - current file open mode flags.
* offset - current file offset.
* cr - credentials of caller.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* ip - ctime|mtime updated
*/
int
zfs_space(znode_t *zp, int cmd, flock64_t *bfp, int flag,
offset_t offset, cred_t *cr)
{
(void) offset;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
uint64_t off, len;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if (cmd != F_FREESP) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Callers might not be able to detect properly that we are read-only,
* so check it explicitly here.
*/
if (zfs_is_readonly(zfsvfs)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EROFS));
}
if (bfp->l_len < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Permissions aren't checked on Solaris because on this OS
* zfs_space() can only be called with an opened file handle.
* On Linux we can get here through truncate_range() which
* operates directly on inodes, so we need to check access rights.
*/
if ((error = zfs_zaccess(zp, ACE_WRITE_DATA, 0, B_FALSE, cr, NULL))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
off = bfp->l_start;
len = bfp->l_len; /* 0 means from off to end of file */
error = zfs_freesp(zp, off, len, flag, TRUE);
zfs_exit(zfsvfs, FTAG);
return (error);
}
static void
zfs_inactive(vnode_t *vp, cred_t *cr, caller_context_t *ct)
{
(void) cr, (void) ct;
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
int error;
ZFS_TEARDOWN_INACTIVE_ENTER_READ(zfsvfs);
if (zp->z_sa_hdl == NULL) {
/*
* The fs has been unmounted, or we did a
* suspend/resume and this file no longer exists.
*/
ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs);
vrecycle(vp);
return;
}
if (zp->z_unlinked) {
/*
* Fast path to recycle a vnode of a removed file.
*/
ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs);
vrecycle(vp);
return;
}
if (zp->z_atime_dirty && zp->z_unlinked == 0) {
dmu_tx_t *tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
} else {
(void) sa_update(zp->z_sa_hdl, SA_ZPL_ATIME(zfsvfs),
(void *)&zp->z_atime, sizeof (zp->z_atime), tx);
zp->z_atime_dirty = 0;
dmu_tx_commit(tx);
}
}
ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs);
}
_Static_assert(sizeof (struct zfid_short) <= sizeof (struct fid),
"struct zfid_short bigger than struct fid");
_Static_assert(sizeof (struct zfid_long) <= sizeof (struct fid),
"struct zfid_long bigger than struct fid");
static int
zfs_fid(vnode_t *vp, fid_t *fidp, caller_context_t *ct)
{
(void) ct;
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
uint32_t gen;
uint64_t gen64;
uint64_t object = zp->z_id;
zfid_short_t *zfid;
int size, i, error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(zfsvfs),
&gen64, sizeof (uint64_t))) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
gen = (uint32_t)gen64;
size = (zfsvfs->z_parent != zfsvfs) ? LONG_FID_LEN : SHORT_FID_LEN;
fidp->fid_len = size;
zfid = (zfid_short_t *)fidp;
zfid->zf_len = size;
for (i = 0; i < sizeof (zfid->zf_object); i++)
zfid->zf_object[i] = (uint8_t)(object >> (8 * i));
/* Must have a non-zero generation number to distinguish from .zfs */
if (gen == 0)
gen = 1;
for (i = 0; i < sizeof (zfid->zf_gen); i++)
zfid->zf_gen[i] = (uint8_t)(gen >> (8 * i));
if (size == LONG_FID_LEN) {
uint64_t objsetid = dmu_objset_id(zfsvfs->z_os);
zfid_long_t *zlfid;
zlfid = (zfid_long_t *)fidp;
for (i = 0; i < sizeof (zlfid->zf_setid); i++)
zlfid->zf_setid[i] = (uint8_t)(objsetid >> (8 * i));
/* XXX - this should be the generation number for the objset */
for (i = 0; i < sizeof (zlfid->zf_setgen); i++)
zlfid->zf_setgen[i] = 0;
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
static int
zfs_pathconf(vnode_t *vp, int cmd, ulong_t *valp, cred_t *cr,
caller_context_t *ct)
{
znode_t *zp;
zfsvfs_t *zfsvfs;
int error;
switch (cmd) {
case _PC_LINK_MAX:
*valp = MIN(LONG_MAX, ZFS_LINK_MAX);
return (0);
case _PC_FILESIZEBITS:
*valp = 64;
return (0);
case _PC_MIN_HOLE_SIZE:
*valp = (int)SPA_MINBLOCKSIZE;
return (0);
case _PC_ACL_EXTENDED:
#if 0 /* POSIX ACLs are not implemented for ZFS on FreeBSD yet. */
zp = VTOZ(vp);
zfsvfs = zp->z_zfsvfs;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
*valp = zfsvfs->z_acl_type == ZFSACLTYPE_POSIX ? 1 : 0;
zfs_exit(zfsvfs, FTAG);
#else
*valp = 0;
#endif
return (0);
case _PC_ACL_NFS4:
zp = VTOZ(vp);
zfsvfs = zp->z_zfsvfs;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
*valp = zfsvfs->z_acl_type == ZFS_ACLTYPE_NFSV4 ? 1 : 0;
zfs_exit(zfsvfs, FTAG);
return (0);
case _PC_ACL_PATH_MAX:
*valp = ACL_MAX_ENTRIES;
return (0);
default:
return (EOPNOTSUPP);
}
}
static int
zfs_getpages(struct vnode *vp, vm_page_t *ma, int count, int *rbehind,
int *rahead)
{
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
zfs_locked_range_t *lr;
vm_object_t object;
off_t start, end, obj_size;
uint_t blksz;
int pgsin_b, pgsin_a;
int error;
if (zfs_enter_verify_zp(zfsvfs, zp, FTAG) != 0)
return (zfs_vm_pagerret_error);
start = IDX_TO_OFF(ma[0]->pindex);
end = IDX_TO_OFF(ma[count - 1]->pindex + 1);
/*
* Lock a range covering all required and optional pages.
* Note that we need to handle the case of the block size growing.
*/
for (;;) {
blksz = zp->z_blksz;
lr = zfs_rangelock_tryenter(&zp->z_rangelock,
rounddown(start, blksz),
roundup(end, blksz) - rounddown(start, blksz), RL_READER);
if (lr == NULL) {
if (rahead != NULL) {
*rahead = 0;
rahead = NULL;
}
if (rbehind != NULL) {
*rbehind = 0;
rbehind = NULL;
}
break;
}
if (blksz == zp->z_blksz)
break;
zfs_rangelock_exit(lr);
}
object = ma[0]->object;
zfs_vmobject_wlock(object);
obj_size = object->un_pager.vnp.vnp_size;
zfs_vmobject_wunlock(object);
if (IDX_TO_OFF(ma[count - 1]->pindex) >= obj_size) {
if (lr != NULL)
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (zfs_vm_pagerret_bad);
}
pgsin_b = 0;
if (rbehind != NULL) {
pgsin_b = OFF_TO_IDX(start - rounddown(start, blksz));
pgsin_b = MIN(*rbehind, pgsin_b);
}
pgsin_a = 0;
if (rahead != NULL) {
pgsin_a = OFF_TO_IDX(roundup(end, blksz) - end);
if (end + IDX_TO_OFF(pgsin_a) >= obj_size)
pgsin_a = OFF_TO_IDX(round_page(obj_size) - end);
pgsin_a = MIN(*rahead, pgsin_a);
}
/*
* NB: we need to pass the exact byte size of the data that we expect
* to read after accounting for the file size. This is required because
* ZFS will panic if we request DMU to read beyond the end of the last
* allocated block.
*/
error = dmu_read_pages(zfsvfs->z_os, zp->z_id, ma, count, &pgsin_b,
&pgsin_a, MIN(end, obj_size) - (end - PAGE_SIZE));
if (lr != NULL)
zfs_rangelock_exit(lr);
ZFS_ACCESSTIME_STAMP(zfsvfs, zp);
dataset_kstats_update_read_kstats(&zfsvfs->z_kstat, count*PAGE_SIZE);
zfs_exit(zfsvfs, FTAG);
if (error != 0)
return (zfs_vm_pagerret_error);
VM_CNT_INC(v_vnodein);
VM_CNT_ADD(v_vnodepgsin, count + pgsin_b + pgsin_a);
if (rbehind != NULL)
*rbehind = pgsin_b;
if (rahead != NULL)
*rahead = pgsin_a;
return (zfs_vm_pagerret_ok);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_getpages_args {
struct vnode *a_vp;
vm_page_t *a_m;
int a_count;
int *a_rbehind;
int *a_rahead;
};
#endif
static int
zfs_freebsd_getpages(struct vop_getpages_args *ap)
{
return (zfs_getpages(ap->a_vp, ap->a_m, ap->a_count, ap->a_rbehind,
ap->a_rahead));
}
static int
zfs_putpages(struct vnode *vp, vm_page_t *ma, size_t len, int flags,
int *rtvals)
{
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
zfs_locked_range_t *lr;
dmu_tx_t *tx;
struct sf_buf *sf;
vm_object_t object;
vm_page_t m;
caddr_t va;
size_t tocopy;
size_t lo_len;
vm_ooffset_t lo_off;
vm_ooffset_t off;
uint_t blksz;
int ncount;
int pcount;
int err;
int i;
object = vp->v_object;
KASSERT(ma[0]->object == object, ("mismatching object"));
KASSERT(len > 0 && (len & PAGE_MASK) == 0, ("unexpected length"));
pcount = btoc(len);
ncount = pcount;
for (i = 0; i < pcount; i++)
rtvals[i] = zfs_vm_pagerret_error;
if (zfs_enter_verify_zp(zfsvfs, zp, FTAG) != 0)
return (zfs_vm_pagerret_error);
off = IDX_TO_OFF(ma[0]->pindex);
blksz = zp->z_blksz;
lo_off = rounddown(off, blksz);
lo_len = roundup(len + (off - lo_off), blksz);
lr = zfs_rangelock_enter(&zp->z_rangelock, lo_off, lo_len, RL_WRITER);
zfs_vmobject_wlock(object);
if (len + off > object->un_pager.vnp.vnp_size) {
if (object->un_pager.vnp.vnp_size > off) {
int pgoff;
len = object->un_pager.vnp.vnp_size - off;
ncount = btoc(len);
if ((pgoff = (int)len & PAGE_MASK) != 0) {
/*
* If the object is locked and the following
* conditions hold, then the page's dirty
* field cannot be concurrently changed by a
* pmap operation.
*/
m = ma[ncount - 1];
vm_page_assert_sbusied(m);
KASSERT(!pmap_page_is_write_mapped(m),
("zfs_putpages: page %p is not read-only",
m));
vm_page_clear_dirty(m, pgoff, PAGE_SIZE -
pgoff);
}
} else {
len = 0;
ncount = 0;
}
if (ncount < pcount) {
for (i = ncount; i < pcount; i++) {
rtvals[i] = zfs_vm_pagerret_bad;
}
}
}
zfs_vmobject_wunlock(object);
if (ncount == 0)
goto out;
if (zfs_id_overblockquota(zfsvfs, DMU_USERUSED_OBJECT, zp->z_uid) ||
zfs_id_overblockquota(zfsvfs, DMU_GROUPUSED_OBJECT, zp->z_gid) ||
(zp->z_projid != ZFS_DEFAULT_PROJID &&
zfs_id_overblockquota(zfsvfs, DMU_PROJECTUSED_OBJECT,
zp->z_projid))) {
goto out;
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_write(tx, zp->z_id, off, len);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
err = dmu_tx_assign(tx, TXG_WAIT);
if (err != 0) {
dmu_tx_abort(tx);
goto out;
}
if (zp->z_blksz < PAGE_SIZE) {
for (i = 0; len > 0; off += tocopy, len -= tocopy, i++) {
tocopy = len > PAGE_SIZE ? PAGE_SIZE : len;
va = zfs_map_page(ma[i], &sf);
dmu_write(zfsvfs->z_os, zp->z_id, off, tocopy, va, tx);
zfs_unmap_page(sf);
}
} else {
err = dmu_write_pages(zfsvfs->z_os, zp->z_id, off, len, ma, tx);
}
if (err == 0) {
uint64_t mtime[2], ctime[2];
sa_bulk_attr_t bulk[3];
int count = 0;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL,
&mtime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
&ctime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, 8);
zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime);
err = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
ASSERT0(err);
/*
* XXX we should be passing a callback to undirty
* but that would make the locking messier
*/
zfs_log_write(zfsvfs->z_log, tx, TX_WRITE, zp, off,
len, 0, NULL, NULL);
zfs_vmobject_wlock(object);
for (i = 0; i < ncount; i++) {
rtvals[i] = zfs_vm_pagerret_ok;
vm_page_undirty(ma[i]);
}
zfs_vmobject_wunlock(object);
VM_CNT_INC(v_vnodeout);
VM_CNT_ADD(v_vnodepgsout, ncount);
}
dmu_tx_commit(tx);
out:
zfs_rangelock_exit(lr);
if ((flags & (zfs_vm_pagerput_sync | zfs_vm_pagerput_inval)) != 0 ||
zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zfsvfs->z_log, zp->z_id);
dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, len);
zfs_exit(zfsvfs, FTAG);
return (rtvals[0]);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_putpages_args {
struct vnode *a_vp;
vm_page_t *a_m;
int a_count;
int a_sync;
int *a_rtvals;
};
#endif
static int
zfs_freebsd_putpages(struct vop_putpages_args *ap)
{
return (zfs_putpages(ap->a_vp, ap->a_m, ap->a_count, ap->a_sync,
ap->a_rtvals));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_bmap_args {
struct vnode *a_vp;
daddr_t a_bn;
struct bufobj **a_bop;
daddr_t *a_bnp;
int *a_runp;
int *a_runb;
};
#endif
static int
zfs_freebsd_bmap(struct vop_bmap_args *ap)
{
if (ap->a_bop != NULL)
*ap->a_bop = &ap->a_vp->v_bufobj;
if (ap->a_bnp != NULL)
*ap->a_bnp = ap->a_bn;
if (ap->a_runp != NULL)
*ap->a_runp = 0;
if (ap->a_runb != NULL)
*ap->a_runb = 0;
return (0);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_open_args {
struct vnode *a_vp;
int a_mode;
struct ucred *a_cred;
struct thread *a_td;
};
#endif
static int
zfs_freebsd_open(struct vop_open_args *ap)
{
vnode_t *vp = ap->a_vp;
znode_t *zp = VTOZ(vp);
int error;
error = zfs_open(&vp, ap->a_mode, ap->a_cred);
if (error == 0)
vnode_create_vobject(vp, zp->z_size, ap->a_td);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_close_args {
struct vnode *a_vp;
int a_fflag;
struct ucred *a_cred;
struct thread *a_td;
};
#endif
static int
zfs_freebsd_close(struct vop_close_args *ap)
{
return (zfs_close(ap->a_vp, ap->a_fflag, 1, 0, ap->a_cred));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_ioctl_args {
struct vnode *a_vp;
ulong_t a_command;
caddr_t a_data;
int a_fflag;
struct ucred *cred;
struct thread *td;
};
#endif
static int
zfs_freebsd_ioctl(struct vop_ioctl_args *ap)
{
return (zfs_ioctl(ap->a_vp, ap->a_command, (intptr_t)ap->a_data,
ap->a_fflag, ap->a_cred, NULL));
}
static int
ioflags(int ioflags)
{
int flags = 0;
if (ioflags & IO_APPEND)
flags |= O_APPEND;
if (ioflags & IO_NDELAY)
flags |= O_NONBLOCK;
if (ioflags & IO_SYNC)
flags |= O_SYNC;
return (flags);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_read_args {
struct vnode *a_vp;
struct uio *a_uio;
int a_ioflag;
struct ucred *a_cred;
};
#endif
static int
zfs_freebsd_read(struct vop_read_args *ap)
{
zfs_uio_t uio;
zfs_uio_init(&uio, ap->a_uio);
return (zfs_read(VTOZ(ap->a_vp), &uio, ioflags(ap->a_ioflag),
ap->a_cred));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_write_args {
struct vnode *a_vp;
struct uio *a_uio;
int a_ioflag;
struct ucred *a_cred;
};
#endif
static int
zfs_freebsd_write(struct vop_write_args *ap)
{
zfs_uio_t uio;
zfs_uio_init(&uio, ap->a_uio);
return (zfs_write(VTOZ(ap->a_vp), &uio, ioflags(ap->a_ioflag),
ap->a_cred));
}
#if __FreeBSD_version >= 1300102
/*
* VOP_FPLOOKUP_VEXEC routines are subject to special circumstances, see
* the comment above cache_fplookup for details.
*/
static int
zfs_freebsd_fplookup_vexec(struct vop_fplookup_vexec_args *v)
{
vnode_t *vp;
znode_t *zp;
uint64_t pflags;
vp = v->a_vp;
zp = VTOZ_SMR(vp);
if (__predict_false(zp == NULL))
return (EAGAIN);
pflags = atomic_load_64(&zp->z_pflags);
if (pflags & ZFS_AV_QUARANTINED)
return (EAGAIN);
if (pflags & ZFS_XATTR)
return (EAGAIN);
if ((pflags & ZFS_NO_EXECS_DENIED) == 0)
return (EAGAIN);
return (0);
}
#endif
#if __FreeBSD_version >= 1300139
static int
zfs_freebsd_fplookup_symlink(struct vop_fplookup_symlink_args *v)
{
vnode_t *vp;
znode_t *zp;
char *target;
vp = v->a_vp;
zp = VTOZ_SMR(vp);
if (__predict_false(zp == NULL)) {
return (EAGAIN);
}
target = atomic_load_consume_ptr(&zp->z_cached_symlink);
if (target == NULL) {
return (EAGAIN);
}
return (cache_symlink_resolve(v->a_fpl, target, strlen(target)));
}
#endif
#ifndef _SYS_SYSPROTO_H_
struct vop_access_args {
struct vnode *a_vp;
accmode_t a_accmode;
struct ucred *a_cred;
struct thread *a_td;
};
#endif
static int
zfs_freebsd_access(struct vop_access_args *ap)
{
vnode_t *vp = ap->a_vp;
znode_t *zp = VTOZ(vp);
accmode_t accmode;
int error = 0;
if (ap->a_accmode == VEXEC) {
if (zfs_fastaccesschk_execute(zp, ap->a_cred) == 0)
return (0);
}
/*
* ZFS itself only knowns about VREAD, VWRITE, VEXEC and VAPPEND,
*/
accmode = ap->a_accmode & (VREAD|VWRITE|VEXEC|VAPPEND);
if (accmode != 0)
error = zfs_access(zp, accmode, 0, ap->a_cred);
/*
* VADMIN has to be handled by vaccess().
*/
if (error == 0) {
accmode = ap->a_accmode & ~(VREAD|VWRITE|VEXEC|VAPPEND);
if (accmode != 0) {
#if __FreeBSD_version >= 1300105
error = vaccess(vp->v_type, zp->z_mode, zp->z_uid,
zp->z_gid, accmode, ap->a_cred);
#else
error = vaccess(vp->v_type, zp->z_mode, zp->z_uid,
zp->z_gid, accmode, ap->a_cred, NULL);
#endif
}
}
/*
* For VEXEC, ensure that at least one execute bit is set for
* non-directories.
*/
if (error == 0 && (ap->a_accmode & VEXEC) != 0 && vp->v_type != VDIR &&
(zp->z_mode & (S_IXUSR | S_IXGRP | S_IXOTH)) == 0) {
error = EACCES;
}
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_lookup_args {
struct vnode *a_dvp;
struct vnode **a_vpp;
struct componentname *a_cnp;
};
#endif
static int
zfs_freebsd_lookup(struct vop_lookup_args *ap, boolean_t cached)
{
struct componentname *cnp = ap->a_cnp;
char nm[NAME_MAX + 1];
ASSERT3U(cnp->cn_namelen, <, sizeof (nm));
strlcpy(nm, cnp->cn_nameptr, MIN(cnp->cn_namelen + 1, sizeof (nm)));
return (zfs_lookup(ap->a_dvp, nm, ap->a_vpp, cnp, cnp->cn_nameiop,
cnp->cn_cred, 0, cached));
}
static int
zfs_freebsd_cachedlookup(struct vop_cachedlookup_args *ap)
{
return (zfs_freebsd_lookup((struct vop_lookup_args *)ap, B_TRUE));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_lookup_args {
struct vnode *a_dvp;
struct vnode **a_vpp;
struct componentname *a_cnp;
};
#endif
static int
zfs_cache_lookup(struct vop_lookup_args *ap)
{
zfsvfs_t *zfsvfs;
zfsvfs = ap->a_dvp->v_mount->mnt_data;
if (zfsvfs->z_use_namecache)
return (vfs_cache_lookup(ap));
else
return (zfs_freebsd_lookup(ap, B_FALSE));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_create_args {
struct vnode *a_dvp;
struct vnode **a_vpp;
struct componentname *a_cnp;
struct vattr *a_vap;
};
#endif
static int
zfs_freebsd_create(struct vop_create_args *ap)
{
zfsvfs_t *zfsvfs;
struct componentname *cnp = ap->a_cnp;
vattr_t *vap = ap->a_vap;
znode_t *zp = NULL;
int rc, mode;
#if __FreeBSD_version < 1400068
ASSERT(cnp->cn_flags & SAVENAME);
#endif
vattr_init_mask(vap);
mode = vap->va_mode & ALLPERMS;
zfsvfs = ap->a_dvp->v_mount->mnt_data;
*ap->a_vpp = NULL;
rc = zfs_create(VTOZ(ap->a_dvp), cnp->cn_nameptr, vap, 0, mode,
&zp, cnp->cn_cred, 0 /* flag */, NULL /* vsecattr */, NULL);
if (rc == 0)
*ap->a_vpp = ZTOV(zp);
if (zfsvfs->z_use_namecache &&
rc == 0 && (cnp->cn_flags & MAKEENTRY) != 0)
cache_enter(ap->a_dvp, *ap->a_vpp, cnp);
return (rc);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_remove_args {
struct vnode *a_dvp;
struct vnode *a_vp;
struct componentname *a_cnp;
};
#endif
static int
zfs_freebsd_remove(struct vop_remove_args *ap)
{
#if __FreeBSD_version < 1400068
ASSERT(ap->a_cnp->cn_flags & SAVENAME);
#endif
return (zfs_remove_(ap->a_dvp, ap->a_vp, ap->a_cnp->cn_nameptr,
ap->a_cnp->cn_cred));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_mkdir_args {
struct vnode *a_dvp;
struct vnode **a_vpp;
struct componentname *a_cnp;
struct vattr *a_vap;
};
#endif
static int
zfs_freebsd_mkdir(struct vop_mkdir_args *ap)
{
vattr_t *vap = ap->a_vap;
znode_t *zp = NULL;
int rc;
#if __FreeBSD_version < 1400068
ASSERT(ap->a_cnp->cn_flags & SAVENAME);
#endif
vattr_init_mask(vap);
*ap->a_vpp = NULL;
rc = zfs_mkdir(VTOZ(ap->a_dvp), ap->a_cnp->cn_nameptr, vap, &zp,
ap->a_cnp->cn_cred, 0, NULL, NULL);
if (rc == 0)
*ap->a_vpp = ZTOV(zp);
return (rc);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_rmdir_args {
struct vnode *a_dvp;
struct vnode *a_vp;
struct componentname *a_cnp;
};
#endif
static int
zfs_freebsd_rmdir(struct vop_rmdir_args *ap)
{
struct componentname *cnp = ap->a_cnp;
#if __FreeBSD_version < 1400068
ASSERT(cnp->cn_flags & SAVENAME);
#endif
return (zfs_rmdir_(ap->a_dvp, ap->a_vp, cnp->cn_nameptr, cnp->cn_cred));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_readdir_args {
struct vnode *a_vp;
struct uio *a_uio;
struct ucred *a_cred;
int *a_eofflag;
int *a_ncookies;
cookie_t **a_cookies;
};
#endif
static int
zfs_freebsd_readdir(struct vop_readdir_args *ap)
{
zfs_uio_t uio;
zfs_uio_init(&uio, ap->a_uio);
return (zfs_readdir(ap->a_vp, &uio, ap->a_cred, ap->a_eofflag,
ap->a_ncookies, ap->a_cookies));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_fsync_args {
struct vnode *a_vp;
int a_waitfor;
struct thread *a_td;
};
#endif
static int
zfs_freebsd_fsync(struct vop_fsync_args *ap)
{
return (zfs_fsync(VTOZ(ap->a_vp), 0, ap->a_td->td_ucred));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_getattr_args {
struct vnode *a_vp;
struct vattr *a_vap;
struct ucred *a_cred;
};
#endif
static int
zfs_freebsd_getattr(struct vop_getattr_args *ap)
{
vattr_t *vap = ap->a_vap;
xvattr_t xvap;
ulong_t fflags = 0;
int error;
xva_init(&xvap);
xvap.xva_vattr = *vap;
xvap.xva_vattr.va_mask |= AT_XVATTR;
/* Convert chflags into ZFS-type flags. */
/* XXX: what about SF_SETTABLE?. */
XVA_SET_REQ(&xvap, XAT_IMMUTABLE);
XVA_SET_REQ(&xvap, XAT_APPENDONLY);
XVA_SET_REQ(&xvap, XAT_NOUNLINK);
XVA_SET_REQ(&xvap, XAT_NODUMP);
XVA_SET_REQ(&xvap, XAT_READONLY);
XVA_SET_REQ(&xvap, XAT_ARCHIVE);
XVA_SET_REQ(&xvap, XAT_SYSTEM);
XVA_SET_REQ(&xvap, XAT_HIDDEN);
XVA_SET_REQ(&xvap, XAT_REPARSE);
XVA_SET_REQ(&xvap, XAT_OFFLINE);
XVA_SET_REQ(&xvap, XAT_SPARSE);
error = zfs_getattr(ap->a_vp, (vattr_t *)&xvap, 0, ap->a_cred);
if (error != 0)
return (error);
/* Convert ZFS xattr into chflags. */
#define FLAG_CHECK(fflag, xflag, xfield) do { \
if (XVA_ISSET_RTN(&xvap, (xflag)) && (xfield) != 0) \
fflags |= (fflag); \
} while (0)
FLAG_CHECK(SF_IMMUTABLE, XAT_IMMUTABLE,
xvap.xva_xoptattrs.xoa_immutable);
FLAG_CHECK(SF_APPEND, XAT_APPENDONLY,
xvap.xva_xoptattrs.xoa_appendonly);
FLAG_CHECK(SF_NOUNLINK, XAT_NOUNLINK,
xvap.xva_xoptattrs.xoa_nounlink);
FLAG_CHECK(UF_ARCHIVE, XAT_ARCHIVE,
xvap.xva_xoptattrs.xoa_archive);
FLAG_CHECK(UF_NODUMP, XAT_NODUMP,
xvap.xva_xoptattrs.xoa_nodump);
FLAG_CHECK(UF_READONLY, XAT_READONLY,
xvap.xva_xoptattrs.xoa_readonly);
FLAG_CHECK(UF_SYSTEM, XAT_SYSTEM,
xvap.xva_xoptattrs.xoa_system);
FLAG_CHECK(UF_HIDDEN, XAT_HIDDEN,
xvap.xva_xoptattrs.xoa_hidden);
FLAG_CHECK(UF_REPARSE, XAT_REPARSE,
xvap.xva_xoptattrs.xoa_reparse);
FLAG_CHECK(UF_OFFLINE, XAT_OFFLINE,
xvap.xva_xoptattrs.xoa_offline);
FLAG_CHECK(UF_SPARSE, XAT_SPARSE,
xvap.xva_xoptattrs.xoa_sparse);
#undef FLAG_CHECK
*vap = xvap.xva_vattr;
vap->va_flags = fflags;
return (0);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_setattr_args {
struct vnode *a_vp;
struct vattr *a_vap;
struct ucred *a_cred;
};
#endif
static int
zfs_freebsd_setattr(struct vop_setattr_args *ap)
{
vnode_t *vp = ap->a_vp;
vattr_t *vap = ap->a_vap;
cred_t *cred = ap->a_cred;
xvattr_t xvap;
ulong_t fflags;
uint64_t zflags;
vattr_init_mask(vap);
vap->va_mask &= ~AT_NOSET;
xva_init(&xvap);
xvap.xva_vattr = *vap;
zflags = VTOZ(vp)->z_pflags;
if (vap->va_flags != VNOVAL) {
zfsvfs_t *zfsvfs = VTOZ(vp)->z_zfsvfs;
int error;
if (zfsvfs->z_use_fuids == B_FALSE)
return (EOPNOTSUPP);
fflags = vap->va_flags;
/*
* XXX KDM
* We need to figure out whether it makes sense to allow
* UF_REPARSE through, since we don't really have other
* facilities to handle reparse points and zfs_setattr()
* doesn't currently allow setting that attribute anyway.
*/
if ((fflags & ~(SF_IMMUTABLE|SF_APPEND|SF_NOUNLINK|UF_ARCHIVE|
UF_NODUMP|UF_SYSTEM|UF_HIDDEN|UF_READONLY|UF_REPARSE|
UF_OFFLINE|UF_SPARSE)) != 0)
return (EOPNOTSUPP);
/*
* Unprivileged processes are not permitted to unset system
* flags, or modify flags if any system flags are set.
* Privileged non-jail processes may not modify system flags
* if securelevel > 0 and any existing system flags are set.
* Privileged jail processes behave like privileged non-jail
* processes if the PR_ALLOW_CHFLAGS permission bit is set;
* otherwise, they behave like unprivileged processes.
*/
if (secpolicy_fs_owner(vp->v_mount, cred) == 0 ||
spl_priv_check_cred(cred, PRIV_VFS_SYSFLAGS) == 0) {
if (zflags &
(ZFS_IMMUTABLE | ZFS_APPENDONLY | ZFS_NOUNLINK)) {
error = securelevel_gt(cred, 0);
if (error != 0)
return (error);
}
} else {
/*
* Callers may only modify the file flags on
* objects they have VADMIN rights for.
*/
if ((error = VOP_ACCESS(vp, VADMIN, cred,
curthread)) != 0)
return (error);
if (zflags &
(ZFS_IMMUTABLE | ZFS_APPENDONLY |
ZFS_NOUNLINK)) {
return (EPERM);
}
if (fflags &
(SF_IMMUTABLE | SF_APPEND | SF_NOUNLINK)) {
return (EPERM);
}
}
#define FLAG_CHANGE(fflag, zflag, xflag, xfield) do { \
if (((fflags & (fflag)) && !(zflags & (zflag))) || \
((zflags & (zflag)) && !(fflags & (fflag)))) { \
XVA_SET_REQ(&xvap, (xflag)); \
(xfield) = ((fflags & (fflag)) != 0); \
} \
} while (0)
/* Convert chflags into ZFS-type flags. */
/* XXX: what about SF_SETTABLE?. */
FLAG_CHANGE(SF_IMMUTABLE, ZFS_IMMUTABLE, XAT_IMMUTABLE,
xvap.xva_xoptattrs.xoa_immutable);
FLAG_CHANGE(SF_APPEND, ZFS_APPENDONLY, XAT_APPENDONLY,
xvap.xva_xoptattrs.xoa_appendonly);
FLAG_CHANGE(SF_NOUNLINK, ZFS_NOUNLINK, XAT_NOUNLINK,
xvap.xva_xoptattrs.xoa_nounlink);
FLAG_CHANGE(UF_ARCHIVE, ZFS_ARCHIVE, XAT_ARCHIVE,
xvap.xva_xoptattrs.xoa_archive);
FLAG_CHANGE(UF_NODUMP, ZFS_NODUMP, XAT_NODUMP,
xvap.xva_xoptattrs.xoa_nodump);
FLAG_CHANGE(UF_READONLY, ZFS_READONLY, XAT_READONLY,
xvap.xva_xoptattrs.xoa_readonly);
FLAG_CHANGE(UF_SYSTEM, ZFS_SYSTEM, XAT_SYSTEM,
xvap.xva_xoptattrs.xoa_system);
FLAG_CHANGE(UF_HIDDEN, ZFS_HIDDEN, XAT_HIDDEN,
xvap.xva_xoptattrs.xoa_hidden);
FLAG_CHANGE(UF_REPARSE, ZFS_REPARSE, XAT_REPARSE,
xvap.xva_xoptattrs.xoa_reparse);
FLAG_CHANGE(UF_OFFLINE, ZFS_OFFLINE, XAT_OFFLINE,
xvap.xva_xoptattrs.xoa_offline);
FLAG_CHANGE(UF_SPARSE, ZFS_SPARSE, XAT_SPARSE,
xvap.xva_xoptattrs.xoa_sparse);
#undef FLAG_CHANGE
}
if (vap->va_birthtime.tv_sec != VNOVAL) {
xvap.xva_vattr.va_mask |= AT_XVATTR;
XVA_SET_REQ(&xvap, XAT_CREATETIME);
}
return (zfs_setattr(VTOZ(vp), (vattr_t *)&xvap, 0, cred, NULL));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_rename_args {
struct vnode *a_fdvp;
struct vnode *a_fvp;
struct componentname *a_fcnp;
struct vnode *a_tdvp;
struct vnode *a_tvp;
struct componentname *a_tcnp;
};
#endif
static int
zfs_freebsd_rename(struct vop_rename_args *ap)
{
vnode_t *fdvp = ap->a_fdvp;
vnode_t *fvp = ap->a_fvp;
vnode_t *tdvp = ap->a_tdvp;
vnode_t *tvp = ap->a_tvp;
int error;
#if __FreeBSD_version < 1400068
ASSERT(ap->a_fcnp->cn_flags & (SAVENAME|SAVESTART));
ASSERT(ap->a_tcnp->cn_flags & (SAVENAME|SAVESTART));
#endif
error = zfs_do_rename(fdvp, &fvp, ap->a_fcnp, tdvp, &tvp,
ap->a_tcnp, ap->a_fcnp->cn_cred);
vrele(fdvp);
vrele(fvp);
vrele(tdvp);
if (tvp != NULL)
vrele(tvp);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_symlink_args {
struct vnode *a_dvp;
struct vnode **a_vpp;
struct componentname *a_cnp;
struct vattr *a_vap;
char *a_target;
};
#endif
static int
zfs_freebsd_symlink(struct vop_symlink_args *ap)
{
struct componentname *cnp = ap->a_cnp;
vattr_t *vap = ap->a_vap;
znode_t *zp = NULL;
#if __FreeBSD_version >= 1300139
char *symlink;
size_t symlink_len;
#endif
int rc;
#if __FreeBSD_version < 1400068
ASSERT(cnp->cn_flags & SAVENAME);
#endif
vap->va_type = VLNK; /* FreeBSD: Syscall only sets va_mode. */
vattr_init_mask(vap);
*ap->a_vpp = NULL;
rc = zfs_symlink(VTOZ(ap->a_dvp), cnp->cn_nameptr, vap,
ap->a_target, &zp, cnp->cn_cred, 0 /* flags */, NULL);
if (rc == 0) {
*ap->a_vpp = ZTOV(zp);
ASSERT_VOP_ELOCKED(ZTOV(zp), __func__);
#if __FreeBSD_version >= 1300139
MPASS(zp->z_cached_symlink == NULL);
symlink_len = strlen(ap->a_target);
symlink = cache_symlink_alloc(symlink_len + 1, M_WAITOK);
if (symlink != NULL) {
memcpy(symlink, ap->a_target, symlink_len);
symlink[symlink_len] = '\0';
atomic_store_rel_ptr((uintptr_t *)&zp->z_cached_symlink,
(uintptr_t)symlink);
}
#endif
}
return (rc);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_readlink_args {
struct vnode *a_vp;
struct uio *a_uio;
struct ucred *a_cred;
};
#endif
static int
zfs_freebsd_readlink(struct vop_readlink_args *ap)
{
zfs_uio_t uio;
int error;
#if __FreeBSD_version >= 1300139
znode_t *zp = VTOZ(ap->a_vp);
char *symlink, *base;
size_t symlink_len;
bool trycache;
#endif
zfs_uio_init(&uio, ap->a_uio);
#if __FreeBSD_version >= 1300139
trycache = false;
if (zfs_uio_segflg(&uio) == UIO_SYSSPACE &&
zfs_uio_iovcnt(&uio) == 1) {
base = zfs_uio_iovbase(&uio, 0);
symlink_len = zfs_uio_iovlen(&uio, 0);
trycache = true;
}
#endif
error = zfs_readlink(ap->a_vp, &uio, ap->a_cred, NULL);
#if __FreeBSD_version >= 1300139
if (atomic_load_ptr(&zp->z_cached_symlink) != NULL ||
error != 0 || !trycache) {
return (error);
}
symlink_len -= zfs_uio_resid(&uio);
symlink = cache_symlink_alloc(symlink_len + 1, M_WAITOK);
if (symlink != NULL) {
memcpy(symlink, base, symlink_len);
symlink[symlink_len] = '\0';
if (!atomic_cmpset_rel_ptr((uintptr_t *)&zp->z_cached_symlink,
(uintptr_t)NULL, (uintptr_t)symlink)) {
cache_symlink_free(symlink, symlink_len + 1);
}
}
#endif
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_link_args {
struct vnode *a_tdvp;
struct vnode *a_vp;
struct componentname *a_cnp;
};
#endif
static int
zfs_freebsd_link(struct vop_link_args *ap)
{
struct componentname *cnp = ap->a_cnp;
vnode_t *vp = ap->a_vp;
vnode_t *tdvp = ap->a_tdvp;
if (tdvp->v_mount != vp->v_mount)
return (EXDEV);
#if __FreeBSD_version < 1400068
ASSERT(cnp->cn_flags & SAVENAME);
#endif
return (zfs_link(VTOZ(tdvp), VTOZ(vp),
cnp->cn_nameptr, cnp->cn_cred, 0));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_inactive_args {
struct vnode *a_vp;
struct thread *a_td;
};
#endif
static int
zfs_freebsd_inactive(struct vop_inactive_args *ap)
{
vnode_t *vp = ap->a_vp;
#if __FreeBSD_version >= 1300123
zfs_inactive(vp, curthread->td_ucred, NULL);
#else
zfs_inactive(vp, ap->a_td->td_ucred, NULL);
#endif
return (0);
}
#if __FreeBSD_version >= 1300042
#ifndef _SYS_SYSPROTO_H_
struct vop_need_inactive_args {
struct vnode *a_vp;
struct thread *a_td;
};
#endif
static int
zfs_freebsd_need_inactive(struct vop_need_inactive_args *ap)
{
vnode_t *vp = ap->a_vp;
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
int need;
if (vn_need_pageq_flush(vp))
return (1);
if (!ZFS_TEARDOWN_INACTIVE_TRY_ENTER_READ(zfsvfs))
return (1);
need = (zp->z_sa_hdl == NULL || zp->z_unlinked || zp->z_atime_dirty);
ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs);
return (need);
}
#endif
#ifndef _SYS_SYSPROTO_H_
struct vop_reclaim_args {
struct vnode *a_vp;
struct thread *a_td;
};
#endif
static int
zfs_freebsd_reclaim(struct vop_reclaim_args *ap)
{
vnode_t *vp = ap->a_vp;
znode_t *zp = VTOZ(vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
ASSERT3P(zp, !=, NULL);
#if __FreeBSD_version < 1300042
/* Destroy the vm object and flush associated pages. */
vnode_destroy_vobject(vp);
#endif
/*
* z_teardown_inactive_lock protects from a race with
* zfs_znode_dmu_fini in zfsvfs_teardown during
* force unmount.
*/
ZFS_TEARDOWN_INACTIVE_ENTER_READ(zfsvfs);
if (zp->z_sa_hdl == NULL)
zfs_znode_free(zp);
else
zfs_zinactive(zp);
ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs);
vp->v_data = NULL;
return (0);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_fid_args {
struct vnode *a_vp;
struct fid *a_fid;
};
#endif
static int
zfs_freebsd_fid(struct vop_fid_args *ap)
{
return (zfs_fid(ap->a_vp, (void *)ap->a_fid, NULL));
}
#ifndef _SYS_SYSPROTO_H_
struct vop_pathconf_args {
struct vnode *a_vp;
int a_name;
register_t *a_retval;
} *ap;
#endif
static int
zfs_freebsd_pathconf(struct vop_pathconf_args *ap)
{
ulong_t val;
int error;
error = zfs_pathconf(ap->a_vp, ap->a_name, &val,
curthread->td_ucred, NULL);
if (error == 0) {
*ap->a_retval = val;
return (error);
}
if (error != EOPNOTSUPP)
return (error);
switch (ap->a_name) {
case _PC_NAME_MAX:
*ap->a_retval = NAME_MAX;
return (0);
#if __FreeBSD_version >= 1400032
case _PC_DEALLOC_PRESENT:
*ap->a_retval = 1;
return (0);
#endif
case _PC_PIPE_BUF:
if (ap->a_vp->v_type == VDIR || ap->a_vp->v_type == VFIFO) {
*ap->a_retval = PIPE_BUF;
return (0);
}
return (EINVAL);
default:
return (vop_stdpathconf(ap));
}
}
static int zfs_xattr_compat = 1;
static int
zfs_check_attrname(const char *name)
{
/* We don't allow '/' character in attribute name. */
if (strchr(name, '/') != NULL)
return (SET_ERROR(EINVAL));
/* We don't allow attribute names that start with a namespace prefix. */
if (ZFS_XA_NS_PREFIX_FORBIDDEN(name))
return (SET_ERROR(EINVAL));
return (0);
}
/*
* FreeBSD's extended attributes namespace defines file name prefix for ZFS'
* extended attribute name:
*
* NAMESPACE XATTR_COMPAT PREFIX
* system * freebsd:system:
* user 1 (none, can be used to access ZFS
* fsattr(5) attributes created on Solaris)
* user 0 user.
*/
static int
zfs_create_attrname(int attrnamespace, const char *name, char *attrname,
size_t size, boolean_t compat)
{
const char *namespace, *prefix, *suffix;
memset(attrname, 0, size);
switch (attrnamespace) {
case EXTATTR_NAMESPACE_USER:
if (compat) {
/*
* This is the default namespace by which we can access
* all attributes created on Solaris.
*/
prefix = namespace = suffix = "";
} else {
/*
* This is compatible with the user namespace encoding
* on Linux prior to xattr_compat, but nothing
* else.
*/
prefix = "";
namespace = "user";
suffix = ".";
}
break;
case EXTATTR_NAMESPACE_SYSTEM:
prefix = "freebsd:";
namespace = EXTATTR_NAMESPACE_SYSTEM_STRING;
suffix = ":";
break;
case EXTATTR_NAMESPACE_EMPTY:
default:
return (SET_ERROR(EINVAL));
}
if (snprintf(attrname, size, "%s%s%s%s", prefix, namespace, suffix,
name) >= size) {
return (SET_ERROR(ENAMETOOLONG));
}
return (0);
}
static int
zfs_ensure_xattr_cached(znode_t *zp)
{
int error = 0;
ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock));
if (zp->z_xattr_cached != NULL)
return (0);
if (rw_write_held(&zp->z_xattr_lock))
return (zfs_sa_get_xattr(zp));
if (!rw_tryupgrade(&zp->z_xattr_lock)) {
rw_exit(&zp->z_xattr_lock);
rw_enter(&zp->z_xattr_lock, RW_WRITER);
}
if (zp->z_xattr_cached == NULL)
error = zfs_sa_get_xattr(zp);
rw_downgrade(&zp->z_xattr_lock);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_getextattr {
IN struct vnode *a_vp;
IN int a_attrnamespace;
IN const char *a_name;
INOUT struct uio *a_uio;
OUT size_t *a_size;
IN struct ucred *a_cred;
IN struct thread *a_td;
};
#endif
static int
zfs_getextattr_dir(struct vop_getextattr_args *ap, const char *attrname)
{
struct thread *td = ap->a_td;
struct nameidata nd;
struct vattr va;
vnode_t *xvp = NULL, *vp;
int error, flags;
error = zfs_lookup(ap->a_vp, NULL, &xvp, NULL, 0, ap->a_cred,
LOOKUP_XATTR, B_FALSE);
if (error != 0)
return (error);
flags = FREAD;
#if __FreeBSD_version < 1400043
NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, attrname,
xvp, td);
#else
NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, attrname, xvp);
#endif
error = vn_open_cred(&nd, &flags, 0, VN_OPEN_INVFS, ap->a_cred, NULL);
if (error != 0)
return (SET_ERROR(error));
vp = nd.ni_vp;
NDFREE_PNBUF(&nd);
if (ap->a_size != NULL) {
error = VOP_GETATTR(vp, &va, ap->a_cred);
if (error == 0)
*ap->a_size = (size_t)va.va_size;
} else if (ap->a_uio != NULL)
error = VOP_READ(vp, ap->a_uio, IO_UNIT, ap->a_cred);
VOP_UNLOCK1(vp);
vn_close(vp, flags, ap->a_cred, td);
return (error);
}
static int
zfs_getextattr_sa(struct vop_getextattr_args *ap, const char *attrname)
{
znode_t *zp = VTOZ(ap->a_vp);
uchar_t *nv_value;
uint_t nv_size;
int error;
error = zfs_ensure_xattr_cached(zp);
if (error != 0)
return (error);
ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock));
ASSERT3P(zp->z_xattr_cached, !=, NULL);
error = nvlist_lookup_byte_array(zp->z_xattr_cached, attrname,
&nv_value, &nv_size);
if (error != 0)
return (SET_ERROR(error));
if (ap->a_size != NULL)
*ap->a_size = nv_size;
else if (ap->a_uio != NULL)
error = uiomove(nv_value, nv_size, ap->a_uio);
if (error != 0)
return (SET_ERROR(error));
return (0);
}
static int
zfs_getextattr_impl(struct vop_getextattr_args *ap, boolean_t compat)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
char attrname[EXTATTR_MAXNAMELEN+1];
int error;
error = zfs_create_attrname(ap->a_attrnamespace, ap->a_name, attrname,
sizeof (attrname), compat);
if (error != 0)
return (error);
error = ENOENT;
if (zfsvfs->z_use_sa && zp->z_is_sa)
error = zfs_getextattr_sa(ap, attrname);
if (error == ENOENT)
error = zfs_getextattr_dir(ap, attrname);
return (error);
}
/*
* Vnode operation to retrieve a named extended attribute.
*/
static int
zfs_getextattr(struct vop_getextattr_args *ap)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
/*
* If the xattr property is off, refuse the request.
*/
if (!(zfsvfs->z_flags & ZSB_XATTR))
return (SET_ERROR(EOPNOTSUPP));
error = extattr_check_cred(ap->a_vp, ap->a_attrnamespace,
ap->a_cred, ap->a_td, VREAD);
if (error != 0)
return (SET_ERROR(error));
error = zfs_check_attrname(ap->a_name);
if (error != 0)
return (error);
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
error = ENOENT;
rw_enter(&zp->z_xattr_lock, RW_READER);
error = zfs_getextattr_impl(ap, zfs_xattr_compat);
if ((error == ENOENT || error == ENOATTR) &&
ap->a_attrnamespace == EXTATTR_NAMESPACE_USER) {
/*
* Fall back to the alternate namespace format if we failed to
* find a user xattr.
*/
error = zfs_getextattr_impl(ap, !zfs_xattr_compat);
}
rw_exit(&zp->z_xattr_lock);
zfs_exit(zfsvfs, FTAG);
if (error == ENOENT)
error = SET_ERROR(ENOATTR);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_deleteextattr {
IN struct vnode *a_vp;
IN int a_attrnamespace;
IN const char *a_name;
IN struct ucred *a_cred;
IN struct thread *a_td;
};
#endif
static int
zfs_deleteextattr_dir(struct vop_deleteextattr_args *ap, const char *attrname)
{
struct nameidata nd;
vnode_t *xvp = NULL, *vp;
int error;
error = zfs_lookup(ap->a_vp, NULL, &xvp, NULL, 0, ap->a_cred,
LOOKUP_XATTR, B_FALSE);
if (error != 0)
return (error);
#if __FreeBSD_version < 1400043
NDINIT_ATVP(&nd, DELETE, NOFOLLOW | LOCKPARENT | LOCKLEAF,
UIO_SYSSPACE, attrname, xvp, ap->a_td);
#else
NDINIT_ATVP(&nd, DELETE, NOFOLLOW | LOCKPARENT | LOCKLEAF,
UIO_SYSSPACE, attrname, xvp);
#endif
error = namei(&nd);
if (error != 0)
return (SET_ERROR(error));
vp = nd.ni_vp;
error = VOP_REMOVE(nd.ni_dvp, vp, &nd.ni_cnd);
NDFREE_PNBUF(&nd);
vput(nd.ni_dvp);
if (vp == nd.ni_dvp)
vrele(vp);
else
vput(vp);
return (error);
}
static int
zfs_deleteextattr_sa(struct vop_deleteextattr_args *ap, const char *attrname)
{
znode_t *zp = VTOZ(ap->a_vp);
nvlist_t *nvl;
int error;
error = zfs_ensure_xattr_cached(zp);
if (error != 0)
return (error);
ASSERT(RW_WRITE_HELD(&zp->z_xattr_lock));
ASSERT3P(zp->z_xattr_cached, !=, NULL);
nvl = zp->z_xattr_cached;
error = nvlist_remove(nvl, attrname, DATA_TYPE_BYTE_ARRAY);
if (error != 0)
error = SET_ERROR(error);
else
error = zfs_sa_set_xattr(zp, attrname, NULL, 0);
if (error != 0) {
zp->z_xattr_cached = NULL;
nvlist_free(nvl);
}
return (error);
}
static int
zfs_deleteextattr_impl(struct vop_deleteextattr_args *ap, boolean_t compat)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
char attrname[EXTATTR_MAXNAMELEN+1];
int error;
error = zfs_create_attrname(ap->a_attrnamespace, ap->a_name, attrname,
sizeof (attrname), compat);
if (error != 0)
return (error);
error = ENOENT;
if (zfsvfs->z_use_sa && zp->z_is_sa)
error = zfs_deleteextattr_sa(ap, attrname);
if (error == ENOENT)
error = zfs_deleteextattr_dir(ap, attrname);
return (error);
}
/*
* Vnode operation to remove a named attribute.
*/
static int
zfs_deleteextattr(struct vop_deleteextattr_args *ap)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
/*
* If the xattr property is off, refuse the request.
*/
if (!(zfsvfs->z_flags & ZSB_XATTR))
return (SET_ERROR(EOPNOTSUPP));
error = extattr_check_cred(ap->a_vp, ap->a_attrnamespace,
ap->a_cred, ap->a_td, VWRITE);
if (error != 0)
return (SET_ERROR(error));
error = zfs_check_attrname(ap->a_name);
if (error != 0)
return (error);
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
rw_enter(&zp->z_xattr_lock, RW_WRITER);
error = zfs_deleteextattr_impl(ap, zfs_xattr_compat);
if ((error == ENOENT || error == ENOATTR) &&
ap->a_attrnamespace == EXTATTR_NAMESPACE_USER) {
/*
* Fall back to the alternate namespace format if we failed to
* find a user xattr.
*/
error = zfs_deleteextattr_impl(ap, !zfs_xattr_compat);
}
rw_exit(&zp->z_xattr_lock);
zfs_exit(zfsvfs, FTAG);
if (error == ENOENT)
error = SET_ERROR(ENOATTR);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_setextattr {
IN struct vnode *a_vp;
IN int a_attrnamespace;
IN const char *a_name;
INOUT struct uio *a_uio;
IN struct ucred *a_cred;
IN struct thread *a_td;
};
#endif
static int
zfs_setextattr_dir(struct vop_setextattr_args *ap, const char *attrname)
{
struct thread *td = ap->a_td;
struct nameidata nd;
struct vattr va;
vnode_t *xvp = NULL, *vp;
int error, flags;
error = zfs_lookup(ap->a_vp, NULL, &xvp, NULL, 0, ap->a_cred,
LOOKUP_XATTR | CREATE_XATTR_DIR, B_FALSE);
if (error != 0)
return (error);
flags = FFLAGS(O_WRONLY | O_CREAT);
#if __FreeBSD_version < 1400043
NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, attrname, xvp, td);
#else
NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, attrname, xvp);
#endif
error = vn_open_cred(&nd, &flags, 0600, VN_OPEN_INVFS, ap->a_cred,
NULL);
if (error != 0)
return (SET_ERROR(error));
vp = nd.ni_vp;
NDFREE_PNBUF(&nd);
VATTR_NULL(&va);
va.va_size = 0;
error = VOP_SETATTR(vp, &va, ap->a_cred);
if (error == 0)
VOP_WRITE(vp, ap->a_uio, IO_UNIT, ap->a_cred);
VOP_UNLOCK1(vp);
vn_close(vp, flags, ap->a_cred, td);
return (error);
}
static int
zfs_setextattr_sa(struct vop_setextattr_args *ap, const char *attrname)
{
znode_t *zp = VTOZ(ap->a_vp);
nvlist_t *nvl;
size_t sa_size;
int error;
error = zfs_ensure_xattr_cached(zp);
if (error != 0)
return (error);
ASSERT(RW_WRITE_HELD(&zp->z_xattr_lock));
ASSERT3P(zp->z_xattr_cached, !=, NULL);
nvl = zp->z_xattr_cached;
size_t entry_size = ap->a_uio->uio_resid;
if (entry_size > DXATTR_MAX_ENTRY_SIZE)
return (SET_ERROR(EFBIG));
error = nvlist_size(nvl, &sa_size, NV_ENCODE_XDR);
if (error != 0)
return (SET_ERROR(error));
if (sa_size > DXATTR_MAX_SA_SIZE)
return (SET_ERROR(EFBIG));
uchar_t *buf = kmem_alloc(entry_size, KM_SLEEP);
error = uiomove(buf, entry_size, ap->a_uio);
if (error != 0) {
error = SET_ERROR(error);
} else {
error = nvlist_add_byte_array(nvl, attrname, buf, entry_size);
if (error != 0)
error = SET_ERROR(error);
}
if (error == 0)
error = zfs_sa_set_xattr(zp, attrname, buf, entry_size);
kmem_free(buf, entry_size);
if (error != 0) {
zp->z_xattr_cached = NULL;
nvlist_free(nvl);
}
return (error);
}
static int
zfs_setextattr_impl(struct vop_setextattr_args *ap, boolean_t compat)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
char attrname[EXTATTR_MAXNAMELEN+1];
int error;
error = zfs_create_attrname(ap->a_attrnamespace, ap->a_name, attrname,
sizeof (attrname), compat);
if (error != 0)
return (error);
struct vop_deleteextattr_args vda = {
.a_vp = ap->a_vp,
.a_attrnamespace = ap->a_attrnamespace,
.a_name = ap->a_name,
.a_cred = ap->a_cred,
.a_td = ap->a_td,
};
error = ENOENT;
if (zfsvfs->z_use_sa && zp->z_is_sa && zfsvfs->z_xattr_sa) {
error = zfs_setextattr_sa(ap, attrname);
if (error == 0) {
/*
* Successfully put into SA, we need to clear the one
* in dir if present.
*/
zfs_deleteextattr_dir(&vda, attrname);
}
}
if (error != 0) {
error = zfs_setextattr_dir(ap, attrname);
if (error == 0 && zp->z_is_sa) {
/*
* Successfully put into dir, we need to clear the one
* in SA if present.
*/
zfs_deleteextattr_sa(&vda, attrname);
}
}
if (error == 0 && ap->a_attrnamespace == EXTATTR_NAMESPACE_USER) {
/*
* Also clear all versions of the alternate compat name.
*/
zfs_deleteextattr_impl(&vda, !compat);
}
return (error);
}
/*
* Vnode operation to set a named attribute.
*/
static int
zfs_setextattr(struct vop_setextattr_args *ap)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
/*
* If the xattr property is off, refuse the request.
*/
if (!(zfsvfs->z_flags & ZSB_XATTR))
return (SET_ERROR(EOPNOTSUPP));
error = extattr_check_cred(ap->a_vp, ap->a_attrnamespace,
ap->a_cred, ap->a_td, VWRITE);
if (error != 0)
return (SET_ERROR(error));
error = zfs_check_attrname(ap->a_name);
if (error != 0)
return (error);
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
rw_enter(&zp->z_xattr_lock, RW_WRITER);
error = zfs_setextattr_impl(ap, zfs_xattr_compat);
rw_exit(&zp->z_xattr_lock);
zfs_exit(zfsvfs, FTAG);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_listextattr {
IN struct vnode *a_vp;
IN int a_attrnamespace;
INOUT struct uio *a_uio;
OUT size_t *a_size;
IN struct ucred *a_cred;
IN struct thread *a_td;
};
#endif
static int
zfs_listextattr_dir(struct vop_listextattr_args *ap, const char *attrprefix)
{
struct thread *td = ap->a_td;
struct nameidata nd;
uint8_t dirbuf[sizeof (struct dirent)];
struct iovec aiov;
struct uio auio;
vnode_t *xvp = NULL, *vp;
int error, eof;
error = zfs_lookup(ap->a_vp, NULL, &xvp, NULL, 0, ap->a_cred,
LOOKUP_XATTR, B_FALSE);
if (error != 0) {
/*
* ENOATTR means that the EA directory does not yet exist,
* i.e. there are no extended attributes there.
*/
if (error == ENOATTR)
error = 0;
return (error);
}
#if __FreeBSD_version < 1400043
NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | LOCKSHARED,
UIO_SYSSPACE, ".", xvp, td);
#else
NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | LOCKSHARED,
UIO_SYSSPACE, ".", xvp);
#endif
error = namei(&nd);
if (error != 0)
return (SET_ERROR(error));
vp = nd.ni_vp;
NDFREE_PNBUF(&nd);
auio.uio_iov = &aiov;
auio.uio_iovcnt = 1;
auio.uio_segflg = UIO_SYSSPACE;
auio.uio_td = td;
auio.uio_rw = UIO_READ;
auio.uio_offset = 0;
size_t plen = strlen(attrprefix);
do {
aiov.iov_base = (void *)dirbuf;
aiov.iov_len = sizeof (dirbuf);
auio.uio_resid = sizeof (dirbuf);
error = VOP_READDIR(vp, &auio, ap->a_cred, &eof, NULL, NULL);
if (error != 0)
break;
int done = sizeof (dirbuf) - auio.uio_resid;
for (int pos = 0; pos < done; ) {
struct dirent *dp = (struct dirent *)(dirbuf + pos);
pos += dp->d_reclen;
/*
* XXX: Temporarily we also accept DT_UNKNOWN, as this
* is what we get when attribute was created on Solaris.
*/
if (dp->d_type != DT_REG && dp->d_type != DT_UNKNOWN)
continue;
else if (plen == 0 &&
ZFS_XA_NS_PREFIX_FORBIDDEN(dp->d_name))
continue;
else if (strncmp(dp->d_name, attrprefix, plen) != 0)
continue;
uint8_t nlen = dp->d_namlen - plen;
if (ap->a_size != NULL) {
*ap->a_size += 1 + nlen;
} else if (ap->a_uio != NULL) {
/*
* Format of extattr name entry is one byte for
* length and the rest for name.
*/
error = uiomove(&nlen, 1, ap->a_uio);
if (error == 0) {
char *namep = dp->d_name + plen;
error = uiomove(namep, nlen, ap->a_uio);
}
if (error != 0) {
error = SET_ERROR(error);
break;
}
}
}
} while (!eof && error == 0);
vput(vp);
return (error);
}
static int
zfs_listextattr_sa(struct vop_listextattr_args *ap, const char *attrprefix)
{
znode_t *zp = VTOZ(ap->a_vp);
int error;
error = zfs_ensure_xattr_cached(zp);
if (error != 0)
return (error);
ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock));
ASSERT3P(zp->z_xattr_cached, !=, NULL);
size_t plen = strlen(attrprefix);
nvpair_t *nvp = NULL;
while ((nvp = nvlist_next_nvpair(zp->z_xattr_cached, nvp)) != NULL) {
ASSERT3U(nvpair_type(nvp), ==, DATA_TYPE_BYTE_ARRAY);
const char *name = nvpair_name(nvp);
if (plen == 0 && ZFS_XA_NS_PREFIX_FORBIDDEN(name))
continue;
else if (strncmp(name, attrprefix, plen) != 0)
continue;
uint8_t nlen = strlen(name) - plen;
if (ap->a_size != NULL) {
*ap->a_size += 1 + nlen;
} else if (ap->a_uio != NULL) {
/*
* Format of extattr name entry is one byte for
* length and the rest for name.
*/
error = uiomove(&nlen, 1, ap->a_uio);
if (error == 0) {
char *namep = __DECONST(char *, name) + plen;
error = uiomove(namep, nlen, ap->a_uio);
}
if (error != 0) {
error = SET_ERROR(error);
break;
}
}
}
return (error);
}
static int
zfs_listextattr_impl(struct vop_listextattr_args *ap, boolean_t compat)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
char attrprefix[16];
int error;
error = zfs_create_attrname(ap->a_attrnamespace, "", attrprefix,
sizeof (attrprefix), compat);
if (error != 0)
return (error);
if (zfsvfs->z_use_sa && zp->z_is_sa)
error = zfs_listextattr_sa(ap, attrprefix);
if (error == 0)
error = zfs_listextattr_dir(ap, attrprefix);
return (error);
}
/*
* Vnode operation to retrieve extended attributes on a vnode.
*/
static int
zfs_listextattr(struct vop_listextattr_args *ap)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
if (ap->a_size != NULL)
*ap->a_size = 0;
/*
* If the xattr property is off, refuse the request.
*/
if (!(zfsvfs->z_flags & ZSB_XATTR))
return (SET_ERROR(EOPNOTSUPP));
error = extattr_check_cred(ap->a_vp, ap->a_attrnamespace,
ap->a_cred, ap->a_td, VREAD);
if (error != 0)
return (SET_ERROR(error));
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
rw_enter(&zp->z_xattr_lock, RW_READER);
error = zfs_listextattr_impl(ap, zfs_xattr_compat);
if (error == 0 && ap->a_attrnamespace == EXTATTR_NAMESPACE_USER) {
/* Also list user xattrs with the alternate format. */
error = zfs_listextattr_impl(ap, !zfs_xattr_compat);
}
rw_exit(&zp->z_xattr_lock);
zfs_exit(zfsvfs, FTAG);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_getacl_args {
struct vnode *vp;
acl_type_t type;
struct acl *aclp;
struct ucred *cred;
struct thread *td;
};
#endif
static int
zfs_freebsd_getacl(struct vop_getacl_args *ap)
{
int error;
vsecattr_t vsecattr;
if (ap->a_type != ACL_TYPE_NFS4)
return (EINVAL);
vsecattr.vsa_mask = VSA_ACE | VSA_ACECNT;
if ((error = zfs_getsecattr(VTOZ(ap->a_vp),
&vsecattr, 0, ap->a_cred)))
return (error);
error = acl_from_aces(ap->a_aclp, vsecattr.vsa_aclentp,
vsecattr.vsa_aclcnt);
if (vsecattr.vsa_aclentp != NULL)
kmem_free(vsecattr.vsa_aclentp, vsecattr.vsa_aclentsz);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_setacl_args {
struct vnode *vp;
acl_type_t type;
struct acl *aclp;
struct ucred *cred;
struct thread *td;
};
#endif
static int
zfs_freebsd_setacl(struct vop_setacl_args *ap)
{
int error;
vsecattr_t vsecattr;
int aclbsize; /* size of acl list in bytes */
aclent_t *aaclp;
if (ap->a_type != ACL_TYPE_NFS4)
return (EINVAL);
if (ap->a_aclp == NULL)
return (EINVAL);
if (ap->a_aclp->acl_cnt < 1 || ap->a_aclp->acl_cnt > MAX_ACL_ENTRIES)
return (EINVAL);
/*
* With NFSv4 ACLs, chmod(2) may need to add additional entries,
* splitting every entry into two and appending "canonical six"
* entries at the end. Don't allow for setting an ACL that would
* cause chmod(2) to run out of ACL entries.
*/
if (ap->a_aclp->acl_cnt * 2 + 6 > ACL_MAX_ENTRIES)
return (ENOSPC);
error = acl_nfs4_check(ap->a_aclp, ap->a_vp->v_type == VDIR);
if (error != 0)
return (error);
vsecattr.vsa_mask = VSA_ACE;
aclbsize = ap->a_aclp->acl_cnt * sizeof (ace_t);
vsecattr.vsa_aclentp = kmem_alloc(aclbsize, KM_SLEEP);
aaclp = vsecattr.vsa_aclentp;
vsecattr.vsa_aclentsz = aclbsize;
aces_from_acl(vsecattr.vsa_aclentp, &vsecattr.vsa_aclcnt, ap->a_aclp);
error = zfs_setsecattr(VTOZ(ap->a_vp), &vsecattr, 0, ap->a_cred);
kmem_free(aaclp, aclbsize);
return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct vop_aclcheck_args {
struct vnode *vp;
acl_type_t type;
struct acl *aclp;
struct ucred *cred;
struct thread *td;
};
#endif
static int
zfs_freebsd_aclcheck(struct vop_aclcheck_args *ap)
{
return (EOPNOTSUPP);
}
static int
zfs_vptocnp(struct vop_vptocnp_args *ap)
{
vnode_t *covered_vp;
vnode_t *vp = ap->a_vp;
zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data;
znode_t *zp = VTOZ(vp);
int ltype;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
/*
* If we are a snapshot mounted under .zfs, run the operation
* on the covered vnode.
*/
if (zp->z_id != zfsvfs->z_root || zfsvfs->z_parent == zfsvfs) {
char name[MAXNAMLEN + 1];
znode_t *dzp;
size_t len;
error = zfs_znode_parent_and_name(zp, &dzp, name);
if (error == 0) {
len = strlen(name);
if (*ap->a_buflen < len)
error = SET_ERROR(ENOMEM);
}
if (error == 0) {
*ap->a_buflen -= len;
memcpy(ap->a_buf + *ap->a_buflen, name, len);
*ap->a_vpp = ZTOV(dzp);
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
zfs_exit(zfsvfs, FTAG);
covered_vp = vp->v_mount->mnt_vnodecovered;
#if __FreeBSD_version >= 1300045
enum vgetstate vs = vget_prep(covered_vp);
#else
vhold(covered_vp);
#endif
ltype = VOP_ISLOCKED(vp);
VOP_UNLOCK1(vp);
#if __FreeBSD_version >= 1300045
error = vget_finish(covered_vp, LK_SHARED, vs);
#else
error = vget(covered_vp, LK_SHARED | LK_VNHELD, curthread);
#endif
if (error == 0) {
#if __FreeBSD_version >= 1300123
error = VOP_VPTOCNP(covered_vp, ap->a_vpp, ap->a_buf,
ap->a_buflen);
#else
error = VOP_VPTOCNP(covered_vp, ap->a_vpp, ap->a_cred,
ap->a_buf, ap->a_buflen);
#endif
vput(covered_vp);
}
vn_lock(vp, ltype | LK_RETRY);
if (VN_IS_DOOMED(vp))
error = SET_ERROR(ENOENT);
return (error);
}
#if __FreeBSD_version >= 1400032
static int
zfs_deallocate(struct vop_deallocate_args *ap)
{
znode_t *zp = VTOZ(ap->a_vp);
zfsvfs_t *zfsvfs = zp->z_zfsvfs;
zilog_t *zilog;
off_t off, len, file_sz;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
/*
* Callers might not be able to detect properly that we are read-only,
* so check it explicitly here.
*/
if (zfs_is_readonly(zfsvfs)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EROFS));
}
zilog = zfsvfs->z_log;
off = *ap->a_offset;
len = *ap->a_len;
file_sz = zp->z_size;
if (off + len > file_sz)
len = file_sz - off;
/* Fast path for out-of-range request. */
if (len <= 0) {
*ap->a_len = 0;
zfs_exit(zfsvfs, FTAG);
return (0);
}
error = zfs_freesp(zp, off, len, O_RDWR, TRUE);
if (error == 0) {
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS ||
(ap->a_ioflag & IO_SYNC) != 0)
zil_commit(zilog, zp->z_id);
*ap->a_offset = off + len;
*ap->a_len = 0;
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
#endif
#ifndef _SYS_SYSPROTO_H_
struct vop_copy_file_range_args {
struct vnode *a_invp;
off_t *a_inoffp;
struct vnode *a_outvp;
off_t *a_outoffp;
size_t *a_lenp;
unsigned int a_flags;
struct ucred *a_incred;
struct ucred *a_outcred;
struct thread *a_fsizetd;
}
#endif
/*
* TODO: FreeBSD will only call file system-specific copy_file_range() if both
* files resides under the same mountpoint. In case of ZFS we want to be called
* even is files are in different datasets (but on the same pools, but we need
* to check that ourselves).
*/
static int
zfs_freebsd_copy_file_range(struct vop_copy_file_range_args *ap)
{
zfsvfs_t *outzfsvfs;
struct vnode *invp = ap->a_invp;
struct vnode *outvp = ap->a_outvp;
struct mount *mp;
struct uio io;
int error;
uint64_t len = *ap->a_lenp;
if (!zfs_bclone_enabled) {
mp = NULL;
goto bad_write_fallback;
}
/*
* TODO: If offset/length is not aligned to recordsize, use
* vn_generic_copy_file_range() on this fragment.
* It would be better to do this after we lock the vnodes, but then we
* need something else than vn_generic_copy_file_range().
*/
vn_start_write(outvp, &mp, V_WAIT);
if (__predict_true(mp == outvp->v_mount)) {
outzfsvfs = (zfsvfs_t *)mp->mnt_data;
if (!spa_feature_is_enabled(dmu_objset_spa(outzfsvfs->z_os),
SPA_FEATURE_BLOCK_CLONING)) {
goto bad_write_fallback;
}
}
if (invp == outvp) {
if (vn_lock(outvp, LK_EXCLUSIVE) != 0) {
goto bad_write_fallback;
}
} else {
+#if __FreeBSD_version >= 1400086
vn_lock_pair(invp, false, LK_EXCLUSIVE, outvp, false,
LK_EXCLUSIVE);
+#else
+ vn_lock_pair(invp, false, outvp, false);
+#endif
if (VN_IS_DOOMED(invp) || VN_IS_DOOMED(outvp)) {
goto bad_locked_fallback;
}
}
#ifdef MAC
error = mac_vnode_check_write(curthread->td_ucred, ap->a_outcred,
outvp);
if (error != 0)
goto out_locked;
#endif
io.uio_offset = *ap->a_outoffp;
io.uio_resid = *ap->a_lenp;
error = vn_rlimit_fsize(outvp, &io, ap->a_fsizetd);
if (error != 0)
goto out_locked;
error = zfs_clone_range(VTOZ(invp), ap->a_inoffp, VTOZ(outvp),
ap->a_outoffp, &len, ap->a_outcred);
if (error == EXDEV)
goto bad_locked_fallback;
*ap->a_lenp = (size_t)len;
out_locked:
if (invp != outvp)
VOP_UNLOCK(invp);
VOP_UNLOCK(outvp);
if (mp != NULL)
vn_finished_write(mp);
return (error);
bad_locked_fallback:
if (invp != outvp)
VOP_UNLOCK(invp);
VOP_UNLOCK(outvp);
bad_write_fallback:
if (mp != NULL)
vn_finished_write(mp);
error = vn_generic_copy_file_range(ap->a_invp, ap->a_inoffp,
ap->a_outvp, ap->a_outoffp, ap->a_lenp, ap->a_flags,
ap->a_incred, ap->a_outcred, ap->a_fsizetd);
return (error);
}
struct vop_vector zfs_vnodeops;
struct vop_vector zfs_fifoops;
struct vop_vector zfs_shareops;
struct vop_vector zfs_vnodeops = {
.vop_default = &default_vnodeops,
.vop_inactive = zfs_freebsd_inactive,
#if __FreeBSD_version >= 1300042
.vop_need_inactive = zfs_freebsd_need_inactive,
#endif
.vop_reclaim = zfs_freebsd_reclaim,
#if __FreeBSD_version >= 1300102
.vop_fplookup_vexec = zfs_freebsd_fplookup_vexec,
#endif
#if __FreeBSD_version >= 1300139
.vop_fplookup_symlink = zfs_freebsd_fplookup_symlink,
#endif
.vop_access = zfs_freebsd_access,
.vop_allocate = VOP_EINVAL,
#if __FreeBSD_version >= 1400032
.vop_deallocate = zfs_deallocate,
#endif
.vop_lookup = zfs_cache_lookup,
.vop_cachedlookup = zfs_freebsd_cachedlookup,
.vop_getattr = zfs_freebsd_getattr,
.vop_setattr = zfs_freebsd_setattr,
.vop_create = zfs_freebsd_create,
.vop_mknod = (vop_mknod_t *)zfs_freebsd_create,
.vop_mkdir = zfs_freebsd_mkdir,
.vop_readdir = zfs_freebsd_readdir,
.vop_fsync = zfs_freebsd_fsync,
.vop_open = zfs_freebsd_open,
.vop_close = zfs_freebsd_close,
.vop_rmdir = zfs_freebsd_rmdir,
.vop_ioctl = zfs_freebsd_ioctl,
.vop_link = zfs_freebsd_link,
.vop_symlink = zfs_freebsd_symlink,
.vop_readlink = zfs_freebsd_readlink,
.vop_read = zfs_freebsd_read,
.vop_write = zfs_freebsd_write,
.vop_remove = zfs_freebsd_remove,
.vop_rename = zfs_freebsd_rename,
.vop_pathconf = zfs_freebsd_pathconf,
.vop_bmap = zfs_freebsd_bmap,
.vop_fid = zfs_freebsd_fid,
.vop_getextattr = zfs_getextattr,
.vop_deleteextattr = zfs_deleteextattr,
.vop_setextattr = zfs_setextattr,
.vop_listextattr = zfs_listextattr,
.vop_getacl = zfs_freebsd_getacl,
.vop_setacl = zfs_freebsd_setacl,
.vop_aclcheck = zfs_freebsd_aclcheck,
.vop_getpages = zfs_freebsd_getpages,
.vop_putpages = zfs_freebsd_putpages,
.vop_vptocnp = zfs_vptocnp,
#if __FreeBSD_version >= 1300064
.vop_lock1 = vop_lock,
.vop_unlock = vop_unlock,
.vop_islocked = vop_islocked,
#endif
#if __FreeBSD_version >= 1400043
.vop_add_writecount = vop_stdadd_writecount_nomsync,
#endif
.vop_copy_file_range = zfs_freebsd_copy_file_range,
};
VFS_VOP_VECTOR_REGISTER(zfs_vnodeops);
struct vop_vector zfs_fifoops = {
.vop_default = &fifo_specops,
.vop_fsync = zfs_freebsd_fsync,
#if __FreeBSD_version >= 1300102
.vop_fplookup_vexec = zfs_freebsd_fplookup_vexec,
#endif
#if __FreeBSD_version >= 1300139
.vop_fplookup_symlink = zfs_freebsd_fplookup_symlink,
#endif
.vop_access = zfs_freebsd_access,
.vop_getattr = zfs_freebsd_getattr,
.vop_inactive = zfs_freebsd_inactive,
.vop_read = VOP_PANIC,
.vop_reclaim = zfs_freebsd_reclaim,
.vop_setattr = zfs_freebsd_setattr,
.vop_write = VOP_PANIC,
.vop_pathconf = zfs_freebsd_pathconf,
.vop_fid = zfs_freebsd_fid,
.vop_getacl = zfs_freebsd_getacl,
.vop_setacl = zfs_freebsd_setacl,
.vop_aclcheck = zfs_freebsd_aclcheck,
#if __FreeBSD_version >= 1400043
.vop_add_writecount = vop_stdadd_writecount_nomsync,
#endif
};
VFS_VOP_VECTOR_REGISTER(zfs_fifoops);
/*
* special share hidden files vnode operations template
*/
struct vop_vector zfs_shareops = {
.vop_default = &default_vnodeops,
#if __FreeBSD_version >= 1300121
.vop_fplookup_vexec = VOP_EAGAIN,
#endif
#if __FreeBSD_version >= 1300139
.vop_fplookup_symlink = VOP_EAGAIN,
#endif
.vop_access = zfs_freebsd_access,
.vop_inactive = zfs_freebsd_inactive,
.vop_reclaim = zfs_freebsd_reclaim,
.vop_fid = zfs_freebsd_fid,
.vop_pathconf = zfs_freebsd_pathconf,
#if __FreeBSD_version >= 1400043
.vop_add_writecount = vop_stdadd_writecount_nomsync,
#endif
};
VFS_VOP_VECTOR_REGISTER(zfs_shareops);
ZFS_MODULE_PARAM(zfs, zfs_, xattr_compat, INT, ZMOD_RW,
"Use legacy ZFS xattr naming for writing new user namespace xattrs");
diff --git a/sys/contrib/openzfs/module/os/linux/spl/spl-cred.c b/sys/contrib/openzfs/module/os/linux/spl/spl-cred.c
index f81b9540a639..d407fc66b2de 100644
--- a/sys/contrib/openzfs/module/os/linux/spl/spl-cred.c
+++ b/sys/contrib/openzfs/module/os/linux/spl/spl-cred.c
@@ -1,155 +1,167 @@
/*
* Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
* Copyright (C) 2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
* Written by Brian Behlendorf <behlendorf1@llnl.gov>.
* UCRL-CODE-235197
*
* This file is part of the SPL, Solaris Porting Layer.
*
* The SPL is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation; either version 2 of the License, or (at your
* option) any later version.
*
* The SPL is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* for more details.
*
* You should have received a copy of the GNU General Public License along
* with the SPL. If not, see <http://www.gnu.org/licenses/>.
*
* Solaris Porting Layer (SPL) Credential Implementation.
*/
#include <sys/cred.h>
static int
cr_groups_search(const struct group_info *group_info, kgid_t grp)
{
unsigned int left, right, mid;
int cmp;
if (!group_info)
return (0);
left = 0;
right = group_info->ngroups;
while (left < right) {
mid = (left + right) / 2;
cmp = KGID_TO_SGID(grp) -
KGID_TO_SGID(GROUP_AT(group_info, mid));
if (cmp > 0)
left = mid + 1;
else if (cmp < 0)
right = mid;
else
return (1);
}
return (0);
}
/* Hold a reference on the credential */
void
crhold(cred_t *cr)
{
(void) get_cred((const cred_t *)cr);
}
/* Free a reference on the credential */
void
crfree(cred_t *cr)
{
put_cred((const cred_t *)cr);
}
/* Return the number of supplemental groups */
int
crgetngroups(const cred_t *cr)
{
struct group_info *gi;
int rc;
gi = cr->group_info;
rc = gi->ngroups;
#ifndef HAVE_GROUP_INFO_GID
/*
* For Linux <= 4.8,
* crgetgroups will only returns gi->blocks[0], which contains only
* the first NGROUPS_PER_BLOCK groups.
*/
if (rc > NGROUPS_PER_BLOCK) {
WARN_ON_ONCE(1);
rc = NGROUPS_PER_BLOCK;
}
#endif
return (rc);
}
/*
* Return an array of supplemental gids. The returned address is safe
* to use as long as the caller has taken a reference with crhold().
*
* Linux 4.9 API change, group_info changed from 2d array via ->blocks to 1d
* array via ->gid.
*/
gid_t *
crgetgroups(const cred_t *cr)
{
struct group_info *gi;
gid_t *gids = NULL;
gi = cr->group_info;
#ifdef HAVE_GROUP_INFO_GID
gids = KGIDP_TO_SGIDP(gi->gid);
#else
if (gi->nblocks > 0)
gids = KGIDP_TO_SGIDP(gi->blocks[0]);
#endif
return (gids);
}
/* Check if the passed gid is available in supplied credential. */
int
groupmember(gid_t gid, const cred_t *cr)
{
struct group_info *gi;
int rc;
gi = cr->group_info;
rc = cr_groups_search(gi, SGID_TO_KGID(gid));
return (rc);
}
/* Return the effective user id */
uid_t
crgetuid(const cred_t *cr)
{
return (KUID_TO_SUID(cr->fsuid));
}
/* Return the real user id */
uid_t
crgetruid(const cred_t *cr)
{
return (KUID_TO_SUID(cr->uid));
}
/* Return the effective group id */
gid_t
crgetgid(const cred_t *cr)
{
return (KGID_TO_SGID(cr->fsgid));
}
+/* Return the initial user ns or nop_mnt_idmap */
+zidmap_t *
+zfs_get_init_idmap(void)
+{
+#ifdef HAVE_IOPS_CREATE_IDMAP
+ return ((zidmap_t *)&nop_mnt_idmap);
+#else
+ return ((zidmap_t *)&init_user_ns);
+#endif
+}
+
+EXPORT_SYMBOL(zfs_get_init_idmap);
EXPORT_SYMBOL(crhold);
EXPORT_SYMBOL(crfree);
EXPORT_SYMBOL(crgetuid);
EXPORT_SYMBOL(crgetruid);
EXPORT_SYMBOL(crgetgid);
EXPORT_SYMBOL(crgetngroups);
EXPORT_SYMBOL(crgetgroups);
EXPORT_SYMBOL(groupmember);
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/policy.c b/sys/contrib/openzfs/module/os/linux/zfs/policy.c
index eaf38df864d3..5d1b4383412a 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/policy.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/policy.c
@@ -1,381 +1,382 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2013, Joyent, Inc. All rights reserved.
* Copyright (C) 2016 Lawrence Livermore National Security, LLC.
*
* For Linux the vast majority of this enforcement is already handled via
* the standard Linux VFS permission checks. However certain administrative
* commands which bypass the standard mechanisms may need to make use of
* this functionality.
*/
#include <sys/policy.h>
#include <linux/security.h>
#include <linux/vfs_compat.h>
/*
* The passed credentials cannot be directly verified because Linux only
* provides and interface to check the *current* process credentials. In
* order to handle this the capable() test is only run when the passed
* credentials match the current process credentials or the kcred. In
* all other cases this function must fail and return the passed err.
*/
static int
priv_policy_ns(const cred_t *cr, int capability, int err,
struct user_namespace *ns)
{
if (cr != CRED() && (cr != kcred))
return (err);
#if defined(CONFIG_USER_NS)
if (!(ns ? ns_capable(ns, capability) : capable(capability)))
#else
if (!capable(capability))
#endif
return (err);
return (0);
}
static int
priv_policy(const cred_t *cr, int capability, int err)
{
return (priv_policy_ns(cr, capability, err, cr->user_ns));
}
static int
priv_policy_user(const cred_t *cr, int capability, int err)
{
/*
* All priv_policy_user checks are preceded by kuid/kgid_has_mapping()
* checks. If we cannot do them, we shouldn't be using ns_capable()
* since we don't know whether the affected files are valid in our
* namespace.
*/
#if defined(CONFIG_USER_NS)
return (priv_policy_ns(cr, capability, err, cr->user_ns));
#else
return (priv_policy_ns(cr, capability, err, NULL));
#endif
}
/*
* Checks for operations that are either client-only or are used by
* both clients and servers.
*/
int
secpolicy_nfs(const cred_t *cr)
{
return (priv_policy(cr, CAP_SYS_ADMIN, EPERM));
}
/*
* Catch all system configuration.
*/
int
secpolicy_sys_config(const cred_t *cr, boolean_t checkonly)
{
return (priv_policy(cr, CAP_SYS_ADMIN, EPERM));
}
/*
* Like secpolicy_vnode_access() but we get the actual wanted mode and the
* current mode of the file, not the missing bits.
*
* Enforced in the Linux VFS.
*/
int
secpolicy_vnode_access2(const cred_t *cr, struct inode *ip, uid_t owner,
mode_t curmode, mode_t wantmode)
{
return (0);
}
/*
* This is a special routine for ZFS; it is used to determine whether
* any of the privileges in effect allow any form of access to the
* file. There's no reason to audit this or any reason to record
* this. More work is needed to do the "KPLD" stuff.
*/
int
secpolicy_vnode_any_access(const cred_t *cr, struct inode *ip, uid_t owner)
{
if (crgetuid(cr) == owner)
return (0);
- if (zpl_inode_owner_or_capable(kcred->user_ns, ip))
+ if (zpl_inode_owner_or_capable(zfs_init_idmap, ip))
return (0);
#if defined(CONFIG_USER_NS)
if (!kuid_has_mapping(cr->user_ns, SUID_TO_KUID(owner)))
return (EPERM);
#endif
if (priv_policy_user(cr, CAP_DAC_OVERRIDE, EPERM) == 0)
return (0);
if (priv_policy_user(cr, CAP_DAC_READ_SEARCH, EPERM) == 0)
return (0);
return (EPERM);
}
/*
* Determine if subject can chown owner of a file.
*/
int
secpolicy_vnode_chown(const cred_t *cr, uid_t owner)
{
if (crgetuid(cr) == owner)
return (0);
#if defined(CONFIG_USER_NS)
if (!kuid_has_mapping(cr->user_ns, SUID_TO_KUID(owner)))
return (EPERM);
#endif
return (priv_policy_user(cr, CAP_FOWNER, EPERM));
}
/*
* Determine if subject can change group ownership of a file.
*/
int
secpolicy_vnode_create_gid(const cred_t *cr)
{
return (priv_policy(cr, CAP_SETGID, EPERM));
}
/*
* Policy determines whether we can remove an entry from a directory,
* regardless of permission bits.
*/
int
secpolicy_vnode_remove(const cred_t *cr)
{
return (priv_policy(cr, CAP_FOWNER, EPERM));
}
/*
* Determine that subject can modify the mode of a file. allzone privilege
* needed when modifying root owned object.
*/
int
secpolicy_vnode_setdac(const cred_t *cr, uid_t owner)
{
if (crgetuid(cr) == owner)
return (0);
#if defined(CONFIG_USER_NS)
if (!kuid_has_mapping(cr->user_ns, SUID_TO_KUID(owner)))
return (EPERM);
#endif
return (priv_policy_user(cr, CAP_FOWNER, EPERM));
}
/*
* Are we allowed to retain the set-uid/set-gid bits when
* changing ownership or when writing to a file?
* "issuid" should be true when set-uid; only in that case
* root ownership is checked (setgid is assumed).
*
* Enforced in the Linux VFS.
*/
int
secpolicy_vnode_setid_retain(struct znode *zp __maybe_unused, const cred_t *cr,
boolean_t issuidroot)
{
return (priv_policy_user(cr, CAP_FSETID, EPERM));
}
/*
* Determine that subject can set the file setgid flag.
*/
int
-secpolicy_vnode_setids_setgids(const cred_t *cr, gid_t gid, zuserns_t *mnt_ns,
- zuserns_t *fs_ns)
+secpolicy_vnode_setids_setgids(const cred_t *cr, gid_t gid, zidmap_t *mnt_ns,
+ struct user_namespace *fs_ns)
{
gid = zfs_gid_to_vfsgid(mnt_ns, fs_ns, gid);
#if defined(CONFIG_USER_NS)
if (!kgid_has_mapping(cr->user_ns, SGID_TO_KGID(gid)))
return (EPERM);
#endif
if (crgetgid(cr) != gid && !groupmember(gid, cr))
return (priv_policy_user(cr, CAP_FSETID, EPERM));
return (0);
}
/*
* Determine if the subject can inject faults in the ZFS fault injection
* framework. Requires all privileges.
*/
int
secpolicy_zinject(const cred_t *cr)
{
return (priv_policy(cr, CAP_SYS_ADMIN, EACCES));
}
/*
* Determine if the subject has permission to manipulate ZFS datasets
* (not pools). Equivalent to the SYS_MOUNT privilege.
*/
int
secpolicy_zfs(const cred_t *cr)
{
return (priv_policy(cr, CAP_SYS_ADMIN, EACCES));
}
/*
* Equivalent to secpolicy_zfs(), but works even if the cred_t is not that of
* the current process. Takes both cred_t and proc_t so that this can work
* easily on all platforms.
*
* The has_capability() function was first exported in the 4.10 Linux kernel
* then backported to some LTS kernels. Prior to this change there was no
* mechanism to perform this check therefore EACCES is returned when the
* functionality is not present in the kernel.
*/
int
secpolicy_zfs_proc(const cred_t *cr, proc_t *proc)
{
#if defined(HAVE_HAS_CAPABILITY)
if (!has_capability(proc, CAP_SYS_ADMIN))
return (EACCES);
return (0);
#else
return (EACCES);
#endif
}
void
secpolicy_setid_clear(vattr_t *vap, cred_t *cr)
{
if ((vap->va_mode & (S_ISUID | S_ISGID)) != 0 &&
secpolicy_vnode_setid_retain(NULL, cr,
(vap->va_mode & S_ISUID) != 0 &&
(vap->va_mask & AT_UID) != 0 && vap->va_uid == 0) != 0) {
vap->va_mask |= AT_MODE;
vap->va_mode &= ~(S_ISUID|S_ISGID);
}
}
/*
* Determine that subject can set the file setid flags.
*/
static int
-secpolicy_vnode_setid_modify(const cred_t *cr, uid_t owner, zuserns_t *mnt_ns,
- zuserns_t *fs_ns)
+secpolicy_vnode_setid_modify(const cred_t *cr, uid_t owner, zidmap_t *mnt_ns,
+ struct user_namespace *fs_ns)
{
owner = zfs_uid_to_vfsuid(mnt_ns, fs_ns, owner);
if (crgetuid(cr) == owner)
return (0);
#if defined(CONFIG_USER_NS)
if (!kuid_has_mapping(cr->user_ns, SUID_TO_KUID(owner)))
return (EPERM);
#endif
return (priv_policy_user(cr, CAP_FSETID, EPERM));
}
/*
* Determine that subject can make a file a "sticky".
*
* Enforced in the Linux VFS.
*/
static int
secpolicy_vnode_stky_modify(const cred_t *cr)
{
return (0);
}
int
secpolicy_setid_setsticky_clear(struct inode *ip, vattr_t *vap,
- const vattr_t *ovap, cred_t *cr, zuserns_t *mnt_ns, zuserns_t *fs_ns)
+ const vattr_t *ovap, cred_t *cr, zidmap_t *mnt_ns,
+ struct user_namespace *fs_ns)
{
int error;
if ((vap->va_mode & S_ISUID) != 0 &&
(error = secpolicy_vnode_setid_modify(cr,
ovap->va_uid, mnt_ns, fs_ns)) != 0) {
return (error);
}
/*
* Check privilege if attempting to set the
* sticky bit on a non-directory.
*/
if (!S_ISDIR(ip->i_mode) && (vap->va_mode & S_ISVTX) != 0 &&
secpolicy_vnode_stky_modify(cr) != 0) {
vap->va_mode &= ~S_ISVTX;
}
/*
* Check for privilege if attempting to set the
* group-id bit.
*/
if ((vap->va_mode & S_ISGID) != 0 &&
secpolicy_vnode_setids_setgids(cr, ovap->va_gid,
mnt_ns, fs_ns) != 0) {
vap->va_mode &= ~S_ISGID;
}
return (0);
}
/*
* Check privileges for setting xvattr attributes
*/
int
secpolicy_xvattr(xvattr_t *xvap, uid_t owner, cred_t *cr, mode_t type)
{
return (secpolicy_vnode_chown(cr, owner));
}
/*
* Check privileges for setattr attributes.
*
* Enforced in the Linux VFS.
*/
int
secpolicy_vnode_setattr(cred_t *cr, struct inode *ip, struct vattr *vap,
const struct vattr *ovap, int flags,
int unlocked_access(void *, int, cred_t *), void *node)
{
return (0);
}
/*
* Check privileges for links.
*
* Enforced in the Linux VFS.
*/
int
secpolicy_basic_link(const cred_t *cr)
{
return (0);
}
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c
index db1bb9577197..ff26f47f2e04 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_acl.c
@@ -1,3038 +1,3036 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013 by Delphix. All rights reserved.
* Copyright 2014 Nexenta Systems, Inc. All rights reserved.
*/
#include <sys/types.h>
#include <sys/param.h>
#include <sys/time.h>
#include <sys/sysmacros.h>
#include <sys/vfs.h>
#include <sys/vnode.h>
#include <sys/sid.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/kmem.h>
#include <sys/cmn_err.h>
#include <sys/errno.h>
#include <sys/fs/zfs.h>
#include <sys/policy.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_fuid.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_quota.h>
#include <sys/zfs_vfsops.h>
#include <sys/dmu.h>
#include <sys/dnode.h>
#include <sys/zap.h>
#include <sys/sa.h>
#include <sys/trace_acl.h>
#include <sys/zpl.h>
#define ALLOW ACE_ACCESS_ALLOWED_ACE_TYPE
#define DENY ACE_ACCESS_DENIED_ACE_TYPE
#define MAX_ACE_TYPE ACE_SYSTEM_ALARM_CALLBACK_OBJECT_ACE_TYPE
#define MIN_ACE_TYPE ALLOW
#define OWNING_GROUP (ACE_GROUP|ACE_IDENTIFIER_GROUP)
#define EVERYONE_ALLOW_MASK (ACE_READ_ACL|ACE_READ_ATTRIBUTES | \
ACE_READ_NAMED_ATTRS|ACE_SYNCHRONIZE)
#define EVERYONE_DENY_MASK (ACE_WRITE_ACL|ACE_WRITE_OWNER | \
ACE_WRITE_ATTRIBUTES|ACE_WRITE_NAMED_ATTRS)
#define OWNER_ALLOW_MASK (ACE_WRITE_ACL | ACE_WRITE_OWNER | \
ACE_WRITE_ATTRIBUTES|ACE_WRITE_NAMED_ATTRS)
#define ZFS_CHECKED_MASKS (ACE_READ_ACL|ACE_READ_ATTRIBUTES|ACE_READ_DATA| \
ACE_READ_NAMED_ATTRS|ACE_WRITE_DATA|ACE_WRITE_ATTRIBUTES| \
ACE_WRITE_NAMED_ATTRS|ACE_APPEND_DATA|ACE_EXECUTE|ACE_WRITE_OWNER| \
ACE_WRITE_ACL|ACE_DELETE|ACE_DELETE_CHILD|ACE_SYNCHRONIZE)
#define WRITE_MASK_DATA (ACE_WRITE_DATA|ACE_APPEND_DATA|ACE_WRITE_NAMED_ATTRS)
#define WRITE_MASK_ATTRS (ACE_WRITE_ACL|ACE_WRITE_OWNER|ACE_WRITE_ATTRIBUTES| \
ACE_DELETE|ACE_DELETE_CHILD)
#define WRITE_MASK (WRITE_MASK_DATA|WRITE_MASK_ATTRS)
#define OGE_CLEAR (ACE_READ_DATA|ACE_LIST_DIRECTORY|ACE_WRITE_DATA| \
ACE_ADD_FILE|ACE_APPEND_DATA|ACE_ADD_SUBDIRECTORY|ACE_EXECUTE)
#define OKAY_MASK_BITS (ACE_READ_DATA|ACE_LIST_DIRECTORY|ACE_WRITE_DATA| \
ACE_ADD_FILE|ACE_APPEND_DATA|ACE_ADD_SUBDIRECTORY|ACE_EXECUTE)
#define ALL_INHERIT (ACE_FILE_INHERIT_ACE|ACE_DIRECTORY_INHERIT_ACE | \
ACE_NO_PROPAGATE_INHERIT_ACE|ACE_INHERIT_ONLY_ACE|ACE_INHERITED_ACE)
#define RESTRICTED_CLEAR (ACE_WRITE_ACL|ACE_WRITE_OWNER)
#define V4_ACL_WIDE_FLAGS (ZFS_ACL_AUTO_INHERIT|ZFS_ACL_DEFAULTED|\
ZFS_ACL_PROTECTED)
#define ZFS_ACL_WIDE_FLAGS (V4_ACL_WIDE_FLAGS|ZFS_ACL_TRIVIAL|ZFS_INHERIT_ACE|\
ZFS_ACL_OBJ_ACE)
#define ALL_MODE_EXECS (S_IXUSR | S_IXGRP | S_IXOTH)
#define IDMAP_WK_CREATOR_OWNER_UID 2147483648U
static uint16_t
zfs_ace_v0_get_type(void *acep)
{
return (((zfs_oldace_t *)acep)->z_type);
}
static uint16_t
zfs_ace_v0_get_flags(void *acep)
{
return (((zfs_oldace_t *)acep)->z_flags);
}
static uint32_t
zfs_ace_v0_get_mask(void *acep)
{
return (((zfs_oldace_t *)acep)->z_access_mask);
}
static uint64_t
zfs_ace_v0_get_who(void *acep)
{
return (((zfs_oldace_t *)acep)->z_fuid);
}
static void
zfs_ace_v0_set_type(void *acep, uint16_t type)
{
((zfs_oldace_t *)acep)->z_type = type;
}
static void
zfs_ace_v0_set_flags(void *acep, uint16_t flags)
{
((zfs_oldace_t *)acep)->z_flags = flags;
}
static void
zfs_ace_v0_set_mask(void *acep, uint32_t mask)
{
((zfs_oldace_t *)acep)->z_access_mask = mask;
}
static void
zfs_ace_v0_set_who(void *acep, uint64_t who)
{
((zfs_oldace_t *)acep)->z_fuid = who;
}
static size_t
zfs_ace_v0_size(void *acep)
{
(void) acep;
return (sizeof (zfs_oldace_t));
}
static size_t
zfs_ace_v0_abstract_size(void)
{
return (sizeof (zfs_oldace_t));
}
static int
zfs_ace_v0_mask_off(void)
{
return (offsetof(zfs_oldace_t, z_access_mask));
}
static int
zfs_ace_v0_data(void *acep, void **datap)
{
(void) acep;
*datap = NULL;
return (0);
}
static const acl_ops_t zfs_acl_v0_ops = {
.ace_mask_get = zfs_ace_v0_get_mask,
.ace_mask_set = zfs_ace_v0_set_mask,
.ace_flags_get = zfs_ace_v0_get_flags,
.ace_flags_set = zfs_ace_v0_set_flags,
.ace_type_get = zfs_ace_v0_get_type,
.ace_type_set = zfs_ace_v0_set_type,
.ace_who_get = zfs_ace_v0_get_who,
.ace_who_set = zfs_ace_v0_set_who,
.ace_size = zfs_ace_v0_size,
.ace_abstract_size = zfs_ace_v0_abstract_size,
.ace_mask_off = zfs_ace_v0_mask_off,
.ace_data = zfs_ace_v0_data
};
static uint16_t
zfs_ace_fuid_get_type(void *acep)
{
return (((zfs_ace_hdr_t *)acep)->z_type);
}
static uint16_t
zfs_ace_fuid_get_flags(void *acep)
{
return (((zfs_ace_hdr_t *)acep)->z_flags);
}
static uint32_t
zfs_ace_fuid_get_mask(void *acep)
{
return (((zfs_ace_hdr_t *)acep)->z_access_mask);
}
static uint64_t
zfs_ace_fuid_get_who(void *args)
{
uint16_t entry_type;
zfs_ace_t *acep = args;
entry_type = acep->z_hdr.z_flags & ACE_TYPE_FLAGS;
if (entry_type == ACE_OWNER || entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE)
return (-1);
return (((zfs_ace_t *)acep)->z_fuid);
}
static void
zfs_ace_fuid_set_type(void *acep, uint16_t type)
{
((zfs_ace_hdr_t *)acep)->z_type = type;
}
static void
zfs_ace_fuid_set_flags(void *acep, uint16_t flags)
{
((zfs_ace_hdr_t *)acep)->z_flags = flags;
}
static void
zfs_ace_fuid_set_mask(void *acep, uint32_t mask)
{
((zfs_ace_hdr_t *)acep)->z_access_mask = mask;
}
static void
zfs_ace_fuid_set_who(void *arg, uint64_t who)
{
zfs_ace_t *acep = arg;
uint16_t entry_type = acep->z_hdr.z_flags & ACE_TYPE_FLAGS;
if (entry_type == ACE_OWNER || entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE)
return;
acep->z_fuid = who;
}
static size_t
zfs_ace_fuid_size(void *acep)
{
zfs_ace_hdr_t *zacep = acep;
uint16_t entry_type;
switch (zacep->z_type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
return (sizeof (zfs_object_ace_t));
case ALLOW:
case DENY:
entry_type =
(((zfs_ace_hdr_t *)acep)->z_flags & ACE_TYPE_FLAGS);
if (entry_type == ACE_OWNER ||
entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE)
return (sizeof (zfs_ace_hdr_t));
zfs_fallthrough;
default:
return (sizeof (zfs_ace_t));
}
}
static size_t
zfs_ace_fuid_abstract_size(void)
{
return (sizeof (zfs_ace_hdr_t));
}
static int
zfs_ace_fuid_mask_off(void)
{
return (offsetof(zfs_ace_hdr_t, z_access_mask));
}
static int
zfs_ace_fuid_data(void *acep, void **datap)
{
zfs_ace_t *zacep = acep;
zfs_object_ace_t *zobjp;
switch (zacep->z_hdr.z_type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
zobjp = acep;
*datap = (caddr_t)zobjp + sizeof (zfs_ace_t);
return (sizeof (zfs_object_ace_t) - sizeof (zfs_ace_t));
default:
*datap = NULL;
return (0);
}
}
static const acl_ops_t zfs_acl_fuid_ops = {
.ace_mask_get = zfs_ace_fuid_get_mask,
.ace_mask_set = zfs_ace_fuid_set_mask,
.ace_flags_get = zfs_ace_fuid_get_flags,
.ace_flags_set = zfs_ace_fuid_set_flags,
.ace_type_get = zfs_ace_fuid_get_type,
.ace_type_set = zfs_ace_fuid_set_type,
.ace_who_get = zfs_ace_fuid_get_who,
.ace_who_set = zfs_ace_fuid_set_who,
.ace_size = zfs_ace_fuid_size,
.ace_abstract_size = zfs_ace_fuid_abstract_size,
.ace_mask_off = zfs_ace_fuid_mask_off,
.ace_data = zfs_ace_fuid_data
};
/*
* The following three functions are provided for compatibility with
* older ZPL version in order to determine if the file use to have
* an external ACL and what version of ACL previously existed on the
* file. Would really be nice to not need this, sigh.
*/
uint64_t
zfs_external_acl(znode_t *zp)
{
zfs_acl_phys_t acl_phys;
int error;
if (zp->z_is_sa)
return (0);
/*
* Need to deal with a potential
* race where zfs_sa_upgrade could cause
* z_isa_sa to change.
*
* If the lookup fails then the state of z_is_sa should have
* changed.
*/
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_ZNODE_ACL(ZTOZSB(zp)),
&acl_phys, sizeof (acl_phys))) == 0)
return (acl_phys.z_acl_extern_obj);
else {
/*
* after upgrade the SA_ZPL_ZNODE_ACL should have been
* removed
*/
VERIFY(zp->z_is_sa && error == ENOENT);
return (0);
}
}
/*
* Determine size of ACL in bytes
*
* This is more complicated than it should be since we have to deal
* with old external ACLs.
*/
static int
zfs_acl_znode_info(znode_t *zp, int *aclsize, int *aclcount,
zfs_acl_phys_t *aclphys)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
uint64_t acl_count;
int size;
int error;
ASSERT(MUTEX_HELD(&zp->z_acl_lock));
if (zp->z_is_sa) {
if ((error = sa_size(zp->z_sa_hdl, SA_ZPL_DACL_ACES(zfsvfs),
&size)) != 0)
return (error);
*aclsize = size;
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_DACL_COUNT(zfsvfs),
&acl_count, sizeof (acl_count))) != 0)
return (error);
*aclcount = acl_count;
} else {
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_ZNODE_ACL(zfsvfs),
aclphys, sizeof (*aclphys))) != 0)
return (error);
if (aclphys->z_acl_version == ZFS_ACL_VERSION_INITIAL) {
*aclsize = ZFS_ACL_SIZE(aclphys->z_acl_size);
*aclcount = aclphys->z_acl_size;
} else {
*aclsize = aclphys->z_acl_size;
*aclcount = aclphys->z_acl_count;
}
}
return (0);
}
int
zfs_znode_acl_version(znode_t *zp)
{
zfs_acl_phys_t acl_phys;
if (zp->z_is_sa)
return (ZFS_ACL_VERSION_FUID);
else {
int error;
/*
* Need to deal with a potential
* race where zfs_sa_upgrade could cause
* z_isa_sa to change.
*
* If the lookup fails then the state of z_is_sa should have
* changed.
*/
if ((error = sa_lookup(zp->z_sa_hdl,
SA_ZPL_ZNODE_ACL(ZTOZSB(zp)),
&acl_phys, sizeof (acl_phys))) == 0)
return (acl_phys.z_acl_version);
else {
/*
* After upgrade SA_ZPL_ZNODE_ACL should have
* been removed.
*/
VERIFY(zp->z_is_sa && error == ENOENT);
return (ZFS_ACL_VERSION_FUID);
}
}
}
static int
zfs_acl_version(int version)
{
if (version < ZPL_VERSION_FUID)
return (ZFS_ACL_VERSION_INITIAL);
else
return (ZFS_ACL_VERSION_FUID);
}
static int
zfs_acl_version_zp(znode_t *zp)
{
return (zfs_acl_version(ZTOZSB(zp)->z_version));
}
zfs_acl_t *
zfs_acl_alloc(int vers)
{
zfs_acl_t *aclp;
aclp = kmem_zalloc(sizeof (zfs_acl_t), KM_SLEEP);
list_create(&aclp->z_acl, sizeof (zfs_acl_node_t),
offsetof(zfs_acl_node_t, z_next));
aclp->z_version = vers;
if (vers == ZFS_ACL_VERSION_FUID)
aclp->z_ops = &zfs_acl_fuid_ops;
else
aclp->z_ops = &zfs_acl_v0_ops;
return (aclp);
}
zfs_acl_node_t *
zfs_acl_node_alloc(size_t bytes)
{
zfs_acl_node_t *aclnode;
aclnode = kmem_zalloc(sizeof (zfs_acl_node_t), KM_SLEEP);
if (bytes) {
aclnode->z_acldata = kmem_alloc(bytes, KM_SLEEP);
aclnode->z_allocdata = aclnode->z_acldata;
aclnode->z_allocsize = bytes;
aclnode->z_size = bytes;
}
return (aclnode);
}
static void
zfs_acl_node_free(zfs_acl_node_t *aclnode)
{
if (aclnode->z_allocsize)
kmem_free(aclnode->z_allocdata, aclnode->z_allocsize);
kmem_free(aclnode, sizeof (zfs_acl_node_t));
}
static void
zfs_acl_release_nodes(zfs_acl_t *aclp)
{
zfs_acl_node_t *aclnode;
while ((aclnode = list_head(&aclp->z_acl))) {
list_remove(&aclp->z_acl, aclnode);
zfs_acl_node_free(aclnode);
}
aclp->z_acl_count = 0;
aclp->z_acl_bytes = 0;
}
void
zfs_acl_free(zfs_acl_t *aclp)
{
zfs_acl_release_nodes(aclp);
list_destroy(&aclp->z_acl);
kmem_free(aclp, sizeof (zfs_acl_t));
}
static boolean_t
zfs_acl_valid_ace_type(uint_t type, uint_t flags)
{
uint16_t entry_type;
switch (type) {
case ALLOW:
case DENY:
case ACE_SYSTEM_AUDIT_ACE_TYPE:
case ACE_SYSTEM_ALARM_ACE_TYPE:
entry_type = flags & ACE_TYPE_FLAGS;
return (entry_type == ACE_OWNER ||
entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE || entry_type == 0 ||
entry_type == ACE_IDENTIFIER_GROUP);
default:
if (type <= MAX_ACE_TYPE)
return (B_TRUE);
}
return (B_FALSE);
}
static boolean_t
zfs_ace_valid(umode_t obj_mode, zfs_acl_t *aclp, uint16_t type, uint16_t iflags)
{
/*
* first check type of entry
*/
if (!zfs_acl_valid_ace_type(type, iflags))
return (B_FALSE);
switch (type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
if (aclp->z_version < ZFS_ACL_VERSION_FUID)
return (B_FALSE);
aclp->z_hints |= ZFS_ACL_OBJ_ACE;
}
/*
* next check inheritance level flags
*/
if (S_ISDIR(obj_mode) &&
(iflags & (ACE_FILE_INHERIT_ACE|ACE_DIRECTORY_INHERIT_ACE)))
aclp->z_hints |= ZFS_INHERIT_ACE;
if (iflags & (ACE_INHERIT_ONLY_ACE|ACE_NO_PROPAGATE_INHERIT_ACE)) {
if ((iflags & (ACE_FILE_INHERIT_ACE|
ACE_DIRECTORY_INHERIT_ACE)) == 0) {
return (B_FALSE);
}
}
return (B_TRUE);
}
static void *
zfs_acl_next_ace(zfs_acl_t *aclp, void *start, uint64_t *who,
uint32_t *access_mask, uint16_t *iflags, uint16_t *type)
{
zfs_acl_node_t *aclnode;
ASSERT(aclp);
if (start == NULL) {
aclnode = list_head(&aclp->z_acl);
if (aclnode == NULL)
return (NULL);
aclp->z_next_ace = aclnode->z_acldata;
aclp->z_curr_node = aclnode;
aclnode->z_ace_idx = 0;
}
aclnode = aclp->z_curr_node;
if (aclnode == NULL)
return (NULL);
if (aclnode->z_ace_idx >= aclnode->z_ace_count) {
aclnode = list_next(&aclp->z_acl, aclnode);
if (aclnode == NULL)
return (NULL);
else {
aclp->z_curr_node = aclnode;
aclnode->z_ace_idx = 0;
aclp->z_next_ace = aclnode->z_acldata;
}
}
if (aclnode->z_ace_idx < aclnode->z_ace_count) {
void *acep = aclp->z_next_ace;
size_t ace_size;
/*
* Make sure we don't overstep our bounds
*/
ace_size = aclp->z_ops->ace_size(acep);
if (((caddr_t)acep + ace_size) >
((caddr_t)aclnode->z_acldata + aclnode->z_size)) {
return (NULL);
}
*iflags = aclp->z_ops->ace_flags_get(acep);
*type = aclp->z_ops->ace_type_get(acep);
*access_mask = aclp->z_ops->ace_mask_get(acep);
*who = aclp->z_ops->ace_who_get(acep);
aclp->z_next_ace = (caddr_t)aclp->z_next_ace + ace_size;
aclnode->z_ace_idx++;
return ((void *)acep);
}
return (NULL);
}
static uintptr_t
zfs_ace_walk(void *datap, uintptr_t cookie, int aclcnt,
uint16_t *flags, uint16_t *type, uint32_t *mask)
{
(void) aclcnt;
zfs_acl_t *aclp = datap;
zfs_ace_hdr_t *acep = (zfs_ace_hdr_t *)cookie;
uint64_t who;
acep = zfs_acl_next_ace(aclp, acep, &who, mask,
flags, type);
return ((uintptr_t)acep);
}
/*
* Copy ACE to internal ZFS format.
* While processing the ACL each ACE will be validated for correctness.
* ACE FUIDs will be created later.
*/
static int
zfs_copy_ace_2_fuid(zfsvfs_t *zfsvfs, umode_t obj_mode, zfs_acl_t *aclp,
void *datap, zfs_ace_t *z_acl, uint64_t aclcnt, size_t *size,
zfs_fuid_info_t **fuidp, cred_t *cr)
{
int i;
uint16_t entry_type;
zfs_ace_t *aceptr = z_acl;
ace_t *acep = datap;
zfs_object_ace_t *zobjacep;
ace_object_t *aceobjp;
for (i = 0; i != aclcnt; i++) {
aceptr->z_hdr.z_access_mask = acep->a_access_mask;
aceptr->z_hdr.z_flags = acep->a_flags;
aceptr->z_hdr.z_type = acep->a_type;
entry_type = aceptr->z_hdr.z_flags & ACE_TYPE_FLAGS;
if (entry_type != ACE_OWNER && entry_type != OWNING_GROUP &&
entry_type != ACE_EVERYONE) {
aceptr->z_fuid = zfs_fuid_create(zfsvfs, acep->a_who,
cr, (entry_type == 0) ?
ZFS_ACE_USER : ZFS_ACE_GROUP, fuidp);
}
/*
* Make sure ACE is valid
*/
if (zfs_ace_valid(obj_mode, aclp, aceptr->z_hdr.z_type,
aceptr->z_hdr.z_flags) != B_TRUE)
return (SET_ERROR(EINVAL));
switch (acep->a_type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
zobjacep = (zfs_object_ace_t *)aceptr;
aceobjp = (ace_object_t *)acep;
memcpy(zobjacep->z_object_type, aceobjp->a_obj_type,
sizeof (aceobjp->a_obj_type));
memcpy(zobjacep->z_inherit_type,
aceobjp->a_inherit_obj_type,
sizeof (aceobjp->a_inherit_obj_type));
acep = (ace_t *)((caddr_t)acep + sizeof (ace_object_t));
break;
default:
acep = (ace_t *)((caddr_t)acep + sizeof (ace_t));
}
aceptr = (zfs_ace_t *)((caddr_t)aceptr +
aclp->z_ops->ace_size(aceptr));
}
*size = (caddr_t)aceptr - (caddr_t)z_acl;
return (0);
}
/*
* Copy ZFS ACEs to fixed size ace_t layout
*/
static void
zfs_copy_fuid_2_ace(zfsvfs_t *zfsvfs, zfs_acl_t *aclp, cred_t *cr,
void *datap, int filter)
{
uint64_t who;
uint32_t access_mask;
uint16_t iflags, type;
zfs_ace_hdr_t *zacep = NULL;
ace_t *acep = datap;
ace_object_t *objacep;
zfs_object_ace_t *zobjacep;
size_t ace_size;
uint16_t entry_type;
while ((zacep = zfs_acl_next_ace(aclp, zacep,
&who, &access_mask, &iflags, &type))) {
switch (type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
if (filter) {
continue;
}
zobjacep = (zfs_object_ace_t *)zacep;
objacep = (ace_object_t *)acep;
memcpy(objacep->a_obj_type,
zobjacep->z_object_type,
sizeof (zobjacep->z_object_type));
memcpy(objacep->a_inherit_obj_type,
zobjacep->z_inherit_type,
sizeof (zobjacep->z_inherit_type));
ace_size = sizeof (ace_object_t);
break;
default:
ace_size = sizeof (ace_t);
break;
}
entry_type = (iflags & ACE_TYPE_FLAGS);
if ((entry_type != ACE_OWNER &&
entry_type != OWNING_GROUP &&
entry_type != ACE_EVERYONE)) {
acep->a_who = zfs_fuid_map_id(zfsvfs, who,
cr, (entry_type & ACE_IDENTIFIER_GROUP) ?
ZFS_ACE_GROUP : ZFS_ACE_USER);
} else {
acep->a_who = (uid_t)(int64_t)who;
}
acep->a_access_mask = access_mask;
acep->a_flags = iflags;
acep->a_type = type;
acep = (ace_t *)((caddr_t)acep + ace_size);
}
}
static int
zfs_copy_ace_2_oldace(umode_t obj_mode, zfs_acl_t *aclp, ace_t *acep,
zfs_oldace_t *z_acl, int aclcnt, size_t *size)
{
int i;
zfs_oldace_t *aceptr = z_acl;
for (i = 0; i != aclcnt; i++, aceptr++) {
aceptr->z_access_mask = acep[i].a_access_mask;
aceptr->z_type = acep[i].a_type;
aceptr->z_flags = acep[i].a_flags;
aceptr->z_fuid = acep[i].a_who;
/*
* Make sure ACE is valid
*/
if (zfs_ace_valid(obj_mode, aclp, aceptr->z_type,
aceptr->z_flags) != B_TRUE)
return (SET_ERROR(EINVAL));
}
*size = (caddr_t)aceptr - (caddr_t)z_acl;
return (0);
}
/*
* convert old ACL format to new
*/
void
zfs_acl_xform(znode_t *zp, zfs_acl_t *aclp, cred_t *cr)
{
zfs_oldace_t *oldaclp;
int i;
uint16_t type, iflags;
uint32_t access_mask;
uint64_t who;
void *cookie = NULL;
zfs_acl_node_t *newaclnode;
ASSERT(aclp->z_version == ZFS_ACL_VERSION_INITIAL);
/*
* First create the ACE in a contiguous piece of memory
* for zfs_copy_ace_2_fuid().
*
* We only convert an ACL once, so this won't happen
* every time.
*/
oldaclp = kmem_alloc(sizeof (zfs_oldace_t) * aclp->z_acl_count,
KM_SLEEP);
i = 0;
while ((cookie = zfs_acl_next_ace(aclp, cookie, &who,
&access_mask, &iflags, &type))) {
oldaclp[i].z_flags = iflags;
oldaclp[i].z_type = type;
oldaclp[i].z_fuid = who;
oldaclp[i++].z_access_mask = access_mask;
}
newaclnode = zfs_acl_node_alloc(aclp->z_acl_count *
sizeof (zfs_object_ace_t));
aclp->z_ops = &zfs_acl_fuid_ops;
VERIFY(zfs_copy_ace_2_fuid(ZTOZSB(zp), ZTOI(zp)->i_mode,
aclp, oldaclp, newaclnode->z_acldata, aclp->z_acl_count,
&newaclnode->z_size, NULL, cr) == 0);
newaclnode->z_ace_count = aclp->z_acl_count;
aclp->z_version = ZFS_ACL_VERSION;
kmem_free(oldaclp, aclp->z_acl_count * sizeof (zfs_oldace_t));
/*
* Release all previous ACL nodes
*/
zfs_acl_release_nodes(aclp);
list_insert_head(&aclp->z_acl, newaclnode);
aclp->z_acl_bytes = newaclnode->z_size;
aclp->z_acl_count = newaclnode->z_ace_count;
}
/*
* Convert unix access mask to v4 access mask
*/
static uint32_t
zfs_unix_to_v4(uint32_t access_mask)
{
uint32_t new_mask = 0;
if (access_mask & S_IXOTH)
new_mask |= ACE_EXECUTE;
if (access_mask & S_IWOTH)
new_mask |= ACE_WRITE_DATA;
if (access_mask & S_IROTH)
new_mask |= ACE_READ_DATA;
return (new_mask);
}
static int
zfs_v4_to_unix(uint32_t access_mask, int *unmapped)
{
int new_mask = 0;
*unmapped = access_mask &
(ACE_WRITE_OWNER | ACE_WRITE_ACL | ACE_DELETE);
if (access_mask & WRITE_MASK)
new_mask |= S_IWOTH;
if (access_mask & ACE_READ_DATA)
new_mask |= S_IROTH;
if (access_mask & ACE_EXECUTE)
new_mask |= S_IXOTH;
return (new_mask);
}
static void
zfs_set_ace(zfs_acl_t *aclp, void *acep, uint32_t access_mask,
uint16_t access_type, uint64_t fuid, uint16_t entry_type)
{
uint16_t type = entry_type & ACE_TYPE_FLAGS;
aclp->z_ops->ace_mask_set(acep, access_mask);
aclp->z_ops->ace_type_set(acep, access_type);
aclp->z_ops->ace_flags_set(acep, entry_type);
if ((type != ACE_OWNER && type != OWNING_GROUP &&
type != ACE_EVERYONE))
aclp->z_ops->ace_who_set(acep, fuid);
}
/*
* Determine mode of file based on ACL.
*/
uint64_t
zfs_mode_compute(uint64_t fmode, zfs_acl_t *aclp,
uint64_t *pflags, uint64_t fuid, uint64_t fgid)
{
int entry_type;
mode_t mode;
mode_t seen = 0;
zfs_ace_hdr_t *acep = NULL;
uint64_t who;
uint16_t iflags, type;
uint32_t access_mask;
boolean_t an_exec_denied = B_FALSE;
mode = (fmode & (S_IFMT | S_ISUID | S_ISGID | S_ISVTX));
while ((acep = zfs_acl_next_ace(aclp, acep, &who,
&access_mask, &iflags, &type))) {
if (!zfs_acl_valid_ace_type(type, iflags))
continue;
entry_type = (iflags & ACE_TYPE_FLAGS);
/*
* Skip over any inherit_only ACEs
*/
if (iflags & ACE_INHERIT_ONLY_ACE)
continue;
if (entry_type == ACE_OWNER || (entry_type == 0 &&
who == fuid)) {
if ((access_mask & ACE_READ_DATA) &&
(!(seen & S_IRUSR))) {
seen |= S_IRUSR;
if (type == ALLOW) {
mode |= S_IRUSR;
}
}
if ((access_mask & ACE_WRITE_DATA) &&
(!(seen & S_IWUSR))) {
seen |= S_IWUSR;
if (type == ALLOW) {
mode |= S_IWUSR;
}
}
if ((access_mask & ACE_EXECUTE) &&
(!(seen & S_IXUSR))) {
seen |= S_IXUSR;
if (type == ALLOW) {
mode |= S_IXUSR;
}
}
} else if (entry_type == OWNING_GROUP ||
(entry_type == ACE_IDENTIFIER_GROUP && who == fgid)) {
if ((access_mask & ACE_READ_DATA) &&
(!(seen & S_IRGRP))) {
seen |= S_IRGRP;
if (type == ALLOW) {
mode |= S_IRGRP;
}
}
if ((access_mask & ACE_WRITE_DATA) &&
(!(seen & S_IWGRP))) {
seen |= S_IWGRP;
if (type == ALLOW) {
mode |= S_IWGRP;
}
}
if ((access_mask & ACE_EXECUTE) &&
(!(seen & S_IXGRP))) {
seen |= S_IXGRP;
if (type == ALLOW) {
mode |= S_IXGRP;
}
}
} else if (entry_type == ACE_EVERYONE) {
if ((access_mask & ACE_READ_DATA)) {
if (!(seen & S_IRUSR)) {
seen |= S_IRUSR;
if (type == ALLOW) {
mode |= S_IRUSR;
}
}
if (!(seen & S_IRGRP)) {
seen |= S_IRGRP;
if (type == ALLOW) {
mode |= S_IRGRP;
}
}
if (!(seen & S_IROTH)) {
seen |= S_IROTH;
if (type == ALLOW) {
mode |= S_IROTH;
}
}
}
if ((access_mask & ACE_WRITE_DATA)) {
if (!(seen & S_IWUSR)) {
seen |= S_IWUSR;
if (type == ALLOW) {
mode |= S_IWUSR;
}
}
if (!(seen & S_IWGRP)) {
seen |= S_IWGRP;
if (type == ALLOW) {
mode |= S_IWGRP;
}
}
if (!(seen & S_IWOTH)) {
seen |= S_IWOTH;
if (type == ALLOW) {
mode |= S_IWOTH;
}
}
}
if ((access_mask & ACE_EXECUTE)) {
if (!(seen & S_IXUSR)) {
seen |= S_IXUSR;
if (type == ALLOW) {
mode |= S_IXUSR;
}
}
if (!(seen & S_IXGRP)) {
seen |= S_IXGRP;
if (type == ALLOW) {
mode |= S_IXGRP;
}
}
if (!(seen & S_IXOTH)) {
seen |= S_IXOTH;
if (type == ALLOW) {
mode |= S_IXOTH;
}
}
}
} else {
/*
* Only care if this IDENTIFIER_GROUP or
* USER ACE denies execute access to someone,
* mode is not affected
*/
if ((access_mask & ACE_EXECUTE) && type == DENY)
an_exec_denied = B_TRUE;
}
}
/*
* Failure to allow is effectively a deny, so execute permission
* is denied if it was never mentioned or if we explicitly
* weren't allowed it.
*/
if (!an_exec_denied &&
((seen & ALL_MODE_EXECS) != ALL_MODE_EXECS ||
(mode & ALL_MODE_EXECS) != ALL_MODE_EXECS))
an_exec_denied = B_TRUE;
if (an_exec_denied)
*pflags &= ~ZFS_NO_EXECS_DENIED;
else
*pflags |= ZFS_NO_EXECS_DENIED;
return (mode);
}
/*
* Read an external acl object. If the intent is to modify, always
* create a new acl and leave any cached acl in place.
*/
int
zfs_acl_node_read(struct znode *zp, boolean_t have_lock, zfs_acl_t **aclpp,
boolean_t will_modify)
{
zfs_acl_t *aclp;
int aclsize = 0;
int acl_count = 0;
zfs_acl_node_t *aclnode;
zfs_acl_phys_t znode_acl;
int version;
int error;
boolean_t drop_lock = B_FALSE;
ASSERT(MUTEX_HELD(&zp->z_acl_lock));
if (zp->z_acl_cached && !will_modify) {
*aclpp = zp->z_acl_cached;
return (0);
}
/*
* close race where znode could be upgrade while trying to
* read the znode attributes.
*
* But this could only happen if the file isn't already an SA
* znode
*/
if (!zp->z_is_sa && !have_lock) {
mutex_enter(&zp->z_lock);
drop_lock = B_TRUE;
}
version = zfs_znode_acl_version(zp);
if ((error = zfs_acl_znode_info(zp, &aclsize,
&acl_count, &znode_acl)) != 0) {
goto done;
}
aclp = zfs_acl_alloc(version);
aclp->z_acl_count = acl_count;
aclp->z_acl_bytes = aclsize;
aclnode = zfs_acl_node_alloc(aclsize);
aclnode->z_ace_count = aclp->z_acl_count;
aclnode->z_size = aclsize;
if (!zp->z_is_sa) {
if (znode_acl.z_acl_extern_obj) {
error = dmu_read(ZTOZSB(zp)->z_os,
znode_acl.z_acl_extern_obj, 0, aclnode->z_size,
aclnode->z_acldata, DMU_READ_PREFETCH);
} else {
memcpy(aclnode->z_acldata, znode_acl.z_ace_data,
aclnode->z_size);
}
} else {
error = sa_lookup(zp->z_sa_hdl, SA_ZPL_DACL_ACES(ZTOZSB(zp)),
aclnode->z_acldata, aclnode->z_size);
}
if (error != 0) {
zfs_acl_free(aclp);
zfs_acl_node_free(aclnode);
/* convert checksum errors into IO errors */
if (error == ECKSUM)
error = SET_ERROR(EIO);
goto done;
}
list_insert_head(&aclp->z_acl, aclnode);
*aclpp = aclp;
if (!will_modify)
zp->z_acl_cached = aclp;
done:
if (drop_lock)
mutex_exit(&zp->z_lock);
return (error);
}
void
zfs_acl_data_locator(void **dataptr, uint32_t *length, uint32_t buflen,
boolean_t start, void *userdata)
{
(void) buflen;
zfs_acl_locator_cb_t *cb = (zfs_acl_locator_cb_t *)userdata;
if (start) {
cb->cb_acl_node = list_head(&cb->cb_aclp->z_acl);
} else {
cb->cb_acl_node = list_next(&cb->cb_aclp->z_acl,
cb->cb_acl_node);
}
ASSERT3P(cb->cb_acl_node, !=, NULL);
*dataptr = cb->cb_acl_node->z_acldata;
*length = cb->cb_acl_node->z_size;
}
int
zfs_acl_chown_setattr(znode_t *zp)
{
int error;
zfs_acl_t *aclp;
if (ZTOZSB(zp)->z_acl_type == ZFS_ACLTYPE_POSIX)
return (0);
ASSERT(MUTEX_HELD(&zp->z_lock));
ASSERT(MUTEX_HELD(&zp->z_acl_lock));
error = zfs_acl_node_read(zp, B_TRUE, &aclp, B_FALSE);
if (error == 0 && aclp->z_acl_count > 0)
zp->z_mode = ZTOI(zp)->i_mode =
zfs_mode_compute(zp->z_mode, aclp,
&zp->z_pflags, KUID_TO_SUID(ZTOI(zp)->i_uid),
KGID_TO_SGID(ZTOI(zp)->i_gid));
/*
* Some ZFS implementations (ZEVO) create neither a ZNODE_ACL
* nor a DACL_ACES SA in which case ENOENT is returned from
* zfs_acl_node_read() when the SA can't be located.
* Allow chown/chgrp to succeed in these cases rather than
* returning an error that makes no sense in the context of
* the caller.
*/
if (error == ENOENT)
return (0);
return (error);
}
typedef struct trivial_acl {
uint32_t allow0; /* allow mask for bits only in owner */
uint32_t deny1; /* deny mask for bits not in owner */
uint32_t deny2; /* deny mask for bits not in group */
uint32_t owner; /* allow mask matching mode */
uint32_t group; /* allow mask matching mode */
uint32_t everyone; /* allow mask matching mode */
} trivial_acl_t;
static void
acl_trivial_access_masks(mode_t mode, boolean_t isdir, trivial_acl_t *masks)
{
uint32_t read_mask = ACE_READ_DATA;
uint32_t write_mask = ACE_WRITE_DATA|ACE_APPEND_DATA;
uint32_t execute_mask = ACE_EXECUTE;
if (isdir)
write_mask |= ACE_DELETE_CHILD;
masks->deny1 = 0;
if (!(mode & S_IRUSR) && (mode & (S_IRGRP|S_IROTH)))
masks->deny1 |= read_mask;
if (!(mode & S_IWUSR) && (mode & (S_IWGRP|S_IWOTH)))
masks->deny1 |= write_mask;
if (!(mode & S_IXUSR) && (mode & (S_IXGRP|S_IXOTH)))
masks->deny1 |= execute_mask;
masks->deny2 = 0;
if (!(mode & S_IRGRP) && (mode & S_IROTH))
masks->deny2 |= read_mask;
if (!(mode & S_IWGRP) && (mode & S_IWOTH))
masks->deny2 |= write_mask;
if (!(mode & S_IXGRP) && (mode & S_IXOTH))
masks->deny2 |= execute_mask;
masks->allow0 = 0;
if ((mode & S_IRUSR) && (!(mode & S_IRGRP) && (mode & S_IROTH)))
masks->allow0 |= read_mask;
if ((mode & S_IWUSR) && (!(mode & S_IWGRP) && (mode & S_IWOTH)))
masks->allow0 |= write_mask;
if ((mode & S_IXUSR) && (!(mode & S_IXGRP) && (mode & S_IXOTH)))
masks->allow0 |= execute_mask;
masks->owner = ACE_WRITE_ATTRIBUTES|ACE_WRITE_OWNER|ACE_WRITE_ACL|
ACE_WRITE_NAMED_ATTRS|ACE_READ_ACL|ACE_READ_ATTRIBUTES|
ACE_READ_NAMED_ATTRS|ACE_SYNCHRONIZE;
if (mode & S_IRUSR)
masks->owner |= read_mask;
if (mode & S_IWUSR)
masks->owner |= write_mask;
if (mode & S_IXUSR)
masks->owner |= execute_mask;
masks->group = ACE_READ_ACL|ACE_READ_ATTRIBUTES|ACE_READ_NAMED_ATTRS|
ACE_SYNCHRONIZE;
if (mode & S_IRGRP)
masks->group |= read_mask;
if (mode & S_IWGRP)
masks->group |= write_mask;
if (mode & S_IXGRP)
masks->group |= execute_mask;
masks->everyone = ACE_READ_ACL|ACE_READ_ATTRIBUTES|ACE_READ_NAMED_ATTRS|
ACE_SYNCHRONIZE;
if (mode & S_IROTH)
masks->everyone |= read_mask;
if (mode & S_IWOTH)
masks->everyone |= write_mask;
if (mode & S_IXOTH)
masks->everyone |= execute_mask;
}
/*
* ace_trivial:
* determine whether an ace_t acl is trivial
*
* Trivialness implies that the acl is composed of only
* owner, group, everyone entries. ACL can't
* have read_acl denied, and write_owner/write_acl/write_attributes
* can only be owner@ entry.
*/
static int
ace_trivial_common(void *acep, int aclcnt,
uintptr_t (*walk)(void *, uintptr_t, int,
uint16_t *, uint16_t *, uint32_t *))
{
uint16_t flags;
uint32_t mask;
uint16_t type;
uint64_t cookie = 0;
while ((cookie = walk(acep, cookie, aclcnt, &flags, &type, &mask))) {
switch (flags & ACE_TYPE_FLAGS) {
case ACE_OWNER:
case ACE_GROUP|ACE_IDENTIFIER_GROUP:
case ACE_EVERYONE:
break;
default:
return (1);
}
if (flags & (ACE_FILE_INHERIT_ACE|
ACE_DIRECTORY_INHERIT_ACE|ACE_NO_PROPAGATE_INHERIT_ACE|
ACE_INHERIT_ONLY_ACE))
return (1);
/*
* Special check for some special bits
*
* Don't allow anybody to deny reading basic
* attributes or a files ACL.
*/
if ((mask & (ACE_READ_ACL|ACE_READ_ATTRIBUTES)) &&
(type == ACE_ACCESS_DENIED_ACE_TYPE))
return (1);
/*
* Delete permission is never set by default
*/
if (mask & ACE_DELETE)
return (1);
/*
* Child delete permission should be accompanied by write
*/
if ((mask & ACE_DELETE_CHILD) && !(mask & ACE_WRITE_DATA))
return (1);
/*
* only allow owner@ to have
* write_acl/write_owner/write_attributes/write_xattr/
*/
if (type == ACE_ACCESS_ALLOWED_ACE_TYPE &&
(!(flags & ACE_OWNER) && (mask &
(ACE_WRITE_OWNER|ACE_WRITE_ACL| ACE_WRITE_ATTRIBUTES|
ACE_WRITE_NAMED_ATTRS))))
return (1);
}
return (0);
}
/*
* common code for setting ACLs.
*
* This function is called from zfs_mode_update, zfs_perm_init, and zfs_setacl.
* zfs_setacl passes a non-NULL inherit pointer (ihp) to indicate that it's
* already checked the acl and knows whether to inherit.
*/
int
zfs_aclset_common(znode_t *zp, zfs_acl_t *aclp, cred_t *cr, dmu_tx_t *tx)
{
int error;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
dmu_object_type_t otype;
zfs_acl_locator_cb_t locate = { 0 };
uint64_t mode;
sa_bulk_attr_t bulk[5];
uint64_t ctime[2];
int count = 0;
zfs_acl_phys_t acl_phys;
mode = zp->z_mode;
mode = zfs_mode_compute(mode, aclp, &zp->z_pflags,
KUID_TO_SUID(ZTOI(zp)->i_uid), KGID_TO_SGID(ZTOI(zp)->i_gid));
zp->z_mode = ZTOI(zp)->i_mode = mode;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL,
&mode, sizeof (mode));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, sizeof (zp->z_pflags));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
&ctime, sizeof (ctime));
if (zp->z_acl_cached) {
zfs_acl_free(zp->z_acl_cached);
zp->z_acl_cached = NULL;
}
/*
* Upgrade needed?
*/
if (!zfsvfs->z_use_fuids) {
otype = DMU_OT_OLDACL;
} else {
if ((aclp->z_version == ZFS_ACL_VERSION_INITIAL) &&
(zfsvfs->z_version >= ZPL_VERSION_FUID))
zfs_acl_xform(zp, aclp, cr);
ASSERT(aclp->z_version >= ZFS_ACL_VERSION_FUID);
otype = DMU_OT_ACL;
}
/*
* Arrgh, we have to handle old on disk format
* as well as newer (preferred) SA format.
*/
if (zp->z_is_sa) { /* the easy case, just update the ACL attribute */
locate.cb_aclp = aclp;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_DACL_ACES(zfsvfs),
zfs_acl_data_locator, &locate, aclp->z_acl_bytes);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_DACL_COUNT(zfsvfs),
NULL, &aclp->z_acl_count, sizeof (uint64_t));
} else { /* Painful legacy way */
zfs_acl_node_t *aclnode;
uint64_t off = 0;
uint64_t aoid;
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_ZNODE_ACL(zfsvfs),
&acl_phys, sizeof (acl_phys))) != 0)
return (error);
aoid = acl_phys.z_acl_extern_obj;
if (aclp->z_acl_bytes > ZFS_ACE_SPACE) {
/*
* If ACL was previously external and we are now
* converting to new ACL format then release old
* ACL object and create a new one.
*/
if (aoid &&
aclp->z_version != acl_phys.z_acl_version) {
error = dmu_object_free(zfsvfs->z_os, aoid, tx);
if (error)
return (error);
aoid = 0;
}
if (aoid == 0) {
aoid = dmu_object_alloc(zfsvfs->z_os,
otype, aclp->z_acl_bytes,
otype == DMU_OT_ACL ?
DMU_OT_SYSACL : DMU_OT_NONE,
otype == DMU_OT_ACL ?
DN_OLD_MAX_BONUSLEN : 0, tx);
} else {
(void) dmu_object_set_blocksize(zfsvfs->z_os,
aoid, aclp->z_acl_bytes, 0, tx);
}
acl_phys.z_acl_extern_obj = aoid;
for (aclnode = list_head(&aclp->z_acl); aclnode;
aclnode = list_next(&aclp->z_acl, aclnode)) {
if (aclnode->z_ace_count == 0)
continue;
dmu_write(zfsvfs->z_os, aoid, off,
aclnode->z_size, aclnode->z_acldata, tx);
off += aclnode->z_size;
}
} else {
void *start = acl_phys.z_ace_data;
/*
* Migrating back embedded?
*/
if (acl_phys.z_acl_extern_obj) {
error = dmu_object_free(zfsvfs->z_os,
acl_phys.z_acl_extern_obj, tx);
if (error)
return (error);
acl_phys.z_acl_extern_obj = 0;
}
for (aclnode = list_head(&aclp->z_acl); aclnode;
aclnode = list_next(&aclp->z_acl, aclnode)) {
if (aclnode->z_ace_count == 0)
continue;
memcpy(start, aclnode->z_acldata,
aclnode->z_size);
start = (caddr_t)start + aclnode->z_size;
}
}
/*
* If Old version then swap count/bytes to match old
* layout of znode_acl_phys_t.
*/
if (aclp->z_version == ZFS_ACL_VERSION_INITIAL) {
acl_phys.z_acl_size = aclp->z_acl_count;
acl_phys.z_acl_count = aclp->z_acl_bytes;
} else {
acl_phys.z_acl_size = aclp->z_acl_bytes;
acl_phys.z_acl_count = aclp->z_acl_count;
}
acl_phys.z_acl_version = aclp->z_version;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ZNODE_ACL(zfsvfs), NULL,
&acl_phys, sizeof (acl_phys));
}
/*
* Replace ACL wide bits, but first clear them.
*/
zp->z_pflags &= ~ZFS_ACL_WIDE_FLAGS;
zp->z_pflags |= aclp->z_hints;
if (ace_trivial_common(aclp, 0, zfs_ace_walk) == 0)
zp->z_pflags |= ZFS_ACL_TRIVIAL;
zfs_tstamp_update_setup(zp, STATE_CHANGED, NULL, ctime);
return (sa_bulk_update(zp->z_sa_hdl, bulk, count, tx));
}
static void
zfs_acl_chmod(boolean_t isdir, uint64_t mode, boolean_t split, boolean_t trim,
zfs_acl_t *aclp)
{
void *acep = NULL;
uint64_t who;
int new_count, new_bytes;
int ace_size;
int entry_type;
uint16_t iflags, type;
uint32_t access_mask;
zfs_acl_node_t *newnode;
size_t abstract_size = aclp->z_ops->ace_abstract_size();
void *zacep;
trivial_acl_t masks;
new_count = new_bytes = 0;
acl_trivial_access_masks((mode_t)mode, isdir, &masks);
newnode = zfs_acl_node_alloc((abstract_size * 6) + aclp->z_acl_bytes);
zacep = newnode->z_acldata;
if (masks.allow0) {
zfs_set_ace(aclp, zacep, masks.allow0, ALLOW, -1, ACE_OWNER);
zacep = (void *)((uintptr_t)zacep + abstract_size);
new_count++;
new_bytes += abstract_size;
}
if (masks.deny1) {
zfs_set_ace(aclp, zacep, masks.deny1, DENY, -1, ACE_OWNER);
zacep = (void *)((uintptr_t)zacep + abstract_size);
new_count++;
new_bytes += abstract_size;
}
if (masks.deny2) {
zfs_set_ace(aclp, zacep, masks.deny2, DENY, -1, OWNING_GROUP);
zacep = (void *)((uintptr_t)zacep + abstract_size);
new_count++;
new_bytes += abstract_size;
}
while ((acep = zfs_acl_next_ace(aclp, acep, &who, &access_mask,
&iflags, &type))) {
entry_type = (iflags & ACE_TYPE_FLAGS);
/*
* ACEs used to represent the file mode may be divided
* into an equivalent pair of inherit-only and regular
* ACEs, if they are inheritable.
* Skip regular ACEs, which are replaced by the new mode.
*/
if (split && (entry_type == ACE_OWNER ||
entry_type == OWNING_GROUP ||
entry_type == ACE_EVERYONE)) {
if (!isdir || !(iflags &
(ACE_FILE_INHERIT_ACE|ACE_DIRECTORY_INHERIT_ACE)))
continue;
/*
* We preserve owner@, group@, or @everyone
* permissions, if they are inheritable, by
* copying them to inherit_only ACEs. This
* prevents inheritable permissions from being
* altered along with the file mode.
*/
iflags |= ACE_INHERIT_ONLY_ACE;
}
/*
* If this ACL has any inheritable ACEs, mark that in
* the hints (which are later masked into the pflags)
* so create knows to do inheritance.
*/
if (isdir && (iflags &
(ACE_FILE_INHERIT_ACE|ACE_DIRECTORY_INHERIT_ACE)))
aclp->z_hints |= ZFS_INHERIT_ACE;
if ((type != ALLOW && type != DENY) ||
(iflags & ACE_INHERIT_ONLY_ACE)) {
switch (type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
aclp->z_hints |= ZFS_ACL_OBJ_ACE;
break;
}
} else {
/*
* Limit permissions to be no greater than
* group permissions.
* The "aclinherit" and "aclmode" properties
* affect policy for create and chmod(2),
* respectively.
*/
if ((type == ALLOW) && trim)
access_mask &= masks.group;
}
zfs_set_ace(aclp, zacep, access_mask, type, who, iflags);
ace_size = aclp->z_ops->ace_size(acep);
zacep = (void *)((uintptr_t)zacep + ace_size);
new_count++;
new_bytes += ace_size;
}
zfs_set_ace(aclp, zacep, masks.owner, ALLOW, -1, ACE_OWNER);
zacep = (void *)((uintptr_t)zacep + abstract_size);
zfs_set_ace(aclp, zacep, masks.group, ALLOW, -1, OWNING_GROUP);
zacep = (void *)((uintptr_t)zacep + abstract_size);
zfs_set_ace(aclp, zacep, masks.everyone, ALLOW, -1, ACE_EVERYONE);
new_count += 3;
new_bytes += abstract_size * 3;
zfs_acl_release_nodes(aclp);
aclp->z_acl_count = new_count;
aclp->z_acl_bytes = new_bytes;
newnode->z_ace_count = new_count;
newnode->z_size = new_bytes;
list_insert_tail(&aclp->z_acl, newnode);
}
int
zfs_acl_chmod_setattr(znode_t *zp, zfs_acl_t **aclp, uint64_t mode)
{
int error = 0;
mutex_enter(&zp->z_acl_lock);
mutex_enter(&zp->z_lock);
if (ZTOZSB(zp)->z_acl_mode == ZFS_ACL_DISCARD)
*aclp = zfs_acl_alloc(zfs_acl_version_zp(zp));
else
error = zfs_acl_node_read(zp, B_TRUE, aclp, B_TRUE);
if (error == 0) {
(*aclp)->z_hints = zp->z_pflags & V4_ACL_WIDE_FLAGS;
zfs_acl_chmod(S_ISDIR(ZTOI(zp)->i_mode), mode, B_TRUE,
(ZTOZSB(zp)->z_acl_mode == ZFS_ACL_GROUPMASK), *aclp);
}
mutex_exit(&zp->z_lock);
mutex_exit(&zp->z_acl_lock);
return (error);
}
/*
* Should ACE be inherited?
*/
static int
zfs_ace_can_use(umode_t obj_mode, uint16_t acep_flags)
{
int iflags = (acep_flags & 0xf);
if (S_ISDIR(obj_mode) && (iflags & ACE_DIRECTORY_INHERIT_ACE))
return (1);
else if (iflags & ACE_FILE_INHERIT_ACE)
return (!(S_ISDIR(obj_mode) &&
(iflags & ACE_NO_PROPAGATE_INHERIT_ACE)));
return (0);
}
/*
* inherit inheritable ACEs from parent
*/
static zfs_acl_t *
zfs_acl_inherit(zfsvfs_t *zfsvfs, umode_t va_mode, zfs_acl_t *paclp,
uint64_t mode, boolean_t *need_chmod)
{
void *pacep = NULL;
void *acep;
zfs_acl_node_t *aclnode;
zfs_acl_t *aclp = NULL;
uint64_t who;
uint32_t access_mask;
uint16_t iflags, newflags, type;
size_t ace_size;
void *data1, *data2;
size_t data1sz, data2sz;
uint_t aclinherit;
boolean_t isdir = S_ISDIR(va_mode);
boolean_t isreg = S_ISREG(va_mode);
*need_chmod = B_TRUE;
aclp = zfs_acl_alloc(paclp->z_version);
aclinherit = zfsvfs->z_acl_inherit;
if (aclinherit == ZFS_ACL_DISCARD || S_ISLNK(va_mode))
return (aclp);
while ((pacep = zfs_acl_next_ace(paclp, pacep, &who,
&access_mask, &iflags, &type))) {
/*
* don't inherit bogus ACEs
*/
if (!zfs_acl_valid_ace_type(type, iflags))
continue;
/*
* Check if ACE is inheritable by this vnode
*/
if ((aclinherit == ZFS_ACL_NOALLOW && type == ALLOW) ||
!zfs_ace_can_use(va_mode, iflags))
continue;
/*
* If owner@, group@, or everyone@ inheritable
* then zfs_acl_chmod() isn't needed.
*/
if ((aclinherit == ZFS_ACL_PASSTHROUGH ||
aclinherit == ZFS_ACL_PASSTHROUGH_X) &&
((iflags & (ACE_OWNER|ACE_EVERYONE)) ||
((iflags & OWNING_GROUP) == OWNING_GROUP)) &&
(isreg || (isdir && (iflags & ACE_DIRECTORY_INHERIT_ACE))))
*need_chmod = B_FALSE;
/*
* Strip inherited execute permission from file if
* not in mode
*/
if (aclinherit == ZFS_ACL_PASSTHROUGH_X && type == ALLOW &&
!isdir && ((mode & (S_IXUSR|S_IXGRP|S_IXOTH)) == 0)) {
access_mask &= ~ACE_EXECUTE;
}
/*
* Strip write_acl and write_owner from permissions
* when inheriting an ACE
*/
if (aclinherit == ZFS_ACL_RESTRICTED && type == ALLOW) {
access_mask &= ~RESTRICTED_CLEAR;
}
ace_size = aclp->z_ops->ace_size(pacep);
aclnode = zfs_acl_node_alloc(ace_size);
list_insert_tail(&aclp->z_acl, aclnode);
acep = aclnode->z_acldata;
zfs_set_ace(aclp, acep, access_mask, type,
who, iflags|ACE_INHERITED_ACE);
/*
* Copy special opaque data if any
*/
if ((data1sz = paclp->z_ops->ace_data(pacep, &data1)) != 0) {
VERIFY((data2sz = aclp->z_ops->ace_data(acep,
&data2)) == data1sz);
memcpy(data2, data1, data2sz);
}
aclp->z_acl_count++;
aclnode->z_ace_count++;
aclp->z_acl_bytes += aclnode->z_size;
newflags = aclp->z_ops->ace_flags_get(acep);
/*
* If ACE is not to be inherited further, or if the vnode is
* not a directory, remove all inheritance flags
*/
if (!isdir || (iflags & ACE_NO_PROPAGATE_INHERIT_ACE)) {
newflags &= ~ALL_INHERIT;
aclp->z_ops->ace_flags_set(acep,
newflags|ACE_INHERITED_ACE);
continue;
}
/*
* This directory has an inheritable ACE
*/
aclp->z_hints |= ZFS_INHERIT_ACE;
/*
* If only FILE_INHERIT is set then turn on
* inherit_only
*/
if ((iflags & (ACE_FILE_INHERIT_ACE |
ACE_DIRECTORY_INHERIT_ACE)) == ACE_FILE_INHERIT_ACE) {
newflags |= ACE_INHERIT_ONLY_ACE;
aclp->z_ops->ace_flags_set(acep,
newflags|ACE_INHERITED_ACE);
} else {
newflags &= ~ACE_INHERIT_ONLY_ACE;
aclp->z_ops->ace_flags_set(acep,
newflags|ACE_INHERITED_ACE);
}
}
if (zfsvfs->z_acl_mode == ZFS_ACL_RESTRICTED &&
aclp->z_acl_count != 0) {
*need_chmod = B_FALSE;
}
return (aclp);
}
/*
* Create file system object initial permissions
* including inheritable ACEs.
* Also, create FUIDs for owner and group.
*/
int
zfs_acl_ids_create(znode_t *dzp, int flag, vattr_t *vap, cred_t *cr,
- vsecattr_t *vsecp, zfs_acl_ids_t *acl_ids, zuserns_t *mnt_ns)
+ vsecattr_t *vsecp, zfs_acl_ids_t *acl_ids, zidmap_t *mnt_ns)
{
int error;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zfs_acl_t *paclp;
gid_t gid = vap->va_gid;
boolean_t need_chmod = B_TRUE;
boolean_t trim = B_FALSE;
boolean_t inherited = B_FALSE;
memset(acl_ids, 0, sizeof (zfs_acl_ids_t));
acl_ids->z_mode = vap->va_mode;
if (vsecp)
if ((error = zfs_vsec_2_aclp(zfsvfs, vap->va_mode, vsecp,
cr, &acl_ids->z_fuidp, &acl_ids->z_aclp)) != 0)
return (error);
acl_ids->z_fuid = vap->va_uid;
acl_ids->z_fgid = vap->va_gid;
#ifdef HAVE_KSID
/*
* Determine uid and gid.
*/
if ((flag & IS_ROOT_NODE) || zfsvfs->z_replay ||
((flag & IS_XATTR) && (S_ISDIR(vap->va_mode)))) {
acl_ids->z_fuid = zfs_fuid_create(zfsvfs, (uint64_t)vap->va_uid,
cr, ZFS_OWNER, &acl_ids->z_fuidp);
acl_ids->z_fgid = zfs_fuid_create(zfsvfs, (uint64_t)vap->va_gid,
cr, ZFS_GROUP, &acl_ids->z_fuidp);
gid = vap->va_gid;
} else {
acl_ids->z_fuid = zfs_fuid_create_cred(zfsvfs, ZFS_OWNER,
cr, &acl_ids->z_fuidp);
acl_ids->z_fgid = 0;
if (vap->va_mask & AT_GID) {
acl_ids->z_fgid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_gid,
cr, ZFS_GROUP, &acl_ids->z_fuidp);
gid = vap->va_gid;
if (acl_ids->z_fgid != KGID_TO_SGID(ZTOI(dzp)->i_gid) &&
!groupmember(vap->va_gid, cr) &&
secpolicy_vnode_create_gid(cr) != 0)
acl_ids->z_fgid = 0;
}
if (acl_ids->z_fgid == 0) {
if (dzp->z_mode & S_ISGID) {
char *domain;
uint32_t rid;
acl_ids->z_fgid = KGID_TO_SGID(
ZTOI(dzp)->i_gid);
gid = zfs_fuid_map_id(zfsvfs, acl_ids->z_fgid,
cr, ZFS_GROUP);
if (zfsvfs->z_use_fuids &&
IS_EPHEMERAL(acl_ids->z_fgid)) {
domain = zfs_fuid_idx_domain(
&zfsvfs->z_fuid_idx,
FUID_INDEX(acl_ids->z_fgid));
rid = FUID_RID(acl_ids->z_fgid);
zfs_fuid_node_add(&acl_ids->z_fuidp,
domain, rid,
FUID_INDEX(acl_ids->z_fgid),
acl_ids->z_fgid, ZFS_GROUP);
}
} else {
acl_ids->z_fgid = zfs_fuid_create_cred(zfsvfs,
ZFS_GROUP, cr, &acl_ids->z_fuidp);
gid = crgetgid(cr);
}
}
}
#endif /* HAVE_KSID */
/*
* If we're creating a directory, and the parent directory has the
* set-GID bit set, set in on the new directory.
* Otherwise, if the user is neither privileged nor a member of the
* file's new group, clear the file's set-GID bit.
*/
if (!(flag & IS_ROOT_NODE) && (dzp->z_mode & S_ISGID) &&
(S_ISDIR(vap->va_mode))) {
acl_ids->z_mode |= S_ISGID;
} else {
if ((acl_ids->z_mode & S_ISGID) &&
secpolicy_vnode_setids_setgids(cr, gid, mnt_ns,
zfs_i_user_ns(ZTOI(dzp))) != 0) {
acl_ids->z_mode &= ~S_ISGID;
}
}
if (acl_ids->z_aclp == NULL) {
mutex_enter(&dzp->z_acl_lock);
mutex_enter(&dzp->z_lock);
if (!(flag & IS_ROOT_NODE) &&
(dzp->z_pflags & ZFS_INHERIT_ACE) &&
!(dzp->z_pflags & ZFS_XATTR)) {
VERIFY(0 == zfs_acl_node_read(dzp, B_TRUE,
&paclp, B_FALSE));
acl_ids->z_aclp = zfs_acl_inherit(zfsvfs,
vap->va_mode, paclp, acl_ids->z_mode, &need_chmod);
inherited = B_TRUE;
} else {
acl_ids->z_aclp =
zfs_acl_alloc(zfs_acl_version_zp(dzp));
acl_ids->z_aclp->z_hints |= ZFS_ACL_TRIVIAL;
}
mutex_exit(&dzp->z_lock);
mutex_exit(&dzp->z_acl_lock);
if (need_chmod) {
if (S_ISDIR(vap->va_mode))
acl_ids->z_aclp->z_hints |=
ZFS_ACL_AUTO_INHERIT;
if (zfsvfs->z_acl_mode == ZFS_ACL_GROUPMASK &&
zfsvfs->z_acl_inherit != ZFS_ACL_PASSTHROUGH &&
zfsvfs->z_acl_inherit != ZFS_ACL_PASSTHROUGH_X)
trim = B_TRUE;
zfs_acl_chmod(vap->va_mode, acl_ids->z_mode, B_FALSE,
trim, acl_ids->z_aclp);
}
}
if (inherited || vsecp) {
acl_ids->z_mode = zfs_mode_compute(acl_ids->z_mode,
acl_ids->z_aclp, &acl_ids->z_aclp->z_hints,
acl_ids->z_fuid, acl_ids->z_fgid);
if (ace_trivial_common(acl_ids->z_aclp, 0, zfs_ace_walk) == 0)
acl_ids->z_aclp->z_hints |= ZFS_ACL_TRIVIAL;
}
return (0);
}
/*
* Free ACL and fuid_infop, but not the acl_ids structure
*/
void
zfs_acl_ids_free(zfs_acl_ids_t *acl_ids)
{
if (acl_ids->z_aclp)
zfs_acl_free(acl_ids->z_aclp);
if (acl_ids->z_fuidp)
zfs_fuid_info_free(acl_ids->z_fuidp);
acl_ids->z_aclp = NULL;
acl_ids->z_fuidp = NULL;
}
boolean_t
zfs_acl_ids_overquota(zfsvfs_t *zv, zfs_acl_ids_t *acl_ids, uint64_t projid)
{
return (zfs_id_overquota(zv, DMU_USERUSED_OBJECT, acl_ids->z_fuid) ||
zfs_id_overquota(zv, DMU_GROUPUSED_OBJECT, acl_ids->z_fgid) ||
(projid != ZFS_DEFAULT_PROJID && projid != ZFS_INVALID_PROJID &&
zfs_id_overquota(zv, DMU_PROJECTUSED_OBJECT, projid)));
}
/*
* Retrieve a file's ACL
*/
int
zfs_getacl(znode_t *zp, vsecattr_t *vsecp, boolean_t skipaclchk, cred_t *cr)
{
zfs_acl_t *aclp;
ulong_t mask;
int error;
int count = 0;
int largeace = 0;
mask = vsecp->vsa_mask & (VSA_ACE | VSA_ACECNT |
VSA_ACE_ACLFLAGS | VSA_ACE_ALLTYPES);
if (mask == 0)
return (SET_ERROR(ENOSYS));
if ((error = zfs_zaccess(zp, ACE_READ_ACL, 0, skipaclchk, cr,
- kcred->user_ns)))
+ zfs_init_idmap)))
return (error);
mutex_enter(&zp->z_acl_lock);
error = zfs_acl_node_read(zp, B_FALSE, &aclp, B_FALSE);
if (error != 0) {
mutex_exit(&zp->z_acl_lock);
return (error);
}
/*
* Scan ACL to determine number of ACEs
*/
if ((zp->z_pflags & ZFS_ACL_OBJ_ACE) && !(mask & VSA_ACE_ALLTYPES)) {
void *zacep = NULL;
uint64_t who;
uint32_t access_mask;
uint16_t type, iflags;
while ((zacep = zfs_acl_next_ace(aclp, zacep,
&who, &access_mask, &iflags, &type))) {
switch (type) {
case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
case ACE_ACCESS_DENIED_OBJECT_ACE_TYPE:
case ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE:
case ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE:
largeace++;
continue;
default:
count++;
}
}
vsecp->vsa_aclcnt = count;
} else
count = (int)aclp->z_acl_count;
if (mask & VSA_ACECNT) {
vsecp->vsa_aclcnt = count;
}
if (mask & VSA_ACE) {
size_t aclsz;
aclsz = count * sizeof (ace_t) +
sizeof (ace_object_t) * largeace;
vsecp->vsa_aclentp = kmem_alloc(aclsz, KM_SLEEP);
vsecp->vsa_aclentsz = aclsz;
if (aclp->z_version == ZFS_ACL_VERSION_FUID)
zfs_copy_fuid_2_ace(ZTOZSB(zp), aclp, cr,
vsecp->vsa_aclentp, !(mask & VSA_ACE_ALLTYPES));
else {
zfs_acl_node_t *aclnode;
void *start = vsecp->vsa_aclentp;
for (aclnode = list_head(&aclp->z_acl); aclnode;
aclnode = list_next(&aclp->z_acl, aclnode)) {
memcpy(start, aclnode->z_acldata,
aclnode->z_size);
start = (caddr_t)start + aclnode->z_size;
}
ASSERT((caddr_t)start - (caddr_t)vsecp->vsa_aclentp ==
aclp->z_acl_bytes);
}
}
if (mask & VSA_ACE_ACLFLAGS) {
vsecp->vsa_aclflags = 0;
if (zp->z_pflags & ZFS_ACL_DEFAULTED)
vsecp->vsa_aclflags |= ACL_DEFAULTED;
if (zp->z_pflags & ZFS_ACL_PROTECTED)
vsecp->vsa_aclflags |= ACL_PROTECTED;
if (zp->z_pflags & ZFS_ACL_AUTO_INHERIT)
vsecp->vsa_aclflags |= ACL_AUTO_INHERIT;
}
mutex_exit(&zp->z_acl_lock);
return (0);
}
int
zfs_vsec_2_aclp(zfsvfs_t *zfsvfs, umode_t obj_mode,
vsecattr_t *vsecp, cred_t *cr, zfs_fuid_info_t **fuidp, zfs_acl_t **zaclp)
{
zfs_acl_t *aclp;
zfs_acl_node_t *aclnode;
int aclcnt = vsecp->vsa_aclcnt;
int error;
if (vsecp->vsa_aclcnt > MAX_ACL_ENTRIES || vsecp->vsa_aclcnt <= 0)
return (SET_ERROR(EINVAL));
aclp = zfs_acl_alloc(zfs_acl_version(zfsvfs->z_version));
aclp->z_hints = 0;
aclnode = zfs_acl_node_alloc(aclcnt * sizeof (zfs_object_ace_t));
if (aclp->z_version == ZFS_ACL_VERSION_INITIAL) {
if ((error = zfs_copy_ace_2_oldace(obj_mode, aclp,
(ace_t *)vsecp->vsa_aclentp, aclnode->z_acldata,
aclcnt, &aclnode->z_size)) != 0) {
zfs_acl_free(aclp);
zfs_acl_node_free(aclnode);
return (error);
}
} else {
if ((error = zfs_copy_ace_2_fuid(zfsvfs, obj_mode, aclp,
vsecp->vsa_aclentp, aclnode->z_acldata, aclcnt,
&aclnode->z_size, fuidp, cr)) != 0) {
zfs_acl_free(aclp);
zfs_acl_node_free(aclnode);
return (error);
}
}
aclp->z_acl_bytes = aclnode->z_size;
aclnode->z_ace_count = aclcnt;
aclp->z_acl_count = aclcnt;
list_insert_head(&aclp->z_acl, aclnode);
/*
* If flags are being set then add them to z_hints
*/
if (vsecp->vsa_mask & VSA_ACE_ACLFLAGS) {
if (vsecp->vsa_aclflags & ACL_PROTECTED)
aclp->z_hints |= ZFS_ACL_PROTECTED;
if (vsecp->vsa_aclflags & ACL_DEFAULTED)
aclp->z_hints |= ZFS_ACL_DEFAULTED;
if (vsecp->vsa_aclflags & ACL_AUTO_INHERIT)
aclp->z_hints |= ZFS_ACL_AUTO_INHERIT;
}
*zaclp = aclp;
return (0);
}
/*
* Set a file's ACL
*/
int
zfs_setacl(znode_t *zp, vsecattr_t *vsecp, boolean_t skipaclchk, cred_t *cr)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
zilog_t *zilog = zfsvfs->z_log;
ulong_t mask = vsecp->vsa_mask & (VSA_ACE | VSA_ACECNT);
dmu_tx_t *tx;
int error;
zfs_acl_t *aclp;
zfs_fuid_info_t *fuidp = NULL;
boolean_t fuid_dirtied;
uint64_t acl_obj;
if (mask == 0)
return (SET_ERROR(ENOSYS));
if (zp->z_pflags & ZFS_IMMUTABLE)
return (SET_ERROR(EPERM));
if ((error = zfs_zaccess(zp, ACE_WRITE_ACL, 0, skipaclchk, cr,
- kcred->user_ns)))
+ zfs_init_idmap)))
return (error);
error = zfs_vsec_2_aclp(zfsvfs, ZTOI(zp)->i_mode, vsecp, cr, &fuidp,
&aclp);
if (error)
return (error);
/*
* If ACL wide flags aren't being set then preserve any
* existing flags.
*/
if (!(vsecp->vsa_mask & VSA_ACE_ACLFLAGS)) {
aclp->z_hints |=
(zp->z_pflags & V4_ACL_WIDE_FLAGS);
}
top:
mutex_enter(&zp->z_acl_lock);
mutex_enter(&zp->z_lock);
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
/*
* If old version and ACL won't fit in bonus and we aren't
* upgrading then take out necessary DMU holds
*/
if ((acl_obj = zfs_external_acl(zp)) != 0) {
if (zfsvfs->z_version >= ZPL_VERSION_FUID &&
zfs_znode_acl_version(zp) <= ZFS_ACL_VERSION_INITIAL) {
dmu_tx_hold_free(tx, acl_obj, 0,
DMU_OBJECT_END);
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0,
aclp->z_acl_bytes);
} else {
dmu_tx_hold_write(tx, acl_obj, 0, aclp->z_acl_bytes);
}
} else if (!zp->z_is_sa && aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, aclp->z_acl_bytes);
}
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_NOWAIT);
if (error) {
mutex_exit(&zp->z_acl_lock);
mutex_exit(&zp->z_lock);
if (error == ERESTART) {
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
dmu_tx_abort(tx);
zfs_acl_free(aclp);
return (error);
}
error = zfs_aclset_common(zp, aclp, cr, tx);
ASSERT(error == 0);
ASSERT(zp->z_acl_cached == NULL);
zp->z_acl_cached = aclp;
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
zfs_log_acl(zilog, tx, zp, vsecp, fuidp);
if (fuidp)
zfs_fuid_info_free(fuidp);
dmu_tx_commit(tx);
mutex_exit(&zp->z_lock);
mutex_exit(&zp->z_acl_lock);
return (error);
}
/*
* Check accesses of interest (AoI) against attributes of the dataset
* such as read-only. Returns zero if no AoI conflict with dataset
* attributes, otherwise an appropriate errno is returned.
*/
static int
zfs_zaccess_dataset_check(znode_t *zp, uint32_t v4_mode)
{
if ((v4_mode & WRITE_MASK) && (zfs_is_readonly(ZTOZSB(zp))) &&
(!Z_ISDEV(ZTOI(zp)->i_mode) || (v4_mode & WRITE_MASK_ATTRS))) {
return (SET_ERROR(EROFS));
}
/*
* Intentionally allow ZFS_READONLY through here.
* See zfs_zaccess_common().
*/
if ((v4_mode & WRITE_MASK_DATA) &&
(zp->z_pflags & ZFS_IMMUTABLE)) {
return (SET_ERROR(EPERM));
}
if ((v4_mode & (ACE_DELETE | ACE_DELETE_CHILD)) &&
(zp->z_pflags & ZFS_NOUNLINK)) {
return (SET_ERROR(EPERM));
}
if (((v4_mode & (ACE_READ_DATA|ACE_EXECUTE)) &&
(zp->z_pflags & ZFS_AV_QUARANTINED))) {
return (SET_ERROR(EACCES));
}
return (0);
}
/*
* The primary usage of this function is to loop through all of the
* ACEs in the znode, determining what accesses of interest (AoI) to
* the caller are allowed or denied. The AoI are expressed as bits in
* the working_mode parameter. As each ACE is processed, bits covered
* by that ACE are removed from the working_mode. This removal
* facilitates two things. The first is that when the working mode is
* empty (= 0), we know we've looked at all the AoI. The second is
* that the ACE interpretation rules don't allow a later ACE to undo
* something granted or denied by an earlier ACE. Removing the
* discovered access or denial enforces this rule. At the end of
* processing the ACEs, all AoI that were found to be denied are
* placed into the working_mode, giving the caller a mask of denied
* accesses. Returns:
* 0 if all AoI granted
* EACCES if the denied mask is non-zero
* other error if abnormal failure (e.g., IO error)
*
* A secondary usage of the function is to determine if any of the
* AoI are granted. If an ACE grants any access in
* the working_mode, we immediately short circuit out of the function.
* This mode is chosen by setting anyaccess to B_TRUE. The
* working_mode is not a denied access mask upon exit if the function
* is used in this manner.
*/
static int
zfs_zaccess_aces_check(znode_t *zp, uint32_t *working_mode,
- boolean_t anyaccess, cred_t *cr, zuserns_t *mnt_ns)
+ boolean_t anyaccess, cred_t *cr, zidmap_t *mnt_ns)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
zfs_acl_t *aclp;
int error;
uid_t uid = crgetuid(cr);
uint64_t who;
uint16_t type, iflags;
uint16_t entry_type;
uint32_t access_mask;
uint32_t deny_mask = 0;
zfs_ace_hdr_t *acep = NULL;
boolean_t checkit;
uid_t gowner;
uid_t fowner;
if (mnt_ns) {
fowner = zfs_uid_to_vfsuid(mnt_ns, zfs_i_user_ns(ZTOI(zp)),
KUID_TO_SUID(ZTOI(zp)->i_uid));
gowner = zfs_gid_to_vfsgid(mnt_ns, zfs_i_user_ns(ZTOI(zp)),
KGID_TO_SGID(ZTOI(zp)->i_gid));
} else
zfs_fuid_map_ids(zp, cr, &fowner, &gowner);
mutex_enter(&zp->z_acl_lock);
error = zfs_acl_node_read(zp, B_FALSE, &aclp, B_FALSE);
if (error != 0) {
mutex_exit(&zp->z_acl_lock);
return (error);
}
ASSERT(zp->z_acl_cached);
while ((acep = zfs_acl_next_ace(aclp, acep, &who, &access_mask,
&iflags, &type))) {
uint32_t mask_matched;
if (!zfs_acl_valid_ace_type(type, iflags))
continue;
if (S_ISDIR(ZTOI(zp)->i_mode) &&
(iflags & ACE_INHERIT_ONLY_ACE))
continue;
/* Skip ACE if it does not affect any AoI */
mask_matched = (access_mask & *working_mode);
if (!mask_matched)
continue;
entry_type = (iflags & ACE_TYPE_FLAGS);
checkit = B_FALSE;
switch (entry_type) {
case ACE_OWNER:
if (uid == fowner)
checkit = B_TRUE;
break;
case OWNING_GROUP:
who = gowner;
zfs_fallthrough;
case ACE_IDENTIFIER_GROUP:
checkit = zfs_groupmember(zfsvfs, who, cr);
break;
case ACE_EVERYONE:
checkit = B_TRUE;
break;
/* USER Entry */
default:
if (entry_type == 0) {
uid_t newid;
newid = zfs_fuid_map_id(zfsvfs, who, cr,
ZFS_ACE_USER);
if (newid != IDMAP_WK_CREATOR_OWNER_UID &&
uid == newid)
checkit = B_TRUE;
break;
} else {
mutex_exit(&zp->z_acl_lock);
return (SET_ERROR(EIO));
}
}
if (checkit) {
if (type == DENY) {
DTRACE_PROBE3(zfs__ace__denies,
znode_t *, zp,
zfs_ace_hdr_t *, acep,
uint32_t, mask_matched);
deny_mask |= mask_matched;
} else {
DTRACE_PROBE3(zfs__ace__allows,
znode_t *, zp,
zfs_ace_hdr_t *, acep,
uint32_t, mask_matched);
if (anyaccess) {
mutex_exit(&zp->z_acl_lock);
return (0);
}
}
*working_mode &= ~mask_matched;
}
/* Are we done? */
if (*working_mode == 0)
break;
}
mutex_exit(&zp->z_acl_lock);
/* Put the found 'denies' back on the working mode */
if (deny_mask) {
*working_mode |= deny_mask;
return (SET_ERROR(EACCES));
} else if (*working_mode) {
return (-1);
}
return (0);
}
/*
* Return true if any access whatsoever granted, we don't actually
* care what access is granted.
*/
boolean_t
zfs_has_access(znode_t *zp, cred_t *cr)
{
uint32_t have = ACE_ALL_PERMS;
if (zfs_zaccess_aces_check(zp, &have, B_TRUE, cr,
- kcred->user_ns) != 0) {
+ zfs_init_idmap) != 0) {
uid_t owner;
owner = zfs_fuid_map_id(ZTOZSB(zp),
KUID_TO_SUID(ZTOI(zp)->i_uid), cr, ZFS_OWNER);
return (secpolicy_vnode_any_access(cr, ZTOI(zp), owner) == 0);
}
return (B_TRUE);
}
/*
* Simplified access check for case where ACL is known to not contain
* information beyond what is defined in the mode. In this case, we
* can pass along to the kernel / vfs generic_permission() check, which
* evaluates the mode and POSIX ACL.
*
* NFSv4 ACLs allow granting permissions that are usually relegated only
* to the file owner or superuser. Examples are ACE_WRITE_OWNER (chown),
* ACE_WRITE_ACL(chmod), and ACE_DELETE. ACE_DELETE requests must fail
* because with conventional posix permissions, right to delete file
* is determined by write bit on the parent dir.
*
* If unmappable perms are requested, then we must return EPERM
* and include those bits in the working_mode so that the caller of
* zfs_zaccess_common() can decide whether to perform additional
* policy / capability checks. EACCES is used in zfs_zaccess_aces_check()
* to indicate access check failed due to explicit DENY entry, and so
* we want to avoid that here.
*/
static int
zfs_zaccess_trivial(znode_t *zp, uint32_t *working_mode, cred_t *cr,
- zuserns_t *mnt_ns)
+ zidmap_t *mnt_ns)
{
int err, mask;
int unmapped = 0;
ASSERT(zp->z_pflags & ZFS_ACL_TRIVIAL);
mask = zfs_v4_to_unix(*working_mode, &unmapped);
if (mask == 0 || unmapped) {
*working_mode = unmapped;
return (unmapped ? SET_ERROR(EPERM) : 0);
}
-#if defined(HAVE_IOPS_PERMISSION_USERNS)
- if (mnt_ns)
- err = generic_permission(mnt_ns, ZTOI(zp), mask);
- else
- err = generic_permission(cr->user_ns, ZTOI(zp), mask);
+#if (defined(HAVE_IOPS_PERMISSION_USERNS) || \
+ defined(HAVE_IOPS_PERMISSION_IDMAP))
+ err = generic_permission(mnt_ns, ZTOI(zp), mask);
#else
err = generic_permission(ZTOI(zp), mask);
#endif
if (err != 0) {
return (SET_ERROR(EPERM));
}
*working_mode = unmapped;
return (0);
}
static int
zfs_zaccess_common(znode_t *zp, uint32_t v4_mode, uint32_t *working_mode,
- boolean_t *check_privs, boolean_t skipaclchk, cred_t *cr, zuserns_t *mnt_ns)
+ boolean_t *check_privs, boolean_t skipaclchk, cred_t *cr, zidmap_t *mnt_ns)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int err;
*working_mode = v4_mode;
*check_privs = B_TRUE;
/*
* Short circuit empty requests
*/
if (v4_mode == 0 || zfsvfs->z_replay) {
*working_mode = 0;
return (0);
}
if ((err = zfs_zaccess_dataset_check(zp, v4_mode)) != 0) {
*check_privs = B_FALSE;
return (err);
}
/*
* The caller requested that the ACL check be skipped. This
* would only happen if the caller checked VOP_ACCESS() with a
* 32 bit ACE mask and already had the appropriate permissions.
*/
if (skipaclchk) {
*working_mode = 0;
return (0);
}
/*
* Note: ZFS_READONLY represents the "DOS R/O" attribute.
* When that flag is set, we should behave as if write access
* were not granted by anything in the ACL. In particular:
* We _must_ allow writes after opening the file r/w, then
* setting the DOS R/O attribute, and writing some more.
* (Similar to how you can write after fchmod(fd, 0444).)
*
* Therefore ZFS_READONLY is ignored in the dataset check
* above, and checked here as if part of the ACL check.
* Also note: DOS R/O is ignored for directories.
*/
if ((v4_mode & WRITE_MASK_DATA) &&
S_ISDIR(ZTOI(zp)->i_mode) &&
(zp->z_pflags & ZFS_READONLY)) {
return (SET_ERROR(EPERM));
}
if (zp->z_pflags & ZFS_ACL_TRIVIAL)
return (zfs_zaccess_trivial(zp, working_mode, cr, mnt_ns));
return (zfs_zaccess_aces_check(zp, working_mode, B_FALSE, cr, mnt_ns));
}
static int
zfs_zaccess_append(znode_t *zp, uint32_t *working_mode, boolean_t *check_privs,
- cred_t *cr, zuserns_t *mnt_ns)
+ cred_t *cr, zidmap_t *mnt_ns)
{
if (*working_mode != ACE_WRITE_DATA)
return (SET_ERROR(EACCES));
return (zfs_zaccess_common(zp, ACE_APPEND_DATA, working_mode,
check_privs, B_FALSE, cr, mnt_ns));
}
int
zfs_fastaccesschk_execute(znode_t *zdp, cred_t *cr)
{
boolean_t owner = B_FALSE;
boolean_t groupmbr = B_FALSE;
boolean_t is_attr;
uid_t uid = crgetuid(cr);
int error;
if (zdp->z_pflags & ZFS_AV_QUARANTINED)
return (SET_ERROR(EACCES));
is_attr = ((zdp->z_pflags & ZFS_XATTR) &&
(S_ISDIR(ZTOI(zdp)->i_mode)));
if (is_attr)
goto slow;
mutex_enter(&zdp->z_acl_lock);
if (zdp->z_pflags & ZFS_NO_EXECS_DENIED) {
mutex_exit(&zdp->z_acl_lock);
return (0);
}
if (KUID_TO_SUID(ZTOI(zdp)->i_uid) != 0 ||
KGID_TO_SGID(ZTOI(zdp)->i_gid) != 0) {
mutex_exit(&zdp->z_acl_lock);
goto slow;
}
if (uid == KUID_TO_SUID(ZTOI(zdp)->i_uid)) {
if (zdp->z_mode & S_IXUSR) {
mutex_exit(&zdp->z_acl_lock);
return (0);
} else {
mutex_exit(&zdp->z_acl_lock);
goto slow;
}
}
if (groupmember(KGID_TO_SGID(ZTOI(zdp)->i_gid), cr)) {
if (zdp->z_mode & S_IXGRP) {
mutex_exit(&zdp->z_acl_lock);
return (0);
} else {
mutex_exit(&zdp->z_acl_lock);
goto slow;
}
}
if (!owner && !groupmbr) {
if (zdp->z_mode & S_IXOTH) {
mutex_exit(&zdp->z_acl_lock);
return (0);
}
}
mutex_exit(&zdp->z_acl_lock);
slow:
DTRACE_PROBE(zfs__fastpath__execute__access__miss);
if ((error = zfs_enter(ZTOZSB(zdp), FTAG)) != 0)
return (error);
error = zfs_zaccess(zdp, ACE_EXECUTE, 0, B_FALSE, cr,
- kcred->user_ns);
+ zfs_init_idmap);
zfs_exit(ZTOZSB(zdp), FTAG);
return (error);
}
/*
* Determine whether Access should be granted/denied.
*
* The least priv subsystem is always consulted as a basic privilege
* can define any form of access.
*/
int
zfs_zaccess(znode_t *zp, int mode, int flags, boolean_t skipaclchk, cred_t *cr,
- zuserns_t *mnt_ns)
+ zidmap_t *mnt_ns)
{
uint32_t working_mode;
int error;
int is_attr;
boolean_t check_privs;
znode_t *xzp;
znode_t *check_zp = zp;
mode_t needed_bits;
uid_t owner;
is_attr = ((zp->z_pflags & ZFS_XATTR) && S_ISDIR(ZTOI(zp)->i_mode));
/*
* If attribute then validate against base file
*/
if (is_attr) {
if ((error = zfs_zget(ZTOZSB(zp),
zp->z_xattr_parent, &xzp)) != 0) {
return (error);
}
check_zp = xzp;
/*
* fixup mode to map to xattr perms
*/
if (mode & (ACE_WRITE_DATA|ACE_APPEND_DATA)) {
mode &= ~(ACE_WRITE_DATA|ACE_APPEND_DATA);
mode |= ACE_WRITE_NAMED_ATTRS;
}
if (mode & (ACE_READ_DATA|ACE_EXECUTE)) {
mode &= ~(ACE_READ_DATA|ACE_EXECUTE);
mode |= ACE_READ_NAMED_ATTRS;
}
}
owner = zfs_uid_to_vfsuid(mnt_ns, zfs_i_user_ns(ZTOI(zp)),
KUID_TO_SUID(ZTOI(zp)->i_uid));
owner = zfs_fuid_map_id(ZTOZSB(zp), owner, cr, ZFS_OWNER);
/*
* Map the bits required to the standard inode flags
* S_IRUSR|S_IWUSR|S_IXUSR in the needed_bits. Map the bits
* mapped by working_mode (currently missing) in missing_bits.
* Call secpolicy_vnode_access2() with (needed_bits & ~checkmode),
* needed_bits.
*/
needed_bits = 0;
working_mode = mode;
if ((working_mode & (ACE_READ_ACL|ACE_READ_ATTRIBUTES)) &&
owner == crgetuid(cr))
working_mode &= ~(ACE_READ_ACL|ACE_READ_ATTRIBUTES);
if (working_mode & (ACE_READ_DATA|ACE_READ_NAMED_ATTRS|
ACE_READ_ACL|ACE_READ_ATTRIBUTES|ACE_SYNCHRONIZE))
needed_bits |= S_IRUSR;
if (working_mode & (ACE_WRITE_DATA|ACE_WRITE_NAMED_ATTRS|
ACE_APPEND_DATA|ACE_WRITE_ATTRIBUTES|ACE_SYNCHRONIZE))
needed_bits |= S_IWUSR;
if (working_mode & ACE_EXECUTE)
needed_bits |= S_IXUSR;
if ((error = zfs_zaccess_common(check_zp, mode, &working_mode,
&check_privs, skipaclchk, cr, mnt_ns)) == 0) {
if (is_attr)
zrele(xzp);
return (secpolicy_vnode_access2(cr, ZTOI(zp), owner,
needed_bits, needed_bits));
}
if (error && !check_privs) {
if (is_attr)
zrele(xzp);
return (error);
}
if (error && (flags & V_APPEND)) {
error = zfs_zaccess_append(zp, &working_mode, &check_privs, cr,
mnt_ns);
}
if (error && check_privs) {
mode_t checkmode = 0;
/*
* First check for implicit owner permission on
* read_acl/read_attributes
*/
ASSERT(working_mode != 0);
if ((working_mode & (ACE_READ_ACL|ACE_READ_ATTRIBUTES) &&
owner == crgetuid(cr)))
working_mode &= ~(ACE_READ_ACL|ACE_READ_ATTRIBUTES);
if (working_mode & (ACE_READ_DATA|ACE_READ_NAMED_ATTRS|
ACE_READ_ACL|ACE_READ_ATTRIBUTES|ACE_SYNCHRONIZE))
checkmode |= S_IRUSR;
if (working_mode & (ACE_WRITE_DATA|ACE_WRITE_NAMED_ATTRS|
ACE_APPEND_DATA|ACE_WRITE_ATTRIBUTES|ACE_SYNCHRONIZE))
checkmode |= S_IWUSR;
if (working_mode & ACE_EXECUTE)
checkmode |= S_IXUSR;
error = secpolicy_vnode_access2(cr, ZTOI(check_zp), owner,
needed_bits & ~checkmode, needed_bits);
if (error == 0 && (working_mode & ACE_WRITE_OWNER))
error = secpolicy_vnode_chown(cr, owner);
if (error == 0 && (working_mode & ACE_WRITE_ACL))
error = secpolicy_vnode_setdac(cr, owner);
if (error == 0 && (working_mode &
(ACE_DELETE|ACE_DELETE_CHILD)))
error = secpolicy_vnode_remove(cr);
if (error == 0 && (working_mode & ACE_SYNCHRONIZE)) {
error = secpolicy_vnode_chown(cr, owner);
}
if (error == 0) {
/*
* See if any bits other than those already checked
* for are still present. If so then return EACCES
*/
if (working_mode & ~(ZFS_CHECKED_MASKS)) {
error = SET_ERROR(EACCES);
}
}
} else if (error == 0) {
error = secpolicy_vnode_access2(cr, ZTOI(zp), owner,
needed_bits, needed_bits);
}
if (is_attr)
zrele(xzp);
return (error);
}
/*
* Translate traditional unix S_IRUSR/S_IWUSR/S_IXUSR mode into
* NFSv4-style ZFS ACL format and call zfs_zaccess()
*/
int
zfs_zaccess_rwx(znode_t *zp, mode_t mode, int flags, cred_t *cr,
- zuserns_t *mnt_ns)
+ zidmap_t *mnt_ns)
{
return (zfs_zaccess(zp, zfs_unix_to_v4(mode >> 6), flags, B_FALSE, cr,
mnt_ns));
}
/*
* Access function for secpolicy_vnode_setattr
*/
int
zfs_zaccess_unix(void *zp, int mode, cred_t *cr)
{
int v4_mode = zfs_unix_to_v4(mode >> 6);
- return (zfs_zaccess(zp, v4_mode, 0, B_FALSE, cr, kcred->user_ns));
+ return (zfs_zaccess(zp, v4_mode, 0, B_FALSE, cr, zfs_init_idmap));
}
/* See zfs_zaccess_delete() */
static const boolean_t zfs_write_implies_delete_child = B_TRUE;
/*
* Determine whether delete access should be granted.
*
* The following chart outlines how we handle delete permissions which is
* how recent versions of windows (Windows 2008) handles it. The efficiency
* comes from not having to check the parent ACL where the object itself grants
* delete:
*
* -------------------------------------------------------
* | Parent Dir | Target Object Permissions |
* | permissions | |
* -------------------------------------------------------
* | | ACL Allows | ACL Denies| Delete |
* | | Delete | Delete | unspecified|
* -------------------------------------------------------
* | ACL Allows | Permit | Deny * | Permit |
* | DELETE_CHILD | | | |
* -------------------------------------------------------
* | ACL Denies | Permit | Deny | Deny |
* | DELETE_CHILD | | | |
* -------------------------------------------------------
* | ACL specifies | | | |
* | only allow | Permit | Deny * | Permit |
* | write and | | | |
* | execute | | | |
* -------------------------------------------------------
* | ACL denies | | | |
* | write and | Permit | Deny | Deny |
* | execute | | | |
* -------------------------------------------------------
* ^
* |
* Re. execute permission on the directory: if that's missing,
* the vnode lookup of the target will fail before we get here.
*
* Re [*] in the table above: NFSv4 would normally Permit delete for
* these two cells of the matrix.
* See acl.h for notes on which ACE_... flags should be checked for which
* operations. Specifically, the NFSv4 committee recommendation is in
* conflict with the Windows interpretation of DENY ACEs, where DENY ACEs
* should take precedence ahead of ALLOW ACEs.
*
* This implementation always consults the target object's ACL first.
* If a DENY ACE is present on the target object that specifies ACE_DELETE,
* delete access is denied. If an ALLOW ACE with ACE_DELETE is present on
* the target object, access is allowed. If and only if no entries with
* ACE_DELETE are present in the object's ACL, check the container's ACL
* for entries with ACE_DELETE_CHILD.
*
* A summary of the logic implemented from the table above is as follows:
*
* First check for DENY ACEs that apply.
* If either target or container has a deny, EACCES.
*
* Delete access can then be summarized as follows:
* 1: The object to be deleted grants ACE_DELETE, or
* 2: The containing directory grants ACE_DELETE_CHILD.
* In a Windows system, that would be the end of the story.
* In this system, (2) has some complications...
* 2a: "sticky" bit on a directory adds restrictions, and
* 2b: existing ACEs from previous versions of ZFS may
* not carry ACE_DELETE_CHILD where they should, so we
* also allow delete when ACE_WRITE_DATA is granted.
*
* Note: 2b is technically a work-around for a prior bug,
* which hopefully can go away some day. For those who
* no longer need the work around, and for testing, this
* work-around is made conditional via the tunable:
* zfs_write_implies_delete_child
*/
int
-zfs_zaccess_delete(znode_t *dzp, znode_t *zp, cred_t *cr, zuserns_t *mnt_ns)
+zfs_zaccess_delete(znode_t *dzp, znode_t *zp, cred_t *cr, zidmap_t *mnt_ns)
{
uint32_t wanted_dirperms;
uint32_t dzp_working_mode = 0;
uint32_t zp_working_mode = 0;
int dzp_error, zp_error;
boolean_t dzpcheck_privs;
boolean_t zpcheck_privs;
if (zp->z_pflags & (ZFS_IMMUTABLE | ZFS_NOUNLINK))
return (SET_ERROR(EPERM));
/*
* Case 1:
* If target object grants ACE_DELETE then we are done. This is
* indicated by a return value of 0. For this case we don't worry
* about the sticky bit because sticky only applies to the parent
* directory and this is the child access result.
*
* If we encounter a DENY ACE here, we're also done (EACCES).
* Note that if we hit a DENY ACE here (on the target) it should
* take precedence over a DENY ACE on the container, so that when
* we have more complete auditing support we will be able to
* report an access failure against the specific target.
* (This is part of why we're checking the target first.)
*/
zp_error = zfs_zaccess_common(zp, ACE_DELETE, &zp_working_mode,
&zpcheck_privs, B_FALSE, cr, mnt_ns);
if (zp_error == EACCES) {
/* We hit a DENY ACE. */
if (!zpcheck_privs)
return (SET_ERROR(zp_error));
return (secpolicy_vnode_remove(cr));
}
if (zp_error == 0)
return (0);
/*
* Case 2:
* If the containing directory grants ACE_DELETE_CHILD,
* or we're in backward compatibility mode and the
* containing directory has ACE_WRITE_DATA, allow.
* Case 2b is handled with wanted_dirperms.
*/
wanted_dirperms = ACE_DELETE_CHILD;
if (zfs_write_implies_delete_child)
wanted_dirperms |= ACE_WRITE_DATA;
dzp_error = zfs_zaccess_common(dzp, wanted_dirperms,
&dzp_working_mode, &dzpcheck_privs, B_FALSE, cr, mnt_ns);
if (dzp_error == EACCES) {
/* We hit a DENY ACE. */
if (!dzpcheck_privs)
return (SET_ERROR(dzp_error));
return (secpolicy_vnode_remove(cr));
}
/*
* Cases 2a, 2b (continued)
*
* Note: dzp_working_mode now contains any permissions
* that were NOT granted. Therefore, if any of the
* wanted_dirperms WERE granted, we will have:
* dzp_working_mode != wanted_dirperms
* We're really asking if ANY of those permissions
* were granted, and if so, grant delete access.
*/
if (dzp_working_mode != wanted_dirperms)
dzp_error = 0;
/*
* dzp_error is 0 if the container granted us permissions to "modify".
* If we do not have permission via one or more ACEs, our current
* privileges may still permit us to modify the container.
*
* dzpcheck_privs is false when i.e. the FS is read-only.
* Otherwise, do privilege checks for the container.
*/
if (dzp_error != 0 && dzpcheck_privs) {
uid_t owner;
/*
* The secpolicy call needs the requested access and
* the current access mode of the container, but it
* only knows about Unix-style modes (VEXEC, VWRITE),
* so this must condense the fine-grained ACE bits into
* Unix modes.
*
* The VEXEC flag is easy, because we know that has
* always been checked before we get here (during the
* lookup of the target vnode). The container has not
* granted us permissions to "modify", so we do not set
* the VWRITE flag in the current access mode.
*/
owner = zfs_fuid_map_id(ZTOZSB(dzp),
KUID_TO_SUID(ZTOI(dzp)->i_uid), cr, ZFS_OWNER);
dzp_error = secpolicy_vnode_access2(cr, ZTOI(dzp),
owner, S_IXUSR, S_IWUSR|S_IXUSR);
}
if (dzp_error != 0) {
/*
* Note: We may have dzp_error = -1 here (from
* zfs_zacess_common). Don't return that.
*/
return (SET_ERROR(EACCES));
}
/*
* At this point, we know that the directory permissions allow
* us to modify, but we still need to check for the additional
* restrictions that apply when the "sticky bit" is set.
*
* Yes, zfs_sticky_remove_access() also checks this bit, but
* checking it here and skipping the call below is nice when
* you're watching all of this with dtrace.
*/
if ((dzp->z_mode & S_ISVTX) == 0)
return (0);
/*
* zfs_sticky_remove_access will succeed if:
* 1. The sticky bit is absent.
* 2. We pass the sticky bit restrictions.
* 3. We have privileges that always allow file removal.
*/
return (zfs_sticky_remove_access(dzp, zp, cr));
}
int
zfs_zaccess_rename(znode_t *sdzp, znode_t *szp, znode_t *tdzp,
- znode_t *tzp, cred_t *cr, zuserns_t *mnt_ns)
+ znode_t *tzp, cred_t *cr, zidmap_t *mnt_ns)
{
int add_perm;
int error;
if (szp->z_pflags & ZFS_AV_QUARANTINED)
return (SET_ERROR(EACCES));
add_perm = S_ISDIR(ZTOI(szp)->i_mode) ?
ACE_ADD_SUBDIRECTORY : ACE_ADD_FILE;
/*
* Rename permissions are combination of delete permission +
* add file/subdir permission.
*/
/*
* first make sure we do the delete portion.
*
* If that succeeds then check for add_file/add_subdir permissions
*/
if ((error = zfs_zaccess_delete(sdzp, szp, cr, mnt_ns)))
return (error);
/*
* If we have a tzp, see if we can delete it?
*/
if (tzp) {
if ((error = zfs_zaccess_delete(tdzp, tzp, cr, mnt_ns)))
return (error);
}
/*
* Now check for add permissions
*/
error = zfs_zaccess(tdzp, add_perm, 0, B_FALSE, cr, mnt_ns);
return (error);
}
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c
index 1fec4ea09317..1eeabe53d23c 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_dir.c
@@ -1,1276 +1,1276 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013, 2016 by Delphix. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc.
*/
#include <sys/types.h>
#include <sys/param.h>
#include <sys/time.h>
#include <sys/sysmacros.h>
#include <sys/vfs.h>
#include <sys/vnode.h>
#include <sys/file.h>
#include <sys/kmem.h>
#include <sys/uio.h>
#include <sys/pathname.h>
#include <sys/cmn_err.h>
#include <sys/errno.h>
#include <sys/stat.h>
#include <sys/sunddi.h>
#include <sys/random.h>
#include <sys/policy.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_vnops.h>
#include <sys/fs/zfs.h>
#include <sys/zap.h>
#include <sys/dmu.h>
#include <sys/atomic.h>
#include <sys/zfs_ctldir.h>
#include <sys/zfs_fuid.h>
#include <sys/sa.h>
#include <sys/zfs_sa.h>
#include <sys/dmu_objset.h>
#include <sys/dsl_dir.h>
/*
* zfs_match_find() is used by zfs_dirent_lock() to perform zap lookups
* of names after deciding which is the appropriate lookup interface.
*/
static int
zfs_match_find(zfsvfs_t *zfsvfs, znode_t *dzp, const char *name,
matchtype_t mt, boolean_t update, int *deflags, pathname_t *rpnp,
uint64_t *zoid)
{
boolean_t conflict = B_FALSE;
int error;
if (zfsvfs->z_norm) {
size_t bufsz = 0;
char *buf = NULL;
if (rpnp) {
buf = rpnp->pn_buf;
bufsz = rpnp->pn_bufsize;
}
/*
* In the non-mixed case we only expect there would ever
* be one match, but we need to use the normalizing lookup.
*/
error = zap_lookup_norm(zfsvfs->z_os, dzp->z_id, name, 8, 1,
zoid, mt, buf, bufsz, &conflict);
} else {
error = zap_lookup(zfsvfs->z_os, dzp->z_id, name, 8, 1, zoid);
}
/*
* Allow multiple entries provided the first entry is
* the object id. Non-zpl consumers may safely make
* use of the additional space.
*
* XXX: This should be a feature flag for compatibility
*/
if (error == EOVERFLOW)
error = 0;
if (zfsvfs->z_norm && !error && deflags)
*deflags = conflict ? ED_CASE_CONFLICT : 0;
*zoid = ZFS_DIRENT_OBJ(*zoid);
return (error);
}
/*
* Lock a directory entry. A dirlock on <dzp, name> protects that name
* in dzp's directory zap object. As long as you hold a dirlock, you can
* assume two things: (1) dzp cannot be reaped, and (2) no other thread
* can change the zap entry for (i.e. link or unlink) this name.
*
* Input arguments:
* dzp - znode for directory
* name - name of entry to lock
* flag - ZNEW: if the entry already exists, fail with EEXIST.
* ZEXISTS: if the entry does not exist, fail with ENOENT.
* ZSHARED: allow concurrent access with other ZSHARED callers.
* ZXATTR: we want dzp's xattr directory
* ZCILOOK: On a mixed sensitivity file system,
* this lookup should be case-insensitive.
* ZCIEXACT: On a purely case-insensitive file system,
* this lookup should be case-sensitive.
* ZRENAMING: we are locking for renaming, force narrow locks
* ZHAVELOCK: Don't grab the z_name_lock for this call. The
* current thread already holds it.
*
* Output arguments:
* zpp - pointer to the znode for the entry (NULL if there isn't one)
* dlpp - pointer to the dirlock for this entry (NULL on error)
* direntflags - (case-insensitive lookup only)
* flags if multiple case-sensitive matches exist in directory
* realpnp - (case-insensitive lookup only)
* actual name matched within the directory
*
* Return value: 0 on success or errno on failure.
*
* NOTE: Always checks for, and rejects, '.' and '..'.
* NOTE: For case-insensitive file systems we take wide locks (see below),
* but return znode pointers to a single match.
*/
int
zfs_dirent_lock(zfs_dirlock_t **dlpp, znode_t *dzp, char *name,
znode_t **zpp, int flag, int *direntflags, pathname_t *realpnp)
{
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zfs_dirlock_t *dl;
boolean_t update;
matchtype_t mt = 0;
uint64_t zoid;
int error = 0;
int cmpflags;
*zpp = NULL;
*dlpp = NULL;
/*
* Verify that we are not trying to lock '.', '..', or '.zfs'
*/
if ((name[0] == '.' &&
(name[1] == '\0' || (name[1] == '.' && name[2] == '\0'))) ||
(zfs_has_ctldir(dzp) && strcmp(name, ZFS_CTLDIR_NAME) == 0))
return (SET_ERROR(EEXIST));
/*
* Case sensitivity and normalization preferences are set when
* the file system is created. These are stored in the
* zfsvfs->z_case and zfsvfs->z_norm fields. These choices
* affect what vnodes can be cached in the DNLC, how we
* perform zap lookups, and the "width" of our dirlocks.
*
* A normal dirlock locks a single name. Note that with
* normalization a name can be composed multiple ways, but
* when normalized, these names all compare equal. A wide
* dirlock locks multiple names. We need these when the file
* system is supporting mixed-mode access. It is sometimes
* necessary to lock all case permutations of file name at
* once so that simultaneous case-insensitive/case-sensitive
* behaves as rationally as possible.
*/
/*
* When matching we may need to normalize & change case according to
* FS settings.
*
* Note that a normalized match is necessary for a case insensitive
* filesystem when the lookup request is not exact because normalization
* can fold case independent of normalizing code point sequences.
*
* See the table above zfs_dropname().
*/
if (zfsvfs->z_norm != 0) {
mt = MT_NORMALIZE;
/*
* Determine if the match needs to honor the case specified in
* lookup, and if so keep track of that so that during
* normalization we don't fold case.
*/
if ((zfsvfs->z_case == ZFS_CASE_INSENSITIVE &&
(flag & ZCIEXACT)) ||
(zfsvfs->z_case == ZFS_CASE_MIXED && !(flag & ZCILOOK))) {
mt |= MT_MATCH_CASE;
}
}
/*
* Only look in or update the DNLC if we are looking for the
* name on a file system that does not require normalization
* or case folding. We can also look there if we happen to be
* on a non-normalizing, mixed sensitivity file system IF we
* are looking for the exact name.
*
* Maybe can add TO-UPPERed version of name to dnlc in ci-only
* case for performance improvement?
*/
update = !zfsvfs->z_norm ||
(zfsvfs->z_case == ZFS_CASE_MIXED &&
!(zfsvfs->z_norm & ~U8_TEXTPREP_TOUPPER) && !(flag & ZCILOOK));
/*
* ZRENAMING indicates we are in a situation where we should
* take narrow locks regardless of the file system's
* preferences for normalizing and case folding. This will
* prevent us deadlocking trying to grab the same wide lock
* twice if the two names happen to be case-insensitive
* matches.
*/
if (flag & ZRENAMING)
cmpflags = 0;
else
cmpflags = zfsvfs->z_norm;
/*
* Wait until there are no locks on this name.
*
* Don't grab the lock if it is already held. However, cannot
* have both ZSHARED and ZHAVELOCK together.
*/
ASSERT(!(flag & ZSHARED) || !(flag & ZHAVELOCK));
if (!(flag & ZHAVELOCK))
rw_enter(&dzp->z_name_lock, RW_READER);
mutex_enter(&dzp->z_lock);
for (;;) {
if (dzp->z_unlinked && !(flag & ZXATTR)) {
mutex_exit(&dzp->z_lock);
if (!(flag & ZHAVELOCK))
rw_exit(&dzp->z_name_lock);
return (SET_ERROR(ENOENT));
}
for (dl = dzp->z_dirlocks; dl != NULL; dl = dl->dl_next) {
if ((u8_strcmp(name, dl->dl_name, 0, cmpflags,
U8_UNICODE_LATEST, &error) == 0) || error != 0)
break;
}
if (error != 0) {
mutex_exit(&dzp->z_lock);
if (!(flag & ZHAVELOCK))
rw_exit(&dzp->z_name_lock);
return (SET_ERROR(ENOENT));
}
if (dl == NULL) {
/*
* Allocate a new dirlock and add it to the list.
*/
dl = kmem_alloc(sizeof (zfs_dirlock_t), KM_SLEEP);
cv_init(&dl->dl_cv, NULL, CV_DEFAULT, NULL);
dl->dl_name = name;
dl->dl_sharecnt = 0;
dl->dl_namelock = 0;
dl->dl_namesize = 0;
dl->dl_dzp = dzp;
dl->dl_next = dzp->z_dirlocks;
dzp->z_dirlocks = dl;
break;
}
if ((flag & ZSHARED) && dl->dl_sharecnt != 0)
break;
cv_wait(&dl->dl_cv, &dzp->z_lock);
}
/*
* If the z_name_lock was NOT held for this dirlock record it.
*/
if (flag & ZHAVELOCK)
dl->dl_namelock = 1;
if ((flag & ZSHARED) && ++dl->dl_sharecnt > 1 && dl->dl_namesize == 0) {
/*
* We're the second shared reference to dl. Make a copy of
* dl_name in case the first thread goes away before we do.
* Note that we initialize the new name before storing its
* pointer into dl_name, because the first thread may load
* dl->dl_name at any time. It'll either see the old value,
* which belongs to it, or the new shared copy; either is OK.
*/
dl->dl_namesize = strlen(dl->dl_name) + 1;
name = kmem_alloc(dl->dl_namesize, KM_SLEEP);
memcpy(name, dl->dl_name, dl->dl_namesize);
dl->dl_name = name;
}
mutex_exit(&dzp->z_lock);
/*
* We have a dirlock on the name. (Note that it is the dirlock,
* not the dzp's z_lock, that protects the name in the zap object.)
* See if there's an object by this name; if so, put a hold on it.
*/
if (flag & ZXATTR) {
error = sa_lookup(dzp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &zoid,
sizeof (zoid));
if (error == 0)
error = (zoid == 0 ? SET_ERROR(ENOENT) : 0);
} else {
error = zfs_match_find(zfsvfs, dzp, name, mt,
update, direntflags, realpnp, &zoid);
}
if (error) {
if (error != ENOENT || (flag & ZEXISTS)) {
zfs_dirent_unlock(dl);
return (error);
}
} else {
if (flag & ZNEW) {
zfs_dirent_unlock(dl);
return (SET_ERROR(EEXIST));
}
error = zfs_zget(zfsvfs, zoid, zpp);
if (error) {
zfs_dirent_unlock(dl);
return (error);
}
}
*dlpp = dl;
return (0);
}
/*
* Unlock this directory entry and wake anyone who was waiting for it.
*/
void
zfs_dirent_unlock(zfs_dirlock_t *dl)
{
znode_t *dzp = dl->dl_dzp;
zfs_dirlock_t **prev_dl, *cur_dl;
mutex_enter(&dzp->z_lock);
if (!dl->dl_namelock)
rw_exit(&dzp->z_name_lock);
if (dl->dl_sharecnt > 1) {
dl->dl_sharecnt--;
mutex_exit(&dzp->z_lock);
return;
}
prev_dl = &dzp->z_dirlocks;
while ((cur_dl = *prev_dl) != dl)
prev_dl = &cur_dl->dl_next;
*prev_dl = dl->dl_next;
cv_broadcast(&dl->dl_cv);
mutex_exit(&dzp->z_lock);
if (dl->dl_namesize != 0)
kmem_free(dl->dl_name, dl->dl_namesize);
cv_destroy(&dl->dl_cv);
kmem_free(dl, sizeof (*dl));
}
/*
* Look up an entry in a directory.
*
* NOTE: '.' and '..' are handled as special cases because
* no directory entries are actually stored for them. If this is
* the root of a filesystem, then '.zfs' is also treated as a
* special pseudo-directory.
*/
int
zfs_dirlook(znode_t *dzp, char *name, znode_t **zpp, int flags,
int *deflg, pathname_t *rpnp)
{
zfs_dirlock_t *dl;
znode_t *zp;
struct inode *ip;
int error = 0;
uint64_t parent;
if (name[0] == 0 || (name[0] == '.' && name[1] == 0)) {
*zpp = dzp;
zhold(*zpp);
} else if (name[0] == '.' && name[1] == '.' && name[2] == 0) {
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
/*
* If we are a snapshot mounted under .zfs, return
* the inode pointer for the snapshot directory.
*/
if ((error = sa_lookup(dzp->z_sa_hdl,
SA_ZPL_PARENT(zfsvfs), &parent, sizeof (parent))) != 0)
return (error);
if (parent == dzp->z_id && zfsvfs->z_parent != zfsvfs) {
error = zfsctl_root_lookup(zfsvfs->z_parent->z_ctldir,
"snapshot", &ip, 0, kcred, NULL, NULL);
*zpp = ITOZ(ip);
return (error);
}
rw_enter(&dzp->z_parent_lock, RW_READER);
error = zfs_zget(zfsvfs, parent, &zp);
if (error == 0)
*zpp = zp;
rw_exit(&dzp->z_parent_lock);
} else if (zfs_has_ctldir(dzp) && strcmp(name, ZFS_CTLDIR_NAME) == 0) {
ip = zfsctl_root(dzp);
*zpp = ITOZ(ip);
} else {
int zf;
zf = ZEXISTS | ZSHARED;
if (flags & FIGNORECASE)
zf |= ZCILOOK;
error = zfs_dirent_lock(&dl, dzp, name, &zp, zf, deflg, rpnp);
if (error == 0) {
*zpp = zp;
zfs_dirent_unlock(dl);
dzp->z_zn_prefetch = B_TRUE; /* enable prefetching */
}
rpnp = NULL;
}
if ((flags & FIGNORECASE) && rpnp && !error)
(void) strlcpy(rpnp->pn_buf, name, rpnp->pn_bufsize);
return (error);
}
/*
* unlinked Set (formerly known as the "delete queue") Error Handling
*
* When dealing with the unlinked set, we dmu_tx_hold_zap(), but we
* don't specify the name of the entry that we will be manipulating. We
* also fib and say that we won't be adding any new entries to the
* unlinked set, even though we might (this is to lower the minimum file
* size that can be deleted in a full filesystem). So on the small
* chance that the nlink list is using a fat zap (ie. has more than
* 2000 entries), we *may* not pre-read a block that's needed.
* Therefore it is remotely possible for some of the assertions
* regarding the unlinked set below to fail due to i/o error. On a
* nondebug system, this will result in the space being leaked.
*/
void
zfs_unlinked_add(znode_t *zp, dmu_tx_t *tx)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
ASSERT(zp->z_unlinked);
ASSERT(ZTOI(zp)->i_nlink == 0);
VERIFY3U(0, ==,
zap_add_int(zfsvfs->z_os, zfsvfs->z_unlinkedobj, zp->z_id, tx));
dataset_kstats_update_nunlinks_kstat(&zfsvfs->z_kstat, 1);
}
/*
* Clean up any znodes that had no links when we either crashed or
* (force) umounted the file system.
*/
static void
zfs_unlinked_drain_task(void *arg)
{
zfsvfs_t *zfsvfs = arg;
zap_cursor_t zc;
zap_attribute_t zap;
dmu_object_info_t doi;
znode_t *zp;
int error;
ASSERT3B(zfsvfs->z_draining, ==, B_TRUE);
/*
* Iterate over the contents of the unlinked set.
*/
for (zap_cursor_init(&zc, zfsvfs->z_os, zfsvfs->z_unlinkedobj);
zap_cursor_retrieve(&zc, &zap) == 0 && !zfsvfs->z_drain_cancel;
zap_cursor_advance(&zc)) {
/*
* See what kind of object we have in list
*/
error = dmu_object_info(zfsvfs->z_os,
zap.za_first_integer, &doi);
if (error != 0)
continue;
ASSERT((doi.doi_type == DMU_OT_PLAIN_FILE_CONTENTS) ||
(doi.doi_type == DMU_OT_DIRECTORY_CONTENTS));
/*
* We need to re-mark these list entries for deletion,
* so we pull them back into core and set zp->z_unlinked.
*/
error = zfs_zget(zfsvfs, zap.za_first_integer, &zp);
/*
* We may pick up znodes that are already marked for deletion.
* This could happen during the purge of an extended attribute
* directory. All we need to do is skip over them, since they
* are already in the system marked z_unlinked.
*/
if (error != 0)
continue;
zp->z_unlinked = B_TRUE;
/*
* zrele() decrements the znode's ref count and may cause
* it to be synchronously freed. We interrupt freeing
* of this znode by checking the return value of
* dmu_objset_zfs_unmounting() in dmu_free_long_range()
* when an unmount is requested.
*/
zrele(zp);
ASSERT3B(zfsvfs->z_unmounted, ==, B_FALSE);
}
zap_cursor_fini(&zc);
zfsvfs->z_draining = B_FALSE;
zfsvfs->z_drain_task = TASKQID_INVALID;
}
/*
* Sets z_draining then tries to dispatch async unlinked drain.
* If that fails executes synchronous unlinked drain.
*/
void
zfs_unlinked_drain(zfsvfs_t *zfsvfs)
{
ASSERT3B(zfsvfs->z_unmounted, ==, B_FALSE);
ASSERT3B(zfsvfs->z_draining, ==, B_FALSE);
zfsvfs->z_draining = B_TRUE;
zfsvfs->z_drain_cancel = B_FALSE;
zfsvfs->z_drain_task = taskq_dispatch(
dsl_pool_unlinked_drain_taskq(dmu_objset_pool(zfsvfs->z_os)),
zfs_unlinked_drain_task, zfsvfs, TQ_SLEEP);
if (zfsvfs->z_drain_task == TASKQID_INVALID) {
zfs_dbgmsg("async zfs_unlinked_drain dispatch failed");
zfs_unlinked_drain_task(zfsvfs);
}
}
/*
* Wait for the unlinked drain taskq task to stop. This will interrupt the
* unlinked set processing if it is in progress.
*/
void
zfs_unlinked_drain_stop_wait(zfsvfs_t *zfsvfs)
{
ASSERT3B(zfsvfs->z_unmounted, ==, B_FALSE);
if (zfsvfs->z_draining) {
zfsvfs->z_drain_cancel = B_TRUE;
taskq_cancel_id(dsl_pool_unlinked_drain_taskq(
dmu_objset_pool(zfsvfs->z_os)), zfsvfs->z_drain_task);
zfsvfs->z_drain_task = TASKQID_INVALID;
zfsvfs->z_draining = B_FALSE;
}
}
/*
* Delete the entire contents of a directory. Return a count
* of the number of entries that could not be deleted. If we encounter
* an error, return a count of at least one so that the directory stays
* in the unlinked set.
*
* NOTE: this function assumes that the directory is inactive,
* so there is no need to lock its entries before deletion.
* Also, it assumes the directory contents is *only* regular
* files.
*/
static int
zfs_purgedir(znode_t *dzp)
{
zap_cursor_t zc;
zap_attribute_t zap;
znode_t *xzp;
dmu_tx_t *tx;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zfs_dirlock_t dl;
int skipped = 0;
int error;
for (zap_cursor_init(&zc, zfsvfs->z_os, dzp->z_id);
(error = zap_cursor_retrieve(&zc, &zap)) == 0;
zap_cursor_advance(&zc)) {
error = zfs_zget(zfsvfs,
ZFS_DIRENT_OBJ(zap.za_first_integer), &xzp);
if (error) {
skipped += 1;
continue;
}
ASSERT(S_ISREG(ZTOI(xzp)->i_mode) ||
S_ISLNK(ZTOI(xzp)->i_mode));
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, dzp->z_id, FALSE, zap.za_name);
dmu_tx_hold_sa(tx, xzp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
/* Is this really needed ? */
zfs_sa_upgrade_txholds(tx, xzp);
dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
zfs_zrele_async(xzp);
skipped += 1;
continue;
}
memset(&dl, 0, sizeof (dl));
dl.dl_dzp = dzp;
dl.dl_name = zap.za_name;
error = zfs_link_destroy(&dl, xzp, tx, 0, NULL);
if (error)
skipped += 1;
dmu_tx_commit(tx);
zfs_zrele_async(xzp);
}
zap_cursor_fini(&zc);
if (error != ENOENT)
skipped += 1;
return (skipped);
}
void
zfs_rmnode(znode_t *zp)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
objset_t *os = zfsvfs->z_os;
znode_t *xzp = NULL;
dmu_tx_t *tx;
znode_hold_t *zh;
uint64_t z_id = zp->z_id;
uint64_t acl_obj;
uint64_t xattr_obj;
uint64_t links;
int error;
ASSERT(ZTOI(zp)->i_nlink == 0);
ASSERT(atomic_read(&ZTOI(zp)->i_count) == 0);
/*
* If this is an attribute directory, purge its contents.
*/
if (S_ISDIR(ZTOI(zp)->i_mode) && (zp->z_pflags & ZFS_XATTR)) {
if (zfs_purgedir(zp) != 0) {
/*
* Not enough space to delete some xattrs.
* Leave it in the unlinked set.
*/
zh = zfs_znode_hold_enter(zfsvfs, z_id);
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
return;
}
}
/*
* Free up all the data in the file. We don't do this for directories
* because we need truncate and remove to be in the same tx, like in
* zfs_znode_delete(). Otherwise, if we crash here we'll end up with
* an inconsistent truncated zap object in the delete queue. Note a
* truncated file is harmless since it only contains user data.
*/
if (S_ISREG(ZTOI(zp)->i_mode)) {
error = dmu_free_long_range(os, zp->z_id, 0, DMU_OBJECT_END);
if (error) {
/*
* Not enough space or we were interrupted by unmount.
* Leave the file in the unlinked set.
*/
zh = zfs_znode_hold_enter(zfsvfs, z_id);
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
return;
}
}
/*
* If the file has extended attributes, we're going to unlink
* the xattr dir.
*/
error = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj, sizeof (xattr_obj));
if (error == 0 && xattr_obj) {
error = zfs_zget(zfsvfs, xattr_obj, &xzp);
ASSERT(error == 0);
}
acl_obj = zfs_external_acl(zp);
/*
* Set up the final transaction.
*/
tx = dmu_tx_create(os);
dmu_tx_hold_free(tx, zp->z_id, 0, DMU_OBJECT_END);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
if (xzp) {
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, TRUE, NULL);
dmu_tx_hold_sa(tx, xzp->z_sa_hdl, B_FALSE);
}
if (acl_obj)
dmu_tx_hold_free(tx, acl_obj, 0, DMU_OBJECT_END);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
/*
* Not enough space to delete the file. Leave it in the
* unlinked set, leaking it until the fs is remounted (at
* which point we'll call zfs_unlinked_drain() to process it).
*/
dmu_tx_abort(tx);
zh = zfs_znode_hold_enter(zfsvfs, z_id);
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
goto out;
}
if (xzp) {
ASSERT(error == 0);
mutex_enter(&xzp->z_lock);
xzp->z_unlinked = B_TRUE; /* mark xzp for deletion */
clear_nlink(ZTOI(xzp)); /* no more links to it */
links = 0;
VERIFY(0 == sa_update(xzp->z_sa_hdl, SA_ZPL_LINKS(zfsvfs),
&links, sizeof (links), tx));
mutex_exit(&xzp->z_lock);
zfs_unlinked_add(xzp, tx);
}
mutex_enter(&os->os_dsl_dataset->ds_dir->dd_activity_lock);
/*
* Remove this znode from the unlinked set. If a has rollback has
* occurred while a file is open and unlinked. Then when the file
* is closed post rollback it will not exist in the rolled back
* version of the unlinked object.
*/
error = zap_remove_int(zfsvfs->z_os, zfsvfs->z_unlinkedobj,
zp->z_id, tx);
VERIFY(error == 0 || error == ENOENT);
uint64_t count;
if (zap_count(os, zfsvfs->z_unlinkedobj, &count) == 0 && count == 0) {
cv_broadcast(&os->os_dsl_dataset->ds_dir->dd_activity_cv);
}
mutex_exit(&os->os_dsl_dataset->ds_dir->dd_activity_lock);
dataset_kstats_update_nunlinked_kstat(&zfsvfs->z_kstat, 1);
zfs_znode_delete(zp, tx);
dmu_tx_commit(tx);
out:
if (xzp)
zfs_zrele_async(xzp);
}
static uint64_t
zfs_dirent(znode_t *zp, uint64_t mode)
{
uint64_t de = zp->z_id;
if (ZTOZSB(zp)->z_version >= ZPL_VERSION_DIRENT_TYPE)
de |= IFTODT(mode) << 60;
return (de);
}
/*
* Link zp into dl. Can fail in the following cases :
* - if zp has been unlinked.
* - if the number of entries with the same hash (aka. colliding entries)
* exceed the capacity of a leaf-block of fatzap and splitting of the
* leaf-block does not help.
*/
int
zfs_link_create(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag)
{
znode_t *dzp = dl->dl_dzp;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
uint64_t value;
int zp_is_dir = S_ISDIR(ZTOI(zp)->i_mode);
sa_bulk_attr_t bulk[5];
uint64_t mtime[2], ctime[2];
uint64_t links;
int count = 0;
int error;
mutex_enter(&zp->z_lock);
if (!(flag & ZRENAMING)) {
if (zp->z_unlinked) { /* no new links to unlinked zp */
ASSERT(!(flag & (ZNEW | ZEXISTS)));
mutex_exit(&zp->z_lock);
return (SET_ERROR(ENOENT));
}
if (!(flag & ZNEW)) {
/*
* ZNEW nodes come from zfs_mknode() where the link
* count has already been initialised
*/
inc_nlink(ZTOI(zp));
links = ZTOI(zp)->i_nlink;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs),
NULL, &links, sizeof (links));
}
}
value = zfs_dirent(zp, zp->z_mode);
error = zap_add(ZTOZSB(zp)->z_os, dzp->z_id, dl->dl_name, 8, 1,
&value, tx);
/*
* zap_add could fail to add the entry if it exceeds the capacity of the
* leaf-block and zap_leaf_split() failed to help.
* The caller of this routine is responsible for failing the transaction
* which will rollback the SA updates done above.
*/
if (error != 0) {
if (!(flag & ZRENAMING) && !(flag & ZNEW))
drop_nlink(ZTOI(zp));
mutex_exit(&zp->z_lock);
return (error);
}
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PARENT(zfsvfs), NULL,
&dzp->z_id, sizeof (dzp->z_id));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, sizeof (zp->z_pflags));
if (!(flag & ZNEW)) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
ctime, sizeof (ctime));
zfs_tstamp_update_setup(zp, STATE_CHANGED, mtime,
ctime);
}
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
ASSERT(error == 0);
mutex_exit(&zp->z_lock);
mutex_enter(&dzp->z_lock);
dzp->z_size++;
if (zp_is_dir)
inc_nlink(ZTOI(dzp));
links = ZTOI(dzp)->i_nlink;
count = 0;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL,
&dzp->z_size, sizeof (dzp->z_size));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL,
&links, sizeof (links));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL,
mtime, sizeof (mtime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
ctime, sizeof (ctime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&dzp->z_pflags, sizeof (dzp->z_pflags));
zfs_tstamp_update_setup(dzp, CONTENT_MODIFIED, mtime, ctime);
error = sa_bulk_update(dzp->z_sa_hdl, bulk, count, tx);
ASSERT(error == 0);
mutex_exit(&dzp->z_lock);
return (0);
}
/*
* The match type in the code for this function should conform to:
*
* ------------------------------------------------------------------------
* fs type | z_norm | lookup type | match type
* ---------|-------------|-------------|----------------------------------
* CS !norm | 0 | 0 | 0 (exact)
* CS norm | formX | 0 | MT_NORMALIZE
* CI !norm | upper | !ZCIEXACT | MT_NORMALIZE
* CI !norm | upper | ZCIEXACT | MT_NORMALIZE | MT_MATCH_CASE
* CI norm | upper|formX | !ZCIEXACT | MT_NORMALIZE
* CI norm | upper|formX | ZCIEXACT | MT_NORMALIZE | MT_MATCH_CASE
* CM !norm | upper | !ZCILOOK | MT_NORMALIZE | MT_MATCH_CASE
* CM !norm | upper | ZCILOOK | MT_NORMALIZE
* CM norm | upper|formX | !ZCILOOK | MT_NORMALIZE | MT_MATCH_CASE
* CM norm | upper|formX | ZCILOOK | MT_NORMALIZE
*
* Abbreviations:
* CS = Case Sensitive, CI = Case Insensitive, CM = Case Mixed
* upper = case folding set by fs type on creation (U8_TEXTPREP_TOUPPER)
* formX = unicode normalization form set on fs creation
*/
static int
zfs_dropname(zfs_dirlock_t *dl, znode_t *zp, znode_t *dzp, dmu_tx_t *tx,
int flag)
{
int error;
if (ZTOZSB(zp)->z_norm) {
matchtype_t mt = MT_NORMALIZE;
if ((ZTOZSB(zp)->z_case == ZFS_CASE_INSENSITIVE &&
(flag & ZCIEXACT)) ||
(ZTOZSB(zp)->z_case == ZFS_CASE_MIXED &&
!(flag & ZCILOOK))) {
mt |= MT_MATCH_CASE;
}
error = zap_remove_norm(ZTOZSB(zp)->z_os, dzp->z_id,
dl->dl_name, mt, tx);
} else {
error = zap_remove(ZTOZSB(zp)->z_os, dzp->z_id, dl->dl_name,
tx);
}
return (error);
}
static int
zfs_drop_nlink_locked(znode_t *zp, dmu_tx_t *tx, boolean_t *unlinkedp)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int zp_is_dir = S_ISDIR(ZTOI(zp)->i_mode);
boolean_t unlinked = B_FALSE;
sa_bulk_attr_t bulk[3];
uint64_t mtime[2], ctime[2];
uint64_t links;
int count = 0;
int error;
if (zp_is_dir && !zfs_dirempty(zp))
return (SET_ERROR(ENOTEMPTY));
if (ZTOI(zp)->i_nlink <= zp_is_dir) {
zfs_panic_recover("zfs: link count on %lu is %u, "
"should be at least %u", zp->z_id,
(int)ZTOI(zp)->i_nlink, zp_is_dir + 1);
set_nlink(ZTOI(zp), zp_is_dir + 1);
}
drop_nlink(ZTOI(zp));
if (ZTOI(zp)->i_nlink == zp_is_dir) {
zp->z_unlinked = B_TRUE;
clear_nlink(ZTOI(zp));
unlinked = B_TRUE;
} else {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs),
NULL, &ctime, sizeof (ctime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs),
NULL, &zp->z_pflags, sizeof (zp->z_pflags));
zfs_tstamp_update_setup(zp, STATE_CHANGED, mtime,
ctime);
}
links = ZTOI(zp)->i_nlink;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs),
NULL, &links, sizeof (links));
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
ASSERT3U(error, ==, 0);
if (unlinkedp != NULL)
*unlinkedp = unlinked;
else if (unlinked)
zfs_unlinked_add(zp, tx);
return (0);
}
/*
* Forcefully drop an nlink reference from (zp) and mark it for deletion if it
* was the last link. This *must* only be done to znodes which have already
* been zfs_link_destroy()'d with ZRENAMING. This is explicitly only used in
* the error path of zfs_rename(), where we have to correct the nlink count if
* we failed to link the target as well as failing to re-link the original
* znodes.
*/
int
zfs_drop_nlink(znode_t *zp, dmu_tx_t *tx, boolean_t *unlinkedp)
{
int error;
mutex_enter(&zp->z_lock);
error = zfs_drop_nlink_locked(zp, tx, unlinkedp);
mutex_exit(&zp->z_lock);
return (error);
}
/*
* Unlink zp from dl, and mark zp for deletion if this was the last link. Can
* fail if zp is a mount point (EBUSY) or a non-empty directory (ENOTEMPTY).
* If 'unlinkedp' is NULL, we put unlinked znodes on the unlinked list.
* If it's non-NULL, we use it to indicate whether the znode needs deletion,
* and it's the caller's job to do it.
*/
int
zfs_link_destroy(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag,
boolean_t *unlinkedp)
{
znode_t *dzp = dl->dl_dzp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
int zp_is_dir = S_ISDIR(ZTOI(zp)->i_mode);
boolean_t unlinked = B_FALSE;
sa_bulk_attr_t bulk[5];
uint64_t mtime[2], ctime[2];
uint64_t links;
int count = 0;
int error;
if (!(flag & ZRENAMING)) {
mutex_enter(&zp->z_lock);
if (zp_is_dir && !zfs_dirempty(zp)) {
mutex_exit(&zp->z_lock);
return (SET_ERROR(ENOTEMPTY));
}
/*
* If we get here, we are going to try to remove the object.
* First try removing the name from the directory; if that
* fails, return the error.
*/
error = zfs_dropname(dl, zp, dzp, tx, flag);
if (error != 0) {
mutex_exit(&zp->z_lock);
return (error);
}
/* The only error is !zfs_dirempty() and we checked earlier. */
error = zfs_drop_nlink_locked(zp, tx, &unlinked);
ASSERT3U(error, ==, 0);
mutex_exit(&zp->z_lock);
} else {
error = zfs_dropname(dl, zp, dzp, tx, flag);
if (error != 0)
return (error);
}
mutex_enter(&dzp->z_lock);
dzp->z_size--; /* one dirent removed */
if (zp_is_dir)
drop_nlink(ZTOI(dzp)); /* ".." link from zp */
links = ZTOI(dzp)->i_nlink;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs),
NULL, &links, sizeof (links));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs),
NULL, &dzp->z_size, sizeof (dzp->z_size));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs),
NULL, ctime, sizeof (ctime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs),
NULL, mtime, sizeof (mtime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs),
NULL, &dzp->z_pflags, sizeof (dzp->z_pflags));
zfs_tstamp_update_setup(dzp, CONTENT_MODIFIED, mtime, ctime);
error = sa_bulk_update(dzp->z_sa_hdl, bulk, count, tx);
ASSERT(error == 0);
mutex_exit(&dzp->z_lock);
if (unlinkedp != NULL)
*unlinkedp = unlinked;
else if (unlinked)
zfs_unlinked_add(zp, tx);
return (0);
}
/*
* Indicate whether the directory is empty. Works with or without z_lock
* held, but can only be consider a hint in the latter case. Returns true
* if only "." and ".." remain and there's no work in progress.
*
* The internal ZAP size, rather than zp->z_size, needs to be checked since
* some consumers (Lustre) do not strictly maintain an accurate SA_ZPL_SIZE.
*/
boolean_t
zfs_dirempty(znode_t *dzp)
{
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
uint64_t count;
int error;
if (dzp->z_dirlocks != NULL)
return (B_FALSE);
error = zap_count(zfsvfs->z_os, dzp->z_id, &count);
if (error != 0 || count != 0)
return (B_FALSE);
return (B_TRUE);
}
int
zfs_make_xattrdir(znode_t *zp, vattr_t *vap, znode_t **xzpp, cred_t *cr)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
znode_t *xzp;
dmu_tx_t *tx;
int error;
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
#ifdef ZFS_DEBUG
uint64_t parent;
#endif
*xzpp = NULL;
if ((error = zfs_acl_ids_create(zp, IS_XATTR, vap, cr, NULL,
- &acl_ids, kcred->user_ns)) != 0)
+ &acl_ids, zfs_init_idmap)) != 0)
return (error);
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, zp->z_projid)) {
zfs_acl_ids_free(&acl_ids);
return (SET_ERROR(EDQUOT));
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
return (error);
}
zfs_mknode(zp, vap, tx, cr, IS_XATTR, &xzp, &acl_ids);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
#ifdef ZFS_DEBUG
error = sa_lookup(xzp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs),
&parent, sizeof (parent));
ASSERT(error == 0 && parent == zp->z_id);
#endif
VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xzp->z_id,
sizeof (xzp->z_id), tx));
if (!zp->z_unlinked)
zfs_log_create(zfsvfs->z_log, tx, TX_MKXATTR, zp, xzp, "", NULL,
acl_ids.z_fuidp, vap);
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
*xzpp = xzp;
return (0);
}
/*
* Return a znode for the extended attribute directory for zp.
* ** If the directory does not already exist, it is created **
*
* IN: zp - znode to obtain attribute directory from
* cr - credentials of caller
* flags - flags from the VOP_LOOKUP call
*
* OUT: xipp - pointer to extended attribute znode
*
* RETURN: 0 on success
* error number on failure
*/
int
zfs_get_xattrdir(znode_t *zp, znode_t **xzpp, cred_t *cr, int flags)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
znode_t *xzp;
zfs_dirlock_t *dl;
vattr_t va;
int error;
top:
error = zfs_dirent_lock(&dl, zp, "", &xzp, ZXATTR, NULL, NULL);
if (error)
return (error);
if (xzp != NULL) {
*xzpp = xzp;
zfs_dirent_unlock(dl);
return (0);
}
if (!(flags & CREATE_XATTR_DIR)) {
zfs_dirent_unlock(dl);
return (SET_ERROR(ENOENT));
}
if (zfs_is_readonly(zfsvfs)) {
zfs_dirent_unlock(dl);
return (SET_ERROR(EROFS));
}
/*
* The ability to 'create' files in an attribute
* directory comes from the write_xattr permission on the base file.
*
* The ability to 'search' an attribute directory requires
* read_xattr permission on the base file.
*
* Once in a directory the ability to read/write attributes
* is controlled by the permissions on the attribute file.
*/
va.va_mask = ATTR_MODE | ATTR_UID | ATTR_GID;
va.va_mode = S_IFDIR | S_ISVTX | 0777;
zfs_fuid_map_ids(zp, cr, &va.va_uid, &va.va_gid);
va.va_dentry = NULL;
error = zfs_make_xattrdir(zp, &va, xzpp, cr);
zfs_dirent_unlock(dl);
if (error == ERESTART) {
/* NB: we already did dmu_tx_wait() if necessary */
goto top;
}
return (error);
}
/*
* Decide whether it is okay to remove within a sticky directory.
*
* In sticky directories, write access is not sufficient;
* you can remove entries from a directory only if:
*
* you own the directory,
* you own the entry,
* you have write access to the entry,
* or you are privileged (checked in secpolicy...).
*
* The function returns 0 if remove access is granted.
*/
int
zfs_sticky_remove_access(znode_t *zdp, znode_t *zp, cred_t *cr)
{
uid_t uid;
uid_t downer;
uid_t fowner;
zfsvfs_t *zfsvfs = ZTOZSB(zdp);
if (zfsvfs->z_replay)
return (0);
if ((zdp->z_mode & S_ISVTX) == 0)
return (0);
downer = zfs_fuid_map_id(zfsvfs, KUID_TO_SUID(ZTOI(zdp)->i_uid),
cr, ZFS_OWNER);
fowner = zfs_fuid_map_id(zfsvfs, KUID_TO_SUID(ZTOI(zp)->i_uid),
cr, ZFS_OWNER);
if ((uid = crgetuid(cr)) == downer || uid == fowner ||
zfs_zaccess(zp, ACE_WRITE_DATA, 0, B_FALSE, cr,
- kcred->user_ns) == 0)
+ zfs_init_idmap) == 0)
return (0);
else
return (secpolicy_vnode_remove(cr));
}
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_ioctl_os.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_ioctl_os.c
index f068f544f0ec..663474ea49ab 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_ioctl_os.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_ioctl_os.c
@@ -1,375 +1,379 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Portions Copyright 2011 Martin Matuska
* Copyright 2015, OmniTI Computer Consulting, Inc. All rights reserved.
* Portions Copyright 2012 Pawel Jakub Dawidek <pawel@dawidek.net>
* Copyright (c) 2014, 2016 Joyent, Inc. All rights reserved.
* Copyright 2016 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2014, Joyent, Inc. All rights reserved.
* Copyright (c) 2011, 2018 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
* Copyright (c) 2013 Steven Hartland. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2016 Toomas Soome <tsoome@me.com>
* Copyright (c) 2016 Actifio, Inc. All rights reserved.
* Copyright (c) 2018, loli10K <ezomori.nozomu@gmail.com>. All rights reserved.
* Copyright 2017 RackTop Systems.
* Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
* Copyright (c) 2019 Datto Inc.
* Copyright (c) 2021 Klara, Inc.
*/
#include <sys/types.h>
#include <sys/param.h>
#include <sys/errno.h>
#include <sys/uio.h>
#include <sys/file.h>
#include <sys/kmem.h>
#include <sys/stat.h>
#include <sys/zfs_ioctl.h>
#include <sys/zfs_vfsops.h>
#include <sys/zap.h>
#include <sys/spa.h>
#include <sys/nvpair.h>
#include <sys/fs/zfs.h>
#include <sys/zfs_ctldir.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_onexit.h>
#include <sys/zvol.h>
#include <sys/fm/util.h>
#include <sys/dsl_crypt.h>
#include <sys/crypto/icp.h>
#include <sys/zstd/zstd.h>
#include <sys/zfs_ioctl_impl.h>
#include <sys/zfs_sysfs.h>
#include <linux/miscdevice.h>
#include <linux/slab.h>
boolean_t
zfs_vfs_held(zfsvfs_t *zfsvfs)
{
return (zfsvfs->z_sb != NULL);
}
int
zfs_vfs_ref(zfsvfs_t **zfvp)
{
if (*zfvp == NULL || (*zfvp)->z_sb == NULL ||
!atomic_inc_not_zero(&((*zfvp)->z_sb->s_active))) {
return (SET_ERROR(ESRCH));
}
return (0);
}
void
zfs_vfs_rele(zfsvfs_t *zfsvfs)
{
deactivate_super(zfsvfs->z_sb);
}
void
zfsdev_private_set_state(void *priv, zfsdev_state_t *zs)
{
struct file *filp = priv;
filp->private_data = zs;
}
zfsdev_state_t *
zfsdev_private_get_state(void *priv)
{
struct file *filp = priv;
return (filp->private_data);
}
static int
zfsdev_open(struct inode *ino, struct file *filp)
{
int error;
mutex_enter(&zfsdev_state_lock);
error = zfsdev_state_init(filp);
mutex_exit(&zfsdev_state_lock);
return (-error);
}
static int
zfsdev_release(struct inode *ino, struct file *filp)
{
zfsdev_state_destroy(filp);
return (0);
}
static long
zfsdev_ioctl(struct file *filp, unsigned cmd, unsigned long arg)
{
uint_t vecnum;
zfs_cmd_t *zc;
int error, rc;
vecnum = cmd - ZFS_IOC_FIRST;
zc = vmem_zalloc(sizeof (zfs_cmd_t), KM_SLEEP);
if (ddi_copyin((void *)(uintptr_t)arg, zc, sizeof (zfs_cmd_t), 0)) {
error = -SET_ERROR(EFAULT);
goto out;
}
error = -zfsdev_ioctl_common(vecnum, zc, 0);
rc = ddi_copyout(zc, (void *)(uintptr_t)arg, sizeof (zfs_cmd_t), 0);
if (error == 0 && rc != 0)
error = -SET_ERROR(EFAULT);
out:
vmem_free(zc, sizeof (zfs_cmd_t));
return (error);
}
static int
zfs_ioc_userns_attach(zfs_cmd_t *zc)
{
int error;
if (zc == NULL)
return (SET_ERROR(EINVAL));
error = zone_dataset_attach(CRED(), zc->zc_name, zc->zc_cleanup_fd);
/*
* Translate ENOTTY to ZFS_ERR_NOT_USER_NAMESPACE as we just arrived
* back from the SPL layer, which does not know about ZFS_ERR_* errors.
* See the comment at the user_ns_get() function in spl-zone.c for
* details.
*/
if (error == ENOTTY)
error = ZFS_ERR_NOT_USER_NAMESPACE;
return (error);
}
static int
zfs_ioc_userns_detach(zfs_cmd_t *zc)
{
int error;
if (zc == NULL)
return (SET_ERROR(EINVAL));
error = zone_dataset_detach(CRED(), zc->zc_name, zc->zc_cleanup_fd);
/*
* See the comment in zfs_ioc_userns_attach() for details on what is
* going on here.
*/
if (error == ENOTTY)
error = ZFS_ERR_NOT_USER_NAMESPACE;
return (error);
}
uint64_t
zfs_max_nvlist_src_size_os(void)
{
if (zfs_max_nvlist_src_size != 0)
return (zfs_max_nvlist_src_size);
return (MIN(ptob(zfs_totalram_pages) / 4, 128 * 1024 * 1024));
}
/* Update the VFS's cache of mountpoint properties */
void
zfs_ioctl_update_mount_cache(const char *dsname)
{
}
void
zfs_ioctl_init_os(void)
{
zfs_ioctl_register_dataset_nolog(ZFS_IOC_USERNS_ATTACH,
zfs_ioc_userns_attach, zfs_secpolicy_config, POOL_CHECK_NONE);
zfs_ioctl_register_dataset_nolog(ZFS_IOC_USERNS_DETACH,
zfs_ioc_userns_detach, zfs_secpolicy_config, POOL_CHECK_NONE);
}
#ifdef CONFIG_COMPAT
static long
zfsdev_compat_ioctl(struct file *filp, unsigned cmd, unsigned long arg)
{
return (zfsdev_ioctl(filp, cmd, arg));
}
#else
#define zfsdev_compat_ioctl NULL
#endif
static const struct file_operations zfsdev_fops = {
.open = zfsdev_open,
.release = zfsdev_release,
.unlocked_ioctl = zfsdev_ioctl,
.compat_ioctl = zfsdev_compat_ioctl,
.owner = THIS_MODULE,
};
static struct miscdevice zfs_misc = {
.minor = ZFS_DEVICE_MINOR,
.name = ZFS_DRIVER,
.fops = &zfsdev_fops,
};
MODULE_ALIAS_MISCDEV(ZFS_DEVICE_MINOR);
MODULE_ALIAS("devname:zfs");
int
zfsdev_attach(void)
{
int error;
error = misc_register(&zfs_misc);
if (error == -EBUSY) {
/*
* Fallback to dynamic minor allocation in the event of a
* collision with a reserved minor in linux/miscdevice.h.
* In this case the kernel modules must be manually loaded.
*/
printk(KERN_INFO "ZFS: misc_register() with static minor %d "
"failed %d, retrying with MISC_DYNAMIC_MINOR\n",
ZFS_DEVICE_MINOR, error);
zfs_misc.minor = MISC_DYNAMIC_MINOR;
error = misc_register(&zfs_misc);
}
if (error)
printk(KERN_INFO "ZFS: misc_register() failed %d\n", error);
return (error);
}
void
zfsdev_detach(void)
{
misc_deregister(&zfs_misc);
}
#ifdef ZFS_DEBUG
#define ZFS_DEBUG_STR " (DEBUG mode)"
#else
#define ZFS_DEBUG_STR ""
#endif
+zidmap_t *zfs_init_idmap;
+
static int
openzfs_init_os(void)
{
int error;
if ((error = zfs_kmod_init()) != 0) {
printk(KERN_NOTICE "ZFS: Failed to Load ZFS Filesystem v%s-%s%s"
", rc = %d\n", ZFS_META_VERSION, ZFS_META_RELEASE,
ZFS_DEBUG_STR, error);
return (-error);
}
zfs_sysfs_init();
printk(KERN_NOTICE "ZFS: Loaded module v%s-%s%s, "
"ZFS pool version %s, ZFS filesystem version %s\n",
ZFS_META_VERSION, ZFS_META_RELEASE, ZFS_DEBUG_STR,
SPA_VERSION_STRING, ZPL_VERSION_STRING);
#ifndef CONFIG_FS_POSIX_ACL
printk(KERN_NOTICE "ZFS: Posix ACLs disabled by kernel\n");
#endif /* CONFIG_FS_POSIX_ACL */
+ zfs_init_idmap = (zidmap_t *)zfs_get_init_idmap();
+
return (0);
}
static void
openzfs_fini_os(void)
{
zfs_sysfs_fini();
zfs_kmod_fini();
printk(KERN_NOTICE "ZFS: Unloaded module v%s-%s%s\n",
ZFS_META_VERSION, ZFS_META_RELEASE, ZFS_DEBUG_STR);
}
extern int __init zcommon_init(void);
extern void zcommon_fini(void);
static int __init
openzfs_init(void)
{
int err;
if ((err = zcommon_init()) != 0)
goto zcommon_failed;
if ((err = icp_init()) != 0)
goto icp_failed;
if ((err = zstd_init()) != 0)
goto zstd_failed;
if ((err = openzfs_init_os()) != 0)
goto openzfs_os_failed;
return (0);
openzfs_os_failed:
zstd_fini();
zstd_failed:
icp_fini();
icp_failed:
zcommon_fini();
zcommon_failed:
return (err);
}
static void __exit
openzfs_fini(void)
{
openzfs_fini_os();
zstd_fini();
icp_fini();
zcommon_fini();
}
#if defined(_KERNEL)
module_init(openzfs_init);
module_exit(openzfs_fini);
#endif
MODULE_ALIAS("zavl");
MODULE_ALIAS("icp");
MODULE_ALIAS("zlua");
MODULE_ALIAS("znvpair");
MODULE_ALIAS("zunicode");
MODULE_ALIAS("zcommon");
MODULE_ALIAS("zzstd");
MODULE_DESCRIPTION("ZFS");
MODULE_AUTHOR(ZFS_META_AUTHOR);
MODULE_LICENSE("Dual MIT/GPL"); /* lua */
MODULE_LICENSE("Dual BSD/GPL"); /* zstd / misc */
MODULE_LICENSE(ZFS_META_LICENSE);
MODULE_VERSION(ZFS_META_VERSION "-" ZFS_META_RELEASE);
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c
index 2d9b27a90884..48945b8af8c1 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vfsops.c
@@ -1,2208 +1,2208 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
*/
/* Portions Copyright 2010 Robert Milkowski */
#include <sys/types.h>
#include <sys/param.h>
#include <sys/sysmacros.h>
#include <sys/kmem.h>
#include <sys/pathname.h>
#include <sys/vnode.h>
#include <sys/vfs.h>
#include <sys/mntent.h>
#include <sys/cmn_err.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_vnops.h>
#include <sys/zfs_dir.h>
#include <sys/zil.h>
#include <sys/fs/zfs.h>
#include <sys/dmu.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_deleg.h>
#include <sys/spa.h>
#include <sys/zap.h>
#include <sys/sa.h>
#include <sys/sa_impl.h>
#include <sys/policy.h>
#include <sys/atomic.h>
#include <sys/zfs_ioctl.h>
#include <sys/zfs_ctldir.h>
#include <sys/zfs_fuid.h>
#include <sys/zfs_quota.h>
#include <sys/sunddi.h>
#include <sys/dmu_objset.h>
#include <sys/dsl_dir.h>
#include <sys/objlist.h>
#include <sys/zpl.h>
#include <linux/vfs_compat.h>
#include "zfs_comutil.h"
enum {
TOKEN_RO,
TOKEN_RW,
TOKEN_SETUID,
TOKEN_NOSETUID,
TOKEN_EXEC,
TOKEN_NOEXEC,
TOKEN_DEVICES,
TOKEN_NODEVICES,
TOKEN_DIRXATTR,
TOKEN_SAXATTR,
TOKEN_XATTR,
TOKEN_NOXATTR,
TOKEN_ATIME,
TOKEN_NOATIME,
TOKEN_RELATIME,
TOKEN_NORELATIME,
TOKEN_NBMAND,
TOKEN_NONBMAND,
TOKEN_MNTPOINT,
TOKEN_LAST,
};
static const match_table_t zpl_tokens = {
{ TOKEN_RO, MNTOPT_RO },
{ TOKEN_RW, MNTOPT_RW },
{ TOKEN_SETUID, MNTOPT_SETUID },
{ TOKEN_NOSETUID, MNTOPT_NOSETUID },
{ TOKEN_EXEC, MNTOPT_EXEC },
{ TOKEN_NOEXEC, MNTOPT_NOEXEC },
{ TOKEN_DEVICES, MNTOPT_DEVICES },
{ TOKEN_NODEVICES, MNTOPT_NODEVICES },
{ TOKEN_DIRXATTR, MNTOPT_DIRXATTR },
{ TOKEN_SAXATTR, MNTOPT_SAXATTR },
{ TOKEN_XATTR, MNTOPT_XATTR },
{ TOKEN_NOXATTR, MNTOPT_NOXATTR },
{ TOKEN_ATIME, MNTOPT_ATIME },
{ TOKEN_NOATIME, MNTOPT_NOATIME },
{ TOKEN_RELATIME, MNTOPT_RELATIME },
{ TOKEN_NORELATIME, MNTOPT_NORELATIME },
{ TOKEN_NBMAND, MNTOPT_NBMAND },
{ TOKEN_NONBMAND, MNTOPT_NONBMAND },
{ TOKEN_MNTPOINT, MNTOPT_MNTPOINT "=%s" },
{ TOKEN_LAST, NULL },
};
static void
zfsvfs_vfs_free(vfs_t *vfsp)
{
if (vfsp != NULL) {
if (vfsp->vfs_mntpoint != NULL)
kmem_strfree(vfsp->vfs_mntpoint);
kmem_free(vfsp, sizeof (vfs_t));
}
}
static int
zfsvfs_parse_option(char *option, int token, substring_t *args, vfs_t *vfsp)
{
switch (token) {
case TOKEN_RO:
vfsp->vfs_readonly = B_TRUE;
vfsp->vfs_do_readonly = B_TRUE;
break;
case TOKEN_RW:
vfsp->vfs_readonly = B_FALSE;
vfsp->vfs_do_readonly = B_TRUE;
break;
case TOKEN_SETUID:
vfsp->vfs_setuid = B_TRUE;
vfsp->vfs_do_setuid = B_TRUE;
break;
case TOKEN_NOSETUID:
vfsp->vfs_setuid = B_FALSE;
vfsp->vfs_do_setuid = B_TRUE;
break;
case TOKEN_EXEC:
vfsp->vfs_exec = B_TRUE;
vfsp->vfs_do_exec = B_TRUE;
break;
case TOKEN_NOEXEC:
vfsp->vfs_exec = B_FALSE;
vfsp->vfs_do_exec = B_TRUE;
break;
case TOKEN_DEVICES:
vfsp->vfs_devices = B_TRUE;
vfsp->vfs_do_devices = B_TRUE;
break;
case TOKEN_NODEVICES:
vfsp->vfs_devices = B_FALSE;
vfsp->vfs_do_devices = B_TRUE;
break;
case TOKEN_DIRXATTR:
vfsp->vfs_xattr = ZFS_XATTR_DIR;
vfsp->vfs_do_xattr = B_TRUE;
break;
case TOKEN_SAXATTR:
vfsp->vfs_xattr = ZFS_XATTR_SA;
vfsp->vfs_do_xattr = B_TRUE;
break;
case TOKEN_XATTR:
vfsp->vfs_xattr = ZFS_XATTR_DIR;
vfsp->vfs_do_xattr = B_TRUE;
break;
case TOKEN_NOXATTR:
vfsp->vfs_xattr = ZFS_XATTR_OFF;
vfsp->vfs_do_xattr = B_TRUE;
break;
case TOKEN_ATIME:
vfsp->vfs_atime = B_TRUE;
vfsp->vfs_do_atime = B_TRUE;
break;
case TOKEN_NOATIME:
vfsp->vfs_atime = B_FALSE;
vfsp->vfs_do_atime = B_TRUE;
break;
case TOKEN_RELATIME:
vfsp->vfs_relatime = B_TRUE;
vfsp->vfs_do_relatime = B_TRUE;
break;
case TOKEN_NORELATIME:
vfsp->vfs_relatime = B_FALSE;
vfsp->vfs_do_relatime = B_TRUE;
break;
case TOKEN_NBMAND:
vfsp->vfs_nbmand = B_TRUE;
vfsp->vfs_do_nbmand = B_TRUE;
break;
case TOKEN_NONBMAND:
vfsp->vfs_nbmand = B_FALSE;
vfsp->vfs_do_nbmand = B_TRUE;
break;
case TOKEN_MNTPOINT:
vfsp->vfs_mntpoint = match_strdup(&args[0]);
if (vfsp->vfs_mntpoint == NULL)
return (SET_ERROR(ENOMEM));
break;
default:
break;
}
return (0);
}
/*
* Parse the raw mntopts and return a vfs_t describing the options.
*/
static int
zfsvfs_parse_options(char *mntopts, vfs_t **vfsp)
{
vfs_t *tmp_vfsp;
int error;
tmp_vfsp = kmem_zalloc(sizeof (vfs_t), KM_SLEEP);
if (mntopts != NULL) {
substring_t args[MAX_OPT_ARGS];
char *tmp_mntopts, *p, *t;
int token;
tmp_mntopts = t = kmem_strdup(mntopts);
if (tmp_mntopts == NULL)
return (SET_ERROR(ENOMEM));
while ((p = strsep(&t, ",")) != NULL) {
if (!*p)
continue;
args[0].to = args[0].from = NULL;
token = match_token(p, zpl_tokens, args);
error = zfsvfs_parse_option(p, token, args, tmp_vfsp);
if (error) {
kmem_strfree(tmp_mntopts);
zfsvfs_vfs_free(tmp_vfsp);
return (error);
}
}
kmem_strfree(tmp_mntopts);
}
*vfsp = tmp_vfsp;
return (0);
}
boolean_t
zfs_is_readonly(zfsvfs_t *zfsvfs)
{
return (!!(zfsvfs->z_sb->s_flags & SB_RDONLY));
}
int
zfs_sync(struct super_block *sb, int wait, cred_t *cr)
{
(void) cr;
zfsvfs_t *zfsvfs = sb->s_fs_info;
/*
* Semantically, the only requirement is that the sync be initiated.
* The DMU syncs out txgs frequently, so there's nothing to do.
*/
if (!wait)
return (0);
if (zfsvfs != NULL) {
/*
* Sync a specific filesystem.
*/
dsl_pool_t *dp;
int error;
if ((error = zfs_enter(zfsvfs, FTAG)) != 0)
return (error);
dp = dmu_objset_pool(zfsvfs->z_os);
/*
* If the system is shutting down, then skip any
* filesystems which may exist on a suspended pool.
*/
if (spa_suspended(dp->dp_spa)) {
zfs_exit(zfsvfs, FTAG);
return (0);
}
if (zfsvfs->z_log != NULL)
zil_commit(zfsvfs->z_log, 0);
zfs_exit(zfsvfs, FTAG);
} else {
/*
* Sync all ZFS filesystems. This is what happens when you
* run sync(1). Unlike other filesystems, ZFS honors the
* request by waiting for all pools to commit all dirty data.
*/
spa_sync_allpools();
}
return (0);
}
static void
atime_changed_cb(void *arg, uint64_t newval)
{
zfsvfs_t *zfsvfs = arg;
struct super_block *sb = zfsvfs->z_sb;
if (sb == NULL)
return;
/*
* Update SB_NOATIME bit in VFS super block. Since atime update is
* determined by atime_needs_update(), atime_needs_update() needs to
* return false if atime is turned off, and not unconditionally return
* false if atime is turned on.
*/
if (newval)
sb->s_flags &= ~SB_NOATIME;
else
sb->s_flags |= SB_NOATIME;
}
static void
relatime_changed_cb(void *arg, uint64_t newval)
{
((zfsvfs_t *)arg)->z_relatime = newval;
}
static void
xattr_changed_cb(void *arg, uint64_t newval)
{
zfsvfs_t *zfsvfs = arg;
if (newval == ZFS_XATTR_OFF) {
zfsvfs->z_flags &= ~ZSB_XATTR;
} else {
zfsvfs->z_flags |= ZSB_XATTR;
if (newval == ZFS_XATTR_SA)
zfsvfs->z_xattr_sa = B_TRUE;
else
zfsvfs->z_xattr_sa = B_FALSE;
}
}
static void
acltype_changed_cb(void *arg, uint64_t newval)
{
zfsvfs_t *zfsvfs = arg;
switch (newval) {
case ZFS_ACLTYPE_NFSV4:
case ZFS_ACLTYPE_OFF:
zfsvfs->z_acl_type = ZFS_ACLTYPE_OFF;
zfsvfs->z_sb->s_flags &= ~SB_POSIXACL;
break;
case ZFS_ACLTYPE_POSIX:
#ifdef CONFIG_FS_POSIX_ACL
zfsvfs->z_acl_type = ZFS_ACLTYPE_POSIX;
zfsvfs->z_sb->s_flags |= SB_POSIXACL;
#else
zfsvfs->z_acl_type = ZFS_ACLTYPE_OFF;
zfsvfs->z_sb->s_flags &= ~SB_POSIXACL;
#endif /* CONFIG_FS_POSIX_ACL */
break;
default:
break;
}
}
static void
blksz_changed_cb(void *arg, uint64_t newval)
{
zfsvfs_t *zfsvfs = arg;
ASSERT3U(newval, <=, spa_maxblocksize(dmu_objset_spa(zfsvfs->z_os)));
ASSERT3U(newval, >=, SPA_MINBLOCKSIZE);
ASSERT(ISP2(newval));
zfsvfs->z_max_blksz = newval;
}
static void
readonly_changed_cb(void *arg, uint64_t newval)
{
zfsvfs_t *zfsvfs = arg;
struct super_block *sb = zfsvfs->z_sb;
if (sb == NULL)
return;
if (newval)
sb->s_flags |= SB_RDONLY;
else
sb->s_flags &= ~SB_RDONLY;
}
static void
devices_changed_cb(void *arg, uint64_t newval)
{
}
static void
setuid_changed_cb(void *arg, uint64_t newval)
{
}
static void
exec_changed_cb(void *arg, uint64_t newval)
{
}
static void
nbmand_changed_cb(void *arg, uint64_t newval)
{
zfsvfs_t *zfsvfs = arg;
struct super_block *sb = zfsvfs->z_sb;
if (sb == NULL)
return;
if (newval == TRUE)
sb->s_flags |= SB_MANDLOCK;
else
sb->s_flags &= ~SB_MANDLOCK;
}
static void
snapdir_changed_cb(void *arg, uint64_t newval)
{
((zfsvfs_t *)arg)->z_show_ctldir = newval;
}
static void
acl_mode_changed_cb(void *arg, uint64_t newval)
{
zfsvfs_t *zfsvfs = arg;
zfsvfs->z_acl_mode = newval;
}
static void
acl_inherit_changed_cb(void *arg, uint64_t newval)
{
((zfsvfs_t *)arg)->z_acl_inherit = newval;
}
static int
zfs_register_callbacks(vfs_t *vfsp)
{
struct dsl_dataset *ds = NULL;
objset_t *os = NULL;
zfsvfs_t *zfsvfs = NULL;
int error = 0;
ASSERT(vfsp);
zfsvfs = vfsp->vfs_data;
ASSERT(zfsvfs);
os = zfsvfs->z_os;
/*
* The act of registering our callbacks will destroy any mount
* options we may have. In order to enable temporary overrides
* of mount options, we stash away the current values and
* restore them after we register the callbacks.
*/
if (zfs_is_readonly(zfsvfs) || !spa_writeable(dmu_objset_spa(os))) {
vfsp->vfs_do_readonly = B_TRUE;
vfsp->vfs_readonly = B_TRUE;
}
/*
* Register property callbacks.
*
* It would probably be fine to just check for i/o error from
* the first prop_register(), but I guess I like to go
* overboard...
*/
ds = dmu_objset_ds(os);
dsl_pool_config_enter(dmu_objset_pool(os), FTAG);
error = dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_ATIME), atime_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_RELATIME), relatime_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_XATTR), xattr_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_RECORDSIZE), blksz_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_READONLY), readonly_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_DEVICES), devices_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_SETUID), setuid_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_EXEC), exec_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_SNAPDIR), snapdir_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_ACLTYPE), acltype_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_ACLMODE), acl_mode_changed_cb, zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_ACLINHERIT), acl_inherit_changed_cb,
zfsvfs);
error = error ? error : dsl_prop_register(ds,
zfs_prop_to_name(ZFS_PROP_NBMAND), nbmand_changed_cb, zfsvfs);
dsl_pool_config_exit(dmu_objset_pool(os), FTAG);
if (error)
goto unregister;
/*
* Invoke our callbacks to restore temporary mount options.
*/
if (vfsp->vfs_do_readonly)
readonly_changed_cb(zfsvfs, vfsp->vfs_readonly);
if (vfsp->vfs_do_setuid)
setuid_changed_cb(zfsvfs, vfsp->vfs_setuid);
if (vfsp->vfs_do_exec)
exec_changed_cb(zfsvfs, vfsp->vfs_exec);
if (vfsp->vfs_do_devices)
devices_changed_cb(zfsvfs, vfsp->vfs_devices);
if (vfsp->vfs_do_xattr)
xattr_changed_cb(zfsvfs, vfsp->vfs_xattr);
if (vfsp->vfs_do_atime)
atime_changed_cb(zfsvfs, vfsp->vfs_atime);
if (vfsp->vfs_do_relatime)
relatime_changed_cb(zfsvfs, vfsp->vfs_relatime);
if (vfsp->vfs_do_nbmand)
nbmand_changed_cb(zfsvfs, vfsp->vfs_nbmand);
return (0);
unregister:
dsl_prop_unregister_all(ds, zfsvfs);
return (error);
}
/*
* Takes a dataset, a property, a value and that value's setpoint as
* found in the ZAP. Checks if the property has been changed in the vfs.
* If so, val and setpoint will be overwritten with updated content.
* Otherwise, they are left unchanged.
*/
int
zfs_get_temporary_prop(dsl_dataset_t *ds, zfs_prop_t zfs_prop, uint64_t *val,
char *setpoint)
{
int error;
zfsvfs_t *zfvp;
vfs_t *vfsp;
objset_t *os;
uint64_t tmp = *val;
error = dmu_objset_from_ds(ds, &os);
if (error != 0)
return (error);
if (dmu_objset_type(os) != DMU_OST_ZFS)
return (EINVAL);
mutex_enter(&os->os_user_ptr_lock);
zfvp = dmu_objset_get_user(os);
mutex_exit(&os->os_user_ptr_lock);
if (zfvp == NULL)
return (ESRCH);
vfsp = zfvp->z_vfs;
switch (zfs_prop) {
case ZFS_PROP_ATIME:
if (vfsp->vfs_do_atime)
tmp = vfsp->vfs_atime;
break;
case ZFS_PROP_RELATIME:
if (vfsp->vfs_do_relatime)
tmp = vfsp->vfs_relatime;
break;
case ZFS_PROP_DEVICES:
if (vfsp->vfs_do_devices)
tmp = vfsp->vfs_devices;
break;
case ZFS_PROP_EXEC:
if (vfsp->vfs_do_exec)
tmp = vfsp->vfs_exec;
break;
case ZFS_PROP_SETUID:
if (vfsp->vfs_do_setuid)
tmp = vfsp->vfs_setuid;
break;
case ZFS_PROP_READONLY:
if (vfsp->vfs_do_readonly)
tmp = vfsp->vfs_readonly;
break;
case ZFS_PROP_XATTR:
if (vfsp->vfs_do_xattr)
tmp = vfsp->vfs_xattr;
break;
case ZFS_PROP_NBMAND:
if (vfsp->vfs_do_nbmand)
tmp = vfsp->vfs_nbmand;
break;
default:
return (ENOENT);
}
if (tmp != *val) {
if (setpoint)
(void) strcpy(setpoint, "temporary");
*val = tmp;
}
return (0);
}
/*
* Associate this zfsvfs with the given objset, which must be owned.
* This will cache a bunch of on-disk state from the objset in the
* zfsvfs.
*/
static int
zfsvfs_init(zfsvfs_t *zfsvfs, objset_t *os)
{
int error;
uint64_t val;
zfsvfs->z_max_blksz = SPA_OLD_MAXBLOCKSIZE;
zfsvfs->z_show_ctldir = ZFS_SNAPDIR_VISIBLE;
zfsvfs->z_os = os;
error = zfs_get_zplprop(os, ZFS_PROP_VERSION, &zfsvfs->z_version);
if (error != 0)
return (error);
if (zfsvfs->z_version >
zfs_zpl_version_map(spa_version(dmu_objset_spa(os)))) {
(void) printk("Can't mount a version %lld file system "
"on a version %lld pool\n. Pool must be upgraded to mount "
"this file system.\n", (u_longlong_t)zfsvfs->z_version,
(u_longlong_t)spa_version(dmu_objset_spa(os)));
return (SET_ERROR(ENOTSUP));
}
error = zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &val);
if (error != 0)
return (error);
zfsvfs->z_norm = (int)val;
error = zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &val);
if (error != 0)
return (error);
zfsvfs->z_utf8 = (val != 0);
error = zfs_get_zplprop(os, ZFS_PROP_CASE, &val);
if (error != 0)
return (error);
zfsvfs->z_case = (uint_t)val;
if ((error = zfs_get_zplprop(os, ZFS_PROP_ACLTYPE, &val)) != 0)
return (error);
zfsvfs->z_acl_type = (uint_t)val;
/*
* Fold case on file systems that are always or sometimes case
* insensitive.
*/
if (zfsvfs->z_case == ZFS_CASE_INSENSITIVE ||
zfsvfs->z_case == ZFS_CASE_MIXED)
zfsvfs->z_norm |= U8_TEXTPREP_TOUPPER;
zfsvfs->z_use_fuids = USE_FUIDS(zfsvfs->z_version, zfsvfs->z_os);
zfsvfs->z_use_sa = USE_SA(zfsvfs->z_version, zfsvfs->z_os);
uint64_t sa_obj = 0;
if (zfsvfs->z_use_sa) {
/* should either have both of these objects or none */
error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_SA_ATTRS, 8, 1,
&sa_obj);
if (error != 0)
return (error);
error = zfs_get_zplprop(os, ZFS_PROP_XATTR, &val);
if ((error == 0) && (val == ZFS_XATTR_SA))
zfsvfs->z_xattr_sa = B_TRUE;
}
error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_ROOT_OBJ, 8, 1,
&zfsvfs->z_root);
if (error != 0)
return (error);
ASSERT(zfsvfs->z_root != 0);
error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_UNLINKED_SET, 8, 1,
&zfsvfs->z_unlinkedobj);
if (error != 0)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ,
zfs_userquota_prop_prefixes[ZFS_PROP_USERQUOTA],
8, 1, &zfsvfs->z_userquota_obj);
if (error == ENOENT)
zfsvfs->z_userquota_obj = 0;
else if (error != 0)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ,
zfs_userquota_prop_prefixes[ZFS_PROP_GROUPQUOTA],
8, 1, &zfsvfs->z_groupquota_obj);
if (error == ENOENT)
zfsvfs->z_groupquota_obj = 0;
else if (error != 0)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ,
zfs_userquota_prop_prefixes[ZFS_PROP_PROJECTQUOTA],
8, 1, &zfsvfs->z_projectquota_obj);
if (error == ENOENT)
zfsvfs->z_projectquota_obj = 0;
else if (error != 0)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ,
zfs_userquota_prop_prefixes[ZFS_PROP_USEROBJQUOTA],
8, 1, &zfsvfs->z_userobjquota_obj);
if (error == ENOENT)
zfsvfs->z_userobjquota_obj = 0;
else if (error != 0)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ,
zfs_userquota_prop_prefixes[ZFS_PROP_GROUPOBJQUOTA],
8, 1, &zfsvfs->z_groupobjquota_obj);
if (error == ENOENT)
zfsvfs->z_groupobjquota_obj = 0;
else if (error != 0)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ,
zfs_userquota_prop_prefixes[ZFS_PROP_PROJECTOBJQUOTA],
8, 1, &zfsvfs->z_projectobjquota_obj);
if (error == ENOENT)
zfsvfs->z_projectobjquota_obj = 0;
else if (error != 0)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES, 8, 1,
&zfsvfs->z_fuid_obj);
if (error == ENOENT)
zfsvfs->z_fuid_obj = 0;
else if (error != 0)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_SHARES_DIR, 8, 1,
&zfsvfs->z_shares_dir);
if (error == ENOENT)
zfsvfs->z_shares_dir = 0;
else if (error != 0)
return (error);
error = sa_setup(os, sa_obj, zfs_attr_table, ZPL_END,
&zfsvfs->z_attr_table);
if (error != 0)
return (error);
if (zfsvfs->z_version >= ZPL_VERSION_SA)
sa_register_update_callback(os, zfs_sa_upgrade);
return (0);
}
int
zfsvfs_create(const char *osname, boolean_t readonly, zfsvfs_t **zfvp)
{
objset_t *os;
zfsvfs_t *zfsvfs;
int error;
boolean_t ro = (readonly || (strchr(osname, '@') != NULL));
zfsvfs = kmem_zalloc(sizeof (zfsvfs_t), KM_SLEEP);
error = dmu_objset_own(osname, DMU_OST_ZFS, ro, B_TRUE, zfsvfs, &os);
if (error != 0) {
kmem_free(zfsvfs, sizeof (zfsvfs_t));
return (error);
}
error = zfsvfs_create_impl(zfvp, zfsvfs, os);
return (error);
}
/*
* Note: zfsvfs is assumed to be malloc'd, and will be freed by this function
* on a failure. Do not pass in a statically allocated zfsvfs.
*/
int
zfsvfs_create_impl(zfsvfs_t **zfvp, zfsvfs_t *zfsvfs, objset_t *os)
{
int error;
zfsvfs->z_vfs = NULL;
zfsvfs->z_sb = NULL;
zfsvfs->z_parent = zfsvfs;
mutex_init(&zfsvfs->z_znodes_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&zfsvfs->z_lock, NULL, MUTEX_DEFAULT, NULL);
list_create(&zfsvfs->z_all_znodes, sizeof (znode_t),
offsetof(znode_t, z_link_node));
ZFS_TEARDOWN_INIT(zfsvfs);
rw_init(&zfsvfs->z_teardown_inactive_lock, NULL, RW_DEFAULT, NULL);
rw_init(&zfsvfs->z_fuid_lock, NULL, RW_DEFAULT, NULL);
int size = MIN(1 << (highbit64(zfs_object_mutex_size) - 1),
ZFS_OBJ_MTX_MAX);
zfsvfs->z_hold_size = size;
zfsvfs->z_hold_trees = vmem_zalloc(sizeof (avl_tree_t) * size,
KM_SLEEP);
zfsvfs->z_hold_locks = vmem_zalloc(sizeof (kmutex_t) * size, KM_SLEEP);
for (int i = 0; i != size; i++) {
avl_create(&zfsvfs->z_hold_trees[i], zfs_znode_hold_compare,
sizeof (znode_hold_t), offsetof(znode_hold_t, zh_node));
mutex_init(&zfsvfs->z_hold_locks[i], NULL, MUTEX_DEFAULT, NULL);
}
error = zfsvfs_init(zfsvfs, os);
if (error != 0) {
dmu_objset_disown(os, B_TRUE, zfsvfs);
*zfvp = NULL;
zfsvfs_free(zfsvfs);
return (error);
}
zfsvfs->z_drain_task = TASKQID_INVALID;
zfsvfs->z_draining = B_FALSE;
zfsvfs->z_drain_cancel = B_TRUE;
*zfvp = zfsvfs;
return (0);
}
static int
zfsvfs_setup(zfsvfs_t *zfsvfs, boolean_t mounting)
{
int error;
boolean_t readonly = zfs_is_readonly(zfsvfs);
error = zfs_register_callbacks(zfsvfs->z_vfs);
if (error)
return (error);
/*
* If we are not mounting (ie: online recv), then we don't
* have to worry about replaying the log as we blocked all
* operations out since we closed the ZIL.
*/
if (mounting) {
ASSERT3P(zfsvfs->z_kstat.dk_kstats, ==, NULL);
error = dataset_kstats_create(&zfsvfs->z_kstat, zfsvfs->z_os);
if (error)
return (error);
zfsvfs->z_log = zil_open(zfsvfs->z_os, zfs_get_data,
&zfsvfs->z_kstat.dk_zil_sums);
/*
* During replay we remove the read only flag to
* allow replays to succeed.
*/
if (readonly != 0) {
readonly_changed_cb(zfsvfs, B_FALSE);
} else {
zap_stats_t zs;
if (zap_get_stats(zfsvfs->z_os, zfsvfs->z_unlinkedobj,
&zs) == 0) {
dataset_kstats_update_nunlinks_kstat(
&zfsvfs->z_kstat, zs.zs_num_entries);
dprintf_ds(zfsvfs->z_os->os_dsl_dataset,
"num_entries in unlinked set: %llu",
zs.zs_num_entries);
}
zfs_unlinked_drain(zfsvfs);
dsl_dir_t *dd = zfsvfs->z_os->os_dsl_dataset->ds_dir;
dd->dd_activity_cancelled = B_FALSE;
}
/*
* Parse and replay the intent log.
*
* Because of ziltest, this must be done after
* zfs_unlinked_drain(). (Further note: ziltest
* doesn't use readonly mounts, where
* zfs_unlinked_drain() isn't called.) This is because
* ziltest causes spa_sync() to think it's committed,
* but actually it is not, so the intent log contains
* many txg's worth of changes.
*
* In particular, if object N is in the unlinked set in
* the last txg to actually sync, then it could be
* actually freed in a later txg and then reallocated
* in a yet later txg. This would write a "create
* object N" record to the intent log. Normally, this
* would be fine because the spa_sync() would have
* written out the fact that object N is free, before
* we could write the "create object N" intent log
* record.
*
* But when we are in ziltest mode, we advance the "open
* txg" without actually spa_sync()-ing the changes to
* disk. So we would see that object N is still
* allocated and in the unlinked set, and there is an
* intent log record saying to allocate it.
*/
if (spa_writeable(dmu_objset_spa(zfsvfs->z_os))) {
if (zil_replay_disable) {
zil_destroy(zfsvfs->z_log, B_FALSE);
} else {
zfsvfs->z_replay = B_TRUE;
zil_replay(zfsvfs->z_os, zfsvfs,
zfs_replay_vector);
zfsvfs->z_replay = B_FALSE;
}
}
/* restore readonly bit */
if (readonly != 0)
readonly_changed_cb(zfsvfs, B_TRUE);
} else {
ASSERT3P(zfsvfs->z_kstat.dk_kstats, !=, NULL);
zfsvfs->z_log = zil_open(zfsvfs->z_os, zfs_get_data,
&zfsvfs->z_kstat.dk_zil_sums);
}
/*
* Set the objset user_ptr to track its zfsvfs.
*/
mutex_enter(&zfsvfs->z_os->os_user_ptr_lock);
dmu_objset_set_user(zfsvfs->z_os, zfsvfs);
mutex_exit(&zfsvfs->z_os->os_user_ptr_lock);
return (0);
}
void
zfsvfs_free(zfsvfs_t *zfsvfs)
{
int i, size = zfsvfs->z_hold_size;
zfs_fuid_destroy(zfsvfs);
mutex_destroy(&zfsvfs->z_znodes_lock);
mutex_destroy(&zfsvfs->z_lock);
list_destroy(&zfsvfs->z_all_znodes);
ZFS_TEARDOWN_DESTROY(zfsvfs);
rw_destroy(&zfsvfs->z_teardown_inactive_lock);
rw_destroy(&zfsvfs->z_fuid_lock);
for (i = 0; i != size; i++) {
avl_destroy(&zfsvfs->z_hold_trees[i]);
mutex_destroy(&zfsvfs->z_hold_locks[i]);
}
vmem_free(zfsvfs->z_hold_trees, sizeof (avl_tree_t) * size);
vmem_free(zfsvfs->z_hold_locks, sizeof (kmutex_t) * size);
zfsvfs_vfs_free(zfsvfs->z_vfs);
dataset_kstats_destroy(&zfsvfs->z_kstat);
kmem_free(zfsvfs, sizeof (zfsvfs_t));
}
static void
zfs_set_fuid_feature(zfsvfs_t *zfsvfs)
{
zfsvfs->z_use_fuids = USE_FUIDS(zfsvfs->z_version, zfsvfs->z_os);
zfsvfs->z_use_sa = USE_SA(zfsvfs->z_version, zfsvfs->z_os);
}
static void
zfs_unregister_callbacks(zfsvfs_t *zfsvfs)
{
objset_t *os = zfsvfs->z_os;
if (!dmu_objset_is_snapshot(os))
dsl_prop_unregister_all(dmu_objset_ds(os), zfsvfs);
}
#ifdef HAVE_MLSLABEL
/*
* Check that the hex label string is appropriate for the dataset being
* mounted into the global_zone proper.
*
* Return an error if the hex label string is not default or
* admin_low/admin_high. For admin_low labels, the corresponding
* dataset must be readonly.
*/
int
zfs_check_global_label(const char *dsname, const char *hexsl)
{
if (strcasecmp(hexsl, ZFS_MLSLABEL_DEFAULT) == 0)
return (0);
if (strcasecmp(hexsl, ADMIN_HIGH) == 0)
return (0);
if (strcasecmp(hexsl, ADMIN_LOW) == 0) {
/* must be readonly */
uint64_t rdonly;
if (dsl_prop_get_integer(dsname,
zfs_prop_to_name(ZFS_PROP_READONLY), &rdonly, NULL))
return (SET_ERROR(EACCES));
return (rdonly ? 0 : SET_ERROR(EACCES));
}
return (SET_ERROR(EACCES));
}
#endif /* HAVE_MLSLABEL */
static int
zfs_statfs_project(zfsvfs_t *zfsvfs, znode_t *zp, struct kstatfs *statp,
uint32_t bshift)
{
char buf[20 + DMU_OBJACCT_PREFIX_LEN];
uint64_t offset = DMU_OBJACCT_PREFIX_LEN;
uint64_t quota;
uint64_t used;
int err;
strlcpy(buf, DMU_OBJACCT_PREFIX, DMU_OBJACCT_PREFIX_LEN + 1);
err = zfs_id_to_fuidstr(zfsvfs, NULL, zp->z_projid, buf + offset,
sizeof (buf) - offset, B_FALSE);
if (err)
return (err);
if (zfsvfs->z_projectquota_obj == 0)
goto objs;
err = zap_lookup(zfsvfs->z_os, zfsvfs->z_projectquota_obj,
buf + offset, 8, 1, &quota);
if (err == ENOENT)
goto objs;
else if (err)
return (err);
err = zap_lookup(zfsvfs->z_os, DMU_PROJECTUSED_OBJECT,
buf + offset, 8, 1, &used);
if (unlikely(err == ENOENT)) {
uint32_t blksize;
u_longlong_t nblocks;
/*
* Quota accounting is async, so it is possible race case.
* There is at least one object with the given project ID.
*/
sa_object_size(zp->z_sa_hdl, &blksize, &nblocks);
if (unlikely(zp->z_blksz == 0))
blksize = zfsvfs->z_max_blksz;
used = blksize * nblocks;
} else if (err) {
return (err);
}
statp->f_blocks = quota >> bshift;
statp->f_bfree = (quota > used) ? ((quota - used) >> bshift) : 0;
statp->f_bavail = statp->f_bfree;
objs:
if (zfsvfs->z_projectobjquota_obj == 0)
return (0);
err = zap_lookup(zfsvfs->z_os, zfsvfs->z_projectobjquota_obj,
buf + offset, 8, 1, &quota);
if (err == ENOENT)
return (0);
else if (err)
return (err);
err = zap_lookup(zfsvfs->z_os, DMU_PROJECTUSED_OBJECT,
buf, 8, 1, &used);
if (unlikely(err == ENOENT)) {
/*
* Quota accounting is async, so it is possible race case.
* There is at least one object with the given project ID.
*/
used = 1;
} else if (err) {
return (err);
}
statp->f_files = quota;
statp->f_ffree = (quota > used) ? (quota - used) : 0;
return (0);
}
int
zfs_statvfs(struct inode *ip, struct kstatfs *statp)
{
zfsvfs_t *zfsvfs = ITOZSB(ip);
uint64_t refdbytes, availbytes, usedobjs, availobjs;
int err = 0;
if ((err = zfs_enter(zfsvfs, FTAG)) != 0)
return (err);
dmu_objset_space(zfsvfs->z_os,
&refdbytes, &availbytes, &usedobjs, &availobjs);
uint64_t fsid = dmu_objset_fsid_guid(zfsvfs->z_os);
/*
* The underlying storage pool actually uses multiple block
* size. Under Solaris frsize (fragment size) is reported as
* the smallest block size we support, and bsize (block size)
* as the filesystem's maximum block size. Unfortunately,
* under Linux the fragment size and block size are often used
* interchangeably. Thus we are forced to report both of them
* as the filesystem's maximum block size.
*/
statp->f_frsize = zfsvfs->z_max_blksz;
statp->f_bsize = zfsvfs->z_max_blksz;
uint32_t bshift = fls(statp->f_bsize) - 1;
/*
* The following report "total" blocks of various kinds in
* the file system, but reported in terms of f_bsize - the
* "preferred" size.
*/
/* Round up so we never have a filesystem using 0 blocks. */
refdbytes = P2ROUNDUP(refdbytes, statp->f_bsize);
statp->f_blocks = (refdbytes + availbytes) >> bshift;
statp->f_bfree = availbytes >> bshift;
statp->f_bavail = statp->f_bfree; /* no root reservation */
/*
* statvfs() should really be called statufs(), because it assumes
* static metadata. ZFS doesn't preallocate files, so the best
* we can do is report the max that could possibly fit in f_files,
* and that minus the number actually used in f_ffree.
* For f_ffree, report the smaller of the number of objects available
* and the number of blocks (each object will take at least a block).
*/
statp->f_ffree = MIN(availobjs, availbytes >> DNODE_SHIFT);
statp->f_files = statp->f_ffree + usedobjs;
statp->f_fsid.val[0] = (uint32_t)fsid;
statp->f_fsid.val[1] = (uint32_t)(fsid >> 32);
statp->f_type = ZFS_SUPER_MAGIC;
statp->f_namelen = MAXNAMELEN - 1;
/*
* We have all of 40 characters to stuff a string here.
* Is there anything useful we could/should provide?
*/
memset(statp->f_spare, 0, sizeof (statp->f_spare));
if (dmu_objset_projectquota_enabled(zfsvfs->z_os) &&
dmu_objset_projectquota_present(zfsvfs->z_os)) {
znode_t *zp = ITOZ(ip);
if (zp->z_pflags & ZFS_PROJINHERIT && zp->z_projid &&
zpl_is_valid_projid(zp->z_projid))
err = zfs_statfs_project(zfsvfs, zp, statp, bshift);
}
zfs_exit(zfsvfs, FTAG);
return (err);
}
static int
zfs_root(zfsvfs_t *zfsvfs, struct inode **ipp)
{
znode_t *rootzp;
int error;
if ((error = zfs_enter(zfsvfs, FTAG)) != 0)
return (error);
error = zfs_zget(zfsvfs, zfsvfs->z_root, &rootzp);
if (error == 0)
*ipp = ZTOI(rootzp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Linux kernels older than 3.1 do not support a per-filesystem shrinker.
* To accommodate this we must improvise and manually walk the list of znodes
* attempting to prune dentries in order to be able to drop the inodes.
*
* To avoid scanning the same znodes multiple times they are always rotated
* to the end of the z_all_znodes list. New znodes are inserted at the
* end of the list so we're always scanning the oldest znodes first.
*/
static int
zfs_prune_aliases(zfsvfs_t *zfsvfs, unsigned long nr_to_scan)
{
znode_t **zp_array, *zp;
int max_array = MIN(nr_to_scan, PAGE_SIZE * 8 / sizeof (znode_t *));
int objects = 0;
int i = 0, j = 0;
- zp_array = kmem_zalloc(max_array * sizeof (znode_t *), KM_SLEEP);
+ zp_array = vmem_zalloc(max_array * sizeof (znode_t *), KM_SLEEP);
mutex_enter(&zfsvfs->z_znodes_lock);
while ((zp = list_head(&zfsvfs->z_all_znodes)) != NULL) {
if ((i++ > nr_to_scan) || (j >= max_array))
break;
ASSERT(list_link_active(&zp->z_link_node));
list_remove(&zfsvfs->z_all_znodes, zp);
list_insert_tail(&zfsvfs->z_all_znodes, zp);
/* Skip active znodes and .zfs entries */
if (MUTEX_HELD(&zp->z_lock) || zp->z_is_ctldir)
continue;
if (igrab(ZTOI(zp)) == NULL)
continue;
zp_array[j] = zp;
j++;
}
mutex_exit(&zfsvfs->z_znodes_lock);
for (i = 0; i < j; i++) {
zp = zp_array[i];
ASSERT3P(zp, !=, NULL);
d_prune_aliases(ZTOI(zp));
if (atomic_read(&ZTOI(zp)->i_count) == 1)
objects++;
zrele(zp);
}
- kmem_free(zp_array, max_array * sizeof (znode_t *));
+ vmem_free(zp_array, max_array * sizeof (znode_t *));
return (objects);
}
/*
* The ARC has requested that the filesystem drop entries from the dentry
* and inode caches. This can occur when the ARC needs to free meta data
* blocks but can't because they are all pinned by entries in these caches.
*/
int
zfs_prune(struct super_block *sb, unsigned long nr_to_scan, int *objects)
{
zfsvfs_t *zfsvfs = sb->s_fs_info;
int error = 0;
struct shrinker *shrinker = &sb->s_shrink;
struct shrink_control sc = {
.nr_to_scan = nr_to_scan,
.gfp_mask = GFP_KERNEL,
};
if ((error = zfs_enter(zfsvfs, FTAG)) != 0)
return (error);
#if defined(HAVE_SPLIT_SHRINKER_CALLBACK) && \
defined(SHRINK_CONTROL_HAS_NID) && \
defined(SHRINKER_NUMA_AWARE)
if (sb->s_shrink.flags & SHRINKER_NUMA_AWARE) {
*objects = 0;
for_each_online_node(sc.nid) {
*objects += (*shrinker->scan_objects)(shrinker, &sc);
/*
* reset sc.nr_to_scan, modified by
* scan_objects == super_cache_scan
*/
sc.nr_to_scan = nr_to_scan;
}
} else {
*objects = (*shrinker->scan_objects)(shrinker, &sc);
}
#elif defined(HAVE_SPLIT_SHRINKER_CALLBACK)
*objects = (*shrinker->scan_objects)(shrinker, &sc);
#elif defined(HAVE_SINGLE_SHRINKER_CALLBACK)
*objects = (*shrinker->shrink)(shrinker, &sc);
#elif defined(HAVE_D_PRUNE_ALIASES)
#define D_PRUNE_ALIASES_IS_DEFAULT
*objects = zfs_prune_aliases(zfsvfs, nr_to_scan);
#else
#error "No available dentry and inode cache pruning mechanism."
#endif
#if defined(HAVE_D_PRUNE_ALIASES) && !defined(D_PRUNE_ALIASES_IS_DEFAULT)
#undef D_PRUNE_ALIASES_IS_DEFAULT
/*
* Fall back to zfs_prune_aliases if the kernel's per-superblock
* shrinker couldn't free anything, possibly due to the inodes being
* allocated in a different memcg.
*/
if (*objects == 0)
*objects = zfs_prune_aliases(zfsvfs, nr_to_scan);
#endif
zfs_exit(zfsvfs, FTAG);
dprintf_ds(zfsvfs->z_os->os_dsl_dataset,
"pruning, nr_to_scan=%lu objects=%d error=%d\n",
nr_to_scan, *objects, error);
return (error);
}
/*
* Teardown the zfsvfs_t.
*
* Note, if 'unmounting' is FALSE, we return with the 'z_teardown_lock'
* and 'z_teardown_inactive_lock' held.
*/
static int
zfsvfs_teardown(zfsvfs_t *zfsvfs, boolean_t unmounting)
{
znode_t *zp;
zfs_unlinked_drain_stop_wait(zfsvfs);
/*
* If someone has not already unmounted this file system,
* drain the zrele_taskq to ensure all active references to the
* zfsvfs_t have been handled only then can it be safely destroyed.
*/
if (zfsvfs->z_os) {
/*
* If we're unmounting we have to wait for the list to
* drain completely.
*
* If we're not unmounting there's no guarantee the list
* will drain completely, but iputs run from the taskq
* may add the parents of dir-based xattrs to the taskq
* so we want to wait for these.
*
* We can safely read z_nr_znodes without locking because the
* VFS has already blocked operations which add to the
* z_all_znodes list and thus increment z_nr_znodes.
*/
int round = 0;
while (zfsvfs->z_nr_znodes > 0) {
taskq_wait_outstanding(dsl_pool_zrele_taskq(
dmu_objset_pool(zfsvfs->z_os)), 0);
if (++round > 1 && !unmounting)
break;
}
}
ZFS_TEARDOWN_ENTER_WRITE(zfsvfs, FTAG);
if (!unmounting) {
/*
* We purge the parent filesystem's super block as the
* parent filesystem and all of its snapshots have their
* inode's super block set to the parent's filesystem's
* super block. Note, 'z_parent' is self referential
* for non-snapshots.
*/
shrink_dcache_sb(zfsvfs->z_parent->z_sb);
}
/*
* Close the zil. NB: Can't close the zil while zfs_inactive
* threads are blocked as zil_close can call zfs_inactive.
*/
if (zfsvfs->z_log) {
zil_close(zfsvfs->z_log);
zfsvfs->z_log = NULL;
}
rw_enter(&zfsvfs->z_teardown_inactive_lock, RW_WRITER);
/*
* If we are not unmounting (ie: online recv) and someone already
* unmounted this file system while we were doing the switcheroo,
* or a reopen of z_os failed then just bail out now.
*/
if (!unmounting && (zfsvfs->z_unmounted || zfsvfs->z_os == NULL)) {
rw_exit(&zfsvfs->z_teardown_inactive_lock);
ZFS_TEARDOWN_EXIT(zfsvfs, FTAG);
return (SET_ERROR(EIO));
}
/*
* At this point there are no VFS ops active, and any new VFS ops
* will fail with EIO since we have z_teardown_lock for writer (only
* relevant for forced unmount).
*
* Release all holds on dbufs. We also grab an extra reference to all
* the remaining inodes so that the kernel does not attempt to free
* any inodes of a suspended fs. This can cause deadlocks since the
* zfs_resume_fs() process may involve starting threads, which might
* attempt to free unreferenced inodes to free up memory for the new
* thread.
*/
if (!unmounting) {
mutex_enter(&zfsvfs->z_znodes_lock);
for (zp = list_head(&zfsvfs->z_all_znodes); zp != NULL;
zp = list_next(&zfsvfs->z_all_znodes, zp)) {
if (zp->z_sa_hdl)
zfs_znode_dmu_fini(zp);
if (igrab(ZTOI(zp)) != NULL)
zp->z_suspended = B_TRUE;
}
mutex_exit(&zfsvfs->z_znodes_lock);
}
/*
* If we are unmounting, set the unmounted flag and let new VFS ops
* unblock. zfs_inactive will have the unmounted behavior, and all
* other VFS ops will fail with EIO.
*/
if (unmounting) {
zfsvfs->z_unmounted = B_TRUE;
rw_exit(&zfsvfs->z_teardown_inactive_lock);
ZFS_TEARDOWN_EXIT(zfsvfs, FTAG);
}
/*
* z_os will be NULL if there was an error in attempting to reopen
* zfsvfs, so just return as the properties had already been
*
* unregistered and cached data had been evicted before.
*/
if (zfsvfs->z_os == NULL)
return (0);
/*
* Unregister properties.
*/
zfs_unregister_callbacks(zfsvfs);
/*
* Evict cached data. We must write out any dirty data before
* disowning the dataset.
*/
objset_t *os = zfsvfs->z_os;
boolean_t os_dirty = B_FALSE;
for (int t = 0; t < TXG_SIZE; t++) {
if (dmu_objset_is_dirty(os, t)) {
os_dirty = B_TRUE;
break;
}
}
if (!zfs_is_readonly(zfsvfs) && os_dirty) {
txg_wait_synced(dmu_objset_pool(zfsvfs->z_os), 0);
}
dmu_objset_evict_dbufs(zfsvfs->z_os);
dsl_dir_t *dd = os->os_dsl_dataset->ds_dir;
dsl_dir_cancel_waiters(dd);
return (0);
}
#if defined(HAVE_SUPER_SETUP_BDI_NAME)
atomic_long_t zfs_bdi_seq = ATOMIC_LONG_INIT(0);
#endif
int
zfs_domount(struct super_block *sb, zfs_mnt_t *zm, int silent)
{
const char *osname = zm->mnt_osname;
struct inode *root_inode = NULL;
uint64_t recordsize;
int error = 0;
zfsvfs_t *zfsvfs = NULL;
vfs_t *vfs = NULL;
int canwrite;
int dataset_visible_zone;
ASSERT(zm);
ASSERT(osname);
dataset_visible_zone = zone_dataset_visible(osname, &canwrite);
/*
* Refuse to mount a filesystem if we are in a namespace and the
* dataset is not visible or writable in that namespace.
*/
if (!INGLOBALZONE(curproc) &&
(!dataset_visible_zone || !canwrite)) {
return (SET_ERROR(EPERM));
}
error = zfsvfs_parse_options(zm->mnt_data, &vfs);
if (error)
return (error);
/*
* If a non-writable filesystem is being mounted without the
* read-only flag, pretend it was set, as done for snapshots.
*/
if (!canwrite)
vfs->vfs_readonly = true;
error = zfsvfs_create(osname, vfs->vfs_readonly, &zfsvfs);
if (error) {
zfsvfs_vfs_free(vfs);
goto out;
}
if ((error = dsl_prop_get_integer(osname, "recordsize",
&recordsize, NULL))) {
zfsvfs_vfs_free(vfs);
goto out;
}
vfs->vfs_data = zfsvfs;
zfsvfs->z_vfs = vfs;
zfsvfs->z_sb = sb;
sb->s_fs_info = zfsvfs;
sb->s_magic = ZFS_SUPER_MAGIC;
sb->s_maxbytes = MAX_LFS_FILESIZE;
sb->s_time_gran = 1;
sb->s_blocksize = recordsize;
sb->s_blocksize_bits = ilog2(recordsize);
error = -zpl_bdi_setup(sb, "zfs");
if (error)
goto out;
sb->s_bdi->ra_pages = 0;
/* Set callback operations for the file system. */
sb->s_op = &zpl_super_operations;
sb->s_xattr = zpl_xattr_handlers;
sb->s_export_op = &zpl_export_operations;
/* Set features for file system. */
zfs_set_fuid_feature(zfsvfs);
if (dmu_objset_is_snapshot(zfsvfs->z_os)) {
uint64_t pval;
atime_changed_cb(zfsvfs, B_FALSE);
readonly_changed_cb(zfsvfs, B_TRUE);
if ((error = dsl_prop_get_integer(osname,
"xattr", &pval, NULL)))
goto out;
xattr_changed_cb(zfsvfs, pval);
if ((error = dsl_prop_get_integer(osname,
"acltype", &pval, NULL)))
goto out;
acltype_changed_cb(zfsvfs, pval);
zfsvfs->z_issnap = B_TRUE;
zfsvfs->z_os->os_sync = ZFS_SYNC_DISABLED;
zfsvfs->z_snap_defer_time = jiffies;
mutex_enter(&zfsvfs->z_os->os_user_ptr_lock);
dmu_objset_set_user(zfsvfs->z_os, zfsvfs);
mutex_exit(&zfsvfs->z_os->os_user_ptr_lock);
} else {
if ((error = zfsvfs_setup(zfsvfs, B_TRUE)))
goto out;
}
/* Allocate a root inode for the filesystem. */
error = zfs_root(zfsvfs, &root_inode);
if (error) {
(void) zfs_umount(sb);
zfsvfs = NULL; /* avoid double-free; first in zfs_umount */
goto out;
}
/* Allocate a root dentry for the filesystem */
sb->s_root = d_make_root(root_inode);
if (sb->s_root == NULL) {
(void) zfs_umount(sb);
zfsvfs = NULL; /* avoid double-free; first in zfs_umount */
error = SET_ERROR(ENOMEM);
goto out;
}
if (!zfsvfs->z_issnap)
zfsctl_create(zfsvfs);
zfsvfs->z_arc_prune = arc_add_prune_callback(zpl_prune_sb, sb);
out:
if (error) {
if (zfsvfs != NULL) {
dmu_objset_disown(zfsvfs->z_os, B_TRUE, zfsvfs);
zfsvfs_free(zfsvfs);
}
/*
* make sure we don't have dangling sb->s_fs_info which
* zfs_preumount will use.
*/
sb->s_fs_info = NULL;
}
return (error);
}
/*
* Called when an unmount is requested and certain sanity checks have
* already passed. At this point no dentries or inodes have been reclaimed
* from their respective caches. We drop the extra reference on the .zfs
* control directory to allow everything to be reclaimed. All snapshots
* must already have been unmounted to reach this point.
*/
void
zfs_preumount(struct super_block *sb)
{
zfsvfs_t *zfsvfs = sb->s_fs_info;
/* zfsvfs is NULL when zfs_domount fails during mount */
if (zfsvfs) {
zfs_unlinked_drain_stop_wait(zfsvfs);
zfsctl_destroy(sb->s_fs_info);
/*
* Wait for zrele_async before entering evict_inodes in
* generic_shutdown_super. The reason we must finish before
* evict_inodes is when lazytime is on, or when zfs_purgedir
* calls zfs_zget, zrele would bump i_count from 0 to 1. This
* would race with the i_count check in evict_inodes. This means
* it could destroy the inode while we are still using it.
*
* We wait for two passes. xattr directories in the first pass
* may add xattr entries in zfs_purgedir, so in the second pass
* we wait for them. We don't use taskq_wait here because it is
* a pool wide taskq. Other mounted filesystems can constantly
* do zrele_async and there's no guarantee when taskq will be
* empty.
*/
taskq_wait_outstanding(dsl_pool_zrele_taskq(
dmu_objset_pool(zfsvfs->z_os)), 0);
taskq_wait_outstanding(dsl_pool_zrele_taskq(
dmu_objset_pool(zfsvfs->z_os)), 0);
}
}
/*
* Called once all other unmount released tear down has occurred.
* It is our responsibility to release any remaining infrastructure.
*/
int
zfs_umount(struct super_block *sb)
{
zfsvfs_t *zfsvfs = sb->s_fs_info;
objset_t *os;
if (zfsvfs->z_arc_prune != NULL)
arc_remove_prune_callback(zfsvfs->z_arc_prune);
VERIFY(zfsvfs_teardown(zfsvfs, B_TRUE) == 0);
os = zfsvfs->z_os;
zpl_bdi_destroy(sb);
/*
* z_os will be NULL if there was an error in
* attempting to reopen zfsvfs.
*/
if (os != NULL) {
/*
* Unset the objset user_ptr.
*/
mutex_enter(&os->os_user_ptr_lock);
dmu_objset_set_user(os, NULL);
mutex_exit(&os->os_user_ptr_lock);
/*
* Finally release the objset
*/
dmu_objset_disown(os, B_TRUE, zfsvfs);
}
zfsvfs_free(zfsvfs);
return (0);
}
int
zfs_remount(struct super_block *sb, int *flags, zfs_mnt_t *zm)
{
zfsvfs_t *zfsvfs = sb->s_fs_info;
vfs_t *vfsp;
boolean_t issnap = dmu_objset_is_snapshot(zfsvfs->z_os);
int error;
if ((issnap || !spa_writeable(dmu_objset_spa(zfsvfs->z_os))) &&
!(*flags & SB_RDONLY)) {
*flags |= SB_RDONLY;
return (EROFS);
}
error = zfsvfs_parse_options(zm->mnt_data, &vfsp);
if (error)
return (error);
if (!zfs_is_readonly(zfsvfs) && (*flags & SB_RDONLY))
txg_wait_synced(dmu_objset_pool(zfsvfs->z_os), 0);
zfs_unregister_callbacks(zfsvfs);
zfsvfs_vfs_free(zfsvfs->z_vfs);
vfsp->vfs_data = zfsvfs;
zfsvfs->z_vfs = vfsp;
if (!issnap)
(void) zfs_register_callbacks(vfsp);
return (error);
}
int
zfs_vget(struct super_block *sb, struct inode **ipp, fid_t *fidp)
{
zfsvfs_t *zfsvfs = sb->s_fs_info;
znode_t *zp;
uint64_t object = 0;
uint64_t fid_gen = 0;
uint64_t gen_mask;
uint64_t zp_gen;
int i, err;
*ipp = NULL;
if (fidp->fid_len == SHORT_FID_LEN || fidp->fid_len == LONG_FID_LEN) {
zfid_short_t *zfid = (zfid_short_t *)fidp;
for (i = 0; i < sizeof (zfid->zf_object); i++)
object |= ((uint64_t)zfid->zf_object[i]) << (8 * i);
for (i = 0; i < sizeof (zfid->zf_gen); i++)
fid_gen |= ((uint64_t)zfid->zf_gen[i]) << (8 * i);
} else {
return (SET_ERROR(EINVAL));
}
/* LONG_FID_LEN means snapdirs */
if (fidp->fid_len == LONG_FID_LEN) {
zfid_long_t *zlfid = (zfid_long_t *)fidp;
uint64_t objsetid = 0;
uint64_t setgen = 0;
for (i = 0; i < sizeof (zlfid->zf_setid); i++)
objsetid |= ((uint64_t)zlfid->zf_setid[i]) << (8 * i);
for (i = 0; i < sizeof (zlfid->zf_setgen); i++)
setgen |= ((uint64_t)zlfid->zf_setgen[i]) << (8 * i);
if (objsetid != ZFSCTL_INO_SNAPDIRS - object) {
dprintf("snapdir fid: objsetid (%llu) != "
"ZFSCTL_INO_SNAPDIRS (%llu) - object (%llu)\n",
objsetid, ZFSCTL_INO_SNAPDIRS, object);
return (SET_ERROR(EINVAL));
}
if (fid_gen > 1 || setgen != 0) {
dprintf("snapdir fid: fid_gen (%llu) and setgen "
"(%llu)\n", fid_gen, setgen);
return (SET_ERROR(EINVAL));
}
return (zfsctl_snapdir_vget(sb, objsetid, fid_gen, ipp));
}
if ((err = zfs_enter(zfsvfs, FTAG)) != 0)
return (err);
/* A zero fid_gen means we are in the .zfs control directories */
if (fid_gen == 0 &&
(object == ZFSCTL_INO_ROOT || object == ZFSCTL_INO_SNAPDIR)) {
*ipp = zfsvfs->z_ctldir;
ASSERT(*ipp != NULL);
if (object == ZFSCTL_INO_SNAPDIR) {
VERIFY(zfsctl_root_lookup(*ipp, "snapshot", ipp,
0, kcred, NULL, NULL) == 0);
} else {
/*
* Must have an existing ref, so igrab()
* cannot return NULL
*/
VERIFY3P(igrab(*ipp), !=, NULL);
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
gen_mask = -1ULL >> (64 - 8 * i);
dprintf("getting %llu [%llu mask %llx]\n", object, fid_gen, gen_mask);
if ((err = zfs_zget(zfsvfs, object, &zp))) {
zfs_exit(zfsvfs, FTAG);
return (err);
}
/* Don't export xattr stuff */
if (zp->z_pflags & ZFS_XATTR) {
zrele(zp);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOENT));
}
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(zfsvfs), &zp_gen,
sizeof (uint64_t));
zp_gen = zp_gen & gen_mask;
if (zp_gen == 0)
zp_gen = 1;
if ((fid_gen == 0) && (zfsvfs->z_root == object))
fid_gen = zp_gen;
if (zp->z_unlinked || zp_gen != fid_gen) {
dprintf("znode gen (%llu) != fid gen (%llu)\n", zp_gen,
fid_gen);
zrele(zp);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOENT));
}
*ipp = ZTOI(zp);
if (*ipp)
zfs_znode_update_vfs(ITOZ(*ipp));
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* Block out VFS ops and close zfsvfs_t
*
* Note, if successful, then we return with the 'z_teardown_lock' and
* 'z_teardown_inactive_lock' write held. We leave ownership of the underlying
* dataset and objset intact so that they can be atomically handed off during
* a subsequent rollback or recv operation and the resume thereafter.
*/
int
zfs_suspend_fs(zfsvfs_t *zfsvfs)
{
int error;
if ((error = zfsvfs_teardown(zfsvfs, B_FALSE)) != 0)
return (error);
return (0);
}
/*
* Rebuild SA and release VOPs. Note that ownership of the underlying dataset
* is an invariant across any of the operations that can be performed while the
* filesystem was suspended. Whether it succeeded or failed, the preconditions
* are the same: the relevant objset and associated dataset are owned by
* zfsvfs, held, and long held on entry.
*/
int
zfs_resume_fs(zfsvfs_t *zfsvfs, dsl_dataset_t *ds)
{
int err, err2;
znode_t *zp;
ASSERT(ZFS_TEARDOWN_WRITE_HELD(zfsvfs));
ASSERT(RW_WRITE_HELD(&zfsvfs->z_teardown_inactive_lock));
/*
* We already own this, so just update the objset_t, as the one we
* had before may have been evicted.
*/
objset_t *os;
VERIFY3P(ds->ds_owner, ==, zfsvfs);
VERIFY(dsl_dataset_long_held(ds));
dsl_pool_t *dp = spa_get_dsl(dsl_dataset_get_spa(ds));
dsl_pool_config_enter(dp, FTAG);
VERIFY0(dmu_objset_from_ds(ds, &os));
dsl_pool_config_exit(dp, FTAG);
err = zfsvfs_init(zfsvfs, os);
if (err != 0)
goto bail;
ds->ds_dir->dd_activity_cancelled = B_FALSE;
VERIFY(zfsvfs_setup(zfsvfs, B_FALSE) == 0);
zfs_set_fuid_feature(zfsvfs);
zfsvfs->z_rollback_time = jiffies;
/*
* Attempt to re-establish all the active inodes with their
* dbufs. If a zfs_rezget() fails, then we unhash the inode
* and mark it stale. This prevents a collision if a new
* inode/object is created which must use the same inode
* number. The stale inode will be be released when the
* VFS prunes the dentry holding the remaining references
* on the stale inode.
*/
mutex_enter(&zfsvfs->z_znodes_lock);
for (zp = list_head(&zfsvfs->z_all_znodes); zp;
zp = list_next(&zfsvfs->z_all_znodes, zp)) {
err2 = zfs_rezget(zp);
if (err2) {
zpl_d_drop_aliases(ZTOI(zp));
remove_inode_hash(ZTOI(zp));
}
/* see comment in zfs_suspend_fs() */
if (zp->z_suspended) {
zfs_zrele_async(zp);
zp->z_suspended = B_FALSE;
}
}
mutex_exit(&zfsvfs->z_znodes_lock);
if (!zfs_is_readonly(zfsvfs) && !zfsvfs->z_unmounted) {
/*
* zfs_suspend_fs() could have interrupted freeing
* of dnodes. We need to restart this freeing so
* that we don't "leak" the space.
*/
zfs_unlinked_drain(zfsvfs);
}
/*
* Most of the time zfs_suspend_fs is used for changing the contents
* of the underlying dataset. ZFS rollback and receive operations
* might create files for which negative dentries are present in
* the cache. Since walking the dcache would require a lot of GPL-only
* code duplication, it's much easier on these rather rare occasions
* just to flush the whole dcache for the given dataset/filesystem.
*/
shrink_dcache_sb(zfsvfs->z_sb);
bail:
if (err != 0)
zfsvfs->z_unmounted = B_TRUE;
/* release the VFS ops */
rw_exit(&zfsvfs->z_teardown_inactive_lock);
ZFS_TEARDOWN_EXIT(zfsvfs, FTAG);
if (err != 0) {
/*
* Since we couldn't setup the sa framework, try to force
* unmount this file system.
*/
if (zfsvfs->z_os)
(void) zfs_umount(zfsvfs->z_sb);
}
return (err);
}
/*
* Release VOPs and unmount a suspended filesystem.
*/
int
zfs_end_fs(zfsvfs_t *zfsvfs, dsl_dataset_t *ds)
{
ASSERT(ZFS_TEARDOWN_WRITE_HELD(zfsvfs));
ASSERT(RW_WRITE_HELD(&zfsvfs->z_teardown_inactive_lock));
/*
* We already own this, so just hold and rele it to update the
* objset_t, as the one we had before may have been evicted.
*/
objset_t *os;
VERIFY3P(ds->ds_owner, ==, zfsvfs);
VERIFY(dsl_dataset_long_held(ds));
dsl_pool_t *dp = spa_get_dsl(dsl_dataset_get_spa(ds));
dsl_pool_config_enter(dp, FTAG);
VERIFY0(dmu_objset_from_ds(ds, &os));
dsl_pool_config_exit(dp, FTAG);
zfsvfs->z_os = os;
/* release the VOPs */
rw_exit(&zfsvfs->z_teardown_inactive_lock);
ZFS_TEARDOWN_EXIT(zfsvfs, FTAG);
/*
* Try to force unmount this file system.
*/
(void) zfs_umount(zfsvfs->z_sb);
zfsvfs->z_unmounted = B_TRUE;
return (0);
}
/*
* Automounted snapshots rely on periodic revalidation
* to defer snapshots from being automatically unmounted.
*/
inline void
zfs_exit_fs(zfsvfs_t *zfsvfs)
{
if (!zfsvfs->z_issnap)
return;
if (time_after(jiffies, zfsvfs->z_snap_defer_time +
MAX(zfs_expire_snapshot * HZ / 2, HZ))) {
zfsvfs->z_snap_defer_time = jiffies;
zfsctl_snapshot_unmount_delay(zfsvfs->z_os->os_spa,
dmu_objset_id(zfsvfs->z_os),
zfs_expire_snapshot);
}
}
int
zfs_set_version(zfsvfs_t *zfsvfs, uint64_t newvers)
{
int error;
objset_t *os = zfsvfs->z_os;
dmu_tx_t *tx;
if (newvers < ZPL_VERSION_INITIAL || newvers > ZPL_VERSION)
return (SET_ERROR(EINVAL));
if (newvers < zfsvfs->z_version)
return (SET_ERROR(EINVAL));
if (zfs_spa_version_map(newvers) >
spa_version(dmu_objset_spa(zfsvfs->z_os)))
return (SET_ERROR(ENOTSUP));
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, MASTER_NODE_OBJ, B_FALSE, ZPL_VERSION_STR);
if (newvers >= ZPL_VERSION_SA && !zfsvfs->z_use_sa) {
dmu_tx_hold_zap(tx, MASTER_NODE_OBJ, B_TRUE,
ZFS_SA_ATTRS);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL);
}
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
return (error);
}
error = zap_update(os, MASTER_NODE_OBJ, ZPL_VERSION_STR,
8, 1, &newvers, tx);
if (error) {
dmu_tx_commit(tx);
return (error);
}
if (newvers >= ZPL_VERSION_SA && !zfsvfs->z_use_sa) {
uint64_t sa_obj;
ASSERT3U(spa_version(dmu_objset_spa(zfsvfs->z_os)), >=,
SPA_VERSION_SA);
sa_obj = zap_create(os, DMU_OT_SA_MASTER_NODE,
DMU_OT_NONE, 0, tx);
error = zap_add(os, MASTER_NODE_OBJ,
ZFS_SA_ATTRS, 8, 1, &sa_obj, tx);
ASSERT0(error);
VERIFY(0 == sa_set_sa_object(os, sa_obj));
sa_register_update_callback(os, zfs_sa_upgrade);
}
spa_history_log_internal_ds(dmu_objset_ds(os), "upgrade", tx,
"from %llu to %llu", zfsvfs->z_version, newvers);
dmu_tx_commit(tx);
zfsvfs->z_version = newvers;
os->os_version = newvers;
zfs_set_fuid_feature(zfsvfs);
return (0);
}
/*
* Read a property stored within the master node.
*/
int
zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value)
{
uint64_t *cached_copy = NULL;
/*
* Figure out where in the objset_t the cached copy would live, if it
* is available for the requested property.
*/
if (os != NULL) {
switch (prop) {
case ZFS_PROP_VERSION:
cached_copy = &os->os_version;
break;
case ZFS_PROP_NORMALIZE:
cached_copy = &os->os_normalization;
break;
case ZFS_PROP_UTF8ONLY:
cached_copy = &os->os_utf8only;
break;
case ZFS_PROP_CASE:
cached_copy = &os->os_casesensitivity;
break;
default:
break;
}
}
if (cached_copy != NULL && *cached_copy != OBJSET_PROP_UNINITIALIZED) {
*value = *cached_copy;
return (0);
}
/*
* If the property wasn't cached, look up the file system's value for
* the property. For the version property, we look up a slightly
* different string.
*/
const char *pname;
int error = ENOENT;
if (prop == ZFS_PROP_VERSION)
pname = ZPL_VERSION_STR;
else
pname = zfs_prop_to_name(prop);
if (os != NULL) {
ASSERT3U(os->os_phys->os_type, ==, DMU_OST_ZFS);
error = zap_lookup(os, MASTER_NODE_OBJ, pname, 8, 1, value);
}
if (error == ENOENT) {
/* No value set, use the default value */
switch (prop) {
case ZFS_PROP_VERSION:
*value = ZPL_VERSION;
break;
case ZFS_PROP_NORMALIZE:
case ZFS_PROP_UTF8ONLY:
*value = 0;
break;
case ZFS_PROP_CASE:
*value = ZFS_CASE_SENSITIVE;
break;
case ZFS_PROP_ACLTYPE:
*value = ZFS_ACLTYPE_OFF;
break;
default:
return (error);
}
error = 0;
}
/*
* If one of the methods for getting the property value above worked,
* copy it into the objset_t's cache.
*/
if (error == 0 && cached_copy != NULL) {
*cached_copy = *value;
}
return (error);
}
/*
* Return true if the corresponding vfs's unmounted flag is set.
* Otherwise return false.
* If this function returns true we know VFS unmount has been initiated.
*/
boolean_t
zfs_get_vfs_flag_unmounted(objset_t *os)
{
zfsvfs_t *zfvp;
boolean_t unmounted = B_FALSE;
ASSERT(dmu_objset_type(os) == DMU_OST_ZFS);
mutex_enter(&os->os_user_ptr_lock);
zfvp = dmu_objset_get_user(os);
if (zfvp != NULL && zfvp->z_unmounted)
unmounted = B_TRUE;
mutex_exit(&os->os_user_ptr_lock);
return (unmounted);
}
void
zfsvfs_update_fromname(const char *oldname, const char *newname)
{
/*
* We don't need to do anything here, the devname is always current by
* virtue of zfsvfs->z_sb->s_op->show_devname.
*/
(void) oldname, (void) newname;
}
void
zfs_init(void)
{
zfsctl_init();
zfs_znode_init();
dmu_objset_register_type(DMU_OST_ZFS, zpl_get_file_info);
register_filesystem(&zpl_fs_type);
}
void
zfs_fini(void)
{
/*
* we don't use outstanding because zpl_posix_acl_free might add more.
*/
taskq_wait(system_delay_taskq);
taskq_wait(system_taskq);
unregister_filesystem(&zpl_fs_type);
zfs_znode_fini();
zfsctl_fini();
}
#if defined(_KERNEL)
EXPORT_SYMBOL(zfs_suspend_fs);
EXPORT_SYMBOL(zfs_resume_fs);
EXPORT_SYMBOL(zfs_set_version);
EXPORT_SYMBOL(zfsvfs_create);
EXPORT_SYMBOL(zfsvfs_free);
EXPORT_SYMBOL(zfs_is_readonly);
EXPORT_SYMBOL(zfs_domount);
EXPORT_SYMBOL(zfs_preumount);
EXPORT_SYMBOL(zfs_umount);
EXPORT_SYMBOL(zfs_remount);
EXPORT_SYMBOL(zfs_statvfs);
EXPORT_SYMBOL(zfs_vget);
EXPORT_SYMBOL(zfs_prune);
#endif
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
index b8af3e3b058d..234c4d5ef0e0 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_vnops_os.c
@@ -1,4233 +1,4232 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
* Copyright (c) 2015 by Chunwei Chen. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc.
*/
/* Portions Copyright 2007 Jeremy Teo */
/* Portions Copyright 2010 Robert Milkowski */
#include <sys/types.h>
#include <sys/param.h>
#include <sys/time.h>
#include <sys/sysmacros.h>
#include <sys/vfs.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/kmem.h>
#include <sys/taskq.h>
#include <sys/uio.h>
#include <sys/vmsystm.h>
#include <sys/atomic.h>
#include <sys/pathname.h>
#include <sys/cmn_err.h>
#include <sys/errno.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_ioctl.h>
#include <sys/fs/zfs.h>
#include <sys/dmu.h>
#include <sys/dmu_objset.h>
#include <sys/spa.h>
#include <sys/txg.h>
#include <sys/dbuf.h>
#include <sys/zap.h>
#include <sys/sa.h>
#include <sys/policy.h>
#include <sys/sunddi.h>
#include <sys/sid.h>
#include <sys/zfs_ctldir.h>
#include <sys/zfs_fuid.h>
#include <sys/zfs_quota.h>
#include <sys/zfs_sa.h>
#include <sys/zfs_vnops.h>
#include <sys/zfs_rlock.h>
#include <sys/cred.h>
#include <sys/zpl.h>
#include <sys/zil.h>
#include <sys/sa_impl.h>
/*
* Programming rules.
*
* Each vnode op performs some logical unit of work. To do this, the ZPL must
* properly lock its in-core state, create a DMU transaction, do the work,
* record this work in the intent log (ZIL), commit the DMU transaction,
* and wait for the intent log to commit if it is a synchronous operation.
* Moreover, the vnode ops must work in both normal and log replay context.
* The ordering of events is important to avoid deadlocks and references
* to freed memory. The example below illustrates the following Big Rules:
*
* (1) A check must be made in each zfs thread for a mounted file system.
* This is done avoiding races using zfs_enter(zfsvfs).
* A zfs_exit(zfsvfs) is needed before all returns. Any znodes
* must be checked with zfs_verify_zp(zp). Both of these macros
* can return EIO from the calling function.
*
* (2) zrele() should always be the last thing except for zil_commit() (if
* necessary) and zfs_exit(). This is for 3 reasons: First, if it's the
* last reference, the vnode/znode can be freed, so the zp may point to
* freed memory. Second, the last reference will call zfs_zinactive(),
* which may induce a lot of work -- pushing cached pages (which acquires
* range locks) and syncing out cached atime changes. Third,
* zfs_zinactive() may require a new tx, which could deadlock the system
* if you were already holding one. This deadlock occurs because the tx
* currently being operated on prevents a txg from syncing, which
* prevents the new tx from progressing, resulting in a deadlock. If you
* must call zrele() within a tx, use zfs_zrele_async(). Note that iput()
* is a synonym for zrele().
*
* (3) All range locks must be grabbed before calling dmu_tx_assign(),
* as they can span dmu_tx_assign() calls.
*
* (4) If ZPL locks are held, pass TXG_NOWAIT as the second argument to
* dmu_tx_assign(). This is critical because we don't want to block
* while holding locks.
*
* If no ZPL locks are held (aside from zfs_enter()), use TXG_WAIT. This
* reduces lock contention and CPU usage when we must wait (note that if
* throughput is constrained by the storage, nearly every transaction
* must wait).
*
* Note, in particular, that if a lock is sometimes acquired before
* the tx assigns, and sometimes after (e.g. z_lock), then failing
* to use a non-blocking assign can deadlock the system. The scenario:
*
* Thread A has grabbed a lock before calling dmu_tx_assign().
* Thread B is in an already-assigned tx, and blocks for this lock.
* Thread A calls dmu_tx_assign(TXG_WAIT) and blocks in txg_wait_open()
* forever, because the previous txg can't quiesce until B's tx commits.
*
* If dmu_tx_assign() returns ERESTART and zfsvfs->z_assign is TXG_NOWAIT,
* then drop all locks, call dmu_tx_wait(), and try again. On subsequent
* calls to dmu_tx_assign(), pass TXG_NOTHROTTLE in addition to TXG_NOWAIT,
* to indicate that this operation has already called dmu_tx_wait().
* This will ensure that we don't retry forever, waiting a short bit
* each time.
*
* (5) If the operation succeeded, generate the intent log entry for it
* before dropping locks. This ensures that the ordering of events
* in the intent log matches the order in which they actually occurred.
* During ZIL replay the zfs_log_* functions will update the sequence
* number to indicate the zil transaction has replayed.
*
* (6) At the end of each vnode op, the DMU tx must always commit,
* regardless of whether there were any errors.
*
* (7) After dropping all locks, invoke zil_commit(zilog, foid)
* to ensure that synchronous semantics are provided when necessary.
*
* In general, this is how things should be ordered in each vnode op:
*
* zfs_enter(zfsvfs); // exit if unmounted
* top:
* zfs_dirent_lock(&dl, ...) // lock directory entry (may igrab())
* rw_enter(...); // grab any other locks you need
* tx = dmu_tx_create(...); // get DMU tx
* dmu_tx_hold_*(); // hold each object you might modify
* error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
* if (error) {
* rw_exit(...); // drop locks
* zfs_dirent_unlock(dl); // unlock directory entry
* zrele(...); // release held znodes
* if (error == ERESTART) {
* waited = B_TRUE;
* dmu_tx_wait(tx);
* dmu_tx_abort(tx);
* goto top;
* }
* dmu_tx_abort(tx); // abort DMU tx
* zfs_exit(zfsvfs); // finished in zfs
* return (error); // really out of space
* }
* error = do_real_work(); // do whatever this VOP does
* if (error == 0)
* zfs_log_*(...); // on success, make ZIL entry
* dmu_tx_commit(tx); // commit DMU tx -- error or not
* rw_exit(...); // drop locks
* zfs_dirent_unlock(dl); // unlock directory entry
* zrele(...); // release held znodes
* zil_commit(zilog, foid); // synchronous when necessary
* zfs_exit(zfsvfs); // finished in zfs
* return (error); // done, report error
*/
int
zfs_open(struct inode *ip, int mode, int flag, cred_t *cr)
{
(void) cr;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
/* Honor ZFS_APPENDONLY file attribute */
if ((mode & FMODE_WRITE) && (zp->z_pflags & ZFS_APPENDONLY) &&
((flag & O_APPEND) == 0)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
/* Keep a count of the synchronous opens in the znode */
if (flag & O_SYNC)
atomic_inc_32(&zp->z_sync_cnt);
zfs_exit(zfsvfs, FTAG);
return (0);
}
int
zfs_close(struct inode *ip, int flag, cred_t *cr)
{
(void) cr;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
/* Decrement the synchronous opens in the znode */
if (flag & O_SYNC)
atomic_dec_32(&zp->z_sync_cnt);
zfs_exit(zfsvfs, FTAG);
return (0);
}
#if defined(_KERNEL)
static int zfs_fillpage(struct inode *ip, struct page *pp);
/*
* When a file is memory mapped, we must keep the IO data synchronized
* between the DMU cache and the memory mapped pages. Update all mapped
* pages with the contents of the coresponding dmu buffer.
*/
void
update_pages(znode_t *zp, int64_t start, int len, objset_t *os)
{
struct address_space *mp = ZTOI(zp)->i_mapping;
int64_t off = start & (PAGE_SIZE - 1);
for (start &= PAGE_MASK; len > 0; start += PAGE_SIZE) {
uint64_t nbytes = MIN(PAGE_SIZE - off, len);
struct page *pp = find_lock_page(mp, start >> PAGE_SHIFT);
if (pp) {
if (mapping_writably_mapped(mp))
flush_dcache_page(pp);
void *pb = kmap(pp);
int error = dmu_read(os, zp->z_id, start + off,
nbytes, pb + off, DMU_READ_PREFETCH);
kunmap(pp);
if (error) {
SetPageError(pp);
ClearPageUptodate(pp);
} else {
ClearPageError(pp);
SetPageUptodate(pp);
if (mapping_writably_mapped(mp))
flush_dcache_page(pp);
mark_page_accessed(pp);
}
unlock_page(pp);
put_page(pp);
}
len -= nbytes;
off = 0;
}
}
/*
* When a file is memory mapped, we must keep the I/O data synchronized
* between the DMU cache and the memory mapped pages. Preferentially read
* from memory mapped pages, otherwise fallback to reading through the dmu.
*/
int
mappedread(znode_t *zp, int nbytes, zfs_uio_t *uio)
{
struct inode *ip = ZTOI(zp);
struct address_space *mp = ip->i_mapping;
int64_t start = uio->uio_loffset;
int64_t off = start & (PAGE_SIZE - 1);
int len = nbytes;
int error = 0;
for (start &= PAGE_MASK; len > 0; start += PAGE_SIZE) {
uint64_t bytes = MIN(PAGE_SIZE - off, len);
struct page *pp = find_lock_page(mp, start >> PAGE_SHIFT);
if (pp) {
/*
* If filemap_fault() retries there exists a window
* where the page will be unlocked and not up to date.
* In this case we must try and fill the page.
*/
if (unlikely(!PageUptodate(pp))) {
error = zfs_fillpage(ip, pp);
if (error) {
unlock_page(pp);
put_page(pp);
return (error);
}
}
ASSERT(PageUptodate(pp) || PageDirty(pp));
unlock_page(pp);
void *pb = kmap(pp);
error = zfs_uiomove(pb + off, bytes, UIO_READ, uio);
kunmap(pp);
if (mapping_writably_mapped(mp))
flush_dcache_page(pp);
mark_page_accessed(pp);
put_page(pp);
} else {
error = dmu_read_uio_dbuf(sa_get_db(zp->z_sa_hdl),
uio, bytes);
}
len -= bytes;
off = 0;
if (error)
break;
}
return (error);
}
#endif /* _KERNEL */
static unsigned long zfs_delete_blocks = DMU_MAX_DELETEBLKCNT;
/*
* Write the bytes to a file.
*
* IN: zp - znode of file to be written to
* data - bytes to write
* len - number of bytes to write
* pos - offset to start writing at
*
* OUT: resid - remaining bytes to write
*
* RETURN: 0 if success
* positive error code if failure. EIO is returned
* for a short write when residp isn't provided.
*
* Timestamps:
* zp - ctime|mtime updated if byte count > 0
*/
int
zfs_write_simple(znode_t *zp, const void *data, size_t len,
loff_t pos, size_t *residp)
{
fstrans_cookie_t cookie;
int error;
struct iovec iov;
iov.iov_base = (void *)data;
iov.iov_len = len;
zfs_uio_t uio;
zfs_uio_iovec_init(&uio, &iov, 1, pos, UIO_SYSSPACE, len, 0);
cookie = spl_fstrans_mark();
error = zfs_write(zp, &uio, 0, kcred);
spl_fstrans_unmark(cookie);
if (error == 0) {
if (residp != NULL)
*residp = zfs_uio_resid(&uio);
else if (zfs_uio_resid(&uio) != 0)
error = SET_ERROR(EIO);
}
return (error);
}
static void
zfs_rele_async_task(void *arg)
{
iput(arg);
}
void
zfs_zrele_async(znode_t *zp)
{
struct inode *ip = ZTOI(zp);
objset_t *os = ITOZSB(ip)->z_os;
ASSERT(atomic_read(&ip->i_count) > 0);
ASSERT(os != NULL);
/*
* If decrementing the count would put us at 0, we can't do it inline
* here, because that would be synchronous. Instead, dispatch an iput
* to run later.
*
* For more information on the dangers of a synchronous iput, see the
* header comment of this file.
*/
if (!atomic_add_unless(&ip->i_count, -1, 1)) {
VERIFY(taskq_dispatch(dsl_pool_zrele_taskq(dmu_objset_pool(os)),
zfs_rele_async_task, ip, TQ_SLEEP) != TASKQID_INVALID);
}
}
/*
* Lookup an entry in a directory, or an extended attribute directory.
* If it exists, return a held inode reference for it.
*
* IN: zdp - znode of directory to search.
* nm - name of entry to lookup.
* flags - LOOKUP_XATTR set if looking for an attribute.
* cr - credentials of caller.
* direntflags - directory lookup flags
* realpnp - returned pathname.
*
* OUT: zpp - znode of located entry, NULL if not found.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* NA
*/
int
zfs_lookup(znode_t *zdp, char *nm, znode_t **zpp, int flags, cred_t *cr,
int *direntflags, pathname_t *realpnp)
{
zfsvfs_t *zfsvfs = ZTOZSB(zdp);
int error = 0;
/*
* Fast path lookup, however we must skip DNLC lookup
* for case folding or normalizing lookups because the
* DNLC code only stores the passed in name. This means
* creating 'a' and removing 'A' on a case insensitive
* file system would work, but DNLC still thinks 'a'
* exists and won't let you create it again on the next
* pass through fast path.
*/
if (!(flags & (LOOKUP_XATTR | FIGNORECASE))) {
if (!S_ISDIR(ZTOI(zdp)->i_mode)) {
return (SET_ERROR(ENOTDIR));
} else if (zdp->z_sa_hdl == NULL) {
return (SET_ERROR(EIO));
}
if (nm[0] == 0 || (nm[0] == '.' && nm[1] == '\0')) {
error = zfs_fastaccesschk_execute(zdp, cr);
if (!error) {
*zpp = zdp;
zhold(*zpp);
return (0);
}
return (error);
}
}
if ((error = zfs_enter_verify_zp(zfsvfs, zdp, FTAG)) != 0)
return (error);
*zpp = NULL;
if (flags & LOOKUP_XATTR) {
/*
* We don't allow recursive attributes..
* Maybe someday we will.
*/
if (zdp->z_pflags & ZFS_XATTR) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if ((error = zfs_get_xattrdir(zdp, zpp, cr, flags))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Do we have permission to get into attribute directory?
*/
if ((error = zfs_zaccess(*zpp, ACE_EXECUTE, 0,
- B_TRUE, cr, kcred->user_ns))) {
+ B_TRUE, cr, zfs_init_idmap))) {
zrele(*zpp);
*zpp = NULL;
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (!S_ISDIR(ZTOI(zdp)->i_mode)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOTDIR));
}
/*
* Check accessibility of directory.
*/
if ((error = zfs_zaccess(zdp, ACE_EXECUTE, 0, B_FALSE, cr,
- kcred->user_ns))) {
+ zfs_init_idmap))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zfsvfs->z_utf8 && u8_validate(nm, strlen(nm),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
error = zfs_dirlook(zdp, nm, zpp, flags, direntflags, realpnp);
if ((error == 0) && (*zpp))
zfs_znode_update_vfs(*zpp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Attempt to create a new entry in a directory. If the entry
* already exists, truncate the file if permissible, else return
* an error. Return the ip of the created or trunc'd file.
*
* IN: dzp - znode of directory to put new file entry in.
* name - name of new file entry.
* vap - attributes of new file.
* excl - flag indicating exclusive or non-exclusive mode.
* mode - mode to open file with.
* cr - credentials of caller.
* flag - file flag.
* vsecp - ACL to be set
* mnt_ns - user namespace of the mount
*
* OUT: zpp - znode of created or trunc'd entry.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dzp - ctime|mtime updated if new entry created
* zp - ctime|mtime always, atime if new
*/
int
zfs_create(znode_t *dzp, char *name, vattr_t *vap, int excl,
int mode, znode_t **zpp, cred_t *cr, int flag, vsecattr_t *vsecp,
- zuserns_t *mnt_ns)
+ zidmap_t *mnt_ns)
{
znode_t *zp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
objset_t *os;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
int error;
uid_t uid;
gid_t gid;
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
boolean_t have_acl = B_FALSE;
boolean_t waited = B_FALSE;
boolean_t skip_acl = (flag & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
/*
* If we have an ephemeral id, ACL, or XVATTR then
* make sure file system is at proper version
*/
gid = crgetgid(cr);
uid = crgetuid(cr);
if (zfsvfs->z_use_fuids == B_FALSE &&
(vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid)))
return (SET_ERROR(EINVAL));
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
os = zfsvfs->z_os;
zilog = zfsvfs->z_log;
if (zfsvfs->z_utf8 && u8_validate(name, strlen(name),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (vap->va_mask & ATTR_XVATTR) {
if ((error = secpolicy_xvattr((xvattr_t *)vap,
crgetuid(cr), cr, vap->va_mode)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
top:
*zpp = NULL;
if (*name == '\0') {
/*
* Null component name refers to the directory itself.
*/
zhold(dzp);
zp = dzp;
dl = NULL;
error = 0;
} else {
/* possible igrab(zp) */
int zflg = 0;
if (flag & FIGNORECASE)
zflg |= ZCILOOK;
error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg,
NULL, NULL);
if (error) {
if (have_acl)
zfs_acl_ids_free(&acl_ids);
if (strcmp(name, "..") == 0)
error = SET_ERROR(EISDIR);
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
if (zp == NULL) {
uint64_t txtype;
uint64_t projid = ZFS_DEFAULT_PROJID;
/*
* Create a new file object and update the directory
* to reference it.
*/
if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, skip_acl, cr,
mnt_ns))) {
if (have_acl)
zfs_acl_ids_free(&acl_ids);
goto out;
}
/*
* We only support the creation of regular files in
* extended attribute directories.
*/
if ((dzp->z_pflags & ZFS_XATTR) && !S_ISREG(vap->va_mode)) {
if (have_acl)
zfs_acl_ids_free(&acl_ids);
error = SET_ERROR(EINVAL);
goto out;
}
if (!have_acl && (error = zfs_acl_ids_create(dzp, 0, vap,
cr, vsecp, &acl_ids, mnt_ns)) != 0)
goto out;
have_acl = B_TRUE;
if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode))
projid = zfs_inherit_projid(dzp);
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, projid)) {
zfs_acl_ids_free(&acl_ids);
error = SET_ERROR(EDQUOT);
goto out;
}
tx = dmu_tx_create(os);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name);
dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE);
if (!zfsvfs->z_use_sa &&
acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, acl_ids.z_aclp->z_acl_bytes);
}
error = dmu_tx_assign(tx,
(waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
error = zfs_link_create(dl, zp, tx, ZNEW);
if (error != 0) {
/*
* Since, we failed to add the directory entry for it,
* delete the newly created dnode.
*/
zfs_znode_delete(zp, tx);
remove_inode_hash(ZTOI(zp));
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
goto out;
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
txtype = zfs_log_create_txtype(Z_FILE, vsecp, vap);
if (flag & FIGNORECASE)
txtype |= TX_CI;
zfs_log_create(zilog, tx, txtype, dzp, zp, name,
vsecp, acl_ids.z_fuidp, vap);
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
} else {
int aflags = (flag & O_APPEND) ? V_APPEND : 0;
if (have_acl)
zfs_acl_ids_free(&acl_ids);
/*
* A directory entry already exists for this name.
*/
/*
* Can't truncate an existing file if in exclusive mode.
*/
if (excl) {
error = SET_ERROR(EEXIST);
goto out;
}
/*
* Can't open a directory for writing.
*/
if (S_ISDIR(ZTOI(zp)->i_mode)) {
error = SET_ERROR(EISDIR);
goto out;
}
/*
* Verify requested access to file.
*/
if (mode && (error = zfs_zaccess_rwx(zp, mode, aflags, cr,
mnt_ns))) {
goto out;
}
mutex_enter(&dzp->z_lock);
dzp->z_seq++;
mutex_exit(&dzp->z_lock);
/*
* Truncate regular files if requested.
*/
if (S_ISREG(ZTOI(zp)->i_mode) &&
(vap->va_mask & ATTR_SIZE) && (vap->va_size == 0)) {
/* we can't hold any locks when calling zfs_freesp() */
if (dl) {
zfs_dirent_unlock(dl);
dl = NULL;
}
error = zfs_freesp(zp, 0, 0, mode, TRUE);
}
}
out:
if (dl)
zfs_dirent_unlock(dl);
if (error) {
if (zp)
zrele(zp);
} else {
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
*zpp = zp;
}
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
int
zfs_tmpfile(struct inode *dip, vattr_t *vap, int excl,
int mode, struct inode **ipp, cred_t *cr, int flag, vsecattr_t *vsecp,
- zuserns_t *mnt_ns)
+ zidmap_t *mnt_ns)
{
(void) excl, (void) mode, (void) flag;
znode_t *zp = NULL, *dzp = ITOZ(dip);
zfsvfs_t *zfsvfs = ITOZSB(dip);
objset_t *os;
dmu_tx_t *tx;
int error;
uid_t uid;
gid_t gid;
zfs_acl_ids_t acl_ids;
uint64_t projid = ZFS_DEFAULT_PROJID;
boolean_t fuid_dirtied;
boolean_t have_acl = B_FALSE;
boolean_t waited = B_FALSE;
/*
* If we have an ephemeral id, ACL, or XVATTR then
* make sure file system is at proper version
*/
gid = crgetgid(cr);
uid = crgetuid(cr);
if (zfsvfs->z_use_fuids == B_FALSE &&
(vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid)))
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
os = zfsvfs->z_os;
if (vap->va_mask & ATTR_XVATTR) {
if ((error = secpolicy_xvattr((xvattr_t *)vap,
crgetuid(cr), cr, vap->va_mode)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
top:
*ipp = NULL;
/*
* Create a new file object and update the directory
* to reference it.
*/
if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) {
if (have_acl)
zfs_acl_ids_free(&acl_ids);
goto out;
}
if (!have_acl && (error = zfs_acl_ids_create(dzp, 0, vap,
cr, vsecp, &acl_ids, mnt_ns)) != 0)
goto out;
have_acl = B_TRUE;
if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode))
projid = zfs_inherit_projid(dzp);
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, projid)) {
zfs_acl_ids_free(&acl_ids);
error = SET_ERROR(EDQUOT);
goto out;
}
tx = dmu_tx_create(os);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
if (!zfsvfs->z_use_sa &&
acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, acl_ids.z_aclp->z_acl_bytes);
}
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
zfs_mknode(dzp, vap, tx, cr, IS_TMPFILE, &zp, &acl_ids);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
/* Add to unlinked set */
zp->z_unlinked = B_TRUE;
zfs_unlinked_add(zp, tx);
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
out:
if (error) {
if (zp)
zrele(zp);
} else {
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
*ipp = ZTOI(zp);
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Remove an entry from a directory.
*
* IN: dzp - znode of directory to remove entry from.
* name - name of entry to remove.
* cr - credentials of caller.
* flags - case flags.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* dzp - ctime|mtime
* ip - ctime (if nlink > 0)
*/
static uint64_t null_xattr = 0;
int
zfs_remove(znode_t *dzp, char *name, cred_t *cr, int flags)
{
znode_t *zp;
znode_t *xzp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
uint64_t acl_obj, xattr_obj;
uint64_t xattr_obj_unlinked = 0;
uint64_t obj = 0;
uint64_t links;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
boolean_t may_delete_now, delete_now = FALSE;
boolean_t unlinked, toobig = FALSE;
uint64_t txtype;
pathname_t *realnmp = NULL;
pathname_t realnm;
int error;
int zflg = ZEXISTS;
boolean_t waited = B_FALSE;
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (flags & FIGNORECASE) {
zflg |= ZCILOOK;
pn_alloc(&realnm);
realnmp = &realnm;
}
top:
xattr_obj = 0;
xzp = NULL;
/*
* Attempt to lock directory; fail if entry doesn't exist.
*/
if ((error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg,
NULL, realnmp))) {
if (realnmp)
pn_free(realnmp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
- if ((error = zfs_zaccess_delete(dzp, zp, cr, kcred->user_ns))) {
+ if ((error = zfs_zaccess_delete(dzp, zp, cr, zfs_init_idmap))) {
goto out;
}
/*
* Need to use rmdir for removing directories.
*/
if (S_ISDIR(ZTOI(zp)->i_mode)) {
error = SET_ERROR(EPERM);
goto out;
}
mutex_enter(&zp->z_lock);
may_delete_now = atomic_read(&ZTOI(zp)->i_count) == 1 &&
!zn_has_cached_data(zp, 0, LLONG_MAX);
mutex_exit(&zp->z_lock);
/*
* We may delete the znode now, or we may put it in the unlinked set;
* it depends on whether we're the last link, and on whether there are
* other holds on the inode. So we dmu_tx_hold() the right things to
* allow for either case.
*/
obj = zp->z_id;
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
zfs_sa_upgrade_txholds(tx, dzp);
if (may_delete_now) {
toobig = zp->z_size > zp->z_blksz * zfs_delete_blocks;
/* if the file is too big, only hold_free a token amount */
dmu_tx_hold_free(tx, zp->z_id, 0,
(toobig ? DMU_MAX_ACCESS : DMU_OBJECT_END));
}
/* are there any extended attributes? */
error = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj, sizeof (xattr_obj));
if (error == 0 && xattr_obj) {
error = zfs_zget(zfsvfs, xattr_obj, &xzp);
ASSERT0(error);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
dmu_tx_hold_sa(tx, xzp->z_sa_hdl, B_FALSE);
}
mutex_enter(&zp->z_lock);
if ((acl_obj = zfs_external_acl(zp)) != 0 && may_delete_now)
dmu_tx_hold_free(tx, acl_obj, 0, DMU_OBJECT_END);
mutex_exit(&zp->z_lock);
/* charge as an update -- would be nice not to charge at all */
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
/*
* Mark this transaction as typically resulting in a net free of space
*/
dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
zrele(zp);
if (xzp)
zrele(xzp);
goto top;
}
if (realnmp)
pn_free(realnmp);
dmu_tx_abort(tx);
zrele(zp);
if (xzp)
zrele(xzp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Remove the directory entry.
*/
error = zfs_link_destroy(dl, zp, tx, zflg, &unlinked);
if (error) {
dmu_tx_commit(tx);
goto out;
}
if (unlinked) {
/*
* Hold z_lock so that we can make sure that the ACL obj
* hasn't changed. Could have been deleted due to
* zfs_sa_upgrade().
*/
mutex_enter(&zp->z_lock);
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj_unlinked, sizeof (xattr_obj_unlinked));
delete_now = may_delete_now && !toobig &&
atomic_read(&ZTOI(zp)->i_count) == 1 &&
!zn_has_cached_data(zp, 0, LLONG_MAX) &&
xattr_obj == xattr_obj_unlinked &&
zfs_external_acl(zp) == acl_obj;
VERIFY_IMPLY(xattr_obj_unlinked, xzp);
}
if (delete_now) {
if (xattr_obj_unlinked) {
ASSERT3U(ZTOI(xzp)->i_nlink, ==, 2);
mutex_enter(&xzp->z_lock);
xzp->z_unlinked = B_TRUE;
clear_nlink(ZTOI(xzp));
links = 0;
error = sa_update(xzp->z_sa_hdl, SA_ZPL_LINKS(zfsvfs),
&links, sizeof (links), tx);
ASSERT3U(error, ==, 0);
mutex_exit(&xzp->z_lock);
zfs_unlinked_add(xzp, tx);
if (zp->z_is_sa)
error = sa_remove(zp->z_sa_hdl,
SA_ZPL_XATTR(zfsvfs), tx);
else
error = sa_update(zp->z_sa_hdl,
SA_ZPL_XATTR(zfsvfs), &null_xattr,
sizeof (uint64_t), tx);
ASSERT0(error);
}
/*
* Add to the unlinked set because a new reference could be
* taken concurrently resulting in a deferred destruction.
*/
zfs_unlinked_add(zp, tx);
mutex_exit(&zp->z_lock);
} else if (unlinked) {
mutex_exit(&zp->z_lock);
zfs_unlinked_add(zp, tx);
}
txtype = TX_REMOVE;
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_remove(zilog, tx, txtype, dzp, name, obj, unlinked);
dmu_tx_commit(tx);
out:
if (realnmp)
pn_free(realnmp);
zfs_dirent_unlock(dl);
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
if (delete_now)
zrele(zp);
else
zfs_zrele_async(zp);
if (xzp) {
zfs_znode_update_vfs(xzp);
zfs_zrele_async(xzp);
}
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Create a new directory and insert it into dzp using the name
* provided. Return a pointer to the inserted directory.
*
* IN: dzp - znode of directory to add subdir to.
* dirname - name of new directory.
* vap - attributes of new directory.
* cr - credentials of caller.
* flags - case flags.
* vsecp - ACL to be set
* mnt_ns - user namespace of the mount
*
* OUT: zpp - znode of created directory.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* dzp - ctime|mtime updated
* zpp - ctime|mtime|atime updated
*/
int
zfs_mkdir(znode_t *dzp, char *dirname, vattr_t *vap, znode_t **zpp,
- cred_t *cr, int flags, vsecattr_t *vsecp, zuserns_t *mnt_ns)
+ cred_t *cr, int flags, vsecattr_t *vsecp, zidmap_t *mnt_ns)
{
znode_t *zp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
zfs_dirlock_t *dl;
uint64_t txtype;
dmu_tx_t *tx;
int error;
int zf = ZNEW;
uid_t uid;
gid_t gid = crgetgid(cr);
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
boolean_t waited = B_FALSE;
ASSERT(S_ISDIR(vap->va_mode));
/*
* If we have an ephemeral id, ACL, or XVATTR then
* make sure file system is at proper version
*/
uid = crgetuid(cr);
if (zfsvfs->z_use_fuids == B_FALSE &&
(vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid)))
return (SET_ERROR(EINVAL));
if (dirname == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (dzp->z_pflags & ZFS_XATTR) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (zfsvfs->z_utf8 && u8_validate(dirname,
strlen(dirname), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (flags & FIGNORECASE)
zf |= ZCILOOK;
if (vap->va_mask & ATTR_XVATTR) {
if ((error = secpolicy_xvattr((xvattr_t *)vap,
crgetuid(cr), cr, vap->va_mode)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
}
if ((error = zfs_acl_ids_create(dzp, 0, vap, cr,
vsecp, &acl_ids, mnt_ns)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* First make sure the new directory doesn't exist.
*
* Existence is checked first to make sure we don't return
* EACCES instead of EEXIST which can cause some applications
* to fail.
*/
top:
*zpp = NULL;
if ((error = zfs_dirent_lock(&dl, dzp, dirname, &zp, zf,
NULL, NULL))) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = zfs_zaccess(dzp, ACE_ADD_SUBDIRECTORY, 0, B_FALSE, cr,
mnt_ns))) {
zfs_acl_ids_free(&acl_ids);
zfs_dirent_unlock(dl);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, zfs_inherit_projid(dzp))) {
zfs_acl_ids_free(&acl_ids);
zfs_dirent_unlock(dl);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EDQUOT));
}
/*
* Add a new entry to the directory.
*/
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, dirname);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0,
acl_ids.z_aclp->z_acl_bytes);
}
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Create new node.
*/
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
/*
* Now put new name in parent dir.
*/
error = zfs_link_create(dl, zp, tx, ZNEW);
if (error != 0) {
zfs_znode_delete(zp, tx);
remove_inode_hash(ZTOI(zp));
goto out;
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
*zpp = zp;
txtype = zfs_log_create_txtype(Z_DIR, vsecp, vap);
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_create(zilog, tx, txtype, dzp, zp, dirname, vsecp,
acl_ids.z_fuidp, vap);
out:
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
zfs_dirent_unlock(dl);
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
if (error != 0) {
zrele(zp);
} else {
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Remove a directory subdir entry. If the current working
* directory is the same as the subdir to be removed, the
* remove will fail.
*
* IN: dzp - znode of directory to remove from.
* name - name of directory to be removed.
* cwd - inode of current working directory.
* cr - credentials of caller.
* flags - case flags
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dzp - ctime|mtime updated
*/
int
zfs_rmdir(znode_t *dzp, char *name, znode_t *cwd, cred_t *cr,
int flags)
{
znode_t *zp;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
int error;
int zflg = ZEXISTS;
boolean_t waited = B_FALSE;
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (flags & FIGNORECASE)
zflg |= ZCILOOK;
top:
zp = NULL;
/*
* Attempt to lock directory; fail if entry doesn't exist.
*/
if ((error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg,
NULL, NULL))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
- if ((error = zfs_zaccess_delete(dzp, zp, cr, kcred->user_ns))) {
+ if ((error = zfs_zaccess_delete(dzp, zp, cr, zfs_init_idmap))) {
goto out;
}
if (!S_ISDIR(ZTOI(zp)->i_mode)) {
error = SET_ERROR(ENOTDIR);
goto out;
}
if (zp == cwd) {
error = SET_ERROR(EINVAL);
goto out;
}
/*
* Grab a lock on the directory to make sure that no one is
* trying to add (or lookup) entries while we are removing it.
*/
rw_enter(&zp->z_name_lock, RW_WRITER);
/*
* Grab a lock on the parent pointer to make sure we play well
* with the treewalk and directory rename code.
*/
rw_enter(&zp->z_parent_lock, RW_WRITER);
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
zfs_sa_upgrade_txholds(tx, zp);
zfs_sa_upgrade_txholds(tx, dzp);
dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
rw_exit(&zp->z_parent_lock);
rw_exit(&zp->z_name_lock);
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
zrele(zp);
goto top;
}
dmu_tx_abort(tx);
zrele(zp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
error = zfs_link_destroy(dl, zp, tx, zflg, NULL);
if (error == 0) {
uint64_t txtype = TX_RMDIR;
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_remove(zilog, tx, txtype, dzp, name, ZFS_NO_OBJECT,
B_FALSE);
}
dmu_tx_commit(tx);
rw_exit(&zp->z_parent_lock);
rw_exit(&zp->z_name_lock);
out:
zfs_dirent_unlock(dl);
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
zrele(zp);
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Read directory entries from the given directory cursor position and emit
* name and position for each entry.
*
* IN: ip - inode of directory to read.
* ctx - directory entry context.
* cr - credentials of caller.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - atime updated
*
* Note that the low 4 bits of the cookie returned by zap is always zero.
* This allows us to use the low range for "special" directory entries:
* We use 0 for '.', and 1 for '..'. If this is the root of the filesystem,
* we use the offset 2 for the '.zfs' directory.
*/
int
zfs_readdir(struct inode *ip, zpl_dir_context_t *ctx, cred_t *cr)
{
(void) cr;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
objset_t *os;
zap_cursor_t zc;
zap_attribute_t zap;
int error;
uint8_t prefetch;
uint8_t type;
int done = 0;
uint64_t parent;
uint64_t offset; /* must be unsigned; checks for < 1 */
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs),
&parent, sizeof (parent))) != 0)
goto out;
/*
* Quit if directory has been removed (posix)
*/
if (zp->z_unlinked)
goto out;
error = 0;
os = zfsvfs->z_os;
offset = ctx->pos;
prefetch = zp->z_zn_prefetch;
/*
* Initialize the iterator cursor.
*/
if (offset <= 3) {
/*
* Start iteration from the beginning of the directory.
*/
zap_cursor_init(&zc, os, zp->z_id);
} else {
/*
* The offset is a serialized cursor.
*/
zap_cursor_init_serialized(&zc, os, zp->z_id, offset);
}
/*
* Transform to file-system independent format
*/
while (!done) {
uint64_t objnum;
/*
* Special case `.', `..', and `.zfs'.
*/
if (offset == 0) {
(void) strcpy(zap.za_name, ".");
zap.za_normalization_conflict = 0;
objnum = zp->z_id;
type = DT_DIR;
} else if (offset == 1) {
(void) strcpy(zap.za_name, "..");
zap.za_normalization_conflict = 0;
objnum = parent;
type = DT_DIR;
} else if (offset == 2 && zfs_show_ctldir(zp)) {
(void) strcpy(zap.za_name, ZFS_CTLDIR_NAME);
zap.za_normalization_conflict = 0;
objnum = ZFSCTL_INO_ROOT;
type = DT_DIR;
} else {
/*
* Grab next entry.
*/
if ((error = zap_cursor_retrieve(&zc, &zap))) {
if (error == ENOENT)
break;
else
goto update;
}
/*
* Allow multiple entries provided the first entry is
* the object id. Non-zpl consumers may safely make
* use of the additional space.
*
* XXX: This should be a feature flag for compatibility
*/
if (zap.za_integer_length != 8 ||
zap.za_num_integers == 0) {
cmn_err(CE_WARN, "zap_readdir: bad directory "
"entry, obj = %lld, offset = %lld, "
"length = %d, num = %lld\n",
(u_longlong_t)zp->z_id,
(u_longlong_t)offset,
zap.za_integer_length,
(u_longlong_t)zap.za_num_integers);
error = SET_ERROR(ENXIO);
goto update;
}
objnum = ZFS_DIRENT_OBJ(zap.za_first_integer);
type = ZFS_DIRENT_TYPE(zap.za_first_integer);
}
done = !zpl_dir_emit(ctx, zap.za_name, strlen(zap.za_name),
objnum, type);
if (done)
break;
/* Prefetch znode */
if (prefetch) {
dmu_prefetch(os, objnum, 0, 0, 0,
ZIO_PRIORITY_SYNC_READ);
}
/*
* Move to the next entry, fill in the previous offset.
*/
if (offset > 2 || (offset == 2 && !zfs_show_ctldir(zp))) {
zap_cursor_advance(&zc);
offset = zap_cursor_serialize(&zc);
} else {
offset += 1;
}
ctx->pos = offset;
}
zp->z_zn_prefetch = B_FALSE; /* a lookup will re-enable pre-fetching */
update:
zap_cursor_fini(&zc);
if (error == ENOENT)
error = 0;
out:
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Get the basic file attributes and place them in the provided kstat
* structure. The inode is assumed to be the authoritative source
* for most of the attributes. However, the znode currently has the
* authoritative atime, blksize, and block count.
*
* IN: ip - inode of file.
*
* OUT: sp - kstat values.
*
* RETURN: 0 (always succeeds)
*/
int
-zfs_getattr_fast(struct user_namespace *user_ns, struct inode *ip,
- struct kstat *sp)
+zfs_getattr_fast(zidmap_t *user_ns, struct inode *ip, struct kstat *sp)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
uint32_t blksize;
u_longlong_t nblocks;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
mutex_enter(&zp->z_lock);
zpl_generic_fillattr(user_ns, ip, sp);
/*
* +1 link count for root inode with visible '.zfs' directory.
*/
if ((zp->z_id == zfsvfs->z_root) && zfs_show_ctldir(zp))
if (sp->nlink < ZFS_LINK_MAX)
sp->nlink++;
sa_object_size(zp->z_sa_hdl, &blksize, &nblocks);
sp->blksize = blksize;
sp->blocks = nblocks;
if (unlikely(zp->z_blksz == 0)) {
/*
* Block size hasn't been set; suggest maximal I/O transfers.
*/
sp->blksize = zfsvfs->z_max_blksz;
}
mutex_exit(&zp->z_lock);
/*
* Required to prevent NFS client from detecting different inode
* numbers of snapshot root dentry before and after snapshot mount.
*/
if (zfsvfs->z_issnap) {
if (ip->i_sb->s_root->d_inode == ip)
sp->ino = ZFSCTL_INO_SNAPDIRS -
dmu_objset_id(zfsvfs->z_os);
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* For the operation of changing file's user/group/project, we need to
* handle not only the main object that is assigned to the file directly,
* but also the ones that are used by the file via hidden xattr directory.
*
* Because the xattr directory may contains many EA entries, as to it may
* be impossible to change all of them via the transaction of changing the
* main object's user/group/project attributes. Then we have to change them
* via other multiple independent transactions one by one. It may be not good
* solution, but we have no better idea yet.
*/
static int
zfs_setattr_dir(znode_t *dzp)
{
struct inode *dxip = ZTOI(dzp);
struct inode *xip = NULL;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
objset_t *os = zfsvfs->z_os;
zap_cursor_t zc;
zap_attribute_t zap;
zfs_dirlock_t *dl;
znode_t *zp = NULL;
dmu_tx_t *tx = NULL;
uint64_t uid, gid;
sa_bulk_attr_t bulk[4];
int count;
int err;
zap_cursor_init(&zc, os, dzp->z_id);
while ((err = zap_cursor_retrieve(&zc, &zap)) == 0) {
count = 0;
if (zap.za_integer_length != 8 || zap.za_num_integers != 1) {
err = ENXIO;
break;
}
err = zfs_dirent_lock(&dl, dzp, (char *)zap.za_name, &zp,
ZEXISTS, NULL, NULL);
if (err == ENOENT)
goto next;
if (err)
break;
xip = ZTOI(zp);
if (KUID_TO_SUID(xip->i_uid) == KUID_TO_SUID(dxip->i_uid) &&
KGID_TO_SGID(xip->i_gid) == KGID_TO_SGID(dxip->i_gid) &&
zp->z_projid == dzp->z_projid)
goto next;
tx = dmu_tx_create(os);
if (!(zp->z_pflags & ZFS_PROJID))
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
else
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
err = dmu_tx_assign(tx, TXG_WAIT);
if (err)
break;
mutex_enter(&dzp->z_lock);
if (KUID_TO_SUID(xip->i_uid) != KUID_TO_SUID(dxip->i_uid)) {
xip->i_uid = dxip->i_uid;
uid = zfs_uid_read(dxip);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL,
&uid, sizeof (uid));
}
if (KGID_TO_SGID(xip->i_gid) != KGID_TO_SGID(dxip->i_gid)) {
xip->i_gid = dxip->i_gid;
gid = zfs_gid_read(dxip);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL,
&gid, sizeof (gid));
}
if (zp->z_projid != dzp->z_projid) {
if (!(zp->z_pflags & ZFS_PROJID)) {
zp->z_pflags |= ZFS_PROJID;
SA_ADD_BULK_ATTR(bulk, count,
SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags,
sizeof (zp->z_pflags));
}
zp->z_projid = dzp->z_projid;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PROJID(zfsvfs),
NULL, &zp->z_projid, sizeof (zp->z_projid));
}
mutex_exit(&dzp->z_lock);
if (likely(count > 0)) {
err = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
dmu_tx_commit(tx);
} else {
dmu_tx_abort(tx);
}
tx = NULL;
if (err != 0 && err != ENOENT)
break;
next:
if (zp) {
zrele(zp);
zp = NULL;
zfs_dirent_unlock(dl);
}
zap_cursor_advance(&zc);
}
if (tx)
dmu_tx_abort(tx);
if (zp) {
zrele(zp);
zfs_dirent_unlock(dl);
}
zap_cursor_fini(&zc);
return (err == ENOENT ? 0 : err);
}
/*
* Set the file attributes to the values contained in the
* vattr structure.
*
* IN: zp - znode of file to be modified.
* vap - new attribute values.
* If ATTR_XVATTR set, then optional attrs are being set
* flags - ATTR_UTIME set if non-default time values provided.
* - ATTR_NOACLCHECK (CIFS context only).
* cr - credentials of caller.
* mnt_ns - user namespace of the mount
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - ctime updated, mtime updated if size changed.
*/
int
-zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zuserns_t *mnt_ns)
+zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zidmap_t *mnt_ns)
{
struct inode *ip;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
objset_t *os = zfsvfs->z_os;
zilog_t *zilog;
dmu_tx_t *tx;
vattr_t oldva;
xvattr_t *tmpxvattr;
uint_t mask = vap->va_mask;
uint_t saved_mask = 0;
int trim_mask = 0;
uint64_t new_mode;
uint64_t new_kuid = 0, new_kgid = 0, new_uid, new_gid;
uint64_t xattr_obj;
uint64_t mtime[2], ctime[2], atime[2];
uint64_t projid = ZFS_INVALID_PROJID;
znode_t *attrzp;
int need_policy = FALSE;
int err, err2 = 0;
zfs_fuid_info_t *fuidp = NULL;
xvattr_t *xvap = (xvattr_t *)vap; /* vap may be an xvattr_t * */
xoptattr_t *xoap;
zfs_acl_t *aclp;
boolean_t skipaclchk = (flags & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
boolean_t fuid_dirtied = B_FALSE;
boolean_t handle_eadir = B_FALSE;
sa_bulk_attr_t *bulk, *xattr_bulk;
int count = 0, xattr_count = 0, bulks = 8;
if (mask == 0)
return (0);
if ((err = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (err);
ip = ZTOI(zp);
/*
* If this is a xvattr_t, then get a pointer to the structure of
* optional attributes. If this is NULL, then we have a vattr_t.
*/
xoap = xva_getxoptattr(xvap);
if (xoap != NULL && (mask & ATTR_XVATTR)) {
if (XVA_ISSET_REQ(xvap, XAT_PROJID)) {
if (!dmu_objset_projectquota_enabled(os) ||
(!S_ISREG(ip->i_mode) && !S_ISDIR(ip->i_mode))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOTSUP));
}
projid = xoap->xoa_projid;
if (unlikely(projid == ZFS_INVALID_PROJID)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (projid == zp->z_projid && zp->z_pflags & ZFS_PROJID)
projid = ZFS_INVALID_PROJID;
else
need_policy = TRUE;
}
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT) &&
(xoap->xoa_projinherit !=
((zp->z_pflags & ZFS_PROJINHERIT) != 0)) &&
(!dmu_objset_projectquota_enabled(os) ||
(!S_ISREG(ip->i_mode) && !S_ISDIR(ip->i_mode)))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOTSUP));
}
}
zilog = zfsvfs->z_log;
/*
* Make sure that if we have ephemeral uid/gid or xvattr specified
* that file system is at proper version level
*/
if (zfsvfs->z_use_fuids == B_FALSE &&
(((mask & ATTR_UID) && IS_EPHEMERAL(vap->va_uid)) ||
((mask & ATTR_GID) && IS_EPHEMERAL(vap->va_gid)) ||
(mask & ATTR_XVATTR))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
if (mask & ATTR_SIZE && S_ISDIR(ip->i_mode)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EISDIR));
}
if (mask & ATTR_SIZE && !S_ISREG(ip->i_mode) && !S_ISFIFO(ip->i_mode)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
tmpxvattr = kmem_alloc(sizeof (xvattr_t), KM_SLEEP);
xva_init(tmpxvattr);
bulk = kmem_alloc(sizeof (sa_bulk_attr_t) * bulks, KM_SLEEP);
xattr_bulk = kmem_alloc(sizeof (sa_bulk_attr_t) * bulks, KM_SLEEP);
/*
* Immutable files can only alter immutable bit and atime
*/
if ((zp->z_pflags & ZFS_IMMUTABLE) &&
((mask & (ATTR_SIZE|ATTR_UID|ATTR_GID|ATTR_MTIME|ATTR_MODE)) ||
((mask & ATTR_XVATTR) && XVA_ISSET_REQ(xvap, XAT_CREATETIME)))) {
err = SET_ERROR(EPERM);
goto out3;
}
if ((mask & ATTR_SIZE) && (zp->z_pflags & ZFS_READONLY)) {
err = SET_ERROR(EPERM);
goto out3;
}
/*
* Verify timestamps doesn't overflow 32 bits.
* ZFS can handle large timestamps, but 32bit syscalls can't
* handle times greater than 2039. This check should be removed
* once large timestamps are fully supported.
*/
if (mask & (ATTR_ATIME | ATTR_MTIME)) {
if (((mask & ATTR_ATIME) &&
TIMESPEC_OVERFLOW(&vap->va_atime)) ||
((mask & ATTR_MTIME) &&
TIMESPEC_OVERFLOW(&vap->va_mtime))) {
err = SET_ERROR(EOVERFLOW);
goto out3;
}
}
top:
attrzp = NULL;
aclp = NULL;
/* Can this be moved to before the top label? */
if (zfs_is_readonly(zfsvfs)) {
err = SET_ERROR(EROFS);
goto out3;
}
/*
* First validate permissions
*/
if (mask & ATTR_SIZE) {
err = zfs_zaccess(zp, ACE_WRITE_DATA, 0, skipaclchk, cr,
mnt_ns);
if (err)
goto out3;
/*
* XXX - Note, we are not providing any open
* mode flags here (like FNDELAY), so we may
* block if there are locks present... this
* should be addressed in openat().
*/
/* XXX - would it be OK to generate a log record here? */
err = zfs_freesp(zp, vap->va_size, 0, 0, FALSE);
if (err)
goto out3;
}
if (mask & (ATTR_ATIME|ATTR_MTIME) ||
((mask & ATTR_XVATTR) && (XVA_ISSET_REQ(xvap, XAT_HIDDEN) ||
XVA_ISSET_REQ(xvap, XAT_READONLY) ||
XVA_ISSET_REQ(xvap, XAT_ARCHIVE) ||
XVA_ISSET_REQ(xvap, XAT_OFFLINE) ||
XVA_ISSET_REQ(xvap, XAT_SPARSE) ||
XVA_ISSET_REQ(xvap, XAT_CREATETIME) ||
XVA_ISSET_REQ(xvap, XAT_SYSTEM)))) {
need_policy = zfs_zaccess(zp, ACE_WRITE_ATTRIBUTES, 0,
skipaclchk, cr, mnt_ns);
}
if (mask & (ATTR_UID|ATTR_GID)) {
int idmask = (mask & (ATTR_UID|ATTR_GID));
int take_owner;
int take_group;
uid_t uid;
gid_t gid;
/*
* NOTE: even if a new mode is being set,
* we may clear S_ISUID/S_ISGID bits.
*/
if (!(mask & ATTR_MODE))
vap->va_mode = zp->z_mode;
/*
* Take ownership or chgrp to group we are a member of
*/
- uid = zfs_uid_to_vfsuid((struct user_namespace *)mnt_ns,
- zfs_i_user_ns(ip), vap->va_uid);
- gid = zfs_gid_to_vfsgid((struct user_namespace *)mnt_ns,
- zfs_i_user_ns(ip), vap->va_gid);
+ uid = zfs_uid_to_vfsuid(mnt_ns, zfs_i_user_ns(ip),
+ vap->va_uid);
+ gid = zfs_gid_to_vfsgid(mnt_ns, zfs_i_user_ns(ip),
+ vap->va_gid);
take_owner = (mask & ATTR_UID) && (uid == crgetuid(cr));
take_group = (mask & ATTR_GID) &&
zfs_groupmember(zfsvfs, gid, cr);
/*
* If both ATTR_UID and ATTR_GID are set then take_owner and
* take_group must both be set in order to allow taking
* ownership.
*
* Otherwise, send the check through secpolicy_vnode_setattr()
*
*/
if (((idmask == (ATTR_UID|ATTR_GID)) &&
take_owner && take_group) ||
((idmask == ATTR_UID) && take_owner) ||
((idmask == ATTR_GID) && take_group)) {
if (zfs_zaccess(zp, ACE_WRITE_OWNER, 0,
skipaclchk, cr, mnt_ns) == 0) {
/*
* Remove setuid/setgid for non-privileged users
*/
(void) secpolicy_setid_clear(vap, cr);
trim_mask = (mask & (ATTR_UID|ATTR_GID));
} else {
need_policy = TRUE;
}
} else {
need_policy = TRUE;
}
}
mutex_enter(&zp->z_lock);
oldva.va_mode = zp->z_mode;
zfs_fuid_map_ids(zp, cr, &oldva.va_uid, &oldva.va_gid);
if (mask & ATTR_XVATTR) {
/*
* Update xvattr mask to include only those attributes
* that are actually changing.
*
* the bits will be restored prior to actually setting
* the attributes so the caller thinks they were set.
*/
if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) {
if (xoap->xoa_appendonly !=
((zp->z_pflags & ZFS_APPENDONLY) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_APPENDONLY);
XVA_SET_REQ(tmpxvattr, XAT_APPENDONLY);
}
}
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT)) {
if (xoap->xoa_projinherit !=
((zp->z_pflags & ZFS_PROJINHERIT) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_PROJINHERIT);
XVA_SET_REQ(tmpxvattr, XAT_PROJINHERIT);
}
}
if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) {
if (xoap->xoa_nounlink !=
((zp->z_pflags & ZFS_NOUNLINK) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_NOUNLINK);
XVA_SET_REQ(tmpxvattr, XAT_NOUNLINK);
}
}
if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) {
if (xoap->xoa_immutable !=
((zp->z_pflags & ZFS_IMMUTABLE) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_IMMUTABLE);
XVA_SET_REQ(tmpxvattr, XAT_IMMUTABLE);
}
}
if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) {
if (xoap->xoa_nodump !=
((zp->z_pflags & ZFS_NODUMP) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_NODUMP);
XVA_SET_REQ(tmpxvattr, XAT_NODUMP);
}
}
if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) {
if (xoap->xoa_av_modified !=
((zp->z_pflags & ZFS_AV_MODIFIED) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_AV_MODIFIED);
XVA_SET_REQ(tmpxvattr, XAT_AV_MODIFIED);
}
}
if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) {
if ((!S_ISREG(ip->i_mode) &&
xoap->xoa_av_quarantined) ||
xoap->xoa_av_quarantined !=
((zp->z_pflags & ZFS_AV_QUARANTINED) != 0)) {
need_policy = TRUE;
} else {
XVA_CLR_REQ(xvap, XAT_AV_QUARANTINED);
XVA_SET_REQ(tmpxvattr, XAT_AV_QUARANTINED);
}
}
if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) {
mutex_exit(&zp->z_lock);
err = SET_ERROR(EPERM);
goto out3;
}
if (need_policy == FALSE &&
(XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP) ||
XVA_ISSET_REQ(xvap, XAT_OPAQUE))) {
need_policy = TRUE;
}
}
mutex_exit(&zp->z_lock);
if (mask & ATTR_MODE) {
if (zfs_zaccess(zp, ACE_WRITE_ACL, 0, skipaclchk, cr,
mnt_ns) == 0) {
err = secpolicy_setid_setsticky_clear(ip, vap,
&oldva, cr, mnt_ns, zfs_i_user_ns(ip));
if (err)
goto out3;
trim_mask |= ATTR_MODE;
} else {
need_policy = TRUE;
}
}
if (need_policy) {
/*
* If trim_mask is set then take ownership
* has been granted or write_acl is present and user
* has the ability to modify mode. In that case remove
* UID|GID and or MODE from mask so that
* secpolicy_vnode_setattr() doesn't revoke it.
*/
if (trim_mask) {
saved_mask = vap->va_mask;
vap->va_mask &= ~trim_mask;
}
err = secpolicy_vnode_setattr(cr, ip, vap, &oldva, flags,
zfs_zaccess_unix, zp);
if (err)
goto out3;
if (trim_mask)
vap->va_mask |= saved_mask;
}
/*
* secpolicy_vnode_setattr, or take ownership may have
* changed va_mask
*/
mask = vap->va_mask;
if ((mask & (ATTR_UID | ATTR_GID)) || projid != ZFS_INVALID_PROJID) {
handle_eadir = B_TRUE;
err = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs),
&xattr_obj, sizeof (xattr_obj));
if (err == 0 && xattr_obj) {
err = zfs_zget(ZTOZSB(zp), xattr_obj, &attrzp);
if (err)
goto out2;
}
if (mask & ATTR_UID) {
new_kuid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_uid, cr, ZFS_OWNER, &fuidp);
if (new_kuid != KUID_TO_SUID(ZTOI(zp)->i_uid) &&
zfs_id_overquota(zfsvfs, DMU_USERUSED_OBJECT,
new_kuid)) {
if (attrzp)
zrele(attrzp);
err = SET_ERROR(EDQUOT);
goto out2;
}
}
if (mask & ATTR_GID) {
new_kgid = zfs_fuid_create(zfsvfs,
(uint64_t)vap->va_gid, cr, ZFS_GROUP, &fuidp);
if (new_kgid != KGID_TO_SGID(ZTOI(zp)->i_gid) &&
zfs_id_overquota(zfsvfs, DMU_GROUPUSED_OBJECT,
new_kgid)) {
if (attrzp)
zrele(attrzp);
err = SET_ERROR(EDQUOT);
goto out2;
}
}
if (projid != ZFS_INVALID_PROJID &&
zfs_id_overquota(zfsvfs, DMU_PROJECTUSED_OBJECT, projid)) {
if (attrzp)
zrele(attrzp);
err = EDQUOT;
goto out2;
}
}
tx = dmu_tx_create(os);
if (mask & ATTR_MODE) {
uint64_t pmode = zp->z_mode;
uint64_t acl_obj;
new_mode = (pmode & S_IFMT) | (vap->va_mode & ~S_IFMT);
if (ZTOZSB(zp)->z_acl_mode == ZFS_ACL_RESTRICTED &&
!(zp->z_pflags & ZFS_ACL_TRIVIAL)) {
err = EPERM;
goto out;
}
if ((err = zfs_acl_chmod_setattr(zp, &aclp, new_mode)))
goto out;
mutex_enter(&zp->z_lock);
if (!zp->z_is_sa && ((acl_obj = zfs_external_acl(zp)) != 0)) {
/*
* Are we upgrading ACL from old V0 format
* to V1 format?
*/
if (zfsvfs->z_version >= ZPL_VERSION_FUID &&
zfs_znode_acl_version(zp) ==
ZFS_ACL_VERSION_INITIAL) {
dmu_tx_hold_free(tx, acl_obj, 0,
DMU_OBJECT_END);
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, aclp->z_acl_bytes);
} else {
dmu_tx_hold_write(tx, acl_obj, 0,
aclp->z_acl_bytes);
}
} else if (!zp->z_is_sa && aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, aclp->z_acl_bytes);
}
mutex_exit(&zp->z_lock);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
} else {
if (((mask & ATTR_XVATTR) &&
XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) ||
(projid != ZFS_INVALID_PROJID &&
!(zp->z_pflags & ZFS_PROJID)))
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE);
else
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
}
if (attrzp) {
dmu_tx_hold_sa(tx, attrzp->z_sa_hdl, B_FALSE);
}
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
zfs_sa_upgrade_txholds(tx, zp);
err = dmu_tx_assign(tx, TXG_WAIT);
if (err)
goto out;
count = 0;
/*
* Set each attribute requested.
* We group settings according to the locks they need to acquire.
*
* Note: you cannot set ctime directly, although it will be
* updated as a side-effect of calling this function.
*/
if (projid != ZFS_INVALID_PROJID && !(zp->z_pflags & ZFS_PROJID)) {
/*
* For the existed object that is upgraded from old system,
* its on-disk layout has no slot for the project ID attribute.
* But quota accounting logic needs to access related slots by
* offset directly. So we need to adjust old objects' layout
* to make the project ID to some unified and fixed offset.
*/
if (attrzp)
err = sa_add_projid(attrzp->z_sa_hdl, tx, projid);
if (err == 0)
err = sa_add_projid(zp->z_sa_hdl, tx, projid);
if (unlikely(err == EEXIST))
err = 0;
else if (err != 0)
goto out;
else
projid = ZFS_INVALID_PROJID;
}
if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))
mutex_enter(&zp->z_acl_lock);
mutex_enter(&zp->z_lock);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, sizeof (zp->z_pflags));
if (attrzp) {
if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))
mutex_enter(&attrzp->z_acl_lock);
mutex_enter(&attrzp->z_lock);
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_FLAGS(zfsvfs), NULL, &attrzp->z_pflags,
sizeof (attrzp->z_pflags));
if (projid != ZFS_INVALID_PROJID) {
attrzp->z_projid = projid;
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_PROJID(zfsvfs), NULL, &attrzp->z_projid,
sizeof (attrzp->z_projid));
}
}
if (mask & (ATTR_UID|ATTR_GID)) {
if (mask & ATTR_UID) {
ZTOI(zp)->i_uid = SUID_TO_KUID(new_kuid);
new_uid = zfs_uid_read(ZTOI(zp));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL,
&new_uid, sizeof (new_uid));
if (attrzp) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_UID(zfsvfs), NULL, &new_uid,
sizeof (new_uid));
ZTOI(attrzp)->i_uid = SUID_TO_KUID(new_uid);
}
}
if (mask & ATTR_GID) {
ZTOI(zp)->i_gid = SGID_TO_KGID(new_kgid);
new_gid = zfs_gid_read(ZTOI(zp));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs),
NULL, &new_gid, sizeof (new_gid));
if (attrzp) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_GID(zfsvfs), NULL, &new_gid,
sizeof (new_gid));
ZTOI(attrzp)->i_gid = SGID_TO_KGID(new_kgid);
}
}
if (!(mask & ATTR_MODE)) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs),
NULL, &new_mode, sizeof (new_mode));
new_mode = zp->z_mode;
}
err = zfs_acl_chown_setattr(zp);
ASSERT(err == 0);
if (attrzp) {
err = zfs_acl_chown_setattr(attrzp);
ASSERT(err == 0);
}
}
if (mask & ATTR_MODE) {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL,
&new_mode, sizeof (new_mode));
zp->z_mode = ZTOI(zp)->i_mode = new_mode;
ASSERT3P(aclp, !=, NULL);
err = zfs_aclset_common(zp, aclp, cr, tx);
ASSERT0(err);
if (zp->z_acl_cached)
zfs_acl_free(zp->z_acl_cached);
zp->z_acl_cached = aclp;
aclp = NULL;
}
if ((mask & ATTR_ATIME) || zp->z_atime_dirty) {
zp->z_atime_dirty = B_FALSE;
ZFS_TIME_ENCODE(&ip->i_atime, atime);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL,
&atime, sizeof (atime));
}
if (mask & (ATTR_MTIME | ATTR_SIZE)) {
ZFS_TIME_ENCODE(&vap->va_mtime, mtime);
ZTOI(zp)->i_mtime = zpl_inode_timestamp_truncate(
vap->va_mtime, ZTOI(zp));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL,
mtime, sizeof (mtime));
}
if (mask & (ATTR_CTIME | ATTR_SIZE)) {
ZFS_TIME_ENCODE(&vap->va_ctime, ctime);
ZTOI(zp)->i_ctime = zpl_inode_timestamp_truncate(vap->va_ctime,
ZTOI(zp));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
ctime, sizeof (ctime));
}
if (projid != ZFS_INVALID_PROJID) {
zp->z_projid = projid;
SA_ADD_BULK_ATTR(bulk, count,
SA_ZPL_PROJID(zfsvfs), NULL, &zp->z_projid,
sizeof (zp->z_projid));
}
if (attrzp && mask) {
SA_ADD_BULK_ATTR(xattr_bulk, xattr_count,
SA_ZPL_CTIME(zfsvfs), NULL, &ctime,
sizeof (ctime));
}
/*
* Do this after setting timestamps to prevent timestamp
* update from toggling bit
*/
if (xoap && (mask & ATTR_XVATTR)) {
/*
* restore trimmed off masks
* so that return masks can be set for caller.
*/
if (XVA_ISSET_REQ(tmpxvattr, XAT_APPENDONLY)) {
XVA_SET_REQ(xvap, XAT_APPENDONLY);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_NOUNLINK)) {
XVA_SET_REQ(xvap, XAT_NOUNLINK);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_IMMUTABLE)) {
XVA_SET_REQ(xvap, XAT_IMMUTABLE);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_NODUMP)) {
XVA_SET_REQ(xvap, XAT_NODUMP);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_AV_MODIFIED)) {
XVA_SET_REQ(xvap, XAT_AV_MODIFIED);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_AV_QUARANTINED)) {
XVA_SET_REQ(xvap, XAT_AV_QUARANTINED);
}
if (XVA_ISSET_REQ(tmpxvattr, XAT_PROJINHERIT)) {
XVA_SET_REQ(xvap, XAT_PROJINHERIT);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP))
ASSERT(S_ISREG(ip->i_mode));
zfs_xvattr_set(zp, xvap, tx);
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
if (mask != 0)
zfs_log_setattr(zilog, tx, TX_SETATTR, zp, vap, mask, fuidp);
mutex_exit(&zp->z_lock);
if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))
mutex_exit(&zp->z_acl_lock);
if (attrzp) {
if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE))
mutex_exit(&attrzp->z_acl_lock);
mutex_exit(&attrzp->z_lock);
}
out:
if (err == 0 && xattr_count > 0) {
err2 = sa_bulk_update(attrzp->z_sa_hdl, xattr_bulk,
xattr_count, tx);
ASSERT(err2 == 0);
}
if (aclp)
zfs_acl_free(aclp);
if (fuidp) {
zfs_fuid_info_free(fuidp);
fuidp = NULL;
}
if (err) {
dmu_tx_abort(tx);
if (attrzp)
zrele(attrzp);
if (err == ERESTART)
goto top;
} else {
if (count > 0)
err2 = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
dmu_tx_commit(tx);
if (attrzp) {
if (err2 == 0 && handle_eadir)
err = zfs_setattr_dir(attrzp);
zrele(attrzp);
}
zfs_znode_update_vfs(zp);
}
out2:
if (os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
out3:
kmem_free(xattr_bulk, sizeof (sa_bulk_attr_t) * bulks);
kmem_free(bulk, sizeof (sa_bulk_attr_t) * bulks);
kmem_free(tmpxvattr, sizeof (xvattr_t));
zfs_exit(zfsvfs, FTAG);
return (err);
}
typedef struct zfs_zlock {
krwlock_t *zl_rwlock; /* lock we acquired */
znode_t *zl_znode; /* znode we held */
struct zfs_zlock *zl_next; /* next in list */
} zfs_zlock_t;
/*
* Drop locks and release vnodes that were held by zfs_rename_lock().
*/
static void
zfs_rename_unlock(zfs_zlock_t **zlpp)
{
zfs_zlock_t *zl;
while ((zl = *zlpp) != NULL) {
if (zl->zl_znode != NULL)
zfs_zrele_async(zl->zl_znode);
rw_exit(zl->zl_rwlock);
*zlpp = zl->zl_next;
kmem_free(zl, sizeof (*zl));
}
}
/*
* Search back through the directory tree, using the ".." entries.
* Lock each directory in the chain to prevent concurrent renames.
* Fail any attempt to move a directory into one of its own descendants.
* XXX - z_parent_lock can overlap with map or grow locks
*/
static int
zfs_rename_lock(znode_t *szp, znode_t *tdzp, znode_t *sdzp, zfs_zlock_t **zlpp)
{
zfs_zlock_t *zl;
znode_t *zp = tdzp;
uint64_t rootid = ZTOZSB(zp)->z_root;
uint64_t oidp = zp->z_id;
krwlock_t *rwlp = &szp->z_parent_lock;
krw_t rw = RW_WRITER;
/*
* First pass write-locks szp and compares to zp->z_id.
* Later passes read-lock zp and compare to zp->z_parent.
*/
do {
if (!rw_tryenter(rwlp, rw)) {
/*
* Another thread is renaming in this path.
* Note that if we are a WRITER, we don't have any
* parent_locks held yet.
*/
if (rw == RW_READER && zp->z_id > szp->z_id) {
/*
* Drop our locks and restart
*/
zfs_rename_unlock(&zl);
*zlpp = NULL;
zp = tdzp;
oidp = zp->z_id;
rwlp = &szp->z_parent_lock;
rw = RW_WRITER;
continue;
} else {
/*
* Wait for other thread to drop its locks
*/
rw_enter(rwlp, rw);
}
}
zl = kmem_alloc(sizeof (*zl), KM_SLEEP);
zl->zl_rwlock = rwlp;
zl->zl_znode = NULL;
zl->zl_next = *zlpp;
*zlpp = zl;
if (oidp == szp->z_id) /* We're a descendant of szp */
return (SET_ERROR(EINVAL));
if (oidp == rootid) /* We've hit the top */
return (0);
if (rw == RW_READER) { /* i.e. not the first pass */
int error = zfs_zget(ZTOZSB(zp), oidp, &zp);
if (error)
return (error);
zl->zl_znode = zp;
}
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(ZTOZSB(zp)),
&oidp, sizeof (oidp));
rwlp = &zp->z_parent_lock;
rw = RW_READER;
} while (zp->z_id != sdzp->z_id);
return (0);
}
/*
* Move an entry from the provided source directory to the target
* directory. Change the entry name as indicated.
*
* IN: sdzp - Source directory containing the "old entry".
* snm - Old entry name.
* tdzp - Target directory to contain the "new entry".
* tnm - New entry name.
* cr - credentials of caller.
* flags - case flags
* rflags - RENAME_* flags
* wa_vap - attributes for RENAME_WHITEOUT (must be a char 0:0).
* mnt_ns - user namespace of the mount
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* sdzp,tdzp - ctime|mtime updated
*/
int
zfs_rename(znode_t *sdzp, char *snm, znode_t *tdzp, char *tnm,
- cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap, zuserns_t *mnt_ns)
+ cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap, zidmap_t *mnt_ns)
{
znode_t *szp, *tzp;
zfsvfs_t *zfsvfs = ZTOZSB(sdzp);
zilog_t *zilog;
zfs_dirlock_t *sdl, *tdl;
dmu_tx_t *tx;
zfs_zlock_t *zl;
int cmp, serr, terr;
int error = 0;
int zflg = 0;
boolean_t waited = B_FALSE;
/* Needed for whiteout inode creation. */
boolean_t fuid_dirtied;
zfs_acl_ids_t acl_ids;
boolean_t have_acl = B_FALSE;
znode_t *wzp = NULL;
if (snm == NULL || tnm == NULL)
return (SET_ERROR(EINVAL));
if (rflags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE | RENAME_WHITEOUT))
return (SET_ERROR(EINVAL));
/* Already checked by Linux VFS, but just to make sure. */
if (rflags & RENAME_EXCHANGE &&
(rflags & (RENAME_NOREPLACE | RENAME_WHITEOUT)))
return (SET_ERROR(EINVAL));
/*
* Make sure we only get wo_vap iff. RENAME_WHITEOUT and that it's the
* right kind of vattr_t for the whiteout file. These are set
* internally by ZFS so should never be incorrect.
*/
VERIFY_EQUIV(rflags & RENAME_WHITEOUT, wo_vap != NULL);
VERIFY_IMPLY(wo_vap, wo_vap->va_mode == S_IFCHR);
VERIFY_IMPLY(wo_vap, wo_vap->va_rdev == makedevice(0, 0));
if ((error = zfs_enter_verify_zp(zfsvfs, sdzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if ((error = zfs_verify_zp(tdzp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* We check i_sb because snapshots and the ctldir must have different
* super blocks.
*/
if (ZTOI(tdzp)->i_sb != ZTOI(sdzp)->i_sb ||
zfsctl_is_node(ZTOI(tdzp))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
if (zfsvfs->z_utf8 && u8_validate(tnm,
strlen(tnm), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (flags & FIGNORECASE)
zflg |= ZCILOOK;
top:
szp = NULL;
tzp = NULL;
zl = NULL;
/*
* This is to prevent the creation of links into attribute space
* by renaming a linked file into/outof an attribute directory.
* See the comment in zfs_link() for why this is considered bad.
*/
if ((tdzp->z_pflags & ZFS_XATTR) != (sdzp->z_pflags & ZFS_XATTR)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Lock source and target directory entries. To prevent deadlock,
* a lock ordering must be defined. We lock the directory with
* the smallest object id first, or if it's a tie, the one with
* the lexically first name.
*/
if (sdzp->z_id < tdzp->z_id) {
cmp = -1;
} else if (sdzp->z_id > tdzp->z_id) {
cmp = 1;
} else {
/*
* First compare the two name arguments without
* considering any case folding.
*/
int nofold = (zfsvfs->z_norm & ~U8_TEXTPREP_TOUPPER);
cmp = u8_strcmp(snm, tnm, 0, nofold, U8_UNICODE_LATEST, &error);
ASSERT(error == 0 || !zfsvfs->z_utf8);
if (cmp == 0) {
/*
* POSIX: "If the old argument and the new argument
* both refer to links to the same existing file,
* the rename() function shall return successfully
* and perform no other action."
*/
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* If the file system is case-folding, then we may
* have some more checking to do. A case-folding file
* system is either supporting mixed case sensitivity
* access or is completely case-insensitive. Note
* that the file system is always case preserving.
*
* In mixed sensitivity mode case sensitive behavior
* is the default. FIGNORECASE must be used to
* explicitly request case insensitive behavior.
*
* If the source and target names provided differ only
* by case (e.g., a request to rename 'tim' to 'Tim'),
* we will treat this as a special case in the
* case-insensitive mode: as long as the source name
* is an exact match, we will allow this to proceed as
* a name-change request.
*/
if ((zfsvfs->z_case == ZFS_CASE_INSENSITIVE ||
(zfsvfs->z_case == ZFS_CASE_MIXED &&
flags & FIGNORECASE)) &&
u8_strcmp(snm, tnm, 0, zfsvfs->z_norm, U8_UNICODE_LATEST,
&error) == 0) {
/*
* case preserving rename request, require exact
* name matches
*/
zflg |= ZCIEXACT;
zflg &= ~ZCILOOK;
}
}
/*
* If the source and destination directories are the same, we should
* grab the z_name_lock of that directory only once.
*/
if (sdzp == tdzp) {
zflg |= ZHAVELOCK;
rw_enter(&sdzp->z_name_lock, RW_READER);
}
if (cmp < 0) {
serr = zfs_dirent_lock(&sdl, sdzp, snm, &szp,
ZEXISTS | zflg, NULL, NULL);
terr = zfs_dirent_lock(&tdl,
tdzp, tnm, &tzp, ZRENAMING | zflg, NULL, NULL);
} else {
terr = zfs_dirent_lock(&tdl,
tdzp, tnm, &tzp, zflg, NULL, NULL);
serr = zfs_dirent_lock(&sdl,
sdzp, snm, &szp, ZEXISTS | ZRENAMING | zflg,
NULL, NULL);
}
if (serr) {
/*
* Source entry invalid or not there.
*/
if (!terr) {
zfs_dirent_unlock(tdl);
if (tzp)
zrele(tzp);
}
if (sdzp == tdzp)
rw_exit(&sdzp->z_name_lock);
if (strcmp(snm, "..") == 0)
serr = EINVAL;
zfs_exit(zfsvfs, FTAG);
return (serr);
}
if (terr) {
zfs_dirent_unlock(sdl);
zrele(szp);
if (sdzp == tdzp)
rw_exit(&sdzp->z_name_lock);
if (strcmp(tnm, "..") == 0)
terr = EINVAL;
zfs_exit(zfsvfs, FTAG);
return (terr);
}
/*
* If we are using project inheritance, means if the directory has
* ZFS_PROJINHERIT set, then its descendant directories will inherit
* not only the project ID, but also the ZFS_PROJINHERIT flag. Under
* such case, we only allow renames into our tree when the project
* IDs are the same.
*/
if (tdzp->z_pflags & ZFS_PROJINHERIT &&
tdzp->z_projid != szp->z_projid) {
error = SET_ERROR(EXDEV);
goto out;
}
/*
* Must have write access at the source to remove the old entry
* and write access at the target to create the new entry.
* Note that if target and source are the same, this can be
* done in a single check.
*/
if ((error = zfs_zaccess_rename(sdzp, szp, tdzp, tzp, cr, mnt_ns)))
goto out;
if (S_ISDIR(ZTOI(szp)->i_mode)) {
/*
* Check to make sure rename is valid.
* Can't do a move like this: /usr/a/b to /usr/a/b/c/d
*/
if ((error = zfs_rename_lock(szp, tdzp, sdzp, &zl)))
goto out;
}
/*
* Does target exist?
*/
if (tzp) {
if (rflags & RENAME_NOREPLACE) {
error = SET_ERROR(EEXIST);
goto out;
}
/*
* Source and target must be the same type (unless exchanging).
*/
if (!(rflags & RENAME_EXCHANGE)) {
boolean_t s_is_dir = S_ISDIR(ZTOI(szp)->i_mode) != 0;
boolean_t t_is_dir = S_ISDIR(ZTOI(tzp)->i_mode) != 0;
if (s_is_dir != t_is_dir) {
error = SET_ERROR(s_is_dir ? ENOTDIR : EISDIR);
goto out;
}
}
/*
* POSIX dictates that when the source and target
* entries refer to the same file object, rename
* must do nothing and exit without error.
*/
if (szp->z_id == tzp->z_id) {
error = 0;
goto out;
}
} else if (rflags & RENAME_EXCHANGE) {
/* Target must exist for RENAME_EXCHANGE. */
error = SET_ERROR(ENOENT);
goto out;
}
/* Set up inode creation for RENAME_WHITEOUT. */
if (rflags & RENAME_WHITEOUT) {
/*
* Whiteout files are not regular files or directories, so to
* match zfs_create() we do not inherit the project id.
*/
uint64_t wo_projid = ZFS_DEFAULT_PROJID;
error = zfs_zaccess(sdzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns);
if (error)
goto out;
if (!have_acl) {
error = zfs_acl_ids_create(sdzp, 0, wo_vap, cr, NULL,
&acl_ids, mnt_ns);
if (error)
goto out;
have_acl = B_TRUE;
}
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, wo_projid)) {
error = SET_ERROR(EDQUOT);
goto out;
}
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE);
dmu_tx_hold_sa(tx, sdzp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, sdzp->z_id,
(rflags & RENAME_EXCHANGE) ? TRUE : FALSE, snm);
dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, tnm);
if (sdzp != tdzp) {
dmu_tx_hold_sa(tx, tdzp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, tdzp);
}
if (tzp) {
dmu_tx_hold_sa(tx, tzp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, tzp);
}
if (rflags & RENAME_WHITEOUT) {
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
dmu_tx_hold_zap(tx, sdzp->z_id, TRUE, snm);
dmu_tx_hold_sa(tx, sdzp->z_sa_hdl, B_FALSE);
if (!zfsvfs->z_use_sa &&
acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT,
0, acl_ids.z_aclp->z_acl_bytes);
}
}
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
zfs_sa_upgrade_txholds(tx, szp);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
if (zl != NULL)
zfs_rename_unlock(&zl);
zfs_dirent_unlock(sdl);
zfs_dirent_unlock(tdl);
if (sdzp == tdzp)
rw_exit(&sdzp->z_name_lock);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
zrele(szp);
if (tzp)
zrele(tzp);
goto top;
}
dmu_tx_abort(tx);
zrele(szp);
if (tzp)
zrele(tzp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Unlink the source.
*/
szp->z_pflags |= ZFS_AV_MODIFIED;
if (tdzp->z_pflags & ZFS_PROJINHERIT)
szp->z_pflags |= ZFS_PROJINHERIT;
error = sa_update(szp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs),
(void *)&szp->z_pflags, sizeof (uint64_t), tx);
VERIFY0(error);
error = zfs_link_destroy(sdl, szp, tx, ZRENAMING, NULL);
if (error)
goto commit;
/*
* Unlink the target.
*/
if (tzp) {
int tzflg = zflg;
if (rflags & RENAME_EXCHANGE) {
/* This inode will be re-linked soon. */
tzflg |= ZRENAMING;
tzp->z_pflags |= ZFS_AV_MODIFIED;
if (sdzp->z_pflags & ZFS_PROJINHERIT)
tzp->z_pflags |= ZFS_PROJINHERIT;
error = sa_update(tzp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs),
(void *)&tzp->z_pflags, sizeof (uint64_t), tx);
ASSERT0(error);
}
error = zfs_link_destroy(tdl, tzp, tx, tzflg, NULL);
if (error)
goto commit_link_szp;
}
/*
* Create the new target links:
* * We always link the target.
* * RENAME_EXCHANGE: Link the old target to the source.
* * RENAME_WHITEOUT: Create a whiteout inode in-place of the source.
*/
error = zfs_link_create(tdl, szp, tx, ZRENAMING);
if (error) {
/*
* If we have removed the existing target, a subsequent call to
* zfs_link_create() to add back the same entry, but with a new
* dnode (szp), should not fail.
*/
ASSERT3P(tzp, ==, NULL);
goto commit_link_tzp;
}
switch (rflags & (RENAME_EXCHANGE | RENAME_WHITEOUT)) {
case RENAME_EXCHANGE:
error = zfs_link_create(sdl, tzp, tx, ZRENAMING);
/*
* The same argument as zfs_link_create() failing for
* szp applies here, since the source directory must
* have had an entry we are replacing.
*/
ASSERT0(error);
if (error)
goto commit_unlink_td_szp;
break;
case RENAME_WHITEOUT:
zfs_mknode(sdzp, wo_vap, tx, cr, 0, &wzp, &acl_ids);
error = zfs_link_create(sdl, wzp, tx, ZNEW);
if (error) {
zfs_znode_delete(wzp, tx);
remove_inode_hash(ZTOI(wzp));
goto commit_unlink_td_szp;
}
break;
}
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
switch (rflags & (RENAME_EXCHANGE | RENAME_WHITEOUT)) {
case RENAME_EXCHANGE:
zfs_log_rename_exchange(zilog, tx,
(flags & FIGNORECASE ? TX_CI : 0), sdzp, sdl->dl_name,
tdzp, tdl->dl_name, szp);
break;
case RENAME_WHITEOUT:
zfs_log_rename_whiteout(zilog, tx,
(flags & FIGNORECASE ? TX_CI : 0), sdzp, sdl->dl_name,
tdzp, tdl->dl_name, szp, wzp);
break;
default:
ASSERT0(rflags & ~RENAME_NOREPLACE);
zfs_log_rename(zilog, tx, (flags & FIGNORECASE ? TX_CI : 0),
sdzp, sdl->dl_name, tdzp, tdl->dl_name, szp);
break;
}
commit:
dmu_tx_commit(tx);
out:
if (have_acl)
zfs_acl_ids_free(&acl_ids);
zfs_znode_update_vfs(sdzp);
if (sdzp == tdzp)
rw_exit(&sdzp->z_name_lock);
if (sdzp != tdzp)
zfs_znode_update_vfs(tdzp);
zfs_znode_update_vfs(szp);
zrele(szp);
if (wzp) {
zfs_znode_update_vfs(wzp);
zrele(wzp);
}
if (tzp) {
zfs_znode_update_vfs(tzp);
zrele(tzp);
}
if (zl != NULL)
zfs_rename_unlock(&zl);
zfs_dirent_unlock(sdl);
zfs_dirent_unlock(tdl);
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
/*
* Clean-up path for broken link state.
*
* At this point we are in a (very) bad state, so we need to do our
* best to correct the state. In particular, all of the nlinks are
* wrong because we were destroying and creating links with ZRENAMING.
*
* In some form, all of these operations have to resolve the state:
*
* * link_destroy() *must* succeed. Fortunately, this is very likely
* since we only just created it.
*
* * link_create()s are allowed to fail (though they shouldn't because
* we only just unlinked them and are putting the entries back
* during clean-up). But if they fail, we can just forcefully drop
* the nlink value to (at the very least) avoid broken nlink values
* -- though in the case of non-empty directories we will have to
* panic (otherwise we'd have a leaked directory with a broken ..).
*/
commit_unlink_td_szp:
VERIFY0(zfs_link_destroy(tdl, szp, tx, ZRENAMING, NULL));
commit_link_tzp:
if (tzp) {
if (zfs_link_create(tdl, tzp, tx, ZRENAMING))
VERIFY0(zfs_drop_nlink(tzp, tx, NULL));
}
commit_link_szp:
if (zfs_link_create(sdl, szp, tx, ZRENAMING))
VERIFY0(zfs_drop_nlink(szp, tx, NULL));
goto commit;
}
/*
* Insert the indicated symbolic reference entry into the directory.
*
* IN: dzp - Directory to contain new symbolic link.
* name - Name of directory entry in dip.
* vap - Attributes of new entry.
* link - Name for new symlink entry.
* cr - credentials of caller.
* flags - case flags
* mnt_ns - user namespace of the mount
*
* OUT: zpp - Znode for new symbolic link.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* dip - ctime|mtime updated
*/
int
zfs_symlink(znode_t *dzp, char *name, vattr_t *vap, char *link,
- znode_t **zpp, cred_t *cr, int flags, zuserns_t *mnt_ns)
+ znode_t **zpp, cred_t *cr, int flags, zidmap_t *mnt_ns)
{
znode_t *zp;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zilog_t *zilog;
uint64_t len = strlen(link);
int error;
int zflg = ZNEW;
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
uint64_t txtype = TX_SYMLINK;
boolean_t waited = B_FALSE;
ASSERT(S_ISLNK(vap->va_mode));
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
if (zfsvfs->z_utf8 && u8_validate(name, strlen(name),
NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (flags & FIGNORECASE)
zflg |= ZCILOOK;
if (len > MAXPATHLEN) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENAMETOOLONG));
}
if ((error = zfs_acl_ids_create(dzp, 0,
vap, cr, NULL, &acl_ids, mnt_ns)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
top:
*zpp = NULL;
/*
* Attempt to lock directory; fail if entry already exists.
*/
error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg, NULL, NULL);
if (error) {
zfs_acl_ids_free(&acl_ids);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) {
zfs_acl_ids_free(&acl_ids);
zfs_dirent_unlock(dl);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, ZFS_DEFAULT_PROJID)) {
zfs_acl_ids_free(&acl_ids);
zfs_dirent_unlock(dl);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EDQUOT));
}
tx = dmu_tx_create(zfsvfs->z_os);
fuid_dirtied = zfsvfs->z_fuid_dirty;
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, MAX(1, len));
dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE + len);
dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE);
if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) {
dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0,
acl_ids.z_aclp->z_acl_bytes);
}
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Create a new object for the symlink.
* for version 4 ZPL datasets the symlink will be an SA attribute
*/
zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids);
if (fuid_dirtied)
zfs_fuid_sync(zfsvfs, tx);
mutex_enter(&zp->z_lock);
if (zp->z_is_sa)
error = sa_update(zp->z_sa_hdl, SA_ZPL_SYMLINK(zfsvfs),
link, len, tx);
else
zfs_sa_symlink(zp, link, len, tx);
mutex_exit(&zp->z_lock);
zp->z_size = len;
(void) sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zfsvfs),
&zp->z_size, sizeof (zp->z_size), tx);
/*
* Insert the new object into the directory.
*/
error = zfs_link_create(dl, zp, tx, ZNEW);
if (error != 0) {
zfs_znode_delete(zp, tx);
remove_inode_hash(ZTOI(zp));
} else {
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_symlink(zilog, tx, txtype, dzp, zp, name, link);
zfs_znode_update_vfs(dzp);
zfs_znode_update_vfs(zp);
}
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
zfs_dirent_unlock(dl);
if (error == 0) {
*zpp = zp;
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
} else {
zrele(zp);
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Return, in the buffer contained in the provided uio structure,
* the symbolic path referred to by ip.
*
* IN: ip - inode of symbolic link
* uio - structure to contain the link path.
* cr - credentials of caller.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - atime updated
*/
int
zfs_readlink(struct inode *ip, zfs_uio_t *uio, cred_t *cr)
{
(void) cr;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
mutex_enter(&zp->z_lock);
if (zp->z_is_sa)
error = sa_lookup_uio(zp->z_sa_hdl,
SA_ZPL_SYMLINK(zfsvfs), uio);
else
error = zfs_sa_readlink(zp, uio);
mutex_exit(&zp->z_lock);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Insert a new entry into directory tdzp referencing szp.
*
* IN: tdzp - Directory to contain new entry.
* szp - znode of new entry.
* name - name of new entry.
* cr - credentials of caller.
* flags - case flags.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* tdzp - ctime|mtime updated
* szp - ctime updated
*/
int
zfs_link(znode_t *tdzp, znode_t *szp, char *name, cred_t *cr,
int flags)
{
struct inode *sip = ZTOI(szp);
znode_t *tzp;
zfsvfs_t *zfsvfs = ZTOZSB(tdzp);
zilog_t *zilog;
zfs_dirlock_t *dl;
dmu_tx_t *tx;
int error;
int zf = ZNEW;
uint64_t parent;
uid_t owner;
boolean_t waited = B_FALSE;
boolean_t is_tmpfile = 0;
uint64_t txg;
#ifdef HAVE_TMPFILE
is_tmpfile = (sip->i_nlink == 0 && (sip->i_state & I_LINKABLE));
#endif
ASSERT(S_ISDIR(ZTOI(tdzp)->i_mode));
if (name == NULL)
return (SET_ERROR(EINVAL));
if ((error = zfs_enter_verify_zp(zfsvfs, tdzp, FTAG)) != 0)
return (error);
zilog = zfsvfs->z_log;
/*
* POSIX dictates that we return EPERM here.
* Better choices include ENOTSUP or EISDIR.
*/
if (S_ISDIR(sip->i_mode)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if ((error = zfs_verify_zp(szp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* If we are using project inheritance, means if the directory has
* ZFS_PROJINHERIT set, then its descendant directories will inherit
* not only the project ID, but also the ZFS_PROJINHERIT flag. Under
* such case, we only allow hard link creation in our tree when the
* project IDs are the same.
*/
if (tdzp->z_pflags & ZFS_PROJINHERIT &&
tdzp->z_projid != szp->z_projid) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
/*
* We check i_sb because snapshots and the ctldir must have different
* super blocks.
*/
if (sip->i_sb != ZTOI(tdzp)->i_sb || zfsctl_is_node(sip)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
/* Prevent links to .zfs/shares files */
if ((error = sa_lookup(szp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs),
&parent, sizeof (uint64_t))) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (parent == zfsvfs->z_shares_dir) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if (zfsvfs->z_utf8 && u8_validate(name,
strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EILSEQ));
}
if (flags & FIGNORECASE)
zf |= ZCILOOK;
/*
* We do not support links between attributes and non-attributes
* because of the potential security risk of creating links
* into "normal" file space in order to circumvent restrictions
* imposed in attribute space.
*/
if ((szp->z_pflags & ZFS_XATTR) != (tdzp->z_pflags & ZFS_XATTR)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
owner = zfs_fuid_map_id(zfsvfs, KUID_TO_SUID(sip->i_uid),
cr, ZFS_OWNER);
if (owner != crgetuid(cr) && secpolicy_basic_link(cr) != 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if ((error = zfs_zaccess(tdzp, ACE_ADD_FILE, 0, B_FALSE, cr,
- kcred->user_ns))) {
+ zfs_init_idmap))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
top:
/*
* Attempt to lock directory; fail if entry already exists.
*/
error = zfs_dirent_lock(&dl, tdzp, name, &tzp, zf, NULL, NULL);
if (error) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, name);
if (is_tmpfile)
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
zfs_sa_upgrade_txholds(tx, szp);
zfs_sa_upgrade_txholds(tx, tdzp);
error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT);
if (error) {
zfs_dirent_unlock(dl);
if (error == ERESTART) {
waited = B_TRUE;
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/* unmark z_unlinked so zfs_link_create will not reject */
if (is_tmpfile)
szp->z_unlinked = B_FALSE;
error = zfs_link_create(dl, szp, tx, 0);
if (error == 0) {
uint64_t txtype = TX_LINK;
/*
* tmpfile is created to be in z_unlinkedobj, so remove it.
* Also, we don't log in ZIL, because all previous file
* operation on the tmpfile are ignored by ZIL. Instead we
* always wait for txg to sync to make sure all previous
* operation are sync safe.
*/
if (is_tmpfile) {
VERIFY(zap_remove_int(zfsvfs->z_os,
zfsvfs->z_unlinkedobj, szp->z_id, tx) == 0);
} else {
if (flags & FIGNORECASE)
txtype |= TX_CI;
zfs_log_link(zilog, tx, txtype, tdzp, szp, name);
}
} else if (is_tmpfile) {
/* restore z_unlinked since when linking failed */
szp->z_unlinked = B_TRUE;
}
txg = dmu_tx_get_txg(tx);
dmu_tx_commit(tx);
zfs_dirent_unlock(dl);
if (!is_tmpfile && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
if (is_tmpfile && zfsvfs->z_os->os_sync != ZFS_SYNC_DISABLED)
txg_wait_synced(dmu_objset_pool(zfsvfs->z_os), txg);
zfs_znode_update_vfs(tdzp);
zfs_znode_update_vfs(szp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
static void
zfs_putpage_sync_commit_cb(void *arg)
{
struct page *pp = arg;
ClearPageError(pp);
end_page_writeback(pp);
}
static void
zfs_putpage_async_commit_cb(void *arg)
{
struct page *pp = arg;
znode_t *zp = ITOZ(pp->mapping->host);
ClearPageError(pp);
end_page_writeback(pp);
atomic_dec_32(&zp->z_async_writes_cnt);
}
/*
* Push a page out to disk, once the page is on stable storage the
* registered commit callback will be run as notification of completion.
*
* IN: ip - page mapped for inode.
* pp - page to push (page is locked)
* wbc - writeback control data
* for_sync - does the caller intend to wait synchronously for the
* page writeback to complete?
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - ctime|mtime updated
*/
int
zfs_putpage(struct inode *ip, struct page *pp, struct writeback_control *wbc,
boolean_t for_sync)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
loff_t offset;
loff_t pgoff;
unsigned int pglen;
dmu_tx_t *tx;
caddr_t va;
int err = 0;
uint64_t mtime[2], ctime[2];
sa_bulk_attr_t bulk[3];
int cnt = 0;
struct address_space *mapping;
if ((err = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (err);
ASSERT(PageLocked(pp));
pgoff = page_offset(pp); /* Page byte-offset in file */
offset = i_size_read(ip); /* File length in bytes */
pglen = MIN(PAGE_SIZE, /* Page length in bytes */
P2ROUNDUP(offset, PAGE_SIZE)-pgoff);
/* Page is beyond end of file */
if (pgoff >= offset) {
unlock_page(pp);
zfs_exit(zfsvfs, FTAG);
return (0);
}
/* Truncate page length to end of file */
if (pgoff + pglen > offset)
pglen = offset - pgoff;
#if 0
/*
* FIXME: Allow mmap writes past its quota. The correct fix
* is to register a page_mkwrite() handler to count the page
* against its quota when it is about to be dirtied.
*/
if (zfs_id_overblockquota(zfsvfs, DMU_USERUSED_OBJECT,
KUID_TO_SUID(ip->i_uid)) ||
zfs_id_overblockquota(zfsvfs, DMU_GROUPUSED_OBJECT,
KGID_TO_SGID(ip->i_gid)) ||
(zp->z_projid != ZFS_DEFAULT_PROJID &&
zfs_id_overblockquota(zfsvfs, DMU_PROJECTUSED_OBJECT,
zp->z_projid))) {
err = EDQUOT;
}
#endif
/*
* The ordering here is critical and must adhere to the following
* rules in order to avoid deadlocking in either zfs_read() or
* zfs_free_range() due to a lock inversion.
*
* 1) The page must be unlocked prior to acquiring the range lock.
* This is critical because zfs_read() calls find_lock_page()
* which may block on the page lock while holding the range lock.
*
* 2) Before setting or clearing write back on a page the range lock
* must be held in order to prevent a lock inversion with the
* zfs_free_range() function.
*
* This presents a problem because upon entering this function the
* page lock is already held. To safely acquire the range lock the
* page lock must be dropped. This creates a window where another
* process could truncate, invalidate, dirty, or write out the page.
*
* Therefore, after successfully reacquiring the range and page locks
* the current page state is checked. In the common case everything
* will be as is expected and it can be written out. However, if
* the page state has changed it must be handled accordingly.
*/
mapping = pp->mapping;
redirty_page_for_writepage(wbc, pp);
unlock_page(pp);
zfs_locked_range_t *lr = zfs_rangelock_enter(&zp->z_rangelock,
pgoff, pglen, RL_WRITER);
lock_page(pp);
/* Page mapping changed or it was no longer dirty, we're done */
if (unlikely((mapping != pp->mapping) || !PageDirty(pp))) {
unlock_page(pp);
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (0);
}
/* Another process started write block if required */
if (PageWriteback(pp)) {
unlock_page(pp);
zfs_rangelock_exit(lr);
if (wbc->sync_mode != WB_SYNC_NONE) {
/*
* Speed up any non-sync page writebacks since
* they may take several seconds to complete.
* Refer to the comment in zpl_fsync() (when
* HAVE_FSYNC_RANGE is defined) for details.
*/
if (atomic_load_32(&zp->z_async_writes_cnt) > 0) {
zil_commit(zfsvfs->z_log, zp->z_id);
}
if (PageWriteback(pp))
#ifdef HAVE_PAGEMAP_FOLIO_WAIT_BIT
folio_wait_bit(page_folio(pp), PG_writeback);
#else
wait_on_page_bit(pp, PG_writeback);
#endif
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
/* Clear the dirty flag the required locks are held */
if (!clear_page_dirty_for_io(pp)) {
unlock_page(pp);
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* Counterpart for redirty_page_for_writepage() above. This page
* was in fact not skipped and should not be counted as if it were.
*/
wbc->pages_skipped--;
if (!for_sync)
atomic_inc_32(&zp->z_async_writes_cnt);
set_page_writeback(pp);
unlock_page(pp);
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_write(tx, zp->z_id, pgoff, pglen);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
err = dmu_tx_assign(tx, TXG_NOWAIT);
if (err != 0) {
if (err == ERESTART)
dmu_tx_wait(tx);
dmu_tx_abort(tx);
#ifdef HAVE_VFS_FILEMAP_DIRTY_FOLIO
filemap_dirty_folio(page_mapping(pp), page_folio(pp));
#else
__set_page_dirty_nobuffers(pp);
#endif
ClearPageError(pp);
end_page_writeback(pp);
if (!for_sync)
atomic_dec_32(&zp->z_async_writes_cnt);
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (err);
}
va = kmap(pp);
ASSERT3U(pglen, <=, PAGE_SIZE);
dmu_write(zfsvfs->z_os, zp->z_id, pgoff, pglen, va, tx);
kunmap(pp);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, 8);
/* Preserve the mtime and ctime provided by the inode */
ZFS_TIME_ENCODE(&ip->i_mtime, mtime);
ZFS_TIME_ENCODE(&ip->i_ctime, ctime);
zp->z_atime_dirty = B_FALSE;
zp->z_seq++;
err = sa_bulk_update(zp->z_sa_hdl, bulk, cnt, tx);
zfs_log_write(zfsvfs->z_log, tx, TX_WRITE, zp, pgoff, pglen, 0,
for_sync ? zfs_putpage_sync_commit_cb :
zfs_putpage_async_commit_cb, pp);
dmu_tx_commit(tx);
zfs_rangelock_exit(lr);
if (wbc->sync_mode != WB_SYNC_NONE) {
/*
* Note that this is rarely called under writepages(), because
* writepages() normally handles the entire commit for
* performance reasons.
*/
zil_commit(zfsvfs->z_log, zp->z_id);
} else if (!for_sync && atomic_load_32(&zp->z_sync_writes_cnt) > 0) {
/*
* If the caller does not intend to wait synchronously
* for this page writeback to complete and there are active
* synchronous calls on this file, do a commit so that
* the latter don't accidentally end up waiting for
* our writeback to complete. Refer to the comment in
* zpl_fsync() (when HAVE_FSYNC_RANGE is defined) for details.
*/
zil_commit(zfsvfs->z_log, zp->z_id);
}
dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, pglen);
zfs_exit(zfsvfs, FTAG);
return (err);
}
/*
* Update the system attributes when the inode has been dirtied. For the
* moment we only update the mode, atime, mtime, and ctime.
*/
int
zfs_dirty_inode(struct inode *ip, int flags)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
dmu_tx_t *tx;
uint64_t mode, atime[2], mtime[2], ctime[2];
sa_bulk_attr_t bulk[4];
int error = 0;
int cnt = 0;
if (zfs_is_readonly(zfsvfs) || dmu_objset_is_snapshot(zfsvfs->z_os))
return (0);
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
#ifdef I_DIRTY_TIME
/*
* This is the lazytime semantic introduced in Linux 4.0
* This flag will only be called from update_time when lazytime is set.
* (Note, I_DIRTY_SYNC will also set if not lazytime)
* Fortunately mtime and ctime are managed within ZFS itself, so we
* only need to dirty atime.
*/
if (flags == I_DIRTY_TIME) {
zp->z_atime_dirty = B_TRUE;
goto out;
}
#endif
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
goto out;
}
mutex_enter(&zp->z_lock);
zp->z_atime_dirty = B_FALSE;
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_ATIME(zfsvfs), NULL, &atime, 16);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16);
SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16);
/* Preserve the mode, mtime and ctime provided by the inode */
ZFS_TIME_ENCODE(&ip->i_atime, atime);
ZFS_TIME_ENCODE(&ip->i_mtime, mtime);
ZFS_TIME_ENCODE(&ip->i_ctime, ctime);
mode = ip->i_mode;
zp->z_mode = mode;
error = sa_bulk_update(zp->z_sa_hdl, bulk, cnt, tx);
mutex_exit(&zp->z_lock);
dmu_tx_commit(tx);
out:
zfs_exit(zfsvfs, FTAG);
return (error);
}
void
zfs_inactive(struct inode *ip)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
uint64_t atime[2];
int error;
int need_unlock = 0;
/* Only read lock if we haven't already write locked, e.g. rollback */
if (!RW_WRITE_HELD(&zfsvfs->z_teardown_inactive_lock)) {
need_unlock = 1;
rw_enter(&zfsvfs->z_teardown_inactive_lock, RW_READER);
}
if (zp->z_sa_hdl == NULL) {
if (need_unlock)
rw_exit(&zfsvfs->z_teardown_inactive_lock);
return;
}
if (zp->z_atime_dirty && zp->z_unlinked == B_FALSE) {
dmu_tx_t *tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
} else {
ZFS_TIME_ENCODE(&ip->i_atime, atime);
mutex_enter(&zp->z_lock);
(void) sa_update(zp->z_sa_hdl, SA_ZPL_ATIME(zfsvfs),
(void *)&atime, sizeof (atime), tx);
zp->z_atime_dirty = B_FALSE;
mutex_exit(&zp->z_lock);
dmu_tx_commit(tx);
}
}
zfs_zinactive(zp);
if (need_unlock)
rw_exit(&zfsvfs->z_teardown_inactive_lock);
}
/*
* Fill pages with data from the disk.
*/
static int
zfs_fillpage(struct inode *ip, struct page *pp)
{
zfsvfs_t *zfsvfs = ITOZSB(ip);
loff_t i_size = i_size_read(ip);
u_offset_t io_off = page_offset(pp);
size_t io_len = PAGE_SIZE;
ASSERT3U(io_off, <, i_size);
if (io_off + io_len > i_size)
io_len = i_size - io_off;
void *va = kmap(pp);
int error = dmu_read(zfsvfs->z_os, ITOZ(ip)->z_id, io_off,
io_len, va, DMU_READ_PREFETCH);
if (io_len != PAGE_SIZE)
memset((char *)va + io_len, 0, PAGE_SIZE - io_len);
kunmap(pp);
if (error) {
/* convert checksum errors into IO errors */
if (error == ECKSUM)
error = SET_ERROR(EIO);
SetPageError(pp);
ClearPageUptodate(pp);
} else {
ClearPageError(pp);
SetPageUptodate(pp);
}
return (error);
}
/*
* Uses zfs_fillpage to read data from the file and fill the page.
*
* IN: ip - inode of file to get data from.
* pp - page to read
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* vp - atime updated
*/
int
zfs_getpage(struct inode *ip, struct page *pp)
{
zfsvfs_t *zfsvfs = ITOZSB(ip);
znode_t *zp = ITOZ(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
error = zfs_fillpage(ip, pp);
if (error == 0)
dataset_kstats_update_read_kstats(&zfsvfs->z_kstat, PAGE_SIZE);
zfs_exit(zfsvfs, FTAG);
return (error);
}
/*
* Check ZFS specific permissions to memory map a section of a file.
*
* IN: ip - inode of the file to mmap
* off - file offset
* addrp - start address in memory region
* len - length of memory region
* vm_flags- address flags
*
* RETURN: 0 if success
* error code if failure
*/
int
zfs_map(struct inode *ip, offset_t off, caddr_t *addrp, size_t len,
unsigned long vm_flags)
{
(void) addrp;
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if ((vm_flags & VM_WRITE) && (zp->z_pflags &
(ZFS_IMMUTABLE | ZFS_READONLY | ZFS_APPENDONLY))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
if ((vm_flags & (VM_READ | VM_EXEC)) &&
(zp->z_pflags & ZFS_AV_QUARANTINED)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EACCES));
}
if (off < 0 || len > MAXOFFSET_T - off) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENXIO));
}
zfs_exit(zfsvfs, FTAG);
return (0);
}
/*
* Free or allocate space in a file. Currently, this function only
* supports the `F_FREESP' command. However, this command is somewhat
* misnamed, as its functionality includes the ability to allocate as
* well as free space.
*
* IN: zp - znode of file to free data in.
* cmd - action to take (only F_FREESP supported).
* bfp - section of file to free/alloc.
* flag - current file open mode flags.
* offset - current file offset.
* cr - credentials of caller.
*
* RETURN: 0 on success, error code on failure.
*
* Timestamps:
* zp - ctime|mtime updated
*/
int
zfs_space(znode_t *zp, int cmd, flock64_t *bfp, int flag,
offset_t offset, cred_t *cr)
{
(void) offset;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
uint64_t off, len;
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if (cmd != F_FREESP) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Callers might not be able to detect properly that we are read-only,
* so check it explicitly here.
*/
if (zfs_is_readonly(zfsvfs)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EROFS));
}
if (bfp->l_len < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Permissions aren't checked on Solaris because on this OS
* zfs_space() can only be called with an opened file handle.
* On Linux we can get here through truncate_range() which
* operates directly on inodes, so we need to check access rights.
*/
if ((error = zfs_zaccess(zp, ACE_WRITE_DATA, 0, B_FALSE, cr,
- kcred->user_ns))) {
+ zfs_init_idmap))) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
off = bfp->l_start;
len = bfp->l_len; /* 0 means from off to end of file */
error = zfs_freesp(zp, off, len, flag, TRUE);
zfs_exit(zfsvfs, FTAG);
return (error);
}
int
zfs_fid(struct inode *ip, fid_t *fidp)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
uint32_t gen;
uint64_t gen64;
uint64_t object = zp->z_id;
zfid_short_t *zfid;
int size, i, error;
if ((error = zfs_enter(zfsvfs, FTAG)) != 0)
return (error);
if (fidp->fid_len < SHORT_FID_LEN) {
fidp->fid_len = SHORT_FID_LEN;
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(ENOSPC));
}
if ((error = zfs_verify_zp(zp)) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(zfsvfs),
&gen64, sizeof (uint64_t))) != 0) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
gen = (uint32_t)gen64;
size = SHORT_FID_LEN;
zfid = (zfid_short_t *)fidp;
zfid->zf_len = size;
for (i = 0; i < sizeof (zfid->zf_object); i++)
zfid->zf_object[i] = (uint8_t)(object >> (8 * i));
/* Must have a non-zero generation number to distinguish from .zfs */
if (gen == 0)
gen = 1;
for (i = 0; i < sizeof (zfid->zf_gen); i++)
zfid->zf_gen[i] = (uint8_t)(gen >> (8 * i));
zfs_exit(zfsvfs, FTAG);
return (0);
}
#if defined(_KERNEL)
EXPORT_SYMBOL(zfs_open);
EXPORT_SYMBOL(zfs_close);
EXPORT_SYMBOL(zfs_lookup);
EXPORT_SYMBOL(zfs_create);
EXPORT_SYMBOL(zfs_tmpfile);
EXPORT_SYMBOL(zfs_remove);
EXPORT_SYMBOL(zfs_mkdir);
EXPORT_SYMBOL(zfs_rmdir);
EXPORT_SYMBOL(zfs_readdir);
EXPORT_SYMBOL(zfs_getattr_fast);
EXPORT_SYMBOL(zfs_setattr);
EXPORT_SYMBOL(zfs_rename);
EXPORT_SYMBOL(zfs_symlink);
EXPORT_SYMBOL(zfs_readlink);
EXPORT_SYMBOL(zfs_link);
EXPORT_SYMBOL(zfs_inactive);
EXPORT_SYMBOL(zfs_space);
EXPORT_SYMBOL(zfs_fid);
EXPORT_SYMBOL(zfs_getpage);
EXPORT_SYMBOL(zfs_putpage);
EXPORT_SYMBOL(zfs_dirty_inode);
EXPORT_SYMBOL(zfs_map);
/* CSTYLED */
module_param(zfs_delete_blocks, ulong, 0644);
MODULE_PARM_DESC(zfs_delete_blocks, "Delete files larger than N blocks async");
#endif
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zfs_znode.c b/sys/contrib/openzfs/module/os/linux/zfs/zfs_znode.c
index 38cdccfd8084..c104cd661bf5 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zfs_znode.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zfs_znode.c
@@ -1,2267 +1,2267 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
*/
/* Portions Copyright 2007 Jeremy Teo */
#ifdef _KERNEL
#include <sys/types.h>
#include <sys/param.h>
#include <sys/time.h>
#include <sys/sysmacros.h>
#include <sys/mntent.h>
#include <sys/u8_textprep.h>
#include <sys/dsl_dataset.h>
#include <sys/vfs.h>
#include <sys/vnode.h>
#include <sys/file.h>
#include <sys/kmem.h>
#include <sys/errno.h>
#include <sys/atomic.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_ioctl.h>
#include <sys/zfs_rlock.h>
#include <sys/zfs_fuid.h>
#include <sys/zfs_vnops.h>
#include <sys/zfs_ctldir.h>
#include <sys/dnode.h>
#include <sys/fs/zfs.h>
#include <sys/zpl.h>
#endif /* _KERNEL */
#include <sys/dmu.h>
#include <sys/dmu_objset.h>
#include <sys/dmu_tx.h>
#include <sys/zfs_refcount.h>
#include <sys/stat.h>
#include <sys/zap.h>
#include <sys/zfs_znode.h>
#include <sys/sa.h>
#include <sys/zfs_sa.h>
#include <sys/zfs_stat.h>
#include "zfs_prop.h"
#include "zfs_comutil.h"
/*
* Functions needed for userland (ie: libzpool) are not put under
* #ifdef_KERNEL; the rest of the functions have dependencies
* (such as VFS logic) that will not compile easily in userland.
*/
#ifdef _KERNEL
static kmem_cache_t *znode_cache = NULL;
static kmem_cache_t *znode_hold_cache = NULL;
unsigned int zfs_object_mutex_size = ZFS_OBJ_MTX_SZ;
/*
* This is used by the test suite so that it can delay znodes from being
* freed in order to inspect the unlinked set.
*/
static int zfs_unlink_suspend_progress = 0;
/*
* This callback is invoked when acquiring a RL_WRITER or RL_APPEND lock on
* z_rangelock. It will modify the offset and length of the lock to reflect
* znode-specific information, and convert RL_APPEND to RL_WRITER. This is
* called with the rangelock_t's rl_lock held, which avoids races.
*/
static void
zfs_rangelock_cb(zfs_locked_range_t *new, void *arg)
{
znode_t *zp = arg;
/*
* If in append mode, convert to writer and lock starting at the
* current end of file.
*/
if (new->lr_type == RL_APPEND) {
new->lr_offset = zp->z_size;
new->lr_type = RL_WRITER;
}
/*
* If we need to grow the block size then lock the whole file range.
*/
uint64_t end_size = MAX(zp->z_size, new->lr_offset + new->lr_length);
if (end_size > zp->z_blksz && (!ISP2(zp->z_blksz) ||
zp->z_blksz < ZTOZSB(zp)->z_max_blksz)) {
new->lr_offset = 0;
new->lr_length = UINT64_MAX;
}
}
static int
zfs_znode_cache_constructor(void *buf, void *arg, int kmflags)
{
(void) arg, (void) kmflags;
znode_t *zp = buf;
inode_init_once(ZTOI(zp));
list_link_init(&zp->z_link_node);
mutex_init(&zp->z_lock, NULL, MUTEX_DEFAULT, NULL);
rw_init(&zp->z_parent_lock, NULL, RW_DEFAULT, NULL);
rw_init(&zp->z_name_lock, NULL, RW_NOLOCKDEP, NULL);
mutex_init(&zp->z_acl_lock, NULL, MUTEX_DEFAULT, NULL);
rw_init(&zp->z_xattr_lock, NULL, RW_DEFAULT, NULL);
zfs_rangelock_init(&zp->z_rangelock, zfs_rangelock_cb, zp);
zp->z_dirlocks = NULL;
zp->z_acl_cached = NULL;
zp->z_xattr_cached = NULL;
zp->z_xattr_parent = 0;
zp->z_sync_writes_cnt = 0;
zp->z_async_writes_cnt = 0;
return (0);
}
static void
zfs_znode_cache_destructor(void *buf, void *arg)
{
(void) arg;
znode_t *zp = buf;
ASSERT(!list_link_active(&zp->z_link_node));
mutex_destroy(&zp->z_lock);
rw_destroy(&zp->z_parent_lock);
rw_destroy(&zp->z_name_lock);
mutex_destroy(&zp->z_acl_lock);
rw_destroy(&zp->z_xattr_lock);
zfs_rangelock_fini(&zp->z_rangelock);
ASSERT3P(zp->z_dirlocks, ==, NULL);
ASSERT3P(zp->z_acl_cached, ==, NULL);
ASSERT3P(zp->z_xattr_cached, ==, NULL);
ASSERT0(atomic_load_32(&zp->z_sync_writes_cnt));
ASSERT0(atomic_load_32(&zp->z_async_writes_cnt));
}
static int
zfs_znode_hold_cache_constructor(void *buf, void *arg, int kmflags)
{
(void) arg, (void) kmflags;
znode_hold_t *zh = buf;
mutex_init(&zh->zh_lock, NULL, MUTEX_DEFAULT, NULL);
zh->zh_refcount = 0;
return (0);
}
static void
zfs_znode_hold_cache_destructor(void *buf, void *arg)
{
(void) arg;
znode_hold_t *zh = buf;
mutex_destroy(&zh->zh_lock);
}
void
zfs_znode_init(void)
{
/*
* Initialize zcache. The KMC_SLAB hint is used in order that it be
* backed by kmalloc() when on the Linux slab in order that any
* wait_on_bit() operations on the related inode operate properly.
*/
ASSERT(znode_cache == NULL);
znode_cache = kmem_cache_create("zfs_znode_cache",
sizeof (znode_t), 0, zfs_znode_cache_constructor,
zfs_znode_cache_destructor, NULL, NULL, NULL, KMC_SLAB);
ASSERT(znode_hold_cache == NULL);
znode_hold_cache = kmem_cache_create("zfs_znode_hold_cache",
sizeof (znode_hold_t), 0, zfs_znode_hold_cache_constructor,
zfs_znode_hold_cache_destructor, NULL, NULL, NULL, 0);
}
void
zfs_znode_fini(void)
{
/*
* Cleanup zcache
*/
if (znode_cache)
kmem_cache_destroy(znode_cache);
znode_cache = NULL;
if (znode_hold_cache)
kmem_cache_destroy(znode_hold_cache);
znode_hold_cache = NULL;
}
/*
* The zfs_znode_hold_enter() / zfs_znode_hold_exit() functions are used to
* serialize access to a znode and its SA buffer while the object is being
* created or destroyed. This kind of locking would normally reside in the
* znode itself but in this case that's impossible because the znode and SA
* buffer may not yet exist. Therefore the locking is handled externally
* with an array of mutexes and AVLs trees which contain per-object locks.
*
* In zfs_znode_hold_enter() a per-object lock is created as needed, inserted
* in to the correct AVL tree and finally the per-object lock is held. In
* zfs_znode_hold_exit() the process is reversed. The per-object lock is
* released, removed from the AVL tree and destroyed if there are no waiters.
*
* This scheme has two important properties:
*
* 1) No memory allocations are performed while holding one of the z_hold_locks.
* This ensures evict(), which can be called from direct memory reclaim, will
* never block waiting on a z_hold_locks which just happens to have hashed
* to the same index.
*
* 2) All locks used to serialize access to an object are per-object and never
* shared. This minimizes lock contention without creating a large number
* of dedicated locks.
*
* On the downside it does require znode_lock_t structures to be frequently
* allocated and freed. However, because these are backed by a kmem cache
* and very short lived this cost is minimal.
*/
int
zfs_znode_hold_compare(const void *a, const void *b)
{
const znode_hold_t *zh_a = (const znode_hold_t *)a;
const znode_hold_t *zh_b = (const znode_hold_t *)b;
return (TREE_CMP(zh_a->zh_obj, zh_b->zh_obj));
}
static boolean_t __maybe_unused
zfs_znode_held(zfsvfs_t *zfsvfs, uint64_t obj)
{
znode_hold_t *zh, search;
int i = ZFS_OBJ_HASH(zfsvfs, obj);
boolean_t held;
search.zh_obj = obj;
mutex_enter(&zfsvfs->z_hold_locks[i]);
zh = avl_find(&zfsvfs->z_hold_trees[i], &search, NULL);
held = (zh && MUTEX_HELD(&zh->zh_lock)) ? B_TRUE : B_FALSE;
mutex_exit(&zfsvfs->z_hold_locks[i]);
return (held);
}
znode_hold_t *
zfs_znode_hold_enter(zfsvfs_t *zfsvfs, uint64_t obj)
{
znode_hold_t *zh, *zh_new, search;
int i = ZFS_OBJ_HASH(zfsvfs, obj);
boolean_t found = B_FALSE;
zh_new = kmem_cache_alloc(znode_hold_cache, KM_SLEEP);
search.zh_obj = obj;
mutex_enter(&zfsvfs->z_hold_locks[i]);
zh = avl_find(&zfsvfs->z_hold_trees[i], &search, NULL);
if (likely(zh == NULL)) {
zh = zh_new;
zh->zh_obj = obj;
avl_add(&zfsvfs->z_hold_trees[i], zh);
} else {
ASSERT3U(zh->zh_obj, ==, obj);
found = B_TRUE;
}
zh->zh_refcount++;
ASSERT3S(zh->zh_refcount, >, 0);
mutex_exit(&zfsvfs->z_hold_locks[i]);
if (found == B_TRUE)
kmem_cache_free(znode_hold_cache, zh_new);
ASSERT(MUTEX_NOT_HELD(&zh->zh_lock));
mutex_enter(&zh->zh_lock);
return (zh);
}
void
zfs_znode_hold_exit(zfsvfs_t *zfsvfs, znode_hold_t *zh)
{
int i = ZFS_OBJ_HASH(zfsvfs, zh->zh_obj);
boolean_t remove = B_FALSE;
ASSERT(zfs_znode_held(zfsvfs, zh->zh_obj));
mutex_exit(&zh->zh_lock);
mutex_enter(&zfsvfs->z_hold_locks[i]);
ASSERT3S(zh->zh_refcount, >, 0);
if (--zh->zh_refcount == 0) {
avl_remove(&zfsvfs->z_hold_trees[i], zh);
remove = B_TRUE;
}
mutex_exit(&zfsvfs->z_hold_locks[i]);
if (remove == B_TRUE)
kmem_cache_free(znode_hold_cache, zh);
}
dev_t
zfs_cmpldev(uint64_t dev)
{
return (dev);
}
static void
zfs_znode_sa_init(zfsvfs_t *zfsvfs, znode_t *zp,
dmu_buf_t *db, dmu_object_type_t obj_type, sa_handle_t *sa_hdl)
{
ASSERT(zfs_znode_held(zfsvfs, zp->z_id));
mutex_enter(&zp->z_lock);
ASSERT(zp->z_sa_hdl == NULL);
ASSERT(zp->z_acl_cached == NULL);
if (sa_hdl == NULL) {
VERIFY(0 == sa_handle_get_from_db(zfsvfs->z_os, db, zp,
SA_HDL_SHARED, &zp->z_sa_hdl));
} else {
zp->z_sa_hdl = sa_hdl;
sa_set_userp(sa_hdl, zp);
}
zp->z_is_sa = (obj_type == DMU_OT_SA) ? B_TRUE : B_FALSE;
mutex_exit(&zp->z_lock);
}
void
zfs_znode_dmu_fini(znode_t *zp)
{
ASSERT(zfs_znode_held(ZTOZSB(zp), zp->z_id) ||
RW_WRITE_HELD(&ZTOZSB(zp)->z_teardown_inactive_lock));
sa_handle_destroy(zp->z_sa_hdl);
zp->z_sa_hdl = NULL;
}
/*
* Called by new_inode() to allocate a new inode.
*/
int
zfs_inode_alloc(struct super_block *sb, struct inode **ip)
{
znode_t *zp;
zp = kmem_cache_alloc(znode_cache, KM_SLEEP);
*ip = ZTOI(zp);
return (0);
}
/*
* Called in multiple places when an inode should be destroyed.
*/
void
zfs_inode_destroy(struct inode *ip)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
mutex_enter(&zfsvfs->z_znodes_lock);
if (list_link_active(&zp->z_link_node)) {
list_remove(&zfsvfs->z_all_znodes, zp);
zfsvfs->z_nr_znodes--;
}
mutex_exit(&zfsvfs->z_znodes_lock);
if (zp->z_acl_cached) {
zfs_acl_free(zp->z_acl_cached);
zp->z_acl_cached = NULL;
}
if (zp->z_xattr_cached) {
nvlist_free(zp->z_xattr_cached);
zp->z_xattr_cached = NULL;
}
kmem_cache_free(znode_cache, zp);
}
static void
zfs_inode_set_ops(zfsvfs_t *zfsvfs, struct inode *ip)
{
uint64_t rdev = 0;
switch (ip->i_mode & S_IFMT) {
case S_IFREG:
ip->i_op = &zpl_inode_operations;
ip->i_fop = &zpl_file_operations;
ip->i_mapping->a_ops = &zpl_address_space_operations;
break;
case S_IFDIR:
#ifdef HAVE_RENAME2_OPERATIONS_WRAPPER
ip->i_flags |= S_IOPS_WRAPPER;
ip->i_op = &zpl_dir_inode_operations.ops;
#else
ip->i_op = &zpl_dir_inode_operations;
#endif
ip->i_fop = &zpl_dir_file_operations;
ITOZ(ip)->z_zn_prefetch = B_TRUE;
break;
case S_IFLNK:
ip->i_op = &zpl_symlink_inode_operations;
break;
/*
* rdev is only stored in a SA only for device files.
*/
case S_IFCHR:
case S_IFBLK:
(void) sa_lookup(ITOZ(ip)->z_sa_hdl, SA_ZPL_RDEV(zfsvfs), &rdev,
sizeof (rdev));
zfs_fallthrough;
case S_IFIFO:
case S_IFSOCK:
init_special_inode(ip, ip->i_mode, rdev);
ip->i_op = &zpl_special_inode_operations;
break;
default:
zfs_panic_recover("inode %llu has invalid mode: 0x%x\n",
(u_longlong_t)ip->i_ino, ip->i_mode);
/* Assume the inode is a file and attempt to continue */
ip->i_mode = S_IFREG | 0644;
ip->i_op = &zpl_inode_operations;
ip->i_fop = &zpl_file_operations;
ip->i_mapping->a_ops = &zpl_address_space_operations;
break;
}
}
static void
zfs_set_inode_flags(znode_t *zp, struct inode *ip)
{
/*
* Linux and Solaris have different sets of file attributes, so we
* restrict this conversion to the intersection of the two.
*/
#ifdef HAVE_INODE_SET_FLAGS
unsigned int flags = 0;
if (zp->z_pflags & ZFS_IMMUTABLE)
flags |= S_IMMUTABLE;
if (zp->z_pflags & ZFS_APPENDONLY)
flags |= S_APPEND;
inode_set_flags(ip, flags, S_IMMUTABLE|S_APPEND);
#else
if (zp->z_pflags & ZFS_IMMUTABLE)
ip->i_flags |= S_IMMUTABLE;
else
ip->i_flags &= ~S_IMMUTABLE;
if (zp->z_pflags & ZFS_APPENDONLY)
ip->i_flags |= S_APPEND;
else
ip->i_flags &= ~S_APPEND;
#endif
}
/*
* Update the embedded inode given the znode.
*/
void
zfs_znode_update_vfs(znode_t *zp)
{
struct inode *ip;
uint32_t blksize;
u_longlong_t i_blocks;
ASSERT(zp != NULL);
ip = ZTOI(zp);
/* Skip .zfs control nodes which do not exist on disk. */
if (zfsctl_is_node(ip))
return;
dmu_object_size_from_db(sa_get_db(zp->z_sa_hdl), &blksize, &i_blocks);
spin_lock(&ip->i_lock);
ip->i_mode = zp->z_mode;
ip->i_blocks = i_blocks;
i_size_write(ip, zp->z_size);
spin_unlock(&ip->i_lock);
}
/*
* Construct a znode+inode and initialize.
*
* This does not do a call to dmu_set_user() that is
* up to the caller to do, in case you don't want to
* return the znode
*/
static znode_t *
zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz,
dmu_object_type_t obj_type, sa_handle_t *hdl)
{
znode_t *zp;
struct inode *ip;
uint64_t mode;
uint64_t parent;
uint64_t tmp_gen;
uint64_t links;
uint64_t z_uid, z_gid;
uint64_t atime[2], mtime[2], ctime[2], btime[2];
uint64_t projid = ZFS_DEFAULT_PROJID;
sa_bulk_attr_t bulk[12];
int count = 0;
ASSERT(zfsvfs != NULL);
ip = new_inode(zfsvfs->z_sb);
if (ip == NULL)
return (NULL);
zp = ITOZ(ip);
ASSERT(zp->z_dirlocks == NULL);
ASSERT3P(zp->z_acl_cached, ==, NULL);
ASSERT3P(zp->z_xattr_cached, ==, NULL);
zp->z_unlinked = B_FALSE;
zp->z_atime_dirty = B_FALSE;
#if !defined(HAVE_FILEMAP_RANGE_HAS_PAGE)
zp->z_is_mapped = B_FALSE;
#endif
zp->z_is_ctldir = B_FALSE;
zp->z_suspended = B_FALSE;
zp->z_sa_hdl = NULL;
zp->z_mapcnt = 0;
zp->z_id = db->db_object;
zp->z_blksz = blksz;
zp->z_seq = 0x7A4653;
zp->z_sync_cnt = 0;
zp->z_sync_writes_cnt = 0;
zp->z_async_writes_cnt = 0;
zfs_znode_sa_init(zfsvfs, zp, db, obj_type, hdl);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GEN(zfsvfs), NULL, &tmp_gen, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL,
&zp->z_size, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL, &links, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PARENT(zfsvfs), NULL,
&parent, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL, &z_uid, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL, &z_gid, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL, &atime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CRTIME(zfsvfs), NULL, &btime, 16);
if (sa_bulk_lookup(zp->z_sa_hdl, bulk, count) != 0 || tmp_gen == 0 ||
(dmu_objset_projectquota_enabled(zfsvfs->z_os) &&
(zp->z_pflags & ZFS_PROJID) &&
sa_lookup(zp->z_sa_hdl, SA_ZPL_PROJID(zfsvfs), &projid, 8) != 0)) {
if (hdl == NULL)
sa_handle_destroy(zp->z_sa_hdl);
zp->z_sa_hdl = NULL;
goto error;
}
zp->z_projid = projid;
zp->z_mode = ip->i_mode = mode;
ip->i_generation = (uint32_t)tmp_gen;
ip->i_blkbits = SPA_MINBLOCKSHIFT;
set_nlink(ip, (uint32_t)links);
zfs_uid_write(ip, z_uid);
zfs_gid_write(ip, z_gid);
zfs_set_inode_flags(zp, ip);
/* Cache the xattr parent id */
if (zp->z_pflags & ZFS_XATTR)
zp->z_xattr_parent = parent;
ZFS_TIME_DECODE(&ip->i_atime, atime);
ZFS_TIME_DECODE(&ip->i_mtime, mtime);
ZFS_TIME_DECODE(&ip->i_ctime, ctime);
ZFS_TIME_DECODE(&zp->z_btime, btime);
ip->i_ino = zp->z_id;
zfs_znode_update_vfs(zp);
zfs_inode_set_ops(zfsvfs, ip);
/*
* The only way insert_inode_locked() can fail is if the ip->i_ino
* number is already hashed for this super block. This can never
* happen because the inode numbers map 1:1 with the object numbers.
*
* Exceptions include rolling back a mounted file system, either
* from the zfs rollback or zfs recv command.
*
* Active inodes are unhashed during the rollback, but since zrele
* can happen asynchronously, we can't guarantee they've been
* unhashed. This can cause hash collisions in unlinked drain
* processing so do not hash unlinked znodes.
*/
if (links > 0)
VERIFY3S(insert_inode_locked(ip), ==, 0);
mutex_enter(&zfsvfs->z_znodes_lock);
list_insert_tail(&zfsvfs->z_all_znodes, zp);
zfsvfs->z_nr_znodes++;
mutex_exit(&zfsvfs->z_znodes_lock);
if (links > 0)
unlock_new_inode(ip);
return (zp);
error:
iput(ip);
return (NULL);
}
/*
* Safely mark an inode dirty. Inodes which are part of a read-only
* file system or snapshot may not be dirtied.
*/
void
zfs_mark_inode_dirty(struct inode *ip)
{
zfsvfs_t *zfsvfs = ITOZSB(ip);
if (zfs_is_readonly(zfsvfs) || dmu_objset_is_snapshot(zfsvfs->z_os))
return;
mark_inode_dirty(ip);
}
static uint64_t empty_xattr;
static uint64_t pad[4];
static zfs_acl_phys_t acl_phys;
/*
* Create a new DMU object to hold a zfs znode.
*
* IN: dzp - parent directory for new znode
* vap - file attributes for new znode
* tx - dmu transaction id for zap operations
* cr - credentials of caller
* flag - flags:
* IS_ROOT_NODE - new object will be root
* IS_TMPFILE - new object is of O_TMPFILE
* IS_XATTR - new object is an attribute
* acl_ids - ACL related attributes
*
* OUT: zpp - allocated znode (set to dzp if IS_ROOT_NODE)
*
*/
void
zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr,
uint_t flag, znode_t **zpp, zfs_acl_ids_t *acl_ids)
{
uint64_t crtime[2], atime[2], mtime[2], ctime[2];
uint64_t mode, size, links, parent, pflags;
uint64_t projid = ZFS_DEFAULT_PROJID;
uint64_t rdev = 0;
zfsvfs_t *zfsvfs = ZTOZSB(dzp);
dmu_buf_t *db;
inode_timespec_t now;
uint64_t gen, obj;
int bonuslen;
int dnodesize;
sa_handle_t *sa_hdl;
dmu_object_type_t obj_type;
sa_bulk_attr_t *sa_attrs;
int cnt = 0;
zfs_acl_locator_cb_t locate = { 0 };
znode_hold_t *zh;
if (zfsvfs->z_replay) {
obj = vap->va_nodeid;
now = vap->va_ctime; /* see zfs_replay_create() */
gen = vap->va_nblocks; /* ditto */
dnodesize = vap->va_fsid; /* ditto */
} else {
obj = 0;
gethrestime(&now);
gen = dmu_tx_get_txg(tx);
dnodesize = dmu_objset_dnodesize(zfsvfs->z_os);
}
if (dnodesize == 0)
dnodesize = DNODE_MIN_SIZE;
obj_type = zfsvfs->z_use_sa ? DMU_OT_SA : DMU_OT_ZNODE;
bonuslen = (obj_type == DMU_OT_SA) ?
DN_BONUS_SIZE(dnodesize) : ZFS_OLD_ZNODE_PHYS_SIZE;
/*
* Create a new DMU object.
*/
/*
* There's currently no mechanism for pre-reading the blocks that will
* be needed to allocate a new object, so we accept the small chance
* that there will be an i/o error and we will fail one of the
* assertions below.
*/
if (S_ISDIR(vap->va_mode)) {
if (zfsvfs->z_replay) {
VERIFY0(zap_create_claim_norm_dnsize(zfsvfs->z_os, obj,
zfsvfs->z_norm, DMU_OT_DIRECTORY_CONTENTS,
obj_type, bonuslen, dnodesize, tx));
} else {
obj = zap_create_norm_dnsize(zfsvfs->z_os,
zfsvfs->z_norm, DMU_OT_DIRECTORY_CONTENTS,
obj_type, bonuslen, dnodesize, tx);
}
} else {
if (zfsvfs->z_replay) {
VERIFY0(dmu_object_claim_dnsize(zfsvfs->z_os, obj,
DMU_OT_PLAIN_FILE_CONTENTS, 0,
obj_type, bonuslen, dnodesize, tx));
} else {
obj = dmu_object_alloc_dnsize(zfsvfs->z_os,
DMU_OT_PLAIN_FILE_CONTENTS, 0,
obj_type, bonuslen, dnodesize, tx);
}
}
zh = zfs_znode_hold_enter(zfsvfs, obj);
VERIFY0(sa_buf_hold(zfsvfs->z_os, obj, NULL, &db));
/*
* If this is the root, fix up the half-initialized parent pointer
* to reference the just-allocated physical data area.
*/
if (flag & IS_ROOT_NODE) {
dzp->z_id = obj;
}
/*
* If parent is an xattr, so am I.
*/
if (dzp->z_pflags & ZFS_XATTR) {
flag |= IS_XATTR;
}
if (zfsvfs->z_use_fuids)
pflags = ZFS_ARCHIVE | ZFS_AV_MODIFIED;
else
pflags = 0;
if (S_ISDIR(vap->va_mode)) {
size = 2; /* contents ("." and "..") */
links = 2;
} else {
size = 0;
links = (flag & IS_TMPFILE) ? 0 : 1;
}
if (S_ISBLK(vap->va_mode) || S_ISCHR(vap->va_mode))
rdev = vap->va_rdev;
parent = dzp->z_id;
mode = acl_ids->z_mode;
if (flag & IS_XATTR)
pflags |= ZFS_XATTR;
if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode)) {
/*
* With ZFS_PROJID flag, we can easily know whether there is
* project ID stored on disk or not. See zfs_space_delta_cb().
*/
if (obj_type != DMU_OT_ZNODE &&
dmu_objset_projectquota_enabled(zfsvfs->z_os))
pflags |= ZFS_PROJID;
/*
* Inherit project ID from parent if required.
*/
projid = zfs_inherit_projid(dzp);
if (dzp->z_pflags & ZFS_PROJINHERIT)
pflags |= ZFS_PROJINHERIT;
}
/*
* No execs denied will be determined when zfs_mode_compute() is called.
*/
pflags |= acl_ids->z_aclp->z_hints &
(ZFS_ACL_TRIVIAL|ZFS_INHERIT_ACE|ZFS_ACL_AUTO_INHERIT|
ZFS_ACL_DEFAULTED|ZFS_ACL_PROTECTED);
ZFS_TIME_ENCODE(&now, crtime);
ZFS_TIME_ENCODE(&now, ctime);
if (vap->va_mask & ATTR_ATIME) {
ZFS_TIME_ENCODE(&vap->va_atime, atime);
} else {
ZFS_TIME_ENCODE(&now, atime);
}
if (vap->va_mask & ATTR_MTIME) {
ZFS_TIME_ENCODE(&vap->va_mtime, mtime);
} else {
ZFS_TIME_ENCODE(&now, mtime);
}
/* Now add in all of the "SA" attributes */
VERIFY(0 == sa_handle_get_from_db(zfsvfs->z_os, db, NULL, SA_HDL_SHARED,
&sa_hdl));
/*
* Setup the array of attributes to be replaced/set on the new file
*
* order for DMU_OT_ZNODE is critical since it needs to be constructed
* in the old znode_phys_t format. Don't change this ordering
*/
sa_attrs = kmem_alloc(sizeof (sa_bulk_attr_t) * ZPL_END, KM_SLEEP);
if (obj_type == DMU_OT_ZNODE) {
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ATIME(zfsvfs),
NULL, &atime, 16);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MTIME(zfsvfs),
NULL, &mtime, 16);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CTIME(zfsvfs),
NULL, &ctime, 16);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CRTIME(zfsvfs),
NULL, &crtime, 16);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GEN(zfsvfs),
NULL, &gen, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MODE(zfsvfs),
NULL, &mode, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_SIZE(zfsvfs),
NULL, &size, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PARENT(zfsvfs),
NULL, &parent, 8);
} else {
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MODE(zfsvfs),
NULL, &mode, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_SIZE(zfsvfs),
NULL, &size, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GEN(zfsvfs),
NULL, &gen, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_UID(zfsvfs),
NULL, &acl_ids->z_fuid, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GID(zfsvfs),
NULL, &acl_ids->z_fgid, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PARENT(zfsvfs),
NULL, &parent, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_FLAGS(zfsvfs),
NULL, &pflags, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ATIME(zfsvfs),
NULL, &atime, 16);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MTIME(zfsvfs),
NULL, &mtime, 16);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CTIME(zfsvfs),
NULL, &ctime, 16);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CRTIME(zfsvfs),
NULL, &crtime, 16);
}
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_LINKS(zfsvfs), NULL, &links, 8);
if (obj_type == DMU_OT_ZNODE) {
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_XATTR(zfsvfs), NULL,
&empty_xattr, 8);
} else if (dmu_objset_projectquota_enabled(zfsvfs->z_os) &&
pflags & ZFS_PROJID) {
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PROJID(zfsvfs),
NULL, &projid, 8);
}
if (obj_type == DMU_OT_ZNODE ||
(S_ISBLK(vap->va_mode) || S_ISCHR(vap->va_mode))) {
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_RDEV(zfsvfs),
NULL, &rdev, 8);
}
if (obj_type == DMU_OT_ZNODE) {
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_FLAGS(zfsvfs),
NULL, &pflags, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_UID(zfsvfs), NULL,
&acl_ids->z_fuid, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GID(zfsvfs), NULL,
&acl_ids->z_fgid, 8);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PAD(zfsvfs), NULL, pad,
sizeof (uint64_t) * 4);
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ZNODE_ACL(zfsvfs), NULL,
&acl_phys, sizeof (zfs_acl_phys_t));
} else if (acl_ids->z_aclp->z_version >= ZFS_ACL_VERSION_FUID) {
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_DACL_COUNT(zfsvfs), NULL,
&acl_ids->z_aclp->z_acl_count, 8);
locate.cb_aclp = acl_ids->z_aclp;
SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_DACL_ACES(zfsvfs),
zfs_acl_data_locator, &locate,
acl_ids->z_aclp->z_acl_bytes);
mode = zfs_mode_compute(mode, acl_ids->z_aclp, &pflags,
acl_ids->z_fuid, acl_ids->z_fgid);
}
VERIFY(sa_replace_all_by_template(sa_hdl, sa_attrs, cnt, tx) == 0);
if (!(flag & IS_ROOT_NODE)) {
/*
* The call to zfs_znode_alloc() may fail if memory is low
* via the call path: alloc_inode() -> inode_init_always() ->
* security_inode_alloc() -> inode_alloc_security(). Since
* the existing code is written such that zfs_mknode() can
* not fail retry until sufficient memory has been reclaimed.
*/
do {
*zpp = zfs_znode_alloc(zfsvfs, db, 0, obj_type, sa_hdl);
} while (*zpp == NULL);
VERIFY(*zpp != NULL);
VERIFY(dzp != NULL);
} else {
/*
* If we are creating the root node, the "parent" we
* passed in is the znode for the root.
*/
*zpp = dzp;
(*zpp)->z_sa_hdl = sa_hdl;
}
(*zpp)->z_pflags = pflags;
(*zpp)->z_mode = ZTOI(*zpp)->i_mode = mode;
(*zpp)->z_dnodesize = dnodesize;
(*zpp)->z_projid = projid;
if (obj_type == DMU_OT_ZNODE ||
acl_ids->z_aclp->z_version < ZFS_ACL_VERSION_FUID) {
VERIFY0(zfs_aclset_common(*zpp, acl_ids->z_aclp, cr, tx));
}
kmem_free(sa_attrs, sizeof (sa_bulk_attr_t) * ZPL_END);
zfs_znode_hold_exit(zfsvfs, zh);
}
/*
* Update in-core attributes. It is assumed the caller will be doing an
* sa_bulk_update to push the changes out.
*/
void
zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx)
{
xoptattr_t *xoap;
boolean_t update_inode = B_FALSE;
xoap = xva_getxoptattr(xvap);
ASSERT(xoap);
if (XVA_ISSET_REQ(xvap, XAT_CREATETIME)) {
uint64_t times[2];
ZFS_TIME_ENCODE(&xoap->xoa_createtime, times);
(void) sa_update(zp->z_sa_hdl, SA_ZPL_CRTIME(ZTOZSB(zp)),
&times, sizeof (times), tx);
XVA_SET_RTN(xvap, XAT_CREATETIME);
}
if (XVA_ISSET_REQ(xvap, XAT_READONLY)) {
ZFS_ATTR_SET(zp, ZFS_READONLY, xoap->xoa_readonly,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_READONLY);
}
if (XVA_ISSET_REQ(xvap, XAT_HIDDEN)) {
ZFS_ATTR_SET(zp, ZFS_HIDDEN, xoap->xoa_hidden,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_HIDDEN);
}
if (XVA_ISSET_REQ(xvap, XAT_SYSTEM)) {
ZFS_ATTR_SET(zp, ZFS_SYSTEM, xoap->xoa_system,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_SYSTEM);
}
if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE)) {
ZFS_ATTR_SET(zp, ZFS_ARCHIVE, xoap->xoa_archive,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_ARCHIVE);
}
if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) {
ZFS_ATTR_SET(zp, ZFS_IMMUTABLE, xoap->xoa_immutable,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_IMMUTABLE);
update_inode = B_TRUE;
}
if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) {
ZFS_ATTR_SET(zp, ZFS_NOUNLINK, xoap->xoa_nounlink,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_NOUNLINK);
}
if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) {
ZFS_ATTR_SET(zp, ZFS_APPENDONLY, xoap->xoa_appendonly,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_APPENDONLY);
update_inode = B_TRUE;
}
if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) {
ZFS_ATTR_SET(zp, ZFS_NODUMP, xoap->xoa_nodump,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_NODUMP);
}
if (XVA_ISSET_REQ(xvap, XAT_OPAQUE)) {
ZFS_ATTR_SET(zp, ZFS_OPAQUE, xoap->xoa_opaque,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_OPAQUE);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) {
ZFS_ATTR_SET(zp, ZFS_AV_QUARANTINED,
xoap->xoa_av_quarantined, zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_AV_QUARANTINED);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) {
ZFS_ATTR_SET(zp, ZFS_AV_MODIFIED, xoap->xoa_av_modified,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_AV_MODIFIED);
}
if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) {
zfs_sa_set_scanstamp(zp, xvap, tx);
XVA_SET_RTN(xvap, XAT_AV_SCANSTAMP);
}
if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) {
ZFS_ATTR_SET(zp, ZFS_REPARSE, xoap->xoa_reparse,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_REPARSE);
}
if (XVA_ISSET_REQ(xvap, XAT_OFFLINE)) {
ZFS_ATTR_SET(zp, ZFS_OFFLINE, xoap->xoa_offline,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_OFFLINE);
}
if (XVA_ISSET_REQ(xvap, XAT_SPARSE)) {
ZFS_ATTR_SET(zp, ZFS_SPARSE, xoap->xoa_sparse,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_SPARSE);
}
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT)) {
ZFS_ATTR_SET(zp, ZFS_PROJINHERIT, xoap->xoa_projinherit,
zp->z_pflags, tx);
XVA_SET_RTN(xvap, XAT_PROJINHERIT);
}
if (update_inode)
zfs_set_inode_flags(zp, ZTOI(zp));
}
int
zfs_zget(zfsvfs_t *zfsvfs, uint64_t obj_num, znode_t **zpp)
{
dmu_object_info_t doi;
dmu_buf_t *db;
znode_t *zp;
znode_hold_t *zh;
int err;
sa_handle_t *hdl;
*zpp = NULL;
again:
zh = zfs_znode_hold_enter(zfsvfs, obj_num);
err = sa_buf_hold(zfsvfs->z_os, obj_num, NULL, &db);
if (err) {
zfs_znode_hold_exit(zfsvfs, zh);
return (err);
}
dmu_object_info_from_db(db, &doi);
if (doi.doi_bonus_type != DMU_OT_SA &&
(doi.doi_bonus_type != DMU_OT_ZNODE ||
(doi.doi_bonus_type == DMU_OT_ZNODE &&
doi.doi_bonus_size < sizeof (znode_phys_t)))) {
sa_buf_rele(db, NULL);
zfs_znode_hold_exit(zfsvfs, zh);
return (SET_ERROR(EINVAL));
}
hdl = dmu_buf_get_user(db);
if (hdl != NULL) {
zp = sa_get_userdata(hdl);
/*
* Since "SA" does immediate eviction we
* should never find a sa handle that doesn't
* know about the znode.
*/
ASSERT3P(zp, !=, NULL);
mutex_enter(&zp->z_lock);
ASSERT3U(zp->z_id, ==, obj_num);
/*
* If zp->z_unlinked is set, the znode is already marked
* for deletion and should not be discovered. Check this
* after checking igrab() due to fsetxattr() & O_TMPFILE.
*
* If igrab() returns NULL the VFS has independently
* determined the inode should be evicted and has
* called iput_final() to start the eviction process.
* The SA handle is still valid but because the VFS
* requires that the eviction succeed we must drop
* our locks and references to allow the eviction to
* complete. The zfs_zget() may then be retried.
*
* This unlikely case could be optimized by registering
* a sops->drop_inode() callback. The callback would
* need to detect the active SA hold thereby informing
* the VFS that this inode should not be evicted.
*/
if (igrab(ZTOI(zp)) == NULL) {
if (zp->z_unlinked)
err = SET_ERROR(ENOENT);
else
err = SET_ERROR(EAGAIN);
} else {
*zpp = zp;
err = 0;
}
mutex_exit(&zp->z_lock);
sa_buf_rele(db, NULL);
zfs_znode_hold_exit(zfsvfs, zh);
if (err == EAGAIN) {
/* inode might need this to finish evict */
cond_resched();
goto again;
}
return (err);
}
/*
* Not found create new znode/vnode but only if file exists.
*
* There is a small window where zfs_vget() could
* find this object while a file create is still in
* progress. This is checked for in zfs_znode_alloc()
*
* if zfs_znode_alloc() fails it will drop the hold on the
* bonus buffer.
*/
zp = zfs_znode_alloc(zfsvfs, db, doi.doi_data_block_size,
doi.doi_bonus_type, NULL);
if (zp == NULL) {
err = SET_ERROR(ENOENT);
} else {
*zpp = zp;
}
zfs_znode_hold_exit(zfsvfs, zh);
return (err);
}
int
zfs_rezget(znode_t *zp)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
dmu_object_info_t doi;
dmu_buf_t *db;
uint64_t obj_num = zp->z_id;
uint64_t mode;
uint64_t links;
sa_bulk_attr_t bulk[11];
int err;
int count = 0;
uint64_t gen;
uint64_t z_uid, z_gid;
uint64_t atime[2], mtime[2], ctime[2], btime[2];
uint64_t projid = ZFS_DEFAULT_PROJID;
znode_hold_t *zh;
/*
* skip ctldir, otherwise they will always get invalidated. This will
* cause funny behaviour for the mounted snapdirs. Especially for
* Linux >= 3.18, d_invalidate will detach the mountpoint and prevent
* anyone automount it again as long as someone is still using the
* detached mount.
*/
if (zp->z_is_ctldir)
return (0);
zh = zfs_znode_hold_enter(zfsvfs, obj_num);
mutex_enter(&zp->z_acl_lock);
if (zp->z_acl_cached) {
zfs_acl_free(zp->z_acl_cached);
zp->z_acl_cached = NULL;
}
mutex_exit(&zp->z_acl_lock);
rw_enter(&zp->z_xattr_lock, RW_WRITER);
if (zp->z_xattr_cached) {
nvlist_free(zp->z_xattr_cached);
zp->z_xattr_cached = NULL;
}
rw_exit(&zp->z_xattr_lock);
ASSERT(zp->z_sa_hdl == NULL);
err = sa_buf_hold(zfsvfs->z_os, obj_num, NULL, &db);
if (err) {
zfs_znode_hold_exit(zfsvfs, zh);
return (err);
}
dmu_object_info_from_db(db, &doi);
if (doi.doi_bonus_type != DMU_OT_SA &&
(doi.doi_bonus_type != DMU_OT_ZNODE ||
(doi.doi_bonus_type == DMU_OT_ZNODE &&
doi.doi_bonus_size < sizeof (znode_phys_t)))) {
sa_buf_rele(db, NULL);
zfs_znode_hold_exit(zfsvfs, zh);
return (SET_ERROR(EINVAL));
}
zfs_znode_sa_init(zfsvfs, zp, db, doi.doi_bonus_type, NULL);
/* reload cached values */
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GEN(zfsvfs), NULL,
&gen, sizeof (gen));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL,
&zp->z_size, sizeof (zp->z_size));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL,
&links, sizeof (links));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, sizeof (zp->z_pflags));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL,
&z_uid, sizeof (z_uid));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL,
&z_gid, sizeof (z_gid));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL,
&mode, sizeof (mode));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL,
&atime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL,
&mtime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
&ctime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CRTIME(zfsvfs), NULL, &btime, 16);
if (sa_bulk_lookup(zp->z_sa_hdl, bulk, count)) {
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
return (SET_ERROR(EIO));
}
if (dmu_objset_projectquota_enabled(zfsvfs->z_os)) {
err = sa_lookup(zp->z_sa_hdl, SA_ZPL_PROJID(zfsvfs),
&projid, 8);
if (err != 0 && err != ENOENT) {
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
return (SET_ERROR(err));
}
}
zp->z_projid = projid;
zp->z_mode = ZTOI(zp)->i_mode = mode;
zfs_uid_write(ZTOI(zp), z_uid);
zfs_gid_write(ZTOI(zp), z_gid);
ZFS_TIME_DECODE(&ZTOI(zp)->i_atime, atime);
ZFS_TIME_DECODE(&ZTOI(zp)->i_mtime, mtime);
ZFS_TIME_DECODE(&ZTOI(zp)->i_ctime, ctime);
ZFS_TIME_DECODE(&zp->z_btime, btime);
if ((uint32_t)gen != ZTOI(zp)->i_generation) {
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
return (SET_ERROR(EIO));
}
set_nlink(ZTOI(zp), (uint32_t)links);
zfs_set_inode_flags(zp, ZTOI(zp));
zp->z_blksz = doi.doi_data_block_size;
zp->z_atime_dirty = B_FALSE;
zfs_znode_update_vfs(zp);
/*
* If the file has zero links, then it has been unlinked on the send
* side and it must be in the received unlinked set.
* We call zfs_znode_dmu_fini() now to prevent any accesses to the
* stale data and to prevent automatic removal of the file in
* zfs_zinactive(). The file will be removed either when it is removed
* on the send side and the next incremental stream is received or
* when the unlinked set gets processed.
*/
zp->z_unlinked = (ZTOI(zp)->i_nlink == 0);
if (zp->z_unlinked)
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
return (0);
}
void
zfs_znode_delete(znode_t *zp, dmu_tx_t *tx)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
objset_t *os = zfsvfs->z_os;
uint64_t obj = zp->z_id;
uint64_t acl_obj = zfs_external_acl(zp);
znode_hold_t *zh;
zh = zfs_znode_hold_enter(zfsvfs, obj);
if (acl_obj) {
VERIFY(!zp->z_is_sa);
VERIFY(0 == dmu_object_free(os, acl_obj, tx));
}
VERIFY(0 == dmu_object_free(os, obj, tx));
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
}
void
zfs_zinactive(znode_t *zp)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
uint64_t z_id = zp->z_id;
znode_hold_t *zh;
ASSERT(zp->z_sa_hdl);
/*
* Don't allow a zfs_zget() while were trying to release this znode.
*/
zh = zfs_znode_hold_enter(zfsvfs, z_id);
mutex_enter(&zp->z_lock);
/*
* If this was the last reference to a file with no links, remove
* the file from the file system unless the file system is mounted
* read-only. That can happen, for example, if the file system was
* originally read-write, the file was opened, then unlinked and
* the file system was made read-only before the file was finally
* closed. The file will remain in the unlinked set.
*/
if (zp->z_unlinked) {
ASSERT(!zfsvfs->z_issnap);
if (!zfs_is_readonly(zfsvfs) && !zfs_unlink_suspend_progress) {
mutex_exit(&zp->z_lock);
zfs_znode_hold_exit(zfsvfs, zh);
zfs_rmnode(zp);
return;
}
}
mutex_exit(&zp->z_lock);
zfs_znode_dmu_fini(zp);
zfs_znode_hold_exit(zfsvfs, zh);
}
#if defined(HAVE_INODE_TIMESPEC64_TIMES)
#define zfs_compare_timespec timespec64_compare
#else
#define zfs_compare_timespec timespec_compare
#endif
/*
* Determine whether the znode's atime must be updated. The logic mostly
* duplicates the Linux kernel's relatime_need_update() functionality.
* This function is only called if the underlying filesystem actually has
* atime updates enabled.
*/
boolean_t
zfs_relatime_need_update(const struct inode *ip)
{
inode_timespec_t now;
gethrestime(&now);
/*
* In relatime mode, only update the atime if the previous atime
* is earlier than either the ctime or mtime or if at least a day
* has passed since the last update of atime.
*/
if (zfs_compare_timespec(&ip->i_mtime, &ip->i_atime) >= 0)
return (B_TRUE);
if (zfs_compare_timespec(&ip->i_ctime, &ip->i_atime) >= 0)
return (B_TRUE);
if ((hrtime_t)now.tv_sec - (hrtime_t)ip->i_atime.tv_sec >= 24*60*60)
return (B_TRUE);
return (B_FALSE);
}
/*
* Prepare to update znode time stamps.
*
* IN: zp - znode requiring timestamp update
* flag - ATTR_MTIME, ATTR_CTIME flags
*
* OUT: zp - z_seq
* mtime - new mtime
* ctime - new ctime
*
* Note: We don't update atime here, because we rely on Linux VFS to do
* atime updating.
*/
void
zfs_tstamp_update_setup(znode_t *zp, uint_t flag, uint64_t mtime[2],
uint64_t ctime[2])
{
inode_timespec_t now;
gethrestime(&now);
zp->z_seq++;
if (flag & ATTR_MTIME) {
ZFS_TIME_ENCODE(&now, mtime);
ZFS_TIME_DECODE(&(ZTOI(zp)->i_mtime), mtime);
if (ZTOZSB(zp)->z_use_fuids) {
zp->z_pflags |= (ZFS_ARCHIVE |
ZFS_AV_MODIFIED);
}
}
if (flag & ATTR_CTIME) {
ZFS_TIME_ENCODE(&now, ctime);
ZFS_TIME_DECODE(&(ZTOI(zp)->i_ctime), ctime);
if (ZTOZSB(zp)->z_use_fuids)
zp->z_pflags |= ZFS_ARCHIVE;
}
}
/*
* Grow the block size for a file.
*
* IN: zp - znode of file to free data in.
* size - requested block size
* tx - open transaction.
*
* NOTE: this function assumes that the znode is write locked.
*/
void
zfs_grow_blocksize(znode_t *zp, uint64_t size, dmu_tx_t *tx)
{
int error;
u_longlong_t dummy;
if (size <= zp->z_blksz)
return;
/*
* If the file size is already greater than the current blocksize,
* we will not grow. If there is more than one block in a file,
* the blocksize cannot change.
*/
if (zp->z_blksz && zp->z_size > zp->z_blksz)
return;
error = dmu_object_set_blocksize(ZTOZSB(zp)->z_os, zp->z_id,
size, 0, tx);
if (error == ENOTSUP)
return;
ASSERT0(error);
/* What blocksize did we actually get? */
dmu_object_size_from_db(sa_get_db(zp->z_sa_hdl), &zp->z_blksz, &dummy);
}
/*
* Increase the file length
*
* IN: zp - znode of file to free data in.
* end - new end-of-file
*
* RETURN: 0 on success, error code on failure
*/
static int
zfs_extend(znode_t *zp, uint64_t end)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
dmu_tx_t *tx;
zfs_locked_range_t *lr;
uint64_t newblksz;
int error;
/*
* We will change zp_size, lock the whole file.
*/
lr = zfs_rangelock_enter(&zp->z_rangelock, 0, UINT64_MAX, RL_WRITER);
/*
* Nothing to do if file already at desired length.
*/
if (end <= zp->z_size) {
zfs_rangelock_exit(lr);
return (0);
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
if (end > zp->z_blksz &&
(!ISP2(zp->z_blksz) || zp->z_blksz < zfsvfs->z_max_blksz)) {
/*
* We are growing the file past the current block size.
*/
if (zp->z_blksz > ZTOZSB(zp)->z_max_blksz) {
/*
* File's blocksize is already larger than the
* "recordsize" property. Only let it grow to
* the next power of 2.
*/
ASSERT(!ISP2(zp->z_blksz));
newblksz = MIN(end, 1 << highbit64(zp->z_blksz));
} else {
newblksz = MIN(end, ZTOZSB(zp)->z_max_blksz);
}
dmu_tx_hold_write(tx, zp->z_id, 0, newblksz);
} else {
newblksz = 0;
}
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
zfs_rangelock_exit(lr);
return (error);
}
if (newblksz)
zfs_grow_blocksize(zp, newblksz, tx);
zp->z_size = end;
VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(ZTOZSB(zp)),
&zp->z_size, sizeof (zp->z_size), tx));
zfs_rangelock_exit(lr);
dmu_tx_commit(tx);
return (0);
}
/*
* zfs_zero_partial_page - Modeled after update_pages() but
* with different arguments and semantics for use by zfs_freesp().
*
* Zeroes a piece of a single page cache entry for zp at offset
* start and length len.
*
* Caller must acquire a range lock on the file for the region
* being zeroed in order that the ARC and page cache stay in sync.
*/
static void
zfs_zero_partial_page(znode_t *zp, uint64_t start, uint64_t len)
{
struct address_space *mp = ZTOI(zp)->i_mapping;
struct page *pp;
int64_t off;
void *pb;
ASSERT((start & PAGE_MASK) == ((start + len - 1) & PAGE_MASK));
off = start & (PAGE_SIZE - 1);
start &= PAGE_MASK;
pp = find_lock_page(mp, start >> PAGE_SHIFT);
if (pp) {
if (mapping_writably_mapped(mp))
flush_dcache_page(pp);
pb = kmap(pp);
memset(pb + off, 0, len);
kunmap(pp);
if (mapping_writably_mapped(mp))
flush_dcache_page(pp);
mark_page_accessed(pp);
SetPageUptodate(pp);
ClearPageError(pp);
unlock_page(pp);
put_page(pp);
}
}
/*
* Free space in a file.
*
* IN: zp - znode of file to free data in.
* off - start of section to free.
* len - length of section to free.
*
* RETURN: 0 on success, error code on failure
*/
static int
zfs_free_range(znode_t *zp, uint64_t off, uint64_t len)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
zfs_locked_range_t *lr;
int error;
/*
* Lock the range being freed.
*/
lr = zfs_rangelock_enter(&zp->z_rangelock, off, len, RL_WRITER);
/*
* Nothing to do if file already at desired length.
*/
if (off >= zp->z_size) {
zfs_rangelock_exit(lr);
return (0);
}
if (off + len > zp->z_size)
len = zp->z_size - off;
error = dmu_free_long_range(zfsvfs->z_os, zp->z_id, off, len);
/*
* Zero partial page cache entries. This must be done under a
* range lock in order to keep the ARC and page cache in sync.
*/
if (zn_has_cached_data(zp, off, off + len - 1)) {
loff_t first_page, last_page, page_len;
loff_t first_page_offset, last_page_offset;
/* first possible full page in hole */
first_page = (off + PAGE_SIZE - 1) >> PAGE_SHIFT;
/* last page of hole */
last_page = (off + len) >> PAGE_SHIFT;
/* offset of first_page */
first_page_offset = first_page << PAGE_SHIFT;
/* offset of last_page */
last_page_offset = last_page << PAGE_SHIFT;
/* truncate whole pages */
if (last_page_offset > first_page_offset) {
truncate_inode_pages_range(ZTOI(zp)->i_mapping,
first_page_offset, last_page_offset - 1);
}
/* truncate sub-page ranges */
if (first_page > last_page) {
/* entire punched area within a single page */
zfs_zero_partial_page(zp, off, len);
} else {
/* beginning of punched area at the end of a page */
page_len = first_page_offset - off;
if (page_len > 0)
zfs_zero_partial_page(zp, off, page_len);
/* end of punched area at the beginning of a page */
page_len = off + len - last_page_offset;
if (page_len > 0)
zfs_zero_partial_page(zp, last_page_offset,
page_len);
}
}
zfs_rangelock_exit(lr);
return (error);
}
/*
* Truncate a file
*
* IN: zp - znode of file to free data in.
* end - new end-of-file.
*
* RETURN: 0 on success, error code on failure
*/
static int
zfs_trunc(znode_t *zp, uint64_t end)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
dmu_tx_t *tx;
zfs_locked_range_t *lr;
int error;
sa_bulk_attr_t bulk[2];
int count = 0;
/*
* We will change zp_size, lock the whole file.
*/
lr = zfs_rangelock_enter(&zp->z_rangelock, 0, UINT64_MAX, RL_WRITER);
/*
* Nothing to do if file already at desired length.
*/
if (end >= zp->z_size) {
zfs_rangelock_exit(lr);
return (0);
}
error = dmu_free_long_range(zfsvfs->z_os, zp->z_id, end,
DMU_OBJECT_END);
if (error) {
zfs_rangelock_exit(lr);
return (error);
}
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
zfs_rangelock_exit(lr);
return (error);
}
zp->z_size = end;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs),
NULL, &zp->z_size, sizeof (zp->z_size));
if (end == 0) {
zp->z_pflags &= ~ZFS_SPARSE;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs),
NULL, &zp->z_pflags, 8);
}
VERIFY(sa_bulk_update(zp->z_sa_hdl, bulk, count, tx) == 0);
dmu_tx_commit(tx);
zfs_rangelock_exit(lr);
return (0);
}
/*
* Free space in a file
*
* IN: zp - znode of file to free data in.
* off - start of range
* len - end of range (0 => EOF)
* flag - current file open mode flags.
* log - TRUE if this action should be logged
*
* RETURN: 0 on success, error code on failure
*/
int
zfs_freesp(znode_t *zp, uint64_t off, uint64_t len, int flag, boolean_t log)
{
dmu_tx_t *tx;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
zilog_t *zilog = zfsvfs->z_log;
uint64_t mode;
uint64_t mtime[2], ctime[2];
sa_bulk_attr_t bulk[3];
int count = 0;
int error;
if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_MODE(zfsvfs), &mode,
sizeof (mode))) != 0)
return (error);
if (off > zp->z_size) {
error = zfs_extend(zp, off+len);
if (error == 0 && log)
goto log;
goto out;
}
if (len == 0) {
error = zfs_trunc(zp, off);
} else {
if ((error = zfs_free_range(zp, off, len)) == 0 &&
off + len > zp->z_size)
error = zfs_extend(zp, off+len);
}
if (error || !log)
goto out;
log:
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
goto out;
}
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, mtime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, ctime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs),
NULL, &zp->z_pflags, 8);
zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime);
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
ASSERT(error == 0);
zfs_log_truncate(zilog, tx, TX_TRUNCATE, zp, off, len);
dmu_tx_commit(tx);
zfs_znode_update_vfs(zp);
error = 0;
out:
/*
* Truncate the page cache - for file truncate operations, use
* the purpose-built API for truncations. For punching operations,
* the truncation is handled under a range lock in zfs_free_range.
*/
if (len == 0)
truncate_setsize(ZTOI(zp), off);
return (error);
}
void
zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
{
struct super_block *sb;
zfsvfs_t *zfsvfs;
uint64_t moid, obj, sa_obj, version;
uint64_t sense = ZFS_CASE_SENSITIVE;
uint64_t norm = 0;
nvpair_t *elem;
int size;
int error;
int i;
znode_t *rootzp = NULL;
vattr_t vattr;
znode_t *zp;
zfs_acl_ids_t acl_ids;
/*
* First attempt to create master node.
*/
/*
* In an empty objset, there are no blocks to read and thus
* there can be no i/o errors (which we assert below).
*/
moid = MASTER_NODE_OBJ;
error = zap_create_claim(os, moid, DMU_OT_MASTER_NODE,
DMU_OT_NONE, 0, tx);
ASSERT(error == 0);
/*
* Set starting attributes.
*/
version = zfs_zpl_version_map(spa_version(dmu_objset_spa(os)));
elem = NULL;
while ((elem = nvlist_next_nvpair(zplprops, elem)) != NULL) {
/* For the moment we expect all zpl props to be uint64_ts */
uint64_t val;
const char *name;
ASSERT(nvpair_type(elem) == DATA_TYPE_UINT64);
VERIFY(nvpair_value_uint64(elem, &val) == 0);
name = nvpair_name(elem);
if (strcmp(name, zfs_prop_to_name(ZFS_PROP_VERSION)) == 0) {
if (val < version)
version = val;
} else {
error = zap_update(os, moid, name, 8, 1, &val, tx);
}
ASSERT(error == 0);
if (strcmp(name, zfs_prop_to_name(ZFS_PROP_NORMALIZE)) == 0)
norm = val;
else if (strcmp(name, zfs_prop_to_name(ZFS_PROP_CASE)) == 0)
sense = val;
}
ASSERT(version != 0);
error = zap_update(os, moid, ZPL_VERSION_STR, 8, 1, &version, tx);
ASSERT(error == 0);
/*
* Create zap object used for SA attribute registration
*/
if (version >= ZPL_VERSION_SA) {
sa_obj = zap_create(os, DMU_OT_SA_MASTER_NODE,
DMU_OT_NONE, 0, tx);
error = zap_add(os, moid, ZFS_SA_ATTRS, 8, 1, &sa_obj, tx);
ASSERT(error == 0);
} else {
sa_obj = 0;
}
/*
* Create a delete queue.
*/
obj = zap_create(os, DMU_OT_UNLINKED_SET, DMU_OT_NONE, 0, tx);
error = zap_add(os, moid, ZFS_UNLINKED_SET, 8, 1, &obj, tx);
ASSERT(error == 0);
/*
* Create root znode. Create minimal znode/inode/zfsvfs/sb
* to allow zfs_mknode to work.
*/
vattr.va_mask = ATTR_MODE|ATTR_UID|ATTR_GID;
vattr.va_mode = S_IFDIR|0755;
vattr.va_uid = crgetuid(cr);
vattr.va_gid = crgetgid(cr);
rootzp = kmem_cache_alloc(znode_cache, KM_SLEEP);
rootzp->z_unlinked = B_FALSE;
rootzp->z_atime_dirty = B_FALSE;
rootzp->z_is_sa = USE_SA(version, os);
rootzp->z_pflags = 0;
zfsvfs = kmem_zalloc(sizeof (zfsvfs_t), KM_SLEEP);
zfsvfs->z_os = os;
zfsvfs->z_parent = zfsvfs;
zfsvfs->z_version = version;
zfsvfs->z_use_fuids = USE_FUIDS(version, os);
zfsvfs->z_use_sa = USE_SA(version, os);
zfsvfs->z_norm = norm;
sb = kmem_zalloc(sizeof (struct super_block), KM_SLEEP);
sb->s_fs_info = zfsvfs;
ZTOI(rootzp)->i_sb = sb;
error = sa_setup(os, sa_obj, zfs_attr_table, ZPL_END,
&zfsvfs->z_attr_table);
ASSERT(error == 0);
/*
* Fold case on file systems that are always or sometimes case
* insensitive.
*/
if (sense == ZFS_CASE_INSENSITIVE || sense == ZFS_CASE_MIXED)
zfsvfs->z_norm |= U8_TEXTPREP_TOUPPER;
mutex_init(&zfsvfs->z_znodes_lock, NULL, MUTEX_DEFAULT, NULL);
list_create(&zfsvfs->z_all_znodes, sizeof (znode_t),
offsetof(znode_t, z_link_node));
size = MIN(1 << (highbit64(zfs_object_mutex_size)-1), ZFS_OBJ_MTX_MAX);
zfsvfs->z_hold_size = size;
zfsvfs->z_hold_trees = vmem_zalloc(sizeof (avl_tree_t) * size,
KM_SLEEP);
zfsvfs->z_hold_locks = vmem_zalloc(sizeof (kmutex_t) * size, KM_SLEEP);
for (i = 0; i != size; i++) {
avl_create(&zfsvfs->z_hold_trees[i], zfs_znode_hold_compare,
sizeof (znode_hold_t), offsetof(znode_hold_t, zh_node));
mutex_init(&zfsvfs->z_hold_locks[i], NULL, MUTEX_DEFAULT, NULL);
}
VERIFY(0 == zfs_acl_ids_create(rootzp, IS_ROOT_NODE, &vattr,
- cr, NULL, &acl_ids, kcred->user_ns));
+ cr, NULL, &acl_ids, zfs_init_idmap));
zfs_mknode(rootzp, &vattr, tx, cr, IS_ROOT_NODE, &zp, &acl_ids);
ASSERT3P(zp, ==, rootzp);
error = zap_add(os, moid, ZFS_ROOT_OBJ, 8, 1, &rootzp->z_id, tx);
ASSERT(error == 0);
zfs_acl_ids_free(&acl_ids);
atomic_set(&ZTOI(rootzp)->i_count, 0);
sa_handle_destroy(rootzp->z_sa_hdl);
kmem_cache_free(znode_cache, rootzp);
for (i = 0; i != size; i++) {
avl_destroy(&zfsvfs->z_hold_trees[i]);
mutex_destroy(&zfsvfs->z_hold_locks[i]);
}
mutex_destroy(&zfsvfs->z_znodes_lock);
vmem_free(zfsvfs->z_hold_trees, sizeof (avl_tree_t) * size);
vmem_free(zfsvfs->z_hold_locks, sizeof (kmutex_t) * size);
kmem_free(sb, sizeof (struct super_block));
kmem_free(zfsvfs, sizeof (zfsvfs_t));
}
#endif /* _KERNEL */
static int
zfs_sa_setup(objset_t *osp, sa_attr_type_t **sa_table)
{
uint64_t sa_obj = 0;
int error;
error = zap_lookup(osp, MASTER_NODE_OBJ, ZFS_SA_ATTRS, 8, 1, &sa_obj);
if (error != 0 && error != ENOENT)
return (error);
error = sa_setup(osp, sa_obj, zfs_attr_table, ZPL_END, sa_table);
return (error);
}
static int
zfs_grab_sa_handle(objset_t *osp, uint64_t obj, sa_handle_t **hdlp,
dmu_buf_t **db, const void *tag)
{
dmu_object_info_t doi;
int error;
if ((error = sa_buf_hold(osp, obj, tag, db)) != 0)
return (error);
dmu_object_info_from_db(*db, &doi);
if ((doi.doi_bonus_type != DMU_OT_SA &&
doi.doi_bonus_type != DMU_OT_ZNODE) ||
(doi.doi_bonus_type == DMU_OT_ZNODE &&
doi.doi_bonus_size < sizeof (znode_phys_t))) {
sa_buf_rele(*db, tag);
return (SET_ERROR(ENOTSUP));
}
error = sa_handle_get(osp, obj, NULL, SA_HDL_PRIVATE, hdlp);
if (error != 0) {
sa_buf_rele(*db, tag);
return (error);
}
return (0);
}
static void
zfs_release_sa_handle(sa_handle_t *hdl, dmu_buf_t *db, const void *tag)
{
sa_handle_destroy(hdl);
sa_buf_rele(db, tag);
}
/*
* Given an object number, return its parent object number and whether
* or not the object is an extended attribute directory.
*/
static int
zfs_obj_to_pobj(objset_t *osp, sa_handle_t *hdl, sa_attr_type_t *sa_table,
uint64_t *pobjp, int *is_xattrdir)
{
uint64_t parent;
uint64_t pflags;
uint64_t mode;
uint64_t parent_mode;
sa_bulk_attr_t bulk[3];
sa_handle_t *sa_hdl;
dmu_buf_t *sa_db;
int count = 0;
int error;
SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_PARENT], NULL,
&parent, sizeof (parent));
SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_FLAGS], NULL,
&pflags, sizeof (pflags));
SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_MODE], NULL,
&mode, sizeof (mode));
if ((error = sa_bulk_lookup(hdl, bulk, count)) != 0)
return (error);
/*
* When a link is removed its parent pointer is not changed and will
* be invalid. There are two cases where a link is removed but the
* file stays around, when it goes to the delete queue and when there
* are additional links.
*/
error = zfs_grab_sa_handle(osp, parent, &sa_hdl, &sa_db, FTAG);
if (error != 0)
return (error);
error = sa_lookup(sa_hdl, ZPL_MODE, &parent_mode, sizeof (parent_mode));
zfs_release_sa_handle(sa_hdl, sa_db, FTAG);
if (error != 0)
return (error);
*is_xattrdir = ((pflags & ZFS_XATTR) != 0) && S_ISDIR(mode);
/*
* Extended attributes can be applied to files, directories, etc.
* Otherwise the parent must be a directory.
*/
if (!*is_xattrdir && !S_ISDIR(parent_mode))
return (SET_ERROR(EINVAL));
*pobjp = parent;
return (0);
}
/*
* Given an object number, return some zpl level statistics
*/
static int
zfs_obj_to_stats_impl(sa_handle_t *hdl, sa_attr_type_t *sa_table,
zfs_stat_t *sb)
{
sa_bulk_attr_t bulk[4];
int count = 0;
SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_MODE], NULL,
&sb->zs_mode, sizeof (sb->zs_mode));
SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_GEN], NULL,
&sb->zs_gen, sizeof (sb->zs_gen));
SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_LINKS], NULL,
&sb->zs_links, sizeof (sb->zs_links));
SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_CTIME], NULL,
&sb->zs_ctime, sizeof (sb->zs_ctime));
return (sa_bulk_lookup(hdl, bulk, count));
}
static int
zfs_obj_to_path_impl(objset_t *osp, uint64_t obj, sa_handle_t *hdl,
sa_attr_type_t *sa_table, char *buf, int len)
{
sa_handle_t *sa_hdl;
sa_handle_t *prevhdl = NULL;
dmu_buf_t *prevdb = NULL;
dmu_buf_t *sa_db = NULL;
char *path = buf + len - 1;
int error;
*path = '\0';
sa_hdl = hdl;
uint64_t deleteq_obj;
VERIFY0(zap_lookup(osp, MASTER_NODE_OBJ,
ZFS_UNLINKED_SET, sizeof (uint64_t), 1, &deleteq_obj));
error = zap_lookup_int(osp, deleteq_obj, obj);
if (error == 0) {
return (ESTALE);
} else if (error != ENOENT) {
return (error);
}
for (;;) {
uint64_t pobj = 0;
char component[MAXNAMELEN + 2];
size_t complen;
int is_xattrdir = 0;
if (prevdb) {
ASSERT(prevhdl != NULL);
zfs_release_sa_handle(prevhdl, prevdb, FTAG);
}
if ((error = zfs_obj_to_pobj(osp, sa_hdl, sa_table, &pobj,
&is_xattrdir)) != 0)
break;
if (pobj == obj) {
if (path[0] != '/')
*--path = '/';
break;
}
component[0] = '/';
if (is_xattrdir) {
strcpy(component + 1, "<xattrdir>");
} else {
error = zap_value_search(osp, pobj, obj,
ZFS_DIRENT_OBJ(-1ULL), component + 1);
if (error != 0)
break;
}
complen = strlen(component);
path -= complen;
ASSERT(path >= buf);
memcpy(path, component, complen);
obj = pobj;
if (sa_hdl != hdl) {
prevhdl = sa_hdl;
prevdb = sa_db;
}
error = zfs_grab_sa_handle(osp, obj, &sa_hdl, &sa_db, FTAG);
if (error != 0) {
sa_hdl = prevhdl;
sa_db = prevdb;
break;
}
}
if (sa_hdl != NULL && sa_hdl != hdl) {
ASSERT(sa_db != NULL);
zfs_release_sa_handle(sa_hdl, sa_db, FTAG);
}
if (error == 0)
(void) memmove(buf, path, buf + len - path);
return (error);
}
int
zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len)
{
sa_attr_type_t *sa_table;
sa_handle_t *hdl;
dmu_buf_t *db;
int error;
error = zfs_sa_setup(osp, &sa_table);
if (error != 0)
return (error);
error = zfs_grab_sa_handle(osp, obj, &hdl, &db, FTAG);
if (error != 0)
return (error);
error = zfs_obj_to_path_impl(osp, obj, hdl, sa_table, buf, len);
zfs_release_sa_handle(hdl, db, FTAG);
return (error);
}
int
zfs_obj_to_stats(objset_t *osp, uint64_t obj, zfs_stat_t *sb,
char *buf, int len)
{
char *path = buf + len - 1;
sa_attr_type_t *sa_table;
sa_handle_t *hdl;
dmu_buf_t *db;
int error;
*path = '\0';
error = zfs_sa_setup(osp, &sa_table);
if (error != 0)
return (error);
error = zfs_grab_sa_handle(osp, obj, &hdl, &db, FTAG);
if (error != 0)
return (error);
error = zfs_obj_to_stats_impl(hdl, sa_table, sb);
if (error != 0) {
zfs_release_sa_handle(hdl, db, FTAG);
return (error);
}
error = zfs_obj_to_path_impl(osp, obj, hdl, sa_table, buf, len);
zfs_release_sa_handle(hdl, db, FTAG);
return (error);
}
#if defined(_KERNEL)
EXPORT_SYMBOL(zfs_create_fs);
EXPORT_SYMBOL(zfs_obj_to_path);
/* CSTYLED */
module_param(zfs_object_mutex_size, uint, 0644);
MODULE_PARM_DESC(zfs_object_mutex_size, "Size of znode hold array");
module_param(zfs_unlink_suspend_progress, int, 0644);
MODULE_PARM_DESC(zfs_unlink_suspend_progress, "Set to prevent async unlinks "
"(debug - leaks space into the unlinked set)");
#endif
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zpl_ctldir.c b/sys/contrib/openzfs/module/os/linux/zfs/zpl_ctldir.c
index f0779c81dc75..68a7de78f471 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zpl_ctldir.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zpl_ctldir.c
@@ -1,639 +1,664 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (C) 2011 Lawrence Livermore National Security, LLC.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
* LLNL-CODE-403049.
* Rewritten for Linux by:
* Rohan Puri <rohan.puri15@gmail.com>
* Brian Behlendorf <behlendorf1@llnl.gov>
*/
#include <sys/zfs_znode.h>
#include <sys/zfs_vfsops.h>
#include <sys/zfs_vnops.h>
#include <sys/zfs_ctldir.h>
#include <sys/zpl.h>
#include <sys/dmu.h>
#include <sys/dsl_dataset.h>
#include <sys/zap.h>
/*
* Common open routine. Disallow any write access.
*/
static int
zpl_common_open(struct inode *ip, struct file *filp)
{
if (filp->f_mode & FMODE_WRITE)
return (-EACCES);
return (generic_file_open(ip, filp));
}
/*
* Get root directory contents.
*/
static int
zpl_root_iterate(struct file *filp, zpl_dir_context_t *ctx)
{
zfsvfs_t *zfsvfs = ITOZSB(file_inode(filp));
int error = 0;
if ((error = zpl_enter(zfsvfs, FTAG)) != 0)
return (error);
if (!zpl_dir_emit_dots(filp, ctx))
goto out;
if (ctx->pos == 2) {
if (!zpl_dir_emit(ctx, ZFS_SNAPDIR_NAME,
strlen(ZFS_SNAPDIR_NAME), ZFSCTL_INO_SNAPDIR, DT_DIR))
goto out;
ctx->pos++;
}
if (ctx->pos == 3) {
if (!zpl_dir_emit(ctx, ZFS_SHAREDIR_NAME,
strlen(ZFS_SHAREDIR_NAME), ZFSCTL_INO_SHARES, DT_DIR))
goto out;
ctx->pos++;
}
out:
zpl_exit(zfsvfs, FTAG);
return (error);
}
#if !defined(HAVE_VFS_ITERATE) && !defined(HAVE_VFS_ITERATE_SHARED)
static int
zpl_root_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
zpl_dir_context_t ctx =
ZPL_DIR_CONTEXT_INIT(dirent, filldir, filp->f_pos);
int error;
error = zpl_root_iterate(filp, &ctx);
filp->f_pos = ctx.pos;
return (error);
}
#endif /* !HAVE_VFS_ITERATE && !HAVE_VFS_ITERATE_SHARED */
/*
* Get root directory attributes.
*/
static int
-#ifdef HAVE_USERNS_IOPS_GETATTR
+#ifdef HAVE_IDMAP_IOPS_GETATTR
+zpl_root_getattr_impl(struct mnt_idmap *user_ns,
+ const struct path *path, struct kstat *stat, u32 request_mask,
+ unsigned int query_flags)
+#elif defined(HAVE_USERNS_IOPS_GETATTR)
zpl_root_getattr_impl(struct user_namespace *user_ns,
const struct path *path, struct kstat *stat, u32 request_mask,
unsigned int query_flags)
#else
zpl_root_getattr_impl(const struct path *path, struct kstat *stat,
u32 request_mask, unsigned int query_flags)
#endif
{
(void) request_mask, (void) query_flags;
struct inode *ip = path->dentry->d_inode;
-#ifdef HAVE_USERNS_IOPS_GETATTR
+#if (defined(HAVE_USERNS_IOPS_GETATTR) || defined(HAVE_IDMAP_IOPS_GETATTR))
#ifdef HAVE_GENERIC_FILLATTR_USERNS
generic_fillattr(user_ns, ip, stat);
+#elif defined(HAVE_GENERIC_FILLATTR_IDMAP)
+ generic_fillattr(user_ns, ip, stat);
#else
(void) user_ns;
#endif
#else
generic_fillattr(ip, stat);
#endif
stat->atime = current_time(ip);
return (0);
}
ZPL_GETATTR_WRAPPER(zpl_root_getattr);
static struct dentry *
zpl_root_lookup(struct inode *dip, struct dentry *dentry, unsigned int flags)
{
cred_t *cr = CRED();
struct inode *ip;
int error;
crhold(cr);
error = -zfsctl_root_lookup(dip, dname(dentry), &ip, 0, cr, NULL, NULL);
ASSERT3S(error, <=, 0);
crfree(cr);
if (error) {
if (error == -ENOENT)
return (d_splice_alias(NULL, dentry));
else
return (ERR_PTR(error));
}
return (d_splice_alias(ip, dentry));
}
/*
* The '.zfs' control directory file and inode operations.
*/
const struct file_operations zpl_fops_root = {
.open = zpl_common_open,
.llseek = generic_file_llseek,
.read = generic_read_dir,
#ifdef HAVE_VFS_ITERATE_SHARED
.iterate_shared = zpl_root_iterate,
#elif defined(HAVE_VFS_ITERATE)
.iterate = zpl_root_iterate,
#else
.readdir = zpl_root_readdir,
#endif
};
const struct inode_operations zpl_ops_root = {
.lookup = zpl_root_lookup,
.getattr = zpl_root_getattr,
};
static struct vfsmount *
zpl_snapdir_automount(struct path *path)
{
int error;
error = -zfsctl_snapshot_mount(path, 0);
if (error)
return (ERR_PTR(error));
/*
* Rather than returning the new vfsmount for the snapshot we must
* return NULL to indicate a mount collision. This is done because
* the user space mount calls do_add_mount() which adds the vfsmount
* to the name space. If we returned the new mount here it would be
* added again to the vfsmount list resulting in list corruption.
*/
return (NULL);
}
/*
* Negative dentries must always be revalidated so newly created snapshots
* can be detected and automounted. Normal dentries should be kept because
* as of the 3.18 kernel revaliding the mountpoint dentry will result in
* the snapshot being immediately unmounted.
*/
static int
#ifdef HAVE_D_REVALIDATE_NAMEIDATA
zpl_snapdir_revalidate(struct dentry *dentry, struct nameidata *i)
#else
zpl_snapdir_revalidate(struct dentry *dentry, unsigned int flags)
#endif
{
return (!!dentry->d_inode);
}
static dentry_operations_t zpl_dops_snapdirs = {
/*
* Auto mounting of snapshots is only supported for 2.6.37 and
* newer kernels. Prior to this kernel the ops->follow_link()
* callback was used as a hack to trigger the mount. The
* resulting vfsmount was then explicitly grafted in to the
* name space. While it might be possible to add compatibility
* code to accomplish this it would require considerable care.
*/
.d_automount = zpl_snapdir_automount,
.d_revalidate = zpl_snapdir_revalidate,
};
static struct dentry *
zpl_snapdir_lookup(struct inode *dip, struct dentry *dentry,
unsigned int flags)
{
fstrans_cookie_t cookie;
cred_t *cr = CRED();
struct inode *ip = NULL;
int error;
crhold(cr);
cookie = spl_fstrans_mark();
error = -zfsctl_snapdir_lookup(dip, dname(dentry), &ip,
0, cr, NULL, NULL);
ASSERT3S(error, <=, 0);
spl_fstrans_unmark(cookie);
crfree(cr);
if (error && error != -ENOENT)
return (ERR_PTR(error));
ASSERT(error == 0 || ip == NULL);
d_clear_d_op(dentry);
d_set_d_op(dentry, &zpl_dops_snapdirs);
dentry->d_flags |= DCACHE_NEED_AUTOMOUNT;
return (d_splice_alias(ip, dentry));
}
static int
zpl_snapdir_iterate(struct file *filp, zpl_dir_context_t *ctx)
{
zfsvfs_t *zfsvfs = ITOZSB(file_inode(filp));
fstrans_cookie_t cookie;
char snapname[MAXNAMELEN];
boolean_t case_conflict;
uint64_t id, pos;
int error = 0;
if ((error = zpl_enter(zfsvfs, FTAG)) != 0)
return (error);
cookie = spl_fstrans_mark();
if (!zpl_dir_emit_dots(filp, ctx))
goto out;
/* Start the position at 0 if it already emitted . and .. */
pos = (ctx->pos == 2 ? 0 : ctx->pos);
while (error == 0) {
dsl_pool_config_enter(dmu_objset_pool(zfsvfs->z_os), FTAG);
error = -dmu_snapshot_list_next(zfsvfs->z_os, MAXNAMELEN,
snapname, &id, &pos, &case_conflict);
dsl_pool_config_exit(dmu_objset_pool(zfsvfs->z_os), FTAG);
if (error)
goto out;
if (!zpl_dir_emit(ctx, snapname, strlen(snapname),
ZFSCTL_INO_SHARES - id, DT_DIR))
goto out;
ctx->pos = pos;
}
out:
spl_fstrans_unmark(cookie);
zpl_exit(zfsvfs, FTAG);
if (error == -ENOENT)
return (0);
return (error);
}
#if !defined(HAVE_VFS_ITERATE) && !defined(HAVE_VFS_ITERATE_SHARED)
static int
zpl_snapdir_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
zpl_dir_context_t ctx =
ZPL_DIR_CONTEXT_INIT(dirent, filldir, filp->f_pos);
int error;
error = zpl_snapdir_iterate(filp, &ctx);
filp->f_pos = ctx.pos;
return (error);
}
#endif /* !HAVE_VFS_ITERATE && !HAVE_VFS_ITERATE_SHARED */
static int
#ifdef HAVE_IOPS_RENAME_USERNS
zpl_snapdir_rename2(struct user_namespace *user_ns, struct inode *sdip,
struct dentry *sdentry, struct inode *tdip, struct dentry *tdentry,
unsigned int flags)
+#elif defined(HAVE_IOPS_RENAME_IDMAP)
+zpl_snapdir_rename2(struct mnt_idmap *user_ns, struct inode *sdip,
+ struct dentry *sdentry, struct inode *tdip, struct dentry *tdentry,
+ unsigned int flags)
#else
zpl_snapdir_rename2(struct inode *sdip, struct dentry *sdentry,
struct inode *tdip, struct dentry *tdentry, unsigned int flags)
#endif
{
cred_t *cr = CRED();
int error;
/* We probably don't want to support renameat2(2) in ctldir */
if (flags)
return (-EINVAL);
crhold(cr);
error = -zfsctl_snapdir_rename(sdip, dname(sdentry),
tdip, dname(tdentry), cr, 0);
ASSERT3S(error, <=, 0);
crfree(cr);
return (error);
}
-#if !defined(HAVE_RENAME_WANTS_FLAGS) && !defined(HAVE_IOPS_RENAME_USERNS)
+#if (!defined(HAVE_RENAME_WANTS_FLAGS) && \
+ !defined(HAVE_IOPS_RENAME_USERNS) && \
+ !defined(HAVE_IOPS_RENAME_IDMAP))
static int
zpl_snapdir_rename(struct inode *sdip, struct dentry *sdentry,
struct inode *tdip, struct dentry *tdentry)
{
return (zpl_snapdir_rename2(sdip, sdentry, tdip, tdentry, 0));
}
#endif
static int
zpl_snapdir_rmdir(struct inode *dip, struct dentry *dentry)
{
cred_t *cr = CRED();
int error;
crhold(cr);
error = -zfsctl_snapdir_remove(dip, dname(dentry), cr, 0);
ASSERT3S(error, <=, 0);
crfree(cr);
return (error);
}
static int
#ifdef HAVE_IOPS_MKDIR_USERNS
zpl_snapdir_mkdir(struct user_namespace *user_ns, struct inode *dip,
struct dentry *dentry, umode_t mode)
+#elif defined(HAVE_IOPS_MKDIR_IDMAP)
+zpl_snapdir_mkdir(struct mnt_idmap *user_ns, struct inode *dip,
+ struct dentry *dentry, umode_t mode)
#else
zpl_snapdir_mkdir(struct inode *dip, struct dentry *dentry, umode_t mode)
#endif
{
cred_t *cr = CRED();
vattr_t *vap;
struct inode *ip;
int error;
crhold(cr);
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
-#ifdef HAVE_IOPS_MKDIR_USERNS
+#if (defined(HAVE_IOPS_MKDIR_USERNS) || defined(HAVE_IOPS_MKDIR_IDMAP))
zpl_vap_init(vap, dip, mode | S_IFDIR, cr, user_ns);
#else
- zpl_vap_init(vap, dip, mode | S_IFDIR, cr, kcred->user_ns);
+ zpl_vap_init(vap, dip, mode | S_IFDIR, cr, zfs_init_idmap);
#endif
error = -zfsctl_snapdir_mkdir(dip, dname(dentry), vap, &ip, cr, 0);
if (error == 0) {
d_clear_d_op(dentry);
d_set_d_op(dentry, &zpl_dops_snapdirs);
d_instantiate(dentry, ip);
}
kmem_free(vap, sizeof (vattr_t));
ASSERT3S(error, <=, 0);
crfree(cr);
return (error);
}
/*
* Get snapshot directory attributes.
*/
static int
-#ifdef HAVE_USERNS_IOPS_GETATTR
+#ifdef HAVE_IDMAP_IOPS_GETATTR
+zpl_snapdir_getattr_impl(struct mnt_idmap *user_ns,
+ const struct path *path, struct kstat *stat, u32 request_mask,
+ unsigned int query_flags)
+#elif defined(HAVE_USERNS_IOPS_GETATTR)
zpl_snapdir_getattr_impl(struct user_namespace *user_ns,
const struct path *path, struct kstat *stat, u32 request_mask,
unsigned int query_flags)
#else
zpl_snapdir_getattr_impl(const struct path *path, struct kstat *stat,
u32 request_mask, unsigned int query_flags)
#endif
{
(void) request_mask, (void) query_flags;
struct inode *ip = path->dentry->d_inode;
zfsvfs_t *zfsvfs = ITOZSB(ip);
int error;
if ((error = zpl_enter(zfsvfs, FTAG)) != 0)
return (error);
-#ifdef HAVE_USERNS_IOPS_GETATTR
+#if (defined(HAVE_USERNS_IOPS_GETATTR) || defined(HAVE_IDMAP_IOPS_GETATTR))
#ifdef HAVE_GENERIC_FILLATTR_USERNS
generic_fillattr(user_ns, ip, stat);
+#elif defined(HAVE_GENERIC_FILLATTR_IDMAP)
+ generic_fillattr(user_ns, ip, stat);
#else
(void) user_ns;
#endif
#else
generic_fillattr(ip, stat);
#endif
stat->nlink = stat->size = 2;
dsl_dataset_t *ds = dmu_objset_ds(zfsvfs->z_os);
if (dsl_dataset_phys(ds)->ds_snapnames_zapobj != 0) {
uint64_t snap_count;
int err = zap_count(
dmu_objset_pool(ds->ds_objset)->dp_meta_objset,
dsl_dataset_phys(ds)->ds_snapnames_zapobj, &snap_count);
if (err != 0) {
zpl_exit(zfsvfs, FTAG);
return (-err);
}
stat->nlink += snap_count;
}
stat->ctime = stat->mtime = dmu_objset_snap_cmtime(zfsvfs->z_os);
stat->atime = current_time(ip);
zpl_exit(zfsvfs, FTAG);
return (0);
}
ZPL_GETATTR_WRAPPER(zpl_snapdir_getattr);
/*
* The '.zfs/snapshot' directory file operations. These mainly control
* generating the list of available snapshots when doing an 'ls' in the
* directory. See zpl_snapdir_readdir().
*/
const struct file_operations zpl_fops_snapdir = {
.open = zpl_common_open,
.llseek = generic_file_llseek,
.read = generic_read_dir,
#ifdef HAVE_VFS_ITERATE_SHARED
.iterate_shared = zpl_snapdir_iterate,
#elif defined(HAVE_VFS_ITERATE)
.iterate = zpl_snapdir_iterate,
#else
.readdir = zpl_snapdir_readdir,
#endif
};
/*
* The '.zfs/snapshot' directory inode operations. These mainly control
* creating an inode for a snapshot directory and initializing the needed
* infrastructure to automount the snapshot. See zpl_snapdir_lookup().
*/
const struct inode_operations zpl_ops_snapdir = {
.lookup = zpl_snapdir_lookup,
.getattr = zpl_snapdir_getattr,
-#if defined(HAVE_RENAME_WANTS_FLAGS) || defined(HAVE_IOPS_RENAME_USERNS)
+#if (defined(HAVE_RENAME_WANTS_FLAGS) || \
+ defined(HAVE_IOPS_RENAME_USERNS) || \
+ defined(HAVE_IOPS_RENAME_IDMAP))
.rename = zpl_snapdir_rename2,
#else
.rename = zpl_snapdir_rename,
#endif
.rmdir = zpl_snapdir_rmdir,
.mkdir = zpl_snapdir_mkdir,
};
static struct dentry *
zpl_shares_lookup(struct inode *dip, struct dentry *dentry,
unsigned int flags)
{
fstrans_cookie_t cookie;
cred_t *cr = CRED();
struct inode *ip = NULL;
int error;
crhold(cr);
cookie = spl_fstrans_mark();
error = -zfsctl_shares_lookup(dip, dname(dentry), &ip,
0, cr, NULL, NULL);
ASSERT3S(error, <=, 0);
spl_fstrans_unmark(cookie);
crfree(cr);
if (error) {
if (error == -ENOENT)
return (d_splice_alias(NULL, dentry));
else
return (ERR_PTR(error));
}
return (d_splice_alias(ip, dentry));
}
static int
zpl_shares_iterate(struct file *filp, zpl_dir_context_t *ctx)
{
fstrans_cookie_t cookie;
cred_t *cr = CRED();
zfsvfs_t *zfsvfs = ITOZSB(file_inode(filp));
znode_t *dzp;
int error = 0;
if ((error = zpl_enter(zfsvfs, FTAG)) != 0)
return (error);
cookie = spl_fstrans_mark();
if (zfsvfs->z_shares_dir == 0) {
zpl_dir_emit_dots(filp, ctx);
goto out;
}
error = -zfs_zget(zfsvfs, zfsvfs->z_shares_dir, &dzp);
if (error)
goto out;
crhold(cr);
error = -zfs_readdir(ZTOI(dzp), ctx, cr);
crfree(cr);
iput(ZTOI(dzp));
out:
spl_fstrans_unmark(cookie);
zpl_exit(zfsvfs, FTAG);
ASSERT3S(error, <=, 0);
return (error);
}
#if !defined(HAVE_VFS_ITERATE) && !defined(HAVE_VFS_ITERATE_SHARED)
static int
zpl_shares_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
zpl_dir_context_t ctx =
ZPL_DIR_CONTEXT_INIT(dirent, filldir, filp->f_pos);
int error;
error = zpl_shares_iterate(filp, &ctx);
filp->f_pos = ctx.pos;
return (error);
}
#endif /* !HAVE_VFS_ITERATE && !HAVE_VFS_ITERATE_SHARED */
static int
#ifdef HAVE_USERNS_IOPS_GETATTR
zpl_shares_getattr_impl(struct user_namespace *user_ns,
const struct path *path, struct kstat *stat, u32 request_mask,
unsigned int query_flags)
+#elif defined(HAVE_IDMAP_IOPS_GETATTR)
+zpl_shares_getattr_impl(struct mnt_idmap *user_ns,
+ const struct path *path, struct kstat *stat, u32 request_mask,
+ unsigned int query_flags)
#else
zpl_shares_getattr_impl(const struct path *path, struct kstat *stat,
u32 request_mask, unsigned int query_flags)
#endif
{
(void) request_mask, (void) query_flags;
struct inode *ip = path->dentry->d_inode;
zfsvfs_t *zfsvfs = ITOZSB(ip);
znode_t *dzp;
int error;
if ((error = zpl_enter(zfsvfs, FTAG)) != 0)
return (error);
if (zfsvfs->z_shares_dir == 0) {
-#ifdef HAVE_USERNS_IOPS_GETATTR
+#if (defined(HAVE_USERNS_IOPS_GETATTR) || defined(HAVE_IDMAP_IOPS_GETATTR))
#ifdef HAVE_GENERIC_FILLATTR_USERNS
generic_fillattr(user_ns, path->dentry->d_inode, stat);
+#elif defined(HAVE_GENERIC_FILLATTR_IDMAP)
+ generic_fillattr(user_ns, path->dentry->d_inode, stat);
#else
(void) user_ns;
#endif
#else
generic_fillattr(path->dentry->d_inode, stat);
#endif
stat->nlink = stat->size = 2;
stat->atime = current_time(ip);
zpl_exit(zfsvfs, FTAG);
return (0);
}
error = -zfs_zget(zfsvfs, zfsvfs->z_shares_dir, &dzp);
if (error == 0) {
-#ifdef HAVE_USERNS_IOPS_GETATTR
-#ifdef HAVE_GENERIC_FILLATTR_USERNS
+#if (defined(HAVE_USERNS_IOPS_GETATTR) || defined(HAVE_IDMAP_IOPS_GETATTR))
error = -zfs_getattr_fast(user_ns, ZTOI(dzp), stat);
-#else
- (void) user_ns;
-#endif
#else
error = -zfs_getattr_fast(kcred->user_ns, ZTOI(dzp), stat);
#endif
iput(ZTOI(dzp));
}
zpl_exit(zfsvfs, FTAG);
ASSERT3S(error, <=, 0);
return (error);
}
ZPL_GETATTR_WRAPPER(zpl_shares_getattr);
/*
* The '.zfs/shares' directory file operations.
*/
const struct file_operations zpl_fops_shares = {
.open = zpl_common_open,
.llseek = generic_file_llseek,
.read = generic_read_dir,
#ifdef HAVE_VFS_ITERATE_SHARED
.iterate_shared = zpl_shares_iterate,
#elif defined(HAVE_VFS_ITERATE)
.iterate = zpl_shares_iterate,
#else
.readdir = zpl_shares_readdir,
#endif
};
/*
* The '.zfs/shares' directory inode operations.
*/
const struct inode_operations zpl_ops_shares = {
.lookup = zpl_shares_lookup,
.getattr = zpl_shares_getattr,
};
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c b/sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c
index 0a50f80ea68d..e690525d3cd4 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zpl_file.c
@@ -1,1343 +1,1365 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2011, Lawrence Livermore National Security, LLC.
* Copyright (c) 2015 by Chunwei Chen. All rights reserved.
*/
#ifdef CONFIG_COMPAT
#include <linux/compat.h>
#endif
#include <linux/fs.h>
#include <sys/file.h>
#include <sys/dmu_objset.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_vfsops.h>
#include <sys/zfs_vnops.h>
#include <sys/zfs_project.h>
#if defined(HAVE_VFS_SET_PAGE_DIRTY_NOBUFFERS) || \
defined(HAVE_VFS_FILEMAP_DIRTY_FOLIO)
#include <linux/pagemap.h>
#endif
#ifdef HAVE_FILE_FADVISE
#include <linux/fadvise.h>
#endif
#ifdef HAVE_VFS_FILEMAP_DIRTY_FOLIO
#include <linux/writeback.h>
#endif
/*
* When using fallocate(2) to preallocate space, inflate the requested
* capacity check by 10% to account for the required metadata blocks.
*/
static unsigned int zfs_fallocate_reserve_percent = 110;
static int
zpl_open(struct inode *ip, struct file *filp)
{
cred_t *cr = CRED();
int error;
fstrans_cookie_t cookie;
error = generic_file_open(ip, filp);
if (error)
return (error);
crhold(cr);
cookie = spl_fstrans_mark();
error = -zfs_open(ip, filp->f_mode, filp->f_flags, cr);
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
static int
zpl_release(struct inode *ip, struct file *filp)
{
cred_t *cr = CRED();
int error;
fstrans_cookie_t cookie;
cookie = spl_fstrans_mark();
if (ITOZ(ip)->z_atime_dirty)
zfs_mark_inode_dirty(ip);
crhold(cr);
error = -zfs_close(ip, filp->f_flags, cr);
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
static int
zpl_iterate(struct file *filp, zpl_dir_context_t *ctx)
{
cred_t *cr = CRED();
int error;
fstrans_cookie_t cookie;
crhold(cr);
cookie = spl_fstrans_mark();
error = -zfs_readdir(file_inode(filp), ctx, cr);
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
#if !defined(HAVE_VFS_ITERATE) && !defined(HAVE_VFS_ITERATE_SHARED)
static int
zpl_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
zpl_dir_context_t ctx =
ZPL_DIR_CONTEXT_INIT(dirent, filldir, filp->f_pos);
int error;
error = zpl_iterate(filp, &ctx);
filp->f_pos = ctx.pos;
return (error);
}
#endif /* !HAVE_VFS_ITERATE && !HAVE_VFS_ITERATE_SHARED */
#if defined(HAVE_FSYNC_WITHOUT_DENTRY)
/*
* Linux 2.6.35 - 3.0 API,
* As of 2.6.35 the dentry argument to the fops->fsync() hook was deemed
* redundant. The dentry is still accessible via filp->f_path.dentry,
* and we are guaranteed that filp will never be NULL.
*/
static int
zpl_fsync(struct file *filp, int datasync)
{
struct inode *inode = filp->f_mapping->host;
cred_t *cr = CRED();
int error;
fstrans_cookie_t cookie;
crhold(cr);
cookie = spl_fstrans_mark();
error = -zfs_fsync(ITOZ(inode), datasync, cr);
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
#ifdef HAVE_FILE_AIO_FSYNC
static int
zpl_aio_fsync(struct kiocb *kiocb, int datasync)
{
return (zpl_fsync(kiocb->ki_filp, datasync));
}
#endif
#elif defined(HAVE_FSYNC_RANGE)
/*
* Linux 3.1 API,
* As of 3.1 the responsibility to call filemap_write_and_wait_range() has
* been pushed down in to the .fsync() vfs hook. Additionally, the i_mutex
* lock is no longer held by the caller, for zfs we don't require the lock
* to be held so we don't acquire it.
*/
static int
zpl_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
{
struct inode *inode = filp->f_mapping->host;
znode_t *zp = ITOZ(inode);
zfsvfs_t *zfsvfs = ITOZSB(inode);
cred_t *cr = CRED();
int error;
fstrans_cookie_t cookie;
/*
* The variables z_sync_writes_cnt and z_async_writes_cnt work in
* tandem so that sync writes can detect if there are any non-sync
* writes going on and vice-versa. The "vice-versa" part to this logic
* is located in zfs_putpage() where non-sync writes check if there are
* any ongoing sync writes. If any sync and non-sync writes overlap,
* we do a commit to complete the non-sync writes since the latter can
* potentially take several seconds to complete and thus block sync
* writes in the upcoming call to filemap_write_and_wait_range().
*/
atomic_inc_32(&zp->z_sync_writes_cnt);
/*
* If the following check does not detect an overlapping non-sync write
* (say because it's just about to start), then it is guaranteed that
* the non-sync write will detect this sync write. This is because we
* always increment z_sync_writes_cnt / z_async_writes_cnt before doing
* the check on z_async_writes_cnt / z_sync_writes_cnt here and in
* zfs_putpage() respectively.
*/
if (atomic_load_32(&zp->z_async_writes_cnt) > 0) {
if ((error = zpl_enter(zfsvfs, FTAG)) != 0) {
atomic_dec_32(&zp->z_sync_writes_cnt);
return (error);
}
zil_commit(zfsvfs->z_log, zp->z_id);
zpl_exit(zfsvfs, FTAG);
}
error = filemap_write_and_wait_range(inode->i_mapping, start, end);
/*
* The sync write is not complete yet but we decrement
* z_sync_writes_cnt since zfs_fsync() increments and decrements
* it internally. If a non-sync write starts just after the decrement
* operation but before we call zfs_fsync(), it may not detect this
* overlapping sync write but it does not matter since we have already
* gone past filemap_write_and_wait_range() and we won't block due to
* the non-sync write.
*/
atomic_dec_32(&zp->z_sync_writes_cnt);
if (error)
return (error);
crhold(cr);
cookie = spl_fstrans_mark();
error = -zfs_fsync(zp, datasync, cr);
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
#ifdef HAVE_FILE_AIO_FSYNC
static int
zpl_aio_fsync(struct kiocb *kiocb, int datasync)
{
return (zpl_fsync(kiocb->ki_filp, kiocb->ki_pos, -1, datasync));
}
#endif
#else
#error "Unsupported fops->fsync() implementation"
#endif
static inline int
zfs_io_flags(struct kiocb *kiocb)
{
int flags = 0;
#if defined(IOCB_DSYNC)
if (kiocb->ki_flags & IOCB_DSYNC)
flags |= O_DSYNC;
#endif
#if defined(IOCB_SYNC)
if (kiocb->ki_flags & IOCB_SYNC)
flags |= O_SYNC;
#endif
#if defined(IOCB_APPEND)
if (kiocb->ki_flags & IOCB_APPEND)
flags |= O_APPEND;
#endif
#if defined(IOCB_DIRECT)
if (kiocb->ki_flags & IOCB_DIRECT)
flags |= O_DIRECT;
#endif
return (flags);
}
/*
* If relatime is enabled, call file_accessed() if zfs_relatime_need_update()
* is true. This is needed since datasets with inherited "relatime" property
* aren't necessarily mounted with the MNT_RELATIME flag (e.g. after
* `zfs set relatime=...`), which is what relatime test in VFS by
* relatime_need_update() is based on.
*/
static inline void
zpl_file_accessed(struct file *filp)
{
struct inode *ip = filp->f_mapping->host;
if (!IS_NOATIME(ip) && ITOZSB(ip)->z_relatime) {
if (zfs_relatime_need_update(ip))
file_accessed(filp);
} else {
file_accessed(filp);
}
}
#if defined(HAVE_VFS_RW_ITERATE)
/*
* When HAVE_VFS_IOV_ITER is defined the iov_iter structure supports
* iovecs, kvevs, bvecs and pipes, plus all the required interfaces to
* manipulate the iov_iter are available. In which case the full iov_iter
* can be attached to the uio and correctly handled in the lower layers.
* Otherwise, for older kernels extract the iovec and pass it instead.
*/
static void
zpl_uio_init(zfs_uio_t *uio, struct kiocb *kiocb, struct iov_iter *to,
loff_t pos, ssize_t count, size_t skip)
{
#if defined(HAVE_VFS_IOV_ITER)
zfs_uio_iov_iter_init(uio, to, pos, count, skip);
#else
#ifdef HAVE_IOV_ITER_TYPE
zfs_uio_iovec_init(uio, to->iov, to->nr_segs, pos,
iov_iter_type(to) & ITER_KVEC ? UIO_SYSSPACE : UIO_USERSPACE,
count, skip);
#else
zfs_uio_iovec_init(uio, to->iov, to->nr_segs, pos,
to->type & ITER_KVEC ? UIO_SYSSPACE : UIO_USERSPACE,
count, skip);
#endif
#endif
}
static ssize_t
zpl_iter_read(struct kiocb *kiocb, struct iov_iter *to)
{
cred_t *cr = CRED();
fstrans_cookie_t cookie;
struct file *filp = kiocb->ki_filp;
ssize_t count = iov_iter_count(to);
zfs_uio_t uio;
zpl_uio_init(&uio, kiocb, to, kiocb->ki_pos, count, 0);
crhold(cr);
cookie = spl_fstrans_mark();
int error = -zfs_read(ITOZ(filp->f_mapping->host), &uio,
filp->f_flags | zfs_io_flags(kiocb), cr);
spl_fstrans_unmark(cookie);
crfree(cr);
if (error < 0)
return (error);
ssize_t read = count - uio.uio_resid;
kiocb->ki_pos += read;
zpl_file_accessed(filp);
return (read);
}
static inline ssize_t
zpl_generic_write_checks(struct kiocb *kiocb, struct iov_iter *from,
size_t *countp)
{
#ifdef HAVE_GENERIC_WRITE_CHECKS_KIOCB
ssize_t ret = generic_write_checks(kiocb, from);
if (ret <= 0)
return (ret);
*countp = ret;
#else
struct file *file = kiocb->ki_filp;
struct address_space *mapping = file->f_mapping;
struct inode *ip = mapping->host;
int isblk = S_ISBLK(ip->i_mode);
*countp = iov_iter_count(from);
ssize_t ret = generic_write_checks(file, &kiocb->ki_pos, countp, isblk);
if (ret)
return (ret);
#endif
return (0);
}
static ssize_t
zpl_iter_write(struct kiocb *kiocb, struct iov_iter *from)
{
cred_t *cr = CRED();
fstrans_cookie_t cookie;
struct file *filp = kiocb->ki_filp;
struct inode *ip = filp->f_mapping->host;
zfs_uio_t uio;
size_t count = 0;
ssize_t ret;
ret = zpl_generic_write_checks(kiocb, from, &count);
if (ret)
return (ret);
zpl_uio_init(&uio, kiocb, from, kiocb->ki_pos, count, from->iov_offset);
crhold(cr);
cookie = spl_fstrans_mark();
int error = -zfs_write(ITOZ(ip), &uio,
filp->f_flags | zfs_io_flags(kiocb), cr);
spl_fstrans_unmark(cookie);
crfree(cr);
if (error < 0)
return (error);
ssize_t wrote = count - uio.uio_resid;
kiocb->ki_pos += wrote;
return (wrote);
}
#else /* !HAVE_VFS_RW_ITERATE */
static ssize_t
zpl_aio_read(struct kiocb *kiocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos)
{
cred_t *cr = CRED();
fstrans_cookie_t cookie;
struct file *filp = kiocb->ki_filp;
size_t count;
ssize_t ret;
ret = generic_segment_checks(iov, &nr_segs, &count, VERIFY_WRITE);
if (ret)
return (ret);
zfs_uio_t uio;
zfs_uio_iovec_init(&uio, iov, nr_segs, kiocb->ki_pos, UIO_USERSPACE,
count, 0);
crhold(cr);
cookie = spl_fstrans_mark();
int error = -zfs_read(ITOZ(filp->f_mapping->host), &uio,
filp->f_flags | zfs_io_flags(kiocb), cr);
spl_fstrans_unmark(cookie);
crfree(cr);
if (error < 0)
return (error);
ssize_t read = count - uio.uio_resid;
kiocb->ki_pos += read;
zpl_file_accessed(filp);
return (read);
}
static ssize_t
zpl_aio_write(struct kiocb *kiocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos)
{
cred_t *cr = CRED();
fstrans_cookie_t cookie;
struct file *filp = kiocb->ki_filp;
struct inode *ip = filp->f_mapping->host;
size_t count;
ssize_t ret;
ret = generic_segment_checks(iov, &nr_segs, &count, VERIFY_READ);
if (ret)
return (ret);
ret = generic_write_checks(filp, &pos, &count, S_ISBLK(ip->i_mode));
if (ret)
return (ret);
kiocb->ki_pos = pos;
zfs_uio_t uio;
zfs_uio_iovec_init(&uio, iov, nr_segs, kiocb->ki_pos, UIO_USERSPACE,
count, 0);
crhold(cr);
cookie = spl_fstrans_mark();
int error = -zfs_write(ITOZ(ip), &uio,
filp->f_flags | zfs_io_flags(kiocb), cr);
spl_fstrans_unmark(cookie);
crfree(cr);
if (error < 0)
return (error);
ssize_t wrote = count - uio.uio_resid;
kiocb->ki_pos += wrote;
return (wrote);
}
#endif /* HAVE_VFS_RW_ITERATE */
#if defined(HAVE_VFS_RW_ITERATE)
static ssize_t
zpl_direct_IO_impl(int rw, struct kiocb *kiocb, struct iov_iter *iter)
{
if (rw == WRITE)
return (zpl_iter_write(kiocb, iter));
else
return (zpl_iter_read(kiocb, iter));
}
#if defined(HAVE_VFS_DIRECT_IO_ITER)
static ssize_t
zpl_direct_IO(struct kiocb *kiocb, struct iov_iter *iter)
{
return (zpl_direct_IO_impl(iov_iter_rw(iter), kiocb, iter));
}
#elif defined(HAVE_VFS_DIRECT_IO_ITER_OFFSET)
static ssize_t
zpl_direct_IO(struct kiocb *kiocb, struct iov_iter *iter, loff_t pos)
{
ASSERT3S(pos, ==, kiocb->ki_pos);
return (zpl_direct_IO_impl(iov_iter_rw(iter), kiocb, iter));
}
#elif defined(HAVE_VFS_DIRECT_IO_ITER_RW_OFFSET)
static ssize_t
zpl_direct_IO(int rw, struct kiocb *kiocb, struct iov_iter *iter, loff_t pos)
{
ASSERT3S(pos, ==, kiocb->ki_pos);
return (zpl_direct_IO_impl(rw, kiocb, iter));
}
#else
#error "Unknown direct IO interface"
#endif
#else /* HAVE_VFS_RW_ITERATE */
#if defined(HAVE_VFS_DIRECT_IO_IOVEC)
static ssize_t
zpl_direct_IO(int rw, struct kiocb *kiocb, const struct iovec *iov,
loff_t pos, unsigned long nr_segs)
{
if (rw == WRITE)
return (zpl_aio_write(kiocb, iov, nr_segs, pos));
else
return (zpl_aio_read(kiocb, iov, nr_segs, pos));
}
#elif defined(HAVE_VFS_DIRECT_IO_ITER_RW_OFFSET)
static ssize_t
zpl_direct_IO(int rw, struct kiocb *kiocb, struct iov_iter *iter, loff_t pos)
{
const struct iovec *iovp = iov_iter_iovec(iter);
unsigned long nr_segs = iter->nr_segs;
ASSERT3S(pos, ==, kiocb->ki_pos);
if (rw == WRITE)
return (zpl_aio_write(kiocb, iovp, nr_segs, pos));
else
return (zpl_aio_read(kiocb, iovp, nr_segs, pos));
}
#else
#error "Unknown direct IO interface"
#endif
#endif /* HAVE_VFS_RW_ITERATE */
static loff_t
zpl_llseek(struct file *filp, loff_t offset, int whence)
{
#if defined(SEEK_HOLE) && defined(SEEK_DATA)
fstrans_cookie_t cookie;
if (whence == SEEK_DATA || whence == SEEK_HOLE) {
struct inode *ip = filp->f_mapping->host;
loff_t maxbytes = ip->i_sb->s_maxbytes;
loff_t error;
spl_inode_lock_shared(ip);
cookie = spl_fstrans_mark();
error = -zfs_holey(ITOZ(ip), whence, &offset);
spl_fstrans_unmark(cookie);
if (error == 0)
error = lseek_execute(filp, ip, offset, maxbytes);
spl_inode_unlock_shared(ip);
return (error);
}
#endif /* SEEK_HOLE && SEEK_DATA */
return (generic_file_llseek(filp, offset, whence));
}
/*
* It's worth taking a moment to describe how mmap is implemented
* for zfs because it differs considerably from other Linux filesystems.
* However, this issue is handled the same way under OpenSolaris.
*
* The issue is that by design zfs bypasses the Linux page cache and
* leaves all caching up to the ARC. This has been shown to work
* well for the common read(2)/write(2) case. However, mmap(2)
* is problem because it relies on being tightly integrated with the
* page cache. To handle this we cache mmap'ed files twice, once in
* the ARC and a second time in the page cache. The code is careful
* to keep both copies synchronized.
*
* When a file with an mmap'ed region is written to using write(2)
* both the data in the ARC and existing pages in the page cache
* are updated. For a read(2) data will be read first from the page
* cache then the ARC if needed. Neither a write(2) or read(2) will
* will ever result in new pages being added to the page cache.
*
* New pages are added to the page cache only via .readpage() which
* is called when the vfs needs to read a page off disk to back the
* virtual memory region. These pages may be modified without
* notifying the ARC and will be written out periodically via
* .writepage(). This will occur due to either a sync or the usual
* page aging behavior. Note because a read(2) of a mmap'ed file
* will always check the page cache first even when the ARC is out
* of date correct data will still be returned.
*
* While this implementation ensures correct behavior it does have
* have some drawbacks. The most obvious of which is that it
* increases the required memory footprint when access mmap'ed
* files. It also adds additional complexity to the code keeping
* both caches synchronized.
*
* Longer term it may be possible to cleanly resolve this wart by
* mapping page cache pages directly on to the ARC buffers. The
* Linux address space operations are flexible enough to allow
* selection of which pages back a particular index. The trick
* would be working out the details of which subsystem is in
* charge, the ARC, the page cache, or both. It may also prove
* helpful to move the ARC buffers to a scatter-gather lists
* rather than a vmalloc'ed region.
*/
static int
zpl_mmap(struct file *filp, struct vm_area_struct *vma)
{
struct inode *ip = filp->f_mapping->host;
int error;
fstrans_cookie_t cookie;
cookie = spl_fstrans_mark();
error = -zfs_map(ip, vma->vm_pgoff, (caddr_t *)vma->vm_start,
(size_t)(vma->vm_end - vma->vm_start), vma->vm_flags);
spl_fstrans_unmark(cookie);
if (error)
return (error);
error = generic_file_mmap(filp, vma);
if (error)
return (error);
#if !defined(HAVE_FILEMAP_RANGE_HAS_PAGE)
znode_t *zp = ITOZ(ip);
mutex_enter(&zp->z_lock);
zp->z_is_mapped = B_TRUE;
mutex_exit(&zp->z_lock);
#endif
return (error);
}
/*
* Populate a page with data for the Linux page cache. This function is
* only used to support mmap(2). There will be an identical copy of the
* data in the ARC which is kept up to date via .write() and .writepage().
*/
static inline int
zpl_readpage_common(struct page *pp)
{
fstrans_cookie_t cookie;
ASSERT(PageLocked(pp));
cookie = spl_fstrans_mark();
int error = -zfs_getpage(pp->mapping->host, pp);
spl_fstrans_unmark(cookie);
unlock_page(pp);
return (error);
}
#ifdef HAVE_VFS_READ_FOLIO
static int
zpl_read_folio(struct file *filp, struct folio *folio)
{
return (zpl_readpage_common(&folio->page));
}
#else
static int
zpl_readpage(struct file *filp, struct page *pp)
{
return (zpl_readpage_common(pp));
}
#endif
static int
zpl_readpage_filler(void *data, struct page *pp)
{
return (zpl_readpage_common(pp));
}
/*
* Populate a set of pages with data for the Linux page cache. This
* function will only be called for read ahead and never for demand
* paging. For simplicity, the code relies on read_cache_pages() to
* correctly lock each page for IO and call zpl_readpage().
*/
#ifdef HAVE_VFS_READPAGES
static int
zpl_readpages(struct file *filp, struct address_space *mapping,
struct list_head *pages, unsigned nr_pages)
{
return (read_cache_pages(mapping, pages, zpl_readpage_filler, NULL));
}
#else
static void
zpl_readahead(struct readahead_control *ractl)
{
struct page *page;
while ((page = readahead_page(ractl)) != NULL) {
int ret;
ret = zpl_readpage_filler(NULL, page);
put_page(page);
if (ret)
break;
}
}
#endif
static int
zpl_putpage(struct page *pp, struct writeback_control *wbc, void *data)
{
boolean_t *for_sync = data;
fstrans_cookie_t cookie;
ASSERT(PageLocked(pp));
ASSERT(!PageWriteback(pp));
cookie = spl_fstrans_mark();
(void) zfs_putpage(pp->mapping->host, pp, wbc, *for_sync);
spl_fstrans_unmark(cookie);
return (0);
}
+#ifdef HAVE_WRITEPAGE_T_FOLIO
+static int
+zpl_putfolio(struct folio *pp, struct writeback_control *wbc, void *data)
+{
+ (void) zpl_putpage(&pp->page, wbc, data);
+ return (0);
+}
+#endif
+
+static inline int
+zpl_write_cache_pages(struct address_space *mapping,
+ struct writeback_control *wbc, void *data)
+{
+ int result;
+
+#ifdef HAVE_WRITEPAGE_T_FOLIO
+ result = write_cache_pages(mapping, wbc, zpl_putfolio, data);
+#else
+ result = write_cache_pages(mapping, wbc, zpl_putpage, data);
+#endif
+ return (result);
+}
+
static int
zpl_writepages(struct address_space *mapping, struct writeback_control *wbc)
{
znode_t *zp = ITOZ(mapping->host);
zfsvfs_t *zfsvfs = ITOZSB(mapping->host);
enum writeback_sync_modes sync_mode;
int result;
if ((result = zpl_enter(zfsvfs, FTAG)) != 0)
return (result);
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
wbc->sync_mode = WB_SYNC_ALL;
zpl_exit(zfsvfs, FTAG);
sync_mode = wbc->sync_mode;
/*
* We don't want to run write_cache_pages() in SYNC mode here, because
* that would make putpage() wait for a single page to be committed to
* disk every single time, resulting in atrocious performance. Instead
* we run it once in non-SYNC mode so that the ZIL gets all the data,
* and then we commit it all in one go.
*/
boolean_t for_sync = (sync_mode == WB_SYNC_ALL);
wbc->sync_mode = WB_SYNC_NONE;
- result = write_cache_pages(mapping, wbc, zpl_putpage, &for_sync);
+ result = zpl_write_cache_pages(mapping, wbc, &for_sync);
if (sync_mode != wbc->sync_mode) {
if ((result = zpl_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (result);
if (zfsvfs->z_log != NULL)
zil_commit(zfsvfs->z_log, zp->z_id);
zpl_exit(zfsvfs, FTAG);
/*
* We need to call write_cache_pages() again (we can't just
* return after the commit) because the previous call in
* non-SYNC mode does not guarantee that we got all the dirty
* pages (see the implementation of write_cache_pages() for
* details). That being said, this is a no-op in most cases.
*/
wbc->sync_mode = sync_mode;
- result = write_cache_pages(mapping, wbc, zpl_putpage,
- &for_sync);
+ result = zpl_write_cache_pages(mapping, wbc, &for_sync);
}
return (result);
}
/*
* Write out dirty pages to the ARC, this function is only required to
* support mmap(2). Mapped pages may be dirtied by memory operations
* which never call .write(). These dirty pages are kept in sync with
* the ARC buffers via this hook.
*/
static int
zpl_writepage(struct page *pp, struct writeback_control *wbc)
{
if (ITOZSB(pp->mapping->host)->z_os->os_sync == ZFS_SYNC_ALWAYS)
wbc->sync_mode = WB_SYNC_ALL;
boolean_t for_sync = (wbc->sync_mode == WB_SYNC_ALL);
return (zpl_putpage(pp, wbc, &for_sync));
}
/*
* The flag combination which matches the behavior of zfs_space() is
* FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE. The FALLOC_FL_PUNCH_HOLE
* flag was introduced in the 2.6.38 kernel.
*
* The original mode=0 (allocate space) behavior can be reasonably emulated
* by checking if enough space exists and creating a sparse file, as real
* persistent space reservation is not possible due to COW, snapshots, etc.
*/
static long
zpl_fallocate_common(struct inode *ip, int mode, loff_t offset, loff_t len)
{
cred_t *cr = CRED();
loff_t olen;
fstrans_cookie_t cookie;
int error = 0;
int test_mode = FALLOC_FL_PUNCH_HOLE;
#ifdef HAVE_FALLOC_FL_ZERO_RANGE
test_mode |= FALLOC_FL_ZERO_RANGE;
#endif
if ((mode & ~(FALLOC_FL_KEEP_SIZE | test_mode)) != 0)
return (-EOPNOTSUPP);
if (offset < 0 || len <= 0)
return (-EINVAL);
spl_inode_lock(ip);
olen = i_size_read(ip);
crhold(cr);
cookie = spl_fstrans_mark();
if (mode & (test_mode)) {
flock64_t bf;
if (mode & FALLOC_FL_KEEP_SIZE) {
if (offset > olen)
goto out_unmark;
if (offset + len > olen)
len = olen - offset;
}
bf.l_type = F_WRLCK;
bf.l_whence = SEEK_SET;
bf.l_start = offset;
bf.l_len = len;
bf.l_pid = 0;
error = -zfs_space(ITOZ(ip), F_FREESP, &bf, O_RDWR, offset, cr);
} else if ((mode & ~FALLOC_FL_KEEP_SIZE) == 0) {
unsigned int percent = zfs_fallocate_reserve_percent;
struct kstatfs statfs;
/* Legacy mode, disable fallocate compatibility. */
if (percent == 0) {
error = -EOPNOTSUPP;
goto out_unmark;
}
/*
* Use zfs_statvfs() instead of dmu_objset_space() since it
* also checks project quota limits, which are relevant here.
*/
error = zfs_statvfs(ip, &statfs);
if (error)
goto out_unmark;
/*
* Shrink available space a bit to account for overhead/races.
* We know the product previously fit into availbytes from
* dmu_objset_space(), so the smaller product will also fit.
*/
if (len > statfs.f_bavail * (statfs.f_bsize * 100 / percent)) {
error = -ENOSPC;
goto out_unmark;
}
if (!(mode & FALLOC_FL_KEEP_SIZE) && offset + len > olen)
error = zfs_freesp(ITOZ(ip), offset + len, 0, 0, FALSE);
}
out_unmark:
spl_fstrans_unmark(cookie);
spl_inode_unlock(ip);
crfree(cr);
return (error);
}
static long
zpl_fallocate(struct file *filp, int mode, loff_t offset, loff_t len)
{
return zpl_fallocate_common(file_inode(filp),
mode, offset, len);
}
static int
zpl_ioctl_getversion(struct file *filp, void __user *arg)
{
uint32_t generation = file_inode(filp)->i_generation;
return (copy_to_user(arg, &generation, sizeof (generation)));
}
#ifdef HAVE_FILE_FADVISE
static int
zpl_fadvise(struct file *filp, loff_t offset, loff_t len, int advice)
{
struct inode *ip = file_inode(filp);
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ITOZSB(ip);
objset_t *os = zfsvfs->z_os;
int error = 0;
if (S_ISFIFO(ip->i_mode))
return (-ESPIPE);
if (offset < 0 || len < 0)
return (-EINVAL);
if ((error = zpl_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
switch (advice) {
case POSIX_FADV_SEQUENTIAL:
case POSIX_FADV_WILLNEED:
#ifdef HAVE_GENERIC_FADVISE
if (zn_has_cached_data(zp, offset, offset + len - 1))
error = generic_fadvise(filp, offset, len, advice);
#endif
/*
* Pass on the caller's size directly, but note that
* dmu_prefetch_max will effectively cap it. If there
* really is a larger sequential access pattern, perhaps
* dmu_zfetch will detect it.
*/
if (len == 0)
len = i_size_read(ip) - offset;
dmu_prefetch(os, zp->z_id, 0, offset, len,
ZIO_PRIORITY_ASYNC_READ);
break;
case POSIX_FADV_NORMAL:
case POSIX_FADV_RANDOM:
case POSIX_FADV_DONTNEED:
case POSIX_FADV_NOREUSE:
/* ignored for now */
break;
default:
error = -EINVAL;
break;
}
zfs_exit(zfsvfs, FTAG);
return (error);
}
#endif /* HAVE_FILE_FADVISE */
#define ZFS_FL_USER_VISIBLE (FS_FL_USER_VISIBLE | ZFS_PROJINHERIT_FL)
#define ZFS_FL_USER_MODIFIABLE (FS_FL_USER_MODIFIABLE | ZFS_PROJINHERIT_FL)
static uint32_t
__zpl_ioctl_getflags(struct inode *ip)
{
uint64_t zfs_flags = ITOZ(ip)->z_pflags;
uint32_t ioctl_flags = 0;
if (zfs_flags & ZFS_IMMUTABLE)
ioctl_flags |= FS_IMMUTABLE_FL;
if (zfs_flags & ZFS_APPENDONLY)
ioctl_flags |= FS_APPEND_FL;
if (zfs_flags & ZFS_NODUMP)
ioctl_flags |= FS_NODUMP_FL;
if (zfs_flags & ZFS_PROJINHERIT)
ioctl_flags |= ZFS_PROJINHERIT_FL;
return (ioctl_flags & ZFS_FL_USER_VISIBLE);
}
/*
* Map zfs file z_pflags (xvattr_t) to linux file attributes. Only file
* attributes common to both Linux and Solaris are mapped.
*/
static int
zpl_ioctl_getflags(struct file *filp, void __user *arg)
{
uint32_t flags;
int err;
flags = __zpl_ioctl_getflags(file_inode(filp));
err = copy_to_user(arg, &flags, sizeof (flags));
return (err);
}
/*
* fchange() is a helper macro to detect if we have been asked to change a
* flag. This is ugly, but the requirement that we do this is a consequence of
* how the Linux file attribute interface was designed. Another consequence is
* that concurrent modification of files suffers from a TOCTOU race. Neither
* are things we can fix without modifying the kernel-userland interface, which
* is outside of our jurisdiction.
*/
#define fchange(f0, f1, b0, b1) (!((f0) & (b0)) != !((f1) & (b1)))
static int
__zpl_ioctl_setflags(struct inode *ip, uint32_t ioctl_flags, xvattr_t *xva)
{
uint64_t zfs_flags = ITOZ(ip)->z_pflags;
xoptattr_t *xoap;
if (ioctl_flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | FS_NODUMP_FL |
ZFS_PROJINHERIT_FL))
return (-EOPNOTSUPP);
if (ioctl_flags & ~ZFS_FL_USER_MODIFIABLE)
return (-EACCES);
if ((fchange(ioctl_flags, zfs_flags, FS_IMMUTABLE_FL, ZFS_IMMUTABLE) ||
fchange(ioctl_flags, zfs_flags, FS_APPEND_FL, ZFS_APPENDONLY)) &&
!capable(CAP_LINUX_IMMUTABLE))
return (-EPERM);
- if (!zpl_inode_owner_or_capable(kcred->user_ns, ip))
+ if (!zpl_inode_owner_or_capable(zfs_init_idmap, ip))
return (-EACCES);
xva_init(xva);
xoap = xva_getxoptattr(xva);
#define FLAG_CHANGE(iflag, zflag, xflag, xfield) do { \
if (((ioctl_flags & (iflag)) && !(zfs_flags & (zflag))) || \
((zfs_flags & (zflag)) && !(ioctl_flags & (iflag)))) { \
XVA_SET_REQ(xva, (xflag)); \
(xfield) = ((ioctl_flags & (iflag)) != 0); \
} \
} while (0)
FLAG_CHANGE(FS_IMMUTABLE_FL, ZFS_IMMUTABLE, XAT_IMMUTABLE,
xoap->xoa_immutable);
FLAG_CHANGE(FS_APPEND_FL, ZFS_APPENDONLY, XAT_APPENDONLY,
xoap->xoa_appendonly);
FLAG_CHANGE(FS_NODUMP_FL, ZFS_NODUMP, XAT_NODUMP,
xoap->xoa_nodump);
FLAG_CHANGE(ZFS_PROJINHERIT_FL, ZFS_PROJINHERIT, XAT_PROJINHERIT,
xoap->xoa_projinherit);
#undef FLAG_CHANGE
return (0);
}
static int
zpl_ioctl_setflags(struct file *filp, void __user *arg)
{
struct inode *ip = file_inode(filp);
uint32_t flags;
cred_t *cr = CRED();
xvattr_t xva;
int err;
fstrans_cookie_t cookie;
if (copy_from_user(&flags, arg, sizeof (flags)))
return (-EFAULT);
err = __zpl_ioctl_setflags(ip, flags, &xva);
if (err)
return (err);
crhold(cr);
cookie = spl_fstrans_mark();
- err = -zfs_setattr(ITOZ(ip), (vattr_t *)&xva, 0, cr, kcred->user_ns);
+ err = -zfs_setattr(ITOZ(ip), (vattr_t *)&xva, 0, cr, zfs_init_idmap);
spl_fstrans_unmark(cookie);
crfree(cr);
return (err);
}
static int
zpl_ioctl_getxattr(struct file *filp, void __user *arg)
{
zfsxattr_t fsx = { 0 };
struct inode *ip = file_inode(filp);
int err;
fsx.fsx_xflags = __zpl_ioctl_getflags(ip);
fsx.fsx_projid = ITOZ(ip)->z_projid;
err = copy_to_user(arg, &fsx, sizeof (fsx));
return (err);
}
static int
zpl_ioctl_setxattr(struct file *filp, void __user *arg)
{
struct inode *ip = file_inode(filp);
zfsxattr_t fsx;
cred_t *cr = CRED();
xvattr_t xva;
xoptattr_t *xoap;
int err;
fstrans_cookie_t cookie;
if (copy_from_user(&fsx, arg, sizeof (fsx)))
return (-EFAULT);
if (!zpl_is_valid_projid(fsx.fsx_projid))
return (-EINVAL);
err = __zpl_ioctl_setflags(ip, fsx.fsx_xflags, &xva);
if (err)
return (err);
xoap = xva_getxoptattr(&xva);
XVA_SET_REQ(&xva, XAT_PROJID);
xoap->xoa_projid = fsx.fsx_projid;
crhold(cr);
cookie = spl_fstrans_mark();
- err = -zfs_setattr(ITOZ(ip), (vattr_t *)&xva, 0, cr, kcred->user_ns);
+ err = -zfs_setattr(ITOZ(ip), (vattr_t *)&xva, 0, cr, zfs_init_idmap);
spl_fstrans_unmark(cookie);
crfree(cr);
return (err);
}
/*
* Expose Additional File Level Attributes of ZFS.
*/
static int
zpl_ioctl_getdosflags(struct file *filp, void __user *arg)
{
struct inode *ip = file_inode(filp);
uint64_t dosflags = ITOZ(ip)->z_pflags;
dosflags &= ZFS_DOS_FL_USER_VISIBLE;
int err = copy_to_user(arg, &dosflags, sizeof (dosflags));
return (err);
}
static int
__zpl_ioctl_setdosflags(struct inode *ip, uint64_t ioctl_flags, xvattr_t *xva)
{
uint64_t zfs_flags = ITOZ(ip)->z_pflags;
xoptattr_t *xoap;
if (ioctl_flags & (~ZFS_DOS_FL_USER_VISIBLE))
return (-EOPNOTSUPP);
if ((fchange(ioctl_flags, zfs_flags, ZFS_IMMUTABLE, ZFS_IMMUTABLE) ||
fchange(ioctl_flags, zfs_flags, ZFS_APPENDONLY, ZFS_APPENDONLY)) &&
!capable(CAP_LINUX_IMMUTABLE))
return (-EPERM);
- if (!zpl_inode_owner_or_capable(kcred->user_ns, ip))
+ if (!zpl_inode_owner_or_capable(zfs_init_idmap, ip))
return (-EACCES);
xva_init(xva);
xoap = xva_getxoptattr(xva);
#define FLAG_CHANGE(iflag, xflag, xfield) do { \
if (((ioctl_flags & (iflag)) && !(zfs_flags & (iflag))) || \
((zfs_flags & (iflag)) && !(ioctl_flags & (iflag)))) { \
XVA_SET_REQ(xva, (xflag)); \
(xfield) = ((ioctl_flags & (iflag)) != 0); \
} \
} while (0)
FLAG_CHANGE(ZFS_IMMUTABLE, XAT_IMMUTABLE, xoap->xoa_immutable);
FLAG_CHANGE(ZFS_APPENDONLY, XAT_APPENDONLY, xoap->xoa_appendonly);
FLAG_CHANGE(ZFS_NODUMP, XAT_NODUMP, xoap->xoa_nodump);
FLAG_CHANGE(ZFS_READONLY, XAT_READONLY, xoap->xoa_readonly);
FLAG_CHANGE(ZFS_HIDDEN, XAT_HIDDEN, xoap->xoa_hidden);
FLAG_CHANGE(ZFS_SYSTEM, XAT_SYSTEM, xoap->xoa_system);
FLAG_CHANGE(ZFS_ARCHIVE, XAT_ARCHIVE, xoap->xoa_archive);
FLAG_CHANGE(ZFS_NOUNLINK, XAT_NOUNLINK, xoap->xoa_nounlink);
FLAG_CHANGE(ZFS_REPARSE, XAT_REPARSE, xoap->xoa_reparse);
FLAG_CHANGE(ZFS_OFFLINE, XAT_OFFLINE, xoap->xoa_offline);
FLAG_CHANGE(ZFS_SPARSE, XAT_SPARSE, xoap->xoa_sparse);
#undef FLAG_CHANGE
return (0);
}
/*
* Set Additional File Level Attributes of ZFS.
*/
static int
zpl_ioctl_setdosflags(struct file *filp, void __user *arg)
{
struct inode *ip = file_inode(filp);
uint64_t dosflags;
cred_t *cr = CRED();
xvattr_t xva;
int err;
fstrans_cookie_t cookie;
if (copy_from_user(&dosflags, arg, sizeof (dosflags)))
return (-EFAULT);
err = __zpl_ioctl_setdosflags(ip, dosflags, &xva);
if (err)
return (err);
crhold(cr);
cookie = spl_fstrans_mark();
- err = -zfs_setattr(ITOZ(ip), (vattr_t *)&xva, 0, cr, kcred->user_ns);
+ err = -zfs_setattr(ITOZ(ip), (vattr_t *)&xva, 0, cr, zfs_init_idmap);
spl_fstrans_unmark(cookie);
crfree(cr);
return (err);
}
static long
zpl_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
switch (cmd) {
case FS_IOC_GETVERSION:
return (zpl_ioctl_getversion(filp, (void *)arg));
case FS_IOC_GETFLAGS:
return (zpl_ioctl_getflags(filp, (void *)arg));
case FS_IOC_SETFLAGS:
return (zpl_ioctl_setflags(filp, (void *)arg));
case ZFS_IOC_FSGETXATTR:
return (zpl_ioctl_getxattr(filp, (void *)arg));
case ZFS_IOC_FSSETXATTR:
return (zpl_ioctl_setxattr(filp, (void *)arg));
case ZFS_IOC_GETDOSFLAGS:
return (zpl_ioctl_getdosflags(filp, (void *)arg));
case ZFS_IOC_SETDOSFLAGS:
return (zpl_ioctl_setdosflags(filp, (void *)arg));
default:
return (-ENOTTY);
}
}
#ifdef CONFIG_COMPAT
static long
zpl_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
switch (cmd) {
case FS_IOC32_GETVERSION:
cmd = FS_IOC_GETVERSION;
break;
case FS_IOC32_GETFLAGS:
cmd = FS_IOC_GETFLAGS;
break;
case FS_IOC32_SETFLAGS:
cmd = FS_IOC_SETFLAGS;
break;
default:
return (-ENOTTY);
}
return (zpl_ioctl(filp, cmd, (unsigned long)compat_ptr(arg)));
}
#endif /* CONFIG_COMPAT */
const struct address_space_operations zpl_address_space_operations = {
#ifdef HAVE_VFS_READPAGES
.readpages = zpl_readpages,
#else
.readahead = zpl_readahead,
#endif
#ifdef HAVE_VFS_READ_FOLIO
.read_folio = zpl_read_folio,
#else
.readpage = zpl_readpage,
#endif
.writepage = zpl_writepage,
.writepages = zpl_writepages,
.direct_IO = zpl_direct_IO,
#ifdef HAVE_VFS_SET_PAGE_DIRTY_NOBUFFERS
.set_page_dirty = __set_page_dirty_nobuffers,
#endif
#ifdef HAVE_VFS_FILEMAP_DIRTY_FOLIO
.dirty_folio = filemap_dirty_folio,
#endif
};
const struct file_operations zpl_file_operations = {
.open = zpl_open,
.release = zpl_release,
.llseek = zpl_llseek,
#ifdef HAVE_VFS_RW_ITERATE
#ifdef HAVE_NEW_SYNC_READ
.read = new_sync_read,
.write = new_sync_write,
#endif
.read_iter = zpl_iter_read,
.write_iter = zpl_iter_write,
#ifdef HAVE_VFS_IOV_ITER
.splice_read = generic_file_splice_read,
.splice_write = iter_file_splice_write,
#endif
#else
.read = do_sync_read,
.write = do_sync_write,
.aio_read = zpl_aio_read,
.aio_write = zpl_aio_write,
#endif
.mmap = zpl_mmap,
.fsync = zpl_fsync,
#ifdef HAVE_FILE_AIO_FSYNC
.aio_fsync = zpl_aio_fsync,
#endif
.fallocate = zpl_fallocate,
#ifdef HAVE_FILE_FADVISE
.fadvise = zpl_fadvise,
#endif
.unlocked_ioctl = zpl_ioctl,
#ifdef CONFIG_COMPAT
.compat_ioctl = zpl_compat_ioctl,
#endif
};
const struct file_operations zpl_dir_file_operations = {
.llseek = generic_file_llseek,
.read = generic_read_dir,
#if defined(HAVE_VFS_ITERATE_SHARED)
.iterate_shared = zpl_iterate,
#elif defined(HAVE_VFS_ITERATE)
.iterate = zpl_iterate,
#else
.readdir = zpl_readdir,
#endif
.fsync = zpl_fsync,
.unlocked_ioctl = zpl_ioctl,
#ifdef CONFIG_COMPAT
.compat_ioctl = zpl_compat_ioctl,
#endif
};
/* CSTYLED */
module_param(zfs_fallocate_reserve_percent, uint, 0644);
MODULE_PARM_DESC(zfs_fallocate_reserve_percent,
"Percentage of length to use for the available capacity check");
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c b/sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c
index 993447e54683..5f5ad186a61c 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zpl_inode.c
@@ -1,876 +1,909 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2011, Lawrence Livermore National Security, LLC.
* Copyright (c) 2015 by Chunwei Chen. All rights reserved.
*/
#include <sys/sysmacros.h>
#include <sys/zfs_ctldir.h>
#include <sys/zfs_vfsops.h>
#include <sys/zfs_vnops.h>
#include <sys/zfs_znode.h>
#include <sys/dmu_objset.h>
#include <sys/vfs.h>
#include <sys/zpl.h>
#include <sys/file.h>
static struct dentry *
zpl_lookup(struct inode *dir, struct dentry *dentry, unsigned int flags)
{
cred_t *cr = CRED();
struct inode *ip;
znode_t *zp;
int error;
fstrans_cookie_t cookie;
pathname_t *ppn = NULL;
pathname_t pn;
int zfs_flags = 0;
zfsvfs_t *zfsvfs = dentry->d_sb->s_fs_info;
if (dlen(dentry) >= ZAP_MAXNAMELEN)
return (ERR_PTR(-ENAMETOOLONG));
crhold(cr);
cookie = spl_fstrans_mark();
/* If we are a case insensitive fs, we need the real name */
if (zfsvfs->z_case == ZFS_CASE_INSENSITIVE) {
zfs_flags = FIGNORECASE;
pn_alloc(&pn);
ppn = &pn;
}
error = -zfs_lookup(ITOZ(dir), dname(dentry), &zp,
zfs_flags, cr, NULL, ppn);
spl_fstrans_unmark(cookie);
ASSERT3S(error, <=, 0);
crfree(cr);
spin_lock(&dentry->d_lock);
dentry->d_time = jiffies;
spin_unlock(&dentry->d_lock);
if (error) {
/*
* If we have a case sensitive fs, we do not want to
* insert negative entries, so return NULL for ENOENT.
* Fall through if the error is not ENOENT. Also free memory.
*/
if (ppn) {
pn_free(ppn);
if (error == -ENOENT)
return (NULL);
}
if (error == -ENOENT)
return (d_splice_alias(NULL, dentry));
else
return (ERR_PTR(error));
}
ip = ZTOI(zp);
/*
* If we are case insensitive, call the correct function
* to install the name.
*/
if (ppn) {
struct dentry *new_dentry;
struct qstr ci_name;
if (strcmp(dname(dentry), pn.pn_buf) == 0) {
new_dentry = d_splice_alias(ip, dentry);
} else {
ci_name.name = pn.pn_buf;
ci_name.len = strlen(pn.pn_buf);
new_dentry = d_add_ci(dentry, ip, &ci_name);
}
pn_free(ppn);
return (new_dentry);
} else {
return (d_splice_alias(ip, dentry));
}
}
void
zpl_vap_init(vattr_t *vap, struct inode *dir, umode_t mode, cred_t *cr,
- zuserns_t *mnt_ns)
+ zidmap_t *mnt_ns)
{
vap->va_mask = ATTR_MODE;
vap->va_mode = mode;
- vap->va_uid = zfs_vfsuid_to_uid((struct user_namespace *)mnt_ns,
+ vap->va_uid = zfs_vfsuid_to_uid(mnt_ns,
zfs_i_user_ns(dir), crgetuid(cr));
if (dir->i_mode & S_ISGID) {
vap->va_gid = KGID_TO_SGID(dir->i_gid);
if (S_ISDIR(mode))
vap->va_mode |= S_ISGID;
} else {
- vap->va_gid = zfs_vfsgid_to_gid((struct user_namespace *)mnt_ns,
+ vap->va_gid = zfs_vfsgid_to_gid(mnt_ns,
zfs_i_user_ns(dir), crgetgid(cr));
}
}
static int
#ifdef HAVE_IOPS_CREATE_USERNS
zpl_create(struct user_namespace *user_ns, struct inode *dir,
struct dentry *dentry, umode_t mode, bool flag)
+#elif defined(HAVE_IOPS_CREATE_IDMAP)
+zpl_create(struct mnt_idmap *user_ns, struct inode *dir,
+ struct dentry *dentry, umode_t mode, bool flag)
#else
zpl_create(struct inode *dir, struct dentry *dentry, umode_t mode, bool flag)
#endif
{
cred_t *cr = CRED();
znode_t *zp;
vattr_t *vap;
int error;
fstrans_cookie_t cookie;
-#ifndef HAVE_IOPS_CREATE_USERNS
- zuserns_t *user_ns = kcred->user_ns;
+#if !(defined(HAVE_IOPS_CREATE_USERNS) || defined(HAVE_IOPS_CREATE_IDMAP))
+ zidmap_t *user_ns = kcred->user_ns;
#endif
crhold(cr);
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
zpl_vap_init(vap, dir, mode, cr, user_ns);
cookie = spl_fstrans_mark();
error = -zfs_create(ITOZ(dir), dname(dentry), vap, 0,
mode, &zp, cr, 0, NULL, user_ns);
if (error == 0) {
error = zpl_xattr_security_init(ZTOI(zp), dir, &dentry->d_name);
if (error == 0)
error = zpl_init_acl(ZTOI(zp), dir);
if (error) {
(void) zfs_remove(ITOZ(dir), dname(dentry), cr, 0);
remove_inode_hash(ZTOI(zp));
iput(ZTOI(zp));
} else {
d_instantiate(dentry, ZTOI(zp));
}
}
spl_fstrans_unmark(cookie);
kmem_free(vap, sizeof (vattr_t));
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
static int
#ifdef HAVE_IOPS_MKNOD_USERNS
zpl_mknod(struct user_namespace *user_ns, struct inode *dir,
struct dentry *dentry, umode_t mode,
+#elif defined(HAVE_IOPS_MKNOD_IDMAP)
+zpl_mknod(struct mnt_idmap *user_ns, struct inode *dir,
+ struct dentry *dentry, umode_t mode,
#else
zpl_mknod(struct inode *dir, struct dentry *dentry, umode_t mode,
#endif
dev_t rdev)
{
cred_t *cr = CRED();
znode_t *zp;
vattr_t *vap;
int error;
fstrans_cookie_t cookie;
-#ifndef HAVE_IOPS_MKNOD_USERNS
- zuserns_t *user_ns = kcred->user_ns;
+#if !(defined(HAVE_IOPS_MKNOD_USERNS) || defined(HAVE_IOPS_MKNOD_IDMAP))
+ zidmap_t *user_ns = kcred->user_ns;
#endif
/*
* We currently expect Linux to supply rdev=0 for all sockets
* and fifos, but we want to know if this behavior ever changes.
*/
if (S_ISSOCK(mode) || S_ISFIFO(mode))
ASSERT(rdev == 0);
crhold(cr);
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
zpl_vap_init(vap, dir, mode, cr, user_ns);
vap->va_rdev = rdev;
cookie = spl_fstrans_mark();
error = -zfs_create(ITOZ(dir), dname(dentry), vap, 0,
mode, &zp, cr, 0, NULL, user_ns);
if (error == 0) {
error = zpl_xattr_security_init(ZTOI(zp), dir, &dentry->d_name);
if (error == 0)
error = zpl_init_acl(ZTOI(zp), dir);
if (error) {
(void) zfs_remove(ITOZ(dir), dname(dentry), cr, 0);
remove_inode_hash(ZTOI(zp));
iput(ZTOI(zp));
} else {
d_instantiate(dentry, ZTOI(zp));
}
}
spl_fstrans_unmark(cookie);
kmem_free(vap, sizeof (vattr_t));
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
#ifdef HAVE_TMPFILE
static int
-#ifndef HAVE_TMPFILE_DENTRY
+#ifdef HAVE_TMPFILE_IDMAP
+zpl_tmpfile(struct mnt_idmap *userns, struct inode *dir,
+ struct file *file, umode_t mode)
+#elif !defined(HAVE_TMPFILE_DENTRY)
zpl_tmpfile(struct user_namespace *userns, struct inode *dir,
struct file *file, umode_t mode)
#else
#ifdef HAVE_TMPFILE_USERNS
zpl_tmpfile(struct user_namespace *userns, struct inode *dir,
struct dentry *dentry, umode_t mode)
#else
zpl_tmpfile(struct inode *dir, struct dentry *dentry, umode_t mode)
#endif
#endif
{
cred_t *cr = CRED();
struct inode *ip;
vattr_t *vap;
int error;
fstrans_cookie_t cookie;
-#ifndef HAVE_TMPFILE_USERNS
- zuserns_t *userns = kcred->user_ns;
+#if !(defined(HAVE_TMPFILE_USERNS) || defined(HAVE_TMPFILE_IDMAP))
+ zidmap_t *userns = kcred->user_ns;
#endif
crhold(cr);
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
/*
* The VFS does not apply the umask, therefore it is applied here
* when POSIX ACLs are not enabled.
*/
if (!IS_POSIXACL(dir))
mode &= ~current_umask();
zpl_vap_init(vap, dir, mode, cr, userns);
cookie = spl_fstrans_mark();
error = -zfs_tmpfile(dir, vap, 0, mode, &ip, cr, 0, NULL, userns);
if (error == 0) {
/* d_tmpfile will do drop_nlink, so we should set it first */
set_nlink(ip, 1);
#ifndef HAVE_TMPFILE_DENTRY
d_tmpfile(file, ip);
error = zpl_xattr_security_init(ip, dir,
&file->f_path.dentry->d_name);
#else
d_tmpfile(dentry, ip);
error = zpl_xattr_security_init(ip, dir, &dentry->d_name);
#endif
if (error == 0)
error = zpl_init_acl(ip, dir);
#ifndef HAVE_TMPFILE_DENTRY
error = finish_open_simple(file, error);
#endif
/*
* don't need to handle error here, file is already in
* unlinked set.
*/
}
spl_fstrans_unmark(cookie);
kmem_free(vap, sizeof (vattr_t));
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
#endif
static int
zpl_unlink(struct inode *dir, struct dentry *dentry)
{
cred_t *cr = CRED();
int error;
fstrans_cookie_t cookie;
zfsvfs_t *zfsvfs = dentry->d_sb->s_fs_info;
crhold(cr);
cookie = spl_fstrans_mark();
error = -zfs_remove(ITOZ(dir), dname(dentry), cr, 0);
/*
* For a CI FS we must invalidate the dentry to prevent the
* creation of negative entries.
*/
if (error == 0 && zfsvfs->z_case == ZFS_CASE_INSENSITIVE)
d_invalidate(dentry);
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
static int
#ifdef HAVE_IOPS_MKDIR_USERNS
zpl_mkdir(struct user_namespace *user_ns, struct inode *dir,
struct dentry *dentry, umode_t mode)
+#elif defined(HAVE_IOPS_MKDIR_IDMAP)
+zpl_mkdir(struct mnt_idmap *user_ns, struct inode *dir,
+ struct dentry *dentry, umode_t mode)
#else
zpl_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
#endif
{
cred_t *cr = CRED();
vattr_t *vap;
znode_t *zp;
int error;
fstrans_cookie_t cookie;
-#ifndef HAVE_IOPS_MKDIR_USERNS
- zuserns_t *user_ns = kcred->user_ns;
+#if !(defined(HAVE_IOPS_MKDIR_USERNS) || defined(HAVE_IOPS_MKDIR_IDMAP))
+ zidmap_t *user_ns = kcred->user_ns;
#endif
crhold(cr);
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
zpl_vap_init(vap, dir, mode | S_IFDIR, cr, user_ns);
cookie = spl_fstrans_mark();
error = -zfs_mkdir(ITOZ(dir), dname(dentry), vap, &zp, cr, 0, NULL,
user_ns);
if (error == 0) {
error = zpl_xattr_security_init(ZTOI(zp), dir, &dentry->d_name);
if (error == 0)
error = zpl_init_acl(ZTOI(zp), dir);
if (error) {
(void) zfs_rmdir(ITOZ(dir), dname(dentry), NULL, cr, 0);
remove_inode_hash(ZTOI(zp));
iput(ZTOI(zp));
} else {
d_instantiate(dentry, ZTOI(zp));
}
}
spl_fstrans_unmark(cookie);
kmem_free(vap, sizeof (vattr_t));
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
static int
zpl_rmdir(struct inode *dir, struct dentry *dentry)
{
cred_t *cr = CRED();
int error;
fstrans_cookie_t cookie;
zfsvfs_t *zfsvfs = dentry->d_sb->s_fs_info;
crhold(cr);
cookie = spl_fstrans_mark();
error = -zfs_rmdir(ITOZ(dir), dname(dentry), NULL, cr, 0);
/*
* For a CI FS we must invalidate the dentry to prevent the
* creation of negative entries.
*/
if (error == 0 && zfsvfs->z_case == ZFS_CASE_INSENSITIVE)
d_invalidate(dentry);
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
static int
#ifdef HAVE_USERNS_IOPS_GETATTR
zpl_getattr_impl(struct user_namespace *user_ns,
const struct path *path, struct kstat *stat, u32 request_mask,
unsigned int query_flags)
+#elif defined(HAVE_IDMAP_IOPS_GETATTR)
+zpl_getattr_impl(struct mnt_idmap *user_ns,
+ const struct path *path, struct kstat *stat, u32 request_mask,
+ unsigned int query_flags)
#else
zpl_getattr_impl(const struct path *path, struct kstat *stat, u32 request_mask,
unsigned int query_flags)
#endif
{
int error;
fstrans_cookie_t cookie;
struct inode *ip = path->dentry->d_inode;
znode_t *zp __maybe_unused = ITOZ(ip);
cookie = spl_fstrans_mark();
/*
* XXX query_flags currently ignored.
*/
-#ifdef HAVE_USERNS_IOPS_GETATTR
+#if (defined(HAVE_USERNS_IOPS_GETATTR) || defined(HAVE_IDMAP_IOPS_GETATTR))
error = -zfs_getattr_fast(user_ns, ip, stat);
#else
error = -zfs_getattr_fast(kcred->user_ns, ip, stat);
#endif
#ifdef STATX_BTIME
if (request_mask & STATX_BTIME) {
stat->btime = zp->z_btime;
stat->result_mask |= STATX_BTIME;
}
#endif
#ifdef STATX_ATTR_IMMUTABLE
if (zp->z_pflags & ZFS_IMMUTABLE)
stat->attributes |= STATX_ATTR_IMMUTABLE;
stat->attributes_mask |= STATX_ATTR_IMMUTABLE;
#endif
#ifdef STATX_ATTR_APPEND
if (zp->z_pflags & ZFS_APPENDONLY)
stat->attributes |= STATX_ATTR_APPEND;
stat->attributes_mask |= STATX_ATTR_APPEND;
#endif
#ifdef STATX_ATTR_NODUMP
if (zp->z_pflags & ZFS_NODUMP)
stat->attributes |= STATX_ATTR_NODUMP;
stat->attributes_mask |= STATX_ATTR_NODUMP;
#endif
spl_fstrans_unmark(cookie);
ASSERT3S(error, <=, 0);
return (error);
}
ZPL_GETATTR_WRAPPER(zpl_getattr);
static int
-#ifdef HAVE_SETATTR_PREPARE_USERNS
+#ifdef HAVE_USERNS_IOPS_SETATTR
zpl_setattr(struct user_namespace *user_ns, struct dentry *dentry,
struct iattr *ia)
+#elif defined(HAVE_IDMAP_IOPS_SETATTR)
+zpl_setattr(struct mnt_idmap *user_ns, struct dentry *dentry,
+ struct iattr *ia)
#else
zpl_setattr(struct dentry *dentry, struct iattr *ia)
#endif
{
struct inode *ip = dentry->d_inode;
cred_t *cr = CRED();
vattr_t *vap;
int error;
fstrans_cookie_t cookie;
#ifdef HAVE_SETATTR_PREPARE_USERNS
error = zpl_setattr_prepare(user_ns, dentry, ia);
+#elif defined(HAVE_SETATTR_PREPARE_IDMAP)
+ error = zpl_setattr_prepare(user_ns, dentry, ia);
#else
- error = zpl_setattr_prepare(kcred->user_ns, dentry, ia);
+ error = zpl_setattr_prepare(zfs_init_idmap, dentry, ia);
#endif
if (error)
return (error);
crhold(cr);
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
vap->va_mask = ia->ia_valid & ATTR_IATTR_MASK;
vap->va_mode = ia->ia_mode;
if (ia->ia_valid & ATTR_UID)
#ifdef HAVE_IATTR_VFSID
vap->va_uid = zfs_vfsuid_to_uid(user_ns, zfs_i_user_ns(ip),
__vfsuid_val(ia->ia_vfsuid));
#else
vap->va_uid = KUID_TO_SUID(ia->ia_uid);
#endif
if (ia->ia_valid & ATTR_GID)
#ifdef HAVE_IATTR_VFSID
vap->va_gid = zfs_vfsgid_to_gid(user_ns, zfs_i_user_ns(ip),
__vfsgid_val(ia->ia_vfsgid));
#else
vap->va_gid = KGID_TO_SGID(ia->ia_gid);
#endif
vap->va_size = ia->ia_size;
vap->va_atime = ia->ia_atime;
vap->va_mtime = ia->ia_mtime;
vap->va_ctime = ia->ia_ctime;
if (vap->va_mask & ATTR_ATIME)
ip->i_atime = zpl_inode_timestamp_truncate(ia->ia_atime, ip);
cookie = spl_fstrans_mark();
-#ifdef HAVE_SETATTR_PREPARE_USERNS
+#ifdef HAVE_USERNS_IOPS_SETATTR
+ error = -zfs_setattr(ITOZ(ip), vap, 0, cr, user_ns);
+#elif defined(HAVE_IDMAP_IOPS_SETATTR)
error = -zfs_setattr(ITOZ(ip), vap, 0, cr, user_ns);
#else
- error = -zfs_setattr(ITOZ(ip), vap, 0, cr, kcred->user_ns);
+ error = -zfs_setattr(ITOZ(ip), vap, 0, cr, zfs_init_idmap);
#endif
if (!error && (ia->ia_valid & ATTR_MODE))
error = zpl_chmod_acl(ip);
spl_fstrans_unmark(cookie);
kmem_free(vap, sizeof (vattr_t));
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
static int
#ifdef HAVE_IOPS_RENAME_USERNS
zpl_rename2(struct user_namespace *user_ns, struct inode *sdip,
struct dentry *sdentry, struct inode *tdip, struct dentry *tdentry,
unsigned int rflags)
+#elif defined(HAVE_IOPS_RENAME_IDMAP)
+zpl_rename2(struct mnt_idmap *user_ns, struct inode *sdip,
+ struct dentry *sdentry, struct inode *tdip, struct dentry *tdentry,
+ unsigned int rflags)
#else
zpl_rename2(struct inode *sdip, struct dentry *sdentry,
struct inode *tdip, struct dentry *tdentry, unsigned int rflags)
#endif
{
cred_t *cr = CRED();
vattr_t *wo_vap = NULL;
int error;
fstrans_cookie_t cookie;
-#ifndef HAVE_IOPS_RENAME_USERNS
- zuserns_t *user_ns = kcred->user_ns;
+#if !(defined(HAVE_IOPS_RENAME_USERNS) || defined(HAVE_IOPS_RENAME_IDMAP))
+ zidmap_t *user_ns = kcred->user_ns;
#endif
crhold(cr);
if (rflags & RENAME_WHITEOUT) {
wo_vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
zpl_vap_init(wo_vap, sdip, S_IFCHR, cr, user_ns);
wo_vap->va_rdev = makedevice(0, 0);
}
cookie = spl_fstrans_mark();
error = -zfs_rename(ITOZ(sdip), dname(sdentry), ITOZ(tdip),
dname(tdentry), cr, 0, rflags, wo_vap, user_ns);
spl_fstrans_unmark(cookie);
if (wo_vap)
kmem_free(wo_vap, sizeof (vattr_t));
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
#if !defined(HAVE_IOPS_RENAME_USERNS) && \
!defined(HAVE_RENAME_WANTS_FLAGS) && \
- !defined(HAVE_RENAME2)
+ !defined(HAVE_RENAME2) && \
+ !defined(HAVE_IOPS_RENAME_IDMAP)
static int
zpl_rename(struct inode *sdip, struct dentry *sdentry,
struct inode *tdip, struct dentry *tdentry)
{
return (zpl_rename2(sdip, sdentry, tdip, tdentry, 0));
}
#endif
static int
#ifdef HAVE_IOPS_SYMLINK_USERNS
zpl_symlink(struct user_namespace *user_ns, struct inode *dir,
struct dentry *dentry, const char *name)
+#elif defined(HAVE_IOPS_SYMLINK_IDMAP)
+zpl_symlink(struct mnt_idmap *user_ns, struct inode *dir,
+ struct dentry *dentry, const char *name)
#else
zpl_symlink(struct inode *dir, struct dentry *dentry, const char *name)
#endif
{
cred_t *cr = CRED();
vattr_t *vap;
znode_t *zp;
int error;
fstrans_cookie_t cookie;
-#ifndef HAVE_IOPS_SYMLINK_USERNS
- zuserns_t *user_ns = kcred->user_ns;
+#if !(defined(HAVE_IOPS_SYMLINK_USERNS) || defined(HAVE_IOPS_SYMLINK_IDMAP))
+ zidmap_t *user_ns = kcred->user_ns;
#endif
crhold(cr);
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
zpl_vap_init(vap, dir, S_IFLNK | S_IRWXUGO, cr, user_ns);
cookie = spl_fstrans_mark();
error = -zfs_symlink(ITOZ(dir), dname(dentry), vap,
(char *)name, &zp, cr, 0, user_ns);
if (error == 0) {
error = zpl_xattr_security_init(ZTOI(zp), dir, &dentry->d_name);
if (error) {
(void) zfs_remove(ITOZ(dir), dname(dentry), cr, 0);
remove_inode_hash(ZTOI(zp));
iput(ZTOI(zp));
} else {
d_instantiate(dentry, ZTOI(zp));
}
}
spl_fstrans_unmark(cookie);
kmem_free(vap, sizeof (vattr_t));
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
#if defined(HAVE_PUT_LINK_COOKIE)
static void
zpl_put_link(struct inode *unused, void *cookie)
{
kmem_free(cookie, MAXPATHLEN);
}
#elif defined(HAVE_PUT_LINK_NAMEIDATA)
static void
zpl_put_link(struct dentry *dentry, struct nameidata *nd, void *ptr)
{
const char *link = nd_get_link(nd);
if (!IS_ERR(link))
kmem_free(link, MAXPATHLEN);
}
#elif defined(HAVE_PUT_LINK_DELAYED)
static void
zpl_put_link(void *ptr)
{
kmem_free(ptr, MAXPATHLEN);
}
#endif
static int
zpl_get_link_common(struct dentry *dentry, struct inode *ip, char **link)
{
fstrans_cookie_t cookie;
cred_t *cr = CRED();
int error;
crhold(cr);
*link = NULL;
struct iovec iov;
iov.iov_len = MAXPATHLEN;
iov.iov_base = kmem_zalloc(MAXPATHLEN, KM_SLEEP);
zfs_uio_t uio;
zfs_uio_iovec_init(&uio, &iov, 1, 0, UIO_SYSSPACE, MAXPATHLEN - 1, 0);
cookie = spl_fstrans_mark();
error = -zfs_readlink(ip, &uio, cr);
spl_fstrans_unmark(cookie);
crfree(cr);
if (error)
kmem_free(iov.iov_base, MAXPATHLEN);
else
*link = iov.iov_base;
return (error);
}
#if defined(HAVE_GET_LINK_DELAYED)
static const char *
zpl_get_link(struct dentry *dentry, struct inode *inode,
struct delayed_call *done)
{
char *link = NULL;
int error;
if (!dentry)
return (ERR_PTR(-ECHILD));
error = zpl_get_link_common(dentry, inode, &link);
if (error)
return (ERR_PTR(error));
set_delayed_call(done, zpl_put_link, link);
return (link);
}
#elif defined(HAVE_GET_LINK_COOKIE)
static const char *
zpl_get_link(struct dentry *dentry, struct inode *inode, void **cookie)
{
char *link = NULL;
int error;
if (!dentry)
return (ERR_PTR(-ECHILD));
error = zpl_get_link_common(dentry, inode, &link);
if (error)
return (ERR_PTR(error));
return (*cookie = link);
}
#elif defined(HAVE_FOLLOW_LINK_COOKIE)
static const char *
zpl_follow_link(struct dentry *dentry, void **cookie)
{
char *link = NULL;
int error;
error = zpl_get_link_common(dentry, dentry->d_inode, &link);
if (error)
return (ERR_PTR(error));
return (*cookie = link);
}
#elif defined(HAVE_FOLLOW_LINK_NAMEIDATA)
static void *
zpl_follow_link(struct dentry *dentry, struct nameidata *nd)
{
char *link = NULL;
int error;
error = zpl_get_link_common(dentry, dentry->d_inode, &link);
if (error)
nd_set_link(nd, ERR_PTR(error));
else
nd_set_link(nd, link);
return (NULL);
}
#endif
static int
zpl_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry)
{
cred_t *cr = CRED();
struct inode *ip = old_dentry->d_inode;
int error;
fstrans_cookie_t cookie;
if (ip->i_nlink >= ZFS_LINK_MAX)
return (-EMLINK);
crhold(cr);
ip->i_ctime = current_time(ip);
/* Must have an existing ref, so igrab() cannot return NULL */
VERIFY3P(igrab(ip), !=, NULL);
cookie = spl_fstrans_mark();
error = -zfs_link(ITOZ(dir), ITOZ(ip), dname(dentry), cr, 0);
if (error) {
iput(ip);
goto out;
}
d_instantiate(dentry, ip);
out:
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
const struct inode_operations zpl_inode_operations = {
.setattr = zpl_setattr,
.getattr = zpl_getattr,
#ifdef HAVE_GENERIC_SETXATTR
.setxattr = generic_setxattr,
.getxattr = generic_getxattr,
.removexattr = generic_removexattr,
#endif
.listxattr = zpl_xattr_list,
#if defined(CONFIG_FS_POSIX_ACL)
#if defined(HAVE_SET_ACL)
.set_acl = zpl_set_acl,
#endif /* HAVE_SET_ACL */
#if defined(HAVE_GET_INODE_ACL)
.get_inode_acl = zpl_get_acl,
#else
.get_acl = zpl_get_acl,
#endif /* HAVE_GET_INODE_ACL */
#endif /* CONFIG_FS_POSIX_ACL */
};
#ifdef HAVE_RENAME2_OPERATIONS_WRAPPER
const struct inode_operations_wrapper zpl_dir_inode_operations = {
.ops = {
#else
const struct inode_operations zpl_dir_inode_operations = {
#endif
.create = zpl_create,
.lookup = zpl_lookup,
.link = zpl_link,
.unlink = zpl_unlink,
.symlink = zpl_symlink,
.mkdir = zpl_mkdir,
.rmdir = zpl_rmdir,
.mknod = zpl_mknod,
#ifdef HAVE_RENAME2
.rename2 = zpl_rename2,
#elif defined(HAVE_RENAME_WANTS_FLAGS) || defined(HAVE_IOPS_RENAME_USERNS)
.rename = zpl_rename2,
+#elif defined(HAVE_IOPS_RENAME_IDMAP)
+ .rename = zpl_rename2,
#else
.rename = zpl_rename,
#endif
#ifdef HAVE_TMPFILE
.tmpfile = zpl_tmpfile,
#endif
.setattr = zpl_setattr,
.getattr = zpl_getattr,
#ifdef HAVE_GENERIC_SETXATTR
.setxattr = generic_setxattr,
.getxattr = generic_getxattr,
.removexattr = generic_removexattr,
#endif
.listxattr = zpl_xattr_list,
#if defined(CONFIG_FS_POSIX_ACL)
#if defined(HAVE_SET_ACL)
.set_acl = zpl_set_acl,
#endif /* HAVE_SET_ACL */
#if defined(HAVE_GET_INODE_ACL)
.get_inode_acl = zpl_get_acl,
#else
.get_acl = zpl_get_acl,
#endif /* HAVE_GET_INODE_ACL */
#endif /* CONFIG_FS_POSIX_ACL */
#ifdef HAVE_RENAME2_OPERATIONS_WRAPPER
},
.rename2 = zpl_rename2,
#endif
};
const struct inode_operations zpl_symlink_inode_operations = {
#ifdef HAVE_GENERIC_READLINK
.readlink = generic_readlink,
#endif
#if defined(HAVE_GET_LINK_DELAYED) || defined(HAVE_GET_LINK_COOKIE)
.get_link = zpl_get_link,
#elif defined(HAVE_FOLLOW_LINK_COOKIE) || defined(HAVE_FOLLOW_LINK_NAMEIDATA)
.follow_link = zpl_follow_link,
#endif
#if defined(HAVE_PUT_LINK_COOKIE) || defined(HAVE_PUT_LINK_NAMEIDATA)
.put_link = zpl_put_link,
#endif
.setattr = zpl_setattr,
.getattr = zpl_getattr,
#ifdef HAVE_GENERIC_SETXATTR
.setxattr = generic_setxattr,
.getxattr = generic_getxattr,
.removexattr = generic_removexattr,
#endif
.listxattr = zpl_xattr_list,
};
const struct inode_operations zpl_special_inode_operations = {
.setattr = zpl_setattr,
.getattr = zpl_getattr,
#ifdef HAVE_GENERIC_SETXATTR
.setxattr = generic_setxattr,
.getxattr = generic_getxattr,
.removexattr = generic_removexattr,
#endif
.listxattr = zpl_xattr_list,
#if defined(CONFIG_FS_POSIX_ACL)
#if defined(HAVE_SET_ACL)
.set_acl = zpl_set_acl,
#endif /* HAVE_SET_ACL */
#if defined(HAVE_GET_INODE_ACL)
.get_inode_acl = zpl_get_acl,
#else
.get_acl = zpl_get_acl,
#endif /* HAVE_GET_INODE_ACL */
#endif /* CONFIG_FS_POSIX_ACL */
};
diff --git a/sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c b/sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c
index 4156d686732a..96d85991811e 100644
--- a/sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c
+++ b/sys/contrib/openzfs/module/os/linux/zfs/zpl_xattr.c
@@ -1,1632 +1,1637 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2011, Lawrence Livermore National Security, LLC.
*
* Extended attributes (xattr) on Solaris are implemented as files
* which exist in a hidden xattr directory. These extended attributes
* can be accessed using the attropen() system call which opens
* the extended attribute. It can then be manipulated just like
* a standard file descriptor. This has a couple advantages such
* as practically no size limit on the file, and the extended
* attributes permissions may differ from those of the parent file.
* This interface is really quite clever, but it's also completely
* different than what is supported on Linux. It also comes with a
* steep performance penalty when accessing small xattrs because they
* are not stored with the parent file.
*
* Under Linux extended attributes are manipulated by the system
* calls getxattr(2), setxattr(2), and listxattr(2). They consider
* extended attributes to be name/value pairs where the name is a
* NULL terminated string. The name must also include one of the
* following namespace prefixes:
*
* user - No restrictions and is available to user applications.
* trusted - Restricted to kernel and root (CAP_SYS_ADMIN) use.
* system - Used for access control lists (system.nfs4_acl, etc).
* security - Used by SELinux to store a files security context.
*
* The value under Linux to limited to 65536 bytes of binary data.
* In practice, individual xattrs tend to be much smaller than this
* and are typically less than 100 bytes. A good example of this
* are the security.selinux xattrs which are less than 100 bytes and
* exist for every file when xattr labeling is enabled.
*
* The Linux xattr implementation has been written to take advantage of
* this typical usage. When the dataset property 'xattr=sa' is set,
* then xattrs will be preferentially stored as System Attributes (SA).
* This allows tiny xattrs (~100 bytes) to be stored with the dnode and
* up to 64k of xattrs to be stored in the spill block. If additional
* xattr space is required, which is unlikely under Linux, they will
* be stored using the traditional directory approach.
*
* This optimization results in roughly a 3x performance improvement
* when accessing xattrs because it avoids the need to perform a seek
* for every xattr value. When multiple xattrs are stored per-file
* the performance improvements are even greater because all of the
* xattrs stored in the spill block will be cached.
*
* However, by default SA based xattrs are disabled in the Linux port
* to maximize compatibility with other implementations. If you do
* enable SA based xattrs then they will not be visible on platforms
* which do not support this feature.
*
* NOTE: One additional consequence of the xattr directory implementation
* is that when an extended attribute is manipulated an inode is created.
* This inode will exist in the Linux inode cache but there will be no
* associated entry in the dentry cache which references it. This is
* safe but it may result in some confusion. Enabling SA based xattrs
* largely avoids the issue except in the overflow case.
*/
#include <sys/zfs_znode.h>
#include <sys/zfs_vfsops.h>
#include <sys/zfs_vnops.h>
#include <sys/zap.h>
#include <sys/vfs.h>
#include <sys/zpl.h>
#include <linux/vfs_compat.h>
enum xattr_permission {
XAPERM_DENY,
XAPERM_ALLOW,
XAPERM_COMPAT,
};
typedef struct xattr_filldir {
size_t size;
size_t offset;
char *buf;
struct dentry *dentry;
} xattr_filldir_t;
static enum xattr_permission zpl_xattr_permission(xattr_filldir_t *,
const char *, int);
static int zfs_xattr_compat = 0;
/*
* Determine is a given xattr name should be visible and if so copy it
* in to the provided buffer (xf->buf).
*/
static int
zpl_xattr_filldir(xattr_filldir_t *xf, const char *name, int name_len)
{
enum xattr_permission perm;
/* Check permissions using the per-namespace list xattr handler. */
perm = zpl_xattr_permission(xf, name, name_len);
if (perm == XAPERM_DENY)
return (0);
/* Prefix the name with "user." if it does not have a namespace. */
if (perm == XAPERM_COMPAT) {
if (xf->buf) {
if (xf->offset + XATTR_USER_PREFIX_LEN + 1 > xf->size)
return (-ERANGE);
memcpy(xf->buf + xf->offset, XATTR_USER_PREFIX,
XATTR_USER_PREFIX_LEN);
xf->buf[xf->offset + XATTR_USER_PREFIX_LEN] = '\0';
}
xf->offset += XATTR_USER_PREFIX_LEN;
}
/* When xf->buf is NULL only calculate the required size. */
if (xf->buf) {
if (xf->offset + name_len + 1 > xf->size)
return (-ERANGE);
memcpy(xf->buf + xf->offset, name, name_len);
xf->buf[xf->offset + name_len] = '\0';
}
xf->offset += (name_len + 1);
return (0);
}
/*
* Read as many directory entry names as will fit in to the provided buffer,
* or when no buffer is provided calculate the required buffer size.
*/
static int
zpl_xattr_readdir(struct inode *dxip, xattr_filldir_t *xf)
{
zap_cursor_t zc;
zap_attribute_t zap;
int error;
zap_cursor_init(&zc, ITOZSB(dxip)->z_os, ITOZ(dxip)->z_id);
while ((error = -zap_cursor_retrieve(&zc, &zap)) == 0) {
if (zap.za_integer_length != 8 || zap.za_num_integers != 1) {
error = -ENXIO;
break;
}
error = zpl_xattr_filldir(xf, zap.za_name, strlen(zap.za_name));
if (error)
break;
zap_cursor_advance(&zc);
}
zap_cursor_fini(&zc);
if (error == -ENOENT)
error = 0;
return (error);
}
static ssize_t
zpl_xattr_list_dir(xattr_filldir_t *xf, cred_t *cr)
{
struct inode *ip = xf->dentry->d_inode;
struct inode *dxip = NULL;
znode_t *dxzp;
int error;
/* Lookup the xattr directory */
error = -zfs_lookup(ITOZ(ip), NULL, &dxzp, LOOKUP_XATTR,
cr, NULL, NULL);
if (error) {
if (error == -ENOENT)
error = 0;
return (error);
}
dxip = ZTOI(dxzp);
error = zpl_xattr_readdir(dxip, xf);
iput(dxip);
return (error);
}
static ssize_t
zpl_xattr_list_sa(xattr_filldir_t *xf)
{
znode_t *zp = ITOZ(xf->dentry->d_inode);
nvpair_t *nvp = NULL;
int error = 0;
mutex_enter(&zp->z_lock);
if (zp->z_xattr_cached == NULL)
error = -zfs_sa_get_xattr(zp);
mutex_exit(&zp->z_lock);
if (error)
return (error);
ASSERT(zp->z_xattr_cached);
while ((nvp = nvlist_next_nvpair(zp->z_xattr_cached, nvp)) != NULL) {
ASSERT3U(nvpair_type(nvp), ==, DATA_TYPE_BYTE_ARRAY);
error = zpl_xattr_filldir(xf, nvpair_name(nvp),
strlen(nvpair_name(nvp)));
if (error)
return (error);
}
return (0);
}
ssize_t
zpl_xattr_list(struct dentry *dentry, char *buffer, size_t buffer_size)
{
znode_t *zp = ITOZ(dentry->d_inode);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
xattr_filldir_t xf = { buffer_size, 0, buffer, dentry };
cred_t *cr = CRED();
fstrans_cookie_t cookie;
int error = 0;
crhold(cr);
cookie = spl_fstrans_mark();
if ((error = zpl_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
goto out1;
rw_enter(&zp->z_xattr_lock, RW_READER);
if (zfsvfs->z_use_sa && zp->z_is_sa) {
error = zpl_xattr_list_sa(&xf);
if (error)
goto out;
}
error = zpl_xattr_list_dir(&xf, cr);
if (error)
goto out;
error = xf.offset;
out:
rw_exit(&zp->z_xattr_lock);
zpl_exit(zfsvfs, FTAG);
out1:
spl_fstrans_unmark(cookie);
crfree(cr);
return (error);
}
static int
zpl_xattr_get_dir(struct inode *ip, const char *name, void *value,
size_t size, cred_t *cr)
{
fstrans_cookie_t cookie;
struct inode *xip = NULL;
znode_t *dxzp = NULL;
znode_t *xzp = NULL;
int error;
/* Lookup the xattr directory */
error = -zfs_lookup(ITOZ(ip), NULL, &dxzp, LOOKUP_XATTR,
cr, NULL, NULL);
if (error)
goto out;
/* Lookup a specific xattr name in the directory */
error = -zfs_lookup(dxzp, (char *)name, &xzp, 0, cr, NULL, NULL);
if (error)
goto out;
xip = ZTOI(xzp);
if (!size) {
error = i_size_read(xip);
goto out;
}
if (size < i_size_read(xip)) {
error = -ERANGE;
goto out;
}
struct iovec iov;
iov.iov_base = (void *)value;
iov.iov_len = size;
zfs_uio_t uio;
zfs_uio_iovec_init(&uio, &iov, 1, 0, UIO_SYSSPACE, size, 0);
cookie = spl_fstrans_mark();
error = -zfs_read(ITOZ(xip), &uio, 0, cr);
spl_fstrans_unmark(cookie);
if (error == 0)
error = size - zfs_uio_resid(&uio);
out:
if (xzp)
zrele(xzp);
if (dxzp)
zrele(dxzp);
return (error);
}
static int
zpl_xattr_get_sa(struct inode *ip, const char *name, void *value, size_t size)
{
znode_t *zp = ITOZ(ip);
uchar_t *nv_value;
uint_t nv_size;
int error = 0;
ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock));
mutex_enter(&zp->z_lock);
if (zp->z_xattr_cached == NULL)
error = -zfs_sa_get_xattr(zp);
mutex_exit(&zp->z_lock);
if (error)
return (error);
ASSERT(zp->z_xattr_cached);
error = -nvlist_lookup_byte_array(zp->z_xattr_cached, name,
&nv_value, &nv_size);
if (error)
return (error);
if (size == 0 || value == NULL)
return (nv_size);
if (size < nv_size)
return (-ERANGE);
memcpy(value, nv_value, nv_size);
return (nv_size);
}
static int
__zpl_xattr_get(struct inode *ip, const char *name, void *value, size_t size,
cred_t *cr)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock));
if (zfsvfs->z_use_sa && zp->z_is_sa) {
error = zpl_xattr_get_sa(ip, name, value, size);
if (error != -ENOENT)
goto out;
}
error = zpl_xattr_get_dir(ip, name, value, size, cr);
out:
if (error == -ENOENT)
error = -ENODATA;
return (error);
}
#define XATTR_NOENT 0x0
#define XATTR_IN_SA 0x1
#define XATTR_IN_DIR 0x2
/* check where the xattr resides */
static int
__zpl_xattr_where(struct inode *ip, const char *name, int *where, cred_t *cr)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
ASSERT(where);
ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock));
*where = XATTR_NOENT;
if (zfsvfs->z_use_sa && zp->z_is_sa) {
error = zpl_xattr_get_sa(ip, name, NULL, 0);
if (error >= 0)
*where |= XATTR_IN_SA;
else if (error != -ENOENT)
return (error);
}
error = zpl_xattr_get_dir(ip, name, NULL, 0, cr);
if (error >= 0)
*where |= XATTR_IN_DIR;
else if (error != -ENOENT)
return (error);
if (*where == (XATTR_IN_SA|XATTR_IN_DIR))
cmn_err(CE_WARN, "ZFS: inode %p has xattr \"%s\""
" in both SA and dir", ip, name);
if (*where == XATTR_NOENT)
error = -ENODATA;
else
error = 0;
return (error);
}
static int
zpl_xattr_get(struct inode *ip, const char *name, void *value, size_t size)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
cred_t *cr = CRED();
fstrans_cookie_t cookie;
int error;
crhold(cr);
cookie = spl_fstrans_mark();
if ((error = zpl_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
goto out;
rw_enter(&zp->z_xattr_lock, RW_READER);
error = __zpl_xattr_get(ip, name, value, size, cr);
rw_exit(&zp->z_xattr_lock);
zpl_exit(zfsvfs, FTAG);
out:
spl_fstrans_unmark(cookie);
crfree(cr);
return (error);
}
static int
zpl_xattr_set_dir(struct inode *ip, const char *name, const void *value,
size_t size, int flags, cred_t *cr)
{
znode_t *dxzp = NULL;
znode_t *xzp = NULL;
vattr_t *vap = NULL;
int lookup_flags, error;
const int xattr_mode = S_IFREG | 0644;
loff_t pos = 0;
/*
* Lookup the xattr directory. When we're adding an entry pass
* CREATE_XATTR_DIR to ensure the xattr directory is created.
* When removing an entry this flag is not passed to avoid
* unnecessarily creating a new xattr directory.
*/
lookup_flags = LOOKUP_XATTR;
if (value != NULL)
lookup_flags |= CREATE_XATTR_DIR;
error = -zfs_lookup(ITOZ(ip), NULL, &dxzp, lookup_flags,
cr, NULL, NULL);
if (error)
goto out;
/* Lookup a specific xattr name in the directory */
error = -zfs_lookup(dxzp, (char *)name, &xzp, 0, cr, NULL, NULL);
if (error && (error != -ENOENT))
goto out;
error = 0;
/* Remove a specific name xattr when value is set to NULL. */
if (value == NULL) {
if (xzp)
error = -zfs_remove(dxzp, (char *)name, cr, 0);
goto out;
}
/* Lookup failed create a new xattr. */
if (xzp == NULL) {
vap = kmem_zalloc(sizeof (vattr_t), KM_SLEEP);
vap->va_mode = xattr_mode;
vap->va_mask = ATTR_MODE;
vap->va_uid = crgetuid(cr);
vap->va_gid = crgetgid(cr);
error = -zfs_create(dxzp, (char *)name, vap, 0, 0644, &xzp,
- cr, ATTR_NOACLCHECK, NULL, kcred->user_ns);
+ cr, ATTR_NOACLCHECK, NULL, zfs_init_idmap);
if (error)
goto out;
}
ASSERT(xzp != NULL);
error = -zfs_freesp(xzp, 0, 0, xattr_mode, TRUE);
if (error)
goto out;
error = -zfs_write_simple(xzp, value, size, pos, NULL);
out:
if (error == 0) {
ip->i_ctime = current_time(ip);
zfs_mark_inode_dirty(ip);
}
if (vap)
kmem_free(vap, sizeof (vattr_t));
if (xzp)
zrele(xzp);
if (dxzp)
zrele(dxzp);
if (error == -ENOENT)
error = -ENODATA;
ASSERT3S(error, <=, 0);
return (error);
}
static int
zpl_xattr_set_sa(struct inode *ip, const char *name, const void *value,
size_t size, int flags, cred_t *cr)
{
znode_t *zp = ITOZ(ip);
nvlist_t *nvl;
size_t sa_size;
int error = 0;
mutex_enter(&zp->z_lock);
if (zp->z_xattr_cached == NULL)
error = -zfs_sa_get_xattr(zp);
mutex_exit(&zp->z_lock);
if (error)
return (error);
ASSERT(zp->z_xattr_cached);
nvl = zp->z_xattr_cached;
if (value == NULL) {
error = -nvlist_remove(nvl, name, DATA_TYPE_BYTE_ARRAY);
if (error == -ENOENT)
error = zpl_xattr_set_dir(ip, name, NULL, 0, flags, cr);
} else {
/* Limited to 32k to keep nvpair memory allocations small */
if (size > DXATTR_MAX_ENTRY_SIZE)
return (-EFBIG);
/* Prevent the DXATTR SA from consuming the entire SA region */
error = -nvlist_size(nvl, &sa_size, NV_ENCODE_XDR);
if (error)
return (error);
if (sa_size > DXATTR_MAX_SA_SIZE)
return (-EFBIG);
error = -nvlist_add_byte_array(nvl, name,
(uchar_t *)value, size);
}
/*
* Update the SA for additions, modifications, and removals. On
* error drop the inconsistent cached version of the nvlist, it
* will be reconstructed from the ARC when next accessed.
*/
if (error == 0)
error = -zfs_sa_set_xattr(zp, name, value, size);
if (error) {
nvlist_free(nvl);
zp->z_xattr_cached = NULL;
}
ASSERT3S(error, <=, 0);
return (error);
}
static int
zpl_xattr_set(struct inode *ip, const char *name, const void *value,
size_t size, int flags)
{
znode_t *zp = ITOZ(ip);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
cred_t *cr = CRED();
fstrans_cookie_t cookie;
int where;
int error;
crhold(cr);
cookie = spl_fstrans_mark();
if ((error = zpl_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
goto out1;
rw_enter(&zp->z_xattr_lock, RW_WRITER);
/*
* Before setting the xattr check to see if it already exists.
* This is done to ensure the following optional flags are honored.
*
* XATTR_CREATE: fail if xattr already exists
* XATTR_REPLACE: fail if xattr does not exist
*
* We also want to know if it resides in sa or dir, so we can make
* sure we don't end up with duplicate in both places.
*/
error = __zpl_xattr_where(ip, name, &where, cr);
if (error < 0) {
if (error != -ENODATA)
goto out;
if (flags & XATTR_REPLACE)
goto out;
/* The xattr to be removed already doesn't exist */
error = 0;
if (value == NULL)
goto out;
} else {
error = -EEXIST;
if (flags & XATTR_CREATE)
goto out;
}
/* Preferentially store the xattr as a SA for better performance */
if (zfsvfs->z_use_sa && zp->z_is_sa &&
(zfsvfs->z_xattr_sa || (value == NULL && where & XATTR_IN_SA))) {
error = zpl_xattr_set_sa(ip, name, value, size, flags, cr);
if (error == 0) {
/*
* Successfully put into SA, we need to clear the one
* in dir.
*/
if (where & XATTR_IN_DIR)
zpl_xattr_set_dir(ip, name, NULL, 0, 0, cr);
goto out;
}
}
error = zpl_xattr_set_dir(ip, name, value, size, flags, cr);
/*
* Successfully put into dir, we need to clear the one in SA.
*/
if (error == 0 && (where & XATTR_IN_SA))
zpl_xattr_set_sa(ip, name, NULL, 0, 0, cr);
out:
rw_exit(&zp->z_xattr_lock);
zpl_exit(zfsvfs, FTAG);
out1:
spl_fstrans_unmark(cookie);
crfree(cr);
ASSERT3S(error, <=, 0);
return (error);
}
/*
* Extended user attributes
*
* "Extended user attributes may be assigned to files and directories for
* storing arbitrary additional information such as the mime type,
* character set or encoding of a file. The access permissions for user
* attributes are defined by the file permission bits: read permission
* is required to retrieve the attribute value, and writer permission is
* required to change it.
*
* The file permission bits of regular files and directories are
* interpreted differently from the file permission bits of special
* files and symbolic links. For regular files and directories the file
* permission bits define access to the file's contents, while for
* device special files they define access to the device described by
* the special file. The file permissions of symbolic links are not
* used in access checks. These differences would allow users to
* consume filesystem resources in a way not controllable by disk quotas
* for group or world writable special files and directories.
*
* For this reason, extended user attributes are allowed only for
* regular files and directories, and access to extended user attributes
* is restricted to the owner and to users with appropriate capabilities
* for directories with the sticky bit set (see the chmod(1) manual page
* for an explanation of the sticky bit)." - xattr(7)
*
* ZFS allows extended user attributes to be disabled administratively
* by setting the 'xattr=off' property on the dataset.
*/
static int
__zpl_xattr_user_list(struct inode *ip, char *list, size_t list_size,
const char *name, size_t name_len)
{
return (ITOZSB(ip)->z_flags & ZSB_XATTR);
}
ZPL_XATTR_LIST_WRAPPER(zpl_xattr_user_list);
static int
__zpl_xattr_user_get(struct inode *ip, const char *name,
void *value, size_t size)
{
int error;
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") == 0)
return (-EINVAL);
#endif
if (ZFS_XA_NS_PREFIX_FORBIDDEN(name))
return (-EINVAL);
if (!(ITOZSB(ip)->z_flags & ZSB_XATTR))
return (-EOPNOTSUPP);
/*
* Try to look up the name with the namespace prefix first for
* compatibility with xattrs from this platform. If that fails,
* try again without the namespace prefix for compatibility with
* other platforms.
*/
char *xattr_name = kmem_asprintf("%s%s", XATTR_USER_PREFIX, name);
error = zpl_xattr_get(ip, xattr_name, value, size);
kmem_strfree(xattr_name);
if (error == -ENODATA)
error = zpl_xattr_get(ip, name, value, size);
return (error);
}
ZPL_XATTR_GET_WRAPPER(zpl_xattr_user_get);
static int
-__zpl_xattr_user_set(struct user_namespace *user_ns,
+__zpl_xattr_user_set(zidmap_t *user_ns,
struct inode *ip, const char *name,
const void *value, size_t size, int flags)
{
(void) user_ns;
int error = 0;
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") == 0)
return (-EINVAL);
#endif
if (ZFS_XA_NS_PREFIX_FORBIDDEN(name))
return (-EINVAL);
if (!(ITOZSB(ip)->z_flags & ZSB_XATTR))
return (-EOPNOTSUPP);
/*
* Remove alternate compat version of the xattr so we only set the
* version specified by the zfs_xattr_compat tunable.
*
* The following flags must be handled correctly:
*
* XATTR_CREATE: fail if xattr already exists
* XATTR_REPLACE: fail if xattr does not exist
*/
char *prefixed_name = kmem_asprintf("%s%s", XATTR_USER_PREFIX, name);
const char *clear_name, *set_name;
if (zfs_xattr_compat) {
clear_name = prefixed_name;
set_name = name;
} else {
clear_name = name;
set_name = prefixed_name;
}
/*
* Clear the old value with the alternative name format, if it exists.
*/
error = zpl_xattr_set(ip, clear_name, NULL, 0, flags);
/*
* XATTR_CREATE was specified and we failed to clear the xattr
* because it already exists. Stop here.
*/
if (error == -EEXIST)
goto out;
/*
* If XATTR_REPLACE was specified and we succeeded to clear
* an xattr, we don't need to replace anything when setting
* the new value. If we failed with -ENODATA that's fine,
* there was nothing to be cleared and we can ignore the error.
*/
if (error == 0)
flags &= ~XATTR_REPLACE;
/*
* Set the new value with the configured name format.
*/
error = zpl_xattr_set(ip, set_name, value, size, flags);
out:
kmem_strfree(prefixed_name);
return (error);
}
ZPL_XATTR_SET_WRAPPER(zpl_xattr_user_set);
static xattr_handler_t zpl_xattr_user_handler =
{
.prefix = XATTR_USER_PREFIX,
.list = zpl_xattr_user_list,
.get = zpl_xattr_user_get,
.set = zpl_xattr_user_set,
};
/*
* Trusted extended attributes
*
* "Trusted extended attributes are visible and accessible only to
* processes that have the CAP_SYS_ADMIN capability. Attributes in this
* class are used to implement mechanisms in user space (i.e., outside
* the kernel) which keep information in extended attributes to which
* ordinary processes should not have access." - xattr(7)
*/
static int
__zpl_xattr_trusted_list(struct inode *ip, char *list, size_t list_size,
const char *name, size_t name_len)
{
return (capable(CAP_SYS_ADMIN));
}
ZPL_XATTR_LIST_WRAPPER(zpl_xattr_trusted_list);
static int
__zpl_xattr_trusted_get(struct inode *ip, const char *name,
void *value, size_t size)
{
char *xattr_name;
int error;
if (!capable(CAP_SYS_ADMIN))
return (-EACCES);
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") == 0)
return (-EINVAL);
#endif
xattr_name = kmem_asprintf("%s%s", XATTR_TRUSTED_PREFIX, name);
error = zpl_xattr_get(ip, xattr_name, value, size);
kmem_strfree(xattr_name);
return (error);
}
ZPL_XATTR_GET_WRAPPER(zpl_xattr_trusted_get);
static int
-__zpl_xattr_trusted_set(struct user_namespace *user_ns,
+__zpl_xattr_trusted_set(zidmap_t *user_ns,
struct inode *ip, const char *name,
const void *value, size_t size, int flags)
{
(void) user_ns;
char *xattr_name;
int error;
if (!capable(CAP_SYS_ADMIN))
return (-EACCES);
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") == 0)
return (-EINVAL);
#endif
xattr_name = kmem_asprintf("%s%s", XATTR_TRUSTED_PREFIX, name);
error = zpl_xattr_set(ip, xattr_name, value, size, flags);
kmem_strfree(xattr_name);
return (error);
}
ZPL_XATTR_SET_WRAPPER(zpl_xattr_trusted_set);
static xattr_handler_t zpl_xattr_trusted_handler = {
.prefix = XATTR_TRUSTED_PREFIX,
.list = zpl_xattr_trusted_list,
.get = zpl_xattr_trusted_get,
.set = zpl_xattr_trusted_set,
};
/*
* Extended security attributes
*
* "The security attribute namespace is used by kernel security modules,
* such as Security Enhanced Linux, and also to implement file
* capabilities (see capabilities(7)). Read and write access
* permissions to security attributes depend on the policy implemented
* for each security attribute by the security module. When no security
* module is loaded, all processes have read access to extended security
* attributes, and write access is limited to processes that have the
* CAP_SYS_ADMIN capability." - xattr(7)
*/
static int
__zpl_xattr_security_list(struct inode *ip, char *list, size_t list_size,
const char *name, size_t name_len)
{
return (1);
}
ZPL_XATTR_LIST_WRAPPER(zpl_xattr_security_list);
static int
__zpl_xattr_security_get(struct inode *ip, const char *name,
void *value, size_t size)
{
char *xattr_name;
int error;
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") == 0)
return (-EINVAL);
#endif
xattr_name = kmem_asprintf("%s%s", XATTR_SECURITY_PREFIX, name);
error = zpl_xattr_get(ip, xattr_name, value, size);
kmem_strfree(xattr_name);
return (error);
}
ZPL_XATTR_GET_WRAPPER(zpl_xattr_security_get);
static int
-__zpl_xattr_security_set(struct user_namespace *user_ns,
+__zpl_xattr_security_set(zidmap_t *user_ns,
struct inode *ip, const char *name,
const void *value, size_t size, int flags)
{
(void) user_ns;
char *xattr_name;
int error;
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") == 0)
return (-EINVAL);
#endif
xattr_name = kmem_asprintf("%s%s", XATTR_SECURITY_PREFIX, name);
error = zpl_xattr_set(ip, xattr_name, value, size, flags);
kmem_strfree(xattr_name);
return (error);
}
ZPL_XATTR_SET_WRAPPER(zpl_xattr_security_set);
static int
zpl_xattr_security_init_impl(struct inode *ip, const struct xattr *xattrs,
void *fs_info)
{
const struct xattr *xattr;
int error = 0;
for (xattr = xattrs; xattr->name != NULL; xattr++) {
error = __zpl_xattr_security_set(NULL, ip,
xattr->name, xattr->value, xattr->value_len, 0);
if (error < 0)
break;
}
return (error);
}
int
zpl_xattr_security_init(struct inode *ip, struct inode *dip,
const struct qstr *qstr)
{
return security_inode_init_security(ip, dip, qstr,
&zpl_xattr_security_init_impl, NULL);
}
/*
* Security xattr namespace handlers.
*/
static xattr_handler_t zpl_xattr_security_handler = {
.prefix = XATTR_SECURITY_PREFIX,
.list = zpl_xattr_security_list,
.get = zpl_xattr_security_get,
.set = zpl_xattr_security_set,
};
/*
* Extended system attributes
*
* "Extended system attributes are used by the kernel to store system
* objects such as Access Control Lists. Read and write access permissions
* to system attributes depend on the policy implemented for each system
* attribute implemented by filesystems in the kernel." - xattr(7)
*/
#ifdef CONFIG_FS_POSIX_ACL
static int
zpl_set_acl_impl(struct inode *ip, struct posix_acl *acl, int type)
{
char *name, *value = NULL;
int error = 0;
size_t size = 0;
if (S_ISLNK(ip->i_mode))
return (-EOPNOTSUPP);
switch (type) {
case ACL_TYPE_ACCESS:
name = XATTR_NAME_POSIX_ACL_ACCESS;
if (acl) {
umode_t mode = ip->i_mode;
error = posix_acl_equiv_mode(acl, &mode);
if (error < 0) {
return (error);
} else {
/*
* The mode bits will have been set by
* ->zfs_setattr()->zfs_acl_chmod_setattr()
* using the ZFS ACL conversion. If they
* differ from the Posix ACL conversion dirty
* the inode to write the Posix mode bits.
*/
if (ip->i_mode != mode) {
ip->i_mode = ITOZ(ip)->z_mode = mode;
ip->i_ctime = current_time(ip);
zfs_mark_inode_dirty(ip);
}
if (error == 0)
acl = NULL;
}
}
break;
case ACL_TYPE_DEFAULT:
name = XATTR_NAME_POSIX_ACL_DEFAULT;
if (!S_ISDIR(ip->i_mode))
return (acl ? -EACCES : 0);
break;
default:
return (-EINVAL);
}
if (acl) {
size = posix_acl_xattr_size(acl->a_count);
value = kmem_alloc(size, KM_SLEEP);
error = zpl_acl_to_xattr(acl, value, size);
if (error < 0) {
kmem_free(value, size);
return (error);
}
}
error = zpl_xattr_set(ip, name, value, size, 0);
if (value)
kmem_free(value, size);
if (!error) {
if (acl)
zpl_set_cached_acl(ip, type, acl);
else
zpl_forget_cached_acl(ip, type);
}
return (error);
}
#ifdef HAVE_SET_ACL
int
#ifdef HAVE_SET_ACL_USERNS
zpl_set_acl(struct user_namespace *userns, struct inode *ip,
struct posix_acl *acl, int type)
+#elif defined(HAVE_SET_ACL_IDMAP_DENTRY)
+zpl_set_acl(struct mnt_idmap *userns, struct dentry *dentry,
+ struct posix_acl *acl, int type)
#elif defined(HAVE_SET_ACL_USERNS_DENTRY_ARG2)
zpl_set_acl(struct user_namespace *userns, struct dentry *dentry,
struct posix_acl *acl, int type)
#else
zpl_set_acl(struct inode *ip, struct posix_acl *acl, int type)
#endif /* HAVE_SET_ACL_USERNS */
{
#ifdef HAVE_SET_ACL_USERNS_DENTRY_ARG2
return (zpl_set_acl_impl(d_inode(dentry), acl, type));
+#elif defined(HAVE_SET_ACL_IDMAP_DENTRY)
+ return (zpl_set_acl_impl(d_inode(dentry), acl, type));
#else
return (zpl_set_acl_impl(ip, acl, type));
#endif /* HAVE_SET_ACL_USERNS_DENTRY_ARG2 */
}
#endif /* HAVE_SET_ACL */
static struct posix_acl *
zpl_get_acl_impl(struct inode *ip, int type)
{
struct posix_acl *acl;
void *value = NULL;
char *name;
/*
* As of Linux 3.14, the kernel get_acl will check this for us.
* Also as of Linux 4.7, comparing against ACL_NOT_CACHED is wrong
* as the kernel get_acl will set it to temporary sentinel value.
*/
#ifndef HAVE_KERNEL_GET_ACL_HANDLE_CACHE
acl = get_cached_acl(ip, type);
if (acl != ACL_NOT_CACHED)
return (acl);
#endif
switch (type) {
case ACL_TYPE_ACCESS:
name = XATTR_NAME_POSIX_ACL_ACCESS;
break;
case ACL_TYPE_DEFAULT:
name = XATTR_NAME_POSIX_ACL_DEFAULT;
break;
default:
return (ERR_PTR(-EINVAL));
}
int size = zpl_xattr_get(ip, name, NULL, 0);
if (size > 0) {
value = kmem_alloc(size, KM_SLEEP);
size = zpl_xattr_get(ip, name, value, size);
}
if (size > 0) {
acl = zpl_acl_from_xattr(value, size);
} else if (size == -ENODATA || size == -ENOSYS) {
acl = NULL;
} else {
acl = ERR_PTR(-EIO);
}
if (size > 0)
kmem_free(value, size);
/* As of Linux 4.7, the kernel get_acl will set this for us */
#ifndef HAVE_KERNEL_GET_ACL_HANDLE_CACHE
if (!IS_ERR(acl))
zpl_set_cached_acl(ip, type, acl);
#endif
return (acl);
}
#if defined(HAVE_GET_ACL_RCU) || defined(HAVE_GET_INODE_ACL)
struct posix_acl *
zpl_get_acl(struct inode *ip, int type, bool rcu)
{
if (rcu)
return (ERR_PTR(-ECHILD));
return (zpl_get_acl_impl(ip, type));
}
#elif defined(HAVE_GET_ACL)
struct posix_acl *
zpl_get_acl(struct inode *ip, int type)
{
return (zpl_get_acl_impl(ip, type));
}
#else
#error "Unsupported iops->get_acl() implementation"
#endif /* HAVE_GET_ACL_RCU */
int
zpl_init_acl(struct inode *ip, struct inode *dir)
{
struct posix_acl *acl = NULL;
int error = 0;
if (ITOZSB(ip)->z_acl_type != ZFS_ACLTYPE_POSIX)
return (0);
if (!S_ISLNK(ip->i_mode)) {
acl = zpl_get_acl_impl(dir, ACL_TYPE_DEFAULT);
if (IS_ERR(acl))
return (PTR_ERR(acl));
if (!acl) {
ITOZ(ip)->z_mode = (ip->i_mode &= ~current_umask());
ip->i_ctime = current_time(ip);
zfs_mark_inode_dirty(ip);
return (0);
}
}
if (acl) {
umode_t mode;
if (S_ISDIR(ip->i_mode)) {
error = zpl_set_acl_impl(ip, acl, ACL_TYPE_DEFAULT);
if (error)
goto out;
}
mode = ip->i_mode;
error = __posix_acl_create(&acl, GFP_KERNEL, &mode);
if (error >= 0) {
ip->i_mode = ITOZ(ip)->z_mode = mode;
zfs_mark_inode_dirty(ip);
if (error > 0) {
error = zpl_set_acl_impl(ip, acl,
ACL_TYPE_ACCESS);
}
}
}
out:
zpl_posix_acl_release(acl);
return (error);
}
int
zpl_chmod_acl(struct inode *ip)
{
struct posix_acl *acl;
int error;
if (ITOZSB(ip)->z_acl_type != ZFS_ACLTYPE_POSIX)
return (0);
if (S_ISLNK(ip->i_mode))
return (-EOPNOTSUPP);
acl = zpl_get_acl_impl(ip, ACL_TYPE_ACCESS);
if (IS_ERR(acl) || !acl)
return (PTR_ERR(acl));
error = __posix_acl_chmod(&acl, GFP_KERNEL, ip->i_mode);
if (!error)
error = zpl_set_acl_impl(ip, acl, ACL_TYPE_ACCESS);
zpl_posix_acl_release(acl);
return (error);
}
static int
__zpl_xattr_acl_list_access(struct inode *ip, char *list, size_t list_size,
const char *name, size_t name_len)
{
char *xattr_name = XATTR_NAME_POSIX_ACL_ACCESS;
size_t xattr_size = sizeof (XATTR_NAME_POSIX_ACL_ACCESS);
if (ITOZSB(ip)->z_acl_type != ZFS_ACLTYPE_POSIX)
return (0);
if (list && xattr_size <= list_size)
memcpy(list, xattr_name, xattr_size);
return (xattr_size);
}
ZPL_XATTR_LIST_WRAPPER(zpl_xattr_acl_list_access);
static int
__zpl_xattr_acl_list_default(struct inode *ip, char *list, size_t list_size,
const char *name, size_t name_len)
{
char *xattr_name = XATTR_NAME_POSIX_ACL_DEFAULT;
size_t xattr_size = sizeof (XATTR_NAME_POSIX_ACL_DEFAULT);
if (ITOZSB(ip)->z_acl_type != ZFS_ACLTYPE_POSIX)
return (0);
if (list && xattr_size <= list_size)
memcpy(list, xattr_name, xattr_size);
return (xattr_size);
}
ZPL_XATTR_LIST_WRAPPER(zpl_xattr_acl_list_default);
static int
__zpl_xattr_acl_get_access(struct inode *ip, const char *name,
void *buffer, size_t size)
{
struct posix_acl *acl;
int type = ACL_TYPE_ACCESS;
int error;
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") != 0)
return (-EINVAL);
#endif
if (ITOZSB(ip)->z_acl_type != ZFS_ACLTYPE_POSIX)
return (-EOPNOTSUPP);
acl = zpl_get_acl_impl(ip, type);
if (IS_ERR(acl))
return (PTR_ERR(acl));
if (acl == NULL)
return (-ENODATA);
error = zpl_acl_to_xattr(acl, buffer, size);
zpl_posix_acl_release(acl);
return (error);
}
ZPL_XATTR_GET_WRAPPER(zpl_xattr_acl_get_access);
static int
__zpl_xattr_acl_get_default(struct inode *ip, const char *name,
void *buffer, size_t size)
{
struct posix_acl *acl;
int type = ACL_TYPE_DEFAULT;
int error;
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") != 0)
return (-EINVAL);
#endif
if (ITOZSB(ip)->z_acl_type != ZFS_ACLTYPE_POSIX)
return (-EOPNOTSUPP);
acl = zpl_get_acl_impl(ip, type);
if (IS_ERR(acl))
return (PTR_ERR(acl));
if (acl == NULL)
return (-ENODATA);
error = zpl_acl_to_xattr(acl, buffer, size);
zpl_posix_acl_release(acl);
return (error);
}
ZPL_XATTR_GET_WRAPPER(zpl_xattr_acl_get_default);
static int
-__zpl_xattr_acl_set_access(struct user_namespace *mnt_ns,
+__zpl_xattr_acl_set_access(zidmap_t *mnt_ns,
struct inode *ip, const char *name,
const void *value, size_t size, int flags)
{
struct posix_acl *acl;
int type = ACL_TYPE_ACCESS;
int error = 0;
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") != 0)
return (-EINVAL);
#endif
if (ITOZSB(ip)->z_acl_type != ZFS_ACLTYPE_POSIX)
return (-EOPNOTSUPP);
-#if defined(HAVE_XATTR_SET_USERNS)
+#if defined(HAVE_XATTR_SET_USERNS) || defined(HAVE_XATTR_SET_IDMAP)
if (!zpl_inode_owner_or_capable(mnt_ns, ip))
return (-EPERM);
#else
(void) mnt_ns;
- if (!zpl_inode_owner_or_capable(kcred->user_ns, ip))
+ if (!zpl_inode_owner_or_capable(zfs_init_idmap, ip))
return (-EPERM);
#endif
if (value) {
acl = zpl_acl_from_xattr(value, size);
if (IS_ERR(acl))
return (PTR_ERR(acl));
else if (acl) {
error = zpl_posix_acl_valid(ip, acl);
if (error) {
zpl_posix_acl_release(acl);
return (error);
}
}
} else {
acl = NULL;
}
error = zpl_set_acl_impl(ip, acl, type);
zpl_posix_acl_release(acl);
return (error);
}
ZPL_XATTR_SET_WRAPPER(zpl_xattr_acl_set_access);
static int
-__zpl_xattr_acl_set_default(struct user_namespace *mnt_ns,
+__zpl_xattr_acl_set_default(zidmap_t *mnt_ns,
struct inode *ip, const char *name,
const void *value, size_t size, int flags)
{
struct posix_acl *acl;
int type = ACL_TYPE_DEFAULT;
int error = 0;
/* xattr_resolve_name will do this for us if this is defined */
#ifndef HAVE_XATTR_HANDLER_NAME
if (strcmp(name, "") != 0)
return (-EINVAL);
#endif
if (ITOZSB(ip)->z_acl_type != ZFS_ACLTYPE_POSIX)
return (-EOPNOTSUPP);
-#if defined(HAVE_XATTR_SET_USERNS)
+#if defined(HAVE_XATTR_SET_USERNS) || defined(HAVE_XATTR_SET_IDMAP)
if (!zpl_inode_owner_or_capable(mnt_ns, ip))
return (-EPERM);
#else
(void) mnt_ns;
- if (!zpl_inode_owner_or_capable(kcred->user_ns, ip))
+ if (!zpl_inode_owner_or_capable(zfs_init_idmap, ip))
return (-EPERM);
#endif
if (value) {
acl = zpl_acl_from_xattr(value, size);
if (IS_ERR(acl))
return (PTR_ERR(acl));
else if (acl) {
error = zpl_posix_acl_valid(ip, acl);
if (error) {
zpl_posix_acl_release(acl);
return (error);
}
}
} else {
acl = NULL;
}
error = zpl_set_acl_impl(ip, acl, type);
zpl_posix_acl_release(acl);
return (error);
}
ZPL_XATTR_SET_WRAPPER(zpl_xattr_acl_set_default);
/*
* ACL access xattr namespace handlers.
*
* Use .name instead of .prefix when available. xattr_resolve_name will match
* whole name and reject anything that has .name only as prefix.
*/
static xattr_handler_t zpl_xattr_acl_access_handler = {
#ifdef HAVE_XATTR_HANDLER_NAME
.name = XATTR_NAME_POSIX_ACL_ACCESS,
#else
.prefix = XATTR_NAME_POSIX_ACL_ACCESS,
#endif
.list = zpl_xattr_acl_list_access,
.get = zpl_xattr_acl_get_access,
.set = zpl_xattr_acl_set_access,
#if defined(HAVE_XATTR_LIST_SIMPLE) || \
defined(HAVE_XATTR_LIST_DENTRY) || \
defined(HAVE_XATTR_LIST_HANDLER)
.flags = ACL_TYPE_ACCESS,
#endif
};
/*
* ACL default xattr namespace handlers.
*
* Use .name instead of .prefix when available. xattr_resolve_name will match
* whole name and reject anything that has .name only as prefix.
*/
static xattr_handler_t zpl_xattr_acl_default_handler = {
#ifdef HAVE_XATTR_HANDLER_NAME
.name = XATTR_NAME_POSIX_ACL_DEFAULT,
#else
.prefix = XATTR_NAME_POSIX_ACL_DEFAULT,
#endif
.list = zpl_xattr_acl_list_default,
.get = zpl_xattr_acl_get_default,
.set = zpl_xattr_acl_set_default,
#if defined(HAVE_XATTR_LIST_SIMPLE) || \
defined(HAVE_XATTR_LIST_DENTRY) || \
defined(HAVE_XATTR_LIST_HANDLER)
.flags = ACL_TYPE_DEFAULT,
#endif
};
#endif /* CONFIG_FS_POSIX_ACL */
xattr_handler_t *zpl_xattr_handlers[] = {
&zpl_xattr_security_handler,
&zpl_xattr_trusted_handler,
&zpl_xattr_user_handler,
#ifdef CONFIG_FS_POSIX_ACL
&zpl_xattr_acl_access_handler,
&zpl_xattr_acl_default_handler,
#endif /* CONFIG_FS_POSIX_ACL */
NULL
};
static const struct xattr_handler *
zpl_xattr_handler(const char *name)
{
if (strncmp(name, XATTR_USER_PREFIX,
XATTR_USER_PREFIX_LEN) == 0)
return (&zpl_xattr_user_handler);
if (strncmp(name, XATTR_TRUSTED_PREFIX,
XATTR_TRUSTED_PREFIX_LEN) == 0)
return (&zpl_xattr_trusted_handler);
if (strncmp(name, XATTR_SECURITY_PREFIX,
XATTR_SECURITY_PREFIX_LEN) == 0)
return (&zpl_xattr_security_handler);
#ifdef CONFIG_FS_POSIX_ACL
if (strncmp(name, XATTR_NAME_POSIX_ACL_ACCESS,
sizeof (XATTR_NAME_POSIX_ACL_ACCESS)) == 0)
return (&zpl_xattr_acl_access_handler);
if (strncmp(name, XATTR_NAME_POSIX_ACL_DEFAULT,
sizeof (XATTR_NAME_POSIX_ACL_DEFAULT)) == 0)
return (&zpl_xattr_acl_default_handler);
#endif /* CONFIG_FS_POSIX_ACL */
return (NULL);
}
static enum xattr_permission
zpl_xattr_permission(xattr_filldir_t *xf, const char *name, int name_len)
{
const struct xattr_handler *handler;
struct dentry *d __maybe_unused = xf->dentry;
enum xattr_permission perm = XAPERM_ALLOW;
handler = zpl_xattr_handler(name);
if (handler == NULL) {
/* Do not expose FreeBSD system namespace xattrs. */
if (ZFS_XA_NS_PREFIX_MATCH(FREEBSD, name))
return (XAPERM_DENY);
/*
* Anything that doesn't match a known namespace gets put in the
* user namespace for compatibility with other platforms.
*/
perm = XAPERM_COMPAT;
handler = &zpl_xattr_user_handler;
}
if (handler->list) {
#if defined(HAVE_XATTR_LIST_SIMPLE)
if (!handler->list(d))
return (XAPERM_DENY);
#elif defined(HAVE_XATTR_LIST_DENTRY)
if (!handler->list(d, NULL, 0, name, name_len, 0))
return (XAPERM_DENY);
#elif defined(HAVE_XATTR_LIST_HANDLER)
if (!handler->list(handler, d, NULL, 0, name, name_len))
return (XAPERM_DENY);
#endif
}
return (perm);
}
#if defined(CONFIG_FS_POSIX_ACL) && \
(!defined(HAVE_POSIX_ACL_RELEASE) || \
defined(HAVE_POSIX_ACL_RELEASE_GPL_ONLY))
struct acl_rel_struct {
struct acl_rel_struct *next;
struct posix_acl *acl;
clock_t time;
};
#define ACL_REL_GRACE (60*HZ)
#define ACL_REL_WINDOW (1*HZ)
#define ACL_REL_SCHED (ACL_REL_GRACE+ACL_REL_WINDOW)
/*
* Lockless multi-producer single-consumer fifo list.
* Nodes are added to tail and removed from head. Tail pointer is our
* synchronization point. It always points to the next pointer of the last
* node, or head if list is empty.
*/
static struct acl_rel_struct *acl_rel_head = NULL;
static struct acl_rel_struct **acl_rel_tail = &acl_rel_head;
static void
zpl_posix_acl_free(void *arg)
{
struct acl_rel_struct *freelist = NULL;
struct acl_rel_struct *a;
clock_t new_time;
boolean_t refire = B_FALSE;
ASSERT3P(acl_rel_head, !=, NULL);
while (acl_rel_head) {
a = acl_rel_head;
if (ddi_get_lbolt() - a->time >= ACL_REL_GRACE) {
/*
* If a is the last node we need to reset tail, but we
* need to use cmpxchg to make sure it is still the
* last node.
*/
if (acl_rel_tail == &a->next) {
acl_rel_head = NULL;
if (cmpxchg(&acl_rel_tail, &a->next,
&acl_rel_head) == &a->next) {
ASSERT3P(a->next, ==, NULL);
a->next = freelist;
freelist = a;
break;
}
}
/*
* a is not last node, make sure next pointer is set
* by the adder and advance the head.
*/
while (READ_ONCE(a->next) == NULL)
cpu_relax();
acl_rel_head = a->next;
a->next = freelist;
freelist = a;
} else {
/*
* a is still in grace period. We are responsible to
* reschedule the free task, since adder will only do
* so if list is empty.
*/
new_time = a->time + ACL_REL_SCHED;
refire = B_TRUE;
break;
}
}
if (refire)
taskq_dispatch_delay(system_delay_taskq, zpl_posix_acl_free,
NULL, TQ_SLEEP, new_time);
while (freelist) {
a = freelist;
freelist = a->next;
kfree(a->acl);
kmem_free(a, sizeof (struct acl_rel_struct));
}
}
void
zpl_posix_acl_release_impl(struct posix_acl *acl)
{
struct acl_rel_struct *a, **prev;
a = kmem_alloc(sizeof (struct acl_rel_struct), KM_SLEEP);
a->next = NULL;
a->acl = acl;
a->time = ddi_get_lbolt();
/* atomically points tail to us and get the previous tail */
prev = xchg(&acl_rel_tail, &a->next);
ASSERT3P(*prev, ==, NULL);
*prev = a;
/* if it was empty before, schedule the free task */
if (prev == &acl_rel_head)
taskq_dispatch_delay(system_delay_taskq, zpl_posix_acl_free,
NULL, TQ_SLEEP, ddi_get_lbolt() + ACL_REL_SCHED);
}
#endif
ZFS_MODULE_PARAM(zfs, zfs_, xattr_compat, INT, ZMOD_RW,
"Use legacy ZFS xattr naming for writing new user namespace xattrs");
diff --git a/sys/contrib/openzfs/module/zcommon/zfeature_common.c b/sys/contrib/openzfs/module/zcommon/zfeature_common.c
index 6fe1da8ed46f..4c9b7ed72a0f 100644
--- a/sys/contrib/openzfs/module/zcommon/zfeature_common.c
+++ b/sys/contrib/openzfs/module/zcommon/zfeature_common.c
@@ -1,745 +1,751 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2011, 2018 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
* Copyright (c) 2013, Joyent, Inc. All rights reserved.
* Copyright (c) 2014, Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, Klara Inc.
* Copyright (c) 2019, Allan Jude
*/
#ifndef _KERNEL
#include <errno.h>
#include <string.h>
#include <dirent.h>
#include <search.h>
#include <sys/stat.h>
#endif
#include <sys/debug.h>
#include <sys/fs/zfs.h>
#include <sys/inttypes.h>
#include <sys/types.h>
#include <sys/param.h>
#include <sys/zfs_sysfs.h>
#include "zfeature_common.h"
/*
* Set to disable all feature checks while opening pools, allowing pools with
* unsupported features to be opened. Set for testing only.
*/
boolean_t zfeature_checks_disable = B_FALSE;
zfeature_info_t spa_feature_table[SPA_FEATURES];
/*
* Valid characters for feature guids. This list is mainly for aesthetic
* purposes and could be expanded in the future. There are different allowed
* characters in the guids reverse dns portion (before the colon) and its
* short name (after the colon).
*/
static int
valid_char(char c, boolean_t after_colon)
{
return ((c >= 'a' && c <= 'z') ||
(c >= '0' && c <= '9') ||
(after_colon && c == '_') ||
(!after_colon && (c == '.' || c == '-')));
}
/*
* Every feature guid must contain exactly one colon which separates a reverse
* dns organization name from the feature's "short" name (e.g.
* "com.company:feature_name").
*/
boolean_t
zfeature_is_valid_guid(const char *name)
{
int i;
boolean_t has_colon = B_FALSE;
i = 0;
while (name[i] != '\0') {
char c = name[i++];
if (c == ':') {
if (has_colon)
return (B_FALSE);
has_colon = B_TRUE;
continue;
}
if (!valid_char(c, has_colon))
return (B_FALSE);
}
return (has_colon);
}
boolean_t
zfeature_is_supported(const char *guid)
{
if (zfeature_checks_disable)
return (B_TRUE);
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t *feature = &spa_feature_table[i];
if (!feature->fi_zfs_mod_supported)
continue;
if (strcmp(guid, feature->fi_guid) == 0)
return (B_TRUE);
}
return (B_FALSE);
}
int
zfeature_lookup_guid(const char *guid, spa_feature_t *res)
{
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t *feature = &spa_feature_table[i];
if (!feature->fi_zfs_mod_supported)
continue;
if (strcmp(guid, feature->fi_guid) == 0) {
if (res != NULL)
*res = i;
return (0);
}
}
return (ENOENT);
}
int
zfeature_lookup_name(const char *name, spa_feature_t *res)
{
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t *feature = &spa_feature_table[i];
if (!feature->fi_zfs_mod_supported)
continue;
if (strcmp(name, feature->fi_uname) == 0) {
if (res != NULL)
*res = i;
return (0);
}
}
return (ENOENT);
}
boolean_t
zfeature_depends_on(spa_feature_t fid, spa_feature_t check)
{
zfeature_info_t *feature = &spa_feature_table[fid];
for (int i = 0; feature->fi_depends[i] != SPA_FEATURE_NONE; i++) {
if (feature->fi_depends[i] == check)
return (B_TRUE);
}
return (B_FALSE);
}
static boolean_t
deps_contains_feature(const spa_feature_t *deps, const spa_feature_t feature)
{
for (int i = 0; deps[i] != SPA_FEATURE_NONE; i++)
if (deps[i] == feature)
return (B_TRUE);
return (B_FALSE);
}
#define STRCMP ((int(*)(const void *, const void *))&strcmp)
struct zfs_mod_supported_features {
void *tree;
boolean_t all_features;
};
struct zfs_mod_supported_features *
zfs_mod_list_supported(const char *scope)
{
#if defined(__FreeBSD__) || defined(_KERNEL) || defined(LIB_ZPOOL_BUILD)
(void) scope;
return (NULL);
#else
struct zfs_mod_supported_features *ret = calloc(1, sizeof (*ret));
if (ret == NULL)
return (NULL);
DIR *sysfs_dir = NULL;
char path[128];
if (snprintf(path, sizeof (path), "%s/%s",
ZFS_SYSFS_DIR, scope) < sizeof (path))
sysfs_dir = opendir(path);
if (sysfs_dir == NULL && errno == ENOENT) {
if (snprintf(path, sizeof (path), "%s/%s",
ZFS_SYSFS_ALT_DIR, scope) < sizeof (path))
sysfs_dir = opendir(path);
}
if (sysfs_dir == NULL) {
ret->all_features = errno == ENOENT &&
(access(ZFS_SYSFS_DIR, F_OK) == 0 ||
access(ZFS_SYSFS_ALT_DIR, F_OK) == 0);
return (ret);
}
struct dirent *node;
while ((node = readdir(sysfs_dir)) != NULL) {
if (strcmp(node->d_name, ".") == 0 ||
strcmp(node->d_name, "..") == 0)
continue;
char *name = strdup(node->d_name);
if (name == NULL) {
goto nomem;
}
if (tsearch(name, &ret->tree, STRCMP) == NULL) {
/*
* Don't bother checking for duplicate entries:
* we're iterating a single directory.
*/
free(name);
goto nomem;
}
}
end:
closedir(sysfs_dir);
return (ret);
nomem:
zfs_mod_list_supported_free(ret);
ret = NULL;
goto end;
#endif
}
void
zfs_mod_list_supported_free(struct zfs_mod_supported_features *list)
{
#if !defined(__FreeBSD__) && !defined(_KERNEL) && !defined(LIB_ZPOOL_BUILD)
if (list) {
tdestroy(list->tree, free);
free(list);
}
#else
(void) list;
#endif
}
#if !defined(_KERNEL) && !defined(LIB_ZPOOL_BUILD)
static boolean_t
zfs_mod_supported_impl(const char *scope, const char *name, const char *sysfs)
{
char path[128];
if (snprintf(path, sizeof (path), "%s%s%s%s%s", sysfs,
scope == NULL ? "" : "/", scope ?: "",
name == NULL ? "" : "/", name ?: "") < sizeof (path))
return (access(path, F_OK) == 0);
else
return (B_FALSE);
}
boolean_t
zfs_mod_supported(const char *scope, const char *name,
const struct zfs_mod_supported_features *sfeatures)
{
boolean_t supported;
if (sfeatures != NULL)
return (sfeatures->all_features ||
tfind(name, &sfeatures->tree, STRCMP));
/*
* Check both the primary and alternate sysfs locations to determine
* if the required functionality is supported.
*/
supported = (zfs_mod_supported_impl(scope, name, ZFS_SYSFS_DIR) ||
zfs_mod_supported_impl(scope, name, ZFS_SYSFS_ALT_DIR));
/*
* For backwards compatibility with kernel modules that predate
* supported feature/property checking. Report the feature/property
* as supported if the kernel module is loaded but the requested
* scope directory does not exist.
*/
if (supported == B_FALSE) {
if ((access(ZFS_SYSFS_DIR, F_OK) == 0 &&
!zfs_mod_supported_impl(scope, NULL, ZFS_SYSFS_DIR)) ||
(access(ZFS_SYSFS_ALT_DIR, F_OK) == 0 &&
!zfs_mod_supported_impl(scope, NULL, ZFS_SYSFS_ALT_DIR))) {
supported = B_TRUE;
}
}
return (supported);
}
#endif
static boolean_t
zfs_mod_supported_feature(const char *name,
const struct zfs_mod_supported_features *sfeatures)
{
/*
* The zfs module spa_feature_table[], whether in-kernel or in
* libzpool, always supports all the features. libzfs needs to
* query the running module, via sysfs, to determine which
* features are supported.
*
* The equivalent _can_ be done on FreeBSD by way of the sysctl
* tree, but this has not been done yet. Therefore, we return
* that all features are supported.
*/
#if defined(_KERNEL) || defined(LIB_ZPOOL_BUILD) || defined(__FreeBSD__)
(void) name, (void) sfeatures;
return (B_TRUE);
#else
return (zfs_mod_supported(ZFS_SYSFS_POOL_FEATURES, name, sfeatures));
#endif
}
static void
zfeature_register(spa_feature_t fid, const char *guid, const char *name,
const char *desc, zfeature_flags_t flags, zfeature_type_t type,
const spa_feature_t *deps,
const struct zfs_mod_supported_features *sfeatures)
{
zfeature_info_t *feature = &spa_feature_table[fid];
static const spa_feature_t nodeps[] = { SPA_FEATURE_NONE };
ASSERT(name != NULL);
ASSERT(desc != NULL);
ASSERT((flags & ZFEATURE_FLAG_READONLY_COMPAT) == 0 ||
(flags & ZFEATURE_FLAG_MOS) == 0);
ASSERT3U(fid, <, SPA_FEATURES);
ASSERT(zfeature_is_valid_guid(guid));
if (deps == NULL)
deps = nodeps;
VERIFY(((flags & ZFEATURE_FLAG_PER_DATASET) == 0) ||
(deps_contains_feature(deps, SPA_FEATURE_EXTENSIBLE_DATASET)));
feature->fi_feature = fid;
feature->fi_guid = guid;
feature->fi_uname = name;
feature->fi_desc = desc;
feature->fi_flags = flags;
feature->fi_type = type;
feature->fi_depends = deps;
feature->fi_zfs_mod_supported =
zfs_mod_supported_feature(guid, sfeatures);
}
/*
* Every feature has a GUID of the form com.example:feature_name. The
* reversed DNS name ensures that the feature's GUID is unique across all ZFS
* implementations. This allows companies to independently develop and
* release features. Examples include org.delphix and org.datto. Previously,
* features developed on one implementation have used that implementation's
* domain name (e.g. org.illumos and org.zfsonlinux). Use of the org.openzfs
* domain name is recommended for new features which are developed by the
* OpenZFS community and its platforms. This domain may optionally be used by
* companies developing features for initial release through an OpenZFS
* implementation. Use of the org.openzfs domain requires reserving the
* feature name in advance with the OpenZFS project.
*/
void
zpool_feature_init(void)
{
struct zfs_mod_supported_features *sfeatures =
zfs_mod_list_supported(ZFS_SYSFS_POOL_FEATURES);
zfeature_register(SPA_FEATURE_ASYNC_DESTROY,
"com.delphix:async_destroy", "async_destroy",
"Destroy filesystems asynchronously.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_EMPTY_BPOBJ,
"com.delphix:empty_bpobj", "empty_bpobj",
"Snapshots use less space.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_LZ4_COMPRESS,
"org.illumos:lz4_compress", "lz4_compress",
"LZ4 compression algorithm support.",
ZFEATURE_FLAG_ACTIVATE_ON_ENABLE, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_MULTI_VDEV_CRASH_DUMP,
"com.joyent:multi_vdev_crash_dump", "multi_vdev_crash_dump",
"Crash dumps to multiple vdev pools.",
0, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
zfeature_register(SPA_FEATURE_SPACEMAP_HISTOGRAM,
"com.delphix:spacemap_histogram", "spacemap_histogram",
"Spacemaps maintain space histograms.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_ENABLED_TXG,
"com.delphix:enabled_txg", "enabled_txg",
"Record txg at which a feature is enabled",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
{
static const spa_feature_t hole_birth_deps[] = {
SPA_FEATURE_ENABLED_TXG,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_HOLE_BIRTH,
"com.delphix:hole_birth", "hole_birth",
"Retain hole birth txg for more precise zfs send",
ZFEATURE_FLAG_MOS | ZFEATURE_FLAG_ACTIVATE_ON_ENABLE,
ZFEATURE_TYPE_BOOLEAN, hole_birth_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_POOL_CHECKPOINT,
"com.delphix:zpool_checkpoint", "zpool_checkpoint",
"Pool state can be checkpointed, allowing rewind later.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_SPACEMAP_V2,
"com.delphix:spacemap_v2", "spacemap_v2",
"Space maps representing large segments are more efficient.",
ZFEATURE_FLAG_READONLY_COMPAT | ZFEATURE_FLAG_ACTIVATE_ON_ENABLE,
ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
zfeature_register(SPA_FEATURE_EXTENSIBLE_DATASET,
"com.delphix:extensible_dataset", "extensible_dataset",
"Enhanced dataset functionality, used by other features.",
0, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
{
static const spa_feature_t bookmarks_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_BOOKMARKS,
"com.delphix:bookmarks", "bookmarks",
"\"zfs bookmark\" command",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
bookmarks_deps, sfeatures);
}
{
static const spa_feature_t filesystem_limits_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_FS_SS_LIMIT,
"com.joyent:filesystem_limits", "filesystem_limits",
"Filesystem and snapshot limits.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
filesystem_limits_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_EMBEDDED_DATA,
"com.delphix:embedded_data", "embedded_data",
"Blocks which compress very well use even less space.",
ZFEATURE_FLAG_MOS | ZFEATURE_FLAG_ACTIVATE_ON_ENABLE,
ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
{
static const spa_feature_t livelist_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_LIVELIST,
"com.delphix:livelist", "livelist",
"Improved clone deletion performance.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
livelist_deps, sfeatures);
}
{
static const spa_feature_t log_spacemap_deps[] = {
SPA_FEATURE_SPACEMAP_V2,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_LOG_SPACEMAP,
"com.delphix:log_spacemap", "log_spacemap",
"Log metaslab changes on a single spacemap and "
"flush them periodically.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
log_spacemap_deps, sfeatures);
}
{
static const spa_feature_t large_blocks_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_LARGE_BLOCKS,
"org.open-zfs:large_blocks", "large_blocks",
"Support for blocks larger than 128KB.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
large_blocks_deps, sfeatures);
}
{
static const spa_feature_t large_dnode_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_LARGE_DNODE,
"org.zfsonlinux:large_dnode", "large_dnode",
"Variable on-disk size of dnodes.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
large_dnode_deps, sfeatures);
}
{
static const spa_feature_t sha512_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_SHA512,
"org.illumos:sha512", "sha512",
"SHA-512/256 hash algorithm.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
sha512_deps, sfeatures);
}
{
static const spa_feature_t skein_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_SKEIN,
"org.illumos:skein", "skein",
"Skein hash algorithm.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
skein_deps, sfeatures);
}
{
static const spa_feature_t edonr_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_EDONR,
"org.illumos:edonr", "edonr",
"Edon-R hash algorithm.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
edonr_deps, sfeatures);
}
{
static const spa_feature_t redact_books_deps[] = {
SPA_FEATURE_BOOKMARK_V2,
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_BOOKMARKS,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_REDACTION_BOOKMARKS,
"com.delphix:redaction_bookmarks", "redaction_bookmarks",
"Support for bookmarks which store redaction lists for zfs "
"redacted send/recv.", 0, ZFEATURE_TYPE_BOOLEAN,
redact_books_deps, sfeatures);
}
{
static const spa_feature_t redact_datasets_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_REDACTED_DATASETS,
"com.delphix:redacted_datasets", "redacted_datasets",
"Support for redacted datasets, produced by receiving "
"a redacted zfs send stream.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_UINT64_ARRAY,
redact_datasets_deps, sfeatures);
}
{
static const spa_feature_t bookmark_written_deps[] = {
SPA_FEATURE_BOOKMARK_V2,
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_BOOKMARKS,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_BOOKMARK_WRITTEN,
"com.delphix:bookmark_written", "bookmark_written",
"Additional accounting, enabling the written#<bookmark> "
"property (space written since a bookmark), "
"and estimates of send stream sizes for incrementals from "
"bookmarks.",
0, ZFEATURE_TYPE_BOOLEAN, bookmark_written_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_DEVICE_REMOVAL,
"com.delphix:device_removal", "device_removal",
"Top-level vdevs can be removed, reducing logical pool size.",
ZFEATURE_FLAG_MOS, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
{
static const spa_feature_t obsolete_counts_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_DEVICE_REMOVAL,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_OBSOLETE_COUNTS,
"com.delphix:obsolete_counts", "obsolete_counts",
"Reduce memory used by removed devices when their blocks "
"are freed or remapped.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN,
obsolete_counts_deps, sfeatures);
}
{
static const spa_feature_t userobj_accounting_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_USEROBJ_ACCOUNTING,
"org.zfsonlinux:userobj_accounting", "userobj_accounting",
"User/Group object accounting.",
ZFEATURE_FLAG_READONLY_COMPAT | ZFEATURE_FLAG_PER_DATASET,
ZFEATURE_TYPE_BOOLEAN, userobj_accounting_deps, sfeatures);
}
{
static const spa_feature_t bookmark_v2_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_BOOKMARKS,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_BOOKMARK_V2,
"com.datto:bookmark_v2", "bookmark_v2",
"Support for larger bookmarks",
0, ZFEATURE_TYPE_BOOLEAN, bookmark_v2_deps, sfeatures);
}
{
static const spa_feature_t encryption_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_BOOKMARK_V2,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_ENCRYPTION,
"com.datto:encryption", "encryption",
"Support for dataset level encryption",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
encryption_deps, sfeatures);
}
{
static const spa_feature_t project_quota_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_PROJECT_QUOTA,
"org.zfsonlinux:project_quota", "project_quota",
"space/object accounting based on project ID.",
ZFEATURE_FLAG_READONLY_COMPAT | ZFEATURE_FLAG_PER_DATASET,
ZFEATURE_TYPE_BOOLEAN, project_quota_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_ALLOCATION_CLASSES,
"org.zfsonlinux:allocation_classes", "allocation_classes",
"Support for separate allocation classes.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_RESILVER_DEFER,
"com.datto:resilver_defer", "resilver_defer",
"Support for deferring new resilvers when one is already running.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
zfeature_register(SPA_FEATURE_DEVICE_REBUILD,
"org.openzfs:device_rebuild", "device_rebuild",
"Support for sequential mirror/dRAID device rebuilds",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
{
static const spa_feature_t zstd_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_ZSTD_COMPRESS,
"org.freebsd:zstd_compress", "zstd_compress",
"zstd compression algorithm support.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN, zstd_deps,
sfeatures);
}
zfeature_register(SPA_FEATURE_DRAID,
"org.openzfs:draid", "draid", "Support for distributed spare RAID",
ZFEATURE_FLAG_MOS, ZFEATURE_TYPE_BOOLEAN, NULL, sfeatures);
{
static const spa_feature_t zilsaxattr_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_ZILSAXATTR,
"org.openzfs:zilsaxattr", "zilsaxattr",
"Support for xattr=sa extended attribute logging in ZIL.",
ZFEATURE_FLAG_PER_DATASET | ZFEATURE_FLAG_READONLY_COMPAT,
ZFEATURE_TYPE_BOOLEAN, zilsaxattr_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_HEAD_ERRLOG,
"com.delphix:head_errlog", "head_errlog",
"Support for per-dataset on-disk error logs.",
ZFEATURE_FLAG_ACTIVATE_ON_ENABLE, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
{
static const spa_feature_t blake3_deps[] = {
SPA_FEATURE_EXTENSIBLE_DATASET,
SPA_FEATURE_NONE
};
zfeature_register(SPA_FEATURE_BLAKE3,
"org.openzfs:blake3", "blake3",
"BLAKE3 hash algorithm.",
ZFEATURE_FLAG_PER_DATASET, ZFEATURE_TYPE_BOOLEAN,
blake3_deps, sfeatures);
}
zfeature_register(SPA_FEATURE_BLOCK_CLONING,
"com.fudosecurity:block_cloning", "block_cloning",
"Support for block cloning via Block Reference Table.",
ZFEATURE_FLAG_READONLY_COMPAT, ZFEATURE_TYPE_BOOLEAN, NULL,
sfeatures);
+ zfeature_register(SPA_FEATURE_AVZ_V2,
+ "com.klarasystems:vdev_zaps_v2", "vdev_zaps_v2",
+ "Support for root vdev ZAP.",
+ ZFEATURE_FLAG_MOS, ZFEATURE_TYPE_BOOLEAN, NULL,
+ sfeatures);
+
zfs_mod_list_supported_free(sfeatures);
}
#if defined(_KERNEL)
EXPORT_SYMBOL(zfeature_lookup_guid);
EXPORT_SYMBOL(zfeature_lookup_name);
EXPORT_SYMBOL(zfeature_is_supported);
EXPORT_SYMBOL(zfeature_is_valid_guid);
EXPORT_SYMBOL(zfeature_depends_on);
EXPORT_SYMBOL(zpool_feature_init);
EXPORT_SYMBOL(spa_feature_table);
#endif
diff --git a/sys/contrib/openzfs/module/zfs/arc.c b/sys/contrib/openzfs/module/zfs/arc.c
index e32707bbe5c3..c50228a2682f 100644
--- a/sys/contrib/openzfs/module/zfs/arc.c
+++ b/sys/contrib/openzfs/module/zfs/arc.c
@@ -1,10919 +1,10919 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2018, Joyent, Inc.
* Copyright (c) 2011, 2020, Delphix. All rights reserved.
* Copyright (c) 2014, Saso Kiselkov. All rights reserved.
* Copyright (c) 2017, Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2019, loli10K <ezomori.nozomu@gmail.com>. All rights reserved.
* Copyright (c) 2020, George Amanakis. All rights reserved.
* Copyright (c) 2019, Klara Inc.
* Copyright (c) 2019, Allan Jude
* Copyright (c) 2020, The FreeBSD Foundation [1]
*
* [1] Portions of this software were developed by Allan Jude
* under sponsorship from the FreeBSD Foundation.
*/
/*
* DVA-based Adjustable Replacement Cache
*
* While much of the theory of operation used here is
* based on the self-tuning, low overhead replacement cache
* presented by Megiddo and Modha at FAST 2003, there are some
* significant differences:
*
* 1. The Megiddo and Modha model assumes any page is evictable.
* Pages in its cache cannot be "locked" into memory. This makes
* the eviction algorithm simple: evict the last page in the list.
* This also make the performance characteristics easy to reason
* about. Our cache is not so simple. At any given moment, some
* subset of the blocks in the cache are un-evictable because we
* have handed out a reference to them. Blocks are only evictable
* when there are no external references active. This makes
* eviction far more problematic: we choose to evict the evictable
* blocks that are the "lowest" in the list.
*
* There are times when it is not possible to evict the requested
* space. In these circumstances we are unable to adjust the cache
* size. To prevent the cache growing unbounded at these times we
* implement a "cache throttle" that slows the flow of new data
* into the cache until we can make space available.
*
* 2. The Megiddo and Modha model assumes a fixed cache size.
* Pages are evicted when the cache is full and there is a cache
* miss. Our model has a variable sized cache. It grows with
* high use, but also tries to react to memory pressure from the
* operating system: decreasing its size when system memory is
* tight.
*
* 3. The Megiddo and Modha model assumes a fixed page size. All
* elements of the cache are therefore exactly the same size. So
* when adjusting the cache size following a cache miss, its simply
* a matter of choosing a single page to evict. In our model, we
* have variable sized cache blocks (ranging from 512 bytes to
* 128K bytes). We therefore choose a set of blocks to evict to make
* space for a cache miss that approximates as closely as possible
* the space used by the new block.
*
* See also: "ARC: A Self-Tuning, Low Overhead Replacement Cache"
* by N. Megiddo & D. Modha, FAST 2003
*/
/*
* The locking model:
*
* A new reference to a cache buffer can be obtained in two
* ways: 1) via a hash table lookup using the DVA as a key,
* or 2) via one of the ARC lists. The arc_read() interface
* uses method 1, while the internal ARC algorithms for
* adjusting the cache use method 2. We therefore provide two
* types of locks: 1) the hash table lock array, and 2) the
* ARC list locks.
*
* Buffers do not have their own mutexes, rather they rely on the
* hash table mutexes for the bulk of their protection (i.e. most
* fields in the arc_buf_hdr_t are protected by these mutexes).
*
* buf_hash_find() returns the appropriate mutex (held) when it
* locates the requested buffer in the hash table. It returns
* NULL for the mutex if the buffer was not in the table.
*
* buf_hash_remove() expects the appropriate hash mutex to be
* already held before it is invoked.
*
* Each ARC state also has a mutex which is used to protect the
* buffer list associated with the state. When attempting to
* obtain a hash table lock while holding an ARC list lock you
* must use: mutex_tryenter() to avoid deadlock. Also note that
* the active state mutex must be held before the ghost state mutex.
*
* It as also possible to register a callback which is run when the
* metadata limit is reached and no buffers can be safely evicted. In
* this case the arc user should drop a reference on some arc buffers so
* they can be reclaimed. For example, when using the ZPL each dentry
* holds a references on a znode. These dentries must be pruned before
* the arc buffer holding the znode can be safely evicted.
*
* Note that the majority of the performance stats are manipulated
* with atomic operations.
*
* The L2ARC uses the l2ad_mtx on each vdev for the following:
*
* - L2ARC buflist creation
* - L2ARC buflist eviction
* - L2ARC write completion, which walks L2ARC buflists
* - ARC header destruction, as it removes from L2ARC buflists
* - ARC header release, as it removes from L2ARC buflists
*/
/*
* ARC operation:
*
* Every block that is in the ARC is tracked by an arc_buf_hdr_t structure.
* This structure can point either to a block that is still in the cache or to
* one that is only accessible in an L2 ARC device, or it can provide
* information about a block that was recently evicted. If a block is
* only accessible in the L2ARC, then the arc_buf_hdr_t only has enough
* information to retrieve it from the L2ARC device. This information is
* stored in the l2arc_buf_hdr_t sub-structure of the arc_buf_hdr_t. A block
* that is in this state cannot access the data directly.
*
* Blocks that are actively being referenced or have not been evicted
* are cached in the L1ARC. The L1ARC (l1arc_buf_hdr_t) is a structure within
* the arc_buf_hdr_t that will point to the data block in memory. A block can
* only be read by a consumer if it has an l1arc_buf_hdr_t. The L1ARC
* caches data in two ways -- in a list of ARC buffers (arc_buf_t) and
* also in the arc_buf_hdr_t's private physical data block pointer (b_pabd).
*
* The L1ARC's data pointer may or may not be uncompressed. The ARC has the
* ability to store the physical data (b_pabd) associated with the DVA of the
* arc_buf_hdr_t. Since the b_pabd is a copy of the on-disk physical block,
* it will match its on-disk compression characteristics. This behavior can be
* disabled by setting 'zfs_compressed_arc_enabled' to B_FALSE. When the
* compressed ARC functionality is disabled, the b_pabd will point to an
* uncompressed version of the on-disk data.
*
* Data in the L1ARC is not accessed by consumers of the ARC directly. Each
* arc_buf_hdr_t can have multiple ARC buffers (arc_buf_t) which reference it.
* Each ARC buffer (arc_buf_t) is being actively accessed by a specific ARC
* consumer. The ARC will provide references to this data and will keep it
* cached until it is no longer in use. The ARC caches only the L1ARC's physical
* data block and will evict any arc_buf_t that is no longer referenced. The
* amount of memory consumed by the arc_buf_ts' data buffers can be seen via the
* "overhead_size" kstat.
*
* Depending on the consumer, an arc_buf_t can be requested in uncompressed or
* compressed form. The typical case is that consumers will want uncompressed
* data, and when that happens a new data buffer is allocated where the data is
* decompressed for them to use. Currently the only consumer who wants
* compressed arc_buf_t's is "zfs send", when it streams data exactly as it
* exists on disk. When this happens, the arc_buf_t's data buffer is shared
* with the arc_buf_hdr_t.
*
* Here is a diagram showing an arc_buf_hdr_t referenced by two arc_buf_t's. The
* first one is owned by a compressed send consumer (and therefore references
* the same compressed data buffer as the arc_buf_hdr_t) and the second could be
* used by any other consumer (and has its own uncompressed copy of the data
* buffer).
*
* arc_buf_hdr_t
* +-----------+
* | fields |
* | common to |
* | L1- and |
* | L2ARC |
* +-----------+
* | l2arc_buf_hdr_t
* | |
* +-----------+
* | l1arc_buf_hdr_t
* | | arc_buf_t
* | b_buf +------------>+-----------+ arc_buf_t
* | b_pabd +-+ |b_next +---->+-----------+
* +-----------+ | |-----------| |b_next +-->NULL
* | |b_comp = T | +-----------+
* | |b_data +-+ |b_comp = F |
* | +-----------+ | |b_data +-+
* +->+------+ | +-----------+ |
* compressed | | | |
* data | |<--------------+ | uncompressed
* +------+ compressed, | data
* shared +-->+------+
* data | |
* | |
* +------+
*
* When a consumer reads a block, the ARC must first look to see if the
* arc_buf_hdr_t is cached. If the hdr is cached then the ARC allocates a new
* arc_buf_t and either copies uncompressed data into a new data buffer from an
* existing uncompressed arc_buf_t, decompresses the hdr's b_pabd buffer into a
* new data buffer, or shares the hdr's b_pabd buffer, depending on whether the
* hdr is compressed and the desired compression characteristics of the
* arc_buf_t consumer. If the arc_buf_t ends up sharing data with the
* arc_buf_hdr_t and both of them are uncompressed then the arc_buf_t must be
* the last buffer in the hdr's b_buf list, however a shared compressed buf can
* be anywhere in the hdr's list.
*
* The diagram below shows an example of an uncompressed ARC hdr that is
* sharing its data with an arc_buf_t (note that the shared uncompressed buf is
* the last element in the buf list):
*
* arc_buf_hdr_t
* +-----------+
* | |
* | |
* | |
* +-----------+
* l2arc_buf_hdr_t| |
* | |
* +-----------+
* l1arc_buf_hdr_t| |
* | | arc_buf_t (shared)
* | b_buf +------------>+---------+ arc_buf_t
* | | |b_next +---->+---------+
* | b_pabd +-+ |---------| |b_next +-->NULL
* +-----------+ | | | +---------+
* | |b_data +-+ | |
* | +---------+ | |b_data +-+
* +->+------+ | +---------+ |
* | | | |
* uncompressed | | | |
* data +------+ | |
* ^ +->+------+ |
* | uncompressed | | |
* | data | | |
* | +------+ |
* +---------------------------------+
*
* Writing to the ARC requires that the ARC first discard the hdr's b_pabd
* since the physical block is about to be rewritten. The new data contents
* will be contained in the arc_buf_t. As the I/O pipeline performs the write,
* it may compress the data before writing it to disk. The ARC will be called
* with the transformed data and will memcpy the transformed on-disk block into
* a newly allocated b_pabd. Writes are always done into buffers which have
* either been loaned (and hence are new and don't have other readers) or
* buffers which have been released (and hence have their own hdr, if there
* were originally other readers of the buf's original hdr). This ensures that
* the ARC only needs to update a single buf and its hdr after a write occurs.
*
* When the L2ARC is in use, it will also take advantage of the b_pabd. The
* L2ARC will always write the contents of b_pabd to the L2ARC. This means
* that when compressed ARC is enabled that the L2ARC blocks are identical
* to the on-disk block in the main data pool. This provides a significant
* advantage since the ARC can leverage the bp's checksum when reading from the
* L2ARC to determine if the contents are valid. However, if the compressed
* ARC is disabled, then the L2ARC's block must be transformed to look
* like the physical block in the main data pool before comparing the
* checksum and determining its validity.
*
* The L1ARC has a slightly different system for storing encrypted data.
* Raw (encrypted + possibly compressed) data has a few subtle differences from
* data that is just compressed. The biggest difference is that it is not
* possible to decrypt encrypted data (or vice-versa) if the keys aren't loaded.
* The other difference is that encryption cannot be treated as a suggestion.
* If a caller would prefer compressed data, but they actually wind up with
* uncompressed data the worst thing that could happen is there might be a
* performance hit. If the caller requests encrypted data, however, we must be
* sure they actually get it or else secret information could be leaked. Raw
* data is stored in hdr->b_crypt_hdr.b_rabd. An encrypted header, therefore,
* may have both an encrypted version and a decrypted version of its data at
* once. When a caller needs a raw arc_buf_t, it is allocated and the data is
* copied out of this header. To avoid complications with b_pabd, raw buffers
* cannot be shared.
*/
#include <sys/spa.h>
#include <sys/zio.h>
#include <sys/spa_impl.h>
#include <sys/zio_compress.h>
#include <sys/zio_checksum.h>
#include <sys/zfs_context.h>
#include <sys/arc.h>
#include <sys/zfs_refcount.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
#include <sys/dsl_pool.h>
#include <sys/multilist.h>
#include <sys/abd.h>
#include <sys/zil.h>
#include <sys/fm/fs/zfs.h>
#include <sys/callb.h>
#include <sys/kstat.h>
#include <sys/zthr.h>
#include <zfs_fletcher.h>
#include <sys/arc_impl.h>
#include <sys/trace_zfs.h>
#include <sys/aggsum.h>
#include <sys/wmsum.h>
#include <cityhash.h>
#include <sys/vdev_trim.h>
#include <sys/zfs_racct.h>
#include <sys/zstd/zstd.h>
#ifndef _KERNEL
/* set with ZFS_DEBUG=watch, to enable watchpoints on frozen buffers */
boolean_t arc_watch = B_FALSE;
#endif
/*
* This thread's job is to keep enough free memory in the system, by
* calling arc_kmem_reap_soon() plus arc_reduce_target_size(), which improves
* arc_available_memory().
*/
static zthr_t *arc_reap_zthr;
/*
* This thread's job is to keep arc_size under arc_c, by calling
* arc_evict(), which improves arc_is_overflowing().
*/
static zthr_t *arc_evict_zthr;
static arc_buf_hdr_t **arc_state_evict_markers;
static int arc_state_evict_marker_count;
static kmutex_t arc_evict_lock;
static boolean_t arc_evict_needed = B_FALSE;
static clock_t arc_last_uncached_flush;
/*
* Count of bytes evicted since boot.
*/
static uint64_t arc_evict_count;
/*
* List of arc_evict_waiter_t's, representing threads waiting for the
* arc_evict_count to reach specific values.
*/
static list_t arc_evict_waiters;
/*
* When arc_is_overflowing(), arc_get_data_impl() waits for this percent of
* the requested amount of data to be evicted. For example, by default for
* every 2KB that's evicted, 1KB of it may be "reused" by a new allocation.
* Since this is above 100%, it ensures that progress is made towards getting
* arc_size under arc_c. Since this is finite, it ensures that allocations
* can still happen, even during the potentially long time that arc_size is
* more than arc_c.
*/
static uint_t zfs_arc_eviction_pct = 200;
/*
* The number of headers to evict in arc_evict_state_impl() before
* dropping the sublist lock and evicting from another sublist. A lower
* value means we're more likely to evict the "correct" header (i.e. the
* oldest header in the arc state), but comes with higher overhead
* (i.e. more invocations of arc_evict_state_impl()).
*/
static uint_t zfs_arc_evict_batch_limit = 10;
/* number of seconds before growing cache again */
uint_t arc_grow_retry = 5;
/*
* Minimum time between calls to arc_kmem_reap_soon().
*/
static const int arc_kmem_cache_reap_retry_ms = 1000;
/* shift of arc_c for calculating overflow limit in arc_get_data_impl */
static int zfs_arc_overflow_shift = 8;
/* log2(fraction of arc to reclaim) */
uint_t arc_shrink_shift = 7;
/* percent of pagecache to reclaim arc to */
#ifdef _KERNEL
uint_t zfs_arc_pc_percent = 0;
#endif
/*
* log2(fraction of ARC which must be free to allow growing).
* I.e. If there is less than arc_c >> arc_no_grow_shift free memory,
* when reading a new block into the ARC, we will evict an equal-sized block
* from the ARC.
*
* This must be less than arc_shrink_shift, so that when we shrink the ARC,
* we will still not allow it to grow.
*/
uint_t arc_no_grow_shift = 5;
/*
* minimum lifespan of a prefetch block in clock ticks
* (initialized in arc_init())
*/
static uint_t arc_min_prefetch_ms;
static uint_t arc_min_prescient_prefetch_ms;
/*
* If this percent of memory is free, don't throttle.
*/
uint_t arc_lotsfree_percent = 10;
/*
* The arc has filled available memory and has now warmed up.
*/
boolean_t arc_warm;
/*
* These tunables are for performance analysis.
*/
uint64_t zfs_arc_max = 0;
uint64_t zfs_arc_min = 0;
static uint64_t zfs_arc_dnode_limit = 0;
static uint_t zfs_arc_dnode_reduce_percent = 10;
static uint_t zfs_arc_grow_retry = 0;
static uint_t zfs_arc_shrink_shift = 0;
uint_t zfs_arc_average_blocksize = 8 * 1024; /* 8KB */
/*
* ARC dirty data constraints for arc_tempreserve_space() throttle:
* * total dirty data limit
* * anon block dirty limit
* * each pool's anon allowance
*/
static const unsigned long zfs_arc_dirty_limit_percent = 50;
static const unsigned long zfs_arc_anon_limit_percent = 25;
static const unsigned long zfs_arc_pool_dirty_percent = 20;
/*
* Enable or disable compressed arc buffers.
*/
int zfs_compressed_arc_enabled = B_TRUE;
/*
* Balance between metadata and data on ghost hits. Values above 100
* increase metadata caching by proportionally reducing effect of ghost
* data hits on target data/metadata rate.
*/
static uint_t zfs_arc_meta_balance = 500;
/*
* Percentage that can be consumed by dnodes of ARC meta buffers.
*/
static uint_t zfs_arc_dnode_limit_percent = 10;
/*
* These tunables are Linux-specific
*/
static uint64_t zfs_arc_sys_free = 0;
static uint_t zfs_arc_min_prefetch_ms = 0;
static uint_t zfs_arc_min_prescient_prefetch_ms = 0;
static uint_t zfs_arc_lotsfree_percent = 10;
/*
* Number of arc_prune threads
*/
static int zfs_arc_prune_task_threads = 1;
/* The 7 states: */
arc_state_t ARC_anon;
arc_state_t ARC_mru;
arc_state_t ARC_mru_ghost;
arc_state_t ARC_mfu;
arc_state_t ARC_mfu_ghost;
arc_state_t ARC_l2c_only;
arc_state_t ARC_uncached;
arc_stats_t arc_stats = {
{ "hits", KSTAT_DATA_UINT64 },
{ "iohits", KSTAT_DATA_UINT64 },
{ "misses", KSTAT_DATA_UINT64 },
{ "demand_data_hits", KSTAT_DATA_UINT64 },
{ "demand_data_iohits", KSTAT_DATA_UINT64 },
{ "demand_data_misses", KSTAT_DATA_UINT64 },
{ "demand_metadata_hits", KSTAT_DATA_UINT64 },
{ "demand_metadata_iohits", KSTAT_DATA_UINT64 },
{ "demand_metadata_misses", KSTAT_DATA_UINT64 },
{ "prefetch_data_hits", KSTAT_DATA_UINT64 },
{ "prefetch_data_iohits", KSTAT_DATA_UINT64 },
{ "prefetch_data_misses", KSTAT_DATA_UINT64 },
{ "prefetch_metadata_hits", KSTAT_DATA_UINT64 },
{ "prefetch_metadata_iohits", KSTAT_DATA_UINT64 },
{ "prefetch_metadata_misses", KSTAT_DATA_UINT64 },
{ "mru_hits", KSTAT_DATA_UINT64 },
{ "mru_ghost_hits", KSTAT_DATA_UINT64 },
{ "mfu_hits", KSTAT_DATA_UINT64 },
{ "mfu_ghost_hits", KSTAT_DATA_UINT64 },
{ "uncached_hits", KSTAT_DATA_UINT64 },
{ "deleted", KSTAT_DATA_UINT64 },
{ "mutex_miss", KSTAT_DATA_UINT64 },
{ "access_skip", KSTAT_DATA_UINT64 },
{ "evict_skip", KSTAT_DATA_UINT64 },
{ "evict_not_enough", KSTAT_DATA_UINT64 },
{ "evict_l2_cached", KSTAT_DATA_UINT64 },
{ "evict_l2_eligible", KSTAT_DATA_UINT64 },
{ "evict_l2_eligible_mfu", KSTAT_DATA_UINT64 },
{ "evict_l2_eligible_mru", KSTAT_DATA_UINT64 },
{ "evict_l2_ineligible", KSTAT_DATA_UINT64 },
{ "evict_l2_skip", KSTAT_DATA_UINT64 },
{ "hash_elements", KSTAT_DATA_UINT64 },
{ "hash_elements_max", KSTAT_DATA_UINT64 },
{ "hash_collisions", KSTAT_DATA_UINT64 },
{ "hash_chains", KSTAT_DATA_UINT64 },
{ "hash_chain_max", KSTAT_DATA_UINT64 },
{ "meta", KSTAT_DATA_UINT64 },
{ "pd", KSTAT_DATA_UINT64 },
{ "pm", KSTAT_DATA_UINT64 },
{ "c", KSTAT_DATA_UINT64 },
{ "c_min", KSTAT_DATA_UINT64 },
{ "c_max", KSTAT_DATA_UINT64 },
{ "size", KSTAT_DATA_UINT64 },
{ "compressed_size", KSTAT_DATA_UINT64 },
{ "uncompressed_size", KSTAT_DATA_UINT64 },
{ "overhead_size", KSTAT_DATA_UINT64 },
{ "hdr_size", KSTAT_DATA_UINT64 },
{ "data_size", KSTAT_DATA_UINT64 },
{ "metadata_size", KSTAT_DATA_UINT64 },
{ "dbuf_size", KSTAT_DATA_UINT64 },
{ "dnode_size", KSTAT_DATA_UINT64 },
{ "bonus_size", KSTAT_DATA_UINT64 },
#if defined(COMPAT_FREEBSD11)
{ "other_size", KSTAT_DATA_UINT64 },
#endif
{ "anon_size", KSTAT_DATA_UINT64 },
{ "anon_data", KSTAT_DATA_UINT64 },
{ "anon_metadata", KSTAT_DATA_UINT64 },
{ "anon_evictable_data", KSTAT_DATA_UINT64 },
{ "anon_evictable_metadata", KSTAT_DATA_UINT64 },
{ "mru_size", KSTAT_DATA_UINT64 },
{ "mru_data", KSTAT_DATA_UINT64 },
{ "mru_metadata", KSTAT_DATA_UINT64 },
{ "mru_evictable_data", KSTAT_DATA_UINT64 },
{ "mru_evictable_metadata", KSTAT_DATA_UINT64 },
{ "mru_ghost_size", KSTAT_DATA_UINT64 },
{ "mru_ghost_data", KSTAT_DATA_UINT64 },
{ "mru_ghost_metadata", KSTAT_DATA_UINT64 },
{ "mru_ghost_evictable_data", KSTAT_DATA_UINT64 },
{ "mru_ghost_evictable_metadata", KSTAT_DATA_UINT64 },
{ "mfu_size", KSTAT_DATA_UINT64 },
{ "mfu_data", KSTAT_DATA_UINT64 },
{ "mfu_metadata", KSTAT_DATA_UINT64 },
{ "mfu_evictable_data", KSTAT_DATA_UINT64 },
{ "mfu_evictable_metadata", KSTAT_DATA_UINT64 },
{ "mfu_ghost_size", KSTAT_DATA_UINT64 },
{ "mfu_ghost_data", KSTAT_DATA_UINT64 },
{ "mfu_ghost_metadata", KSTAT_DATA_UINT64 },
{ "mfu_ghost_evictable_data", KSTAT_DATA_UINT64 },
{ "mfu_ghost_evictable_metadata", KSTAT_DATA_UINT64 },
{ "uncached_size", KSTAT_DATA_UINT64 },
{ "uncached_data", KSTAT_DATA_UINT64 },
{ "uncached_metadata", KSTAT_DATA_UINT64 },
{ "uncached_evictable_data", KSTAT_DATA_UINT64 },
{ "uncached_evictable_metadata", KSTAT_DATA_UINT64 },
{ "l2_hits", KSTAT_DATA_UINT64 },
{ "l2_misses", KSTAT_DATA_UINT64 },
{ "l2_prefetch_asize", KSTAT_DATA_UINT64 },
{ "l2_mru_asize", KSTAT_DATA_UINT64 },
{ "l2_mfu_asize", KSTAT_DATA_UINT64 },
{ "l2_bufc_data_asize", KSTAT_DATA_UINT64 },
{ "l2_bufc_metadata_asize", KSTAT_DATA_UINT64 },
{ "l2_feeds", KSTAT_DATA_UINT64 },
{ "l2_rw_clash", KSTAT_DATA_UINT64 },
{ "l2_read_bytes", KSTAT_DATA_UINT64 },
{ "l2_write_bytes", KSTAT_DATA_UINT64 },
{ "l2_writes_sent", KSTAT_DATA_UINT64 },
{ "l2_writes_done", KSTAT_DATA_UINT64 },
{ "l2_writes_error", KSTAT_DATA_UINT64 },
{ "l2_writes_lock_retry", KSTAT_DATA_UINT64 },
{ "l2_evict_lock_retry", KSTAT_DATA_UINT64 },
{ "l2_evict_reading", KSTAT_DATA_UINT64 },
{ "l2_evict_l1cached", KSTAT_DATA_UINT64 },
{ "l2_free_on_write", KSTAT_DATA_UINT64 },
{ "l2_abort_lowmem", KSTAT_DATA_UINT64 },
{ "l2_cksum_bad", KSTAT_DATA_UINT64 },
{ "l2_io_error", KSTAT_DATA_UINT64 },
{ "l2_size", KSTAT_DATA_UINT64 },
{ "l2_asize", KSTAT_DATA_UINT64 },
{ "l2_hdr_size", KSTAT_DATA_UINT64 },
{ "l2_log_blk_writes", KSTAT_DATA_UINT64 },
{ "l2_log_blk_avg_asize", KSTAT_DATA_UINT64 },
{ "l2_log_blk_asize", KSTAT_DATA_UINT64 },
{ "l2_log_blk_count", KSTAT_DATA_UINT64 },
{ "l2_data_to_meta_ratio", KSTAT_DATA_UINT64 },
{ "l2_rebuild_success", KSTAT_DATA_UINT64 },
{ "l2_rebuild_unsupported", KSTAT_DATA_UINT64 },
{ "l2_rebuild_io_errors", KSTAT_DATA_UINT64 },
{ "l2_rebuild_dh_errors", KSTAT_DATA_UINT64 },
{ "l2_rebuild_cksum_lb_errors", KSTAT_DATA_UINT64 },
{ "l2_rebuild_lowmem", KSTAT_DATA_UINT64 },
{ "l2_rebuild_size", KSTAT_DATA_UINT64 },
{ "l2_rebuild_asize", KSTAT_DATA_UINT64 },
{ "l2_rebuild_bufs", KSTAT_DATA_UINT64 },
{ "l2_rebuild_bufs_precached", KSTAT_DATA_UINT64 },
{ "l2_rebuild_log_blks", KSTAT_DATA_UINT64 },
{ "memory_throttle_count", KSTAT_DATA_UINT64 },
{ "memory_direct_count", KSTAT_DATA_UINT64 },
{ "memory_indirect_count", KSTAT_DATA_UINT64 },
{ "memory_all_bytes", KSTAT_DATA_UINT64 },
{ "memory_free_bytes", KSTAT_DATA_UINT64 },
{ "memory_available_bytes", KSTAT_DATA_INT64 },
{ "arc_no_grow", KSTAT_DATA_UINT64 },
{ "arc_tempreserve", KSTAT_DATA_UINT64 },
{ "arc_loaned_bytes", KSTAT_DATA_UINT64 },
{ "arc_prune", KSTAT_DATA_UINT64 },
{ "arc_meta_used", KSTAT_DATA_UINT64 },
{ "arc_dnode_limit", KSTAT_DATA_UINT64 },
{ "async_upgrade_sync", KSTAT_DATA_UINT64 },
{ "predictive_prefetch", KSTAT_DATA_UINT64 },
{ "demand_hit_predictive_prefetch", KSTAT_DATA_UINT64 },
{ "demand_iohit_predictive_prefetch", KSTAT_DATA_UINT64 },
{ "prescient_prefetch", KSTAT_DATA_UINT64 },
{ "demand_hit_prescient_prefetch", KSTAT_DATA_UINT64 },
{ "demand_iohit_prescient_prefetch", KSTAT_DATA_UINT64 },
{ "arc_need_free", KSTAT_DATA_UINT64 },
{ "arc_sys_free", KSTAT_DATA_UINT64 },
{ "arc_raw_size", KSTAT_DATA_UINT64 },
{ "cached_only_in_progress", KSTAT_DATA_UINT64 },
{ "abd_chunk_waste_size", KSTAT_DATA_UINT64 },
};
arc_sums_t arc_sums;
#define ARCSTAT_MAX(stat, val) { \
uint64_t m; \
while ((val) > (m = arc_stats.stat.value.ui64) && \
(m != atomic_cas_64(&arc_stats.stat.value.ui64, m, (val)))) \
continue; \
}
/*
* We define a macro to allow ARC hits/misses to be easily broken down by
* two separate conditions, giving a total of four different subtypes for
* each of hits and misses (so eight statistics total).
*/
#define ARCSTAT_CONDSTAT(cond1, stat1, notstat1, cond2, stat2, notstat2, stat) \
if (cond1) { \
if (cond2) { \
ARCSTAT_BUMP(arcstat_##stat1##_##stat2##_##stat); \
} else { \
ARCSTAT_BUMP(arcstat_##stat1##_##notstat2##_##stat); \
} \
} else { \
if (cond2) { \
ARCSTAT_BUMP(arcstat_##notstat1##_##stat2##_##stat); \
} else { \
ARCSTAT_BUMP(arcstat_##notstat1##_##notstat2##_##stat);\
} \
}
/*
* This macro allows us to use kstats as floating averages. Each time we
* update this kstat, we first factor it and the update value by
* ARCSTAT_AVG_FACTOR to shrink the new value's contribution to the overall
* average. This macro assumes that integer loads and stores are atomic, but
* is not safe for multiple writers updating the kstat in parallel (only the
* last writer's update will remain).
*/
#define ARCSTAT_F_AVG_FACTOR 3
#define ARCSTAT_F_AVG(stat, value) \
do { \
uint64_t x = ARCSTAT(stat); \
x = x - x / ARCSTAT_F_AVG_FACTOR + \
(value) / ARCSTAT_F_AVG_FACTOR; \
ARCSTAT(stat) = x; \
} while (0)
static kstat_t *arc_ksp;
/*
* There are several ARC variables that are critical to export as kstats --
* but we don't want to have to grovel around in the kstat whenever we wish to
* manipulate them. For these variables, we therefore define them to be in
* terms of the statistic variable. This assures that we are not introducing
* the possibility of inconsistency by having shadow copies of the variables,
* while still allowing the code to be readable.
*/
#define arc_tempreserve ARCSTAT(arcstat_tempreserve)
#define arc_loaned_bytes ARCSTAT(arcstat_loaned_bytes)
#define arc_dnode_limit ARCSTAT(arcstat_dnode_limit) /* max size for dnodes */
#define arc_need_free ARCSTAT(arcstat_need_free) /* waiting to be evicted */
hrtime_t arc_growtime;
list_t arc_prune_list;
kmutex_t arc_prune_mtx;
taskq_t *arc_prune_taskq;
#define GHOST_STATE(state) \
((state) == arc_mru_ghost || (state) == arc_mfu_ghost || \
(state) == arc_l2c_only)
#define HDR_IN_HASH_TABLE(hdr) ((hdr)->b_flags & ARC_FLAG_IN_HASH_TABLE)
#define HDR_IO_IN_PROGRESS(hdr) ((hdr)->b_flags & ARC_FLAG_IO_IN_PROGRESS)
#define HDR_IO_ERROR(hdr) ((hdr)->b_flags & ARC_FLAG_IO_ERROR)
#define HDR_PREFETCH(hdr) ((hdr)->b_flags & ARC_FLAG_PREFETCH)
#define HDR_PRESCIENT_PREFETCH(hdr) \
((hdr)->b_flags & ARC_FLAG_PRESCIENT_PREFETCH)
#define HDR_COMPRESSION_ENABLED(hdr) \
((hdr)->b_flags & ARC_FLAG_COMPRESSED_ARC)
#define HDR_L2CACHE(hdr) ((hdr)->b_flags & ARC_FLAG_L2CACHE)
#define HDR_UNCACHED(hdr) ((hdr)->b_flags & ARC_FLAG_UNCACHED)
#define HDR_L2_READING(hdr) \
(((hdr)->b_flags & ARC_FLAG_IO_IN_PROGRESS) && \
((hdr)->b_flags & ARC_FLAG_HAS_L2HDR))
#define HDR_L2_WRITING(hdr) ((hdr)->b_flags & ARC_FLAG_L2_WRITING)
#define HDR_L2_EVICTED(hdr) ((hdr)->b_flags & ARC_FLAG_L2_EVICTED)
#define HDR_L2_WRITE_HEAD(hdr) ((hdr)->b_flags & ARC_FLAG_L2_WRITE_HEAD)
#define HDR_PROTECTED(hdr) ((hdr)->b_flags & ARC_FLAG_PROTECTED)
#define HDR_NOAUTH(hdr) ((hdr)->b_flags & ARC_FLAG_NOAUTH)
#define HDR_SHARED_DATA(hdr) ((hdr)->b_flags & ARC_FLAG_SHARED_DATA)
#define HDR_ISTYPE_METADATA(hdr) \
((hdr)->b_flags & ARC_FLAG_BUFC_METADATA)
#define HDR_ISTYPE_DATA(hdr) (!HDR_ISTYPE_METADATA(hdr))
#define HDR_HAS_L1HDR(hdr) ((hdr)->b_flags & ARC_FLAG_HAS_L1HDR)
#define HDR_HAS_L2HDR(hdr) ((hdr)->b_flags & ARC_FLAG_HAS_L2HDR)
#define HDR_HAS_RABD(hdr) \
(HDR_HAS_L1HDR(hdr) && HDR_PROTECTED(hdr) && \
(hdr)->b_crypt_hdr.b_rabd != NULL)
#define HDR_ENCRYPTED(hdr) \
(HDR_PROTECTED(hdr) && DMU_OT_IS_ENCRYPTED((hdr)->b_crypt_hdr.b_ot))
#define HDR_AUTHENTICATED(hdr) \
(HDR_PROTECTED(hdr) && !DMU_OT_IS_ENCRYPTED((hdr)->b_crypt_hdr.b_ot))
/* For storing compression mode in b_flags */
#define HDR_COMPRESS_OFFSET (highbit64(ARC_FLAG_COMPRESS_0) - 1)
#define HDR_GET_COMPRESS(hdr) ((enum zio_compress)BF32_GET((hdr)->b_flags, \
HDR_COMPRESS_OFFSET, SPA_COMPRESSBITS))
#define HDR_SET_COMPRESS(hdr, cmp) BF32_SET((hdr)->b_flags, \
HDR_COMPRESS_OFFSET, SPA_COMPRESSBITS, (cmp));
#define ARC_BUF_LAST(buf) ((buf)->b_next == NULL)
#define ARC_BUF_SHARED(buf) ((buf)->b_flags & ARC_BUF_FLAG_SHARED)
#define ARC_BUF_COMPRESSED(buf) ((buf)->b_flags & ARC_BUF_FLAG_COMPRESSED)
#define ARC_BUF_ENCRYPTED(buf) ((buf)->b_flags & ARC_BUF_FLAG_ENCRYPTED)
/*
* Other sizes
*/
#define HDR_FULL_CRYPT_SIZE ((int64_t)sizeof (arc_buf_hdr_t))
#define HDR_FULL_SIZE ((int64_t)offsetof(arc_buf_hdr_t, b_crypt_hdr))
#define HDR_L2ONLY_SIZE ((int64_t)offsetof(arc_buf_hdr_t, b_l1hdr))
/*
* Hash table routines
*/
#define BUF_LOCKS 2048
typedef struct buf_hash_table {
uint64_t ht_mask;
arc_buf_hdr_t **ht_table;
kmutex_t ht_locks[BUF_LOCKS] ____cacheline_aligned;
} buf_hash_table_t;
static buf_hash_table_t buf_hash_table;
#define BUF_HASH_INDEX(spa, dva, birth) \
(buf_hash(spa, dva, birth) & buf_hash_table.ht_mask)
#define BUF_HASH_LOCK(idx) (&buf_hash_table.ht_locks[idx & (BUF_LOCKS-1)])
#define HDR_LOCK(hdr) \
(BUF_HASH_LOCK(BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth)))
uint64_t zfs_crc64_table[256];
/*
* Level 2 ARC
*/
#define L2ARC_WRITE_SIZE (8 * 1024 * 1024) /* initial write max */
#define L2ARC_HEADROOM 2 /* num of writes */
/*
* If we discover during ARC scan any buffers to be compressed, we boost
* our headroom for the next scanning cycle by this percentage multiple.
*/
#define L2ARC_HEADROOM_BOOST 200
#define L2ARC_FEED_SECS 1 /* caching interval secs */
#define L2ARC_FEED_MIN_MS 200 /* min caching interval ms */
/*
* We can feed L2ARC from two states of ARC buffers, mru and mfu,
* and each of the state has two types: data and metadata.
*/
#define L2ARC_FEED_TYPES 4
/* L2ARC Performance Tunables */
uint64_t l2arc_write_max = L2ARC_WRITE_SIZE; /* def max write size */
uint64_t l2arc_write_boost = L2ARC_WRITE_SIZE; /* extra warmup write */
uint64_t l2arc_headroom = L2ARC_HEADROOM; /* # of dev writes */
uint64_t l2arc_headroom_boost = L2ARC_HEADROOM_BOOST;
uint64_t l2arc_feed_secs = L2ARC_FEED_SECS; /* interval seconds */
uint64_t l2arc_feed_min_ms = L2ARC_FEED_MIN_MS; /* min interval msecs */
int l2arc_noprefetch = B_TRUE; /* don't cache prefetch bufs */
int l2arc_feed_again = B_TRUE; /* turbo warmup */
int l2arc_norw = B_FALSE; /* no reads during writes */
static uint_t l2arc_meta_percent = 33; /* limit on headers size */
/*
* L2ARC Internals
*/
static list_t L2ARC_dev_list; /* device list */
static list_t *l2arc_dev_list; /* device list pointer */
static kmutex_t l2arc_dev_mtx; /* device list mutex */
static l2arc_dev_t *l2arc_dev_last; /* last device used */
static list_t L2ARC_free_on_write; /* free after write buf list */
static list_t *l2arc_free_on_write; /* free after write list ptr */
static kmutex_t l2arc_free_on_write_mtx; /* mutex for list */
static uint64_t l2arc_ndev; /* number of devices */
typedef struct l2arc_read_callback {
arc_buf_hdr_t *l2rcb_hdr; /* read header */
blkptr_t l2rcb_bp; /* original blkptr */
zbookmark_phys_t l2rcb_zb; /* original bookmark */
int l2rcb_flags; /* original flags */
abd_t *l2rcb_abd; /* temporary buffer */
} l2arc_read_callback_t;
typedef struct l2arc_data_free {
/* protected by l2arc_free_on_write_mtx */
abd_t *l2df_abd;
size_t l2df_size;
arc_buf_contents_t l2df_type;
list_node_t l2df_list_node;
} l2arc_data_free_t;
typedef enum arc_fill_flags {
ARC_FILL_LOCKED = 1 << 0, /* hdr lock is held */
ARC_FILL_COMPRESSED = 1 << 1, /* fill with compressed data */
ARC_FILL_ENCRYPTED = 1 << 2, /* fill with encrypted data */
ARC_FILL_NOAUTH = 1 << 3, /* don't attempt to authenticate */
ARC_FILL_IN_PLACE = 1 << 4 /* fill in place (special case) */
} arc_fill_flags_t;
typedef enum arc_ovf_level {
ARC_OVF_NONE, /* ARC within target size. */
ARC_OVF_SOME, /* ARC is slightly overflowed. */
ARC_OVF_SEVERE /* ARC is severely overflowed. */
} arc_ovf_level_t;
static kmutex_t l2arc_feed_thr_lock;
static kcondvar_t l2arc_feed_thr_cv;
static uint8_t l2arc_thread_exit;
static kmutex_t l2arc_rebuild_thr_lock;
static kcondvar_t l2arc_rebuild_thr_cv;
enum arc_hdr_alloc_flags {
ARC_HDR_ALLOC_RDATA = 0x1,
ARC_HDR_USE_RESERVE = 0x4,
ARC_HDR_ALLOC_LINEAR = 0x8,
};
static abd_t *arc_get_data_abd(arc_buf_hdr_t *, uint64_t, const void *, int);
static void *arc_get_data_buf(arc_buf_hdr_t *, uint64_t, const void *);
static void arc_get_data_impl(arc_buf_hdr_t *, uint64_t, const void *, int);
static void arc_free_data_abd(arc_buf_hdr_t *, abd_t *, uint64_t, const void *);
static void arc_free_data_buf(arc_buf_hdr_t *, void *, uint64_t, const void *);
static void arc_free_data_impl(arc_buf_hdr_t *hdr, uint64_t size,
const void *tag);
static void arc_hdr_free_abd(arc_buf_hdr_t *, boolean_t);
static void arc_hdr_alloc_abd(arc_buf_hdr_t *, int);
static void arc_hdr_destroy(arc_buf_hdr_t *);
static void arc_access(arc_buf_hdr_t *, arc_flags_t, boolean_t);
static void arc_buf_watch(arc_buf_t *);
static void arc_change_state(arc_state_t *, arc_buf_hdr_t *);
static arc_buf_contents_t arc_buf_type(arc_buf_hdr_t *);
static uint32_t arc_bufc_to_flags(arc_buf_contents_t);
static inline void arc_hdr_set_flags(arc_buf_hdr_t *hdr, arc_flags_t flags);
static inline void arc_hdr_clear_flags(arc_buf_hdr_t *hdr, arc_flags_t flags);
static boolean_t l2arc_write_eligible(uint64_t, arc_buf_hdr_t *);
static void l2arc_read_done(zio_t *);
static void l2arc_do_free_on_write(void);
static void l2arc_hdr_arcstats_update(arc_buf_hdr_t *hdr, boolean_t incr,
boolean_t state_only);
#define l2arc_hdr_arcstats_increment(hdr) \
l2arc_hdr_arcstats_update((hdr), B_TRUE, B_FALSE)
#define l2arc_hdr_arcstats_decrement(hdr) \
l2arc_hdr_arcstats_update((hdr), B_FALSE, B_FALSE)
#define l2arc_hdr_arcstats_increment_state(hdr) \
l2arc_hdr_arcstats_update((hdr), B_TRUE, B_TRUE)
#define l2arc_hdr_arcstats_decrement_state(hdr) \
l2arc_hdr_arcstats_update((hdr), B_FALSE, B_TRUE)
/*
* l2arc_exclude_special : A zfs module parameter that controls whether buffers
* present on special vdevs are eligibile for caching in L2ARC. If
* set to 1, exclude dbufs on special vdevs from being cached to
* L2ARC.
*/
int l2arc_exclude_special = 0;
/*
* l2arc_mfuonly : A ZFS module parameter that controls whether only MFU
* metadata and data are cached from ARC into L2ARC.
*/
static int l2arc_mfuonly = 0;
/*
* L2ARC TRIM
* l2arc_trim_ahead : A ZFS module parameter that controls how much ahead of
* the current write size (l2arc_write_max) we should TRIM if we
* have filled the device. It is defined as a percentage of the
* write size. If set to 100 we trim twice the space required to
* accommodate upcoming writes. A minimum of 64MB will be trimmed.
* It also enables TRIM of the whole L2ARC device upon creation or
* addition to an existing pool or if the header of the device is
* invalid upon importing a pool or onlining a cache device. The
* default is 0, which disables TRIM on L2ARC altogether as it can
* put significant stress on the underlying storage devices. This
* will vary depending of how well the specific device handles
* these commands.
*/
static uint64_t l2arc_trim_ahead = 0;
/*
* Performance tuning of L2ARC persistence:
*
* l2arc_rebuild_enabled : A ZFS module parameter that controls whether adding
* an L2ARC device (either at pool import or later) will attempt
* to rebuild L2ARC buffer contents.
* l2arc_rebuild_blocks_min_l2size : A ZFS module parameter that controls
* whether log blocks are written to the L2ARC device. If the L2ARC
* device is less than 1GB, the amount of data l2arc_evict()
* evicts is significant compared to the amount of restored L2ARC
* data. In this case do not write log blocks in L2ARC in order
* not to waste space.
*/
static int l2arc_rebuild_enabled = B_TRUE;
static uint64_t l2arc_rebuild_blocks_min_l2size = 1024 * 1024 * 1024;
/* L2ARC persistence rebuild control routines. */
void l2arc_rebuild_vdev(vdev_t *vd, boolean_t reopen);
static __attribute__((noreturn)) void l2arc_dev_rebuild_thread(void *arg);
static int l2arc_rebuild(l2arc_dev_t *dev);
/* L2ARC persistence read I/O routines. */
static int l2arc_dev_hdr_read(l2arc_dev_t *dev);
static int l2arc_log_blk_read(l2arc_dev_t *dev,
const l2arc_log_blkptr_t *this_lp, const l2arc_log_blkptr_t *next_lp,
l2arc_log_blk_phys_t *this_lb, l2arc_log_blk_phys_t *next_lb,
zio_t *this_io, zio_t **next_io);
static zio_t *l2arc_log_blk_fetch(vdev_t *vd,
const l2arc_log_blkptr_t *lp, l2arc_log_blk_phys_t *lb);
static void l2arc_log_blk_fetch_abort(zio_t *zio);
/* L2ARC persistence block restoration routines. */
static void l2arc_log_blk_restore(l2arc_dev_t *dev,
const l2arc_log_blk_phys_t *lb, uint64_t lb_asize);
static void l2arc_hdr_restore(const l2arc_log_ent_phys_t *le,
l2arc_dev_t *dev);
/* L2ARC persistence write I/O routines. */
static void l2arc_log_blk_commit(l2arc_dev_t *dev, zio_t *pio,
l2arc_write_callback_t *cb);
/* L2ARC persistence auxiliary routines. */
boolean_t l2arc_log_blkptr_valid(l2arc_dev_t *dev,
const l2arc_log_blkptr_t *lbp);
static boolean_t l2arc_log_blk_insert(l2arc_dev_t *dev,
const arc_buf_hdr_t *ab);
boolean_t l2arc_range_check_overlap(uint64_t bottom,
uint64_t top, uint64_t check);
static void l2arc_blk_fetch_done(zio_t *zio);
static inline uint64_t
l2arc_log_blk_overhead(uint64_t write_sz, l2arc_dev_t *dev);
/*
* We use Cityhash for this. It's fast, and has good hash properties without
* requiring any large static buffers.
*/
static uint64_t
buf_hash(uint64_t spa, const dva_t *dva, uint64_t birth)
{
return (cityhash4(spa, dva->dva_word[0], dva->dva_word[1], birth));
}
#define HDR_EMPTY(hdr) \
((hdr)->b_dva.dva_word[0] == 0 && \
(hdr)->b_dva.dva_word[1] == 0)
#define HDR_EMPTY_OR_LOCKED(hdr) \
(HDR_EMPTY(hdr) || MUTEX_HELD(HDR_LOCK(hdr)))
#define HDR_EQUAL(spa, dva, birth, hdr) \
((hdr)->b_dva.dva_word[0] == (dva)->dva_word[0]) && \
((hdr)->b_dva.dva_word[1] == (dva)->dva_word[1]) && \
((hdr)->b_birth == birth) && ((hdr)->b_spa == spa)
static void
buf_discard_identity(arc_buf_hdr_t *hdr)
{
hdr->b_dva.dva_word[0] = 0;
hdr->b_dva.dva_word[1] = 0;
hdr->b_birth = 0;
}
static arc_buf_hdr_t *
buf_hash_find(uint64_t spa, const blkptr_t *bp, kmutex_t **lockp)
{
const dva_t *dva = BP_IDENTITY(bp);
uint64_t birth = BP_PHYSICAL_BIRTH(bp);
uint64_t idx = BUF_HASH_INDEX(spa, dva, birth);
kmutex_t *hash_lock = BUF_HASH_LOCK(idx);
arc_buf_hdr_t *hdr;
mutex_enter(hash_lock);
for (hdr = buf_hash_table.ht_table[idx]; hdr != NULL;
hdr = hdr->b_hash_next) {
if (HDR_EQUAL(spa, dva, birth, hdr)) {
*lockp = hash_lock;
return (hdr);
}
}
mutex_exit(hash_lock);
*lockp = NULL;
return (NULL);
}
/*
* Insert an entry into the hash table. If there is already an element
* equal to elem in the hash table, then the already existing element
* will be returned and the new element will not be inserted.
* Otherwise returns NULL.
* If lockp == NULL, the caller is assumed to already hold the hash lock.
*/
static arc_buf_hdr_t *
buf_hash_insert(arc_buf_hdr_t *hdr, kmutex_t **lockp)
{
uint64_t idx = BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth);
kmutex_t *hash_lock = BUF_HASH_LOCK(idx);
arc_buf_hdr_t *fhdr;
uint32_t i;
ASSERT(!DVA_IS_EMPTY(&hdr->b_dva));
ASSERT(hdr->b_birth != 0);
ASSERT(!HDR_IN_HASH_TABLE(hdr));
if (lockp != NULL) {
*lockp = hash_lock;
mutex_enter(hash_lock);
} else {
ASSERT(MUTEX_HELD(hash_lock));
}
for (fhdr = buf_hash_table.ht_table[idx], i = 0; fhdr != NULL;
fhdr = fhdr->b_hash_next, i++) {
if (HDR_EQUAL(hdr->b_spa, &hdr->b_dva, hdr->b_birth, fhdr))
return (fhdr);
}
hdr->b_hash_next = buf_hash_table.ht_table[idx];
buf_hash_table.ht_table[idx] = hdr;
arc_hdr_set_flags(hdr, ARC_FLAG_IN_HASH_TABLE);
/* collect some hash table performance data */
if (i > 0) {
ARCSTAT_BUMP(arcstat_hash_collisions);
if (i == 1)
ARCSTAT_BUMP(arcstat_hash_chains);
ARCSTAT_MAX(arcstat_hash_chain_max, i);
}
uint64_t he = atomic_inc_64_nv(
&arc_stats.arcstat_hash_elements.value.ui64);
ARCSTAT_MAX(arcstat_hash_elements_max, he);
return (NULL);
}
static void
buf_hash_remove(arc_buf_hdr_t *hdr)
{
arc_buf_hdr_t *fhdr, **hdrp;
uint64_t idx = BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth);
ASSERT(MUTEX_HELD(BUF_HASH_LOCK(idx)));
ASSERT(HDR_IN_HASH_TABLE(hdr));
hdrp = &buf_hash_table.ht_table[idx];
while ((fhdr = *hdrp) != hdr) {
ASSERT3P(fhdr, !=, NULL);
hdrp = &fhdr->b_hash_next;
}
*hdrp = hdr->b_hash_next;
hdr->b_hash_next = NULL;
arc_hdr_clear_flags(hdr, ARC_FLAG_IN_HASH_TABLE);
/* collect some hash table performance data */
atomic_dec_64(&arc_stats.arcstat_hash_elements.value.ui64);
if (buf_hash_table.ht_table[idx] &&
buf_hash_table.ht_table[idx]->b_hash_next == NULL)
ARCSTAT_BUMPDOWN(arcstat_hash_chains);
}
/*
* Global data structures and functions for the buf kmem cache.
*/
static kmem_cache_t *hdr_full_cache;
static kmem_cache_t *hdr_full_crypt_cache;
static kmem_cache_t *hdr_l2only_cache;
static kmem_cache_t *buf_cache;
static void
buf_fini(void)
{
#if defined(_KERNEL)
/*
* Large allocations which do not require contiguous pages
* should be using vmem_free() in the linux kernel\
*/
vmem_free(buf_hash_table.ht_table,
(buf_hash_table.ht_mask + 1) * sizeof (void *));
#else
kmem_free(buf_hash_table.ht_table,
(buf_hash_table.ht_mask + 1) * sizeof (void *));
#endif
for (int i = 0; i < BUF_LOCKS; i++)
mutex_destroy(BUF_HASH_LOCK(i));
kmem_cache_destroy(hdr_full_cache);
kmem_cache_destroy(hdr_full_crypt_cache);
kmem_cache_destroy(hdr_l2only_cache);
kmem_cache_destroy(buf_cache);
}
/*
* Constructor callback - called when the cache is empty
* and a new buf is requested.
*/
static int
hdr_full_cons(void *vbuf, void *unused, int kmflag)
{
(void) unused, (void) kmflag;
arc_buf_hdr_t *hdr = vbuf;
memset(hdr, 0, HDR_FULL_SIZE);
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_NUMFUNCS;
cv_init(&hdr->b_l1hdr.b_cv, NULL, CV_DEFAULT, NULL);
zfs_refcount_create(&hdr->b_l1hdr.b_refcnt);
#ifdef ZFS_DEBUG
mutex_init(&hdr->b_l1hdr.b_freeze_lock, NULL, MUTEX_DEFAULT, NULL);
#endif
multilist_link_init(&hdr->b_l1hdr.b_arc_node);
list_link_init(&hdr->b_l2hdr.b_l2node);
arc_space_consume(HDR_FULL_SIZE, ARC_SPACE_HDRS);
return (0);
}
static int
hdr_full_crypt_cons(void *vbuf, void *unused, int kmflag)
{
(void) unused;
arc_buf_hdr_t *hdr = vbuf;
hdr_full_cons(vbuf, unused, kmflag);
memset(&hdr->b_crypt_hdr, 0, sizeof (hdr->b_crypt_hdr));
arc_space_consume(sizeof (hdr->b_crypt_hdr), ARC_SPACE_HDRS);
return (0);
}
static int
hdr_l2only_cons(void *vbuf, void *unused, int kmflag)
{
(void) unused, (void) kmflag;
arc_buf_hdr_t *hdr = vbuf;
memset(hdr, 0, HDR_L2ONLY_SIZE);
arc_space_consume(HDR_L2ONLY_SIZE, ARC_SPACE_L2HDRS);
return (0);
}
static int
buf_cons(void *vbuf, void *unused, int kmflag)
{
(void) unused, (void) kmflag;
arc_buf_t *buf = vbuf;
memset(buf, 0, sizeof (arc_buf_t));
arc_space_consume(sizeof (arc_buf_t), ARC_SPACE_HDRS);
return (0);
}
/*
* Destructor callback - called when a cached buf is
* no longer required.
*/
static void
hdr_full_dest(void *vbuf, void *unused)
{
(void) unused;
arc_buf_hdr_t *hdr = vbuf;
ASSERT(HDR_EMPTY(hdr));
cv_destroy(&hdr->b_l1hdr.b_cv);
zfs_refcount_destroy(&hdr->b_l1hdr.b_refcnt);
#ifdef ZFS_DEBUG
mutex_destroy(&hdr->b_l1hdr.b_freeze_lock);
#endif
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
arc_space_return(HDR_FULL_SIZE, ARC_SPACE_HDRS);
}
static void
hdr_full_crypt_dest(void *vbuf, void *unused)
{
(void) vbuf, (void) unused;
hdr_full_dest(vbuf, unused);
arc_space_return(sizeof (((arc_buf_hdr_t *)NULL)->b_crypt_hdr),
ARC_SPACE_HDRS);
}
static void
hdr_l2only_dest(void *vbuf, void *unused)
{
(void) unused;
arc_buf_hdr_t *hdr = vbuf;
ASSERT(HDR_EMPTY(hdr));
arc_space_return(HDR_L2ONLY_SIZE, ARC_SPACE_L2HDRS);
}
static void
buf_dest(void *vbuf, void *unused)
{
(void) unused;
(void) vbuf;
arc_space_return(sizeof (arc_buf_t), ARC_SPACE_HDRS);
}
static void
buf_init(void)
{
uint64_t *ct = NULL;
uint64_t hsize = 1ULL << 12;
int i, j;
/*
* The hash table is big enough to fill all of physical memory
* with an average block size of zfs_arc_average_blocksize (default 8K).
* By default, the table will take up
* totalmem * sizeof(void*) / 8K (1MB per GB with 8-byte pointers).
*/
while (hsize * zfs_arc_average_blocksize < arc_all_memory())
hsize <<= 1;
retry:
buf_hash_table.ht_mask = hsize - 1;
#if defined(_KERNEL)
/*
* Large allocations which do not require contiguous pages
* should be using vmem_alloc() in the linux kernel
*/
buf_hash_table.ht_table =
vmem_zalloc(hsize * sizeof (void*), KM_SLEEP);
#else
buf_hash_table.ht_table =
kmem_zalloc(hsize * sizeof (void*), KM_NOSLEEP);
#endif
if (buf_hash_table.ht_table == NULL) {
ASSERT(hsize > (1ULL << 8));
hsize >>= 1;
goto retry;
}
hdr_full_cache = kmem_cache_create("arc_buf_hdr_t_full", HDR_FULL_SIZE,
0, hdr_full_cons, hdr_full_dest, NULL, NULL, NULL, 0);
hdr_full_crypt_cache = kmem_cache_create("arc_buf_hdr_t_full_crypt",
HDR_FULL_CRYPT_SIZE, 0, hdr_full_crypt_cons, hdr_full_crypt_dest,
NULL, NULL, NULL, 0);
hdr_l2only_cache = kmem_cache_create("arc_buf_hdr_t_l2only",
HDR_L2ONLY_SIZE, 0, hdr_l2only_cons, hdr_l2only_dest, NULL,
NULL, NULL, 0);
buf_cache = kmem_cache_create("arc_buf_t", sizeof (arc_buf_t),
0, buf_cons, buf_dest, NULL, NULL, NULL, 0);
for (i = 0; i < 256; i++)
for (ct = zfs_crc64_table + i, *ct = i, j = 8; j > 0; j--)
*ct = (*ct >> 1) ^ (-(*ct & 1) & ZFS_CRC64_POLY);
for (i = 0; i < BUF_LOCKS; i++)
mutex_init(BUF_HASH_LOCK(i), NULL, MUTEX_DEFAULT, NULL);
}
#define ARC_MINTIME (hz>>4) /* 62 ms */
/*
* This is the size that the buf occupies in memory. If the buf is compressed,
* it will correspond to the compressed size. You should use this method of
* getting the buf size unless you explicitly need the logical size.
*/
uint64_t
arc_buf_size(arc_buf_t *buf)
{
return (ARC_BUF_COMPRESSED(buf) ?
HDR_GET_PSIZE(buf->b_hdr) : HDR_GET_LSIZE(buf->b_hdr));
}
uint64_t
arc_buf_lsize(arc_buf_t *buf)
{
return (HDR_GET_LSIZE(buf->b_hdr));
}
/*
* This function will return B_TRUE if the buffer is encrypted in memory.
* This buffer can be decrypted by calling arc_untransform().
*/
boolean_t
arc_is_encrypted(arc_buf_t *buf)
{
return (ARC_BUF_ENCRYPTED(buf) != 0);
}
/*
* Returns B_TRUE if the buffer represents data that has not had its MAC
* verified yet.
*/
boolean_t
arc_is_unauthenticated(arc_buf_t *buf)
{
return (HDR_NOAUTH(buf->b_hdr) != 0);
}
void
arc_get_raw_params(arc_buf_t *buf, boolean_t *byteorder, uint8_t *salt,
uint8_t *iv, uint8_t *mac)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT(HDR_PROTECTED(hdr));
memcpy(salt, hdr->b_crypt_hdr.b_salt, ZIO_DATA_SALT_LEN);
memcpy(iv, hdr->b_crypt_hdr.b_iv, ZIO_DATA_IV_LEN);
memcpy(mac, hdr->b_crypt_hdr.b_mac, ZIO_DATA_MAC_LEN);
*byteorder = (hdr->b_l1hdr.b_byteswap == DMU_BSWAP_NUMFUNCS) ?
ZFS_HOST_BYTEORDER : !ZFS_HOST_BYTEORDER;
}
/*
* Indicates how this buffer is compressed in memory. If it is not compressed
* the value will be ZIO_COMPRESS_OFF. It can be made normally readable with
* arc_untransform() as long as it is also unencrypted.
*/
enum zio_compress
arc_get_compression(arc_buf_t *buf)
{
return (ARC_BUF_COMPRESSED(buf) ?
HDR_GET_COMPRESS(buf->b_hdr) : ZIO_COMPRESS_OFF);
}
/*
* Return the compression algorithm used to store this data in the ARC. If ARC
* compression is enabled or this is an encrypted block, this will be the same
* as what's used to store it on-disk. Otherwise, this will be ZIO_COMPRESS_OFF.
*/
static inline enum zio_compress
arc_hdr_get_compress(arc_buf_hdr_t *hdr)
{
return (HDR_COMPRESSION_ENABLED(hdr) ?
HDR_GET_COMPRESS(hdr) : ZIO_COMPRESS_OFF);
}
uint8_t
arc_get_complevel(arc_buf_t *buf)
{
return (buf->b_hdr->b_complevel);
}
static inline boolean_t
arc_buf_is_shared(arc_buf_t *buf)
{
boolean_t shared = (buf->b_data != NULL &&
buf->b_hdr->b_l1hdr.b_pabd != NULL &&
abd_is_linear(buf->b_hdr->b_l1hdr.b_pabd) &&
buf->b_data == abd_to_buf(buf->b_hdr->b_l1hdr.b_pabd));
IMPLY(shared, HDR_SHARED_DATA(buf->b_hdr));
IMPLY(shared, ARC_BUF_SHARED(buf));
IMPLY(shared, ARC_BUF_COMPRESSED(buf) || ARC_BUF_LAST(buf));
/*
* It would be nice to assert arc_can_share() too, but the "hdr isn't
* already being shared" requirement prevents us from doing that.
*/
return (shared);
}
/*
* Free the checksum associated with this header. If there is no checksum, this
* is a no-op.
*/
static inline void
arc_cksum_free(arc_buf_hdr_t *hdr)
{
#ifdef ZFS_DEBUG
ASSERT(HDR_HAS_L1HDR(hdr));
mutex_enter(&hdr->b_l1hdr.b_freeze_lock);
if (hdr->b_l1hdr.b_freeze_cksum != NULL) {
kmem_free(hdr->b_l1hdr.b_freeze_cksum, sizeof (zio_cksum_t));
hdr->b_l1hdr.b_freeze_cksum = NULL;
}
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
#endif
}
/*
* Return true iff at least one of the bufs on hdr is not compressed.
* Encrypted buffers count as compressed.
*/
static boolean_t
arc_hdr_has_uncompressed_buf(arc_buf_hdr_t *hdr)
{
ASSERT(hdr->b_l1hdr.b_state == arc_anon || HDR_EMPTY_OR_LOCKED(hdr));
for (arc_buf_t *b = hdr->b_l1hdr.b_buf; b != NULL; b = b->b_next) {
if (!ARC_BUF_COMPRESSED(b)) {
return (B_TRUE);
}
}
return (B_FALSE);
}
/*
* If we've turned on the ZFS_DEBUG_MODIFY flag, verify that the buf's data
* matches the checksum that is stored in the hdr. If there is no checksum,
* or if the buf is compressed, this is a no-op.
*/
static void
arc_cksum_verify(arc_buf_t *buf)
{
#ifdef ZFS_DEBUG
arc_buf_hdr_t *hdr = buf->b_hdr;
zio_cksum_t zc;
if (!(zfs_flags & ZFS_DEBUG_MODIFY))
return;
if (ARC_BUF_COMPRESSED(buf))
return;
ASSERT(HDR_HAS_L1HDR(hdr));
mutex_enter(&hdr->b_l1hdr.b_freeze_lock);
if (hdr->b_l1hdr.b_freeze_cksum == NULL || HDR_IO_ERROR(hdr)) {
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
return;
}
fletcher_2_native(buf->b_data, arc_buf_size(buf), NULL, &zc);
if (!ZIO_CHECKSUM_EQUAL(*hdr->b_l1hdr.b_freeze_cksum, zc))
panic("buffer modified while frozen!");
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
#endif
}
/*
* This function makes the assumption that data stored in the L2ARC
* will be transformed exactly as it is in the main pool. Because of
* this we can verify the checksum against the reading process's bp.
*/
static boolean_t
arc_cksum_is_equal(arc_buf_hdr_t *hdr, zio_t *zio)
{
ASSERT(!BP_IS_EMBEDDED(zio->io_bp));
VERIFY3U(BP_GET_PSIZE(zio->io_bp), ==, HDR_GET_PSIZE(hdr));
/*
* Block pointers always store the checksum for the logical data.
* If the block pointer has the gang bit set, then the checksum
* it represents is for the reconstituted data and not for an
* individual gang member. The zio pipeline, however, must be able to
* determine the checksum of each of the gang constituents so it
* treats the checksum comparison differently than what we need
* for l2arc blocks. This prevents us from using the
* zio_checksum_error() interface directly. Instead we must call the
* zio_checksum_error_impl() so that we can ensure the checksum is
* generated using the correct checksum algorithm and accounts for the
* logical I/O size and not just a gang fragment.
*/
return (zio_checksum_error_impl(zio->io_spa, zio->io_bp,
BP_GET_CHECKSUM(zio->io_bp), zio->io_abd, zio->io_size,
zio->io_offset, NULL) == 0);
}
/*
* Given a buf full of data, if ZFS_DEBUG_MODIFY is enabled this computes a
* checksum and attaches it to the buf's hdr so that we can ensure that the buf
* isn't modified later on. If buf is compressed or there is already a checksum
* on the hdr, this is a no-op (we only checksum uncompressed bufs).
*/
static void
arc_cksum_compute(arc_buf_t *buf)
{
if (!(zfs_flags & ZFS_DEBUG_MODIFY))
return;
#ifdef ZFS_DEBUG
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT(HDR_HAS_L1HDR(hdr));
mutex_enter(&hdr->b_l1hdr.b_freeze_lock);
if (hdr->b_l1hdr.b_freeze_cksum != NULL || ARC_BUF_COMPRESSED(buf)) {
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
return;
}
ASSERT(!ARC_BUF_ENCRYPTED(buf));
ASSERT(!ARC_BUF_COMPRESSED(buf));
hdr->b_l1hdr.b_freeze_cksum = kmem_alloc(sizeof (zio_cksum_t),
KM_SLEEP);
fletcher_2_native(buf->b_data, arc_buf_size(buf), NULL,
hdr->b_l1hdr.b_freeze_cksum);
mutex_exit(&hdr->b_l1hdr.b_freeze_lock);
#endif
arc_buf_watch(buf);
}
#ifndef _KERNEL
void
arc_buf_sigsegv(int sig, siginfo_t *si, void *unused)
{
(void) sig, (void) unused;
panic("Got SIGSEGV at address: 0x%lx\n", (long)si->si_addr);
}
#endif
static void
arc_buf_unwatch(arc_buf_t *buf)
{
#ifndef _KERNEL
if (arc_watch) {
ASSERT0(mprotect(buf->b_data, arc_buf_size(buf),
PROT_READ | PROT_WRITE));
}
#else
(void) buf;
#endif
}
static void
arc_buf_watch(arc_buf_t *buf)
{
#ifndef _KERNEL
if (arc_watch)
ASSERT0(mprotect(buf->b_data, arc_buf_size(buf),
PROT_READ));
#else
(void) buf;
#endif
}
static arc_buf_contents_t
arc_buf_type(arc_buf_hdr_t *hdr)
{
arc_buf_contents_t type;
if (HDR_ISTYPE_METADATA(hdr)) {
type = ARC_BUFC_METADATA;
} else {
type = ARC_BUFC_DATA;
}
VERIFY3U(hdr->b_type, ==, type);
return (type);
}
boolean_t
arc_is_metadata(arc_buf_t *buf)
{
return (HDR_ISTYPE_METADATA(buf->b_hdr) != 0);
}
static uint32_t
arc_bufc_to_flags(arc_buf_contents_t type)
{
switch (type) {
case ARC_BUFC_DATA:
/* metadata field is 0 if buffer contains normal data */
return (0);
case ARC_BUFC_METADATA:
return (ARC_FLAG_BUFC_METADATA);
default:
break;
}
panic("undefined ARC buffer type!");
return ((uint32_t)-1);
}
void
arc_buf_thaw(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon);
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
arc_cksum_verify(buf);
/*
* Compressed buffers do not manipulate the b_freeze_cksum.
*/
if (ARC_BUF_COMPRESSED(buf))
return;
ASSERT(HDR_HAS_L1HDR(hdr));
arc_cksum_free(hdr);
arc_buf_unwatch(buf);
}
void
arc_buf_freeze(arc_buf_t *buf)
{
if (!(zfs_flags & ZFS_DEBUG_MODIFY))
return;
if (ARC_BUF_COMPRESSED(buf))
return;
ASSERT(HDR_HAS_L1HDR(buf->b_hdr));
arc_cksum_compute(buf);
}
/*
* The arc_buf_hdr_t's b_flags should never be modified directly. Instead,
* the following functions should be used to ensure that the flags are
* updated in a thread-safe way. When manipulating the flags either
* the hash_lock must be held or the hdr must be undiscoverable. This
* ensures that we're not racing with any other threads when updating
* the flags.
*/
static inline void
arc_hdr_set_flags(arc_buf_hdr_t *hdr, arc_flags_t flags)
{
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
hdr->b_flags |= flags;
}
static inline void
arc_hdr_clear_flags(arc_buf_hdr_t *hdr, arc_flags_t flags)
{
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
hdr->b_flags &= ~flags;
}
/*
* Setting the compression bits in the arc_buf_hdr_t's b_flags is
* done in a special way since we have to clear and set bits
* at the same time. Consumers that wish to set the compression bits
* must use this function to ensure that the flags are updated in
* thread-safe manner.
*/
static void
arc_hdr_set_compress(arc_buf_hdr_t *hdr, enum zio_compress cmp)
{
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
/*
* Holes and embedded blocks will always have a psize = 0 so
* we ignore the compression of the blkptr and set the
* want to uncompress them. Mark them as uncompressed.
*/
if (!zfs_compressed_arc_enabled || HDR_GET_PSIZE(hdr) == 0) {
arc_hdr_clear_flags(hdr, ARC_FLAG_COMPRESSED_ARC);
ASSERT(!HDR_COMPRESSION_ENABLED(hdr));
} else {
arc_hdr_set_flags(hdr, ARC_FLAG_COMPRESSED_ARC);
ASSERT(HDR_COMPRESSION_ENABLED(hdr));
}
HDR_SET_COMPRESS(hdr, cmp);
ASSERT3U(HDR_GET_COMPRESS(hdr), ==, cmp);
}
/*
* Looks for another buf on the same hdr which has the data decompressed, copies
* from it, and returns true. If no such buf exists, returns false.
*/
static boolean_t
arc_buf_try_copy_decompressed_data(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
boolean_t copied = B_FALSE;
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT3P(buf->b_data, !=, NULL);
ASSERT(!ARC_BUF_COMPRESSED(buf));
for (arc_buf_t *from = hdr->b_l1hdr.b_buf; from != NULL;
from = from->b_next) {
/* can't use our own data buffer */
if (from == buf) {
continue;
}
if (!ARC_BUF_COMPRESSED(from)) {
memcpy(buf->b_data, from->b_data, arc_buf_size(buf));
copied = B_TRUE;
break;
}
}
#ifdef ZFS_DEBUG
/*
* There were no decompressed bufs, so there should not be a
* checksum on the hdr either.
*/
if (zfs_flags & ZFS_DEBUG_MODIFY)
EQUIV(!copied, hdr->b_l1hdr.b_freeze_cksum == NULL);
#endif
return (copied);
}
/*
* Allocates an ARC buf header that's in an evicted & L2-cached state.
* This is used during l2arc reconstruction to make empty ARC buffers
* which circumvent the regular disk->arc->l2arc path and instead come
* into being in the reverse order, i.e. l2arc->arc.
*/
static arc_buf_hdr_t *
arc_buf_alloc_l2only(size_t size, arc_buf_contents_t type, l2arc_dev_t *dev,
dva_t dva, uint64_t daddr, int32_t psize, uint64_t birth,
enum zio_compress compress, uint8_t complevel, boolean_t protected,
boolean_t prefetch, arc_state_type_t arcs_state)
{
arc_buf_hdr_t *hdr;
ASSERT(size != 0);
hdr = kmem_cache_alloc(hdr_l2only_cache, KM_SLEEP);
hdr->b_birth = birth;
hdr->b_type = type;
hdr->b_flags = 0;
arc_hdr_set_flags(hdr, arc_bufc_to_flags(type) | ARC_FLAG_HAS_L2HDR);
HDR_SET_LSIZE(hdr, size);
HDR_SET_PSIZE(hdr, psize);
arc_hdr_set_compress(hdr, compress);
hdr->b_complevel = complevel;
if (protected)
arc_hdr_set_flags(hdr, ARC_FLAG_PROTECTED);
if (prefetch)
arc_hdr_set_flags(hdr, ARC_FLAG_PREFETCH);
hdr->b_spa = spa_load_guid(dev->l2ad_vdev->vdev_spa);
hdr->b_dva = dva;
hdr->b_l2hdr.b_dev = dev;
hdr->b_l2hdr.b_daddr = daddr;
hdr->b_l2hdr.b_arcs_state = arcs_state;
return (hdr);
}
/*
* Return the size of the block, b_pabd, that is stored in the arc_buf_hdr_t.
*/
static uint64_t
arc_hdr_size(arc_buf_hdr_t *hdr)
{
uint64_t size;
if (arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF &&
HDR_GET_PSIZE(hdr) > 0) {
size = HDR_GET_PSIZE(hdr);
} else {
ASSERT3U(HDR_GET_LSIZE(hdr), !=, 0);
size = HDR_GET_LSIZE(hdr);
}
return (size);
}
static int
arc_hdr_authenticate(arc_buf_hdr_t *hdr, spa_t *spa, uint64_t dsobj)
{
int ret;
uint64_t csize;
uint64_t lsize = HDR_GET_LSIZE(hdr);
uint64_t psize = HDR_GET_PSIZE(hdr);
void *tmpbuf = NULL;
abd_t *abd = hdr->b_l1hdr.b_pabd;
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
ASSERT(HDR_AUTHENTICATED(hdr));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
/*
* The MAC is calculated on the compressed data that is stored on disk.
* However, if compressed arc is disabled we will only have the
* decompressed data available to us now. Compress it into a temporary
* abd so we can verify the MAC. The performance overhead of this will
* be relatively low, since most objects in an encrypted objset will
* be encrypted (instead of authenticated) anyway.
*/
if (HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr)) {
csize = zio_compress_data(HDR_GET_COMPRESS(hdr),
hdr->b_l1hdr.b_pabd, &tmpbuf, lsize, hdr->b_complevel);
ASSERT3P(tmpbuf, !=, NULL);
ASSERT3U(csize, <=, psize);
abd = abd_get_from_buf(tmpbuf, lsize);
abd_take_ownership_of_buf(abd, B_TRUE);
abd_zero_off(abd, csize, psize - csize);
}
/*
* Authentication is best effort. We authenticate whenever the key is
* available. If we succeed we clear ARC_FLAG_NOAUTH.
*/
if (hdr->b_crypt_hdr.b_ot == DMU_OT_OBJSET) {
ASSERT3U(HDR_GET_COMPRESS(hdr), ==, ZIO_COMPRESS_OFF);
ASSERT3U(lsize, ==, psize);
ret = spa_do_crypt_objset_mac_abd(B_FALSE, spa, dsobj, abd,
psize, hdr->b_l1hdr.b_byteswap != DMU_BSWAP_NUMFUNCS);
} else {
ret = spa_do_crypt_mac_abd(B_FALSE, spa, dsobj, abd, psize,
hdr->b_crypt_hdr.b_mac);
}
if (ret == 0)
arc_hdr_clear_flags(hdr, ARC_FLAG_NOAUTH);
else if (ret != ENOENT)
goto error;
if (tmpbuf != NULL)
abd_free(abd);
return (0);
error:
if (tmpbuf != NULL)
abd_free(abd);
return (ret);
}
/*
* This function will take a header that only has raw encrypted data in
* b_crypt_hdr.b_rabd and decrypt it into a new buffer which is stored in
* b_l1hdr.b_pabd. If designated in the header flags, this function will
* also decompress the data.
*/
static int
arc_hdr_decrypt(arc_buf_hdr_t *hdr, spa_t *spa, const zbookmark_phys_t *zb)
{
int ret;
abd_t *cabd = NULL;
void *tmp = NULL;
boolean_t no_crypt = B_FALSE;
boolean_t bswap = (hdr->b_l1hdr.b_byteswap != DMU_BSWAP_NUMFUNCS);
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
ASSERT(HDR_ENCRYPTED(hdr));
arc_hdr_alloc_abd(hdr, 0);
ret = spa_do_crypt_abd(B_FALSE, spa, zb, hdr->b_crypt_hdr.b_ot,
B_FALSE, bswap, hdr->b_crypt_hdr.b_salt, hdr->b_crypt_hdr.b_iv,
hdr->b_crypt_hdr.b_mac, HDR_GET_PSIZE(hdr), hdr->b_l1hdr.b_pabd,
hdr->b_crypt_hdr.b_rabd, &no_crypt);
if (ret != 0)
goto error;
if (no_crypt) {
abd_copy(hdr->b_l1hdr.b_pabd, hdr->b_crypt_hdr.b_rabd,
HDR_GET_PSIZE(hdr));
}
/*
* If this header has disabled arc compression but the b_pabd is
* compressed after decrypting it, we need to decompress the newly
* decrypted data.
*/
if (HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr)) {
/*
* We want to make sure that we are correctly honoring the
* zfs_abd_scatter_enabled setting, so we allocate an abd here
* and then loan a buffer from it, rather than allocating a
* linear buffer and wrapping it in an abd later.
*/
cabd = arc_get_data_abd(hdr, arc_hdr_size(hdr), hdr, 0);
tmp = abd_borrow_buf(cabd, arc_hdr_size(hdr));
ret = zio_decompress_data(HDR_GET_COMPRESS(hdr),
hdr->b_l1hdr.b_pabd, tmp, HDR_GET_PSIZE(hdr),
HDR_GET_LSIZE(hdr), &hdr->b_complevel);
if (ret != 0) {
abd_return_buf(cabd, tmp, arc_hdr_size(hdr));
goto error;
}
abd_return_buf_copy(cabd, tmp, arc_hdr_size(hdr));
arc_free_data_abd(hdr, hdr->b_l1hdr.b_pabd,
arc_hdr_size(hdr), hdr);
hdr->b_l1hdr.b_pabd = cabd;
}
return (0);
error:
arc_hdr_free_abd(hdr, B_FALSE);
if (cabd != NULL)
arc_free_data_buf(hdr, cabd, arc_hdr_size(hdr), hdr);
return (ret);
}
/*
* This function is called during arc_buf_fill() to prepare the header's
* abd plaintext pointer for use. This involves authenticated protected
* data and decrypting encrypted data into the plaintext abd.
*/
static int
arc_fill_hdr_crypt(arc_buf_hdr_t *hdr, kmutex_t *hash_lock, spa_t *spa,
const zbookmark_phys_t *zb, boolean_t noauth)
{
int ret;
ASSERT(HDR_PROTECTED(hdr));
if (hash_lock != NULL)
mutex_enter(hash_lock);
if (HDR_NOAUTH(hdr) && !noauth) {
/*
* The caller requested authenticated data but our data has
* not been authenticated yet. Verify the MAC now if we can.
*/
ret = arc_hdr_authenticate(hdr, spa, zb->zb_objset);
if (ret != 0)
goto error;
} else if (HDR_HAS_RABD(hdr) && hdr->b_l1hdr.b_pabd == NULL) {
/*
* If we only have the encrypted version of the data, but the
* unencrypted version was requested we take this opportunity
* to store the decrypted version in the header for future use.
*/
ret = arc_hdr_decrypt(hdr, spa, zb);
if (ret != 0)
goto error;
}
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
if (hash_lock != NULL)
mutex_exit(hash_lock);
return (0);
error:
if (hash_lock != NULL)
mutex_exit(hash_lock);
return (ret);
}
/*
* This function is used by the dbuf code to decrypt bonus buffers in place.
* The dbuf code itself doesn't have any locking for decrypting a shared dnode
* block, so we use the hash lock here to protect against concurrent calls to
* arc_buf_fill().
*/
static void
arc_buf_untransform_in_place(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT(HDR_ENCRYPTED(hdr));
ASSERT3U(hdr->b_crypt_hdr.b_ot, ==, DMU_OT_DNODE);
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
zio_crypt_copy_dnode_bonus(hdr->b_l1hdr.b_pabd, buf->b_data,
arc_buf_size(buf));
buf->b_flags &= ~ARC_BUF_FLAG_ENCRYPTED;
buf->b_flags &= ~ARC_BUF_FLAG_COMPRESSED;
hdr->b_crypt_hdr.b_ebufcnt -= 1;
}
/*
* Given a buf that has a data buffer attached to it, this function will
* efficiently fill the buf with data of the specified compression setting from
* the hdr and update the hdr's b_freeze_cksum if necessary. If the buf and hdr
* are already sharing a data buf, no copy is performed.
*
* If the buf is marked as compressed but uncompressed data was requested, this
* will allocate a new data buffer for the buf, remove that flag, and fill the
* buf with uncompressed data. You can't request a compressed buf on a hdr with
* uncompressed data, and (since we haven't added support for it yet) if you
* want compressed data your buf must already be marked as compressed and have
* the correct-sized data buffer.
*/
static int
arc_buf_fill(arc_buf_t *buf, spa_t *spa, const zbookmark_phys_t *zb,
arc_fill_flags_t flags)
{
int error = 0;
arc_buf_hdr_t *hdr = buf->b_hdr;
boolean_t hdr_compressed =
(arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF);
boolean_t compressed = (flags & ARC_FILL_COMPRESSED) != 0;
boolean_t encrypted = (flags & ARC_FILL_ENCRYPTED) != 0;
dmu_object_byteswap_t bswap = hdr->b_l1hdr.b_byteswap;
kmutex_t *hash_lock = (flags & ARC_FILL_LOCKED) ? NULL : HDR_LOCK(hdr);
ASSERT3P(buf->b_data, !=, NULL);
IMPLY(compressed, hdr_compressed || ARC_BUF_ENCRYPTED(buf));
IMPLY(compressed, ARC_BUF_COMPRESSED(buf));
IMPLY(encrypted, HDR_ENCRYPTED(hdr));
IMPLY(encrypted, ARC_BUF_ENCRYPTED(buf));
IMPLY(encrypted, ARC_BUF_COMPRESSED(buf));
IMPLY(encrypted, !ARC_BUF_SHARED(buf));
/*
* If the caller wanted encrypted data we just need to copy it from
* b_rabd and potentially byteswap it. We won't be able to do any
* further transforms on it.
*/
if (encrypted) {
ASSERT(HDR_HAS_RABD(hdr));
abd_copy_to_buf(buf->b_data, hdr->b_crypt_hdr.b_rabd,
HDR_GET_PSIZE(hdr));
goto byteswap;
}
/*
* Adjust encrypted and authenticated headers to accommodate
* the request if needed. Dnode blocks (ARC_FILL_IN_PLACE) are
* allowed to fail decryption due to keys not being loaded
* without being marked as an IO error.
*/
if (HDR_PROTECTED(hdr)) {
error = arc_fill_hdr_crypt(hdr, hash_lock, spa,
zb, !!(flags & ARC_FILL_NOAUTH));
if (error == EACCES && (flags & ARC_FILL_IN_PLACE) != 0) {
return (error);
} else if (error != 0) {
if (hash_lock != NULL)
mutex_enter(hash_lock);
arc_hdr_set_flags(hdr, ARC_FLAG_IO_ERROR);
if (hash_lock != NULL)
mutex_exit(hash_lock);
return (error);
}
}
/*
* There is a special case here for dnode blocks which are
* decrypting their bonus buffers. These blocks may request to
* be decrypted in-place. This is necessary because there may
* be many dnodes pointing into this buffer and there is
* currently no method to synchronize replacing the backing
* b_data buffer and updating all of the pointers. Here we use
* the hash lock to ensure there are no races. If the need
* arises for other types to be decrypted in-place, they must
* add handling here as well.
*/
if ((flags & ARC_FILL_IN_PLACE) != 0) {
ASSERT(!hdr_compressed);
ASSERT(!compressed);
ASSERT(!encrypted);
if (HDR_ENCRYPTED(hdr) && ARC_BUF_ENCRYPTED(buf)) {
ASSERT3U(hdr->b_crypt_hdr.b_ot, ==, DMU_OT_DNODE);
if (hash_lock != NULL)
mutex_enter(hash_lock);
arc_buf_untransform_in_place(buf);
if (hash_lock != NULL)
mutex_exit(hash_lock);
/* Compute the hdr's checksum if necessary */
arc_cksum_compute(buf);
}
return (0);
}
if (hdr_compressed == compressed) {
if (!arc_buf_is_shared(buf)) {
abd_copy_to_buf(buf->b_data, hdr->b_l1hdr.b_pabd,
arc_buf_size(buf));
}
} else {
ASSERT(hdr_compressed);
ASSERT(!compressed);
/*
* If the buf is sharing its data with the hdr, unlink it and
* allocate a new data buffer for the buf.
*/
if (arc_buf_is_shared(buf)) {
ASSERT(ARC_BUF_COMPRESSED(buf));
/* We need to give the buf its own b_data */
buf->b_flags &= ~ARC_BUF_FLAG_SHARED;
buf->b_data =
arc_get_data_buf(hdr, HDR_GET_LSIZE(hdr), buf);
arc_hdr_clear_flags(hdr, ARC_FLAG_SHARED_DATA);
/* Previously overhead was 0; just add new overhead */
ARCSTAT_INCR(arcstat_overhead_size, HDR_GET_LSIZE(hdr));
} else if (ARC_BUF_COMPRESSED(buf)) {
/* We need to reallocate the buf's b_data */
arc_free_data_buf(hdr, buf->b_data, HDR_GET_PSIZE(hdr),
buf);
buf->b_data =
arc_get_data_buf(hdr, HDR_GET_LSIZE(hdr), buf);
/* We increased the size of b_data; update overhead */
ARCSTAT_INCR(arcstat_overhead_size,
HDR_GET_LSIZE(hdr) - HDR_GET_PSIZE(hdr));
}
/*
* Regardless of the buf's previous compression settings, it
* should not be compressed at the end of this function.
*/
buf->b_flags &= ~ARC_BUF_FLAG_COMPRESSED;
/*
* Try copying the data from another buf which already has a
* decompressed version. If that's not possible, it's time to
* bite the bullet and decompress the data from the hdr.
*/
if (arc_buf_try_copy_decompressed_data(buf)) {
/* Skip byteswapping and checksumming (already done) */
return (0);
} else {
error = zio_decompress_data(HDR_GET_COMPRESS(hdr),
hdr->b_l1hdr.b_pabd, buf->b_data,
HDR_GET_PSIZE(hdr), HDR_GET_LSIZE(hdr),
&hdr->b_complevel);
/*
* Absent hardware errors or software bugs, this should
* be impossible, but log it anyway so we can debug it.
*/
if (error != 0) {
zfs_dbgmsg(
"hdr %px, compress %d, psize %d, lsize %d",
hdr, arc_hdr_get_compress(hdr),
HDR_GET_PSIZE(hdr), HDR_GET_LSIZE(hdr));
if (hash_lock != NULL)
mutex_enter(hash_lock);
arc_hdr_set_flags(hdr, ARC_FLAG_IO_ERROR);
if (hash_lock != NULL)
mutex_exit(hash_lock);
return (SET_ERROR(EIO));
}
}
}
byteswap:
/* Byteswap the buf's data if necessary */
if (bswap != DMU_BSWAP_NUMFUNCS) {
ASSERT(!HDR_SHARED_DATA(hdr));
ASSERT3U(bswap, <, DMU_BSWAP_NUMFUNCS);
dmu_ot_byteswap[bswap].ob_func(buf->b_data, HDR_GET_LSIZE(hdr));
}
/* Compute the hdr's checksum if necessary */
arc_cksum_compute(buf);
return (0);
}
/*
* If this function is being called to decrypt an encrypted buffer or verify an
* authenticated one, the key must be loaded and a mapping must be made
* available in the keystore via spa_keystore_create_mapping() or one of its
* callers.
*/
int
arc_untransform(arc_buf_t *buf, spa_t *spa, const zbookmark_phys_t *zb,
boolean_t in_place)
{
int ret;
arc_fill_flags_t flags = 0;
if (in_place)
flags |= ARC_FILL_IN_PLACE;
ret = arc_buf_fill(buf, spa, zb, flags);
if (ret == ECKSUM) {
/*
* Convert authentication and decryption errors to EIO
* (and generate an ereport) before leaving the ARC.
*/
ret = SET_ERROR(EIO);
spa_log_error(spa, zb, &buf->b_hdr->b_birth);
(void) zfs_ereport_post(FM_EREPORT_ZFS_AUTHENTICATION,
spa, NULL, zb, NULL, 0);
}
return (ret);
}
/*
* Increment the amount of evictable space in the arc_state_t's refcount.
* We account for the space used by the hdr and the arc buf individually
* so that we can add and remove them from the refcount individually.
*/
static void
arc_evictable_space_increment(arc_buf_hdr_t *hdr, arc_state_t *state)
{
arc_buf_contents_t type = arc_buf_type(hdr);
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(state)) {
ASSERT0(hdr->b_l1hdr.b_bufcnt);
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
(void) zfs_refcount_add_many(&state->arcs_esize[type],
HDR_GET_LSIZE(hdr), hdr);
return;
}
if (hdr->b_l1hdr.b_pabd != NULL) {
(void) zfs_refcount_add_many(&state->arcs_esize[type],
arc_hdr_size(hdr), hdr);
}
if (HDR_HAS_RABD(hdr)) {
(void) zfs_refcount_add_many(&state->arcs_esize[type],
HDR_GET_PSIZE(hdr), hdr);
}
for (arc_buf_t *buf = hdr->b_l1hdr.b_buf; buf != NULL;
buf = buf->b_next) {
if (arc_buf_is_shared(buf))
continue;
(void) zfs_refcount_add_many(&state->arcs_esize[type],
arc_buf_size(buf), buf);
}
}
/*
* Decrement the amount of evictable space in the arc_state_t's refcount.
* We account for the space used by the hdr and the arc buf individually
* so that we can add and remove them from the refcount individually.
*/
static void
arc_evictable_space_decrement(arc_buf_hdr_t *hdr, arc_state_t *state)
{
arc_buf_contents_t type = arc_buf_type(hdr);
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(state)) {
ASSERT0(hdr->b_l1hdr.b_bufcnt);
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
HDR_GET_LSIZE(hdr), hdr);
return;
}
if (hdr->b_l1hdr.b_pabd != NULL) {
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
arc_hdr_size(hdr), hdr);
}
if (HDR_HAS_RABD(hdr)) {
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
HDR_GET_PSIZE(hdr), hdr);
}
for (arc_buf_t *buf = hdr->b_l1hdr.b_buf; buf != NULL;
buf = buf->b_next) {
if (arc_buf_is_shared(buf))
continue;
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
arc_buf_size(buf), buf);
}
}
/*
* Add a reference to this hdr indicating that someone is actively
* referencing that memory. When the refcount transitions from 0 to 1,
* we remove it from the respective arc_state_t list to indicate that
* it is not evictable.
*/
static void
add_reference(arc_buf_hdr_t *hdr, const void *tag)
{
arc_state_t *state = hdr->b_l1hdr.b_state;
ASSERT(HDR_HAS_L1HDR(hdr));
if (!HDR_EMPTY(hdr) && !MUTEX_HELD(HDR_LOCK(hdr))) {
ASSERT(state == arc_anon);
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
}
if ((zfs_refcount_add(&hdr->b_l1hdr.b_refcnt, tag) == 1) &&
state != arc_anon && state != arc_l2c_only) {
/* We don't use the L2-only state list. */
multilist_remove(&state->arcs_list[arc_buf_type(hdr)], hdr);
arc_evictable_space_decrement(hdr, state);
}
}
/*
* Remove a reference from this hdr. When the reference transitions from
* 1 to 0 and we're not anonymous, then we add this hdr to the arc_state_t's
* list making it eligible for eviction.
*/
static int
remove_reference(arc_buf_hdr_t *hdr, const void *tag)
{
int cnt;
arc_state_t *state = hdr->b_l1hdr.b_state;
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(state == arc_anon || MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT(!GHOST_STATE(state)); /* arc_l2c_only counts as a ghost. */
if ((cnt = zfs_refcount_remove(&hdr->b_l1hdr.b_refcnt, tag)) != 0)
return (cnt);
if (state == arc_anon) {
arc_hdr_destroy(hdr);
return (0);
}
if (state == arc_uncached && !HDR_PREFETCH(hdr)) {
arc_change_state(arc_anon, hdr);
arc_hdr_destroy(hdr);
return (0);
}
multilist_insert(&state->arcs_list[arc_buf_type(hdr)], hdr);
arc_evictable_space_increment(hdr, state);
return (0);
}
/*
* Returns detailed information about a specific arc buffer. When the
* state_index argument is set the function will calculate the arc header
* list position for its arc state. Since this requires a linear traversal
* callers are strongly encourage not to do this. However, it can be helpful
* for targeted analysis so the functionality is provided.
*/
void
arc_buf_info(arc_buf_t *ab, arc_buf_info_t *abi, int state_index)
{
(void) state_index;
arc_buf_hdr_t *hdr = ab->b_hdr;
l1arc_buf_hdr_t *l1hdr = NULL;
l2arc_buf_hdr_t *l2hdr = NULL;
arc_state_t *state = NULL;
memset(abi, 0, sizeof (arc_buf_info_t));
if (hdr == NULL)
return;
abi->abi_flags = hdr->b_flags;
if (HDR_HAS_L1HDR(hdr)) {
l1hdr = &hdr->b_l1hdr;
state = l1hdr->b_state;
}
if (HDR_HAS_L2HDR(hdr))
l2hdr = &hdr->b_l2hdr;
if (l1hdr) {
abi->abi_bufcnt = l1hdr->b_bufcnt;
abi->abi_access = l1hdr->b_arc_access;
abi->abi_mru_hits = l1hdr->b_mru_hits;
abi->abi_mru_ghost_hits = l1hdr->b_mru_ghost_hits;
abi->abi_mfu_hits = l1hdr->b_mfu_hits;
abi->abi_mfu_ghost_hits = l1hdr->b_mfu_ghost_hits;
abi->abi_holds = zfs_refcount_count(&l1hdr->b_refcnt);
}
if (l2hdr) {
abi->abi_l2arc_dattr = l2hdr->b_daddr;
abi->abi_l2arc_hits = l2hdr->b_hits;
}
abi->abi_state_type = state ? state->arcs_state : ARC_STATE_ANON;
abi->abi_state_contents = arc_buf_type(hdr);
abi->abi_size = arc_hdr_size(hdr);
}
/*
* Move the supplied buffer to the indicated state. The hash lock
* for the buffer must be held by the caller.
*/
static void
arc_change_state(arc_state_t *new_state, arc_buf_hdr_t *hdr)
{
arc_state_t *old_state;
int64_t refcnt;
uint32_t bufcnt;
boolean_t update_old, update_new;
arc_buf_contents_t type = arc_buf_type(hdr);
/*
* We almost always have an L1 hdr here, since we call arc_hdr_realloc()
* in arc_read() when bringing a buffer out of the L2ARC. However, the
* L1 hdr doesn't always exist when we change state to arc_anon before
* destroying a header, in which case reallocating to add the L1 hdr is
* pointless.
*/
if (HDR_HAS_L1HDR(hdr)) {
old_state = hdr->b_l1hdr.b_state;
refcnt = zfs_refcount_count(&hdr->b_l1hdr.b_refcnt);
bufcnt = hdr->b_l1hdr.b_bufcnt;
update_old = (bufcnt > 0 || hdr->b_l1hdr.b_pabd != NULL ||
HDR_HAS_RABD(hdr));
IMPLY(GHOST_STATE(old_state), bufcnt == 0);
IMPLY(GHOST_STATE(new_state), bufcnt == 0);
IMPLY(GHOST_STATE(old_state), hdr->b_l1hdr.b_buf == NULL);
IMPLY(GHOST_STATE(new_state), hdr->b_l1hdr.b_buf == NULL);
IMPLY(old_state == arc_anon, bufcnt <= 1);
} else {
old_state = arc_l2c_only;
refcnt = 0;
bufcnt = 0;
update_old = B_FALSE;
}
update_new = update_old;
if (GHOST_STATE(old_state))
update_old = B_TRUE;
if (GHOST_STATE(new_state))
update_new = B_TRUE;
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT3P(new_state, !=, old_state);
/*
* If this buffer is evictable, transfer it from the
* old state list to the new state list.
*/
if (refcnt == 0) {
if (old_state != arc_anon && old_state != arc_l2c_only) {
ASSERT(HDR_HAS_L1HDR(hdr));
/* remove_reference() saves on insert. */
if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) {
multilist_remove(&old_state->arcs_list[type],
hdr);
arc_evictable_space_decrement(hdr, old_state);
}
}
if (new_state != arc_anon && new_state != arc_l2c_only) {
/*
* An L1 header always exists here, since if we're
* moving to some L1-cached state (i.e. not l2c_only or
* anonymous), we realloc the header to add an L1hdr
* beforehand.
*/
ASSERT(HDR_HAS_L1HDR(hdr));
multilist_insert(&new_state->arcs_list[type], hdr);
arc_evictable_space_increment(hdr, new_state);
}
}
ASSERT(!HDR_EMPTY(hdr));
if (new_state == arc_anon && HDR_IN_HASH_TABLE(hdr))
buf_hash_remove(hdr);
/* adjust state sizes (ignore arc_l2c_only) */
if (update_new && new_state != arc_l2c_only) {
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(new_state)) {
ASSERT0(bufcnt);
/*
* When moving a header to a ghost state, we first
* remove all arc buffers. Thus, we'll have a
* bufcnt of zero, and no arc buffer to use for
* the reference. As a result, we use the arc
* header pointer for the reference.
*/
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
HDR_GET_LSIZE(hdr), hdr);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
} else {
uint32_t buffers = 0;
/*
* Each individual buffer holds a unique reference,
* thus we must remove each of these references one
* at a time.
*/
for (arc_buf_t *buf = hdr->b_l1hdr.b_buf; buf != NULL;
buf = buf->b_next) {
ASSERT3U(bufcnt, !=, 0);
buffers++;
/*
* When the arc_buf_t is sharing the data
* block with the hdr, the owner of the
* reference belongs to the hdr. Only
* add to the refcount if the arc_buf_t is
* not shared.
*/
if (arc_buf_is_shared(buf))
continue;
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
arc_buf_size(buf), buf);
}
ASSERT3U(bufcnt, ==, buffers);
if (hdr->b_l1hdr.b_pabd != NULL) {
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
arc_hdr_size(hdr), hdr);
}
if (HDR_HAS_RABD(hdr)) {
(void) zfs_refcount_add_many(
&new_state->arcs_size[type],
HDR_GET_PSIZE(hdr), hdr);
}
}
}
if (update_old && old_state != arc_l2c_only) {
ASSERT(HDR_HAS_L1HDR(hdr));
if (GHOST_STATE(old_state)) {
ASSERT0(bufcnt);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
/*
* When moving a header off of a ghost state,
* the header will not contain any arc buffers.
* We use the arc header pointer for the reference
* which is exactly what we did when we put the
* header on the ghost state.
*/
(void) zfs_refcount_remove_many(
&old_state->arcs_size[type],
HDR_GET_LSIZE(hdr), hdr);
} else {
uint32_t buffers = 0;
/*
* Each individual buffer holds a unique reference,
* thus we must remove each of these references one
* at a time.
*/
for (arc_buf_t *buf = hdr->b_l1hdr.b_buf; buf != NULL;
buf = buf->b_next) {
ASSERT3U(bufcnt, !=, 0);
buffers++;
/*
* When the arc_buf_t is sharing the data
* block with the hdr, the owner of the
* reference belongs to the hdr. Only
* add to the refcount if the arc_buf_t is
* not shared.
*/
if (arc_buf_is_shared(buf))
continue;
(void) zfs_refcount_remove_many(
&old_state->arcs_size[type],
arc_buf_size(buf), buf);
}
ASSERT3U(bufcnt, ==, buffers);
ASSERT(hdr->b_l1hdr.b_pabd != NULL ||
HDR_HAS_RABD(hdr));
if (hdr->b_l1hdr.b_pabd != NULL) {
(void) zfs_refcount_remove_many(
&old_state->arcs_size[type],
arc_hdr_size(hdr), hdr);
}
if (HDR_HAS_RABD(hdr)) {
(void) zfs_refcount_remove_many(
&old_state->arcs_size[type],
HDR_GET_PSIZE(hdr), hdr);
}
}
}
if (HDR_HAS_L1HDR(hdr)) {
hdr->b_l1hdr.b_state = new_state;
if (HDR_HAS_L2HDR(hdr) && new_state != arc_l2c_only) {
l2arc_hdr_arcstats_decrement_state(hdr);
hdr->b_l2hdr.b_arcs_state = new_state->arcs_state;
l2arc_hdr_arcstats_increment_state(hdr);
}
}
}
void
arc_space_consume(uint64_t space, arc_space_type_t type)
{
ASSERT(type >= 0 && type < ARC_SPACE_NUMTYPES);
switch (type) {
default:
break;
case ARC_SPACE_DATA:
ARCSTAT_INCR(arcstat_data_size, space);
break;
case ARC_SPACE_META:
ARCSTAT_INCR(arcstat_metadata_size, space);
break;
case ARC_SPACE_BONUS:
ARCSTAT_INCR(arcstat_bonus_size, space);
break;
case ARC_SPACE_DNODE:
ARCSTAT_INCR(arcstat_dnode_size, space);
break;
case ARC_SPACE_DBUF:
ARCSTAT_INCR(arcstat_dbuf_size, space);
break;
case ARC_SPACE_HDRS:
ARCSTAT_INCR(arcstat_hdr_size, space);
break;
case ARC_SPACE_L2HDRS:
aggsum_add(&arc_sums.arcstat_l2_hdr_size, space);
break;
case ARC_SPACE_ABD_CHUNK_WASTE:
/*
* Note: this includes space wasted by all scatter ABD's, not
* just those allocated by the ARC. But the vast majority of
* scatter ABD's come from the ARC, because other users are
* very short-lived.
*/
ARCSTAT_INCR(arcstat_abd_chunk_waste_size, space);
break;
}
if (type != ARC_SPACE_DATA && type != ARC_SPACE_ABD_CHUNK_WASTE)
ARCSTAT_INCR(arcstat_meta_used, space);
aggsum_add(&arc_sums.arcstat_size, space);
}
void
arc_space_return(uint64_t space, arc_space_type_t type)
{
ASSERT(type >= 0 && type < ARC_SPACE_NUMTYPES);
switch (type) {
default:
break;
case ARC_SPACE_DATA:
ARCSTAT_INCR(arcstat_data_size, -space);
break;
case ARC_SPACE_META:
ARCSTAT_INCR(arcstat_metadata_size, -space);
break;
case ARC_SPACE_BONUS:
ARCSTAT_INCR(arcstat_bonus_size, -space);
break;
case ARC_SPACE_DNODE:
ARCSTAT_INCR(arcstat_dnode_size, -space);
break;
case ARC_SPACE_DBUF:
ARCSTAT_INCR(arcstat_dbuf_size, -space);
break;
case ARC_SPACE_HDRS:
ARCSTAT_INCR(arcstat_hdr_size, -space);
break;
case ARC_SPACE_L2HDRS:
aggsum_add(&arc_sums.arcstat_l2_hdr_size, -space);
break;
case ARC_SPACE_ABD_CHUNK_WASTE:
ARCSTAT_INCR(arcstat_abd_chunk_waste_size, -space);
break;
}
if (type != ARC_SPACE_DATA && type != ARC_SPACE_ABD_CHUNK_WASTE)
ARCSTAT_INCR(arcstat_meta_used, -space);
ASSERT(aggsum_compare(&arc_sums.arcstat_size, space) >= 0);
aggsum_add(&arc_sums.arcstat_size, -space);
}
/*
* Given a hdr and a buf, returns whether that buf can share its b_data buffer
* with the hdr's b_pabd.
*/
static boolean_t
arc_can_share(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
/*
* The criteria for sharing a hdr's data are:
* 1. the buffer is not encrypted
* 2. the hdr's compression matches the buf's compression
* 3. the hdr doesn't need to be byteswapped
* 4. the hdr isn't already being shared
* 5. the buf is either compressed or it is the last buf in the hdr list
*
* Criterion #5 maintains the invariant that shared uncompressed
* bufs must be the final buf in the hdr's b_buf list. Reading this, you
* might ask, "if a compressed buf is allocated first, won't that be the
* last thing in the list?", but in that case it's impossible to create
* a shared uncompressed buf anyway (because the hdr must be compressed
* to have the compressed buf). You might also think that #3 is
* sufficient to make this guarantee, however it's possible
* (specifically in the rare L2ARC write race mentioned in
* arc_buf_alloc_impl()) there will be an existing uncompressed buf that
* is shareable, but wasn't at the time of its allocation. Rather than
* allow a new shared uncompressed buf to be created and then shuffle
* the list around to make it the last element, this simply disallows
* sharing if the new buf isn't the first to be added.
*/
ASSERT3P(buf->b_hdr, ==, hdr);
boolean_t hdr_compressed =
arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF;
boolean_t buf_compressed = ARC_BUF_COMPRESSED(buf) != 0;
return (!ARC_BUF_ENCRYPTED(buf) &&
buf_compressed == hdr_compressed &&
hdr->b_l1hdr.b_byteswap == DMU_BSWAP_NUMFUNCS &&
!HDR_SHARED_DATA(hdr) &&
(ARC_BUF_LAST(buf) || ARC_BUF_COMPRESSED(buf)));
}
/*
* Allocate a buf for this hdr. If you care about the data that's in the hdr,
* or if you want a compressed buffer, pass those flags in. Returns 0 if the
* copy was made successfully, or an error code otherwise.
*/
static int
arc_buf_alloc_impl(arc_buf_hdr_t *hdr, spa_t *spa, const zbookmark_phys_t *zb,
const void *tag, boolean_t encrypted, boolean_t compressed,
boolean_t noauth, boolean_t fill, arc_buf_t **ret)
{
arc_buf_t *buf;
arc_fill_flags_t flags = ARC_FILL_LOCKED;
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT3U(HDR_GET_LSIZE(hdr), >, 0);
VERIFY(hdr->b_type == ARC_BUFC_DATA ||
hdr->b_type == ARC_BUFC_METADATA);
ASSERT3P(ret, !=, NULL);
ASSERT3P(*ret, ==, NULL);
IMPLY(encrypted, compressed);
buf = *ret = kmem_cache_alloc(buf_cache, KM_PUSHPAGE);
buf->b_hdr = hdr;
buf->b_data = NULL;
buf->b_next = hdr->b_l1hdr.b_buf;
buf->b_flags = 0;
add_reference(hdr, tag);
/*
* We're about to change the hdr's b_flags. We must either
* hold the hash_lock or be undiscoverable.
*/
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
/*
* Only honor requests for compressed bufs if the hdr is actually
* compressed. This must be overridden if the buffer is encrypted since
* encrypted buffers cannot be decompressed.
*/
if (encrypted) {
buf->b_flags |= ARC_BUF_FLAG_COMPRESSED;
buf->b_flags |= ARC_BUF_FLAG_ENCRYPTED;
flags |= ARC_FILL_COMPRESSED | ARC_FILL_ENCRYPTED;
} else if (compressed &&
arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF) {
buf->b_flags |= ARC_BUF_FLAG_COMPRESSED;
flags |= ARC_FILL_COMPRESSED;
}
if (noauth) {
ASSERT0(encrypted);
flags |= ARC_FILL_NOAUTH;
}
/*
* If the hdr's data can be shared then we share the data buffer and
* set the appropriate bit in the hdr's b_flags to indicate the hdr is
* sharing it's b_pabd with the arc_buf_t. Otherwise, we allocate a new
* buffer to store the buf's data.
*
* There are two additional restrictions here because we're sharing
* hdr -> buf instead of the usual buf -> hdr. First, the hdr can't be
* actively involved in an L2ARC write, because if this buf is used by
* an arc_write() then the hdr's data buffer will be released when the
* write completes, even though the L2ARC write might still be using it.
* Second, the hdr's ABD must be linear so that the buf's user doesn't
* need to be ABD-aware. It must be allocated via
* zio_[data_]buf_alloc(), not as a page, because we need to be able
* to abd_release_ownership_of_buf(), which isn't allowed on "linear
* page" buffers because the ABD code needs to handle freeing them
* specially.
*/
boolean_t can_share = arc_can_share(hdr, buf) &&
!HDR_L2_WRITING(hdr) &&
hdr->b_l1hdr.b_pabd != NULL &&
abd_is_linear(hdr->b_l1hdr.b_pabd) &&
!abd_is_linear_page(hdr->b_l1hdr.b_pabd);
/* Set up b_data and sharing */
if (can_share) {
buf->b_data = abd_to_buf(hdr->b_l1hdr.b_pabd);
buf->b_flags |= ARC_BUF_FLAG_SHARED;
arc_hdr_set_flags(hdr, ARC_FLAG_SHARED_DATA);
} else {
buf->b_data =
arc_get_data_buf(hdr, arc_buf_size(buf), buf);
ARCSTAT_INCR(arcstat_overhead_size, arc_buf_size(buf));
}
VERIFY3P(buf->b_data, !=, NULL);
hdr->b_l1hdr.b_buf = buf;
hdr->b_l1hdr.b_bufcnt += 1;
if (encrypted)
hdr->b_crypt_hdr.b_ebufcnt += 1;
/*
* If the user wants the data from the hdr, we need to either copy or
* decompress the data.
*/
if (fill) {
ASSERT3P(zb, !=, NULL);
return (arc_buf_fill(buf, spa, zb, flags));
}
return (0);
}
static const char *arc_onloan_tag = "onloan";
static inline void
arc_loaned_bytes_update(int64_t delta)
{
atomic_add_64(&arc_loaned_bytes, delta);
/* assert that it did not wrap around */
ASSERT3S(atomic_add_64_nv(&arc_loaned_bytes, 0), >=, 0);
}
/*
* Loan out an anonymous arc buffer. Loaned buffers are not counted as in
* flight data by arc_tempreserve_space() until they are "returned". Loaned
* buffers must be returned to the arc before they can be used by the DMU or
* freed.
*/
arc_buf_t *
arc_loan_buf(spa_t *spa, boolean_t is_metadata, int size)
{
arc_buf_t *buf = arc_alloc_buf(spa, arc_onloan_tag,
is_metadata ? ARC_BUFC_METADATA : ARC_BUFC_DATA, size);
arc_loaned_bytes_update(arc_buf_size(buf));
return (buf);
}
arc_buf_t *
arc_loan_compressed_buf(spa_t *spa, uint64_t psize, uint64_t lsize,
enum zio_compress compression_type, uint8_t complevel)
{
arc_buf_t *buf = arc_alloc_compressed_buf(spa, arc_onloan_tag,
psize, lsize, compression_type, complevel);
arc_loaned_bytes_update(arc_buf_size(buf));
return (buf);
}
arc_buf_t *
arc_loan_raw_buf(spa_t *spa, uint64_t dsobj, boolean_t byteorder,
const uint8_t *salt, const uint8_t *iv, const uint8_t *mac,
dmu_object_type_t ot, uint64_t psize, uint64_t lsize,
enum zio_compress compression_type, uint8_t complevel)
{
arc_buf_t *buf = arc_alloc_raw_buf(spa, arc_onloan_tag, dsobj,
byteorder, salt, iv, mac, ot, psize, lsize, compression_type,
complevel);
atomic_add_64(&arc_loaned_bytes, psize);
return (buf);
}
/*
* Return a loaned arc buffer to the arc.
*/
void
arc_return_buf(arc_buf_t *buf, const void *tag)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT3P(buf->b_data, !=, NULL);
ASSERT(HDR_HAS_L1HDR(hdr));
(void) zfs_refcount_add(&hdr->b_l1hdr.b_refcnt, tag);
(void) zfs_refcount_remove(&hdr->b_l1hdr.b_refcnt, arc_onloan_tag);
arc_loaned_bytes_update(-arc_buf_size(buf));
}
/* Detach an arc_buf from a dbuf (tag) */
void
arc_loan_inuse_buf(arc_buf_t *buf, const void *tag)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT3P(buf->b_data, !=, NULL);
ASSERT(HDR_HAS_L1HDR(hdr));
(void) zfs_refcount_add(&hdr->b_l1hdr.b_refcnt, arc_onloan_tag);
(void) zfs_refcount_remove(&hdr->b_l1hdr.b_refcnt, tag);
arc_loaned_bytes_update(arc_buf_size(buf));
}
static void
l2arc_free_abd_on_write(abd_t *abd, size_t size, arc_buf_contents_t type)
{
l2arc_data_free_t *df = kmem_alloc(sizeof (*df), KM_SLEEP);
df->l2df_abd = abd;
df->l2df_size = size;
df->l2df_type = type;
mutex_enter(&l2arc_free_on_write_mtx);
list_insert_head(l2arc_free_on_write, df);
mutex_exit(&l2arc_free_on_write_mtx);
}
static void
arc_hdr_free_on_write(arc_buf_hdr_t *hdr, boolean_t free_rdata)
{
arc_state_t *state = hdr->b_l1hdr.b_state;
arc_buf_contents_t type = arc_buf_type(hdr);
uint64_t size = (free_rdata) ? HDR_GET_PSIZE(hdr) : arc_hdr_size(hdr);
/* protected by hash lock, if in the hash table */
if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) {
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
ASSERT(state != arc_anon && state != arc_l2c_only);
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
size, hdr);
}
(void) zfs_refcount_remove_many(&state->arcs_size[type], size, hdr);
if (type == ARC_BUFC_METADATA) {
arc_space_return(size, ARC_SPACE_META);
} else {
ASSERT(type == ARC_BUFC_DATA);
arc_space_return(size, ARC_SPACE_DATA);
}
if (free_rdata) {
l2arc_free_abd_on_write(hdr->b_crypt_hdr.b_rabd, size, type);
} else {
l2arc_free_abd_on_write(hdr->b_l1hdr.b_pabd, size, type);
}
}
/*
* Share the arc_buf_t's data with the hdr. Whenever we are sharing the
* data buffer, we transfer the refcount ownership to the hdr and update
* the appropriate kstats.
*/
static void
arc_share_buf(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
ASSERT(arc_can_share(hdr, buf));
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!ARC_BUF_ENCRYPTED(buf));
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
/*
* Start sharing the data buffer. We transfer the
* refcount ownership to the hdr since it always owns
* the refcount whenever an arc_buf_t is shared.
*/
zfs_refcount_transfer_ownership_many(
&hdr->b_l1hdr.b_state->arcs_size[arc_buf_type(hdr)],
arc_hdr_size(hdr), buf, hdr);
hdr->b_l1hdr.b_pabd = abd_get_from_buf(buf->b_data, arc_buf_size(buf));
abd_take_ownership_of_buf(hdr->b_l1hdr.b_pabd,
HDR_ISTYPE_METADATA(hdr));
arc_hdr_set_flags(hdr, ARC_FLAG_SHARED_DATA);
buf->b_flags |= ARC_BUF_FLAG_SHARED;
/*
* Since we've transferred ownership to the hdr we need
* to increment its compressed and uncompressed kstats and
* decrement the overhead size.
*/
ARCSTAT_INCR(arcstat_compressed_size, arc_hdr_size(hdr));
ARCSTAT_INCR(arcstat_uncompressed_size, HDR_GET_LSIZE(hdr));
ARCSTAT_INCR(arcstat_overhead_size, -arc_buf_size(buf));
}
static void
arc_unshare_buf(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
ASSERT(arc_buf_is_shared(buf));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
/*
* We are no longer sharing this buffer so we need
* to transfer its ownership to the rightful owner.
*/
zfs_refcount_transfer_ownership_many(
&hdr->b_l1hdr.b_state->arcs_size[arc_buf_type(hdr)],
arc_hdr_size(hdr), hdr, buf);
arc_hdr_clear_flags(hdr, ARC_FLAG_SHARED_DATA);
abd_release_ownership_of_buf(hdr->b_l1hdr.b_pabd);
abd_free(hdr->b_l1hdr.b_pabd);
hdr->b_l1hdr.b_pabd = NULL;
buf->b_flags &= ~ARC_BUF_FLAG_SHARED;
/*
* Since the buffer is no longer shared between
* the arc buf and the hdr, count it as overhead.
*/
ARCSTAT_INCR(arcstat_compressed_size, -arc_hdr_size(hdr));
ARCSTAT_INCR(arcstat_uncompressed_size, -HDR_GET_LSIZE(hdr));
ARCSTAT_INCR(arcstat_overhead_size, arc_buf_size(buf));
}
/*
* Remove an arc_buf_t from the hdr's buf list and return the last
* arc_buf_t on the list. If no buffers remain on the list then return
* NULL.
*/
static arc_buf_t *
arc_buf_remove(arc_buf_hdr_t *hdr, arc_buf_t *buf)
{
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
arc_buf_t **bufp = &hdr->b_l1hdr.b_buf;
arc_buf_t *lastbuf = NULL;
/*
* Remove the buf from the hdr list and locate the last
* remaining buffer on the list.
*/
while (*bufp != NULL) {
if (*bufp == buf)
*bufp = buf->b_next;
/*
* If we've removed a buffer in the middle of
* the list then update the lastbuf and update
* bufp.
*/
if (*bufp != NULL) {
lastbuf = *bufp;
bufp = &(*bufp)->b_next;
}
}
buf->b_next = NULL;
ASSERT3P(lastbuf, !=, buf);
IMPLY(hdr->b_l1hdr.b_bufcnt > 0, lastbuf != NULL);
IMPLY(hdr->b_l1hdr.b_bufcnt > 0, hdr->b_l1hdr.b_buf != NULL);
IMPLY(lastbuf != NULL, ARC_BUF_LAST(lastbuf));
return (lastbuf);
}
/*
* Free up buf->b_data and pull the arc_buf_t off of the arc_buf_hdr_t's
* list and free it.
*/
static void
arc_buf_destroy_impl(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
/*
* Free up the data associated with the buf but only if we're not
* sharing this with the hdr. If we are sharing it with the hdr, the
* hdr is responsible for doing the free.
*/
if (buf->b_data != NULL) {
/*
* We're about to change the hdr's b_flags. We must either
* hold the hash_lock or be undiscoverable.
*/
ASSERT(HDR_EMPTY_OR_LOCKED(hdr));
arc_cksum_verify(buf);
arc_buf_unwatch(buf);
if (arc_buf_is_shared(buf)) {
arc_hdr_clear_flags(hdr, ARC_FLAG_SHARED_DATA);
} else {
uint64_t size = arc_buf_size(buf);
arc_free_data_buf(hdr, buf->b_data, size, buf);
ARCSTAT_INCR(arcstat_overhead_size, -size);
}
buf->b_data = NULL;
ASSERT(hdr->b_l1hdr.b_bufcnt > 0);
hdr->b_l1hdr.b_bufcnt -= 1;
if (ARC_BUF_ENCRYPTED(buf)) {
hdr->b_crypt_hdr.b_ebufcnt -= 1;
/*
* If we have no more encrypted buffers and we've
* already gotten a copy of the decrypted data we can
* free b_rabd to save some space.
*/
if (hdr->b_crypt_hdr.b_ebufcnt == 0 &&
HDR_HAS_RABD(hdr) && hdr->b_l1hdr.b_pabd != NULL &&
!HDR_IO_IN_PROGRESS(hdr)) {
arc_hdr_free_abd(hdr, B_TRUE);
}
}
}
arc_buf_t *lastbuf = arc_buf_remove(hdr, buf);
if (ARC_BUF_SHARED(buf) && !ARC_BUF_COMPRESSED(buf)) {
/*
* If the current arc_buf_t is sharing its data buffer with the
* hdr, then reassign the hdr's b_pabd to share it with the new
* buffer at the end of the list. The shared buffer is always
* the last one on the hdr's buffer list.
*
* There is an equivalent case for compressed bufs, but since
* they aren't guaranteed to be the last buf in the list and
* that is an exceedingly rare case, we just allow that space be
* wasted temporarily. We must also be careful not to share
* encrypted buffers, since they cannot be shared.
*/
if (lastbuf != NULL && !ARC_BUF_ENCRYPTED(lastbuf)) {
/* Only one buf can be shared at once */
VERIFY(!arc_buf_is_shared(lastbuf));
/* hdr is uncompressed so can't have compressed buf */
VERIFY(!ARC_BUF_COMPRESSED(lastbuf));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
arc_hdr_free_abd(hdr, B_FALSE);
/*
* We must setup a new shared block between the
* last buffer and the hdr. The data would have
* been allocated by the arc buf so we need to transfer
* ownership to the hdr since it's now being shared.
*/
arc_share_buf(hdr, lastbuf);
}
} else if (HDR_SHARED_DATA(hdr)) {
/*
* Uncompressed shared buffers are always at the end
* of the list. Compressed buffers don't have the
* same requirements. This makes it hard to
* simply assert that the lastbuf is shared so
* we rely on the hdr's compression flags to determine
* if we have a compressed, shared buffer.
*/
ASSERT3P(lastbuf, !=, NULL);
ASSERT(arc_buf_is_shared(lastbuf) ||
arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF);
}
/*
* Free the checksum if we're removing the last uncompressed buf from
* this hdr.
*/
if (!arc_hdr_has_uncompressed_buf(hdr)) {
arc_cksum_free(hdr);
}
/* clean up the buf */
buf->b_hdr = NULL;
kmem_cache_free(buf_cache, buf);
}
static void
arc_hdr_alloc_abd(arc_buf_hdr_t *hdr, int alloc_flags)
{
uint64_t size;
boolean_t alloc_rdata = ((alloc_flags & ARC_HDR_ALLOC_RDATA) != 0);
ASSERT3U(HDR_GET_LSIZE(hdr), >, 0);
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(!HDR_SHARED_DATA(hdr) || alloc_rdata);
IMPLY(alloc_rdata, HDR_PROTECTED(hdr));
if (alloc_rdata) {
size = HDR_GET_PSIZE(hdr);
ASSERT3P(hdr->b_crypt_hdr.b_rabd, ==, NULL);
hdr->b_crypt_hdr.b_rabd = arc_get_data_abd(hdr, size, hdr,
alloc_flags);
ASSERT3P(hdr->b_crypt_hdr.b_rabd, !=, NULL);
ARCSTAT_INCR(arcstat_raw_size, size);
} else {
size = arc_hdr_size(hdr);
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
hdr->b_l1hdr.b_pabd = arc_get_data_abd(hdr, size, hdr,
alloc_flags);
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
}
ARCSTAT_INCR(arcstat_compressed_size, size);
ARCSTAT_INCR(arcstat_uncompressed_size, HDR_GET_LSIZE(hdr));
}
static void
arc_hdr_free_abd(arc_buf_hdr_t *hdr, boolean_t free_rdata)
{
uint64_t size = (free_rdata) ? HDR_GET_PSIZE(hdr) : arc_hdr_size(hdr);
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(hdr->b_l1hdr.b_pabd != NULL || HDR_HAS_RABD(hdr));
IMPLY(free_rdata, HDR_HAS_RABD(hdr));
/*
* If the hdr is currently being written to the l2arc then
* we defer freeing the data by adding it to the l2arc_free_on_write
* list. The l2arc will free the data once it's finished
* writing it to the l2arc device.
*/
if (HDR_L2_WRITING(hdr)) {
arc_hdr_free_on_write(hdr, free_rdata);
ARCSTAT_BUMP(arcstat_l2_free_on_write);
} else if (free_rdata) {
arc_free_data_abd(hdr, hdr->b_crypt_hdr.b_rabd, size, hdr);
} else {
arc_free_data_abd(hdr, hdr->b_l1hdr.b_pabd, size, hdr);
}
if (free_rdata) {
hdr->b_crypt_hdr.b_rabd = NULL;
ARCSTAT_INCR(arcstat_raw_size, -size);
} else {
hdr->b_l1hdr.b_pabd = NULL;
}
if (hdr->b_l1hdr.b_pabd == NULL && !HDR_HAS_RABD(hdr))
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_NUMFUNCS;
ARCSTAT_INCR(arcstat_compressed_size, -size);
ARCSTAT_INCR(arcstat_uncompressed_size, -HDR_GET_LSIZE(hdr));
}
/*
* Allocate empty anonymous ARC header. The header will get its identity
* assigned and buffers attached later as part of read or write operations.
*
* In case of read arc_read() assigns header its identify (b_dva + b_birth),
* inserts it into ARC hash to become globally visible and allocates physical
* (b_pabd) or raw (b_rabd) ABD buffer to read into from disk. On disk read
* completion arc_read_done() allocates ARC buffer(s) as needed, potentially
* sharing one of them with the physical ABD buffer.
*
* In case of write arc_alloc_buf() allocates ARC buffer to be filled with
* data. Then after compression and/or encryption arc_write_ready() allocates
* and fills (or potentially shares) physical (b_pabd) or raw (b_rabd) ABD
* buffer. On disk write completion arc_write_done() assigns the header its
* new identity (b_dva + b_birth) and inserts into ARC hash.
*
* In case of partial overwrite the old data is read first as described. Then
* arc_release() either allocates new anonymous ARC header and moves the ARC
* buffer to it, or reuses the old ARC header by discarding its identity and
* removing it from ARC hash. After buffer modification normal write process
* follows as described.
*/
static arc_buf_hdr_t *
arc_hdr_alloc(uint64_t spa, int32_t psize, int32_t lsize,
boolean_t protected, enum zio_compress compression_type, uint8_t complevel,
arc_buf_contents_t type)
{
arc_buf_hdr_t *hdr;
VERIFY(type == ARC_BUFC_DATA || type == ARC_BUFC_METADATA);
if (protected) {
hdr = kmem_cache_alloc(hdr_full_crypt_cache, KM_PUSHPAGE);
} else {
hdr = kmem_cache_alloc(hdr_full_cache, KM_PUSHPAGE);
}
ASSERT(HDR_EMPTY(hdr));
#ifdef ZFS_DEBUG
ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
#endif
HDR_SET_PSIZE(hdr, psize);
HDR_SET_LSIZE(hdr, lsize);
hdr->b_spa = spa;
hdr->b_type = type;
hdr->b_flags = 0;
arc_hdr_set_flags(hdr, arc_bufc_to_flags(type) | ARC_FLAG_HAS_L1HDR);
arc_hdr_set_compress(hdr, compression_type);
hdr->b_complevel = complevel;
if (protected)
arc_hdr_set_flags(hdr, ARC_FLAG_PROTECTED);
hdr->b_l1hdr.b_state = arc_anon;
hdr->b_l1hdr.b_arc_access = 0;
hdr->b_l1hdr.b_mru_hits = 0;
hdr->b_l1hdr.b_mru_ghost_hits = 0;
hdr->b_l1hdr.b_mfu_hits = 0;
hdr->b_l1hdr.b_mfu_ghost_hits = 0;
hdr->b_l1hdr.b_bufcnt = 0;
hdr->b_l1hdr.b_buf = NULL;
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
return (hdr);
}
/*
* Transition between the two allocation states for the arc_buf_hdr struct.
* The arc_buf_hdr struct can be allocated with (hdr_full_cache) or without
* (hdr_l2only_cache) the fields necessary for the L1 cache - the smaller
* version is used when a cache buffer is only in the L2ARC in order to reduce
* memory usage.
*/
static arc_buf_hdr_t *
arc_hdr_realloc(arc_buf_hdr_t *hdr, kmem_cache_t *old, kmem_cache_t *new)
{
ASSERT(HDR_HAS_L2HDR(hdr));
arc_buf_hdr_t *nhdr;
l2arc_dev_t *dev = hdr->b_l2hdr.b_dev;
ASSERT((old == hdr_full_cache && new == hdr_l2only_cache) ||
(old == hdr_l2only_cache && new == hdr_full_cache));
/*
* if the caller wanted a new full header and the header is to be
* encrypted we will actually allocate the header from the full crypt
* cache instead. The same applies to freeing from the old cache.
*/
if (HDR_PROTECTED(hdr) && new == hdr_full_cache)
new = hdr_full_crypt_cache;
if (HDR_PROTECTED(hdr) && old == hdr_full_cache)
old = hdr_full_crypt_cache;
nhdr = kmem_cache_alloc(new, KM_PUSHPAGE);
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
buf_hash_remove(hdr);
memcpy(nhdr, hdr, HDR_L2ONLY_SIZE);
if (new == hdr_full_cache || new == hdr_full_crypt_cache) {
arc_hdr_set_flags(nhdr, ARC_FLAG_HAS_L1HDR);
/*
* arc_access and arc_change_state need to be aware that a
* header has just come out of L2ARC, so we set its state to
* l2c_only even though it's about to change.
*/
nhdr->b_l1hdr.b_state = arc_l2c_only;
/* Verify previous threads set to NULL before freeing */
ASSERT3P(nhdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
} else {
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
ASSERT0(hdr->b_l1hdr.b_bufcnt);
#ifdef ZFS_DEBUG
ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
#endif
/*
* If we've reached here, We must have been called from
* arc_evict_hdr(), as such we should have already been
* removed from any ghost list we were previously on
* (which protects us from racing with arc_evict_state),
* thus no locking is needed during this check.
*/
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
/*
* A buffer must not be moved into the arc_l2c_only
* state if it's not finished being written out to the
* l2arc device. Otherwise, the b_l1hdr.b_pabd field
* might try to be accessed, even though it was removed.
*/
VERIFY(!HDR_L2_WRITING(hdr));
VERIFY3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
arc_hdr_clear_flags(nhdr, ARC_FLAG_HAS_L1HDR);
}
/*
* The header has been reallocated so we need to re-insert it into any
* lists it was on.
*/
(void) buf_hash_insert(nhdr, NULL);
ASSERT(list_link_active(&hdr->b_l2hdr.b_l2node));
mutex_enter(&dev->l2ad_mtx);
/*
* We must place the realloc'ed header back into the list at
* the same spot. Otherwise, if it's placed earlier in the list,
* l2arc_write_buffers() could find it during the function's
* write phase, and try to write it out to the l2arc.
*/
list_insert_after(&dev->l2ad_buflist, hdr, nhdr);
list_remove(&dev->l2ad_buflist, hdr);
mutex_exit(&dev->l2ad_mtx);
/*
* Since we're using the pointer address as the tag when
* incrementing and decrementing the l2ad_alloc refcount, we
* must remove the old pointer (that we're about to destroy) and
* add the new pointer to the refcount. Otherwise we'd remove
* the wrong pointer address when calling arc_hdr_destroy() later.
*/
(void) zfs_refcount_remove_many(&dev->l2ad_alloc,
arc_hdr_size(hdr), hdr);
(void) zfs_refcount_add_many(&dev->l2ad_alloc,
arc_hdr_size(nhdr), nhdr);
buf_discard_identity(hdr);
kmem_cache_free(old, hdr);
return (nhdr);
}
/*
* This function allows an L1 header to be reallocated as a crypt
* header and vice versa. If we are going to a crypt header, the
* new fields will be zeroed out.
*/
static arc_buf_hdr_t *
arc_hdr_realloc_crypt(arc_buf_hdr_t *hdr, boolean_t need_crypt)
{
arc_buf_hdr_t *nhdr;
arc_buf_t *buf;
kmem_cache_t *ncache, *ocache;
/*
* This function requires that hdr is in the arc_anon state.
* Therefore it won't have any L2ARC data for us to worry
* about copying.
*/
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(!HDR_HAS_L2HDR(hdr));
ASSERT3U(!!HDR_PROTECTED(hdr), !=, need_crypt);
ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon);
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
ASSERT(!list_link_active(&hdr->b_l2hdr.b_l2node));
ASSERT3P(hdr->b_hash_next, ==, NULL);
if (need_crypt) {
ncache = hdr_full_crypt_cache;
ocache = hdr_full_cache;
} else {
ncache = hdr_full_cache;
ocache = hdr_full_crypt_cache;
}
nhdr = kmem_cache_alloc(ncache, KM_PUSHPAGE);
/*
* Copy all members that aren't locks or condvars to the new header.
* No lists are pointing to us (as we asserted above), so we don't
* need to worry about the list nodes.
*/
nhdr->b_dva = hdr->b_dva;
nhdr->b_birth = hdr->b_birth;
nhdr->b_type = hdr->b_type;
nhdr->b_flags = hdr->b_flags;
nhdr->b_psize = hdr->b_psize;
nhdr->b_lsize = hdr->b_lsize;
nhdr->b_spa = hdr->b_spa;
#ifdef ZFS_DEBUG
nhdr->b_l1hdr.b_freeze_cksum = hdr->b_l1hdr.b_freeze_cksum;
#endif
nhdr->b_l1hdr.b_bufcnt = hdr->b_l1hdr.b_bufcnt;
nhdr->b_l1hdr.b_byteswap = hdr->b_l1hdr.b_byteswap;
nhdr->b_l1hdr.b_state = hdr->b_l1hdr.b_state;
nhdr->b_l1hdr.b_arc_access = hdr->b_l1hdr.b_arc_access;
nhdr->b_l1hdr.b_mru_hits = hdr->b_l1hdr.b_mru_hits;
nhdr->b_l1hdr.b_mru_ghost_hits = hdr->b_l1hdr.b_mru_ghost_hits;
nhdr->b_l1hdr.b_mfu_hits = hdr->b_l1hdr.b_mfu_hits;
nhdr->b_l1hdr.b_mfu_ghost_hits = hdr->b_l1hdr.b_mfu_ghost_hits;
nhdr->b_l1hdr.b_acb = hdr->b_l1hdr.b_acb;
nhdr->b_l1hdr.b_pabd = hdr->b_l1hdr.b_pabd;
/*
* This zfs_refcount_add() exists only to ensure that the individual
* arc buffers always point to a header that is referenced, avoiding
* a small race condition that could trigger ASSERTs.
*/
(void) zfs_refcount_add(&nhdr->b_l1hdr.b_refcnt, FTAG);
nhdr->b_l1hdr.b_buf = hdr->b_l1hdr.b_buf;
for (buf = nhdr->b_l1hdr.b_buf; buf != NULL; buf = buf->b_next)
buf->b_hdr = nhdr;
zfs_refcount_transfer(&nhdr->b_l1hdr.b_refcnt, &hdr->b_l1hdr.b_refcnt);
(void) zfs_refcount_remove(&nhdr->b_l1hdr.b_refcnt, FTAG);
ASSERT0(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt));
if (need_crypt) {
arc_hdr_set_flags(nhdr, ARC_FLAG_PROTECTED);
} else {
arc_hdr_clear_flags(nhdr, ARC_FLAG_PROTECTED);
}
/* unset all members of the original hdr */
memset(&hdr->b_dva, 0, sizeof (dva_t));
hdr->b_birth = 0;
hdr->b_type = 0;
hdr->b_flags = 0;
hdr->b_psize = 0;
hdr->b_lsize = 0;
hdr->b_spa = 0;
#ifdef ZFS_DEBUG
hdr->b_l1hdr.b_freeze_cksum = NULL;
#endif
hdr->b_l1hdr.b_buf = NULL;
hdr->b_l1hdr.b_bufcnt = 0;
hdr->b_l1hdr.b_byteswap = 0;
hdr->b_l1hdr.b_state = NULL;
hdr->b_l1hdr.b_arc_access = 0;
hdr->b_l1hdr.b_mru_hits = 0;
hdr->b_l1hdr.b_mru_ghost_hits = 0;
hdr->b_l1hdr.b_mfu_hits = 0;
hdr->b_l1hdr.b_mfu_ghost_hits = 0;
hdr->b_l1hdr.b_acb = NULL;
hdr->b_l1hdr.b_pabd = NULL;
if (ocache == hdr_full_crypt_cache) {
ASSERT(!HDR_HAS_RABD(hdr));
hdr->b_crypt_hdr.b_ot = DMU_OT_NONE;
hdr->b_crypt_hdr.b_ebufcnt = 0;
hdr->b_crypt_hdr.b_dsobj = 0;
memset(hdr->b_crypt_hdr.b_salt, 0, ZIO_DATA_SALT_LEN);
memset(hdr->b_crypt_hdr.b_iv, 0, ZIO_DATA_IV_LEN);
memset(hdr->b_crypt_hdr.b_mac, 0, ZIO_DATA_MAC_LEN);
}
buf_discard_identity(hdr);
kmem_cache_free(ocache, hdr);
return (nhdr);
}
/*
* This function is used by the send / receive code to convert a newly
* allocated arc_buf_t to one that is suitable for a raw encrypted write. It
* is also used to allow the root objset block to be updated without altering
* its embedded MACs. Both block types will always be uncompressed so we do not
* have to worry about compression type or psize.
*/
void
arc_convert_to_raw(arc_buf_t *buf, uint64_t dsobj, boolean_t byteorder,
dmu_object_type_t ot, const uint8_t *salt, const uint8_t *iv,
const uint8_t *mac)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT(ot == DMU_OT_DNODE || ot == DMU_OT_OBJSET);
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon);
buf->b_flags |= (ARC_BUF_FLAG_COMPRESSED | ARC_BUF_FLAG_ENCRYPTED);
if (!HDR_PROTECTED(hdr))
hdr = arc_hdr_realloc_crypt(hdr, B_TRUE);
hdr->b_crypt_hdr.b_dsobj = dsobj;
hdr->b_crypt_hdr.b_ot = ot;
hdr->b_l1hdr.b_byteswap = (byteorder == ZFS_HOST_BYTEORDER) ?
DMU_BSWAP_NUMFUNCS : DMU_OT_BYTESWAP(ot);
if (!arc_hdr_has_uncompressed_buf(hdr))
arc_cksum_free(hdr);
if (salt != NULL)
memcpy(hdr->b_crypt_hdr.b_salt, salt, ZIO_DATA_SALT_LEN);
if (iv != NULL)
memcpy(hdr->b_crypt_hdr.b_iv, iv, ZIO_DATA_IV_LEN);
if (mac != NULL)
memcpy(hdr->b_crypt_hdr.b_mac, mac, ZIO_DATA_MAC_LEN);
}
/*
* Allocate a new arc_buf_hdr_t and arc_buf_t and return the buf to the caller.
* The buf is returned thawed since we expect the consumer to modify it.
*/
arc_buf_t *
arc_alloc_buf(spa_t *spa, const void *tag, arc_buf_contents_t type,
int32_t size)
{
arc_buf_hdr_t *hdr = arc_hdr_alloc(spa_load_guid(spa), size, size,
B_FALSE, ZIO_COMPRESS_OFF, 0, type);
arc_buf_t *buf = NULL;
VERIFY0(arc_buf_alloc_impl(hdr, spa, NULL, tag, B_FALSE, B_FALSE,
B_FALSE, B_FALSE, &buf));
arc_buf_thaw(buf);
return (buf);
}
/*
* Allocate a compressed buf in the same manner as arc_alloc_buf. Don't use this
* for bufs containing metadata.
*/
arc_buf_t *
arc_alloc_compressed_buf(spa_t *spa, const void *tag, uint64_t psize,
uint64_t lsize, enum zio_compress compression_type, uint8_t complevel)
{
ASSERT3U(lsize, >, 0);
ASSERT3U(lsize, >=, psize);
ASSERT3U(compression_type, >, ZIO_COMPRESS_OFF);
ASSERT3U(compression_type, <, ZIO_COMPRESS_FUNCTIONS);
arc_buf_hdr_t *hdr = arc_hdr_alloc(spa_load_guid(spa), psize, lsize,
B_FALSE, compression_type, complevel, ARC_BUFC_DATA);
arc_buf_t *buf = NULL;
VERIFY0(arc_buf_alloc_impl(hdr, spa, NULL, tag, B_FALSE,
B_TRUE, B_FALSE, B_FALSE, &buf));
arc_buf_thaw(buf);
/*
* To ensure that the hdr has the correct data in it if we call
* arc_untransform() on this buf before it's been written to disk,
* it's easiest if we just set up sharing between the buf and the hdr.
*/
arc_share_buf(hdr, buf);
return (buf);
}
arc_buf_t *
arc_alloc_raw_buf(spa_t *spa, const void *tag, uint64_t dsobj,
boolean_t byteorder, const uint8_t *salt, const uint8_t *iv,
const uint8_t *mac, dmu_object_type_t ot, uint64_t psize, uint64_t lsize,
enum zio_compress compression_type, uint8_t complevel)
{
arc_buf_hdr_t *hdr;
arc_buf_t *buf;
arc_buf_contents_t type = DMU_OT_IS_METADATA(ot) ?
ARC_BUFC_METADATA : ARC_BUFC_DATA;
ASSERT3U(lsize, >, 0);
ASSERT3U(lsize, >=, psize);
ASSERT3U(compression_type, >=, ZIO_COMPRESS_OFF);
ASSERT3U(compression_type, <, ZIO_COMPRESS_FUNCTIONS);
hdr = arc_hdr_alloc(spa_load_guid(spa), psize, lsize, B_TRUE,
compression_type, complevel, type);
hdr->b_crypt_hdr.b_dsobj = dsobj;
hdr->b_crypt_hdr.b_ot = ot;
hdr->b_l1hdr.b_byteswap = (byteorder == ZFS_HOST_BYTEORDER) ?
DMU_BSWAP_NUMFUNCS : DMU_OT_BYTESWAP(ot);
memcpy(hdr->b_crypt_hdr.b_salt, salt, ZIO_DATA_SALT_LEN);
memcpy(hdr->b_crypt_hdr.b_iv, iv, ZIO_DATA_IV_LEN);
memcpy(hdr->b_crypt_hdr.b_mac, mac, ZIO_DATA_MAC_LEN);
/*
* This buffer will be considered encrypted even if the ot is not an
* encrypted type. It will become authenticated instead in
* arc_write_ready().
*/
buf = NULL;
VERIFY0(arc_buf_alloc_impl(hdr, spa, NULL, tag, B_TRUE, B_TRUE,
B_FALSE, B_FALSE, &buf));
arc_buf_thaw(buf);
return (buf);
}
static void
l2arc_hdr_arcstats_update(arc_buf_hdr_t *hdr, boolean_t incr,
boolean_t state_only)
{
l2arc_buf_hdr_t *l2hdr = &hdr->b_l2hdr;
l2arc_dev_t *dev = l2hdr->b_dev;
uint64_t lsize = HDR_GET_LSIZE(hdr);
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t asize = vdev_psize_to_asize(dev->l2ad_vdev, psize);
arc_buf_contents_t type = hdr->b_type;
int64_t lsize_s;
int64_t psize_s;
int64_t asize_s;
if (incr) {
lsize_s = lsize;
psize_s = psize;
asize_s = asize;
} else {
lsize_s = -lsize;
psize_s = -psize;
asize_s = -asize;
}
/* If the buffer is a prefetch, count it as such. */
if (HDR_PREFETCH(hdr)) {
ARCSTAT_INCR(arcstat_l2_prefetch_asize, asize_s);
} else {
/*
* We use the value stored in the L2 header upon initial
* caching in L2ARC. This value will be updated in case
* an MRU/MRU_ghost buffer transitions to MFU but the L2ARC
* metadata (log entry) cannot currently be updated. Having
* the ARC state in the L2 header solves the problem of a
* possibly absent L1 header (apparent in buffers restored
* from persistent L2ARC).
*/
switch (hdr->b_l2hdr.b_arcs_state) {
case ARC_STATE_MRU_GHOST:
case ARC_STATE_MRU:
ARCSTAT_INCR(arcstat_l2_mru_asize, asize_s);
break;
case ARC_STATE_MFU_GHOST:
case ARC_STATE_MFU:
ARCSTAT_INCR(arcstat_l2_mfu_asize, asize_s);
break;
default:
break;
}
}
if (state_only)
return;
ARCSTAT_INCR(arcstat_l2_psize, psize_s);
ARCSTAT_INCR(arcstat_l2_lsize, lsize_s);
switch (type) {
case ARC_BUFC_DATA:
ARCSTAT_INCR(arcstat_l2_bufc_data_asize, asize_s);
break;
case ARC_BUFC_METADATA:
ARCSTAT_INCR(arcstat_l2_bufc_metadata_asize, asize_s);
break;
default:
break;
}
}
static void
arc_hdr_l2hdr_destroy(arc_buf_hdr_t *hdr)
{
l2arc_buf_hdr_t *l2hdr = &hdr->b_l2hdr;
l2arc_dev_t *dev = l2hdr->b_dev;
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t asize = vdev_psize_to_asize(dev->l2ad_vdev, psize);
ASSERT(MUTEX_HELD(&dev->l2ad_mtx));
ASSERT(HDR_HAS_L2HDR(hdr));
list_remove(&dev->l2ad_buflist, hdr);
l2arc_hdr_arcstats_decrement(hdr);
vdev_space_update(dev->l2ad_vdev, -asize, 0, 0);
(void) zfs_refcount_remove_many(&dev->l2ad_alloc, arc_hdr_size(hdr),
hdr);
arc_hdr_clear_flags(hdr, ARC_FLAG_HAS_L2HDR);
}
static void
arc_hdr_destroy(arc_buf_hdr_t *hdr)
{
if (HDR_HAS_L1HDR(hdr)) {
ASSERT(hdr->b_l1hdr.b_buf == NULL ||
hdr->b_l1hdr.b_bufcnt > 0);
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon);
}
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT(!HDR_IN_HASH_TABLE(hdr));
if (HDR_HAS_L2HDR(hdr)) {
l2arc_dev_t *dev = hdr->b_l2hdr.b_dev;
boolean_t buflist_held = MUTEX_HELD(&dev->l2ad_mtx);
if (!buflist_held)
mutex_enter(&dev->l2ad_mtx);
/*
* Even though we checked this conditional above, we
* need to check this again now that we have the
* l2ad_mtx. This is because we could be racing with
* another thread calling l2arc_evict() which might have
* destroyed this header's L2 portion as we were waiting
* to acquire the l2ad_mtx. If that happens, we don't
* want to re-destroy the header's L2 portion.
*/
if (HDR_HAS_L2HDR(hdr)) {
if (!HDR_EMPTY(hdr))
buf_discard_identity(hdr);
arc_hdr_l2hdr_destroy(hdr);
}
if (!buflist_held)
mutex_exit(&dev->l2ad_mtx);
}
/*
* The header's identify can only be safely discarded once it is no
* longer discoverable. This requires removing it from the hash table
* and the l2arc header list. After this point the hash lock can not
* be used to protect the header.
*/
if (!HDR_EMPTY(hdr))
buf_discard_identity(hdr);
if (HDR_HAS_L1HDR(hdr)) {
arc_cksum_free(hdr);
while (hdr->b_l1hdr.b_buf != NULL)
arc_buf_destroy_impl(hdr->b_l1hdr.b_buf);
if (hdr->b_l1hdr.b_pabd != NULL)
arc_hdr_free_abd(hdr, B_FALSE);
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
}
ASSERT3P(hdr->b_hash_next, ==, NULL);
if (HDR_HAS_L1HDR(hdr)) {
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
#ifdef ZFS_DEBUG
ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
#endif
if (!HDR_PROTECTED(hdr)) {
kmem_cache_free(hdr_full_cache, hdr);
} else {
kmem_cache_free(hdr_full_crypt_cache, hdr);
}
} else {
kmem_cache_free(hdr_l2only_cache, hdr);
}
}
void
arc_buf_destroy(arc_buf_t *buf, const void *tag)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
if (hdr->b_l1hdr.b_state == arc_anon) {
ASSERT3U(hdr->b_l1hdr.b_bufcnt, ==, 1);
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
VERIFY0(remove_reference(hdr, tag));
return;
}
kmutex_t *hash_lock = HDR_LOCK(hdr);
mutex_enter(hash_lock);
ASSERT3P(hdr, ==, buf->b_hdr);
ASSERT(hdr->b_l1hdr.b_bufcnt > 0);
ASSERT3P(hash_lock, ==, HDR_LOCK(hdr));
ASSERT3P(hdr->b_l1hdr.b_state, !=, arc_anon);
ASSERT3P(buf->b_data, !=, NULL);
arc_buf_destroy_impl(buf);
(void) remove_reference(hdr, tag);
mutex_exit(hash_lock);
}
/*
* Evict the arc_buf_hdr that is provided as a parameter. The resultant
* state of the header is dependent on its state prior to entering this
* function. The following transitions are possible:
*
* - arc_mru -> arc_mru_ghost
* - arc_mfu -> arc_mfu_ghost
* - arc_mru_ghost -> arc_l2c_only
* - arc_mru_ghost -> deleted
* - arc_mfu_ghost -> arc_l2c_only
* - arc_mfu_ghost -> deleted
* - arc_uncached -> deleted
*
* Return total size of evicted data buffers for eviction progress tracking.
* When evicting from ghost states return logical buffer size to make eviction
* progress at the same (or at least comparable) rate as from non-ghost states.
*
* Return *real_evicted for actual ARC size reduction to wake up threads
* waiting for it. For non-ghost states it includes size of evicted data
* buffers (the headers are not freed there). For ghost states it includes
* only the evicted headers size.
*/
static int64_t
arc_evict_hdr(arc_buf_hdr_t *hdr, uint64_t *real_evicted)
{
arc_state_t *evicted_state, *state;
int64_t bytes_evicted = 0;
uint_t min_lifetime = HDR_PRESCIENT_PREFETCH(hdr) ?
arc_min_prescient_prefetch_ms : arc_min_prefetch_ms;
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT0(hdr->b_l1hdr.b_bufcnt);
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
ASSERT0(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt));
*real_evicted = 0;
state = hdr->b_l1hdr.b_state;
if (GHOST_STATE(state)) {
/*
* l2arc_write_buffers() relies on a header's L1 portion
* (i.e. its b_pabd field) during it's write phase.
* Thus, we cannot push a header onto the arc_l2c_only
* state (removing its L1 piece) until the header is
* done being written to the l2arc.
*/
if (HDR_HAS_L2HDR(hdr) && HDR_L2_WRITING(hdr)) {
ARCSTAT_BUMP(arcstat_evict_l2_skip);
return (bytes_evicted);
}
ARCSTAT_BUMP(arcstat_deleted);
bytes_evicted += HDR_GET_LSIZE(hdr);
DTRACE_PROBE1(arc__delete, arc_buf_hdr_t *, hdr);
if (HDR_HAS_L2HDR(hdr)) {
ASSERT(hdr->b_l1hdr.b_pabd == NULL);
ASSERT(!HDR_HAS_RABD(hdr));
/*
* This buffer is cached on the 2nd Level ARC;
* don't destroy the header.
*/
arc_change_state(arc_l2c_only, hdr);
/*
* dropping from L1+L2 cached to L2-only,
* realloc to remove the L1 header.
*/
(void) arc_hdr_realloc(hdr, hdr_full_cache,
hdr_l2only_cache);
*real_evicted += HDR_FULL_SIZE - HDR_L2ONLY_SIZE;
} else {
arc_change_state(arc_anon, hdr);
arc_hdr_destroy(hdr);
*real_evicted += HDR_FULL_SIZE;
}
return (bytes_evicted);
}
ASSERT(state == arc_mru || state == arc_mfu || state == arc_uncached);
evicted_state = (state == arc_uncached) ? arc_anon :
((state == arc_mru) ? arc_mru_ghost : arc_mfu_ghost);
/* prefetch buffers have a minimum lifespan */
if ((hdr->b_flags & (ARC_FLAG_PREFETCH | ARC_FLAG_INDIRECT)) &&
ddi_get_lbolt() - hdr->b_l1hdr.b_arc_access <
MSEC_TO_TICK(min_lifetime)) {
ARCSTAT_BUMP(arcstat_evict_skip);
return (bytes_evicted);
}
if (HDR_HAS_L2HDR(hdr)) {
ARCSTAT_INCR(arcstat_evict_l2_cached, HDR_GET_LSIZE(hdr));
} else {
if (l2arc_write_eligible(hdr->b_spa, hdr)) {
ARCSTAT_INCR(arcstat_evict_l2_eligible,
HDR_GET_LSIZE(hdr));
switch (state->arcs_state) {
case ARC_STATE_MRU:
ARCSTAT_INCR(
arcstat_evict_l2_eligible_mru,
HDR_GET_LSIZE(hdr));
break;
case ARC_STATE_MFU:
ARCSTAT_INCR(
arcstat_evict_l2_eligible_mfu,
HDR_GET_LSIZE(hdr));
break;
default:
break;
}
} else {
ARCSTAT_INCR(arcstat_evict_l2_ineligible,
HDR_GET_LSIZE(hdr));
}
}
bytes_evicted += arc_hdr_size(hdr);
*real_evicted += arc_hdr_size(hdr);
/*
* If this hdr is being evicted and has a compressed buffer then we
* discard it here before we change states. This ensures that the
* accounting is updated correctly in arc_free_data_impl().
*/
if (hdr->b_l1hdr.b_pabd != NULL)
arc_hdr_free_abd(hdr, B_FALSE);
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
arc_change_state(evicted_state, hdr);
DTRACE_PROBE1(arc__evict, arc_buf_hdr_t *, hdr);
if (evicted_state == arc_anon) {
arc_hdr_destroy(hdr);
*real_evicted += HDR_FULL_SIZE;
} else {
ASSERT(HDR_IN_HASH_TABLE(hdr));
}
return (bytes_evicted);
}
static void
arc_set_need_free(void)
{
ASSERT(MUTEX_HELD(&arc_evict_lock));
int64_t remaining = arc_free_memory() - arc_sys_free / 2;
arc_evict_waiter_t *aw = list_tail(&arc_evict_waiters);
if (aw == NULL) {
arc_need_free = MAX(-remaining, 0);
} else {
arc_need_free =
MAX(-remaining, (int64_t)(aw->aew_count - arc_evict_count));
}
}
static uint64_t
arc_evict_state_impl(multilist_t *ml, int idx, arc_buf_hdr_t *marker,
uint64_t spa, uint64_t bytes)
{
multilist_sublist_t *mls;
uint64_t bytes_evicted = 0, real_evicted = 0;
arc_buf_hdr_t *hdr;
kmutex_t *hash_lock;
uint_t evict_count = zfs_arc_evict_batch_limit;
ASSERT3P(marker, !=, NULL);
mls = multilist_sublist_lock(ml, idx);
for (hdr = multilist_sublist_prev(mls, marker); likely(hdr != NULL);
hdr = multilist_sublist_prev(mls, marker)) {
if ((evict_count == 0) || (bytes_evicted >= bytes))
break;
/*
* To keep our iteration location, move the marker
* forward. Since we're not holding hdr's hash lock, we
* must be very careful and not remove 'hdr' from the
* sublist. Otherwise, other consumers might mistake the
* 'hdr' as not being on a sublist when they call the
* multilist_link_active() function (they all rely on
* the hash lock protecting concurrent insertions and
* removals). multilist_sublist_move_forward() was
* specifically implemented to ensure this is the case
* (only 'marker' will be removed and re-inserted).
*/
multilist_sublist_move_forward(mls, marker);
/*
* The only case where the b_spa field should ever be
* zero, is the marker headers inserted by
* arc_evict_state(). It's possible for multiple threads
* to be calling arc_evict_state() concurrently (e.g.
* dsl_pool_close() and zio_inject_fault()), so we must
* skip any markers we see from these other threads.
*/
if (hdr->b_spa == 0)
continue;
/* we're only interested in evicting buffers of a certain spa */
if (spa != 0 && hdr->b_spa != spa) {
ARCSTAT_BUMP(arcstat_evict_skip);
continue;
}
hash_lock = HDR_LOCK(hdr);
/*
* We aren't calling this function from any code path
* that would already be holding a hash lock, so we're
* asserting on this assumption to be defensive in case
* this ever changes. Without this check, it would be
* possible to incorrectly increment arcstat_mutex_miss
* below (e.g. if the code changed such that we called
* this function with a hash lock held).
*/
ASSERT(!MUTEX_HELD(hash_lock));
if (mutex_tryenter(hash_lock)) {
uint64_t revicted;
uint64_t evicted = arc_evict_hdr(hdr, &revicted);
mutex_exit(hash_lock);
bytes_evicted += evicted;
real_evicted += revicted;
/*
* If evicted is zero, arc_evict_hdr() must have
* decided to skip this header, don't increment
* evict_count in this case.
*/
if (evicted != 0)
evict_count--;
} else {
ARCSTAT_BUMP(arcstat_mutex_miss);
}
}
multilist_sublist_unlock(mls);
/*
* Increment the count of evicted bytes, and wake up any threads that
* are waiting for the count to reach this value. Since the list is
* ordered by ascending aew_count, we pop off the beginning of the
* list until we reach the end, or a waiter that's past the current
* "count". Doing this outside the loop reduces the number of times
* we need to acquire the global arc_evict_lock.
*
* Only wake when there's sufficient free memory in the system
* (specifically, arc_sys_free/2, which by default is a bit more than
* 1/64th of RAM). See the comments in arc_wait_for_eviction().
*/
mutex_enter(&arc_evict_lock);
arc_evict_count += real_evicted;
if (arc_free_memory() > arc_sys_free / 2) {
arc_evict_waiter_t *aw;
while ((aw = list_head(&arc_evict_waiters)) != NULL &&
aw->aew_count <= arc_evict_count) {
list_remove(&arc_evict_waiters, aw);
cv_broadcast(&aw->aew_cv);
}
}
arc_set_need_free();
mutex_exit(&arc_evict_lock);
/*
* If the ARC size is reduced from arc_c_max to arc_c_min (especially
* if the average cached block is small), eviction can be on-CPU for
* many seconds. To ensure that other threads that may be bound to
* this CPU are able to make progress, make a voluntary preemption
* call here.
*/
kpreempt(KPREEMPT_SYNC);
return (bytes_evicted);
}
/*
* Allocate an array of buffer headers used as placeholders during arc state
* eviction.
*/
static arc_buf_hdr_t **
arc_state_alloc_markers(int count)
{
arc_buf_hdr_t **markers;
markers = kmem_zalloc(sizeof (*markers) * count, KM_SLEEP);
for (int i = 0; i < count; i++) {
markers[i] = kmem_cache_alloc(hdr_full_cache, KM_SLEEP);
/*
* A b_spa of 0 is used to indicate that this header is
* a marker. This fact is used in arc_evict_state_impl().
*/
markers[i]->b_spa = 0;
}
return (markers);
}
static void
arc_state_free_markers(arc_buf_hdr_t **markers, int count)
{
for (int i = 0; i < count; i++)
kmem_cache_free(hdr_full_cache, markers[i]);
kmem_free(markers, sizeof (*markers) * count);
}
/*
* Evict buffers from the given arc state, until we've removed the
* specified number of bytes. Move the removed buffers to the
* appropriate evict state.
*
* This function makes a "best effort". It skips over any buffers
* it can't get a hash_lock on, and so, may not catch all candidates.
* It may also return without evicting as much space as requested.
*
* If bytes is specified using the special value ARC_EVICT_ALL, this
* will evict all available (i.e. unlocked and evictable) buffers from
* the given arc state; which is used by arc_flush().
*/
static uint64_t
arc_evict_state(arc_state_t *state, arc_buf_contents_t type, uint64_t spa,
uint64_t bytes)
{
uint64_t total_evicted = 0;
multilist_t *ml = &state->arcs_list[type];
int num_sublists;
arc_buf_hdr_t **markers;
num_sublists = multilist_get_num_sublists(ml);
/*
* If we've tried to evict from each sublist, made some
* progress, but still have not hit the target number of bytes
* to evict, we want to keep trying. The markers allow us to
* pick up where we left off for each individual sublist, rather
* than starting from the tail each time.
*/
if (zthr_iscurthread(arc_evict_zthr)) {
markers = arc_state_evict_markers;
ASSERT3S(num_sublists, <=, arc_state_evict_marker_count);
} else {
markers = arc_state_alloc_markers(num_sublists);
}
for (int i = 0; i < num_sublists; i++) {
multilist_sublist_t *mls;
mls = multilist_sublist_lock(ml, i);
multilist_sublist_insert_tail(mls, markers[i]);
multilist_sublist_unlock(mls);
}
/*
* While we haven't hit our target number of bytes to evict, or
* we're evicting all available buffers.
*/
while (total_evicted < bytes) {
int sublist_idx = multilist_get_random_index(ml);
uint64_t scan_evicted = 0;
/*
* Start eviction using a randomly selected sublist,
* this is to try and evenly balance eviction across all
* sublists. Always starting at the same sublist
* (e.g. index 0) would cause evictions to favor certain
* sublists over others.
*/
for (int i = 0; i < num_sublists; i++) {
uint64_t bytes_remaining;
uint64_t bytes_evicted;
if (total_evicted < bytes)
bytes_remaining = bytes - total_evicted;
else
break;
bytes_evicted = arc_evict_state_impl(ml, sublist_idx,
markers[sublist_idx], spa, bytes_remaining);
scan_evicted += bytes_evicted;
total_evicted += bytes_evicted;
/* we've reached the end, wrap to the beginning */
if (++sublist_idx >= num_sublists)
sublist_idx = 0;
}
/*
* If we didn't evict anything during this scan, we have
* no reason to believe we'll evict more during another
* scan, so break the loop.
*/
if (scan_evicted == 0) {
/* This isn't possible, let's make that obvious */
ASSERT3S(bytes, !=, 0);
/*
* When bytes is ARC_EVICT_ALL, the only way to
* break the loop is when scan_evicted is zero.
* In that case, we actually have evicted enough,
* so we don't want to increment the kstat.
*/
if (bytes != ARC_EVICT_ALL) {
ASSERT3S(total_evicted, <, bytes);
ARCSTAT_BUMP(arcstat_evict_not_enough);
}
break;
}
}
for (int i = 0; i < num_sublists; i++) {
multilist_sublist_t *mls = multilist_sublist_lock(ml, i);
multilist_sublist_remove(mls, markers[i]);
multilist_sublist_unlock(mls);
}
if (markers != arc_state_evict_markers)
arc_state_free_markers(markers, num_sublists);
return (total_evicted);
}
/*
* Flush all "evictable" data of the given type from the arc state
* specified. This will not evict any "active" buffers (i.e. referenced).
*
* When 'retry' is set to B_FALSE, the function will make a single pass
* over the state and evict any buffers that it can. Since it doesn't
* continually retry the eviction, it might end up leaving some buffers
* in the ARC due to lock misses.
*
* When 'retry' is set to B_TRUE, the function will continually retry the
* eviction until *all* evictable buffers have been removed from the
* state. As a result, if concurrent insertions into the state are
* allowed (e.g. if the ARC isn't shutting down), this function might
* wind up in an infinite loop, continually trying to evict buffers.
*/
static uint64_t
arc_flush_state(arc_state_t *state, uint64_t spa, arc_buf_contents_t type,
boolean_t retry)
{
uint64_t evicted = 0;
while (zfs_refcount_count(&state->arcs_esize[type]) != 0) {
evicted += arc_evict_state(state, type, spa, ARC_EVICT_ALL);
if (!retry)
break;
}
return (evicted);
}
/*
* Evict the specified number of bytes from the state specified. This
* function prevents us from trying to evict more from a state's list
* than is "evictable", and to skip evicting altogether when passed a
* negative value for "bytes". In contrast, arc_evict_state() will
* evict everything it can, when passed a negative value for "bytes".
*/
static uint64_t
arc_evict_impl(arc_state_t *state, arc_buf_contents_t type, int64_t bytes)
{
uint64_t delta;
if (bytes > 0 && zfs_refcount_count(&state->arcs_esize[type]) > 0) {
delta = MIN(zfs_refcount_count(&state->arcs_esize[type]),
bytes);
return (arc_evict_state(state, type, 0, delta));
}
return (0);
}
/*
* Adjust specified fraction, taking into account initial ghost state(s) size,
* ghost hit bytes towards increasing the fraction, ghost hit bytes towards
* decreasing it, plus a balance factor, controlling the decrease rate, used
* to balance metadata vs data.
*/
static uint64_t
arc_evict_adj(uint64_t frac, uint64_t total, uint64_t up, uint64_t down,
uint_t balance)
{
if (total < 8 || up + down == 0)
return (frac);
/*
* We should not have more ghost hits than ghost size, but they
* may get close. Restrict maximum adjustment in that case.
*/
if (up + down >= total / 4) {
uint64_t scale = (up + down) / (total / 8);
up /= scale;
down /= scale;
}
/* Get maximal dynamic range by choosing optimal shifts. */
int s = highbit64(total);
s = MIN(64 - s, 32);
uint64_t ofrac = (1ULL << 32) - frac;
if (frac >= 4 * ofrac)
up /= frac / (2 * ofrac + 1);
up = (up << s) / (total >> (32 - s));
if (ofrac >= 4 * frac)
down /= ofrac / (2 * frac + 1);
down = (down << s) / (total >> (32 - s));
down = down * 100 / balance;
return (frac + up - down);
}
/*
* Evict buffers from the cache, such that arcstat_size is capped by arc_c.
*/
static uint64_t
arc_evict(void)
{
uint64_t asize, bytes, total_evicted = 0;
int64_t e, mrud, mrum, mfud, mfum, w;
static uint64_t ogrd, ogrm, ogfd, ogfm;
static uint64_t gsrd, gsrm, gsfd, gsfm;
uint64_t ngrd, ngrm, ngfd, ngfm;
/* Get current size of ARC states we can evict from. */
mrud = zfs_refcount_count(&arc_mru->arcs_size[ARC_BUFC_DATA]) +
zfs_refcount_count(&arc_anon->arcs_size[ARC_BUFC_DATA]);
mrum = zfs_refcount_count(&arc_mru->arcs_size[ARC_BUFC_METADATA]) +
zfs_refcount_count(&arc_anon->arcs_size[ARC_BUFC_METADATA]);
mfud = zfs_refcount_count(&arc_mfu->arcs_size[ARC_BUFC_DATA]);
mfum = zfs_refcount_count(&arc_mfu->arcs_size[ARC_BUFC_METADATA]);
uint64_t d = mrud + mfud;
uint64_t m = mrum + mfum;
uint64_t t = d + m;
/* Get ARC ghost hits since last eviction. */
ngrd = wmsum_value(&arc_mru_ghost->arcs_hits[ARC_BUFC_DATA]);
uint64_t grd = ngrd - ogrd;
ogrd = ngrd;
ngrm = wmsum_value(&arc_mru_ghost->arcs_hits[ARC_BUFC_METADATA]);
uint64_t grm = ngrm - ogrm;
ogrm = ngrm;
ngfd = wmsum_value(&arc_mfu_ghost->arcs_hits[ARC_BUFC_DATA]);
uint64_t gfd = ngfd - ogfd;
ogfd = ngfd;
ngfm = wmsum_value(&arc_mfu_ghost->arcs_hits[ARC_BUFC_METADATA]);
uint64_t gfm = ngfm - ogfm;
ogfm = ngfm;
/* Adjust ARC states balance based on ghost hits. */
arc_meta = arc_evict_adj(arc_meta, gsrd + gsrm + gsfd + gsfm,
grm + gfm, grd + gfd, zfs_arc_meta_balance);
arc_pd = arc_evict_adj(arc_pd, gsrd + gsfd, grd, gfd, 100);
arc_pm = arc_evict_adj(arc_pm, gsrm + gsfm, grm, gfm, 100);
asize = aggsum_value(&arc_sums.arcstat_size);
int64_t wt = t - (asize - arc_c);
/*
* Try to reduce pinned dnodes if more than 3/4 of wanted metadata
* target is not evictable or if they go over arc_dnode_limit.
*/
int64_t prune = 0;
int64_t dn = wmsum_value(&arc_sums.arcstat_dnode_size);
- w = wt * (arc_meta >> 16) >> 16;
+ w = wt * (int64_t)(arc_meta >> 16) >> 16;
if (zfs_refcount_count(&arc_mru->arcs_size[ARC_BUFC_METADATA]) +
zfs_refcount_count(&arc_mfu->arcs_size[ARC_BUFC_METADATA]) -
zfs_refcount_count(&arc_mru->arcs_esize[ARC_BUFC_METADATA]) -
zfs_refcount_count(&arc_mfu->arcs_esize[ARC_BUFC_METADATA]) >
w * 3 / 4) {
prune = dn / sizeof (dnode_t) *
zfs_arc_dnode_reduce_percent / 100;
} else if (dn > arc_dnode_limit) {
prune = (dn - arc_dnode_limit) / sizeof (dnode_t) *
zfs_arc_dnode_reduce_percent / 100;
}
if (prune > 0)
arc_prune_async(prune);
/* Evict MRU metadata. */
- w = wt * (arc_meta * arc_pm >> 48) >> 16;
+ w = wt * (int64_t)(arc_meta * arc_pm >> 48) >> 16;
e = MIN((int64_t)(asize - arc_c), (int64_t)(mrum - w));
bytes = arc_evict_impl(arc_mru, ARC_BUFC_METADATA, e);
total_evicted += bytes;
mrum -= bytes;
asize -= bytes;
/* Evict MFU metadata. */
- w = wt * (arc_meta >> 16) >> 16;
+ w = wt * (int64_t)(arc_meta >> 16) >> 16;
e = MIN((int64_t)(asize - arc_c), (int64_t)(m - w));
bytes = arc_evict_impl(arc_mfu, ARC_BUFC_METADATA, e);
total_evicted += bytes;
mfum -= bytes;
asize -= bytes;
/* Evict MRU data. */
wt -= m - total_evicted;
- w = wt * (arc_pd >> 16) >> 16;
+ w = wt * (int64_t)(arc_pd >> 16) >> 16;
e = MIN((int64_t)(asize - arc_c), (int64_t)(mrud - w));
bytes = arc_evict_impl(arc_mru, ARC_BUFC_DATA, e);
total_evicted += bytes;
mrud -= bytes;
asize -= bytes;
/* Evict MFU data. */
e = asize - arc_c;
bytes = arc_evict_impl(arc_mfu, ARC_BUFC_DATA, e);
mfud -= bytes;
total_evicted += bytes;
/*
* Evict ghost lists
*
* Size of each state's ghost list represents how much that state
* may grow by shrinking the other states. Would it need to shrink
* other states to zero (that is unlikely), its ghost size would be
* equal to sum of other three state sizes. But excessive ghost
* size may result in false ghost hits (too far back), that may
* never result in real cache hits if several states are competing.
* So choose some arbitraty point of 1/2 of other state sizes.
*/
gsrd = (mrum + mfud + mfum) / 2;
e = zfs_refcount_count(&arc_mru_ghost->arcs_size[ARC_BUFC_DATA]) -
gsrd;
(void) arc_evict_impl(arc_mru_ghost, ARC_BUFC_DATA, e);
gsrm = (mrud + mfud + mfum) / 2;
e = zfs_refcount_count(&arc_mru_ghost->arcs_size[ARC_BUFC_METADATA]) -
gsrm;
(void) arc_evict_impl(arc_mru_ghost, ARC_BUFC_METADATA, e);
gsfd = (mrud + mrum + mfum) / 2;
e = zfs_refcount_count(&arc_mfu_ghost->arcs_size[ARC_BUFC_DATA]) -
gsfd;
(void) arc_evict_impl(arc_mfu_ghost, ARC_BUFC_DATA, e);
gsfm = (mrud + mrum + mfud) / 2;
e = zfs_refcount_count(&arc_mfu_ghost->arcs_size[ARC_BUFC_METADATA]) -
gsfm;
(void) arc_evict_impl(arc_mfu_ghost, ARC_BUFC_METADATA, e);
return (total_evicted);
}
void
arc_flush(spa_t *spa, boolean_t retry)
{
uint64_t guid = 0;
/*
* If retry is B_TRUE, a spa must not be specified since we have
* no good way to determine if all of a spa's buffers have been
* evicted from an arc state.
*/
ASSERT(!retry || spa == NULL);
if (spa != NULL)
guid = spa_load_guid(spa);
(void) arc_flush_state(arc_mru, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_mru, guid, ARC_BUFC_METADATA, retry);
(void) arc_flush_state(arc_mfu, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_mfu, guid, ARC_BUFC_METADATA, retry);
(void) arc_flush_state(arc_mru_ghost, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_mru_ghost, guid, ARC_BUFC_METADATA, retry);
(void) arc_flush_state(arc_mfu_ghost, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_mfu_ghost, guid, ARC_BUFC_METADATA, retry);
(void) arc_flush_state(arc_uncached, guid, ARC_BUFC_DATA, retry);
(void) arc_flush_state(arc_uncached, guid, ARC_BUFC_METADATA, retry);
}
void
arc_reduce_target_size(int64_t to_free)
{
uint64_t c = arc_c;
if (c <= arc_c_min)
return;
/*
* All callers want the ARC to actually evict (at least) this much
* memory. Therefore we reduce from the lower of the current size and
* the target size. This way, even if arc_c is much higher than
* arc_size (as can be the case after many calls to arc_freed(), we will
* immediately have arc_c < arc_size and therefore the arc_evict_zthr
* will evict.
*/
uint64_t asize = aggsum_value(&arc_sums.arcstat_size);
if (asize < c)
to_free += c - asize;
arc_c = MAX((int64_t)c - to_free, (int64_t)arc_c_min);
/* See comment in arc_evict_cb_check() on why lock+flag */
mutex_enter(&arc_evict_lock);
arc_evict_needed = B_TRUE;
mutex_exit(&arc_evict_lock);
zthr_wakeup(arc_evict_zthr);
}
/*
* Determine if the system is under memory pressure and is asking
* to reclaim memory. A return value of B_TRUE indicates that the system
* is under memory pressure and that the arc should adjust accordingly.
*/
boolean_t
arc_reclaim_needed(void)
{
return (arc_available_memory() < 0);
}
void
arc_kmem_reap_soon(void)
{
size_t i;
kmem_cache_t *prev_cache = NULL;
kmem_cache_t *prev_data_cache = NULL;
#ifdef _KERNEL
#if defined(_ILP32)
/*
* Reclaim unused memory from all kmem caches.
*/
kmem_reap();
#endif
#endif
for (i = 0; i < SPA_MAXBLOCKSIZE >> SPA_MINBLOCKSHIFT; i++) {
#if defined(_ILP32)
/* reach upper limit of cache size on 32-bit */
if (zio_buf_cache[i] == NULL)
break;
#endif
if (zio_buf_cache[i] != prev_cache) {
prev_cache = zio_buf_cache[i];
kmem_cache_reap_now(zio_buf_cache[i]);
}
if (zio_data_buf_cache[i] != prev_data_cache) {
prev_data_cache = zio_data_buf_cache[i];
kmem_cache_reap_now(zio_data_buf_cache[i]);
}
}
kmem_cache_reap_now(buf_cache);
kmem_cache_reap_now(hdr_full_cache);
kmem_cache_reap_now(hdr_l2only_cache);
kmem_cache_reap_now(zfs_btree_leaf_cache);
abd_cache_reap_now();
}
static boolean_t
arc_evict_cb_check(void *arg, zthr_t *zthr)
{
(void) arg, (void) zthr;
#ifdef ZFS_DEBUG
/*
* This is necessary in order to keep the kstat information
* up to date for tools that display kstat data such as the
* mdb ::arc dcmd and the Linux crash utility. These tools
* typically do not call kstat's update function, but simply
* dump out stats from the most recent update. Without
* this call, these commands may show stale stats for the
* anon, mru, mru_ghost, mfu, and mfu_ghost lists. Even
* with this call, the data might be out of date if the
* evict thread hasn't been woken recently; but that should
* suffice. The arc_state_t structures can be queried
* directly if more accurate information is needed.
*/
if (arc_ksp != NULL)
arc_ksp->ks_update(arc_ksp, KSTAT_READ);
#endif
/*
* We have to rely on arc_wait_for_eviction() to tell us when to
* evict, rather than checking if we are overflowing here, so that we
* are sure to not leave arc_wait_for_eviction() waiting on aew_cv.
* If we have become "not overflowing" since arc_wait_for_eviction()
* checked, we need to wake it up. We could broadcast the CV here,
* but arc_wait_for_eviction() may have not yet gone to sleep. We
* would need to use a mutex to ensure that this function doesn't
* broadcast until arc_wait_for_eviction() has gone to sleep (e.g.
* the arc_evict_lock). However, the lock ordering of such a lock
* would necessarily be incorrect with respect to the zthr_lock,
* which is held before this function is called, and is held by
* arc_wait_for_eviction() when it calls zthr_wakeup().
*/
if (arc_evict_needed)
return (B_TRUE);
/*
* If we have buffers in uncached state, evict them periodically.
*/
return ((zfs_refcount_count(&arc_uncached->arcs_esize[ARC_BUFC_DATA]) +
zfs_refcount_count(&arc_uncached->arcs_esize[ARC_BUFC_METADATA]) &&
ddi_get_lbolt() - arc_last_uncached_flush >
MSEC_TO_TICK(arc_min_prefetch_ms / 2)));
}
/*
* Keep arc_size under arc_c by running arc_evict which evicts data
* from the ARC.
*/
static void
arc_evict_cb(void *arg, zthr_t *zthr)
{
(void) arg, (void) zthr;
uint64_t evicted = 0;
fstrans_cookie_t cookie = spl_fstrans_mark();
/* Always try to evict from uncached state. */
arc_last_uncached_flush = ddi_get_lbolt();
evicted += arc_flush_state(arc_uncached, 0, ARC_BUFC_DATA, B_FALSE);
evicted += arc_flush_state(arc_uncached, 0, ARC_BUFC_METADATA, B_FALSE);
/* Evict from other states only if told to. */
if (arc_evict_needed)
evicted += arc_evict();
/*
* If evicted is zero, we couldn't evict anything
* via arc_evict(). This could be due to hash lock
* collisions, but more likely due to the majority of
* arc buffers being unevictable. Therefore, even if
* arc_size is above arc_c, another pass is unlikely to
* be helpful and could potentially cause us to enter an
* infinite loop. Additionally, zthr_iscancelled() is
* checked here so that if the arc is shutting down, the
* broadcast will wake any remaining arc evict waiters.
*/
mutex_enter(&arc_evict_lock);
arc_evict_needed = !zthr_iscancelled(arc_evict_zthr) &&
evicted > 0 && aggsum_compare(&arc_sums.arcstat_size, arc_c) > 0;
if (!arc_evict_needed) {
/*
* We're either no longer overflowing, or we
* can't evict anything more, so we should wake
* arc_get_data_impl() sooner.
*/
arc_evict_waiter_t *aw;
while ((aw = list_remove_head(&arc_evict_waiters)) != NULL) {
cv_broadcast(&aw->aew_cv);
}
arc_set_need_free();
}
mutex_exit(&arc_evict_lock);
spl_fstrans_unmark(cookie);
}
static boolean_t
arc_reap_cb_check(void *arg, zthr_t *zthr)
{
(void) arg, (void) zthr;
int64_t free_memory = arc_available_memory();
static int reap_cb_check_counter = 0;
/*
* If a kmem reap is already active, don't schedule more. We must
* check for this because kmem_cache_reap_soon() won't actually
* block on the cache being reaped (this is to prevent callers from
* becoming implicitly blocked by a system-wide kmem reap -- which,
* on a system with many, many full magazines, can take minutes).
*/
if (!kmem_cache_reap_active() && free_memory < 0) {
arc_no_grow = B_TRUE;
arc_warm = B_TRUE;
/*
* Wait at least zfs_grow_retry (default 5) seconds
* before considering growing.
*/
arc_growtime = gethrtime() + SEC2NSEC(arc_grow_retry);
return (B_TRUE);
} else if (free_memory < arc_c >> arc_no_grow_shift) {
arc_no_grow = B_TRUE;
} else if (gethrtime() >= arc_growtime) {
arc_no_grow = B_FALSE;
}
/*
* Called unconditionally every 60 seconds to reclaim unused
* zstd compression and decompression context. This is done
* here to avoid the need for an independent thread.
*/
if (!((reap_cb_check_counter++) % 60))
zfs_zstd_cache_reap_now();
return (B_FALSE);
}
/*
* Keep enough free memory in the system by reaping the ARC's kmem
* caches. To cause more slabs to be reapable, we may reduce the
* target size of the cache (arc_c), causing the arc_evict_cb()
* to free more buffers.
*/
static void
arc_reap_cb(void *arg, zthr_t *zthr)
{
(void) arg, (void) zthr;
int64_t free_memory;
fstrans_cookie_t cookie = spl_fstrans_mark();
/*
* Kick off asynchronous kmem_reap()'s of all our caches.
*/
arc_kmem_reap_soon();
/*
* Wait at least arc_kmem_cache_reap_retry_ms between
* arc_kmem_reap_soon() calls. Without this check it is possible to
* end up in a situation where we spend lots of time reaping
* caches, while we're near arc_c_min. Waiting here also gives the
* subsequent free memory check a chance of finding that the
* asynchronous reap has already freed enough memory, and we don't
* need to call arc_reduce_target_size().
*/
delay((hz * arc_kmem_cache_reap_retry_ms + 999) / 1000);
/*
* Reduce the target size as needed to maintain the amount of free
* memory in the system at a fraction of the arc_size (1/128th by
* default). If oversubscribed (free_memory < 0) then reduce the
* target arc_size by the deficit amount plus the fractional
* amount. If free memory is positive but less than the fractional
* amount, reduce by what is needed to hit the fractional amount.
*/
free_memory = arc_available_memory();
int64_t can_free = arc_c - arc_c_min;
if (can_free > 0) {
int64_t to_free = (can_free >> arc_shrink_shift) - free_memory;
if (to_free > 0)
arc_reduce_target_size(to_free);
}
spl_fstrans_unmark(cookie);
}
#ifdef _KERNEL
/*
* Determine the amount of memory eligible for eviction contained in the
* ARC. All clean data reported by the ghost lists can always be safely
* evicted. Due to arc_c_min, the same does not hold for all clean data
* contained by the regular mru and mfu lists.
*
* In the case of the regular mru and mfu lists, we need to report as
* much clean data as possible, such that evicting that same reported
* data will not bring arc_size below arc_c_min. Thus, in certain
* circumstances, the total amount of clean data in the mru and mfu
* lists might not actually be evictable.
*
* The following two distinct cases are accounted for:
*
* 1. The sum of the amount of dirty data contained by both the mru and
* mfu lists, plus the ARC's other accounting (e.g. the anon list),
* is greater than or equal to arc_c_min.
* (i.e. amount of dirty data >= arc_c_min)
*
* This is the easy case; all clean data contained by the mru and mfu
* lists is evictable. Evicting all clean data can only drop arc_size
* to the amount of dirty data, which is greater than arc_c_min.
*
* 2. The sum of the amount of dirty data contained by both the mru and
* mfu lists, plus the ARC's other accounting (e.g. the anon list),
* is less than arc_c_min.
* (i.e. arc_c_min > amount of dirty data)
*
* 2.1. arc_size is greater than or equal arc_c_min.
* (i.e. arc_size >= arc_c_min > amount of dirty data)
*
* In this case, not all clean data from the regular mru and mfu
* lists is actually evictable; we must leave enough clean data
* to keep arc_size above arc_c_min. Thus, the maximum amount of
* evictable data from the two lists combined, is exactly the
* difference between arc_size and arc_c_min.
*
* 2.2. arc_size is less than arc_c_min
* (i.e. arc_c_min > arc_size > amount of dirty data)
*
* In this case, none of the data contained in the mru and mfu
* lists is evictable, even if it's clean. Since arc_size is
* already below arc_c_min, evicting any more would only
* increase this negative difference.
*/
#endif /* _KERNEL */
/*
* Adapt arc info given the number of bytes we are trying to add and
* the state that we are coming from. This function is only called
* when we are adding new content to the cache.
*/
static void
arc_adapt(uint64_t bytes)
{
/*
* Wake reap thread if we do not have any available memory
*/
if (arc_reclaim_needed()) {
zthr_wakeup(arc_reap_zthr);
return;
}
if (arc_no_grow)
return;
if (arc_c >= arc_c_max)
return;
/*
* If we're within (2 * maxblocksize) bytes of the target
* cache size, increment the target cache size
*/
if (aggsum_upper_bound(&arc_sums.arcstat_size) +
2 * SPA_MAXBLOCKSIZE >= arc_c) {
uint64_t dc = MAX(bytes, SPA_OLD_MAXBLOCKSIZE);
if (atomic_add_64_nv(&arc_c, dc) > arc_c_max)
arc_c = arc_c_max;
}
}
/*
* Check if arc_size has grown past our upper threshold, determined by
* zfs_arc_overflow_shift.
*/
static arc_ovf_level_t
arc_is_overflowing(boolean_t use_reserve)
{
/* Always allow at least one block of overflow */
int64_t overflow = MAX(SPA_MAXBLOCKSIZE,
arc_c >> zfs_arc_overflow_shift);
/*
* We just compare the lower bound here for performance reasons. Our
* primary goals are to make sure that the arc never grows without
* bound, and that it can reach its maximum size. This check
* accomplishes both goals. The maximum amount we could run over by is
* 2 * aggsum_borrow_multiplier * NUM_CPUS * the average size of a block
* in the ARC. In practice, that's in the tens of MB, which is low
* enough to be safe.
*/
int64_t over = aggsum_lower_bound(&arc_sums.arcstat_size) -
arc_c - overflow / 2;
if (!use_reserve)
overflow /= 2;
return (over < 0 ? ARC_OVF_NONE :
over < overflow ? ARC_OVF_SOME : ARC_OVF_SEVERE);
}
static abd_t *
arc_get_data_abd(arc_buf_hdr_t *hdr, uint64_t size, const void *tag,
int alloc_flags)
{
arc_buf_contents_t type = arc_buf_type(hdr);
arc_get_data_impl(hdr, size, tag, alloc_flags);
if (alloc_flags & ARC_HDR_ALLOC_LINEAR)
return (abd_alloc_linear(size, type == ARC_BUFC_METADATA));
else
return (abd_alloc(size, type == ARC_BUFC_METADATA));
}
static void *
arc_get_data_buf(arc_buf_hdr_t *hdr, uint64_t size, const void *tag)
{
arc_buf_contents_t type = arc_buf_type(hdr);
arc_get_data_impl(hdr, size, tag, 0);
if (type == ARC_BUFC_METADATA) {
return (zio_buf_alloc(size));
} else {
ASSERT(type == ARC_BUFC_DATA);
return (zio_data_buf_alloc(size));
}
}
/*
* Wait for the specified amount of data (in bytes) to be evicted from the
* ARC, and for there to be sufficient free memory in the system. Waiting for
* eviction ensures that the memory used by the ARC decreases. Waiting for
* free memory ensures that the system won't run out of free pages, regardless
* of ARC behavior and settings. See arc_lowmem_init().
*/
void
arc_wait_for_eviction(uint64_t amount, boolean_t use_reserve)
{
switch (arc_is_overflowing(use_reserve)) {
case ARC_OVF_NONE:
return;
case ARC_OVF_SOME:
/*
* This is a bit racy without taking arc_evict_lock, but the
* worst that can happen is we either call zthr_wakeup() extra
* time due to race with other thread here, or the set flag
* get cleared by arc_evict_cb(), which is unlikely due to
* big hysteresis, but also not important since at this level
* of overflow the eviction is purely advisory. Same time
* taking the global lock here every time without waiting for
* the actual eviction creates a significant lock contention.
*/
if (!arc_evict_needed) {
arc_evict_needed = B_TRUE;
zthr_wakeup(arc_evict_zthr);
}
return;
case ARC_OVF_SEVERE:
default:
{
arc_evict_waiter_t aw;
list_link_init(&aw.aew_node);
cv_init(&aw.aew_cv, NULL, CV_DEFAULT, NULL);
uint64_t last_count = 0;
mutex_enter(&arc_evict_lock);
if (!list_is_empty(&arc_evict_waiters)) {
arc_evict_waiter_t *last =
list_tail(&arc_evict_waiters);
last_count = last->aew_count;
} else if (!arc_evict_needed) {
arc_evict_needed = B_TRUE;
zthr_wakeup(arc_evict_zthr);
}
/*
* Note, the last waiter's count may be less than
* arc_evict_count if we are low on memory in which
* case arc_evict_state_impl() may have deferred
* wakeups (but still incremented arc_evict_count).
*/
aw.aew_count = MAX(last_count, arc_evict_count) + amount;
list_insert_tail(&arc_evict_waiters, &aw);
arc_set_need_free();
DTRACE_PROBE3(arc__wait__for__eviction,
uint64_t, amount,
uint64_t, arc_evict_count,
uint64_t, aw.aew_count);
/*
* We will be woken up either when arc_evict_count reaches
* aew_count, or when the ARC is no longer overflowing and
* eviction completes.
* In case of "false" wakeup, we will still be on the list.
*/
do {
cv_wait(&aw.aew_cv, &arc_evict_lock);
} while (list_link_active(&aw.aew_node));
mutex_exit(&arc_evict_lock);
cv_destroy(&aw.aew_cv);
}
}
}
/*
* Allocate a block and return it to the caller. If we are hitting the
* hard limit for the cache size, we must sleep, waiting for the eviction
* thread to catch up. If we're past the target size but below the hard
* limit, we'll only signal the reclaim thread and continue on.
*/
static void
arc_get_data_impl(arc_buf_hdr_t *hdr, uint64_t size, const void *tag,
int alloc_flags)
{
arc_adapt(size);
/*
* If arc_size is currently overflowing, we must be adding data
* faster than we are evicting. To ensure we don't compound the
* problem by adding more data and forcing arc_size to grow even
* further past it's target size, we wait for the eviction thread to
* make some progress. We also wait for there to be sufficient free
* memory in the system, as measured by arc_free_memory().
*
* Specifically, we wait for zfs_arc_eviction_pct percent of the
* requested size to be evicted. This should be more than 100%, to
* ensure that that progress is also made towards getting arc_size
* under arc_c. See the comment above zfs_arc_eviction_pct.
*/
arc_wait_for_eviction(size * zfs_arc_eviction_pct / 100,
alloc_flags & ARC_HDR_USE_RESERVE);
arc_buf_contents_t type = arc_buf_type(hdr);
if (type == ARC_BUFC_METADATA) {
arc_space_consume(size, ARC_SPACE_META);
} else {
arc_space_consume(size, ARC_SPACE_DATA);
}
/*
* Update the state size. Note that ghost states have a
* "ghost size" and so don't need to be updated.
*/
arc_state_t *state = hdr->b_l1hdr.b_state;
if (!GHOST_STATE(state)) {
(void) zfs_refcount_add_many(&state->arcs_size[type], size,
tag);
/*
* If this is reached via arc_read, the link is
* protected by the hash lock. If reached via
* arc_buf_alloc, the header should not be accessed by
* any other thread. And, if reached via arc_read_done,
* the hash lock will protect it if it's found in the
* hash table; otherwise no other thread should be
* trying to [add|remove]_reference it.
*/
if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) {
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
(void) zfs_refcount_add_many(&state->arcs_esize[type],
size, tag);
}
}
}
static void
arc_free_data_abd(arc_buf_hdr_t *hdr, abd_t *abd, uint64_t size,
const void *tag)
{
arc_free_data_impl(hdr, size, tag);
abd_free(abd);
}
static void
arc_free_data_buf(arc_buf_hdr_t *hdr, void *buf, uint64_t size, const void *tag)
{
arc_buf_contents_t type = arc_buf_type(hdr);
arc_free_data_impl(hdr, size, tag);
if (type == ARC_BUFC_METADATA) {
zio_buf_free(buf, size);
} else {
ASSERT(type == ARC_BUFC_DATA);
zio_data_buf_free(buf, size);
}
}
/*
* Free the arc data buffer.
*/
static void
arc_free_data_impl(arc_buf_hdr_t *hdr, uint64_t size, const void *tag)
{
arc_state_t *state = hdr->b_l1hdr.b_state;
arc_buf_contents_t type = arc_buf_type(hdr);
/* protected by hash lock, if in the hash table */
if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) {
ASSERT(zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt));
ASSERT(state != arc_anon && state != arc_l2c_only);
(void) zfs_refcount_remove_many(&state->arcs_esize[type],
size, tag);
}
(void) zfs_refcount_remove_many(&state->arcs_size[type], size, tag);
VERIFY3U(hdr->b_type, ==, type);
if (type == ARC_BUFC_METADATA) {
arc_space_return(size, ARC_SPACE_META);
} else {
ASSERT(type == ARC_BUFC_DATA);
arc_space_return(size, ARC_SPACE_DATA);
}
}
/*
* This routine is called whenever a buffer is accessed.
*/
static void
arc_access(arc_buf_hdr_t *hdr, arc_flags_t arc_flags, boolean_t hit)
{
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT(HDR_HAS_L1HDR(hdr));
/*
* Update buffer prefetch status.
*/
boolean_t was_prefetch = HDR_PREFETCH(hdr);
boolean_t now_prefetch = arc_flags & ARC_FLAG_PREFETCH;
if (was_prefetch != now_prefetch) {
if (was_prefetch) {
ARCSTAT_CONDSTAT(hit, demand_hit, demand_iohit,
HDR_PRESCIENT_PREFETCH(hdr), prescient, predictive,
prefetch);
}
if (HDR_HAS_L2HDR(hdr))
l2arc_hdr_arcstats_decrement_state(hdr);
if (was_prefetch) {
arc_hdr_clear_flags(hdr,
ARC_FLAG_PREFETCH | ARC_FLAG_PRESCIENT_PREFETCH);
} else {
arc_hdr_set_flags(hdr, ARC_FLAG_PREFETCH);
}
if (HDR_HAS_L2HDR(hdr))
l2arc_hdr_arcstats_increment_state(hdr);
}
if (now_prefetch) {
if (arc_flags & ARC_FLAG_PRESCIENT_PREFETCH) {
arc_hdr_set_flags(hdr, ARC_FLAG_PRESCIENT_PREFETCH);
ARCSTAT_BUMP(arcstat_prescient_prefetch);
} else {
ARCSTAT_BUMP(arcstat_predictive_prefetch);
}
}
if (arc_flags & ARC_FLAG_L2CACHE)
arc_hdr_set_flags(hdr, ARC_FLAG_L2CACHE);
clock_t now = ddi_get_lbolt();
if (hdr->b_l1hdr.b_state == arc_anon) {
arc_state_t *new_state;
/*
* This buffer is not in the cache, and does not appear in
* our "ghost" lists. Add it to the MRU or uncached state.
*/
ASSERT0(hdr->b_l1hdr.b_arc_access);
hdr->b_l1hdr.b_arc_access = now;
if (HDR_UNCACHED(hdr)) {
new_state = arc_uncached;
DTRACE_PROBE1(new_state__uncached, arc_buf_hdr_t *,
hdr);
} else {
new_state = arc_mru;
DTRACE_PROBE1(new_state__mru, arc_buf_hdr_t *, hdr);
}
arc_change_state(new_state, hdr);
} else if (hdr->b_l1hdr.b_state == arc_mru) {
/*
* This buffer has been accessed once recently and either
* its read is still in progress or it is in the cache.
*/
if (HDR_IO_IN_PROGRESS(hdr)) {
hdr->b_l1hdr.b_arc_access = now;
return;
}
hdr->b_l1hdr.b_mru_hits++;
ARCSTAT_BUMP(arcstat_mru_hits);
/*
* If the previous access was a prefetch, then it already
* handled possible promotion, so nothing more to do for now.
*/
if (was_prefetch) {
hdr->b_l1hdr.b_arc_access = now;
return;
}
/*
* If more than ARC_MINTIME have passed from the previous
* hit, promote the buffer to the MFU state.
*/
if (ddi_time_after(now, hdr->b_l1hdr.b_arc_access +
ARC_MINTIME)) {
hdr->b_l1hdr.b_arc_access = now;
DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr);
arc_change_state(arc_mfu, hdr);
}
} else if (hdr->b_l1hdr.b_state == arc_mru_ghost) {
arc_state_t *new_state;
/*
* This buffer has been accessed once recently, but was
* evicted from the cache. Would we have bigger MRU, it
* would be an MRU hit, so handle it the same way, except
* we don't need to check the previous access time.
*/
hdr->b_l1hdr.b_mru_ghost_hits++;
ARCSTAT_BUMP(arcstat_mru_ghost_hits);
hdr->b_l1hdr.b_arc_access = now;
wmsum_add(&arc_mru_ghost->arcs_hits[arc_buf_type(hdr)],
arc_hdr_size(hdr));
if (was_prefetch) {
new_state = arc_mru;
DTRACE_PROBE1(new_state__mru, arc_buf_hdr_t *, hdr);
} else {
new_state = arc_mfu;
DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr);
}
arc_change_state(new_state, hdr);
} else if (hdr->b_l1hdr.b_state == arc_mfu) {
/*
* This buffer has been accessed more than once and either
* still in the cache or being restored from one of ghosts.
*/
if (!HDR_IO_IN_PROGRESS(hdr)) {
hdr->b_l1hdr.b_mfu_hits++;
ARCSTAT_BUMP(arcstat_mfu_hits);
}
hdr->b_l1hdr.b_arc_access = now;
} else if (hdr->b_l1hdr.b_state == arc_mfu_ghost) {
/*
* This buffer has been accessed more than once recently, but
* has been evicted from the cache. Would we have bigger MFU
* it would stay in cache, so move it back to MFU state.
*/
hdr->b_l1hdr.b_mfu_ghost_hits++;
ARCSTAT_BUMP(arcstat_mfu_ghost_hits);
hdr->b_l1hdr.b_arc_access = now;
wmsum_add(&arc_mfu_ghost->arcs_hits[arc_buf_type(hdr)],
arc_hdr_size(hdr));
DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr);
arc_change_state(arc_mfu, hdr);
} else if (hdr->b_l1hdr.b_state == arc_uncached) {
/*
* This buffer is uncacheable, but we got a hit. Probably
* a demand read after prefetch. Nothing more to do here.
*/
if (!HDR_IO_IN_PROGRESS(hdr))
ARCSTAT_BUMP(arcstat_uncached_hits);
hdr->b_l1hdr.b_arc_access = now;
} else if (hdr->b_l1hdr.b_state == arc_l2c_only) {
/*
* This buffer is on the 2nd Level ARC and was not accessed
* for a long time, so treat it as new and put into MRU.
*/
hdr->b_l1hdr.b_arc_access = now;
DTRACE_PROBE1(new_state__mru, arc_buf_hdr_t *, hdr);
arc_change_state(arc_mru, hdr);
} else {
cmn_err(CE_PANIC, "invalid arc state 0x%p",
hdr->b_l1hdr.b_state);
}
}
/*
* This routine is called by dbuf_hold() to update the arc_access() state
* which otherwise would be skipped for entries in the dbuf cache.
*/
void
arc_buf_access(arc_buf_t *buf)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
/*
* Avoid taking the hash_lock when possible as an optimization.
* The header must be checked again under the hash_lock in order
* to handle the case where it is concurrently being released.
*/
if (hdr->b_l1hdr.b_state == arc_anon || HDR_EMPTY(hdr))
return;
kmutex_t *hash_lock = HDR_LOCK(hdr);
mutex_enter(hash_lock);
if (hdr->b_l1hdr.b_state == arc_anon || HDR_EMPTY(hdr)) {
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_access_skip);
return;
}
ASSERT(hdr->b_l1hdr.b_state == arc_mru ||
hdr->b_l1hdr.b_state == arc_mfu ||
hdr->b_l1hdr.b_state == arc_uncached);
DTRACE_PROBE1(arc__hit, arc_buf_hdr_t *, hdr);
arc_access(hdr, 0, B_TRUE);
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_hits);
ARCSTAT_CONDSTAT(B_TRUE /* demand */, demand, prefetch,
!HDR_ISTYPE_METADATA(hdr), data, metadata, hits);
}
/* a generic arc_read_done_func_t which you can use */
void
arc_bcopy_func(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp,
arc_buf_t *buf, void *arg)
{
(void) zio, (void) zb, (void) bp;
if (buf == NULL)
return;
memcpy(arg, buf->b_data, arc_buf_size(buf));
arc_buf_destroy(buf, arg);
}
/* a generic arc_read_done_func_t */
void
arc_getbuf_func(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp,
arc_buf_t *buf, void *arg)
{
(void) zb, (void) bp;
arc_buf_t **bufp = arg;
if (buf == NULL) {
ASSERT(zio == NULL || zio->io_error != 0);
*bufp = NULL;
} else {
ASSERT(zio == NULL || zio->io_error == 0);
*bufp = buf;
ASSERT(buf->b_data != NULL);
}
}
static void
arc_hdr_verify(arc_buf_hdr_t *hdr, blkptr_t *bp)
{
if (BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp)) {
ASSERT3U(HDR_GET_PSIZE(hdr), ==, 0);
ASSERT3U(arc_hdr_get_compress(hdr), ==, ZIO_COMPRESS_OFF);
} else {
if (HDR_COMPRESSION_ENABLED(hdr)) {
ASSERT3U(arc_hdr_get_compress(hdr), ==,
BP_GET_COMPRESS(bp));
}
ASSERT3U(HDR_GET_LSIZE(hdr), ==, BP_GET_LSIZE(bp));
ASSERT3U(HDR_GET_PSIZE(hdr), ==, BP_GET_PSIZE(bp));
ASSERT3U(!!HDR_PROTECTED(hdr), ==, BP_IS_PROTECTED(bp));
}
}
static void
arc_read_done(zio_t *zio)
{
blkptr_t *bp = zio->io_bp;
arc_buf_hdr_t *hdr = zio->io_private;
kmutex_t *hash_lock = NULL;
arc_callback_t *callback_list;
arc_callback_t *acb;
/*
* The hdr was inserted into hash-table and removed from lists
* prior to starting I/O. We should find this header, since
* it's in the hash table, and it should be legit since it's
* not possible to evict it during the I/O. The only possible
* reason for it not to be found is if we were freed during the
* read.
*/
if (HDR_IN_HASH_TABLE(hdr)) {
arc_buf_hdr_t *found;
ASSERT3U(hdr->b_birth, ==, BP_PHYSICAL_BIRTH(zio->io_bp));
ASSERT3U(hdr->b_dva.dva_word[0], ==,
BP_IDENTITY(zio->io_bp)->dva_word[0]);
ASSERT3U(hdr->b_dva.dva_word[1], ==,
BP_IDENTITY(zio->io_bp)->dva_word[1]);
found = buf_hash_find(hdr->b_spa, zio->io_bp, &hash_lock);
ASSERT((found == hdr &&
DVA_EQUAL(&hdr->b_dva, BP_IDENTITY(zio->io_bp))) ||
(found == hdr && HDR_L2_READING(hdr)));
ASSERT3P(hash_lock, !=, NULL);
}
if (BP_IS_PROTECTED(bp)) {
hdr->b_crypt_hdr.b_ot = BP_GET_TYPE(bp);
hdr->b_crypt_hdr.b_dsobj = zio->io_bookmark.zb_objset;
zio_crypt_decode_params_bp(bp, hdr->b_crypt_hdr.b_salt,
hdr->b_crypt_hdr.b_iv);
if (zio->io_error == 0) {
if (BP_GET_TYPE(bp) == DMU_OT_INTENT_LOG) {
void *tmpbuf;
tmpbuf = abd_borrow_buf_copy(zio->io_abd,
sizeof (zil_chain_t));
zio_crypt_decode_mac_zil(tmpbuf,
hdr->b_crypt_hdr.b_mac);
abd_return_buf(zio->io_abd, tmpbuf,
sizeof (zil_chain_t));
} else {
zio_crypt_decode_mac_bp(bp,
hdr->b_crypt_hdr.b_mac);
}
}
}
if (zio->io_error == 0) {
/* byteswap if necessary */
if (BP_SHOULD_BYTESWAP(zio->io_bp)) {
if (BP_GET_LEVEL(zio->io_bp) > 0) {
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_UINT64;
} else {
hdr->b_l1hdr.b_byteswap =
DMU_OT_BYTESWAP(BP_GET_TYPE(zio->io_bp));
}
} else {
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_NUMFUNCS;
}
if (!HDR_L2_READING(hdr)) {
hdr->b_complevel = zio->io_prop.zp_complevel;
}
}
arc_hdr_clear_flags(hdr, ARC_FLAG_L2_EVICTED);
if (l2arc_noprefetch && HDR_PREFETCH(hdr))
arc_hdr_clear_flags(hdr, ARC_FLAG_L2CACHE);
callback_list = hdr->b_l1hdr.b_acb;
ASSERT3P(callback_list, !=, NULL);
hdr->b_l1hdr.b_acb = NULL;
/*
* If a read request has a callback (i.e. acb_done is not NULL), then we
* make a buf containing the data according to the parameters which were
* passed in. The implementation of arc_buf_alloc_impl() ensures that we
* aren't needlessly decompressing the data multiple times.
*/
int callback_cnt = 0;
for (acb = callback_list; acb != NULL; acb = acb->acb_next) {
/* We need the last one to call below in original order. */
callback_list = acb;
if (!acb->acb_done || acb->acb_nobuf)
continue;
callback_cnt++;
if (zio->io_error != 0)
continue;
int error = arc_buf_alloc_impl(hdr, zio->io_spa,
&acb->acb_zb, acb->acb_private, acb->acb_encrypted,
acb->acb_compressed, acb->acb_noauth, B_TRUE,
&acb->acb_buf);
/*
* Assert non-speculative zios didn't fail because an
* encryption key wasn't loaded
*/
ASSERT((zio->io_flags & ZIO_FLAG_SPECULATIVE) ||
error != EACCES);
/*
* If we failed to decrypt, report an error now (as the zio
* layer would have done if it had done the transforms).
*/
if (error == ECKSUM) {
ASSERT(BP_IS_PROTECTED(bp));
error = SET_ERROR(EIO);
if ((zio->io_flags & ZIO_FLAG_SPECULATIVE) == 0) {
spa_log_error(zio->io_spa, &acb->acb_zb,
&zio->io_bp->blk_birth);
(void) zfs_ereport_post(
FM_EREPORT_ZFS_AUTHENTICATION,
zio->io_spa, NULL, &acb->acb_zb, zio, 0);
}
}
if (error != 0) {
/*
* Decompression or decryption failed. Set
* io_error so that when we call acb_done
* (below), we will indicate that the read
* failed. Note that in the unusual case
* where one callback is compressed and another
* uncompressed, we will mark all of them
* as failed, even though the uncompressed
* one can't actually fail. In this case,
* the hdr will not be anonymous, because
* if there are multiple callbacks, it's
* because multiple threads found the same
* arc buf in the hash table.
*/
zio->io_error = error;
}
}
/*
* If there are multiple callbacks, we must have the hash lock,
* because the only way for multiple threads to find this hdr is
* in the hash table. This ensures that if there are multiple
* callbacks, the hdr is not anonymous. If it were anonymous,
* we couldn't use arc_buf_destroy() in the error case below.
*/
ASSERT(callback_cnt < 2 || hash_lock != NULL);
if (zio->io_error == 0) {
arc_hdr_verify(hdr, zio->io_bp);
} else {
arc_hdr_set_flags(hdr, ARC_FLAG_IO_ERROR);
if (hdr->b_l1hdr.b_state != arc_anon)
arc_change_state(arc_anon, hdr);
if (HDR_IN_HASH_TABLE(hdr))
buf_hash_remove(hdr);
}
/*
* Broadcast before we drop the hash_lock to avoid the possibility
* that the hdr (and hence the cv) might be freed before we get to
* the cv_broadcast().
*/
cv_broadcast(&hdr->b_l1hdr.b_cv);
arc_hdr_clear_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
(void) remove_reference(hdr, hdr);
if (hash_lock != NULL)
mutex_exit(hash_lock);
/* execute each callback and free its structure */
while ((acb = callback_list) != NULL) {
if (acb->acb_done != NULL) {
if (zio->io_error != 0 && acb->acb_buf != NULL) {
/*
* If arc_buf_alloc_impl() fails during
* decompression, the buf will still be
* allocated, and needs to be freed here.
*/
arc_buf_destroy(acb->acb_buf,
acb->acb_private);
acb->acb_buf = NULL;
}
acb->acb_done(zio, &zio->io_bookmark, zio->io_bp,
acb->acb_buf, acb->acb_private);
}
if (acb->acb_zio_dummy != NULL) {
acb->acb_zio_dummy->io_error = zio->io_error;
zio_nowait(acb->acb_zio_dummy);
}
callback_list = acb->acb_prev;
if (acb->acb_wait) {
mutex_enter(&acb->acb_wait_lock);
acb->acb_wait_error = zio->io_error;
acb->acb_wait = B_FALSE;
cv_signal(&acb->acb_wait_cv);
mutex_exit(&acb->acb_wait_lock);
/* acb will be freed by the waiting thread. */
} else {
kmem_free(acb, sizeof (arc_callback_t));
}
}
}
/*
* "Read" the block at the specified DVA (in bp) via the
* cache. If the block is found in the cache, invoke the provided
* callback immediately and return. Note that the `zio' parameter
* in the callback will be NULL in this case, since no IO was
* required. If the block is not in the cache pass the read request
* on to the spa with a substitute callback function, so that the
* requested block will be added to the cache.
*
* If a read request arrives for a block that has a read in-progress,
* either wait for the in-progress read to complete (and return the
* results); or, if this is a read with a "done" func, add a record
* to the read to invoke the "done" func when the read completes,
* and return; or just return.
*
* arc_read_done() will invoke all the requested "done" functions
* for readers of this block.
*/
int
arc_read(zio_t *pio, spa_t *spa, const blkptr_t *bp,
arc_read_done_func_t *done, void *private, zio_priority_t priority,
int zio_flags, arc_flags_t *arc_flags, const zbookmark_phys_t *zb)
{
arc_buf_hdr_t *hdr = NULL;
kmutex_t *hash_lock = NULL;
zio_t *rzio;
uint64_t guid = spa_load_guid(spa);
boolean_t compressed_read = (zio_flags & ZIO_FLAG_RAW_COMPRESS) != 0;
boolean_t encrypted_read = BP_IS_ENCRYPTED(bp) &&
(zio_flags & ZIO_FLAG_RAW_ENCRYPT) != 0;
boolean_t noauth_read = BP_IS_AUTHENTICATED(bp) &&
(zio_flags & ZIO_FLAG_RAW_ENCRYPT) != 0;
boolean_t embedded_bp = !!BP_IS_EMBEDDED(bp);
boolean_t no_buf = *arc_flags & ARC_FLAG_NO_BUF;
arc_buf_t *buf = NULL;
int rc = 0;
ASSERT(!embedded_bp ||
BPE_GET_ETYPE(bp) == BP_EMBEDDED_TYPE_DATA);
ASSERT(!BP_IS_HOLE(bp));
ASSERT(!BP_IS_REDACTED(bp));
/*
* Normally SPL_FSTRANS will already be set since kernel threads which
* expect to call the DMU interfaces will set it when created. System
* calls are similarly handled by setting/cleaning the bit in the
* registered callback (module/os/.../zfs/zpl_*).
*
* External consumers such as Lustre which call the exported DMU
* interfaces may not have set SPL_FSTRANS. To avoid a deadlock
* on the hash_lock always set and clear the bit.
*/
fstrans_cookie_t cookie = spl_fstrans_mark();
top:
/*
* Verify the block pointer contents are reasonable. This should
* always be the case since the blkptr is protected by a checksum.
* However, if there is damage it's desirable to detect this early
* and treat it as a checksum error. This allows an alternate blkptr
* to be tried when one is available (e.g. ditto blocks).
*/
if (!zfs_blkptr_verify(spa, bp, zio_flags & ZIO_FLAG_CONFIG_WRITER,
BLK_VERIFY_LOG)) {
rc = SET_ERROR(ECKSUM);
goto done;
}
if (!embedded_bp) {
/*
* Embedded BP's have no DVA and require no I/O to "read".
* Create an anonymous arc buf to back it.
*/
hdr = buf_hash_find(guid, bp, &hash_lock);
}
/*
* Determine if we have an L1 cache hit or a cache miss. For simplicity
* we maintain encrypted data separately from compressed / uncompressed
* data. If the user is requesting raw encrypted data and we don't have
* that in the header we will read from disk to guarantee that we can
* get it even if the encryption keys aren't loaded.
*/
if (hdr != NULL && HDR_HAS_L1HDR(hdr) && (HDR_HAS_RABD(hdr) ||
(hdr->b_l1hdr.b_pabd != NULL && !encrypted_read))) {
boolean_t is_data = !HDR_ISTYPE_METADATA(hdr);
if (HDR_IO_IN_PROGRESS(hdr)) {
if (*arc_flags & ARC_FLAG_CACHED_ONLY) {
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_cached_only_in_progress);
rc = SET_ERROR(ENOENT);
goto done;
}
zio_t *head_zio = hdr->b_l1hdr.b_acb->acb_zio_head;
ASSERT3P(head_zio, !=, NULL);
if ((hdr->b_flags & ARC_FLAG_PRIO_ASYNC_READ) &&
priority == ZIO_PRIORITY_SYNC_READ) {
/*
* This is a sync read that needs to wait for
* an in-flight async read. Request that the
* zio have its priority upgraded.
*/
zio_change_priority(head_zio, priority);
DTRACE_PROBE1(arc__async__upgrade__sync,
arc_buf_hdr_t *, hdr);
ARCSTAT_BUMP(arcstat_async_upgrade_sync);
}
DTRACE_PROBE1(arc__iohit, arc_buf_hdr_t *, hdr);
arc_access(hdr, *arc_flags, B_FALSE);
/*
* If there are multiple threads reading the same block
* and that block is not yet in the ARC, then only one
* thread will do the physical I/O and all other
* threads will wait until that I/O completes.
* Synchronous reads use the acb_wait_cv whereas nowait
* reads register a callback. Both are signalled/called
* in arc_read_done.
*
* Errors of the physical I/O may need to be propagated.
* Synchronous read errors are returned here from
* arc_read_done via acb_wait_error. Nowait reads
* attach the acb_zio_dummy zio to pio and
* arc_read_done propagates the physical I/O's io_error
* to acb_zio_dummy, and thereby to pio.
*/
arc_callback_t *acb = NULL;
if (done || pio || *arc_flags & ARC_FLAG_WAIT) {
acb = kmem_zalloc(sizeof (arc_callback_t),
KM_SLEEP);
acb->acb_done = done;
acb->acb_private = private;
acb->acb_compressed = compressed_read;
acb->acb_encrypted = encrypted_read;
acb->acb_noauth = noauth_read;
acb->acb_nobuf = no_buf;
if (*arc_flags & ARC_FLAG_WAIT) {
acb->acb_wait = B_TRUE;
mutex_init(&acb->acb_wait_lock, NULL,
MUTEX_DEFAULT, NULL);
cv_init(&acb->acb_wait_cv, NULL,
CV_DEFAULT, NULL);
}
acb->acb_zb = *zb;
if (pio != NULL) {
acb->acb_zio_dummy = zio_null(pio,
spa, NULL, NULL, NULL, zio_flags);
}
acb->acb_zio_head = head_zio;
acb->acb_next = hdr->b_l1hdr.b_acb;
if (hdr->b_l1hdr.b_acb)
hdr->b_l1hdr.b_acb->acb_prev = acb;
hdr->b_l1hdr.b_acb = acb;
}
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_iohits);
ARCSTAT_CONDSTAT(!(*arc_flags & ARC_FLAG_PREFETCH),
demand, prefetch, is_data, data, metadata, iohits);
if (*arc_flags & ARC_FLAG_WAIT) {
mutex_enter(&acb->acb_wait_lock);
while (acb->acb_wait) {
cv_wait(&acb->acb_wait_cv,
&acb->acb_wait_lock);
}
rc = acb->acb_wait_error;
mutex_exit(&acb->acb_wait_lock);
mutex_destroy(&acb->acb_wait_lock);
cv_destroy(&acb->acb_wait_cv);
kmem_free(acb, sizeof (arc_callback_t));
}
goto out;
}
ASSERT(hdr->b_l1hdr.b_state == arc_mru ||
hdr->b_l1hdr.b_state == arc_mfu ||
hdr->b_l1hdr.b_state == arc_uncached);
DTRACE_PROBE1(arc__hit, arc_buf_hdr_t *, hdr);
arc_access(hdr, *arc_flags, B_TRUE);
if (done && !no_buf) {
ASSERT(!embedded_bp || !BP_IS_HOLE(bp));
/* Get a buf with the desired data in it. */
rc = arc_buf_alloc_impl(hdr, spa, zb, private,
encrypted_read, compressed_read, noauth_read,
B_TRUE, &buf);
if (rc == ECKSUM) {
/*
* Convert authentication and decryption errors
* to EIO (and generate an ereport if needed)
* before leaving the ARC.
*/
rc = SET_ERROR(EIO);
if ((zio_flags & ZIO_FLAG_SPECULATIVE) == 0) {
spa_log_error(spa, zb, &hdr->b_birth);
(void) zfs_ereport_post(
FM_EREPORT_ZFS_AUTHENTICATION,
spa, NULL, zb, NULL, 0);
}
}
if (rc != 0) {
arc_buf_destroy_impl(buf);
buf = NULL;
(void) remove_reference(hdr, private);
}
/* assert any errors weren't due to unloaded keys */
ASSERT((zio_flags & ZIO_FLAG_SPECULATIVE) ||
rc != EACCES);
}
mutex_exit(hash_lock);
ARCSTAT_BUMP(arcstat_hits);
ARCSTAT_CONDSTAT(!(*arc_flags & ARC_FLAG_PREFETCH),
demand, prefetch, is_data, data, metadata, hits);
*arc_flags |= ARC_FLAG_CACHED;
goto done;
} else {
uint64_t lsize = BP_GET_LSIZE(bp);
uint64_t psize = BP_GET_PSIZE(bp);
arc_callback_t *acb;
vdev_t *vd = NULL;
uint64_t addr = 0;
boolean_t devw = B_FALSE;
uint64_t size;
abd_t *hdr_abd;
int alloc_flags = encrypted_read ? ARC_HDR_ALLOC_RDATA : 0;
arc_buf_contents_t type = BP_GET_BUFC_TYPE(bp);
if (*arc_flags & ARC_FLAG_CACHED_ONLY) {
if (hash_lock != NULL)
mutex_exit(hash_lock);
rc = SET_ERROR(ENOENT);
goto done;
}
if (hdr == NULL) {
/*
* This block is not in the cache or it has
* embedded data.
*/
arc_buf_hdr_t *exists = NULL;
hdr = arc_hdr_alloc(spa_load_guid(spa), psize, lsize,
BP_IS_PROTECTED(bp), BP_GET_COMPRESS(bp), 0, type);
if (!embedded_bp) {
hdr->b_dva = *BP_IDENTITY(bp);
hdr->b_birth = BP_PHYSICAL_BIRTH(bp);
exists = buf_hash_insert(hdr, &hash_lock);
}
if (exists != NULL) {
/* somebody beat us to the hash insert */
mutex_exit(hash_lock);
buf_discard_identity(hdr);
arc_hdr_destroy(hdr);
goto top; /* restart the IO request */
}
} else {
/*
* This block is in the ghost cache or encrypted data
* was requested and we didn't have it. If it was
* L2-only (and thus didn't have an L1 hdr),
* we realloc the header to add an L1 hdr.
*/
if (!HDR_HAS_L1HDR(hdr)) {
hdr = arc_hdr_realloc(hdr, hdr_l2only_cache,
hdr_full_cache);
}
if (GHOST_STATE(hdr->b_l1hdr.b_state)) {
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT0(zfs_refcount_count(
&hdr->b_l1hdr.b_refcnt));
ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL);
#ifdef ZFS_DEBUG
ASSERT3P(hdr->b_l1hdr.b_freeze_cksum, ==, NULL);
#endif
} else if (HDR_IO_IN_PROGRESS(hdr)) {
/*
* If this header already had an IO in progress
* and we are performing another IO to fetch
* encrypted data we must wait until the first
* IO completes so as not to confuse
* arc_read_done(). This should be very rare
* and so the performance impact shouldn't
* matter.
*/
cv_wait(&hdr->b_l1hdr.b_cv, hash_lock);
mutex_exit(hash_lock);
goto top;
}
}
if (*arc_flags & ARC_FLAG_UNCACHED) {
arc_hdr_set_flags(hdr, ARC_FLAG_UNCACHED);
if (!encrypted_read)
alloc_flags |= ARC_HDR_ALLOC_LINEAR;
}
/*
* Take additional reference for IO_IN_PROGRESS. It stops
* arc_access() from putting this header without any buffers
* and so other references but obviously nonevictable onto
* the evictable list of MRU or MFU state.
*/
add_reference(hdr, hdr);
if (!embedded_bp)
arc_access(hdr, *arc_flags, B_FALSE);
arc_hdr_set_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
arc_hdr_alloc_abd(hdr, alloc_flags);
if (encrypted_read) {
ASSERT(HDR_HAS_RABD(hdr));
size = HDR_GET_PSIZE(hdr);
hdr_abd = hdr->b_crypt_hdr.b_rabd;
zio_flags |= ZIO_FLAG_RAW;
} else {
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
size = arc_hdr_size(hdr);
hdr_abd = hdr->b_l1hdr.b_pabd;
if (arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF) {
zio_flags |= ZIO_FLAG_RAW_COMPRESS;
}
/*
* For authenticated bp's, we do not ask the ZIO layer
* to authenticate them since this will cause the entire
* IO to fail if the key isn't loaded. Instead, we
* defer authentication until arc_buf_fill(), which will
* verify the data when the key is available.
*/
if (BP_IS_AUTHENTICATED(bp))
zio_flags |= ZIO_FLAG_RAW_ENCRYPT;
}
if (BP_IS_AUTHENTICATED(bp))
arc_hdr_set_flags(hdr, ARC_FLAG_NOAUTH);
if (BP_GET_LEVEL(bp) > 0)
arc_hdr_set_flags(hdr, ARC_FLAG_INDIRECT);
ASSERT(!GHOST_STATE(hdr->b_l1hdr.b_state));
acb = kmem_zalloc(sizeof (arc_callback_t), KM_SLEEP);
acb->acb_done = done;
acb->acb_private = private;
acb->acb_compressed = compressed_read;
acb->acb_encrypted = encrypted_read;
acb->acb_noauth = noauth_read;
acb->acb_zb = *zb;
ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
hdr->b_l1hdr.b_acb = acb;
if (HDR_HAS_L2HDR(hdr) &&
(vd = hdr->b_l2hdr.b_dev->l2ad_vdev) != NULL) {
devw = hdr->b_l2hdr.b_dev->l2ad_writing;
addr = hdr->b_l2hdr.b_daddr;
/*
* Lock out L2ARC device removal.
*/
if (vdev_is_dead(vd) ||
!spa_config_tryenter(spa, SCL_L2ARC, vd, RW_READER))
vd = NULL;
}
/*
* We count both async reads and scrub IOs as asynchronous so
* that both can be upgraded in the event of a cache hit while
* the read IO is still in-flight.
*/
if (priority == ZIO_PRIORITY_ASYNC_READ ||
priority == ZIO_PRIORITY_SCRUB)
arc_hdr_set_flags(hdr, ARC_FLAG_PRIO_ASYNC_READ);
else
arc_hdr_clear_flags(hdr, ARC_FLAG_PRIO_ASYNC_READ);
/*
* At this point, we have a level 1 cache miss or a blkptr
* with embedded data. Try again in L2ARC if possible.
*/
ASSERT3U(HDR_GET_LSIZE(hdr), ==, lsize);
/*
* Skip ARC stat bump for block pointers with embedded
* data. The data are read from the blkptr itself via
* decode_embedded_bp_compressed().
*/
if (!embedded_bp) {
DTRACE_PROBE4(arc__miss, arc_buf_hdr_t *, hdr,
blkptr_t *, bp, uint64_t, lsize,
zbookmark_phys_t *, zb);
ARCSTAT_BUMP(arcstat_misses);
ARCSTAT_CONDSTAT(!(*arc_flags & ARC_FLAG_PREFETCH),
demand, prefetch, !HDR_ISTYPE_METADATA(hdr), data,
metadata, misses);
zfs_racct_read(size, 1);
}
/* Check if the spa even has l2 configured */
const boolean_t spa_has_l2 = l2arc_ndev != 0 &&
spa->spa_l2cache.sav_count > 0;
if (vd != NULL && spa_has_l2 && !(l2arc_norw && devw)) {
/*
* Read from the L2ARC if the following are true:
* 1. The L2ARC vdev was previously cached.
* 2. This buffer still has L2ARC metadata.
* 3. This buffer isn't currently writing to the L2ARC.
* 4. The L2ARC entry wasn't evicted, which may
* also have invalidated the vdev.
* 5. This isn't prefetch or l2arc_noprefetch is 0.
*/
if (HDR_HAS_L2HDR(hdr) &&
!HDR_L2_WRITING(hdr) && !HDR_L2_EVICTED(hdr) &&
!(l2arc_noprefetch &&
(*arc_flags & ARC_FLAG_PREFETCH))) {
l2arc_read_callback_t *cb;
abd_t *abd;
uint64_t asize;
DTRACE_PROBE1(l2arc__hit, arc_buf_hdr_t *, hdr);
ARCSTAT_BUMP(arcstat_l2_hits);
hdr->b_l2hdr.b_hits++;
cb = kmem_zalloc(sizeof (l2arc_read_callback_t),
KM_SLEEP);
cb->l2rcb_hdr = hdr;
cb->l2rcb_bp = *bp;
cb->l2rcb_zb = *zb;
cb->l2rcb_flags = zio_flags;
/*
* When Compressed ARC is disabled, but the
* L2ARC block is compressed, arc_hdr_size()
* will have returned LSIZE rather than PSIZE.
*/
if (HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr) &&
HDR_GET_PSIZE(hdr) != 0) {
size = HDR_GET_PSIZE(hdr);
}
asize = vdev_psize_to_asize(vd, size);
if (asize != size) {
abd = abd_alloc_for_io(asize,
HDR_ISTYPE_METADATA(hdr));
cb->l2rcb_abd = abd;
} else {
abd = hdr_abd;
}
ASSERT(addr >= VDEV_LABEL_START_SIZE &&
addr + asize <= vd->vdev_psize -
VDEV_LABEL_END_SIZE);
/*
* l2arc read. The SCL_L2ARC lock will be
* released by l2arc_read_done().
* Issue a null zio if the underlying buffer
* was squashed to zero size by compression.
*/
ASSERT3U(arc_hdr_get_compress(hdr), !=,
ZIO_COMPRESS_EMPTY);
rzio = zio_read_phys(pio, vd, addr,
asize, abd,
ZIO_CHECKSUM_OFF,
l2arc_read_done, cb, priority,
zio_flags | ZIO_FLAG_DONT_CACHE |
ZIO_FLAG_CANFAIL |
ZIO_FLAG_DONT_PROPAGATE |
ZIO_FLAG_DONT_RETRY, B_FALSE);
acb->acb_zio_head = rzio;
if (hash_lock != NULL)
mutex_exit(hash_lock);
DTRACE_PROBE2(l2arc__read, vdev_t *, vd,
zio_t *, rzio);
ARCSTAT_INCR(arcstat_l2_read_bytes,
HDR_GET_PSIZE(hdr));
if (*arc_flags & ARC_FLAG_NOWAIT) {
zio_nowait(rzio);
goto out;
}
ASSERT(*arc_flags & ARC_FLAG_WAIT);
if (zio_wait(rzio) == 0)
goto out;
/* l2arc read error; goto zio_read() */
if (hash_lock != NULL)
mutex_enter(hash_lock);
} else {
DTRACE_PROBE1(l2arc__miss,
arc_buf_hdr_t *, hdr);
ARCSTAT_BUMP(arcstat_l2_misses);
if (HDR_L2_WRITING(hdr))
ARCSTAT_BUMP(arcstat_l2_rw_clash);
spa_config_exit(spa, SCL_L2ARC, vd);
}
} else {
if (vd != NULL)
spa_config_exit(spa, SCL_L2ARC, vd);
/*
* Only a spa with l2 should contribute to l2
* miss stats. (Including the case of having a
* faulted cache device - that's also a miss.)
*/
if (spa_has_l2) {
/*
* Skip ARC stat bump for block pointers with
* embedded data. The data are read from the
* blkptr itself via
* decode_embedded_bp_compressed().
*/
if (!embedded_bp) {
DTRACE_PROBE1(l2arc__miss,
arc_buf_hdr_t *, hdr);
ARCSTAT_BUMP(arcstat_l2_misses);
}
}
}
rzio = zio_read(pio, spa, bp, hdr_abd, size,
arc_read_done, hdr, priority, zio_flags, zb);
acb->acb_zio_head = rzio;
if (hash_lock != NULL)
mutex_exit(hash_lock);
if (*arc_flags & ARC_FLAG_WAIT) {
rc = zio_wait(rzio);
goto out;
}
ASSERT(*arc_flags & ARC_FLAG_NOWAIT);
zio_nowait(rzio);
}
out:
/* embedded bps don't actually go to disk */
if (!embedded_bp)
spa_read_history_add(spa, zb, *arc_flags);
spl_fstrans_unmark(cookie);
return (rc);
done:
if (done)
done(NULL, zb, bp, buf, private);
if (pio && rc != 0) {
zio_t *zio = zio_null(pio, spa, NULL, NULL, NULL, zio_flags);
zio->io_error = rc;
zio_nowait(zio);
}
goto out;
}
arc_prune_t *
arc_add_prune_callback(arc_prune_func_t *func, void *private)
{
arc_prune_t *p;
p = kmem_alloc(sizeof (*p), KM_SLEEP);
p->p_pfunc = func;
p->p_private = private;
list_link_init(&p->p_node);
zfs_refcount_create(&p->p_refcnt);
mutex_enter(&arc_prune_mtx);
zfs_refcount_add(&p->p_refcnt, &arc_prune_list);
list_insert_head(&arc_prune_list, p);
mutex_exit(&arc_prune_mtx);
return (p);
}
void
arc_remove_prune_callback(arc_prune_t *p)
{
boolean_t wait = B_FALSE;
mutex_enter(&arc_prune_mtx);
list_remove(&arc_prune_list, p);
if (zfs_refcount_remove(&p->p_refcnt, &arc_prune_list) > 0)
wait = B_TRUE;
mutex_exit(&arc_prune_mtx);
/* wait for arc_prune_task to finish */
if (wait)
taskq_wait_outstanding(arc_prune_taskq, 0);
ASSERT0(zfs_refcount_count(&p->p_refcnt));
zfs_refcount_destroy(&p->p_refcnt);
kmem_free(p, sizeof (*p));
}
/*
* Notify the arc that a block was freed, and thus will never be used again.
*/
void
arc_freed(spa_t *spa, const blkptr_t *bp)
{
arc_buf_hdr_t *hdr;
kmutex_t *hash_lock;
uint64_t guid = spa_load_guid(spa);
ASSERT(!BP_IS_EMBEDDED(bp));
hdr = buf_hash_find(guid, bp, &hash_lock);
if (hdr == NULL)
return;
/*
* We might be trying to free a block that is still doing I/O
* (i.e. prefetch) or has some other reference (i.e. a dedup-ed,
* dmu_sync-ed block). A block may also have a reference if it is
* part of a dedup-ed, dmu_synced write. The dmu_sync() function would
* have written the new block to its final resting place on disk but
* without the dedup flag set. This would have left the hdr in the MRU
* state and discoverable. When the txg finally syncs it detects that
* the block was overridden in open context and issues an override I/O.
* Since this is a dedup block, the override I/O will determine if the
* block is already in the DDT. If so, then it will replace the io_bp
* with the bp from the DDT and allow the I/O to finish. When the I/O
* reaches the done callback, dbuf_write_override_done, it will
* check to see if the io_bp and io_bp_override are identical.
* If they are not, then it indicates that the bp was replaced with
* the bp in the DDT and the override bp is freed. This allows
* us to arrive here with a reference on a block that is being
* freed. So if we have an I/O in progress, or a reference to
* this hdr, then we don't destroy the hdr.
*/
if (!HDR_HAS_L1HDR(hdr) ||
zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt)) {
arc_change_state(arc_anon, hdr);
arc_hdr_destroy(hdr);
mutex_exit(hash_lock);
} else {
mutex_exit(hash_lock);
}
}
/*
* Release this buffer from the cache, making it an anonymous buffer. This
* must be done after a read and prior to modifying the buffer contents.
* If the buffer has more than one reference, we must make
* a new hdr for the buffer.
*/
void
arc_release(arc_buf_t *buf, const void *tag)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
/*
* It would be nice to assert that if its DMU metadata (level >
* 0 || it's the dnode file), then it must be syncing context.
* But we don't know that information at this level.
*/
ASSERT(HDR_HAS_L1HDR(hdr));
/*
* We don't grab the hash lock prior to this check, because if
* the buffer's header is in the arc_anon state, it won't be
* linked into the hash table.
*/
if (hdr->b_l1hdr.b_state == arc_anon) {
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT(!HDR_IN_HASH_TABLE(hdr));
ASSERT(!HDR_HAS_L2HDR(hdr));
ASSERT3U(hdr->b_l1hdr.b_bufcnt, ==, 1);
ASSERT3S(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt), ==, 1);
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
hdr->b_l1hdr.b_arc_access = 0;
/*
* If the buf is being overridden then it may already
* have a hdr that is not empty.
*/
buf_discard_identity(hdr);
arc_buf_thaw(buf);
return;
}
kmutex_t *hash_lock = HDR_LOCK(hdr);
mutex_enter(hash_lock);
/*
* This assignment is only valid as long as the hash_lock is
* held, we must be careful not to reference state or the
* b_state field after dropping the lock.
*/
arc_state_t *state = hdr->b_l1hdr.b_state;
ASSERT3P(hash_lock, ==, HDR_LOCK(hdr));
ASSERT3P(state, !=, arc_anon);
/* this buffer is not on any list */
ASSERT3S(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt), >, 0);
if (HDR_HAS_L2HDR(hdr)) {
mutex_enter(&hdr->b_l2hdr.b_dev->l2ad_mtx);
/*
* We have to recheck this conditional again now that
* we're holding the l2ad_mtx to prevent a race with
* another thread which might be concurrently calling
* l2arc_evict(). In that case, l2arc_evict() might have
* destroyed the header's L2 portion as we were waiting
* to acquire the l2ad_mtx.
*/
if (HDR_HAS_L2HDR(hdr))
arc_hdr_l2hdr_destroy(hdr);
mutex_exit(&hdr->b_l2hdr.b_dev->l2ad_mtx);
}
/*
* Do we have more than one buf?
*/
if (hdr->b_l1hdr.b_bufcnt > 1) {
arc_buf_hdr_t *nhdr;
uint64_t spa = hdr->b_spa;
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t lsize = HDR_GET_LSIZE(hdr);
boolean_t protected = HDR_PROTECTED(hdr);
enum zio_compress compress = arc_hdr_get_compress(hdr);
arc_buf_contents_t type = arc_buf_type(hdr);
VERIFY3U(hdr->b_type, ==, type);
ASSERT(hdr->b_l1hdr.b_buf != buf || buf->b_next != NULL);
VERIFY3S(remove_reference(hdr, tag), >, 0);
if (arc_buf_is_shared(buf) && !ARC_BUF_COMPRESSED(buf)) {
ASSERT3P(hdr->b_l1hdr.b_buf, !=, buf);
ASSERT(ARC_BUF_LAST(buf));
}
/*
* Pull the data off of this hdr and attach it to
* a new anonymous hdr. Also find the last buffer
* in the hdr's buffer list.
*/
arc_buf_t *lastbuf = arc_buf_remove(hdr, buf);
ASSERT3P(lastbuf, !=, NULL);
/*
* If the current arc_buf_t and the hdr are sharing their data
* buffer, then we must stop sharing that block.
*/
if (arc_buf_is_shared(buf)) {
ASSERT3P(hdr->b_l1hdr.b_buf, !=, buf);
VERIFY(!arc_buf_is_shared(lastbuf));
/*
* First, sever the block sharing relationship between
* buf and the arc_buf_hdr_t.
*/
arc_unshare_buf(hdr, buf);
/*
* Now we need to recreate the hdr's b_pabd. Since we
* have lastbuf handy, we try to share with it, but if
* we can't then we allocate a new b_pabd and copy the
* data from buf into it.
*/
if (arc_can_share(hdr, lastbuf)) {
arc_share_buf(hdr, lastbuf);
} else {
arc_hdr_alloc_abd(hdr, 0);
abd_copy_from_buf(hdr->b_l1hdr.b_pabd,
buf->b_data, psize);
}
VERIFY3P(lastbuf->b_data, !=, NULL);
} else if (HDR_SHARED_DATA(hdr)) {
/*
* Uncompressed shared buffers are always at the end
* of the list. Compressed buffers don't have the
* same requirements. This makes it hard to
* simply assert that the lastbuf is shared so
* we rely on the hdr's compression flags to determine
* if we have a compressed, shared buffer.
*/
ASSERT(arc_buf_is_shared(lastbuf) ||
arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF);
ASSERT(!ARC_BUF_SHARED(buf));
}
ASSERT(hdr->b_l1hdr.b_pabd != NULL || HDR_HAS_RABD(hdr));
ASSERT3P(state, !=, arc_l2c_only);
(void) zfs_refcount_remove_many(&state->arcs_size[type],
arc_buf_size(buf), buf);
if (zfs_refcount_is_zero(&hdr->b_l1hdr.b_refcnt)) {
ASSERT3P(state, !=, arc_l2c_only);
(void) zfs_refcount_remove_many(
&state->arcs_esize[type],
arc_buf_size(buf), buf);
}
hdr->b_l1hdr.b_bufcnt -= 1;
if (ARC_BUF_ENCRYPTED(buf))
hdr->b_crypt_hdr.b_ebufcnt -= 1;
arc_cksum_verify(buf);
arc_buf_unwatch(buf);
/* if this is the last uncompressed buf free the checksum */
if (!arc_hdr_has_uncompressed_buf(hdr))
arc_cksum_free(hdr);
mutex_exit(hash_lock);
nhdr = arc_hdr_alloc(spa, psize, lsize, protected,
compress, hdr->b_complevel, type);
ASSERT3P(nhdr->b_l1hdr.b_buf, ==, NULL);
ASSERT0(nhdr->b_l1hdr.b_bufcnt);
ASSERT0(zfs_refcount_count(&nhdr->b_l1hdr.b_refcnt));
VERIFY3U(nhdr->b_type, ==, type);
ASSERT(!HDR_SHARED_DATA(nhdr));
nhdr->b_l1hdr.b_buf = buf;
nhdr->b_l1hdr.b_bufcnt = 1;
if (ARC_BUF_ENCRYPTED(buf))
nhdr->b_crypt_hdr.b_ebufcnt = 1;
(void) zfs_refcount_add(&nhdr->b_l1hdr.b_refcnt, tag);
buf->b_hdr = nhdr;
(void) zfs_refcount_add_many(&arc_anon->arcs_size[type],
arc_buf_size(buf), buf);
} else {
ASSERT(zfs_refcount_count(&hdr->b_l1hdr.b_refcnt) == 1);
/* protected by hash lock, or hdr is on arc_anon */
ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
hdr->b_l1hdr.b_mru_hits = 0;
hdr->b_l1hdr.b_mru_ghost_hits = 0;
hdr->b_l1hdr.b_mfu_hits = 0;
hdr->b_l1hdr.b_mfu_ghost_hits = 0;
arc_change_state(arc_anon, hdr);
hdr->b_l1hdr.b_arc_access = 0;
mutex_exit(hash_lock);
buf_discard_identity(hdr);
arc_buf_thaw(buf);
}
}
int
arc_released(arc_buf_t *buf)
{
return (buf->b_data != NULL &&
buf->b_hdr->b_l1hdr.b_state == arc_anon);
}
#ifdef ZFS_DEBUG
int
arc_referenced(arc_buf_t *buf)
{
return (zfs_refcount_count(&buf->b_hdr->b_l1hdr.b_refcnt));
}
#endif
static void
arc_write_ready(zio_t *zio)
{
arc_write_callback_t *callback = zio->io_private;
arc_buf_t *buf = callback->awcb_buf;
arc_buf_hdr_t *hdr = buf->b_hdr;
blkptr_t *bp = zio->io_bp;
uint64_t psize = BP_IS_HOLE(bp) ? 0 : BP_GET_PSIZE(bp);
fstrans_cookie_t cookie = spl_fstrans_mark();
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT(!zfs_refcount_is_zero(&buf->b_hdr->b_l1hdr.b_refcnt));
ASSERT(hdr->b_l1hdr.b_bufcnt > 0);
/*
* If we're reexecuting this zio because the pool suspended, then
* cleanup any state that was previously set the first time the
* callback was invoked.
*/
if (zio->io_flags & ZIO_FLAG_REEXECUTED) {
arc_cksum_free(hdr);
arc_buf_unwatch(buf);
if (hdr->b_l1hdr.b_pabd != NULL) {
if (arc_buf_is_shared(buf)) {
arc_unshare_buf(hdr, buf);
} else {
arc_hdr_free_abd(hdr, B_FALSE);
}
}
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
}
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
ASSERT(!HDR_HAS_RABD(hdr));
ASSERT(!HDR_SHARED_DATA(hdr));
ASSERT(!arc_buf_is_shared(buf));
callback->awcb_ready(zio, buf, callback->awcb_private);
if (HDR_IO_IN_PROGRESS(hdr)) {
ASSERT(zio->io_flags & ZIO_FLAG_REEXECUTED);
} else {
arc_hdr_set_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
add_reference(hdr, hdr); /* For IO_IN_PROGRESS. */
}
if (BP_IS_PROTECTED(bp) != !!HDR_PROTECTED(hdr))
hdr = arc_hdr_realloc_crypt(hdr, BP_IS_PROTECTED(bp));
if (BP_IS_PROTECTED(bp)) {
/* ZIL blocks are written through zio_rewrite */
ASSERT3U(BP_GET_TYPE(bp), !=, DMU_OT_INTENT_LOG);
ASSERT(HDR_PROTECTED(hdr));
if (BP_SHOULD_BYTESWAP(bp)) {
if (BP_GET_LEVEL(bp) > 0) {
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_UINT64;
} else {
hdr->b_l1hdr.b_byteswap =
DMU_OT_BYTESWAP(BP_GET_TYPE(bp));
}
} else {
hdr->b_l1hdr.b_byteswap = DMU_BSWAP_NUMFUNCS;
}
hdr->b_crypt_hdr.b_ot = BP_GET_TYPE(bp);
hdr->b_crypt_hdr.b_dsobj = zio->io_bookmark.zb_objset;
zio_crypt_decode_params_bp(bp, hdr->b_crypt_hdr.b_salt,
hdr->b_crypt_hdr.b_iv);
zio_crypt_decode_mac_bp(bp, hdr->b_crypt_hdr.b_mac);
}
/*
* If this block was written for raw encryption but the zio layer
* ended up only authenticating it, adjust the buffer flags now.
*/
if (BP_IS_AUTHENTICATED(bp) && ARC_BUF_ENCRYPTED(buf)) {
arc_hdr_set_flags(hdr, ARC_FLAG_NOAUTH);
buf->b_flags &= ~ARC_BUF_FLAG_ENCRYPTED;
if (BP_GET_COMPRESS(bp) == ZIO_COMPRESS_OFF)
buf->b_flags &= ~ARC_BUF_FLAG_COMPRESSED;
} else if (BP_IS_HOLE(bp) && ARC_BUF_ENCRYPTED(buf)) {
buf->b_flags &= ~ARC_BUF_FLAG_ENCRYPTED;
buf->b_flags &= ~ARC_BUF_FLAG_COMPRESSED;
}
/* this must be done after the buffer flags are adjusted */
arc_cksum_compute(buf);
enum zio_compress compress;
if (BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp)) {
compress = ZIO_COMPRESS_OFF;
} else {
ASSERT3U(HDR_GET_LSIZE(hdr), ==, BP_GET_LSIZE(bp));
compress = BP_GET_COMPRESS(bp);
}
HDR_SET_PSIZE(hdr, psize);
arc_hdr_set_compress(hdr, compress);
hdr->b_complevel = zio->io_prop.zp_complevel;
if (zio->io_error != 0 || psize == 0)
goto out;
/*
* Fill the hdr with data. If the buffer is encrypted we have no choice
* but to copy the data into b_radb. If the hdr is compressed, the data
* we want is available from the zio, otherwise we can take it from
* the buf.
*
* We might be able to share the buf's data with the hdr here. However,
* doing so would cause the ARC to be full of linear ABDs if we write a
* lot of shareable data. As a compromise, we check whether scattered
* ABDs are allowed, and assume that if they are then the user wants
* the ARC to be primarily filled with them regardless of the data being
* written. Therefore, if they're allowed then we allocate one and copy
* the data into it; otherwise, we share the data directly if we can.
*/
if (ARC_BUF_ENCRYPTED(buf)) {
ASSERT3U(psize, >, 0);
ASSERT(ARC_BUF_COMPRESSED(buf));
arc_hdr_alloc_abd(hdr, ARC_HDR_ALLOC_RDATA |
ARC_HDR_USE_RESERVE);
abd_copy(hdr->b_crypt_hdr.b_rabd, zio->io_abd, psize);
} else if (!(HDR_UNCACHED(hdr) ||
abd_size_alloc_linear(arc_buf_size(buf))) ||
!arc_can_share(hdr, buf)) {
/*
* Ideally, we would always copy the io_abd into b_pabd, but the
* user may have disabled compressed ARC, thus we must check the
* hdr's compression setting rather than the io_bp's.
*/
if (BP_IS_ENCRYPTED(bp)) {
ASSERT3U(psize, >, 0);
arc_hdr_alloc_abd(hdr, ARC_HDR_ALLOC_RDATA |
ARC_HDR_USE_RESERVE);
abd_copy(hdr->b_crypt_hdr.b_rabd, zio->io_abd, psize);
} else if (arc_hdr_get_compress(hdr) != ZIO_COMPRESS_OFF &&
!ARC_BUF_COMPRESSED(buf)) {
ASSERT3U(psize, >, 0);
arc_hdr_alloc_abd(hdr, ARC_HDR_USE_RESERVE);
abd_copy(hdr->b_l1hdr.b_pabd, zio->io_abd, psize);
} else {
ASSERT3U(zio->io_orig_size, ==, arc_hdr_size(hdr));
arc_hdr_alloc_abd(hdr, ARC_HDR_USE_RESERVE);
abd_copy_from_buf(hdr->b_l1hdr.b_pabd, buf->b_data,
arc_buf_size(buf));
}
} else {
ASSERT3P(buf->b_data, ==, abd_to_buf(zio->io_orig_abd));
ASSERT3U(zio->io_orig_size, ==, arc_buf_size(buf));
ASSERT3U(hdr->b_l1hdr.b_bufcnt, ==, 1);
arc_share_buf(hdr, buf);
}
out:
arc_hdr_verify(hdr, bp);
spl_fstrans_unmark(cookie);
}
static void
arc_write_children_ready(zio_t *zio)
{
arc_write_callback_t *callback = zio->io_private;
arc_buf_t *buf = callback->awcb_buf;
callback->awcb_children_ready(zio, buf, callback->awcb_private);
}
/*
* The SPA calls this callback for each physical write that happens on behalf
* of a logical write. See the comment in dbuf_write_physdone() for details.
*/
static void
arc_write_physdone(zio_t *zio)
{
arc_write_callback_t *cb = zio->io_private;
if (cb->awcb_physdone != NULL)
cb->awcb_physdone(zio, cb->awcb_buf, cb->awcb_private);
}
static void
arc_write_done(zio_t *zio)
{
arc_write_callback_t *callback = zio->io_private;
arc_buf_t *buf = callback->awcb_buf;
arc_buf_hdr_t *hdr = buf->b_hdr;
ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
if (zio->io_error == 0) {
arc_hdr_verify(hdr, zio->io_bp);
if (BP_IS_HOLE(zio->io_bp) || BP_IS_EMBEDDED(zio->io_bp)) {
buf_discard_identity(hdr);
} else {
hdr->b_dva = *BP_IDENTITY(zio->io_bp);
hdr->b_birth = BP_PHYSICAL_BIRTH(zio->io_bp);
}
} else {
ASSERT(HDR_EMPTY(hdr));
}
/*
* If the block to be written was all-zero or compressed enough to be
* embedded in the BP, no write was performed so there will be no
* dva/birth/checksum. The buffer must therefore remain anonymous
* (and uncached).
*/
if (!HDR_EMPTY(hdr)) {
arc_buf_hdr_t *exists;
kmutex_t *hash_lock;
ASSERT3U(zio->io_error, ==, 0);
arc_cksum_verify(buf);
exists = buf_hash_insert(hdr, &hash_lock);
if (exists != NULL) {
/*
* This can only happen if we overwrite for
* sync-to-convergence, because we remove
* buffers from the hash table when we arc_free().
*/
if (zio->io_flags & ZIO_FLAG_IO_REWRITE) {
if (!BP_EQUAL(&zio->io_bp_orig, zio->io_bp))
panic("bad overwrite, hdr=%p exists=%p",
(void *)hdr, (void *)exists);
ASSERT(zfs_refcount_is_zero(
&exists->b_l1hdr.b_refcnt));
arc_change_state(arc_anon, exists);
arc_hdr_destroy(exists);
mutex_exit(hash_lock);
exists = buf_hash_insert(hdr, &hash_lock);
ASSERT3P(exists, ==, NULL);
} else if (zio->io_flags & ZIO_FLAG_NOPWRITE) {
/* nopwrite */
ASSERT(zio->io_prop.zp_nopwrite);
if (!BP_EQUAL(&zio->io_bp_orig, zio->io_bp))
panic("bad nopwrite, hdr=%p exists=%p",
(void *)hdr, (void *)exists);
} else {
/* Dedup */
ASSERT(hdr->b_l1hdr.b_bufcnt == 1);
ASSERT(hdr->b_l1hdr.b_state == arc_anon);
ASSERT(BP_GET_DEDUP(zio->io_bp));
ASSERT(BP_GET_LEVEL(zio->io_bp) == 0);
}
}
arc_hdr_clear_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
VERIFY3S(remove_reference(hdr, hdr), >, 0);
/* if it's not anon, we are doing a scrub */
if (exists == NULL && hdr->b_l1hdr.b_state == arc_anon)
arc_access(hdr, 0, B_FALSE);
mutex_exit(hash_lock);
} else {
arc_hdr_clear_flags(hdr, ARC_FLAG_IO_IN_PROGRESS);
VERIFY3S(remove_reference(hdr, hdr), >, 0);
}
callback->awcb_done(zio, buf, callback->awcb_private);
abd_free(zio->io_abd);
kmem_free(callback, sizeof (arc_write_callback_t));
}
zio_t *
arc_write(zio_t *pio, spa_t *spa, uint64_t txg,
blkptr_t *bp, arc_buf_t *buf, boolean_t uncached, boolean_t l2arc,
const zio_prop_t *zp, arc_write_done_func_t *ready,
arc_write_done_func_t *children_ready, arc_write_done_func_t *physdone,
arc_write_done_func_t *done, void *private, zio_priority_t priority,
int zio_flags, const zbookmark_phys_t *zb)
{
arc_buf_hdr_t *hdr = buf->b_hdr;
arc_write_callback_t *callback;
zio_t *zio;
zio_prop_t localprop = *zp;
ASSERT3P(ready, !=, NULL);
ASSERT3P(done, !=, NULL);
ASSERT(!HDR_IO_ERROR(hdr));
ASSERT(!HDR_IO_IN_PROGRESS(hdr));
ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL);
ASSERT3U(hdr->b_l1hdr.b_bufcnt, >, 0);
if (uncached)
arc_hdr_set_flags(hdr, ARC_FLAG_UNCACHED);
else if (l2arc)
arc_hdr_set_flags(hdr, ARC_FLAG_L2CACHE);
if (ARC_BUF_ENCRYPTED(buf)) {
ASSERT(ARC_BUF_COMPRESSED(buf));
localprop.zp_encrypt = B_TRUE;
localprop.zp_compress = HDR_GET_COMPRESS(hdr);
localprop.zp_complevel = hdr->b_complevel;
localprop.zp_byteorder =
(hdr->b_l1hdr.b_byteswap == DMU_BSWAP_NUMFUNCS) ?
ZFS_HOST_BYTEORDER : !ZFS_HOST_BYTEORDER;
memcpy(localprop.zp_salt, hdr->b_crypt_hdr.b_salt,
ZIO_DATA_SALT_LEN);
memcpy(localprop.zp_iv, hdr->b_crypt_hdr.b_iv,
ZIO_DATA_IV_LEN);
memcpy(localprop.zp_mac, hdr->b_crypt_hdr.b_mac,
ZIO_DATA_MAC_LEN);
if (DMU_OT_IS_ENCRYPTED(localprop.zp_type)) {
localprop.zp_nopwrite = B_FALSE;
localprop.zp_copies =
MIN(localprop.zp_copies, SPA_DVAS_PER_BP - 1);
}
zio_flags |= ZIO_FLAG_RAW;
} else if (ARC_BUF_COMPRESSED(buf)) {
ASSERT3U(HDR_GET_LSIZE(hdr), !=, arc_buf_size(buf));
localprop.zp_compress = HDR_GET_COMPRESS(hdr);
localprop.zp_complevel = hdr->b_complevel;
zio_flags |= ZIO_FLAG_RAW_COMPRESS;
}
callback = kmem_zalloc(sizeof (arc_write_callback_t), KM_SLEEP);
callback->awcb_ready = ready;
callback->awcb_children_ready = children_ready;
callback->awcb_physdone = physdone;
callback->awcb_done = done;
callback->awcb_private = private;
callback->awcb_buf = buf;
/*
* The hdr's b_pabd is now stale, free it now. A new data block
* will be allocated when the zio pipeline calls arc_write_ready().
*/
if (hdr->b_l1hdr.b_pabd != NULL) {
/*
* If the buf is currently sharing the data block with
* the hdr then we need to break that relationship here.
* The hdr will remain with a NULL data pointer and the
* buf will take sole ownership of the block.
*/
if (arc_buf_is_shared(buf)) {
arc_unshare_buf(hdr, buf);
} else {
arc_hdr_free_abd(hdr, B_FALSE);
}
VERIFY3P(buf->b_data, !=, NULL);
}
if (HDR_HAS_RABD(hdr))
arc_hdr_free_abd(hdr, B_TRUE);
if (!(zio_flags & ZIO_FLAG_RAW))
arc_hdr_set_compress(hdr, ZIO_COMPRESS_OFF);
ASSERT(!arc_buf_is_shared(buf));
ASSERT3P(hdr->b_l1hdr.b_pabd, ==, NULL);
zio = zio_write(pio, spa, txg, bp,
abd_get_from_buf(buf->b_data, HDR_GET_LSIZE(hdr)),
HDR_GET_LSIZE(hdr), arc_buf_size(buf), &localprop, arc_write_ready,
(children_ready != NULL) ? arc_write_children_ready : NULL,
arc_write_physdone, arc_write_done, callback,
priority, zio_flags, zb);
return (zio);
}
void
arc_tempreserve_clear(uint64_t reserve)
{
atomic_add_64(&arc_tempreserve, -reserve);
ASSERT((int64_t)arc_tempreserve >= 0);
}
int
arc_tempreserve_space(spa_t *spa, uint64_t reserve, uint64_t txg)
{
int error;
uint64_t anon_size;
if (!arc_no_grow &&
reserve > arc_c/4 &&
reserve * 4 > (2ULL << SPA_MAXBLOCKSHIFT))
arc_c = MIN(arc_c_max, reserve * 4);
/*
* Throttle when the calculated memory footprint for the TXG
* exceeds the target ARC size.
*/
if (reserve > arc_c) {
DMU_TX_STAT_BUMP(dmu_tx_memory_reserve);
return (SET_ERROR(ERESTART));
}
/*
* Don't count loaned bufs as in flight dirty data to prevent long
* network delays from blocking transactions that are ready to be
* assigned to a txg.
*/
/* assert that it has not wrapped around */
ASSERT3S(atomic_add_64_nv(&arc_loaned_bytes, 0), >=, 0);
anon_size = MAX((int64_t)
(zfs_refcount_count(&arc_anon->arcs_size[ARC_BUFC_DATA]) +
zfs_refcount_count(&arc_anon->arcs_size[ARC_BUFC_METADATA]) -
arc_loaned_bytes), 0);
/*
* Writes will, almost always, require additional memory allocations
* in order to compress/encrypt/etc the data. We therefore need to
* make sure that there is sufficient available memory for this.
*/
error = arc_memory_throttle(spa, reserve, txg);
if (error != 0)
return (error);
/*
* Throttle writes when the amount of dirty data in the cache
* gets too large. We try to keep the cache less than half full
* of dirty blocks so that our sync times don't grow too large.
*
* In the case of one pool being built on another pool, we want
* to make sure we don't end up throttling the lower (backing)
* pool when the upper pool is the majority contributor to dirty
* data. To insure we make forward progress during throttling, we
* also check the current pool's net dirty data and only throttle
* if it exceeds zfs_arc_pool_dirty_percent of the anonymous dirty
* data in the cache.
*
* Note: if two requests come in concurrently, we might let them
* both succeed, when one of them should fail. Not a huge deal.
*/
uint64_t total_dirty = reserve + arc_tempreserve + anon_size;
uint64_t spa_dirty_anon = spa_dirty_data(spa);
uint64_t rarc_c = arc_warm ? arc_c : arc_c_max;
if (total_dirty > rarc_c * zfs_arc_dirty_limit_percent / 100 &&
anon_size > rarc_c * zfs_arc_anon_limit_percent / 100 &&
spa_dirty_anon > anon_size * zfs_arc_pool_dirty_percent / 100) {
#ifdef ZFS_DEBUG
uint64_t meta_esize = zfs_refcount_count(
&arc_anon->arcs_esize[ARC_BUFC_METADATA]);
uint64_t data_esize =
zfs_refcount_count(&arc_anon->arcs_esize[ARC_BUFC_DATA]);
dprintf("failing, arc_tempreserve=%lluK anon_meta=%lluK "
"anon_data=%lluK tempreserve=%lluK rarc_c=%lluK\n",
(u_longlong_t)arc_tempreserve >> 10,
(u_longlong_t)meta_esize >> 10,
(u_longlong_t)data_esize >> 10,
(u_longlong_t)reserve >> 10,
(u_longlong_t)rarc_c >> 10);
#endif
DMU_TX_STAT_BUMP(dmu_tx_dirty_throttle);
return (SET_ERROR(ERESTART));
}
atomic_add_64(&arc_tempreserve, reserve);
return (0);
}
static void
arc_kstat_update_state(arc_state_t *state, kstat_named_t *size,
kstat_named_t *data, kstat_named_t *metadata,
kstat_named_t *evict_data, kstat_named_t *evict_metadata)
{
data->value.ui64 =
zfs_refcount_count(&state->arcs_size[ARC_BUFC_DATA]);
metadata->value.ui64 =
zfs_refcount_count(&state->arcs_size[ARC_BUFC_METADATA]);
size->value.ui64 = data->value.ui64 + metadata->value.ui64;
evict_data->value.ui64 =
zfs_refcount_count(&state->arcs_esize[ARC_BUFC_DATA]);
evict_metadata->value.ui64 =
zfs_refcount_count(&state->arcs_esize[ARC_BUFC_METADATA]);
}
static int
arc_kstat_update(kstat_t *ksp, int rw)
{
arc_stats_t *as = ksp->ks_data;
if (rw == KSTAT_WRITE)
return (SET_ERROR(EACCES));
as->arcstat_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_hits);
as->arcstat_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_iohits);
as->arcstat_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_misses);
as->arcstat_demand_data_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_data_hits);
as->arcstat_demand_data_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_data_iohits);
as->arcstat_demand_data_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_data_misses);
as->arcstat_demand_metadata_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_metadata_hits);
as->arcstat_demand_metadata_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_metadata_iohits);
as->arcstat_demand_metadata_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_metadata_misses);
as->arcstat_prefetch_data_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_data_hits);
as->arcstat_prefetch_data_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_data_iohits);
as->arcstat_prefetch_data_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_data_misses);
as->arcstat_prefetch_metadata_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_metadata_hits);
as->arcstat_prefetch_metadata_iohits.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_metadata_iohits);
as->arcstat_prefetch_metadata_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_prefetch_metadata_misses);
as->arcstat_mru_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_mru_hits);
as->arcstat_mru_ghost_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_mru_ghost_hits);
as->arcstat_mfu_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_mfu_hits);
as->arcstat_mfu_ghost_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_mfu_ghost_hits);
as->arcstat_uncached_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_uncached_hits);
as->arcstat_deleted.value.ui64 =
wmsum_value(&arc_sums.arcstat_deleted);
as->arcstat_mutex_miss.value.ui64 =
wmsum_value(&arc_sums.arcstat_mutex_miss);
as->arcstat_access_skip.value.ui64 =
wmsum_value(&arc_sums.arcstat_access_skip);
as->arcstat_evict_skip.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_skip);
as->arcstat_evict_not_enough.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_not_enough);
as->arcstat_evict_l2_cached.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_cached);
as->arcstat_evict_l2_eligible.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_eligible);
as->arcstat_evict_l2_eligible_mfu.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_eligible_mfu);
as->arcstat_evict_l2_eligible_mru.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_eligible_mru);
as->arcstat_evict_l2_ineligible.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_ineligible);
as->arcstat_evict_l2_skip.value.ui64 =
wmsum_value(&arc_sums.arcstat_evict_l2_skip);
as->arcstat_hash_collisions.value.ui64 =
wmsum_value(&arc_sums.arcstat_hash_collisions);
as->arcstat_hash_chains.value.ui64 =
wmsum_value(&arc_sums.arcstat_hash_chains);
as->arcstat_size.value.ui64 =
aggsum_value(&arc_sums.arcstat_size);
as->arcstat_compressed_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_compressed_size);
as->arcstat_uncompressed_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_uncompressed_size);
as->arcstat_overhead_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_overhead_size);
as->arcstat_hdr_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_hdr_size);
as->arcstat_data_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_data_size);
as->arcstat_metadata_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_metadata_size);
as->arcstat_dbuf_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_dbuf_size);
#if defined(COMPAT_FREEBSD11)
as->arcstat_other_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_bonus_size) +
wmsum_value(&arc_sums.arcstat_dnode_size) +
wmsum_value(&arc_sums.arcstat_dbuf_size);
#endif
arc_kstat_update_state(arc_anon,
&as->arcstat_anon_size,
&as->arcstat_anon_data,
&as->arcstat_anon_metadata,
&as->arcstat_anon_evictable_data,
&as->arcstat_anon_evictable_metadata);
arc_kstat_update_state(arc_mru,
&as->arcstat_mru_size,
&as->arcstat_mru_data,
&as->arcstat_mru_metadata,
&as->arcstat_mru_evictable_data,
&as->arcstat_mru_evictable_metadata);
arc_kstat_update_state(arc_mru_ghost,
&as->arcstat_mru_ghost_size,
&as->arcstat_mru_ghost_data,
&as->arcstat_mru_ghost_metadata,
&as->arcstat_mru_ghost_evictable_data,
&as->arcstat_mru_ghost_evictable_metadata);
arc_kstat_update_state(arc_mfu,
&as->arcstat_mfu_size,
&as->arcstat_mfu_data,
&as->arcstat_mfu_metadata,
&as->arcstat_mfu_evictable_data,
&as->arcstat_mfu_evictable_metadata);
arc_kstat_update_state(arc_mfu_ghost,
&as->arcstat_mfu_ghost_size,
&as->arcstat_mfu_ghost_data,
&as->arcstat_mfu_ghost_metadata,
&as->arcstat_mfu_ghost_evictable_data,
&as->arcstat_mfu_ghost_evictable_metadata);
arc_kstat_update_state(arc_uncached,
&as->arcstat_uncached_size,
&as->arcstat_uncached_data,
&as->arcstat_uncached_metadata,
&as->arcstat_uncached_evictable_data,
&as->arcstat_uncached_evictable_metadata);
as->arcstat_dnode_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_dnode_size);
as->arcstat_bonus_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_bonus_size);
as->arcstat_l2_hits.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_hits);
as->arcstat_l2_misses.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_misses);
as->arcstat_l2_prefetch_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_prefetch_asize);
as->arcstat_l2_mru_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_mru_asize);
as->arcstat_l2_mfu_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_mfu_asize);
as->arcstat_l2_bufc_data_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_bufc_data_asize);
as->arcstat_l2_bufc_metadata_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_bufc_metadata_asize);
as->arcstat_l2_feeds.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_feeds);
as->arcstat_l2_rw_clash.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rw_clash);
as->arcstat_l2_read_bytes.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_read_bytes);
as->arcstat_l2_write_bytes.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_write_bytes);
as->arcstat_l2_writes_sent.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_writes_sent);
as->arcstat_l2_writes_done.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_writes_done);
as->arcstat_l2_writes_error.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_writes_error);
as->arcstat_l2_writes_lock_retry.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_writes_lock_retry);
as->arcstat_l2_evict_lock_retry.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_evict_lock_retry);
as->arcstat_l2_evict_reading.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_evict_reading);
as->arcstat_l2_evict_l1cached.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_evict_l1cached);
as->arcstat_l2_free_on_write.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_free_on_write);
as->arcstat_l2_abort_lowmem.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_abort_lowmem);
as->arcstat_l2_cksum_bad.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_cksum_bad);
as->arcstat_l2_io_error.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_io_error);
as->arcstat_l2_lsize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_lsize);
as->arcstat_l2_psize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_psize);
as->arcstat_l2_hdr_size.value.ui64 =
aggsum_value(&arc_sums.arcstat_l2_hdr_size);
as->arcstat_l2_log_blk_writes.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_log_blk_writes);
as->arcstat_l2_log_blk_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_log_blk_asize);
as->arcstat_l2_log_blk_count.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_log_blk_count);
as->arcstat_l2_rebuild_success.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_success);
as->arcstat_l2_rebuild_abort_unsupported.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_unsupported);
as->arcstat_l2_rebuild_abort_io_errors.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_io_errors);
as->arcstat_l2_rebuild_abort_dh_errors.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_dh_errors);
as->arcstat_l2_rebuild_abort_cksum_lb_errors.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_cksum_lb_errors);
as->arcstat_l2_rebuild_abort_lowmem.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_abort_lowmem);
as->arcstat_l2_rebuild_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_size);
as->arcstat_l2_rebuild_asize.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_asize);
as->arcstat_l2_rebuild_bufs.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_bufs);
as->arcstat_l2_rebuild_bufs_precached.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_bufs_precached);
as->arcstat_l2_rebuild_log_blks.value.ui64 =
wmsum_value(&arc_sums.arcstat_l2_rebuild_log_blks);
as->arcstat_memory_throttle_count.value.ui64 =
wmsum_value(&arc_sums.arcstat_memory_throttle_count);
as->arcstat_memory_direct_count.value.ui64 =
wmsum_value(&arc_sums.arcstat_memory_direct_count);
as->arcstat_memory_indirect_count.value.ui64 =
wmsum_value(&arc_sums.arcstat_memory_indirect_count);
as->arcstat_memory_all_bytes.value.ui64 =
arc_all_memory();
as->arcstat_memory_free_bytes.value.ui64 =
arc_free_memory();
as->arcstat_memory_available_bytes.value.i64 =
arc_available_memory();
as->arcstat_prune.value.ui64 =
wmsum_value(&arc_sums.arcstat_prune);
as->arcstat_meta_used.value.ui64 =
wmsum_value(&arc_sums.arcstat_meta_used);
as->arcstat_async_upgrade_sync.value.ui64 =
wmsum_value(&arc_sums.arcstat_async_upgrade_sync);
as->arcstat_predictive_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_predictive_prefetch);
as->arcstat_demand_hit_predictive_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_hit_predictive_prefetch);
as->arcstat_demand_iohit_predictive_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_iohit_predictive_prefetch);
as->arcstat_prescient_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_prescient_prefetch);
as->arcstat_demand_hit_prescient_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_hit_prescient_prefetch);
as->arcstat_demand_iohit_prescient_prefetch.value.ui64 =
wmsum_value(&arc_sums.arcstat_demand_iohit_prescient_prefetch);
as->arcstat_raw_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_raw_size);
as->arcstat_cached_only_in_progress.value.ui64 =
wmsum_value(&arc_sums.arcstat_cached_only_in_progress);
as->arcstat_abd_chunk_waste_size.value.ui64 =
wmsum_value(&arc_sums.arcstat_abd_chunk_waste_size);
return (0);
}
/*
* This function *must* return indices evenly distributed between all
* sublists of the multilist. This is needed due to how the ARC eviction
* code is laid out; arc_evict_state() assumes ARC buffers are evenly
* distributed between all sublists and uses this assumption when
* deciding which sublist to evict from and how much to evict from it.
*/
static unsigned int
arc_state_multilist_index_func(multilist_t *ml, void *obj)
{
arc_buf_hdr_t *hdr = obj;
/*
* We rely on b_dva to generate evenly distributed index
* numbers using buf_hash below. So, as an added precaution,
* let's make sure we never add empty buffers to the arc lists.
*/
ASSERT(!HDR_EMPTY(hdr));
/*
* The assumption here, is the hash value for a given
* arc_buf_hdr_t will remain constant throughout its lifetime
* (i.e. its b_spa, b_dva, and b_birth fields don't change).
* Thus, we don't need to store the header's sublist index
* on insertion, as this index can be recalculated on removal.
*
* Also, the low order bits of the hash value are thought to be
* distributed evenly. Otherwise, in the case that the multilist
* has a power of two number of sublists, each sublists' usage
* would not be evenly distributed. In this context full 64bit
* division would be a waste of time, so limit it to 32 bits.
*/
return ((unsigned int)buf_hash(hdr->b_spa, &hdr->b_dva, hdr->b_birth) %
multilist_get_num_sublists(ml));
}
static unsigned int
arc_state_l2c_multilist_index_func(multilist_t *ml, void *obj)
{
panic("Header %p insert into arc_l2c_only %p", obj, ml);
}
#define WARN_IF_TUNING_IGNORED(tuning, value, do_warn) do { \
if ((do_warn) && (tuning) && ((tuning) != (value))) { \
cmn_err(CE_WARN, \
"ignoring tunable %s (using %llu instead)", \
(#tuning), (u_longlong_t)(value)); \
} \
} while (0)
/*
* Called during module initialization and periodically thereafter to
* apply reasonable changes to the exposed performance tunings. Can also be
* called explicitly by param_set_arc_*() functions when ARC tunables are
* updated manually. Non-zero zfs_* values which differ from the currently set
* values will be applied.
*/
void
arc_tuning_update(boolean_t verbose)
{
uint64_t allmem = arc_all_memory();
/* Valid range: 32M - <arc_c_max> */
if ((zfs_arc_min) && (zfs_arc_min != arc_c_min) &&
(zfs_arc_min >= 2ULL << SPA_MAXBLOCKSHIFT) &&
(zfs_arc_min <= arc_c_max)) {
arc_c_min = zfs_arc_min;
arc_c = MAX(arc_c, arc_c_min);
}
WARN_IF_TUNING_IGNORED(zfs_arc_min, arc_c_min, verbose);
/* Valid range: 64M - <all physical memory> */
if ((zfs_arc_max) && (zfs_arc_max != arc_c_max) &&
(zfs_arc_max >= MIN_ARC_MAX) && (zfs_arc_max < allmem) &&
(zfs_arc_max > arc_c_min)) {
arc_c_max = zfs_arc_max;
arc_c = MIN(arc_c, arc_c_max);
if (arc_dnode_limit > arc_c_max)
arc_dnode_limit = arc_c_max;
}
WARN_IF_TUNING_IGNORED(zfs_arc_max, arc_c_max, verbose);
/* Valid range: 0 - <all physical memory> */
arc_dnode_limit = zfs_arc_dnode_limit ? zfs_arc_dnode_limit :
MIN(zfs_arc_dnode_limit_percent, 100) * arc_c_max / 100;
WARN_IF_TUNING_IGNORED(zfs_arc_dnode_limit, arc_dnode_limit, verbose);
/* Valid range: 1 - N */
if (zfs_arc_grow_retry)
arc_grow_retry = zfs_arc_grow_retry;
/* Valid range: 1 - N */
if (zfs_arc_shrink_shift) {
arc_shrink_shift = zfs_arc_shrink_shift;
arc_no_grow_shift = MIN(arc_no_grow_shift, arc_shrink_shift -1);
}
/* Valid range: 1 - N ms */
if (zfs_arc_min_prefetch_ms)
arc_min_prefetch_ms = zfs_arc_min_prefetch_ms;
/* Valid range: 1 - N ms */
if (zfs_arc_min_prescient_prefetch_ms) {
arc_min_prescient_prefetch_ms =
zfs_arc_min_prescient_prefetch_ms;
}
/* Valid range: 0 - 100 */
if (zfs_arc_lotsfree_percent <= 100)
arc_lotsfree_percent = zfs_arc_lotsfree_percent;
WARN_IF_TUNING_IGNORED(zfs_arc_lotsfree_percent, arc_lotsfree_percent,
verbose);
/* Valid range: 0 - <all physical memory> */
if ((zfs_arc_sys_free) && (zfs_arc_sys_free != arc_sys_free))
arc_sys_free = MIN(zfs_arc_sys_free, allmem);
WARN_IF_TUNING_IGNORED(zfs_arc_sys_free, arc_sys_free, verbose);
}
static void
arc_state_multilist_init(multilist_t *ml,
multilist_sublist_index_func_t *index_func, int *maxcountp)
{
multilist_create(ml, sizeof (arc_buf_hdr_t),
offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), index_func);
*maxcountp = MAX(*maxcountp, multilist_get_num_sublists(ml));
}
static void
arc_state_init(void)
{
int num_sublists = 0;
arc_state_multilist_init(&arc_mru->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mru->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mru_ghost->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mru_ghost->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mfu->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mfu->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mfu_ghost->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_mfu_ghost->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_uncached->arcs_list[ARC_BUFC_METADATA],
arc_state_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_uncached->arcs_list[ARC_BUFC_DATA],
arc_state_multilist_index_func, &num_sublists);
/*
* L2 headers should never be on the L2 state list since they don't
* have L1 headers allocated. Special index function asserts that.
*/
arc_state_multilist_init(&arc_l2c_only->arcs_list[ARC_BUFC_METADATA],
arc_state_l2c_multilist_index_func, &num_sublists);
arc_state_multilist_init(&arc_l2c_only->arcs_list[ARC_BUFC_DATA],
arc_state_l2c_multilist_index_func, &num_sublists);
/*
* Keep track of the number of markers needed to reclaim buffers from
* any ARC state. The markers will be pre-allocated so as to minimize
* the number of memory allocations performed by the eviction thread.
*/
arc_state_evict_marker_count = num_sublists;
zfs_refcount_create(&arc_anon->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_anon->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mru->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mru->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mru_ghost->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mru_ghost->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mfu->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mfu->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mfu_ghost->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mfu_ghost->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_l2c_only->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_l2c_only->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_uncached->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_uncached->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_anon->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_anon->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mru->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mru->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mru_ghost->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mru_ghost->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mfu->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mfu->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_mfu_ghost->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_mfu_ghost->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_l2c_only->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_l2c_only->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_create(&arc_uncached->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_create(&arc_uncached->arcs_size[ARC_BUFC_METADATA]);
wmsum_init(&arc_mru_ghost->arcs_hits[ARC_BUFC_DATA], 0);
wmsum_init(&arc_mru_ghost->arcs_hits[ARC_BUFC_METADATA], 0);
wmsum_init(&arc_mfu_ghost->arcs_hits[ARC_BUFC_DATA], 0);
wmsum_init(&arc_mfu_ghost->arcs_hits[ARC_BUFC_METADATA], 0);
wmsum_init(&arc_sums.arcstat_hits, 0);
wmsum_init(&arc_sums.arcstat_iohits, 0);
wmsum_init(&arc_sums.arcstat_misses, 0);
wmsum_init(&arc_sums.arcstat_demand_data_hits, 0);
wmsum_init(&arc_sums.arcstat_demand_data_iohits, 0);
wmsum_init(&arc_sums.arcstat_demand_data_misses, 0);
wmsum_init(&arc_sums.arcstat_demand_metadata_hits, 0);
wmsum_init(&arc_sums.arcstat_demand_metadata_iohits, 0);
wmsum_init(&arc_sums.arcstat_demand_metadata_misses, 0);
wmsum_init(&arc_sums.arcstat_prefetch_data_hits, 0);
wmsum_init(&arc_sums.arcstat_prefetch_data_iohits, 0);
wmsum_init(&arc_sums.arcstat_prefetch_data_misses, 0);
wmsum_init(&arc_sums.arcstat_prefetch_metadata_hits, 0);
wmsum_init(&arc_sums.arcstat_prefetch_metadata_iohits, 0);
wmsum_init(&arc_sums.arcstat_prefetch_metadata_misses, 0);
wmsum_init(&arc_sums.arcstat_mru_hits, 0);
wmsum_init(&arc_sums.arcstat_mru_ghost_hits, 0);
wmsum_init(&arc_sums.arcstat_mfu_hits, 0);
wmsum_init(&arc_sums.arcstat_mfu_ghost_hits, 0);
wmsum_init(&arc_sums.arcstat_uncached_hits, 0);
wmsum_init(&arc_sums.arcstat_deleted, 0);
wmsum_init(&arc_sums.arcstat_mutex_miss, 0);
wmsum_init(&arc_sums.arcstat_access_skip, 0);
wmsum_init(&arc_sums.arcstat_evict_skip, 0);
wmsum_init(&arc_sums.arcstat_evict_not_enough, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_cached, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_eligible, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_eligible_mfu, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_eligible_mru, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_ineligible, 0);
wmsum_init(&arc_sums.arcstat_evict_l2_skip, 0);
wmsum_init(&arc_sums.arcstat_hash_collisions, 0);
wmsum_init(&arc_sums.arcstat_hash_chains, 0);
aggsum_init(&arc_sums.arcstat_size, 0);
wmsum_init(&arc_sums.arcstat_compressed_size, 0);
wmsum_init(&arc_sums.arcstat_uncompressed_size, 0);
wmsum_init(&arc_sums.arcstat_overhead_size, 0);
wmsum_init(&arc_sums.arcstat_hdr_size, 0);
wmsum_init(&arc_sums.arcstat_data_size, 0);
wmsum_init(&arc_sums.arcstat_metadata_size, 0);
wmsum_init(&arc_sums.arcstat_dbuf_size, 0);
wmsum_init(&arc_sums.arcstat_dnode_size, 0);
wmsum_init(&arc_sums.arcstat_bonus_size, 0);
wmsum_init(&arc_sums.arcstat_l2_hits, 0);
wmsum_init(&arc_sums.arcstat_l2_misses, 0);
wmsum_init(&arc_sums.arcstat_l2_prefetch_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_mru_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_mfu_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_bufc_data_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_bufc_metadata_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_feeds, 0);
wmsum_init(&arc_sums.arcstat_l2_rw_clash, 0);
wmsum_init(&arc_sums.arcstat_l2_read_bytes, 0);
wmsum_init(&arc_sums.arcstat_l2_write_bytes, 0);
wmsum_init(&arc_sums.arcstat_l2_writes_sent, 0);
wmsum_init(&arc_sums.arcstat_l2_writes_done, 0);
wmsum_init(&arc_sums.arcstat_l2_writes_error, 0);
wmsum_init(&arc_sums.arcstat_l2_writes_lock_retry, 0);
wmsum_init(&arc_sums.arcstat_l2_evict_lock_retry, 0);
wmsum_init(&arc_sums.arcstat_l2_evict_reading, 0);
wmsum_init(&arc_sums.arcstat_l2_evict_l1cached, 0);
wmsum_init(&arc_sums.arcstat_l2_free_on_write, 0);
wmsum_init(&arc_sums.arcstat_l2_abort_lowmem, 0);
wmsum_init(&arc_sums.arcstat_l2_cksum_bad, 0);
wmsum_init(&arc_sums.arcstat_l2_io_error, 0);
wmsum_init(&arc_sums.arcstat_l2_lsize, 0);
wmsum_init(&arc_sums.arcstat_l2_psize, 0);
aggsum_init(&arc_sums.arcstat_l2_hdr_size, 0);
wmsum_init(&arc_sums.arcstat_l2_log_blk_writes, 0);
wmsum_init(&arc_sums.arcstat_l2_log_blk_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_log_blk_count, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_success, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_unsupported, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_io_errors, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_dh_errors, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_cksum_lb_errors, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_abort_lowmem, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_size, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_asize, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_bufs, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_bufs_precached, 0);
wmsum_init(&arc_sums.arcstat_l2_rebuild_log_blks, 0);
wmsum_init(&arc_sums.arcstat_memory_throttle_count, 0);
wmsum_init(&arc_sums.arcstat_memory_direct_count, 0);
wmsum_init(&arc_sums.arcstat_memory_indirect_count, 0);
wmsum_init(&arc_sums.arcstat_prune, 0);
wmsum_init(&arc_sums.arcstat_meta_used, 0);
wmsum_init(&arc_sums.arcstat_async_upgrade_sync, 0);
wmsum_init(&arc_sums.arcstat_predictive_prefetch, 0);
wmsum_init(&arc_sums.arcstat_demand_hit_predictive_prefetch, 0);
wmsum_init(&arc_sums.arcstat_demand_iohit_predictive_prefetch, 0);
wmsum_init(&arc_sums.arcstat_prescient_prefetch, 0);
wmsum_init(&arc_sums.arcstat_demand_hit_prescient_prefetch, 0);
wmsum_init(&arc_sums.arcstat_demand_iohit_prescient_prefetch, 0);
wmsum_init(&arc_sums.arcstat_raw_size, 0);
wmsum_init(&arc_sums.arcstat_cached_only_in_progress, 0);
wmsum_init(&arc_sums.arcstat_abd_chunk_waste_size, 0);
arc_anon->arcs_state = ARC_STATE_ANON;
arc_mru->arcs_state = ARC_STATE_MRU;
arc_mru_ghost->arcs_state = ARC_STATE_MRU_GHOST;
arc_mfu->arcs_state = ARC_STATE_MFU;
arc_mfu_ghost->arcs_state = ARC_STATE_MFU_GHOST;
arc_l2c_only->arcs_state = ARC_STATE_L2C_ONLY;
arc_uncached->arcs_state = ARC_STATE_UNCACHED;
}
static void
arc_state_fini(void)
{
zfs_refcount_destroy(&arc_anon->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_anon->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mru->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mru->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mru_ghost->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mru_ghost->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mfu->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mfu->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mfu_ghost->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mfu_ghost->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_l2c_only->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_l2c_only->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_uncached->arcs_esize[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_uncached->arcs_esize[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_anon->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_anon->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mru->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mru->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mru_ghost->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mru_ghost->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mfu->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mfu->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_mfu_ghost->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_mfu_ghost->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_l2c_only->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_l2c_only->arcs_size[ARC_BUFC_METADATA]);
zfs_refcount_destroy(&arc_uncached->arcs_size[ARC_BUFC_DATA]);
zfs_refcount_destroy(&arc_uncached->arcs_size[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mru->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mru_ghost->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mfu->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mfu_ghost->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_mru->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_mru_ghost->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_mfu->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_mfu_ghost->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_l2c_only->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_l2c_only->arcs_list[ARC_BUFC_DATA]);
multilist_destroy(&arc_uncached->arcs_list[ARC_BUFC_METADATA]);
multilist_destroy(&arc_uncached->arcs_list[ARC_BUFC_DATA]);
wmsum_fini(&arc_mru_ghost->arcs_hits[ARC_BUFC_DATA]);
wmsum_fini(&arc_mru_ghost->arcs_hits[ARC_BUFC_METADATA]);
wmsum_fini(&arc_mfu_ghost->arcs_hits[ARC_BUFC_DATA]);
wmsum_fini(&arc_mfu_ghost->arcs_hits[ARC_BUFC_METADATA]);
wmsum_fini(&arc_sums.arcstat_hits);
wmsum_fini(&arc_sums.arcstat_iohits);
wmsum_fini(&arc_sums.arcstat_misses);
wmsum_fini(&arc_sums.arcstat_demand_data_hits);
wmsum_fini(&arc_sums.arcstat_demand_data_iohits);
wmsum_fini(&arc_sums.arcstat_demand_data_misses);
wmsum_fini(&arc_sums.arcstat_demand_metadata_hits);
wmsum_fini(&arc_sums.arcstat_demand_metadata_iohits);
wmsum_fini(&arc_sums.arcstat_demand_metadata_misses);
wmsum_fini(&arc_sums.arcstat_prefetch_data_hits);
wmsum_fini(&arc_sums.arcstat_prefetch_data_iohits);
wmsum_fini(&arc_sums.arcstat_prefetch_data_misses);
wmsum_fini(&arc_sums.arcstat_prefetch_metadata_hits);
wmsum_fini(&arc_sums.arcstat_prefetch_metadata_iohits);
wmsum_fini(&arc_sums.arcstat_prefetch_metadata_misses);
wmsum_fini(&arc_sums.arcstat_mru_hits);
wmsum_fini(&arc_sums.arcstat_mru_ghost_hits);
wmsum_fini(&arc_sums.arcstat_mfu_hits);
wmsum_fini(&arc_sums.arcstat_mfu_ghost_hits);
wmsum_fini(&arc_sums.arcstat_uncached_hits);
wmsum_fini(&arc_sums.arcstat_deleted);
wmsum_fini(&arc_sums.arcstat_mutex_miss);
wmsum_fini(&arc_sums.arcstat_access_skip);
wmsum_fini(&arc_sums.arcstat_evict_skip);
wmsum_fini(&arc_sums.arcstat_evict_not_enough);
wmsum_fini(&arc_sums.arcstat_evict_l2_cached);
wmsum_fini(&arc_sums.arcstat_evict_l2_eligible);
wmsum_fini(&arc_sums.arcstat_evict_l2_eligible_mfu);
wmsum_fini(&arc_sums.arcstat_evict_l2_eligible_mru);
wmsum_fini(&arc_sums.arcstat_evict_l2_ineligible);
wmsum_fini(&arc_sums.arcstat_evict_l2_skip);
wmsum_fini(&arc_sums.arcstat_hash_collisions);
wmsum_fini(&arc_sums.arcstat_hash_chains);
aggsum_fini(&arc_sums.arcstat_size);
wmsum_fini(&arc_sums.arcstat_compressed_size);
wmsum_fini(&arc_sums.arcstat_uncompressed_size);
wmsum_fini(&arc_sums.arcstat_overhead_size);
wmsum_fini(&arc_sums.arcstat_hdr_size);
wmsum_fini(&arc_sums.arcstat_data_size);
wmsum_fini(&arc_sums.arcstat_metadata_size);
wmsum_fini(&arc_sums.arcstat_dbuf_size);
wmsum_fini(&arc_sums.arcstat_dnode_size);
wmsum_fini(&arc_sums.arcstat_bonus_size);
wmsum_fini(&arc_sums.arcstat_l2_hits);
wmsum_fini(&arc_sums.arcstat_l2_misses);
wmsum_fini(&arc_sums.arcstat_l2_prefetch_asize);
wmsum_fini(&arc_sums.arcstat_l2_mru_asize);
wmsum_fini(&arc_sums.arcstat_l2_mfu_asize);
wmsum_fini(&arc_sums.arcstat_l2_bufc_data_asize);
wmsum_fini(&arc_sums.arcstat_l2_bufc_metadata_asize);
wmsum_fini(&arc_sums.arcstat_l2_feeds);
wmsum_fini(&arc_sums.arcstat_l2_rw_clash);
wmsum_fini(&arc_sums.arcstat_l2_read_bytes);
wmsum_fini(&arc_sums.arcstat_l2_write_bytes);
wmsum_fini(&arc_sums.arcstat_l2_writes_sent);
wmsum_fini(&arc_sums.arcstat_l2_writes_done);
wmsum_fini(&arc_sums.arcstat_l2_writes_error);
wmsum_fini(&arc_sums.arcstat_l2_writes_lock_retry);
wmsum_fini(&arc_sums.arcstat_l2_evict_lock_retry);
wmsum_fini(&arc_sums.arcstat_l2_evict_reading);
wmsum_fini(&arc_sums.arcstat_l2_evict_l1cached);
wmsum_fini(&arc_sums.arcstat_l2_free_on_write);
wmsum_fini(&arc_sums.arcstat_l2_abort_lowmem);
wmsum_fini(&arc_sums.arcstat_l2_cksum_bad);
wmsum_fini(&arc_sums.arcstat_l2_io_error);
wmsum_fini(&arc_sums.arcstat_l2_lsize);
wmsum_fini(&arc_sums.arcstat_l2_psize);
aggsum_fini(&arc_sums.arcstat_l2_hdr_size);
wmsum_fini(&arc_sums.arcstat_l2_log_blk_writes);
wmsum_fini(&arc_sums.arcstat_l2_log_blk_asize);
wmsum_fini(&arc_sums.arcstat_l2_log_blk_count);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_success);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_unsupported);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_io_errors);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_dh_errors);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_cksum_lb_errors);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_abort_lowmem);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_size);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_asize);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_bufs);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_bufs_precached);
wmsum_fini(&arc_sums.arcstat_l2_rebuild_log_blks);
wmsum_fini(&arc_sums.arcstat_memory_throttle_count);
wmsum_fini(&arc_sums.arcstat_memory_direct_count);
wmsum_fini(&arc_sums.arcstat_memory_indirect_count);
wmsum_fini(&arc_sums.arcstat_prune);
wmsum_fini(&arc_sums.arcstat_meta_used);
wmsum_fini(&arc_sums.arcstat_async_upgrade_sync);
wmsum_fini(&arc_sums.arcstat_predictive_prefetch);
wmsum_fini(&arc_sums.arcstat_demand_hit_predictive_prefetch);
wmsum_fini(&arc_sums.arcstat_demand_iohit_predictive_prefetch);
wmsum_fini(&arc_sums.arcstat_prescient_prefetch);
wmsum_fini(&arc_sums.arcstat_demand_hit_prescient_prefetch);
wmsum_fini(&arc_sums.arcstat_demand_iohit_prescient_prefetch);
wmsum_fini(&arc_sums.arcstat_raw_size);
wmsum_fini(&arc_sums.arcstat_cached_only_in_progress);
wmsum_fini(&arc_sums.arcstat_abd_chunk_waste_size);
}
uint64_t
arc_target_bytes(void)
{
return (arc_c);
}
void
arc_set_limits(uint64_t allmem)
{
/* Set min cache to 1/32 of all memory, or 32MB, whichever is more. */
arc_c_min = MAX(allmem / 32, 2ULL << SPA_MAXBLOCKSHIFT);
/* How to set default max varies by platform. */
arc_c_max = arc_default_max(arc_c_min, allmem);
}
void
arc_init(void)
{
uint64_t percent, allmem = arc_all_memory();
mutex_init(&arc_evict_lock, NULL, MUTEX_DEFAULT, NULL);
list_create(&arc_evict_waiters, sizeof (arc_evict_waiter_t),
offsetof(arc_evict_waiter_t, aew_node));
arc_min_prefetch_ms = 1000;
arc_min_prescient_prefetch_ms = 6000;
#if defined(_KERNEL)
arc_lowmem_init();
#endif
arc_set_limits(allmem);
#ifdef _KERNEL
/*
* If zfs_arc_max is non-zero at init, meaning it was set in the kernel
* environment before the module was loaded, don't block setting the
* maximum because it is less than arc_c_min, instead, reset arc_c_min
* to a lower value.
* zfs_arc_min will be handled by arc_tuning_update().
*/
if (zfs_arc_max != 0 && zfs_arc_max >= MIN_ARC_MAX &&
zfs_arc_max < allmem) {
arc_c_max = zfs_arc_max;
if (arc_c_min >= arc_c_max) {
arc_c_min = MAX(zfs_arc_max / 2,
2ULL << SPA_MAXBLOCKSHIFT);
}
}
#else
/*
* In userland, there's only the memory pressure that we artificially
* create (see arc_available_memory()). Don't let arc_c get too
* small, because it can cause transactions to be larger than
* arc_c, causing arc_tempreserve_space() to fail.
*/
arc_c_min = MAX(arc_c_max / 2, 2ULL << SPA_MAXBLOCKSHIFT);
#endif
arc_c = arc_c_min;
/*
* 32-bit fixed point fractions of metadata from total ARC size,
* MRU data from all data and MRU metadata from all metadata.
*/
arc_meta = (1ULL << 32) / 4; /* Metadata is 25% of arc_c. */
arc_pd = (1ULL << 32) / 2; /* Data MRU is 50% of data. */
arc_pm = (1ULL << 32) / 2; /* Metadata MRU is 50% of metadata. */
percent = MIN(zfs_arc_dnode_limit_percent, 100);
arc_dnode_limit = arc_c_max * percent / 100;
/* Apply user specified tunings */
arc_tuning_update(B_TRUE);
/* if kmem_flags are set, lets try to use less memory */
if (kmem_debugging())
arc_c = arc_c / 2;
if (arc_c < arc_c_min)
arc_c = arc_c_min;
arc_register_hotplug();
arc_state_init();
buf_init();
list_create(&arc_prune_list, sizeof (arc_prune_t),
offsetof(arc_prune_t, p_node));
mutex_init(&arc_prune_mtx, NULL, MUTEX_DEFAULT, NULL);
arc_prune_taskq = taskq_create("arc_prune", zfs_arc_prune_task_threads,
defclsyspri, 100, INT_MAX, TASKQ_PREPOPULATE | TASKQ_DYNAMIC);
arc_ksp = kstat_create("zfs", 0, "arcstats", "misc", KSTAT_TYPE_NAMED,
sizeof (arc_stats) / sizeof (kstat_named_t), KSTAT_FLAG_VIRTUAL);
if (arc_ksp != NULL) {
arc_ksp->ks_data = &arc_stats;
arc_ksp->ks_update = arc_kstat_update;
kstat_install(arc_ksp);
}
arc_state_evict_markers =
arc_state_alloc_markers(arc_state_evict_marker_count);
arc_evict_zthr = zthr_create_timer("arc_evict",
arc_evict_cb_check, arc_evict_cb, NULL, SEC2NSEC(1), defclsyspri);
arc_reap_zthr = zthr_create_timer("arc_reap",
arc_reap_cb_check, arc_reap_cb, NULL, SEC2NSEC(1), minclsyspri);
arc_warm = B_FALSE;
/*
* Calculate maximum amount of dirty data per pool.
*
* If it has been set by a module parameter, take that.
* Otherwise, use a percentage of physical memory defined by
* zfs_dirty_data_max_percent (default 10%) with a cap at
* zfs_dirty_data_max_max (default 4G or 25% of physical memory).
*/
#ifdef __LP64__
if (zfs_dirty_data_max_max == 0)
zfs_dirty_data_max_max = MIN(4ULL * 1024 * 1024 * 1024,
allmem * zfs_dirty_data_max_max_percent / 100);
#else
if (zfs_dirty_data_max_max == 0)
zfs_dirty_data_max_max = MIN(1ULL * 1024 * 1024 * 1024,
allmem * zfs_dirty_data_max_max_percent / 100);
#endif
if (zfs_dirty_data_max == 0) {
zfs_dirty_data_max = allmem *
zfs_dirty_data_max_percent / 100;
zfs_dirty_data_max = MIN(zfs_dirty_data_max,
zfs_dirty_data_max_max);
}
if (zfs_wrlog_data_max == 0) {
/*
* dp_wrlog_total is reduced for each txg at the end of
* spa_sync(). However, dp_dirty_total is reduced every time
* a block is written out. Thus under normal operation,
* dp_wrlog_total could grow 2 times as big as
* zfs_dirty_data_max.
*/
zfs_wrlog_data_max = zfs_dirty_data_max * 2;
}
}
void
arc_fini(void)
{
arc_prune_t *p;
#ifdef _KERNEL
arc_lowmem_fini();
#endif /* _KERNEL */
/* Use B_TRUE to ensure *all* buffers are evicted */
arc_flush(NULL, B_TRUE);
if (arc_ksp != NULL) {
kstat_delete(arc_ksp);
arc_ksp = NULL;
}
taskq_wait(arc_prune_taskq);
taskq_destroy(arc_prune_taskq);
mutex_enter(&arc_prune_mtx);
while ((p = list_head(&arc_prune_list)) != NULL) {
list_remove(&arc_prune_list, p);
zfs_refcount_remove(&p->p_refcnt, &arc_prune_list);
zfs_refcount_destroy(&p->p_refcnt);
kmem_free(p, sizeof (*p));
}
mutex_exit(&arc_prune_mtx);
list_destroy(&arc_prune_list);
mutex_destroy(&arc_prune_mtx);
(void) zthr_cancel(arc_evict_zthr);
(void) zthr_cancel(arc_reap_zthr);
arc_state_free_markers(arc_state_evict_markers,
arc_state_evict_marker_count);
mutex_destroy(&arc_evict_lock);
list_destroy(&arc_evict_waiters);
/*
* Free any buffers that were tagged for destruction. This needs
* to occur before arc_state_fini() runs and destroys the aggsum
* values which are updated when freeing scatter ABDs.
*/
l2arc_do_free_on_write();
/*
* buf_fini() must proceed arc_state_fini() because buf_fin() may
* trigger the release of kmem magazines, which can callback to
* arc_space_return() which accesses aggsums freed in act_state_fini().
*/
buf_fini();
arc_state_fini();
arc_unregister_hotplug();
/*
* We destroy the zthrs after all the ARC state has been
* torn down to avoid the case of them receiving any
* wakeup() signals after they are destroyed.
*/
zthr_destroy(arc_evict_zthr);
zthr_destroy(arc_reap_zthr);
ASSERT0(arc_loaned_bytes);
}
/*
* Level 2 ARC
*
* The level 2 ARC (L2ARC) is a cache layer in-between main memory and disk.
* It uses dedicated storage devices to hold cached data, which are populated
* using large infrequent writes. The main role of this cache is to boost
* the performance of random read workloads. The intended L2ARC devices
* include short-stroked disks, solid state disks, and other media with
* substantially faster read latency than disk.
*
* +-----------------------+
* | ARC |
* +-----------------------+
* | ^ ^
* | | |
* l2arc_feed_thread() arc_read()
* | | |
* | l2arc read |
* V | |
* +---------------+ |
* | L2ARC | |
* +---------------+ |
* | ^ |
* l2arc_write() | |
* | | |
* V | |
* +-------+ +-------+
* | vdev | | vdev |
* | cache | | cache |
* +-------+ +-------+
* +=========+ .-----.
* : L2ARC : |-_____-|
* : devices : | Disks |
* +=========+ `-_____-'
*
* Read requests are satisfied from the following sources, in order:
*
* 1) ARC
* 2) vdev cache of L2ARC devices
* 3) L2ARC devices
* 4) vdev cache of disks
* 5) disks
*
* Some L2ARC device types exhibit extremely slow write performance.
* To accommodate for this there are some significant differences between
* the L2ARC and traditional cache design:
*
* 1. There is no eviction path from the ARC to the L2ARC. Evictions from
* the ARC behave as usual, freeing buffers and placing headers on ghost
* lists. The ARC does not send buffers to the L2ARC during eviction as
* this would add inflated write latencies for all ARC memory pressure.
*
* 2. The L2ARC attempts to cache data from the ARC before it is evicted.
* It does this by periodically scanning buffers from the eviction-end of
* the MFU and MRU ARC lists, copying them to the L2ARC devices if they are
* not already there. It scans until a headroom of buffers is satisfied,
* which itself is a buffer for ARC eviction. If a compressible buffer is
* found during scanning and selected for writing to an L2ARC device, we
* temporarily boost scanning headroom during the next scan cycle to make
* sure we adapt to compression effects (which might significantly reduce
* the data volume we write to L2ARC). The thread that does this is
* l2arc_feed_thread(), illustrated below; example sizes are included to
* provide a better sense of ratio than this diagram:
*
* head --> tail
* +---------------------+----------+
* ARC_mfu |:::::#:::::::::::::::|o#o###o###|-->. # already on L2ARC
* +---------------------+----------+ | o L2ARC eligible
* ARC_mru |:#:::::::::::::::::::|#o#ooo####|-->| : ARC buffer
* +---------------------+----------+ |
* 15.9 Gbytes ^ 32 Mbytes |
* headroom |
* l2arc_feed_thread()
* |
* l2arc write hand <--[oooo]--'
* | 8 Mbyte
* | write max
* V
* +==============================+
* L2ARC dev |####|#|###|###| |####| ... |
* +==============================+
* 32 Gbytes
*
* 3. If an ARC buffer is copied to the L2ARC but then hit instead of
* evicted, then the L2ARC has cached a buffer much sooner than it probably
* needed to, potentially wasting L2ARC device bandwidth and storage. It is
* safe to say that this is an uncommon case, since buffers at the end of
* the ARC lists have moved there due to inactivity.
*
* 4. If the ARC evicts faster than the L2ARC can maintain a headroom,
* then the L2ARC simply misses copying some buffers. This serves as a
* pressure valve to prevent heavy read workloads from both stalling the ARC
* with waits and clogging the L2ARC with writes. This also helps prevent
* the potential for the L2ARC to churn if it attempts to cache content too
* quickly, such as during backups of the entire pool.
*
* 5. After system boot and before the ARC has filled main memory, there are
* no evictions from the ARC and so the tails of the ARC_mfu and ARC_mru
* lists can remain mostly static. Instead of searching from tail of these
* lists as pictured, the l2arc_feed_thread() will search from the list heads
* for eligible buffers, greatly increasing its chance of finding them.
*
* The L2ARC device write speed is also boosted during this time so that
* the L2ARC warms up faster. Since there have been no ARC evictions yet,
* there are no L2ARC reads, and no fear of degrading read performance
* through increased writes.
*
* 6. Writes to the L2ARC devices are grouped and sent in-sequence, so that
* the vdev queue can aggregate them into larger and fewer writes. Each
* device is written to in a rotor fashion, sweeping writes through
* available space then repeating.
*
* 7. The L2ARC does not store dirty content. It never needs to flush
* write buffers back to disk based storage.
*
* 8. If an ARC buffer is written (and dirtied) which also exists in the
* L2ARC, the now stale L2ARC buffer is immediately dropped.
*
* The performance of the L2ARC can be tweaked by a number of tunables, which
* may be necessary for different workloads:
*
* l2arc_write_max max write bytes per interval
* l2arc_write_boost extra write bytes during device warmup
* l2arc_noprefetch skip caching prefetched buffers
* l2arc_headroom number of max device writes to precache
* l2arc_headroom_boost when we find compressed buffers during ARC
* scanning, we multiply headroom by this
* percentage factor for the next scan cycle,
* since more compressed buffers are likely to
* be present
* l2arc_feed_secs seconds between L2ARC writing
*
* Tunables may be removed or added as future performance improvements are
* integrated, and also may become zpool properties.
*
* There are three key functions that control how the L2ARC warms up:
*
* l2arc_write_eligible() check if a buffer is eligible to cache
* l2arc_write_size() calculate how much to write
* l2arc_write_interval() calculate sleep delay between writes
*
* These three functions determine what to write, how much, and how quickly
* to send writes.
*
* L2ARC persistence:
*
* When writing buffers to L2ARC, we periodically add some metadata to
* make sure we can pick them up after reboot, thus dramatically reducing
* the impact that any downtime has on the performance of storage systems
* with large caches.
*
* The implementation works fairly simply by integrating the following two
* modifications:
*
* *) When writing to the L2ARC, we occasionally write a "l2arc log block",
* which is an additional piece of metadata which describes what's been
* written. This allows us to rebuild the arc_buf_hdr_t structures of the
* main ARC buffers. There are 2 linked-lists of log blocks headed by
* dh_start_lbps[2]. We alternate which chain we append to, so they are
* time-wise and offset-wise interleaved, but that is an optimization rather
* than for correctness. The log block also includes a pointer to the
* previous block in its chain.
*
* *) We reserve SPA_MINBLOCKSIZE of space at the start of each L2ARC device
* for our header bookkeeping purposes. This contains a device header,
* which contains our top-level reference structures. We update it each
* time we write a new log block, so that we're able to locate it in the
* L2ARC device. If this write results in an inconsistent device header
* (e.g. due to power failure), we detect this by verifying the header's
* checksum and simply fail to reconstruct the L2ARC after reboot.
*
* Implementation diagram:
*
* +=== L2ARC device (not to scale) ======================================+
* | ___two newest log block pointers__.__________ |
* | / \dh_start_lbps[1] |
* | / \ \dh_start_lbps[0]|
* |.___/__. V V |
* ||L2 dev|....|lb |bufs |lb |bufs |lb |bufs |lb |bufs |lb |---(empty)---|
* || hdr| ^ /^ /^ / / |
* |+------+ ...--\-------/ \-----/--\------/ / |
* | \--------------/ \--------------/ |
* +======================================================================+
*
* As can be seen on the diagram, rather than using a simple linked list,
* we use a pair of linked lists with alternating elements. This is a
* performance enhancement due to the fact that we only find out the
* address of the next log block access once the current block has been
* completely read in. Obviously, this hurts performance, because we'd be
* keeping the device's I/O queue at only a 1 operation deep, thus
* incurring a large amount of I/O round-trip latency. Having two lists
* allows us to fetch two log blocks ahead of where we are currently
* rebuilding L2ARC buffers.
*
* On-device data structures:
*
* L2ARC device header: l2arc_dev_hdr_phys_t
* L2ARC log block: l2arc_log_blk_phys_t
*
* L2ARC reconstruction:
*
* When writing data, we simply write in the standard rotary fashion,
* evicting buffers as we go and simply writing new data over them (writing
* a new log block every now and then). This obviously means that once we
* loop around the end of the device, we will start cutting into an already
* committed log block (and its referenced data buffers), like so:
*
* current write head__ __old tail
* \ /
* V V
* <--|bufs |lb |bufs |lb | |bufs |lb |bufs |lb |-->
* ^ ^^^^^^^^^___________________________________
* | \
* <<nextwrite>> may overwrite this blk and/or its bufs --'
*
* When importing the pool, we detect this situation and use it to stop
* our scanning process (see l2arc_rebuild).
*
* There is one significant caveat to consider when rebuilding ARC contents
* from an L2ARC device: what about invalidated buffers? Given the above
* construction, we cannot update blocks which we've already written to amend
* them to remove buffers which were invalidated. Thus, during reconstruction,
* we might be populating the cache with buffers for data that's not on the
* main pool anymore, or may have been overwritten!
*
* As it turns out, this isn't a problem. Every arc_read request includes
* both the DVA and, crucially, the birth TXG of the BP the caller is
* looking for. So even if the cache were populated by completely rotten
* blocks for data that had been long deleted and/or overwritten, we'll
* never actually return bad data from the cache, since the DVA with the
* birth TXG uniquely identify a block in space and time - once created,
* a block is immutable on disk. The worst thing we have done is wasted
* some time and memory at l2arc rebuild to reconstruct outdated ARC
* entries that will get dropped from the l2arc as it is being updated
* with new blocks.
*
* L2ARC buffers that have been evicted by l2arc_evict() ahead of the write
* hand are not restored. This is done by saving the offset (in bytes)
* l2arc_evict() has evicted to in the L2ARC device header and taking it
* into account when restoring buffers.
*/
static boolean_t
l2arc_write_eligible(uint64_t spa_guid, arc_buf_hdr_t *hdr)
{
/*
* A buffer is *not* eligible for the L2ARC if it:
* 1. belongs to a different spa.
* 2. is already cached on the L2ARC.
* 3. has an I/O in progress (it may be an incomplete read).
* 4. is flagged not eligible (zfs property).
*/
if (hdr->b_spa != spa_guid || HDR_HAS_L2HDR(hdr) ||
HDR_IO_IN_PROGRESS(hdr) || !HDR_L2CACHE(hdr))
return (B_FALSE);
return (B_TRUE);
}
static uint64_t
l2arc_write_size(l2arc_dev_t *dev)
{
uint64_t size, dev_size, tsize;
/*
* Make sure our globals have meaningful values in case the user
* altered them.
*/
size = l2arc_write_max;
if (size == 0) {
cmn_err(CE_NOTE, "Bad value for l2arc_write_max, value must "
"be greater than zero, resetting it to the default (%d)",
L2ARC_WRITE_SIZE);
size = l2arc_write_max = L2ARC_WRITE_SIZE;
}
if (arc_warm == B_FALSE)
size += l2arc_write_boost;
/*
* Make sure the write size does not exceed the size of the cache
* device. This is important in l2arc_evict(), otherwise infinite
* iteration can occur.
*/
dev_size = dev->l2ad_end - dev->l2ad_start;
tsize = size + l2arc_log_blk_overhead(size, dev);
if (dev->l2ad_vdev->vdev_has_trim && l2arc_trim_ahead > 0)
tsize += MAX(64 * 1024 * 1024,
(tsize * l2arc_trim_ahead) / 100);
if (tsize >= dev_size) {
cmn_err(CE_NOTE, "l2arc_write_max or l2arc_write_boost "
"plus the overhead of log blocks (persistent L2ARC, "
"%llu bytes) exceeds the size of the cache device "
"(guid %llu), resetting them to the default (%d)",
(u_longlong_t)l2arc_log_blk_overhead(size, dev),
(u_longlong_t)dev->l2ad_vdev->vdev_guid, L2ARC_WRITE_SIZE);
size = l2arc_write_max = l2arc_write_boost = L2ARC_WRITE_SIZE;
if (arc_warm == B_FALSE)
size += l2arc_write_boost;
}
return (size);
}
static clock_t
l2arc_write_interval(clock_t began, uint64_t wanted, uint64_t wrote)
{
clock_t interval, next, now;
/*
* If the ARC lists are busy, increase our write rate; if the
* lists are stale, idle back. This is achieved by checking
* how much we previously wrote - if it was more than half of
* what we wanted, schedule the next write much sooner.
*/
if (l2arc_feed_again && wrote > (wanted / 2))
interval = (hz * l2arc_feed_min_ms) / 1000;
else
interval = hz * l2arc_feed_secs;
now = ddi_get_lbolt();
next = MAX(now, MIN(now + interval, began + interval));
return (next);
}
/*
* Cycle through L2ARC devices. This is how L2ARC load balances.
* If a device is returned, this also returns holding the spa config lock.
*/
static l2arc_dev_t *
l2arc_dev_get_next(void)
{
l2arc_dev_t *first, *next = NULL;
/*
* Lock out the removal of spas (spa_namespace_lock), then removal
* of cache devices (l2arc_dev_mtx). Once a device has been selected,
* both locks will be dropped and a spa config lock held instead.
*/
mutex_enter(&spa_namespace_lock);
mutex_enter(&l2arc_dev_mtx);
/* if there are no vdevs, there is nothing to do */
if (l2arc_ndev == 0)
goto out;
first = NULL;
next = l2arc_dev_last;
do {
/* loop around the list looking for a non-faulted vdev */
if (next == NULL) {
next = list_head(l2arc_dev_list);
} else {
next = list_next(l2arc_dev_list, next);
if (next == NULL)
next = list_head(l2arc_dev_list);
}
/* if we have come back to the start, bail out */
if (first == NULL)
first = next;
else if (next == first)
break;
ASSERT3P(next, !=, NULL);
} while (vdev_is_dead(next->l2ad_vdev) || next->l2ad_rebuild ||
next->l2ad_trim_all);
/* if we were unable to find any usable vdevs, return NULL */
if (vdev_is_dead(next->l2ad_vdev) || next->l2ad_rebuild ||
next->l2ad_trim_all)
next = NULL;
l2arc_dev_last = next;
out:
mutex_exit(&l2arc_dev_mtx);
/*
* Grab the config lock to prevent the 'next' device from being
* removed while we are writing to it.
*/
if (next != NULL)
spa_config_enter(next->l2ad_spa, SCL_L2ARC, next, RW_READER);
mutex_exit(&spa_namespace_lock);
return (next);
}
/*
* Free buffers that were tagged for destruction.
*/
static void
l2arc_do_free_on_write(void)
{
list_t *buflist;
l2arc_data_free_t *df, *df_prev;
mutex_enter(&l2arc_free_on_write_mtx);
buflist = l2arc_free_on_write;
for (df = list_tail(buflist); df; df = df_prev) {
df_prev = list_prev(buflist, df);
ASSERT3P(df->l2df_abd, !=, NULL);
abd_free(df->l2df_abd);
list_remove(buflist, df);
kmem_free(df, sizeof (l2arc_data_free_t));
}
mutex_exit(&l2arc_free_on_write_mtx);
}
/*
* A write to a cache device has completed. Update all headers to allow
* reads from these buffers to begin.
*/
static void
l2arc_write_done(zio_t *zio)
{
l2arc_write_callback_t *cb;
l2arc_lb_abd_buf_t *abd_buf;
l2arc_lb_ptr_buf_t *lb_ptr_buf;
l2arc_dev_t *dev;
l2arc_dev_hdr_phys_t *l2dhdr;
list_t *buflist;
arc_buf_hdr_t *head, *hdr, *hdr_prev;
kmutex_t *hash_lock;
int64_t bytes_dropped = 0;
cb = zio->io_private;
ASSERT3P(cb, !=, NULL);
dev = cb->l2wcb_dev;
l2dhdr = dev->l2ad_dev_hdr;
ASSERT3P(dev, !=, NULL);
head = cb->l2wcb_head;
ASSERT3P(head, !=, NULL);
buflist = &dev->l2ad_buflist;
ASSERT3P(buflist, !=, NULL);
DTRACE_PROBE2(l2arc__iodone, zio_t *, zio,
l2arc_write_callback_t *, cb);
/*
* All writes completed, or an error was hit.
*/
top:
mutex_enter(&dev->l2ad_mtx);
for (hdr = list_prev(buflist, head); hdr; hdr = hdr_prev) {
hdr_prev = list_prev(buflist, hdr);
hash_lock = HDR_LOCK(hdr);
/*
* We cannot use mutex_enter or else we can deadlock
* with l2arc_write_buffers (due to swapping the order
* the hash lock and l2ad_mtx are taken).
*/
if (!mutex_tryenter(hash_lock)) {
/*
* Missed the hash lock. We must retry so we
* don't leave the ARC_FLAG_L2_WRITING bit set.
*/
ARCSTAT_BUMP(arcstat_l2_writes_lock_retry);
/*
* We don't want to rescan the headers we've
* already marked as having been written out, so
* we reinsert the head node so we can pick up
* where we left off.
*/
list_remove(buflist, head);
list_insert_after(buflist, hdr, head);
mutex_exit(&dev->l2ad_mtx);
/*
* We wait for the hash lock to become available
* to try and prevent busy waiting, and increase
* the chance we'll be able to acquire the lock
* the next time around.
*/
mutex_enter(hash_lock);
mutex_exit(hash_lock);
goto top;
}
/*
* We could not have been moved into the arc_l2c_only
* state while in-flight due to our ARC_FLAG_L2_WRITING
* bit being set. Let's just ensure that's being enforced.
*/
ASSERT(HDR_HAS_L1HDR(hdr));
/*
* Skipped - drop L2ARC entry and mark the header as no
* longer L2 eligibile.
*/
if (zio->io_error != 0) {
/*
* Error - drop L2ARC entry.
*/
list_remove(buflist, hdr);
arc_hdr_clear_flags(hdr, ARC_FLAG_HAS_L2HDR);
uint64_t psize = HDR_GET_PSIZE(hdr);
l2arc_hdr_arcstats_decrement(hdr);
bytes_dropped +=
vdev_psize_to_asize(dev->l2ad_vdev, psize);
(void) zfs_refcount_remove_many(&dev->l2ad_alloc,
arc_hdr_size(hdr), hdr);
}
/*
* Allow ARC to begin reads and ghost list evictions to
* this L2ARC entry.
*/
arc_hdr_clear_flags(hdr, ARC_FLAG_L2_WRITING);
mutex_exit(hash_lock);
}
/*
* Free the allocated abd buffers for writing the log blocks.
* If the zio failed reclaim the allocated space and remove the
* pointers to these log blocks from the log block pointer list
* of the L2ARC device.
*/
while ((abd_buf = list_remove_tail(&cb->l2wcb_abd_list)) != NULL) {
abd_free(abd_buf->abd);
zio_buf_free(abd_buf, sizeof (*abd_buf));
if (zio->io_error != 0) {
lb_ptr_buf = list_remove_head(&dev->l2ad_lbptr_list);
/*
* L2BLK_GET_PSIZE returns aligned size for log
* blocks.
*/
uint64_t asize =
L2BLK_GET_PSIZE((lb_ptr_buf->lb_ptr)->lbp_prop);
bytes_dropped += asize;
ARCSTAT_INCR(arcstat_l2_log_blk_asize, -asize);
ARCSTAT_BUMPDOWN(arcstat_l2_log_blk_count);
zfs_refcount_remove_many(&dev->l2ad_lb_asize, asize,
lb_ptr_buf);
zfs_refcount_remove(&dev->l2ad_lb_count, lb_ptr_buf);
kmem_free(lb_ptr_buf->lb_ptr,
sizeof (l2arc_log_blkptr_t));
kmem_free(lb_ptr_buf, sizeof (l2arc_lb_ptr_buf_t));
}
}
list_destroy(&cb->l2wcb_abd_list);
if (zio->io_error != 0) {
ARCSTAT_BUMP(arcstat_l2_writes_error);
/*
* Restore the lbps array in the header to its previous state.
* If the list of log block pointers is empty, zero out the
* log block pointers in the device header.
*/
lb_ptr_buf = list_head(&dev->l2ad_lbptr_list);
for (int i = 0; i < 2; i++) {
if (lb_ptr_buf == NULL) {
/*
* If the list is empty zero out the device
* header. Otherwise zero out the second log
* block pointer in the header.
*/
if (i == 0) {
memset(l2dhdr, 0,
dev->l2ad_dev_hdr_asize);
} else {
memset(&l2dhdr->dh_start_lbps[i], 0,
sizeof (l2arc_log_blkptr_t));
}
break;
}
memcpy(&l2dhdr->dh_start_lbps[i], lb_ptr_buf->lb_ptr,
sizeof (l2arc_log_blkptr_t));
lb_ptr_buf = list_next(&dev->l2ad_lbptr_list,
lb_ptr_buf);
}
}
ARCSTAT_BUMP(arcstat_l2_writes_done);
list_remove(buflist, head);
ASSERT(!HDR_HAS_L1HDR(head));
kmem_cache_free(hdr_l2only_cache, head);
mutex_exit(&dev->l2ad_mtx);
ASSERT(dev->l2ad_vdev != NULL);
vdev_space_update(dev->l2ad_vdev, -bytes_dropped, 0, 0);
l2arc_do_free_on_write();
kmem_free(cb, sizeof (l2arc_write_callback_t));
}
static int
l2arc_untransform(zio_t *zio, l2arc_read_callback_t *cb)
{
int ret;
spa_t *spa = zio->io_spa;
arc_buf_hdr_t *hdr = cb->l2rcb_hdr;
blkptr_t *bp = zio->io_bp;
uint8_t salt[ZIO_DATA_SALT_LEN];
uint8_t iv[ZIO_DATA_IV_LEN];
uint8_t mac[ZIO_DATA_MAC_LEN];
boolean_t no_crypt = B_FALSE;
/*
* ZIL data is never be written to the L2ARC, so we don't need
* special handling for its unique MAC storage.
*/
ASSERT3U(BP_GET_TYPE(bp), !=, DMU_OT_INTENT_LOG);
ASSERT(MUTEX_HELD(HDR_LOCK(hdr)));
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
/*
* If the data was encrypted, decrypt it now. Note that
* we must check the bp here and not the hdr, since the
* hdr does not have its encryption parameters updated
* until arc_read_done().
*/
if (BP_IS_ENCRYPTED(bp)) {
abd_t *eabd = arc_get_data_abd(hdr, arc_hdr_size(hdr), hdr,
ARC_HDR_USE_RESERVE);
zio_crypt_decode_params_bp(bp, salt, iv);
zio_crypt_decode_mac_bp(bp, mac);
ret = spa_do_crypt_abd(B_FALSE, spa, &cb->l2rcb_zb,
BP_GET_TYPE(bp), BP_GET_DEDUP(bp), BP_SHOULD_BYTESWAP(bp),
salt, iv, mac, HDR_GET_PSIZE(hdr), eabd,
hdr->b_l1hdr.b_pabd, &no_crypt);
if (ret != 0) {
arc_free_data_abd(hdr, eabd, arc_hdr_size(hdr), hdr);
goto error;
}
/*
* If we actually performed decryption, replace b_pabd
* with the decrypted data. Otherwise we can just throw
* our decryption buffer away.
*/
if (!no_crypt) {
arc_free_data_abd(hdr, hdr->b_l1hdr.b_pabd,
arc_hdr_size(hdr), hdr);
hdr->b_l1hdr.b_pabd = eabd;
zio->io_abd = eabd;
} else {
arc_free_data_abd(hdr, eabd, arc_hdr_size(hdr), hdr);
}
}
/*
* If the L2ARC block was compressed, but ARC compression
* is disabled we decompress the data into a new buffer and
* replace the existing data.
*/
if (HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr)) {
abd_t *cabd = arc_get_data_abd(hdr, arc_hdr_size(hdr), hdr,
ARC_HDR_USE_RESERVE);
void *tmp = abd_borrow_buf(cabd, arc_hdr_size(hdr));
ret = zio_decompress_data(HDR_GET_COMPRESS(hdr),
hdr->b_l1hdr.b_pabd, tmp, HDR_GET_PSIZE(hdr),
HDR_GET_LSIZE(hdr), &hdr->b_complevel);
if (ret != 0) {
abd_return_buf_copy(cabd, tmp, arc_hdr_size(hdr));
arc_free_data_abd(hdr, cabd, arc_hdr_size(hdr), hdr);
goto error;
}
abd_return_buf_copy(cabd, tmp, arc_hdr_size(hdr));
arc_free_data_abd(hdr, hdr->b_l1hdr.b_pabd,
arc_hdr_size(hdr), hdr);
hdr->b_l1hdr.b_pabd = cabd;
zio->io_abd = cabd;
zio->io_size = HDR_GET_LSIZE(hdr);
}
return (0);
error:
return (ret);
}
/*
* A read to a cache device completed. Validate buffer contents before
* handing over to the regular ARC routines.
*/
static void
l2arc_read_done(zio_t *zio)
{
int tfm_error = 0;
l2arc_read_callback_t *cb = zio->io_private;
arc_buf_hdr_t *hdr;
kmutex_t *hash_lock;
boolean_t valid_cksum;
boolean_t using_rdata = (BP_IS_ENCRYPTED(&cb->l2rcb_bp) &&
(cb->l2rcb_flags & ZIO_FLAG_RAW_ENCRYPT));
ASSERT3P(zio->io_vd, !=, NULL);
ASSERT(zio->io_flags & ZIO_FLAG_DONT_PROPAGATE);
spa_config_exit(zio->io_spa, SCL_L2ARC, zio->io_vd);
ASSERT3P(cb, !=, NULL);
hdr = cb->l2rcb_hdr;
ASSERT3P(hdr, !=, NULL);
hash_lock = HDR_LOCK(hdr);
mutex_enter(hash_lock);
ASSERT3P(hash_lock, ==, HDR_LOCK(hdr));
/*
* If the data was read into a temporary buffer,
* move it and free the buffer.
*/
if (cb->l2rcb_abd != NULL) {
ASSERT3U(arc_hdr_size(hdr), <, zio->io_size);
if (zio->io_error == 0) {
if (using_rdata) {
abd_copy(hdr->b_crypt_hdr.b_rabd,
cb->l2rcb_abd, arc_hdr_size(hdr));
} else {
abd_copy(hdr->b_l1hdr.b_pabd,
cb->l2rcb_abd, arc_hdr_size(hdr));
}
}
/*
* The following must be done regardless of whether
* there was an error:
* - free the temporary buffer
* - point zio to the real ARC buffer
* - set zio size accordingly
* These are required because zio is either re-used for
* an I/O of the block in the case of the error
* or the zio is passed to arc_read_done() and it
* needs real data.
*/
abd_free(cb->l2rcb_abd);
zio->io_size = zio->io_orig_size = arc_hdr_size(hdr);
if (using_rdata) {
ASSERT(HDR_HAS_RABD(hdr));
zio->io_abd = zio->io_orig_abd =
hdr->b_crypt_hdr.b_rabd;
} else {
ASSERT3P(hdr->b_l1hdr.b_pabd, !=, NULL);
zio->io_abd = zio->io_orig_abd = hdr->b_l1hdr.b_pabd;
}
}
ASSERT3P(zio->io_abd, !=, NULL);
/*
* Check this survived the L2ARC journey.
*/
ASSERT(zio->io_abd == hdr->b_l1hdr.b_pabd ||
(HDR_HAS_RABD(hdr) && zio->io_abd == hdr->b_crypt_hdr.b_rabd));
zio->io_bp_copy = cb->l2rcb_bp; /* XXX fix in L2ARC 2.0 */
zio->io_bp = &zio->io_bp_copy; /* XXX fix in L2ARC 2.0 */
zio->io_prop.zp_complevel = hdr->b_complevel;
valid_cksum = arc_cksum_is_equal(hdr, zio);
/*
* b_rabd will always match the data as it exists on disk if it is
* being used. Therefore if we are reading into b_rabd we do not
* attempt to untransform the data.
*/
if (valid_cksum && !using_rdata)
tfm_error = l2arc_untransform(zio, cb);
if (valid_cksum && tfm_error == 0 && zio->io_error == 0 &&
!HDR_L2_EVICTED(hdr)) {
mutex_exit(hash_lock);
zio->io_private = hdr;
arc_read_done(zio);
} else {
/*
* Buffer didn't survive caching. Increment stats and
* reissue to the original storage device.
*/
if (zio->io_error != 0) {
ARCSTAT_BUMP(arcstat_l2_io_error);
} else {
zio->io_error = SET_ERROR(EIO);
}
if (!valid_cksum || tfm_error != 0)
ARCSTAT_BUMP(arcstat_l2_cksum_bad);
/*
* If there's no waiter, issue an async i/o to the primary
* storage now. If there *is* a waiter, the caller must
* issue the i/o in a context where it's OK to block.
*/
if (zio->io_waiter == NULL) {
zio_t *pio = zio_unique_parent(zio);
void *abd = (using_rdata) ?
hdr->b_crypt_hdr.b_rabd : hdr->b_l1hdr.b_pabd;
ASSERT(!pio || pio->io_child_type == ZIO_CHILD_LOGICAL);
zio = zio_read(pio, zio->io_spa, zio->io_bp,
abd, zio->io_size, arc_read_done,
hdr, zio->io_priority, cb->l2rcb_flags,
&cb->l2rcb_zb);
/*
* Original ZIO will be freed, so we need to update
* ARC header with the new ZIO pointer to be used
* by zio_change_priority() in arc_read().
*/
for (struct arc_callback *acb = hdr->b_l1hdr.b_acb;
acb != NULL; acb = acb->acb_next)
acb->acb_zio_head = zio;
mutex_exit(hash_lock);
zio_nowait(zio);
} else {
mutex_exit(hash_lock);
}
}
kmem_free(cb, sizeof (l2arc_read_callback_t));
}
/*
* This is the list priority from which the L2ARC will search for pages to
* cache. This is used within loops (0..3) to cycle through lists in the
* desired order. This order can have a significant effect on cache
* performance.
*
* Currently the metadata lists are hit first, MFU then MRU, followed by
* the data lists. This function returns a locked list, and also returns
* the lock pointer.
*/
static multilist_sublist_t *
l2arc_sublist_lock(int list_num)
{
multilist_t *ml = NULL;
unsigned int idx;
ASSERT(list_num >= 0 && list_num < L2ARC_FEED_TYPES);
switch (list_num) {
case 0:
ml = &arc_mfu->arcs_list[ARC_BUFC_METADATA];
break;
case 1:
ml = &arc_mru->arcs_list[ARC_BUFC_METADATA];
break;
case 2:
ml = &arc_mfu->arcs_list[ARC_BUFC_DATA];
break;
case 3:
ml = &arc_mru->arcs_list[ARC_BUFC_DATA];
break;
default:
return (NULL);
}
/*
* Return a randomly-selected sublist. This is acceptable
* because the caller feeds only a little bit of data for each
* call (8MB). Subsequent calls will result in different
* sublists being selected.
*/
idx = multilist_get_random_index(ml);
return (multilist_sublist_lock(ml, idx));
}
/*
* Calculates the maximum overhead of L2ARC metadata log blocks for a given
* L2ARC write size. l2arc_evict and l2arc_write_size need to include this
* overhead in processing to make sure there is enough headroom available
* when writing buffers.
*/
static inline uint64_t
l2arc_log_blk_overhead(uint64_t write_sz, l2arc_dev_t *dev)
{
if (dev->l2ad_log_entries == 0) {
return (0);
} else {
uint64_t log_entries = write_sz >> SPA_MINBLOCKSHIFT;
uint64_t log_blocks = (log_entries +
dev->l2ad_log_entries - 1) /
dev->l2ad_log_entries;
return (vdev_psize_to_asize(dev->l2ad_vdev,
sizeof (l2arc_log_blk_phys_t)) * log_blocks);
}
}
/*
* Evict buffers from the device write hand to the distance specified in
* bytes. This distance may span populated buffers, it may span nothing.
* This is clearing a region on the L2ARC device ready for writing.
* If the 'all' boolean is set, every buffer is evicted.
*/
static void
l2arc_evict(l2arc_dev_t *dev, uint64_t distance, boolean_t all)
{
list_t *buflist;
arc_buf_hdr_t *hdr, *hdr_prev;
kmutex_t *hash_lock;
uint64_t taddr;
l2arc_lb_ptr_buf_t *lb_ptr_buf, *lb_ptr_buf_prev;
vdev_t *vd = dev->l2ad_vdev;
boolean_t rerun;
buflist = &dev->l2ad_buflist;
/*
* We need to add in the worst case scenario of log block overhead.
*/
distance += l2arc_log_blk_overhead(distance, dev);
if (vd->vdev_has_trim && l2arc_trim_ahead > 0) {
/*
* Trim ahead of the write size 64MB or (l2arc_trim_ahead/100)
* times the write size, whichever is greater.
*/
distance += MAX(64 * 1024 * 1024,
(distance * l2arc_trim_ahead) / 100);
}
top:
rerun = B_FALSE;
if (dev->l2ad_hand >= (dev->l2ad_end - distance)) {
/*
* When there is no space to accommodate upcoming writes,
* evict to the end. Then bump the write and evict hands
* to the start and iterate. This iteration does not
* happen indefinitely as we make sure in
* l2arc_write_size() that when the write hand is reset,
* the write size does not exceed the end of the device.
*/
rerun = B_TRUE;
taddr = dev->l2ad_end;
} else {
taddr = dev->l2ad_hand + distance;
}
DTRACE_PROBE4(l2arc__evict, l2arc_dev_t *, dev, list_t *, buflist,
uint64_t, taddr, boolean_t, all);
if (!all) {
/*
* This check has to be placed after deciding whether to
* iterate (rerun).
*/
if (dev->l2ad_first) {
/*
* This is the first sweep through the device. There is
* nothing to evict. We have already trimmmed the
* whole device.
*/
goto out;
} else {
/*
* Trim the space to be evicted.
*/
if (vd->vdev_has_trim && dev->l2ad_evict < taddr &&
l2arc_trim_ahead > 0) {
/*
* We have to drop the spa_config lock because
* vdev_trim_range() will acquire it.
* l2ad_evict already accounts for the label
* size. To prevent vdev_trim_ranges() from
* adding it again, we subtract it from
* l2ad_evict.
*/
spa_config_exit(dev->l2ad_spa, SCL_L2ARC, dev);
vdev_trim_simple(vd,
dev->l2ad_evict - VDEV_LABEL_START_SIZE,
taddr - dev->l2ad_evict);
spa_config_enter(dev->l2ad_spa, SCL_L2ARC, dev,
RW_READER);
}
/*
* When rebuilding L2ARC we retrieve the evict hand
* from the header of the device. Of note, l2arc_evict()
* does not actually delete buffers from the cache
* device, but trimming may do so depending on the
* hardware implementation. Thus keeping track of the
* evict hand is useful.
*/
dev->l2ad_evict = MAX(dev->l2ad_evict, taddr);
}
}
retry:
mutex_enter(&dev->l2ad_mtx);
/*
* We have to account for evicted log blocks. Run vdev_space_update()
* on log blocks whose offset (in bytes) is before the evicted offset
* (in bytes) by searching in the list of pointers to log blocks
* present in the L2ARC device.
*/
for (lb_ptr_buf = list_tail(&dev->l2ad_lbptr_list); lb_ptr_buf;
lb_ptr_buf = lb_ptr_buf_prev) {
lb_ptr_buf_prev = list_prev(&dev->l2ad_lbptr_list, lb_ptr_buf);
/* L2BLK_GET_PSIZE returns aligned size for log blocks */
uint64_t asize = L2BLK_GET_PSIZE(
(lb_ptr_buf->lb_ptr)->lbp_prop);
/*
* We don't worry about log blocks left behind (ie
* lbp_payload_start < l2ad_hand) because l2arc_write_buffers()
* will never write more than l2arc_evict() evicts.
*/
if (!all && l2arc_log_blkptr_valid(dev, lb_ptr_buf->lb_ptr)) {
break;
} else {
vdev_space_update(vd, -asize, 0, 0);
ARCSTAT_INCR(arcstat_l2_log_blk_asize, -asize);
ARCSTAT_BUMPDOWN(arcstat_l2_log_blk_count);
zfs_refcount_remove_many(&dev->l2ad_lb_asize, asize,
lb_ptr_buf);
zfs_refcount_remove(&dev->l2ad_lb_count, lb_ptr_buf);
list_remove(&dev->l2ad_lbptr_list, lb_ptr_buf);
kmem_free(lb_ptr_buf->lb_ptr,
sizeof (l2arc_log_blkptr_t));
kmem_free(lb_ptr_buf, sizeof (l2arc_lb_ptr_buf_t));
}
}
for (hdr = list_tail(buflist); hdr; hdr = hdr_prev) {
hdr_prev = list_prev(buflist, hdr);
ASSERT(!HDR_EMPTY(hdr));
hash_lock = HDR_LOCK(hdr);
/*
* We cannot use mutex_enter or else we can deadlock
* with l2arc_write_buffers (due to swapping the order
* the hash lock and l2ad_mtx are taken).
*/
if (!mutex_tryenter(hash_lock)) {
/*
* Missed the hash lock. Retry.
*/
ARCSTAT_BUMP(arcstat_l2_evict_lock_retry);
mutex_exit(&dev->l2ad_mtx);
mutex_enter(hash_lock);
mutex_exit(hash_lock);
goto retry;
}
/*
* A header can't be on this list if it doesn't have L2 header.
*/
ASSERT(HDR_HAS_L2HDR(hdr));
/* Ensure this header has finished being written. */
ASSERT(!HDR_L2_WRITING(hdr));
ASSERT(!HDR_L2_WRITE_HEAD(hdr));
if (!all && (hdr->b_l2hdr.b_daddr >= dev->l2ad_evict ||
hdr->b_l2hdr.b_daddr < dev->l2ad_hand)) {
/*
* We've evicted to the target address,
* or the end of the device.
*/
mutex_exit(hash_lock);
break;
}
if (!HDR_HAS_L1HDR(hdr)) {
ASSERT(!HDR_L2_READING(hdr));
/*
* This doesn't exist in the ARC. Destroy.
* arc_hdr_destroy() will call list_remove()
* and decrement arcstat_l2_lsize.
*/
arc_change_state(arc_anon, hdr);
arc_hdr_destroy(hdr);
} else {
ASSERT(hdr->b_l1hdr.b_state != arc_l2c_only);
ARCSTAT_BUMP(arcstat_l2_evict_l1cached);
/*
* Invalidate issued or about to be issued
* reads, since we may be about to write
* over this location.
*/
if (HDR_L2_READING(hdr)) {
ARCSTAT_BUMP(arcstat_l2_evict_reading);
arc_hdr_set_flags(hdr, ARC_FLAG_L2_EVICTED);
}
arc_hdr_l2hdr_destroy(hdr);
}
mutex_exit(hash_lock);
}
mutex_exit(&dev->l2ad_mtx);
out:
/*
* We need to check if we evict all buffers, otherwise we may iterate
* unnecessarily.
*/
if (!all && rerun) {
/*
* Bump device hand to the device start if it is approaching the
* end. l2arc_evict() has already evicted ahead for this case.
*/
dev->l2ad_hand = dev->l2ad_start;
dev->l2ad_evict = dev->l2ad_start;
dev->l2ad_first = B_FALSE;
goto top;
}
if (!all) {
/*
* In case of cache device removal (all) the following
* assertions may be violated without functional consequences
* as the device is about to be removed.
*/
ASSERT3U(dev->l2ad_hand + distance, <, dev->l2ad_end);
if (!dev->l2ad_first)
ASSERT3U(dev->l2ad_hand, <, dev->l2ad_evict);
}
}
/*
* Handle any abd transforms that might be required for writing to the L2ARC.
* If successful, this function will always return an abd with the data
* transformed as it is on disk in a new abd of asize bytes.
*/
static int
l2arc_apply_transforms(spa_t *spa, arc_buf_hdr_t *hdr, uint64_t asize,
abd_t **abd_out)
{
int ret;
void *tmp = NULL;
abd_t *cabd = NULL, *eabd = NULL, *to_write = hdr->b_l1hdr.b_pabd;
enum zio_compress compress = HDR_GET_COMPRESS(hdr);
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t size = arc_hdr_size(hdr);
boolean_t ismd = HDR_ISTYPE_METADATA(hdr);
boolean_t bswap = (hdr->b_l1hdr.b_byteswap != DMU_BSWAP_NUMFUNCS);
dsl_crypto_key_t *dck = NULL;
uint8_t mac[ZIO_DATA_MAC_LEN] = { 0 };
boolean_t no_crypt = B_FALSE;
ASSERT((HDR_GET_COMPRESS(hdr) != ZIO_COMPRESS_OFF &&
!HDR_COMPRESSION_ENABLED(hdr)) ||
HDR_ENCRYPTED(hdr) || HDR_SHARED_DATA(hdr) || psize != asize);
ASSERT3U(psize, <=, asize);
/*
* If this data simply needs its own buffer, we simply allocate it
* and copy the data. This may be done to eliminate a dependency on a
* shared buffer or to reallocate the buffer to match asize.
*/
if (HDR_HAS_RABD(hdr) && asize != psize) {
ASSERT3U(asize, >=, psize);
to_write = abd_alloc_for_io(asize, ismd);
abd_copy(to_write, hdr->b_crypt_hdr.b_rabd, psize);
if (psize != asize)
abd_zero_off(to_write, psize, asize - psize);
goto out;
}
if ((compress == ZIO_COMPRESS_OFF || HDR_COMPRESSION_ENABLED(hdr)) &&
!HDR_ENCRYPTED(hdr)) {
ASSERT3U(size, ==, psize);
to_write = abd_alloc_for_io(asize, ismd);
abd_copy(to_write, hdr->b_l1hdr.b_pabd, size);
if (size != asize)
abd_zero_off(to_write, size, asize - size);
goto out;
}
if (compress != ZIO_COMPRESS_OFF && !HDR_COMPRESSION_ENABLED(hdr)) {
/*
* In some cases, we can wind up with size > asize, so
* we need to opt for the larger allocation option here.
*
* (We also need abd_return_buf_copy in all cases because
* it's an ASSERT() to modify the buffer before returning it
* with arc_return_buf(), and all the compressors
* write things before deciding to fail compression in nearly
* every case.)
*/
cabd = abd_alloc_for_io(size, ismd);
tmp = abd_borrow_buf(cabd, size);
psize = zio_compress_data(compress, to_write, &tmp, size,
hdr->b_complevel);
if (psize >= asize) {
psize = HDR_GET_PSIZE(hdr);
abd_return_buf_copy(cabd, tmp, size);
HDR_SET_COMPRESS(hdr, ZIO_COMPRESS_OFF);
to_write = cabd;
abd_copy(to_write, hdr->b_l1hdr.b_pabd, psize);
if (psize != asize)
abd_zero_off(to_write, psize, asize - psize);
goto encrypt;
}
ASSERT3U(psize, <=, HDR_GET_PSIZE(hdr));
if (psize < asize)
memset((char *)tmp + psize, 0, asize - psize);
psize = HDR_GET_PSIZE(hdr);
abd_return_buf_copy(cabd, tmp, size);
to_write = cabd;
}
encrypt:
if (HDR_ENCRYPTED(hdr)) {
eabd = abd_alloc_for_io(asize, ismd);
/*
* If the dataset was disowned before the buffer
* made it to this point, the key to re-encrypt
* it won't be available. In this case we simply
* won't write the buffer to the L2ARC.
*/
ret = spa_keystore_lookup_key(spa, hdr->b_crypt_hdr.b_dsobj,
FTAG, &dck);
if (ret != 0)
goto error;
ret = zio_do_crypt_abd(B_TRUE, &dck->dck_key,
hdr->b_crypt_hdr.b_ot, bswap, hdr->b_crypt_hdr.b_salt,
hdr->b_crypt_hdr.b_iv, mac, psize, to_write, eabd,
&no_crypt);
if (ret != 0)
goto error;
if (no_crypt)
abd_copy(eabd, to_write, psize);
if (psize != asize)
abd_zero_off(eabd, psize, asize - psize);
/* assert that the MAC we got here matches the one we saved */
ASSERT0(memcmp(mac, hdr->b_crypt_hdr.b_mac, ZIO_DATA_MAC_LEN));
spa_keystore_dsl_key_rele(spa, dck, FTAG);
if (to_write == cabd)
abd_free(cabd);
to_write = eabd;
}
out:
ASSERT3P(to_write, !=, hdr->b_l1hdr.b_pabd);
*abd_out = to_write;
return (0);
error:
if (dck != NULL)
spa_keystore_dsl_key_rele(spa, dck, FTAG);
if (cabd != NULL)
abd_free(cabd);
if (eabd != NULL)
abd_free(eabd);
*abd_out = NULL;
return (ret);
}
static void
l2arc_blk_fetch_done(zio_t *zio)
{
l2arc_read_callback_t *cb;
cb = zio->io_private;
if (cb->l2rcb_abd != NULL)
abd_free(cb->l2rcb_abd);
kmem_free(cb, sizeof (l2arc_read_callback_t));
}
/*
* Find and write ARC buffers to the L2ARC device.
*
* An ARC_FLAG_L2_WRITING flag is set so that the L2ARC buffers are not valid
* for reading until they have completed writing.
* The headroom_boost is an in-out parameter used to maintain headroom boost
* state between calls to this function.
*
* Returns the number of bytes actually written (which may be smaller than
* the delta by which the device hand has changed due to alignment and the
* writing of log blocks).
*/
static uint64_t
l2arc_write_buffers(spa_t *spa, l2arc_dev_t *dev, uint64_t target_sz)
{
arc_buf_hdr_t *hdr, *hdr_prev, *head;
uint64_t write_asize, write_psize, write_lsize, headroom;
boolean_t full;
l2arc_write_callback_t *cb = NULL;
zio_t *pio, *wzio;
uint64_t guid = spa_load_guid(spa);
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
ASSERT3P(dev->l2ad_vdev, !=, NULL);
pio = NULL;
write_lsize = write_asize = write_psize = 0;
full = B_FALSE;
head = kmem_cache_alloc(hdr_l2only_cache, KM_PUSHPAGE);
arc_hdr_set_flags(head, ARC_FLAG_L2_WRITE_HEAD | ARC_FLAG_HAS_L2HDR);
/*
* Copy buffers for L2ARC writing.
*/
for (int pass = 0; pass < L2ARC_FEED_TYPES; pass++) {
/*
* If pass == 1 or 3, we cache MRU metadata and data
* respectively.
*/
if (l2arc_mfuonly) {
if (pass == 1 || pass == 3)
continue;
}
multilist_sublist_t *mls = l2arc_sublist_lock(pass);
uint64_t passed_sz = 0;
VERIFY3P(mls, !=, NULL);
/*
* L2ARC fast warmup.
*
* Until the ARC is warm and starts to evict, read from the
* head of the ARC lists rather than the tail.
*/
if (arc_warm == B_FALSE)
hdr = multilist_sublist_head(mls);
else
hdr = multilist_sublist_tail(mls);
headroom = target_sz * l2arc_headroom;
if (zfs_compressed_arc_enabled)
headroom = (headroom * l2arc_headroom_boost) / 100;
for (; hdr; hdr = hdr_prev) {
kmutex_t *hash_lock;
abd_t *to_write = NULL;
if (arc_warm == B_FALSE)
hdr_prev = multilist_sublist_next(mls, hdr);
else
hdr_prev = multilist_sublist_prev(mls, hdr);
hash_lock = HDR_LOCK(hdr);
if (!mutex_tryenter(hash_lock)) {
/*
* Skip this buffer rather than waiting.
*/
continue;
}
passed_sz += HDR_GET_LSIZE(hdr);
if (l2arc_headroom != 0 && passed_sz > headroom) {
/*
* Searched too far.
*/
mutex_exit(hash_lock);
break;
}
if (!l2arc_write_eligible(guid, hdr)) {
mutex_exit(hash_lock);
continue;
}
ASSERT(HDR_HAS_L1HDR(hdr));
ASSERT3U(HDR_GET_PSIZE(hdr), >, 0);
ASSERT3U(arc_hdr_size(hdr), >, 0);
ASSERT(hdr->b_l1hdr.b_pabd != NULL ||
HDR_HAS_RABD(hdr));
uint64_t psize = HDR_GET_PSIZE(hdr);
uint64_t asize = vdev_psize_to_asize(dev->l2ad_vdev,
psize);
if ((write_asize + asize) > target_sz) {
full = B_TRUE;
mutex_exit(hash_lock);
break;
}
/*
* We rely on the L1 portion of the header below, so
* it's invalid for this header to have been evicted out
* of the ghost cache, prior to being written out. The
* ARC_FLAG_L2_WRITING bit ensures this won't happen.
*/
arc_hdr_set_flags(hdr, ARC_FLAG_L2_WRITING);
/*
* If this header has b_rabd, we can use this since it
* must always match the data exactly as it exists on
* disk. Otherwise, the L2ARC can normally use the
* hdr's data, but if we're sharing data between the
* hdr and one of its bufs, L2ARC needs its own copy of
* the data so that the ZIO below can't race with the
* buf consumer. To ensure that this copy will be
* available for the lifetime of the ZIO and be cleaned
* up afterwards, we add it to the l2arc_free_on_write
* queue. If we need to apply any transforms to the
* data (compression, encryption) we will also need the
* extra buffer.
*/
if (HDR_HAS_RABD(hdr) && psize == asize) {
to_write = hdr->b_crypt_hdr.b_rabd;
} else if ((HDR_COMPRESSION_ENABLED(hdr) ||
HDR_GET_COMPRESS(hdr) == ZIO_COMPRESS_OFF) &&
!HDR_ENCRYPTED(hdr) && !HDR_SHARED_DATA(hdr) &&
psize == asize) {
to_write = hdr->b_l1hdr.b_pabd;
} else {
int ret;
arc_buf_contents_t type = arc_buf_type(hdr);
ret = l2arc_apply_transforms(spa, hdr, asize,
&to_write);
if (ret != 0) {
arc_hdr_clear_flags(hdr,
ARC_FLAG_L2_WRITING);
mutex_exit(hash_lock);
continue;
}
l2arc_free_abd_on_write(to_write, asize, type);
}
if (pio == NULL) {
/*
* Insert a dummy header on the buflist so
* l2arc_write_done() can find where the
* write buffers begin without searching.
*/
mutex_enter(&dev->l2ad_mtx);
list_insert_head(&dev->l2ad_buflist, head);
mutex_exit(&dev->l2ad_mtx);
cb = kmem_alloc(
sizeof (l2arc_write_callback_t), KM_SLEEP);
cb->l2wcb_dev = dev;
cb->l2wcb_head = head;
/*
* Create a list to save allocated abd buffers
* for l2arc_log_blk_commit().
*/
list_create(&cb->l2wcb_abd_list,
sizeof (l2arc_lb_abd_buf_t),
offsetof(l2arc_lb_abd_buf_t, node));
pio = zio_root(spa, l2arc_write_done, cb,
ZIO_FLAG_CANFAIL);
}
hdr->b_l2hdr.b_dev = dev;
hdr->b_l2hdr.b_hits = 0;
hdr->b_l2hdr.b_daddr = dev->l2ad_hand;
hdr->b_l2hdr.b_arcs_state =
hdr->b_l1hdr.b_state->arcs_state;
arc_hdr_set_flags(hdr, ARC_FLAG_HAS_L2HDR);
mutex_enter(&dev->l2ad_mtx);
list_insert_head(&dev->l2ad_buflist, hdr);
mutex_exit(&dev->l2ad_mtx);
(void) zfs_refcount_add_many(&dev->l2ad_alloc,
arc_hdr_size(hdr), hdr);
wzio = zio_write_phys(pio, dev->l2ad_vdev,
hdr->b_l2hdr.b_daddr, asize, to_write,
ZIO_CHECKSUM_OFF, NULL, hdr,
ZIO_PRIORITY_ASYNC_WRITE,
ZIO_FLAG_CANFAIL, B_FALSE);
write_lsize += HDR_GET_LSIZE(hdr);
DTRACE_PROBE2(l2arc__write, vdev_t *, dev->l2ad_vdev,
zio_t *, wzio);
write_psize += psize;
write_asize += asize;
dev->l2ad_hand += asize;
l2arc_hdr_arcstats_increment(hdr);
vdev_space_update(dev->l2ad_vdev, asize, 0, 0);
mutex_exit(hash_lock);
/*
* Append buf info to current log and commit if full.
* arcstat_l2_{size,asize} kstats are updated
* internally.
*/
if (l2arc_log_blk_insert(dev, hdr))
l2arc_log_blk_commit(dev, pio, cb);
zio_nowait(wzio);
}
multilist_sublist_unlock(mls);
if (full == B_TRUE)
break;
}
/* No buffers selected for writing? */
if (pio == NULL) {
ASSERT0(write_lsize);
ASSERT(!HDR_HAS_L1HDR(head));
kmem_cache_free(hdr_l2only_cache, head);
/*
* Although we did not write any buffers l2ad_evict may
* have advanced.
*/
if (dev->l2ad_evict != l2dhdr->dh_evict)
l2arc_dev_hdr_update(dev);
return (0);
}
if (!dev->l2ad_first)
ASSERT3U(dev->l2ad_hand, <=, dev->l2ad_evict);
ASSERT3U(write_asize, <=, target_sz);
ARCSTAT_BUMP(arcstat_l2_writes_sent);
ARCSTAT_INCR(arcstat_l2_write_bytes, write_psize);
dev->l2ad_writing = B_TRUE;
(void) zio_wait(pio);
dev->l2ad_writing = B_FALSE;
/*
* Update the device header after the zio completes as
* l2arc_write_done() may have updated the memory holding the log block
* pointers in the device header.
*/
l2arc_dev_hdr_update(dev);
return (write_asize);
}
static boolean_t
l2arc_hdr_limit_reached(void)
{
int64_t s = aggsum_upper_bound(&arc_sums.arcstat_l2_hdr_size);
return (arc_reclaim_needed() ||
(s > (arc_warm ? arc_c : arc_c_max) * l2arc_meta_percent / 100));
}
/*
* This thread feeds the L2ARC at regular intervals. This is the beating
* heart of the L2ARC.
*/
static __attribute__((noreturn)) void
l2arc_feed_thread(void *unused)
{
(void) unused;
callb_cpr_t cpr;
l2arc_dev_t *dev;
spa_t *spa;
uint64_t size, wrote;
clock_t begin, next = ddi_get_lbolt();
fstrans_cookie_t cookie;
CALLB_CPR_INIT(&cpr, &l2arc_feed_thr_lock, callb_generic_cpr, FTAG);
mutex_enter(&l2arc_feed_thr_lock);
cookie = spl_fstrans_mark();
while (l2arc_thread_exit == 0) {
CALLB_CPR_SAFE_BEGIN(&cpr);
(void) cv_timedwait_idle(&l2arc_feed_thr_cv,
&l2arc_feed_thr_lock, next);
CALLB_CPR_SAFE_END(&cpr, &l2arc_feed_thr_lock);
next = ddi_get_lbolt() + hz;
/*
* Quick check for L2ARC devices.
*/
mutex_enter(&l2arc_dev_mtx);
if (l2arc_ndev == 0) {
mutex_exit(&l2arc_dev_mtx);
continue;
}
mutex_exit(&l2arc_dev_mtx);
begin = ddi_get_lbolt();
/*
* This selects the next l2arc device to write to, and in
* doing so the next spa to feed from: dev->l2ad_spa. This
* will return NULL if there are now no l2arc devices or if
* they are all faulted.
*
* If a device is returned, its spa's config lock is also
* held to prevent device removal. l2arc_dev_get_next()
* will grab and release l2arc_dev_mtx.
*/
if ((dev = l2arc_dev_get_next()) == NULL)
continue;
spa = dev->l2ad_spa;
ASSERT3P(spa, !=, NULL);
/*
* If the pool is read-only then force the feed thread to
* sleep a little longer.
*/
if (!spa_writeable(spa)) {
next = ddi_get_lbolt() + 5 * l2arc_feed_secs * hz;
spa_config_exit(spa, SCL_L2ARC, dev);
continue;
}
/*
* Avoid contributing to memory pressure.
*/
if (l2arc_hdr_limit_reached()) {
ARCSTAT_BUMP(arcstat_l2_abort_lowmem);
spa_config_exit(spa, SCL_L2ARC, dev);
continue;
}
ARCSTAT_BUMP(arcstat_l2_feeds);
size = l2arc_write_size(dev);
/*
* Evict L2ARC buffers that will be overwritten.
*/
l2arc_evict(dev, size, B_FALSE);
/*
* Write ARC buffers.
*/
wrote = l2arc_write_buffers(spa, dev, size);
/*
* Calculate interval between writes.
*/
next = l2arc_write_interval(begin, size, wrote);
spa_config_exit(spa, SCL_L2ARC, dev);
}
spl_fstrans_unmark(cookie);
l2arc_thread_exit = 0;
cv_broadcast(&l2arc_feed_thr_cv);
CALLB_CPR_EXIT(&cpr); /* drops l2arc_feed_thr_lock */
thread_exit();
}
boolean_t
l2arc_vdev_present(vdev_t *vd)
{
return (l2arc_vdev_get(vd) != NULL);
}
/*
* Returns the l2arc_dev_t associated with a particular vdev_t or NULL if
* the vdev_t isn't an L2ARC device.
*/
l2arc_dev_t *
l2arc_vdev_get(vdev_t *vd)
{
l2arc_dev_t *dev;
mutex_enter(&l2arc_dev_mtx);
for (dev = list_head(l2arc_dev_list); dev != NULL;
dev = list_next(l2arc_dev_list, dev)) {
if (dev->l2ad_vdev == vd)
break;
}
mutex_exit(&l2arc_dev_mtx);
return (dev);
}
static void
l2arc_rebuild_dev(l2arc_dev_t *dev, boolean_t reopen)
{
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
uint64_t l2dhdr_asize = dev->l2ad_dev_hdr_asize;
spa_t *spa = dev->l2ad_spa;
/*
* The L2ARC has to hold at least the payload of one log block for
* them to be restored (persistent L2ARC). The payload of a log block
* depends on the amount of its log entries. We always write log blocks
* with 1022 entries. How many of them are committed or restored depends
* on the size of the L2ARC device. Thus the maximum payload of
* one log block is 1022 * SPA_MAXBLOCKSIZE = 16GB. If the L2ARC device
* is less than that, we reduce the amount of committed and restored
* log entries per block so as to enable persistence.
*/
if (dev->l2ad_end < l2arc_rebuild_blocks_min_l2size) {
dev->l2ad_log_entries = 0;
} else {
dev->l2ad_log_entries = MIN((dev->l2ad_end -
dev->l2ad_start) >> SPA_MAXBLOCKSHIFT,
L2ARC_LOG_BLK_MAX_ENTRIES);
}
/*
* Read the device header, if an error is returned do not rebuild L2ARC.
*/
if (l2arc_dev_hdr_read(dev) == 0 && dev->l2ad_log_entries > 0) {
/*
* If we are onlining a cache device (vdev_reopen) that was
* still present (l2arc_vdev_present()) and rebuild is enabled,
* we should evict all ARC buffers and pointers to log blocks
* and reclaim their space before restoring its contents to
* L2ARC.
*/
if (reopen) {
if (!l2arc_rebuild_enabled) {
return;
} else {
l2arc_evict(dev, 0, B_TRUE);
/* start a new log block */
dev->l2ad_log_ent_idx = 0;
dev->l2ad_log_blk_payload_asize = 0;
dev->l2ad_log_blk_payload_start = 0;
}
}
/*
* Just mark the device as pending for a rebuild. We won't
* be starting a rebuild in line here as it would block pool
* import. Instead spa_load_impl will hand that off to an
* async task which will call l2arc_spa_rebuild_start.
*/
dev->l2ad_rebuild = B_TRUE;
} else if (spa_writeable(spa)) {
/*
* In this case TRIM the whole device if l2arc_trim_ahead > 0,
* otherwise create a new header. We zero out the memory holding
* the header to reset dh_start_lbps. If we TRIM the whole
* device the new header will be written by
* vdev_trim_l2arc_thread() at the end of the TRIM to update the
* trim_state in the header too. When reading the header, if
* trim_state is not VDEV_TRIM_COMPLETE and l2arc_trim_ahead > 0
* we opt to TRIM the whole device again.
*/
if (l2arc_trim_ahead > 0) {
dev->l2ad_trim_all = B_TRUE;
} else {
memset(l2dhdr, 0, l2dhdr_asize);
l2arc_dev_hdr_update(dev);
}
}
}
/*
* Add a vdev for use by the L2ARC. By this point the spa has already
* validated the vdev and opened it.
*/
void
l2arc_add_vdev(spa_t *spa, vdev_t *vd)
{
l2arc_dev_t *adddev;
uint64_t l2dhdr_asize;
ASSERT(!l2arc_vdev_present(vd));
/*
* Create a new l2arc device entry.
*/
adddev = vmem_zalloc(sizeof (l2arc_dev_t), KM_SLEEP);
adddev->l2ad_spa = spa;
adddev->l2ad_vdev = vd;
/* leave extra size for an l2arc device header */
l2dhdr_asize = adddev->l2ad_dev_hdr_asize =
MAX(sizeof (*adddev->l2ad_dev_hdr), 1 << vd->vdev_ashift);
adddev->l2ad_start = VDEV_LABEL_START_SIZE + l2dhdr_asize;
adddev->l2ad_end = VDEV_LABEL_START_SIZE + vdev_get_min_asize(vd);
ASSERT3U(adddev->l2ad_start, <, adddev->l2ad_end);
adddev->l2ad_hand = adddev->l2ad_start;
adddev->l2ad_evict = adddev->l2ad_start;
adddev->l2ad_first = B_TRUE;
adddev->l2ad_writing = B_FALSE;
adddev->l2ad_trim_all = B_FALSE;
list_link_init(&adddev->l2ad_node);
adddev->l2ad_dev_hdr = kmem_zalloc(l2dhdr_asize, KM_SLEEP);
mutex_init(&adddev->l2ad_mtx, NULL, MUTEX_DEFAULT, NULL);
/*
* This is a list of all ARC buffers that are still valid on the
* device.
*/
list_create(&adddev->l2ad_buflist, sizeof (arc_buf_hdr_t),
offsetof(arc_buf_hdr_t, b_l2hdr.b_l2node));
/*
* This is a list of pointers to log blocks that are still present
* on the device.
*/
list_create(&adddev->l2ad_lbptr_list, sizeof (l2arc_lb_ptr_buf_t),
offsetof(l2arc_lb_ptr_buf_t, node));
vdev_space_update(vd, 0, 0, adddev->l2ad_end - adddev->l2ad_hand);
zfs_refcount_create(&adddev->l2ad_alloc);
zfs_refcount_create(&adddev->l2ad_lb_asize);
zfs_refcount_create(&adddev->l2ad_lb_count);
/*
* Decide if dev is eligible for L2ARC rebuild or whole device
* trimming. This has to happen before the device is added in the
* cache device list and l2arc_dev_mtx is released. Otherwise
* l2arc_feed_thread() might already start writing on the
* device.
*/
l2arc_rebuild_dev(adddev, B_FALSE);
/*
* Add device to global list
*/
mutex_enter(&l2arc_dev_mtx);
list_insert_head(l2arc_dev_list, adddev);
atomic_inc_64(&l2arc_ndev);
mutex_exit(&l2arc_dev_mtx);
}
/*
* Decide if a vdev is eligible for L2ARC rebuild, called from vdev_reopen()
* in case of onlining a cache device.
*/
void
l2arc_rebuild_vdev(vdev_t *vd, boolean_t reopen)
{
l2arc_dev_t *dev = NULL;
dev = l2arc_vdev_get(vd);
ASSERT3P(dev, !=, NULL);
/*
* In contrast to l2arc_add_vdev() we do not have to worry about
* l2arc_feed_thread() invalidating previous content when onlining a
* cache device. The device parameters (l2ad*) are not cleared when
* offlining the device and writing new buffers will not invalidate
* all previous content. In worst case only buffers that have not had
* their log block written to the device will be lost.
* When onlining the cache device (ie offline->online without exporting
* the pool in between) this happens:
* vdev_reopen() -> vdev_open() -> l2arc_rebuild_vdev()
* | |
* vdev_is_dead() = B_FALSE l2ad_rebuild = B_TRUE
* During the time where vdev_is_dead = B_FALSE and until l2ad_rebuild
* is set to B_TRUE we might write additional buffers to the device.
*/
l2arc_rebuild_dev(dev, reopen);
}
/*
* Remove a vdev from the L2ARC.
*/
void
l2arc_remove_vdev(vdev_t *vd)
{
l2arc_dev_t *remdev = NULL;
/*
* Find the device by vdev
*/
remdev = l2arc_vdev_get(vd);
ASSERT3P(remdev, !=, NULL);
/*
* Cancel any ongoing or scheduled rebuild.
*/
mutex_enter(&l2arc_rebuild_thr_lock);
if (remdev->l2ad_rebuild_began == B_TRUE) {
remdev->l2ad_rebuild_cancel = B_TRUE;
while (remdev->l2ad_rebuild == B_TRUE)
cv_wait(&l2arc_rebuild_thr_cv, &l2arc_rebuild_thr_lock);
}
mutex_exit(&l2arc_rebuild_thr_lock);
/*
* Remove device from global list
*/
mutex_enter(&l2arc_dev_mtx);
list_remove(l2arc_dev_list, remdev);
l2arc_dev_last = NULL; /* may have been invalidated */
atomic_dec_64(&l2arc_ndev);
mutex_exit(&l2arc_dev_mtx);
/*
* Clear all buflists and ARC references. L2ARC device flush.
*/
l2arc_evict(remdev, 0, B_TRUE);
list_destroy(&remdev->l2ad_buflist);
ASSERT(list_is_empty(&remdev->l2ad_lbptr_list));
list_destroy(&remdev->l2ad_lbptr_list);
mutex_destroy(&remdev->l2ad_mtx);
zfs_refcount_destroy(&remdev->l2ad_alloc);
zfs_refcount_destroy(&remdev->l2ad_lb_asize);
zfs_refcount_destroy(&remdev->l2ad_lb_count);
kmem_free(remdev->l2ad_dev_hdr, remdev->l2ad_dev_hdr_asize);
vmem_free(remdev, sizeof (l2arc_dev_t));
}
void
l2arc_init(void)
{
l2arc_thread_exit = 0;
l2arc_ndev = 0;
mutex_init(&l2arc_feed_thr_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&l2arc_feed_thr_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&l2arc_rebuild_thr_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&l2arc_rebuild_thr_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&l2arc_dev_mtx, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&l2arc_free_on_write_mtx, NULL, MUTEX_DEFAULT, NULL);
l2arc_dev_list = &L2ARC_dev_list;
l2arc_free_on_write = &L2ARC_free_on_write;
list_create(l2arc_dev_list, sizeof (l2arc_dev_t),
offsetof(l2arc_dev_t, l2ad_node));
list_create(l2arc_free_on_write, sizeof (l2arc_data_free_t),
offsetof(l2arc_data_free_t, l2df_list_node));
}
void
l2arc_fini(void)
{
mutex_destroy(&l2arc_feed_thr_lock);
cv_destroy(&l2arc_feed_thr_cv);
mutex_destroy(&l2arc_rebuild_thr_lock);
cv_destroy(&l2arc_rebuild_thr_cv);
mutex_destroy(&l2arc_dev_mtx);
mutex_destroy(&l2arc_free_on_write_mtx);
list_destroy(l2arc_dev_list);
list_destroy(l2arc_free_on_write);
}
void
l2arc_start(void)
{
if (!(spa_mode_global & SPA_MODE_WRITE))
return;
(void) thread_create(NULL, 0, l2arc_feed_thread, NULL, 0, &p0,
TS_RUN, defclsyspri);
}
void
l2arc_stop(void)
{
if (!(spa_mode_global & SPA_MODE_WRITE))
return;
mutex_enter(&l2arc_feed_thr_lock);
cv_signal(&l2arc_feed_thr_cv); /* kick thread out of startup */
l2arc_thread_exit = 1;
while (l2arc_thread_exit != 0)
cv_wait(&l2arc_feed_thr_cv, &l2arc_feed_thr_lock);
mutex_exit(&l2arc_feed_thr_lock);
}
/*
* Punches out rebuild threads for the L2ARC devices in a spa. This should
* be called after pool import from the spa async thread, since starting
* these threads directly from spa_import() will make them part of the
* "zpool import" context and delay process exit (and thus pool import).
*/
void
l2arc_spa_rebuild_start(spa_t *spa)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
/*
* Locate the spa's l2arc devices and kick off rebuild threads.
*/
for (int i = 0; i < spa->spa_l2cache.sav_count; i++) {
l2arc_dev_t *dev =
l2arc_vdev_get(spa->spa_l2cache.sav_vdevs[i]);
if (dev == NULL) {
/* Don't attempt a rebuild if the vdev is UNAVAIL */
continue;
}
mutex_enter(&l2arc_rebuild_thr_lock);
if (dev->l2ad_rebuild && !dev->l2ad_rebuild_cancel) {
dev->l2ad_rebuild_began = B_TRUE;
(void) thread_create(NULL, 0, l2arc_dev_rebuild_thread,
dev, 0, &p0, TS_RUN, minclsyspri);
}
mutex_exit(&l2arc_rebuild_thr_lock);
}
}
/*
* Main entry point for L2ARC rebuilding.
*/
static __attribute__((noreturn)) void
l2arc_dev_rebuild_thread(void *arg)
{
l2arc_dev_t *dev = arg;
VERIFY(!dev->l2ad_rebuild_cancel);
VERIFY(dev->l2ad_rebuild);
(void) l2arc_rebuild(dev);
mutex_enter(&l2arc_rebuild_thr_lock);
dev->l2ad_rebuild_began = B_FALSE;
dev->l2ad_rebuild = B_FALSE;
mutex_exit(&l2arc_rebuild_thr_lock);
thread_exit();
}
/*
* This function implements the actual L2ARC metadata rebuild. It:
* starts reading the log block chain and restores each block's contents
* to memory (reconstructing arc_buf_hdr_t's).
*
* Operation stops under any of the following conditions:
*
* 1) We reach the end of the log block chain.
* 2) We encounter *any* error condition (cksum errors, io errors)
*/
static int
l2arc_rebuild(l2arc_dev_t *dev)
{
vdev_t *vd = dev->l2ad_vdev;
spa_t *spa = vd->vdev_spa;
int err = 0;
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
l2arc_log_blk_phys_t *this_lb, *next_lb;
zio_t *this_io = NULL, *next_io = NULL;
l2arc_log_blkptr_t lbps[2];
l2arc_lb_ptr_buf_t *lb_ptr_buf;
boolean_t lock_held;
this_lb = vmem_zalloc(sizeof (*this_lb), KM_SLEEP);
next_lb = vmem_zalloc(sizeof (*next_lb), KM_SLEEP);
/*
* We prevent device removal while issuing reads to the device,
* then during the rebuilding phases we drop this lock again so
* that a spa_unload or device remove can be initiated - this is
* safe, because the spa will signal us to stop before removing
* our device and wait for us to stop.
*/
spa_config_enter(spa, SCL_L2ARC, vd, RW_READER);
lock_held = B_TRUE;
/*
* Retrieve the persistent L2ARC device state.
* L2BLK_GET_PSIZE returns aligned size for log blocks.
*/
dev->l2ad_evict = MAX(l2dhdr->dh_evict, dev->l2ad_start);
dev->l2ad_hand = MAX(l2dhdr->dh_start_lbps[0].lbp_daddr +
L2BLK_GET_PSIZE((&l2dhdr->dh_start_lbps[0])->lbp_prop),
dev->l2ad_start);
dev->l2ad_first = !!(l2dhdr->dh_flags & L2ARC_DEV_HDR_EVICT_FIRST);
vd->vdev_trim_action_time = l2dhdr->dh_trim_action_time;
vd->vdev_trim_state = l2dhdr->dh_trim_state;
/*
* In case the zfs module parameter l2arc_rebuild_enabled is false
* we do not start the rebuild process.
*/
if (!l2arc_rebuild_enabled)
goto out;
/* Prepare the rebuild process */
memcpy(lbps, l2dhdr->dh_start_lbps, sizeof (lbps));
/* Start the rebuild process */
for (;;) {
if (!l2arc_log_blkptr_valid(dev, &lbps[0]))
break;
if ((err = l2arc_log_blk_read(dev, &lbps[0], &lbps[1],
this_lb, next_lb, this_io, &next_io)) != 0)
goto out;
/*
* Our memory pressure valve. If the system is running low
* on memory, rather than swamping memory with new ARC buf
* hdrs, we opt not to rebuild the L2ARC. At this point,
* however, we have already set up our L2ARC dev to chain in
* new metadata log blocks, so the user may choose to offline/
* online the L2ARC dev at a later time (or re-import the pool)
* to reconstruct it (when there's less memory pressure).
*/
if (l2arc_hdr_limit_reached()) {
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_lowmem);
cmn_err(CE_NOTE, "System running low on memory, "
"aborting L2ARC rebuild.");
err = SET_ERROR(ENOMEM);
goto out;
}
spa_config_exit(spa, SCL_L2ARC, vd);
lock_held = B_FALSE;
/*
* Now that we know that the next_lb checks out alright, we
* can start reconstruction from this log block.
* L2BLK_GET_PSIZE returns aligned size for log blocks.
*/
uint64_t asize = L2BLK_GET_PSIZE((&lbps[0])->lbp_prop);
l2arc_log_blk_restore(dev, this_lb, asize);
/*
* log block restored, include its pointer in the list of
* pointers to log blocks present in the L2ARC device.
*/
lb_ptr_buf = kmem_zalloc(sizeof (l2arc_lb_ptr_buf_t), KM_SLEEP);
lb_ptr_buf->lb_ptr = kmem_zalloc(sizeof (l2arc_log_blkptr_t),
KM_SLEEP);
memcpy(lb_ptr_buf->lb_ptr, &lbps[0],
sizeof (l2arc_log_blkptr_t));
mutex_enter(&dev->l2ad_mtx);
list_insert_tail(&dev->l2ad_lbptr_list, lb_ptr_buf);
ARCSTAT_INCR(arcstat_l2_log_blk_asize, asize);
ARCSTAT_BUMP(arcstat_l2_log_blk_count);
zfs_refcount_add_many(&dev->l2ad_lb_asize, asize, lb_ptr_buf);
zfs_refcount_add(&dev->l2ad_lb_count, lb_ptr_buf);
mutex_exit(&dev->l2ad_mtx);
vdev_space_update(vd, asize, 0, 0);
/*
* Protection against loops of log blocks:
*
* l2ad_hand l2ad_evict
* V V
* l2ad_start |=======================================| l2ad_end
* -----|||----|||---|||----|||
* (3) (2) (1) (0)
* ---|||---|||----|||---|||
* (7) (6) (5) (4)
*
* In this situation the pointer of log block (4) passes
* l2arc_log_blkptr_valid() but the log block should not be
* restored as it is overwritten by the payload of log block
* (0). Only log blocks (0)-(3) should be restored. We check
* whether l2ad_evict lies in between the payload starting
* offset of the next log block (lbps[1].lbp_payload_start)
* and the payload starting offset of the present log block
* (lbps[0].lbp_payload_start). If true and this isn't the
* first pass, we are looping from the beginning and we should
* stop.
*/
if (l2arc_range_check_overlap(lbps[1].lbp_payload_start,
lbps[0].lbp_payload_start, dev->l2ad_evict) &&
!dev->l2ad_first)
goto out;
kpreempt(KPREEMPT_SYNC);
for (;;) {
mutex_enter(&l2arc_rebuild_thr_lock);
if (dev->l2ad_rebuild_cancel) {
dev->l2ad_rebuild = B_FALSE;
cv_signal(&l2arc_rebuild_thr_cv);
mutex_exit(&l2arc_rebuild_thr_lock);
err = SET_ERROR(ECANCELED);
goto out;
}
mutex_exit(&l2arc_rebuild_thr_lock);
if (spa_config_tryenter(spa, SCL_L2ARC, vd,
RW_READER)) {
lock_held = B_TRUE;
break;
}
/*
* L2ARC config lock held by somebody in writer,
* possibly due to them trying to remove us. They'll
* likely to want us to shut down, so after a little
* delay, we check l2ad_rebuild_cancel and retry
* the lock again.
*/
delay(1);
}
/*
* Continue with the next log block.
*/
lbps[0] = lbps[1];
lbps[1] = this_lb->lb_prev_lbp;
PTR_SWAP(this_lb, next_lb);
this_io = next_io;
next_io = NULL;
}
if (this_io != NULL)
l2arc_log_blk_fetch_abort(this_io);
out:
if (next_io != NULL)
l2arc_log_blk_fetch_abort(next_io);
vmem_free(this_lb, sizeof (*this_lb));
vmem_free(next_lb, sizeof (*next_lb));
if (!l2arc_rebuild_enabled) {
spa_history_log_internal(spa, "L2ARC rebuild", NULL,
"disabled");
} else if (err == 0 && zfs_refcount_count(&dev->l2ad_lb_count) > 0) {
ARCSTAT_BUMP(arcstat_l2_rebuild_success);
spa_history_log_internal(spa, "L2ARC rebuild", NULL,
"successful, restored %llu blocks",
(u_longlong_t)zfs_refcount_count(&dev->l2ad_lb_count));
} else if (err == 0 && zfs_refcount_count(&dev->l2ad_lb_count) == 0) {
/*
* No error but also nothing restored, meaning the lbps array
* in the device header points to invalid/non-present log
* blocks. Reset the header.
*/
spa_history_log_internal(spa, "L2ARC rebuild", NULL,
"no valid log blocks");
memset(l2dhdr, 0, dev->l2ad_dev_hdr_asize);
l2arc_dev_hdr_update(dev);
} else if (err == ECANCELED) {
/*
* In case the rebuild was canceled do not log to spa history
* log as the pool may be in the process of being removed.
*/
zfs_dbgmsg("L2ARC rebuild aborted, restored %llu blocks",
(u_longlong_t)zfs_refcount_count(&dev->l2ad_lb_count));
} else if (err != 0) {
spa_history_log_internal(spa, "L2ARC rebuild", NULL,
"aborted, restored %llu blocks",
(u_longlong_t)zfs_refcount_count(&dev->l2ad_lb_count));
}
if (lock_held)
spa_config_exit(spa, SCL_L2ARC, vd);
return (err);
}
/*
* Attempts to read the device header on the provided L2ARC device and writes
* it to `hdr'. On success, this function returns 0, otherwise the appropriate
* error code is returned.
*/
static int
l2arc_dev_hdr_read(l2arc_dev_t *dev)
{
int err;
uint64_t guid;
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
const uint64_t l2dhdr_asize = dev->l2ad_dev_hdr_asize;
abd_t *abd;
guid = spa_guid(dev->l2ad_vdev->vdev_spa);
abd = abd_get_from_buf(l2dhdr, l2dhdr_asize);
err = zio_wait(zio_read_phys(NULL, dev->l2ad_vdev,
VDEV_LABEL_START_SIZE, l2dhdr_asize, abd,
ZIO_CHECKSUM_LABEL, NULL, NULL, ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_DONT_CACHE | ZIO_FLAG_CANFAIL |
ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY |
ZIO_FLAG_SPECULATIVE, B_FALSE));
abd_free(abd);
if (err != 0) {
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_dh_errors);
zfs_dbgmsg("L2ARC IO error (%d) while reading device header, "
"vdev guid: %llu", err,
(u_longlong_t)dev->l2ad_vdev->vdev_guid);
return (err);
}
if (l2dhdr->dh_magic == BSWAP_64(L2ARC_DEV_HDR_MAGIC))
byteswap_uint64_array(l2dhdr, sizeof (*l2dhdr));
if (l2dhdr->dh_magic != L2ARC_DEV_HDR_MAGIC ||
l2dhdr->dh_spa_guid != guid ||
l2dhdr->dh_vdev_guid != dev->l2ad_vdev->vdev_guid ||
l2dhdr->dh_version != L2ARC_PERSISTENT_VERSION ||
l2dhdr->dh_log_entries != dev->l2ad_log_entries ||
l2dhdr->dh_end != dev->l2ad_end ||
!l2arc_range_check_overlap(dev->l2ad_start, dev->l2ad_end,
l2dhdr->dh_evict) ||
(l2dhdr->dh_trim_state != VDEV_TRIM_COMPLETE &&
l2arc_trim_ahead > 0)) {
/*
* Attempt to rebuild a device containing no actual dev hdr
* or containing a header from some other pool or from another
* version of persistent L2ARC.
*/
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_unsupported);
return (SET_ERROR(ENOTSUP));
}
return (0);
}
/*
* Reads L2ARC log blocks from storage and validates their contents.
*
* This function implements a simple fetcher to make sure that while
* we're processing one buffer the L2ARC is already fetching the next
* one in the chain.
*
* The arguments this_lp and next_lp point to the current and next log block
* address in the block chain. Similarly, this_lb and next_lb hold the
* l2arc_log_blk_phys_t's of the current and next L2ARC blk.
*
* The `this_io' and `next_io' arguments are used for block fetching.
* When issuing the first blk IO during rebuild, you should pass NULL for
* `this_io'. This function will then issue a sync IO to read the block and
* also issue an async IO to fetch the next block in the block chain. The
* fetched IO is returned in `next_io'. On subsequent calls to this
* function, pass the value returned in `next_io' from the previous call
* as `this_io' and a fresh `next_io' pointer to hold the next fetch IO.
* Prior to the call, you should initialize your `next_io' pointer to be
* NULL. If no fetch IO was issued, the pointer is left set at NULL.
*
* On success, this function returns 0, otherwise it returns an appropriate
* error code. On error the fetching IO is aborted and cleared before
* returning from this function. Therefore, if we return `success', the
* caller can assume that we have taken care of cleanup of fetch IOs.
*/
static int
l2arc_log_blk_read(l2arc_dev_t *dev,
const l2arc_log_blkptr_t *this_lbp, const l2arc_log_blkptr_t *next_lbp,
l2arc_log_blk_phys_t *this_lb, l2arc_log_blk_phys_t *next_lb,
zio_t *this_io, zio_t **next_io)
{
int err = 0;
zio_cksum_t cksum;
abd_t *abd = NULL;
uint64_t asize;
ASSERT(this_lbp != NULL && next_lbp != NULL);
ASSERT(this_lb != NULL && next_lb != NULL);
ASSERT(next_io != NULL && *next_io == NULL);
ASSERT(l2arc_log_blkptr_valid(dev, this_lbp));
/*
* Check to see if we have issued the IO for this log block in a
* previous run. If not, this is the first call, so issue it now.
*/
if (this_io == NULL) {
this_io = l2arc_log_blk_fetch(dev->l2ad_vdev, this_lbp,
this_lb);
}
/*
* Peek to see if we can start issuing the next IO immediately.
*/
if (l2arc_log_blkptr_valid(dev, next_lbp)) {
/*
* Start issuing IO for the next log block early - this
* should help keep the L2ARC device busy while we
* decompress and restore this log block.
*/
*next_io = l2arc_log_blk_fetch(dev->l2ad_vdev, next_lbp,
next_lb);
}
/* Wait for the IO to read this log block to complete */
if ((err = zio_wait(this_io)) != 0) {
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_io_errors);
zfs_dbgmsg("L2ARC IO error (%d) while reading log block, "
"offset: %llu, vdev guid: %llu", err,
(u_longlong_t)this_lbp->lbp_daddr,
(u_longlong_t)dev->l2ad_vdev->vdev_guid);
goto cleanup;
}
/*
* Make sure the buffer checks out.
* L2BLK_GET_PSIZE returns aligned size for log blocks.
*/
asize = L2BLK_GET_PSIZE((this_lbp)->lbp_prop);
fletcher_4_native(this_lb, asize, NULL, &cksum);
if (!ZIO_CHECKSUM_EQUAL(cksum, this_lbp->lbp_cksum)) {
ARCSTAT_BUMP(arcstat_l2_rebuild_abort_cksum_lb_errors);
zfs_dbgmsg("L2ARC log block cksum failed, offset: %llu, "
"vdev guid: %llu, l2ad_hand: %llu, l2ad_evict: %llu",
(u_longlong_t)this_lbp->lbp_daddr,
(u_longlong_t)dev->l2ad_vdev->vdev_guid,
(u_longlong_t)dev->l2ad_hand,
(u_longlong_t)dev->l2ad_evict);
err = SET_ERROR(ECKSUM);
goto cleanup;
}
/* Now we can take our time decoding this buffer */
switch (L2BLK_GET_COMPRESS((this_lbp)->lbp_prop)) {
case ZIO_COMPRESS_OFF:
break;
case ZIO_COMPRESS_LZ4:
abd = abd_alloc_for_io(asize, B_TRUE);
abd_copy_from_buf_off(abd, this_lb, 0, asize);
if ((err = zio_decompress_data(
L2BLK_GET_COMPRESS((this_lbp)->lbp_prop),
abd, this_lb, asize, sizeof (*this_lb), NULL)) != 0) {
err = SET_ERROR(EINVAL);
goto cleanup;
}
break;
default:
err = SET_ERROR(EINVAL);
goto cleanup;
}
if (this_lb->lb_magic == BSWAP_64(L2ARC_LOG_BLK_MAGIC))
byteswap_uint64_array(this_lb, sizeof (*this_lb));
if (this_lb->lb_magic != L2ARC_LOG_BLK_MAGIC) {
err = SET_ERROR(EINVAL);
goto cleanup;
}
cleanup:
/* Abort an in-flight fetch I/O in case of error */
if (err != 0 && *next_io != NULL) {
l2arc_log_blk_fetch_abort(*next_io);
*next_io = NULL;
}
if (abd != NULL)
abd_free(abd);
return (err);
}
/*
* Restores the payload of a log block to ARC. This creates empty ARC hdr
* entries which only contain an l2arc hdr, essentially restoring the
* buffers to their L2ARC evicted state. This function also updates space
* usage on the L2ARC vdev to make sure it tracks restored buffers.
*/
static void
l2arc_log_blk_restore(l2arc_dev_t *dev, const l2arc_log_blk_phys_t *lb,
uint64_t lb_asize)
{
uint64_t size = 0, asize = 0;
uint64_t log_entries = dev->l2ad_log_entries;
/*
* Usually arc_adapt() is called only for data, not headers, but
* since we may allocate significant amount of memory here, let ARC
* grow its arc_c.
*/
arc_adapt(log_entries * HDR_L2ONLY_SIZE);
for (int i = log_entries - 1; i >= 0; i--) {
/*
* Restore goes in the reverse temporal direction to preserve
* correct temporal ordering of buffers in the l2ad_buflist.
* l2arc_hdr_restore also does a list_insert_tail instead of
* list_insert_head on the l2ad_buflist:
*
* LIST l2ad_buflist LIST
* HEAD <------ (time) ------ TAIL
* direction +-----+-----+-----+-----+-----+ direction
* of l2arc <== | buf | buf | buf | buf | buf | ===> of rebuild
* fill +-----+-----+-----+-----+-----+
* ^ ^
* | |
* | |
* l2arc_feed_thread l2arc_rebuild
* will place new bufs here restores bufs here
*
* During l2arc_rebuild() the device is not used by
* l2arc_feed_thread() as dev->l2ad_rebuild is set to true.
*/
size += L2BLK_GET_LSIZE((&lb->lb_entries[i])->le_prop);
asize += vdev_psize_to_asize(dev->l2ad_vdev,
L2BLK_GET_PSIZE((&lb->lb_entries[i])->le_prop));
l2arc_hdr_restore(&lb->lb_entries[i], dev);
}
/*
* Record rebuild stats:
* size Logical size of restored buffers in the L2ARC
* asize Aligned size of restored buffers in the L2ARC
*/
ARCSTAT_INCR(arcstat_l2_rebuild_size, size);
ARCSTAT_INCR(arcstat_l2_rebuild_asize, asize);
ARCSTAT_INCR(arcstat_l2_rebuild_bufs, log_entries);
ARCSTAT_F_AVG(arcstat_l2_log_blk_avg_asize, lb_asize);
ARCSTAT_F_AVG(arcstat_l2_data_to_meta_ratio, asize / lb_asize);
ARCSTAT_BUMP(arcstat_l2_rebuild_log_blks);
}
/*
* Restores a single ARC buf hdr from a log entry. The ARC buffer is put
* into a state indicating that it has been evicted to L2ARC.
*/
static void
l2arc_hdr_restore(const l2arc_log_ent_phys_t *le, l2arc_dev_t *dev)
{
arc_buf_hdr_t *hdr, *exists;
kmutex_t *hash_lock;
arc_buf_contents_t type = L2BLK_GET_TYPE((le)->le_prop);
uint64_t asize;
/*
* Do all the allocation before grabbing any locks, this lets us
* sleep if memory is full and we don't have to deal with failed
* allocations.
*/
hdr = arc_buf_alloc_l2only(L2BLK_GET_LSIZE((le)->le_prop), type,
dev, le->le_dva, le->le_daddr,
L2BLK_GET_PSIZE((le)->le_prop), le->le_birth,
L2BLK_GET_COMPRESS((le)->le_prop), le->le_complevel,
L2BLK_GET_PROTECTED((le)->le_prop),
L2BLK_GET_PREFETCH((le)->le_prop),
L2BLK_GET_STATE((le)->le_prop));
asize = vdev_psize_to_asize(dev->l2ad_vdev,
L2BLK_GET_PSIZE((le)->le_prop));
/*
* vdev_space_update() has to be called before arc_hdr_destroy() to
* avoid underflow since the latter also calls vdev_space_update().
*/
l2arc_hdr_arcstats_increment(hdr);
vdev_space_update(dev->l2ad_vdev, asize, 0, 0);
mutex_enter(&dev->l2ad_mtx);
list_insert_tail(&dev->l2ad_buflist, hdr);
(void) zfs_refcount_add_many(&dev->l2ad_alloc, arc_hdr_size(hdr), hdr);
mutex_exit(&dev->l2ad_mtx);
exists = buf_hash_insert(hdr, &hash_lock);
if (exists) {
/* Buffer was already cached, no need to restore it. */
arc_hdr_destroy(hdr);
/*
* If the buffer is already cached, check whether it has
* L2ARC metadata. If not, enter them and update the flag.
* This is important is case of onlining a cache device, since
* we previously evicted all L2ARC metadata from ARC.
*/
if (!HDR_HAS_L2HDR(exists)) {
arc_hdr_set_flags(exists, ARC_FLAG_HAS_L2HDR);
exists->b_l2hdr.b_dev = dev;
exists->b_l2hdr.b_daddr = le->le_daddr;
exists->b_l2hdr.b_arcs_state =
L2BLK_GET_STATE((le)->le_prop);
mutex_enter(&dev->l2ad_mtx);
list_insert_tail(&dev->l2ad_buflist, exists);
(void) zfs_refcount_add_many(&dev->l2ad_alloc,
arc_hdr_size(exists), exists);
mutex_exit(&dev->l2ad_mtx);
l2arc_hdr_arcstats_increment(exists);
vdev_space_update(dev->l2ad_vdev, asize, 0, 0);
}
ARCSTAT_BUMP(arcstat_l2_rebuild_bufs_precached);
}
mutex_exit(hash_lock);
}
/*
* Starts an asynchronous read IO to read a log block. This is used in log
* block reconstruction to start reading the next block before we are done
* decoding and reconstructing the current block, to keep the l2arc device
* nice and hot with read IO to process.
* The returned zio will contain a newly allocated memory buffers for the IO
* data which should then be freed by the caller once the zio is no longer
* needed (i.e. due to it having completed). If you wish to abort this
* zio, you should do so using l2arc_log_blk_fetch_abort, which takes
* care of disposing of the allocated buffers correctly.
*/
static zio_t *
l2arc_log_blk_fetch(vdev_t *vd, const l2arc_log_blkptr_t *lbp,
l2arc_log_blk_phys_t *lb)
{
uint32_t asize;
zio_t *pio;
l2arc_read_callback_t *cb;
/* L2BLK_GET_PSIZE returns aligned size for log blocks */
asize = L2BLK_GET_PSIZE((lbp)->lbp_prop);
ASSERT(asize <= sizeof (l2arc_log_blk_phys_t));
cb = kmem_zalloc(sizeof (l2arc_read_callback_t), KM_SLEEP);
cb->l2rcb_abd = abd_get_from_buf(lb, asize);
pio = zio_root(vd->vdev_spa, l2arc_blk_fetch_done, cb,
ZIO_FLAG_DONT_CACHE | ZIO_FLAG_CANFAIL | ZIO_FLAG_DONT_PROPAGATE |
ZIO_FLAG_DONT_RETRY);
(void) zio_nowait(zio_read_phys(pio, vd, lbp->lbp_daddr, asize,
cb->l2rcb_abd, ZIO_CHECKSUM_OFF, NULL, NULL,
ZIO_PRIORITY_ASYNC_READ, ZIO_FLAG_DONT_CACHE | ZIO_FLAG_CANFAIL |
ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY, B_FALSE));
return (pio);
}
/*
* Aborts a zio returned from l2arc_log_blk_fetch and frees the data
* buffers allocated for it.
*/
static void
l2arc_log_blk_fetch_abort(zio_t *zio)
{
(void) zio_wait(zio);
}
/*
* Creates a zio to update the device header on an l2arc device.
*/
void
l2arc_dev_hdr_update(l2arc_dev_t *dev)
{
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
const uint64_t l2dhdr_asize = dev->l2ad_dev_hdr_asize;
abd_t *abd;
int err;
VERIFY(spa_config_held(dev->l2ad_spa, SCL_STATE_ALL, RW_READER));
l2dhdr->dh_magic = L2ARC_DEV_HDR_MAGIC;
l2dhdr->dh_version = L2ARC_PERSISTENT_VERSION;
l2dhdr->dh_spa_guid = spa_guid(dev->l2ad_vdev->vdev_spa);
l2dhdr->dh_vdev_guid = dev->l2ad_vdev->vdev_guid;
l2dhdr->dh_log_entries = dev->l2ad_log_entries;
l2dhdr->dh_evict = dev->l2ad_evict;
l2dhdr->dh_start = dev->l2ad_start;
l2dhdr->dh_end = dev->l2ad_end;
l2dhdr->dh_lb_asize = zfs_refcount_count(&dev->l2ad_lb_asize);
l2dhdr->dh_lb_count = zfs_refcount_count(&dev->l2ad_lb_count);
l2dhdr->dh_flags = 0;
l2dhdr->dh_trim_action_time = dev->l2ad_vdev->vdev_trim_action_time;
l2dhdr->dh_trim_state = dev->l2ad_vdev->vdev_trim_state;
if (dev->l2ad_first)
l2dhdr->dh_flags |= L2ARC_DEV_HDR_EVICT_FIRST;
abd = abd_get_from_buf(l2dhdr, l2dhdr_asize);
err = zio_wait(zio_write_phys(NULL, dev->l2ad_vdev,
VDEV_LABEL_START_SIZE, l2dhdr_asize, abd, ZIO_CHECKSUM_LABEL, NULL,
NULL, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_CANFAIL, B_FALSE));
abd_free(abd);
if (err != 0) {
zfs_dbgmsg("L2ARC IO error (%d) while writing device header, "
"vdev guid: %llu", err,
(u_longlong_t)dev->l2ad_vdev->vdev_guid);
}
}
/*
* Commits a log block to the L2ARC device. This routine is invoked from
* l2arc_write_buffers when the log block fills up.
* This function allocates some memory to temporarily hold the serialized
* buffer to be written. This is then released in l2arc_write_done.
*/
static void
l2arc_log_blk_commit(l2arc_dev_t *dev, zio_t *pio, l2arc_write_callback_t *cb)
{
l2arc_log_blk_phys_t *lb = &dev->l2ad_log_blk;
l2arc_dev_hdr_phys_t *l2dhdr = dev->l2ad_dev_hdr;
uint64_t psize, asize;
zio_t *wzio;
l2arc_lb_abd_buf_t *abd_buf;
uint8_t *tmpbuf = NULL;
l2arc_lb_ptr_buf_t *lb_ptr_buf;
VERIFY3S(dev->l2ad_log_ent_idx, ==, dev->l2ad_log_entries);
abd_buf = zio_buf_alloc(sizeof (*abd_buf));
abd_buf->abd = abd_get_from_buf(lb, sizeof (*lb));
lb_ptr_buf = kmem_zalloc(sizeof (l2arc_lb_ptr_buf_t), KM_SLEEP);
lb_ptr_buf->lb_ptr = kmem_zalloc(sizeof (l2arc_log_blkptr_t), KM_SLEEP);
/* link the buffer into the block chain */
lb->lb_prev_lbp = l2dhdr->dh_start_lbps[1];
lb->lb_magic = L2ARC_LOG_BLK_MAGIC;
/*
* l2arc_log_blk_commit() may be called multiple times during a single
* l2arc_write_buffers() call. Save the allocated abd buffers in a list
* so we can free them in l2arc_write_done() later on.
*/
list_insert_tail(&cb->l2wcb_abd_list, abd_buf);
/* try to compress the buffer */
psize = zio_compress_data(ZIO_COMPRESS_LZ4,
abd_buf->abd, (void **) &tmpbuf, sizeof (*lb), 0);
/* a log block is never entirely zero */
ASSERT(psize != 0);
asize = vdev_psize_to_asize(dev->l2ad_vdev, psize);
ASSERT(asize <= sizeof (*lb));
/*
* Update the start log block pointer in the device header to point
* to the log block we're about to write.
*/
l2dhdr->dh_start_lbps[1] = l2dhdr->dh_start_lbps[0];
l2dhdr->dh_start_lbps[0].lbp_daddr = dev->l2ad_hand;
l2dhdr->dh_start_lbps[0].lbp_payload_asize =
dev->l2ad_log_blk_payload_asize;
l2dhdr->dh_start_lbps[0].lbp_payload_start =
dev->l2ad_log_blk_payload_start;
L2BLK_SET_LSIZE(
(&l2dhdr->dh_start_lbps[0])->lbp_prop, sizeof (*lb));
L2BLK_SET_PSIZE(
(&l2dhdr->dh_start_lbps[0])->lbp_prop, asize);
L2BLK_SET_CHECKSUM(
(&l2dhdr->dh_start_lbps[0])->lbp_prop,
ZIO_CHECKSUM_FLETCHER_4);
if (asize < sizeof (*lb)) {
/* compression succeeded */
memset(tmpbuf + psize, 0, asize - psize);
L2BLK_SET_COMPRESS(
(&l2dhdr->dh_start_lbps[0])->lbp_prop,
ZIO_COMPRESS_LZ4);
} else {
/* compression failed */
memcpy(tmpbuf, lb, sizeof (*lb));
L2BLK_SET_COMPRESS(
(&l2dhdr->dh_start_lbps[0])->lbp_prop,
ZIO_COMPRESS_OFF);
}
/* checksum what we're about to write */
fletcher_4_native(tmpbuf, asize, NULL,
&l2dhdr->dh_start_lbps[0].lbp_cksum);
abd_free(abd_buf->abd);
/* perform the write itself */
abd_buf->abd = abd_get_from_buf(tmpbuf, sizeof (*lb));
abd_take_ownership_of_buf(abd_buf->abd, B_TRUE);
wzio = zio_write_phys(pio, dev->l2ad_vdev, dev->l2ad_hand,
asize, abd_buf->abd, ZIO_CHECKSUM_OFF, NULL, NULL,
ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_CANFAIL, B_FALSE);
DTRACE_PROBE2(l2arc__write, vdev_t *, dev->l2ad_vdev, zio_t *, wzio);
(void) zio_nowait(wzio);
dev->l2ad_hand += asize;
/*
* Include the committed log block's pointer in the list of pointers
* to log blocks present in the L2ARC device.
*/
memcpy(lb_ptr_buf->lb_ptr, &l2dhdr->dh_start_lbps[0],
sizeof (l2arc_log_blkptr_t));
mutex_enter(&dev->l2ad_mtx);
list_insert_head(&dev->l2ad_lbptr_list, lb_ptr_buf);
ARCSTAT_INCR(arcstat_l2_log_blk_asize, asize);
ARCSTAT_BUMP(arcstat_l2_log_blk_count);
zfs_refcount_add_many(&dev->l2ad_lb_asize, asize, lb_ptr_buf);
zfs_refcount_add(&dev->l2ad_lb_count, lb_ptr_buf);
mutex_exit(&dev->l2ad_mtx);
vdev_space_update(dev->l2ad_vdev, asize, 0, 0);
/* bump the kstats */
ARCSTAT_INCR(arcstat_l2_write_bytes, asize);
ARCSTAT_BUMP(arcstat_l2_log_blk_writes);
ARCSTAT_F_AVG(arcstat_l2_log_blk_avg_asize, asize);
ARCSTAT_F_AVG(arcstat_l2_data_to_meta_ratio,
dev->l2ad_log_blk_payload_asize / asize);
/* start a new log block */
dev->l2ad_log_ent_idx = 0;
dev->l2ad_log_blk_payload_asize = 0;
dev->l2ad_log_blk_payload_start = 0;
}
/*
* Validates an L2ARC log block address to make sure that it can be read
* from the provided L2ARC device.
*/
boolean_t
l2arc_log_blkptr_valid(l2arc_dev_t *dev, const l2arc_log_blkptr_t *lbp)
{
/* L2BLK_GET_PSIZE returns aligned size for log blocks */
uint64_t asize = L2BLK_GET_PSIZE((lbp)->lbp_prop);
uint64_t end = lbp->lbp_daddr + asize - 1;
uint64_t start = lbp->lbp_payload_start;
boolean_t evicted = B_FALSE;
/*
* A log block is valid if all of the following conditions are true:
* - it fits entirely (including its payload) between l2ad_start and
* l2ad_end
* - it has a valid size
* - neither the log block itself nor part of its payload was evicted
* by l2arc_evict():
*
* l2ad_hand l2ad_evict
* | | lbp_daddr
* | start | | end
* | | | | |
* V V V V V
* l2ad_start ============================================ l2ad_end
* --------------------------||||
* ^ ^
* | log block
* payload
*/
evicted =
l2arc_range_check_overlap(start, end, dev->l2ad_hand) ||
l2arc_range_check_overlap(start, end, dev->l2ad_evict) ||
l2arc_range_check_overlap(dev->l2ad_hand, dev->l2ad_evict, start) ||
l2arc_range_check_overlap(dev->l2ad_hand, dev->l2ad_evict, end);
return (start >= dev->l2ad_start && end <= dev->l2ad_end &&
asize > 0 && asize <= sizeof (l2arc_log_blk_phys_t) &&
(!evicted || dev->l2ad_first));
}
/*
* Inserts ARC buffer header `hdr' into the current L2ARC log block on
* the device. The buffer being inserted must be present in L2ARC.
* Returns B_TRUE if the L2ARC log block is full and needs to be committed
* to L2ARC, or B_FALSE if it still has room for more ARC buffers.
*/
static boolean_t
l2arc_log_blk_insert(l2arc_dev_t *dev, const arc_buf_hdr_t *hdr)
{
l2arc_log_blk_phys_t *lb = &dev->l2ad_log_blk;
l2arc_log_ent_phys_t *le;
if (dev->l2ad_log_entries == 0)
return (B_FALSE);
int index = dev->l2ad_log_ent_idx++;
ASSERT3S(index, <, dev->l2ad_log_entries);
ASSERT(HDR_HAS_L2HDR(hdr));
le = &lb->lb_entries[index];
memset(le, 0, sizeof (*le));
le->le_dva = hdr->b_dva;
le->le_birth = hdr->b_birth;
le->le_daddr = hdr->b_l2hdr.b_daddr;
if (index == 0)
dev->l2ad_log_blk_payload_start = le->le_daddr;
L2BLK_SET_LSIZE((le)->le_prop, HDR_GET_LSIZE(hdr));
L2BLK_SET_PSIZE((le)->le_prop, HDR_GET_PSIZE(hdr));
L2BLK_SET_COMPRESS((le)->le_prop, HDR_GET_COMPRESS(hdr));
le->le_complevel = hdr->b_complevel;
L2BLK_SET_TYPE((le)->le_prop, hdr->b_type);
L2BLK_SET_PROTECTED((le)->le_prop, !!(HDR_PROTECTED(hdr)));
L2BLK_SET_PREFETCH((le)->le_prop, !!(HDR_PREFETCH(hdr)));
L2BLK_SET_STATE((le)->le_prop, hdr->b_l1hdr.b_state->arcs_state);
dev->l2ad_log_blk_payload_asize += vdev_psize_to_asize(dev->l2ad_vdev,
HDR_GET_PSIZE(hdr));
return (dev->l2ad_log_ent_idx == dev->l2ad_log_entries);
}
/*
* Checks whether a given L2ARC device address sits in a time-sequential
* range. The trick here is that the L2ARC is a rotary buffer, so we can't
* just do a range comparison, we need to handle the situation in which the
* range wraps around the end of the L2ARC device. Arguments:
* bottom -- Lower end of the range to check (written to earlier).
* top -- Upper end of the range to check (written to later).
* check -- The address for which we want to determine if it sits in
* between the top and bottom.
*
* The 3-way conditional below represents the following cases:
*
* bottom < top : Sequentially ordered case:
* <check>--------+-------------------+
* | (overlap here?) |
* L2ARC dev V V
* |---------------<bottom>============<top>--------------|
*
* bottom > top: Looped-around case:
* <check>--------+------------------+
* | (overlap here?) |
* L2ARC dev V V
* |===============<top>---------------<bottom>===========|
* ^ ^
* | (or here?) |
* +---------------+---------<check>
*
* top == bottom : Just a single address comparison.
*/
boolean_t
l2arc_range_check_overlap(uint64_t bottom, uint64_t top, uint64_t check)
{
if (bottom < top)
return (bottom <= check && check <= top);
else if (bottom > top)
return (check <= top || bottom <= check);
else
return (check == top);
}
EXPORT_SYMBOL(arc_buf_size);
EXPORT_SYMBOL(arc_write);
EXPORT_SYMBOL(arc_read);
EXPORT_SYMBOL(arc_buf_info);
EXPORT_SYMBOL(arc_getbuf_func);
EXPORT_SYMBOL(arc_add_prune_callback);
EXPORT_SYMBOL(arc_remove_prune_callback);
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, min, param_set_arc_min,
spl_param_get_u64, ZMOD_RW, "Minimum ARC size in bytes");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, max, param_set_arc_max,
spl_param_get_u64, ZMOD_RW, "Maximum ARC size in bytes");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, meta_balance, UINT, ZMOD_RW,
"Balance between metadata and data on ghost hits.");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, grow_retry, param_set_arc_int,
param_get_uint, ZMOD_RW, "Seconds before growing ARC size");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, shrink_shift, param_set_arc_int,
param_get_uint, ZMOD_RW, "log2(fraction of ARC to reclaim)");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, pc_percent, UINT, ZMOD_RW,
"Percent of pagecache to reclaim ARC to");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, average_blocksize, UINT, ZMOD_RD,
"Target average block size");
ZFS_MODULE_PARAM(zfs, zfs_, compressed_arc_enabled, INT, ZMOD_RW,
"Disable compressed ARC buffers");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, min_prefetch_ms, param_set_arc_int,
param_get_uint, ZMOD_RW, "Min life of prefetch block in ms");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, min_prescient_prefetch_ms,
param_set_arc_int, param_get_uint, ZMOD_RW,
"Min life of prescient prefetched block in ms");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, write_max, U64, ZMOD_RW,
"Max write bytes per interval");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, write_boost, U64, ZMOD_RW,
"Extra write bytes during device warmup");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, headroom, U64, ZMOD_RW,
"Number of max device writes to precache");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, headroom_boost, U64, ZMOD_RW,
"Compressed l2arc_headroom multiplier");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, trim_ahead, U64, ZMOD_RW,
"TRIM ahead L2ARC write size multiplier");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, feed_secs, U64, ZMOD_RW,
"Seconds between L2ARC writing");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, feed_min_ms, U64, ZMOD_RW,
"Min feed interval in milliseconds");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, noprefetch, INT, ZMOD_RW,
"Skip caching prefetched buffers");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, feed_again, INT, ZMOD_RW,
"Turbo L2ARC warmup");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, norw, INT, ZMOD_RW,
"No reads during writes");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, meta_percent, UINT, ZMOD_RW,
"Percent of ARC size allowed for L2ARC-only headers");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, rebuild_enabled, INT, ZMOD_RW,
"Rebuild the L2ARC when importing a pool");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, rebuild_blocks_min_l2size, U64, ZMOD_RW,
"Min size in bytes to write rebuild log blocks in L2ARC");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, mfuonly, INT, ZMOD_RW,
"Cache only MFU data from ARC into L2ARC");
ZFS_MODULE_PARAM(zfs_l2arc, l2arc_, exclude_special, INT, ZMOD_RW,
"Exclude dbufs on special vdevs from being cached to L2ARC if set.");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, lotsfree_percent, param_set_arc_int,
param_get_uint, ZMOD_RW, "System free memory I/O throttle in bytes");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, sys_free, param_set_arc_u64,
spl_param_get_u64, ZMOD_RW, "System free memory target size in bytes");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, dnode_limit, param_set_arc_u64,
spl_param_get_u64, ZMOD_RW, "Minimum bytes of dnodes in ARC");
ZFS_MODULE_PARAM_CALL(zfs_arc, zfs_arc_, dnode_limit_percent,
param_set_arc_int, param_get_uint, ZMOD_RW,
"Percent of ARC meta buffers for dnodes");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, dnode_reduce_percent, UINT, ZMOD_RW,
"Percentage of excess dnodes to try to unpin");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, eviction_pct, UINT, ZMOD_RW,
"When full, ARC allocation waits for eviction of this % of alloc size");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, evict_batch_limit, UINT, ZMOD_RW,
"The number of headers to evict per sublist before moving to the next");
ZFS_MODULE_PARAM(zfs_arc, zfs_arc_, prune_task_threads, INT, ZMOD_RW,
"Number of arc_prune threads");
diff --git a/sys/contrib/openzfs/module/zfs/dbuf.c b/sys/contrib/openzfs/module/zfs/dbuf.c
index c7f76e8d96f8..8193fb244079 100644
--- a/sys/contrib/openzfs/module/zfs/dbuf.c
+++ b/sys/contrib/openzfs/module/zfs/dbuf.c
@@ -1,5184 +1,5183 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2012, 2020 by Delphix. All rights reserved.
* Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
* Copyright (c) 2019, Klara Inc.
* Copyright (c) 2019, Allan Jude
* Copyright (c) 2021, 2022 by Pawel Jakub Dawidek
*/
#include <sys/zfs_context.h>
#include <sys/arc.h>
#include <sys/dmu.h>
#include <sys/dmu_send.h>
#include <sys/dmu_impl.h>
#include <sys/dbuf.h>
#include <sys/dmu_objset.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_dir.h>
#include <sys/dmu_tx.h>
#include <sys/spa.h>
#include <sys/zio.h>
#include <sys/dmu_zfetch.h>
#include <sys/sa.h>
#include <sys/sa_impl.h>
#include <sys/zfeature.h>
#include <sys/blkptr.h>
#include <sys/range_tree.h>
#include <sys/trace_zfs.h>
#include <sys/callb.h>
#include <sys/abd.h>
#include <sys/brt.h>
#include <sys/vdev.h>
#include <cityhash.h>
#include <sys/spa_impl.h>
#include <sys/wmsum.h>
#include <sys/vdev_impl.h>
static kstat_t *dbuf_ksp;
typedef struct dbuf_stats {
/*
* Various statistics about the size of the dbuf cache.
*/
kstat_named_t cache_count;
kstat_named_t cache_size_bytes;
kstat_named_t cache_size_bytes_max;
/*
* Statistics regarding the bounds on the dbuf cache size.
*/
kstat_named_t cache_target_bytes;
kstat_named_t cache_lowater_bytes;
kstat_named_t cache_hiwater_bytes;
/*
* Total number of dbuf cache evictions that have occurred.
*/
kstat_named_t cache_total_evicts;
/*
* The distribution of dbuf levels in the dbuf cache and
* the total size of all dbufs at each level.
*/
kstat_named_t cache_levels[DN_MAX_LEVELS];
kstat_named_t cache_levels_bytes[DN_MAX_LEVELS];
/*
* Statistics about the dbuf hash table.
*/
kstat_named_t hash_hits;
kstat_named_t hash_misses;
kstat_named_t hash_collisions;
kstat_named_t hash_elements;
kstat_named_t hash_elements_max;
/*
* Number of sublists containing more than one dbuf in the dbuf
* hash table. Keep track of the longest hash chain.
*/
kstat_named_t hash_chains;
kstat_named_t hash_chain_max;
/*
* Number of times a dbuf_create() discovers that a dbuf was
* already created and in the dbuf hash table.
*/
kstat_named_t hash_insert_race;
/*
* Number of entries in the hash table dbuf and mutex arrays.
*/
kstat_named_t hash_table_count;
kstat_named_t hash_mutex_count;
/*
* Statistics about the size of the metadata dbuf cache.
*/
kstat_named_t metadata_cache_count;
kstat_named_t metadata_cache_size_bytes;
kstat_named_t metadata_cache_size_bytes_max;
/*
* For diagnostic purposes, this is incremented whenever we can't add
* something to the metadata cache because it's full, and instead put
* the data in the regular dbuf cache.
*/
kstat_named_t metadata_cache_overflow;
} dbuf_stats_t;
dbuf_stats_t dbuf_stats = {
{ "cache_count", KSTAT_DATA_UINT64 },
{ "cache_size_bytes", KSTAT_DATA_UINT64 },
{ "cache_size_bytes_max", KSTAT_DATA_UINT64 },
{ "cache_target_bytes", KSTAT_DATA_UINT64 },
{ "cache_lowater_bytes", KSTAT_DATA_UINT64 },
{ "cache_hiwater_bytes", KSTAT_DATA_UINT64 },
{ "cache_total_evicts", KSTAT_DATA_UINT64 },
{ { "cache_levels_N", KSTAT_DATA_UINT64 } },
{ { "cache_levels_bytes_N", KSTAT_DATA_UINT64 } },
{ "hash_hits", KSTAT_DATA_UINT64 },
{ "hash_misses", KSTAT_DATA_UINT64 },
{ "hash_collisions", KSTAT_DATA_UINT64 },
{ "hash_elements", KSTAT_DATA_UINT64 },
{ "hash_elements_max", KSTAT_DATA_UINT64 },
{ "hash_chains", KSTAT_DATA_UINT64 },
{ "hash_chain_max", KSTAT_DATA_UINT64 },
{ "hash_insert_race", KSTAT_DATA_UINT64 },
{ "hash_table_count", KSTAT_DATA_UINT64 },
{ "hash_mutex_count", KSTAT_DATA_UINT64 },
{ "metadata_cache_count", KSTAT_DATA_UINT64 },
{ "metadata_cache_size_bytes", KSTAT_DATA_UINT64 },
{ "metadata_cache_size_bytes_max", KSTAT_DATA_UINT64 },
{ "metadata_cache_overflow", KSTAT_DATA_UINT64 }
};
struct {
wmsum_t cache_count;
wmsum_t cache_total_evicts;
wmsum_t cache_levels[DN_MAX_LEVELS];
wmsum_t cache_levels_bytes[DN_MAX_LEVELS];
wmsum_t hash_hits;
wmsum_t hash_misses;
wmsum_t hash_collisions;
wmsum_t hash_chains;
wmsum_t hash_insert_race;
wmsum_t metadata_cache_count;
wmsum_t metadata_cache_overflow;
} dbuf_sums;
#define DBUF_STAT_INCR(stat, val) \
wmsum_add(&dbuf_sums.stat, val);
#define DBUF_STAT_DECR(stat, val) \
DBUF_STAT_INCR(stat, -(val));
#define DBUF_STAT_BUMP(stat) \
DBUF_STAT_INCR(stat, 1);
#define DBUF_STAT_BUMPDOWN(stat) \
DBUF_STAT_INCR(stat, -1);
#define DBUF_STAT_MAX(stat, v) { \
uint64_t _m; \
while ((v) > (_m = dbuf_stats.stat.value.ui64) && \
(_m != atomic_cas_64(&dbuf_stats.stat.value.ui64, _m, (v))))\
continue; \
}
static void dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx);
static void dbuf_sync_leaf_verify_bonus_dnode(dbuf_dirty_record_t *dr);
static int dbuf_read_verify_dnode_crypt(dmu_buf_impl_t *db, uint32_t flags);
/*
* Global data structures and functions for the dbuf cache.
*/
static kmem_cache_t *dbuf_kmem_cache;
static taskq_t *dbu_evict_taskq;
static kthread_t *dbuf_cache_evict_thread;
static kmutex_t dbuf_evict_lock;
static kcondvar_t dbuf_evict_cv;
static boolean_t dbuf_evict_thread_exit;
/*
* There are two dbuf caches; each dbuf can only be in one of them at a time.
*
* 1. Cache of metadata dbufs, to help make read-heavy administrative commands
* from /sbin/zfs run faster. The "metadata cache" specifically stores dbufs
* that represent the metadata that describes filesystems/snapshots/
* bookmarks/properties/etc. We only evict from this cache when we export a
* pool, to short-circuit as much I/O as possible for all administrative
* commands that need the metadata. There is no eviction policy for this
* cache, because we try to only include types in it which would occupy a
* very small amount of space per object but create a large impact on the
* performance of these commands. Instead, after it reaches a maximum size
* (which should only happen on very small memory systems with a very large
* number of filesystem objects), we stop taking new dbufs into the
* metadata cache, instead putting them in the normal dbuf cache.
*
* 2. LRU cache of dbufs. The dbuf cache maintains a list of dbufs that
* are not currently held but have been recently released. These dbufs
* are not eligible for arc eviction until they are aged out of the cache.
* Dbufs that are aged out of the cache will be immediately destroyed and
* become eligible for arc eviction.
*
* Dbufs are added to these caches once the last hold is released. If a dbuf is
* later accessed and still exists in the dbuf cache, then it will be removed
* from the cache and later re-added to the head of the cache.
*
* If a given dbuf meets the requirements for the metadata cache, it will go
* there, otherwise it will be considered for the generic LRU dbuf cache. The
* caches and the refcounts tracking their sizes are stored in an array indexed
* by those caches' matching enum values (from dbuf_cached_state_t).
*/
typedef struct dbuf_cache {
multilist_t cache;
zfs_refcount_t size ____cacheline_aligned;
} dbuf_cache_t;
dbuf_cache_t dbuf_caches[DB_CACHE_MAX];
/* Size limits for the caches */
static uint64_t dbuf_cache_max_bytes = UINT64_MAX;
static uint64_t dbuf_metadata_cache_max_bytes = UINT64_MAX;
/* Set the default sizes of the caches to log2 fraction of arc size */
static uint_t dbuf_cache_shift = 5;
static uint_t dbuf_metadata_cache_shift = 6;
/* Set the dbuf hash mutex count as log2 shift (dynamic by default) */
static uint_t dbuf_mutex_cache_shift = 0;
static unsigned long dbuf_cache_target_bytes(void);
static unsigned long dbuf_metadata_cache_target_bytes(void);
/*
* The LRU dbuf cache uses a three-stage eviction policy:
* - A low water marker designates when the dbuf eviction thread
* should stop evicting from the dbuf cache.
* - When we reach the maximum size (aka mid water mark), we
* signal the eviction thread to run.
* - The high water mark indicates when the eviction thread
* is unable to keep up with the incoming load and eviction must
* happen in the context of the calling thread.
*
* The dbuf cache:
* (max size)
* low water mid water hi water
* +----------------------------------------+----------+----------+
* | | | |
* | | | |
* | | | |
* | | | |
* +----------------------------------------+----------+----------+
* stop signal evict
* evicting eviction directly
* thread
*
* The high and low water marks indicate the operating range for the eviction
* thread. The low water mark is, by default, 90% of the total size of the
* cache and the high water mark is at 110% (both of these percentages can be
* changed by setting dbuf_cache_lowater_pct and dbuf_cache_hiwater_pct,
* respectively). The eviction thread will try to ensure that the cache remains
* within this range by waking up every second and checking if the cache is
* above the low water mark. The thread can also be woken up by callers adding
* elements into the cache if the cache is larger than the mid water (i.e max
* cache size). Once the eviction thread is woken up and eviction is required,
* it will continue evicting buffers until it's able to reduce the cache size
* to the low water mark. If the cache size continues to grow and hits the high
* water mark, then callers adding elements to the cache will begin to evict
* directly from the cache until the cache is no longer above the high water
* mark.
*/
/*
* The percentage above and below the maximum cache size.
*/
static uint_t dbuf_cache_hiwater_pct = 10;
static uint_t dbuf_cache_lowater_pct = 10;
static int
dbuf_cons(void *vdb, void *unused, int kmflag)
{
(void) unused, (void) kmflag;
dmu_buf_impl_t *db = vdb;
memset(db, 0, sizeof (dmu_buf_impl_t));
mutex_init(&db->db_mtx, NULL, MUTEX_DEFAULT, NULL);
rw_init(&db->db_rwlock, NULL, RW_DEFAULT, NULL);
cv_init(&db->db_changed, NULL, CV_DEFAULT, NULL);
multilist_link_init(&db->db_cache_link);
zfs_refcount_create(&db->db_holds);
return (0);
}
static void
dbuf_dest(void *vdb, void *unused)
{
(void) unused;
dmu_buf_impl_t *db = vdb;
mutex_destroy(&db->db_mtx);
rw_destroy(&db->db_rwlock);
cv_destroy(&db->db_changed);
ASSERT(!multilist_link_active(&db->db_cache_link));
zfs_refcount_destroy(&db->db_holds);
}
/*
* dbuf hash table routines
*/
static dbuf_hash_table_t dbuf_hash_table;
/*
* We use Cityhash for this. It's fast, and has good hash properties without
* requiring any large static buffers.
*/
static uint64_t
dbuf_hash(void *os, uint64_t obj, uint8_t lvl, uint64_t blkid)
{
return (cityhash4((uintptr_t)os, obj, (uint64_t)lvl, blkid));
}
#define DTRACE_SET_STATE(db, why) \
DTRACE_PROBE2(dbuf__state_change, dmu_buf_impl_t *, db, \
const char *, why)
#define DBUF_EQUAL(dbuf, os, obj, level, blkid) \
((dbuf)->db.db_object == (obj) && \
(dbuf)->db_objset == (os) && \
(dbuf)->db_level == (level) && \
(dbuf)->db_blkid == (blkid))
dmu_buf_impl_t *
dbuf_find(objset_t *os, uint64_t obj, uint8_t level, uint64_t blkid,
uint64_t *hash_out)
{
dbuf_hash_table_t *h = &dbuf_hash_table;
uint64_t hv;
uint64_t idx;
dmu_buf_impl_t *db;
hv = dbuf_hash(os, obj, level, blkid);
idx = hv & h->hash_table_mask;
mutex_enter(DBUF_HASH_MUTEX(h, idx));
for (db = h->hash_table[idx]; db != NULL; db = db->db_hash_next) {
if (DBUF_EQUAL(db, os, obj, level, blkid)) {
mutex_enter(&db->db_mtx);
if (db->db_state != DB_EVICTING) {
mutex_exit(DBUF_HASH_MUTEX(h, idx));
return (db);
}
mutex_exit(&db->db_mtx);
}
}
mutex_exit(DBUF_HASH_MUTEX(h, idx));
if (hash_out != NULL)
*hash_out = hv;
return (NULL);
}
static dmu_buf_impl_t *
dbuf_find_bonus(objset_t *os, uint64_t object)
{
dnode_t *dn;
dmu_buf_impl_t *db = NULL;
if (dnode_hold(os, object, FTAG, &dn) == 0) {
rw_enter(&dn->dn_struct_rwlock, RW_READER);
if (dn->dn_bonus != NULL) {
db = dn->dn_bonus;
mutex_enter(&db->db_mtx);
}
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
}
return (db);
}
/*
* Insert an entry into the hash table. If there is already an element
* equal to elem in the hash table, then the already existing element
* will be returned and the new element will not be inserted.
* Otherwise returns NULL.
*/
static dmu_buf_impl_t *
dbuf_hash_insert(dmu_buf_impl_t *db)
{
dbuf_hash_table_t *h = &dbuf_hash_table;
objset_t *os = db->db_objset;
uint64_t obj = db->db.db_object;
int level = db->db_level;
uint64_t blkid, idx;
dmu_buf_impl_t *dbf;
uint32_t i;
blkid = db->db_blkid;
ASSERT3U(dbuf_hash(os, obj, level, blkid), ==, db->db_hash);
idx = db->db_hash & h->hash_table_mask;
mutex_enter(DBUF_HASH_MUTEX(h, idx));
for (dbf = h->hash_table[idx], i = 0; dbf != NULL;
dbf = dbf->db_hash_next, i++) {
if (DBUF_EQUAL(dbf, os, obj, level, blkid)) {
mutex_enter(&dbf->db_mtx);
if (dbf->db_state != DB_EVICTING) {
mutex_exit(DBUF_HASH_MUTEX(h, idx));
return (dbf);
}
mutex_exit(&dbf->db_mtx);
}
}
if (i > 0) {
DBUF_STAT_BUMP(hash_collisions);
if (i == 1)
DBUF_STAT_BUMP(hash_chains);
DBUF_STAT_MAX(hash_chain_max, i);
}
mutex_enter(&db->db_mtx);
db->db_hash_next = h->hash_table[idx];
h->hash_table[idx] = db;
mutex_exit(DBUF_HASH_MUTEX(h, idx));
uint64_t he = atomic_inc_64_nv(&dbuf_stats.hash_elements.value.ui64);
DBUF_STAT_MAX(hash_elements_max, he);
return (NULL);
}
/*
* This returns whether this dbuf should be stored in the metadata cache, which
* is based on whether it's from one of the dnode types that store data related
* to traversing dataset hierarchies.
*/
static boolean_t
dbuf_include_in_metadata_cache(dmu_buf_impl_t *db)
{
DB_DNODE_ENTER(db);
dmu_object_type_t type = DB_DNODE(db)->dn_type;
DB_DNODE_EXIT(db);
/* Check if this dbuf is one of the types we care about */
if (DMU_OT_IS_METADATA_CACHED(type)) {
/* If we hit this, then we set something up wrong in dmu_ot */
ASSERT(DMU_OT_IS_METADATA(type));
/*
* Sanity check for small-memory systems: don't allocate too
* much memory for this purpose.
*/
if (zfs_refcount_count(
&dbuf_caches[DB_DBUF_METADATA_CACHE].size) >
dbuf_metadata_cache_target_bytes()) {
DBUF_STAT_BUMP(metadata_cache_overflow);
return (B_FALSE);
}
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Remove an entry from the hash table. It must be in the EVICTING state.
*/
static void
dbuf_hash_remove(dmu_buf_impl_t *db)
{
dbuf_hash_table_t *h = &dbuf_hash_table;
uint64_t idx;
dmu_buf_impl_t *dbf, **dbp;
ASSERT3U(dbuf_hash(db->db_objset, db->db.db_object, db->db_level,
db->db_blkid), ==, db->db_hash);
idx = db->db_hash & h->hash_table_mask;
/*
* We mustn't hold db_mtx to maintain lock ordering:
* DBUF_HASH_MUTEX > db_mtx.
*/
ASSERT(zfs_refcount_is_zero(&db->db_holds));
ASSERT(db->db_state == DB_EVICTING);
ASSERT(!MUTEX_HELD(&db->db_mtx));
mutex_enter(DBUF_HASH_MUTEX(h, idx));
dbp = &h->hash_table[idx];
while ((dbf = *dbp) != db) {
dbp = &dbf->db_hash_next;
ASSERT(dbf != NULL);
}
*dbp = db->db_hash_next;
db->db_hash_next = NULL;
if (h->hash_table[idx] &&
h->hash_table[idx]->db_hash_next == NULL)
DBUF_STAT_BUMPDOWN(hash_chains);
mutex_exit(DBUF_HASH_MUTEX(h, idx));
atomic_dec_64(&dbuf_stats.hash_elements.value.ui64);
}
typedef enum {
DBVU_EVICTING,
DBVU_NOT_EVICTING
} dbvu_verify_type_t;
static void
dbuf_verify_user(dmu_buf_impl_t *db, dbvu_verify_type_t verify_type)
{
#ifdef ZFS_DEBUG
int64_t holds;
if (db->db_user == NULL)
return;
/* Only data blocks support the attachment of user data. */
ASSERT(db->db_level == 0);
/* Clients must resolve a dbuf before attaching user data. */
ASSERT(db->db.db_data != NULL);
ASSERT3U(db->db_state, ==, DB_CACHED);
holds = zfs_refcount_count(&db->db_holds);
if (verify_type == DBVU_EVICTING) {
/*
* Immediate eviction occurs when holds == dirtycnt.
* For normal eviction buffers, holds is zero on
* eviction, except when dbuf_fix_old_data() calls
* dbuf_clear_data(). However, the hold count can grow
* during eviction even though db_mtx is held (see
* dmu_bonus_hold() for an example), so we can only
* test the generic invariant that holds >= dirtycnt.
*/
ASSERT3U(holds, >=, db->db_dirtycnt);
} else {
if (db->db_user_immediate_evict == TRUE)
ASSERT3U(holds, >=, db->db_dirtycnt);
else
ASSERT3U(holds, >, 0);
}
#endif
}
static void
dbuf_evict_user(dmu_buf_impl_t *db)
{
dmu_buf_user_t *dbu = db->db_user;
ASSERT(MUTEX_HELD(&db->db_mtx));
if (dbu == NULL)
return;
dbuf_verify_user(db, DBVU_EVICTING);
db->db_user = NULL;
#ifdef ZFS_DEBUG
if (dbu->dbu_clear_on_evict_dbufp != NULL)
*dbu->dbu_clear_on_evict_dbufp = NULL;
#endif
/*
* There are two eviction callbacks - one that we call synchronously
* and one that we invoke via a taskq. The async one is useful for
* avoiding lock order reversals and limiting stack depth.
*
* Note that if we have a sync callback but no async callback,
* it's likely that the sync callback will free the structure
* containing the dbu. In that case we need to take care to not
* dereference dbu after calling the sync evict func.
*/
boolean_t has_async = (dbu->dbu_evict_func_async != NULL);
if (dbu->dbu_evict_func_sync != NULL)
dbu->dbu_evict_func_sync(dbu);
if (has_async) {
taskq_dispatch_ent(dbu_evict_taskq, dbu->dbu_evict_func_async,
dbu, 0, &dbu->dbu_tqent);
}
}
boolean_t
dbuf_is_metadata(dmu_buf_impl_t *db)
{
/*
* Consider indirect blocks and spill blocks to be meta data.
*/
if (db->db_level > 0 || db->db_blkid == DMU_SPILL_BLKID) {
return (B_TRUE);
} else {
boolean_t is_metadata;
DB_DNODE_ENTER(db);
is_metadata = DMU_OT_IS_METADATA(DB_DNODE(db)->dn_type);
DB_DNODE_EXIT(db);
return (is_metadata);
}
}
/*
* We want to exclude buffers that are on a special allocation class from
* L2ARC.
*/
boolean_t
dbuf_is_l2cacheable(dmu_buf_impl_t *db)
{
if (db->db_objset->os_secondary_cache == ZFS_CACHE_ALL ||
(db->db_objset->os_secondary_cache ==
ZFS_CACHE_METADATA && dbuf_is_metadata(db))) {
if (l2arc_exclude_special == 0)
return (B_TRUE);
blkptr_t *bp = db->db_blkptr;
if (bp == NULL || BP_IS_HOLE(bp))
return (B_FALSE);
uint64_t vdev = DVA_GET_VDEV(bp->blk_dva);
vdev_t *rvd = db->db_objset->os_spa->spa_root_vdev;
vdev_t *vd = NULL;
if (vdev < rvd->vdev_children)
vd = rvd->vdev_child[vdev];
if (vd == NULL)
return (B_TRUE);
if (vd->vdev_alloc_bias != VDEV_BIAS_SPECIAL &&
vd->vdev_alloc_bias != VDEV_BIAS_DEDUP)
return (B_TRUE);
}
return (B_FALSE);
}
static inline boolean_t
dnode_level_is_l2cacheable(blkptr_t *bp, dnode_t *dn, int64_t level)
{
if (dn->dn_objset->os_secondary_cache == ZFS_CACHE_ALL ||
(dn->dn_objset->os_secondary_cache == ZFS_CACHE_METADATA &&
(level > 0 ||
DMU_OT_IS_METADATA(dn->dn_handle->dnh_dnode->dn_type)))) {
if (l2arc_exclude_special == 0)
return (B_TRUE);
if (bp == NULL || BP_IS_HOLE(bp))
return (B_FALSE);
uint64_t vdev = DVA_GET_VDEV(bp->blk_dva);
vdev_t *rvd = dn->dn_objset->os_spa->spa_root_vdev;
vdev_t *vd = NULL;
if (vdev < rvd->vdev_children)
vd = rvd->vdev_child[vdev];
if (vd == NULL)
return (B_TRUE);
if (vd->vdev_alloc_bias != VDEV_BIAS_SPECIAL &&
vd->vdev_alloc_bias != VDEV_BIAS_DEDUP)
return (B_TRUE);
}
return (B_FALSE);
}
/*
* This function *must* return indices evenly distributed between all
* sublists of the multilist. This is needed due to how the dbuf eviction
* code is laid out; dbuf_evict_thread() assumes dbufs are evenly
* distributed between all sublists and uses this assumption when
* deciding which sublist to evict from and how much to evict from it.
*/
static unsigned int
dbuf_cache_multilist_index_func(multilist_t *ml, void *obj)
{
dmu_buf_impl_t *db = obj;
/*
* The assumption here, is the hash value for a given
* dmu_buf_impl_t will remain constant throughout it's lifetime
* (i.e. it's objset, object, level and blkid fields don't change).
* Thus, we don't need to store the dbuf's sublist index
* on insertion, as this index can be recalculated on removal.
*
* Also, the low order bits of the hash value are thought to be
* distributed evenly. Otherwise, in the case that the multilist
* has a power of two number of sublists, each sublists' usage
* would not be evenly distributed. In this context full 64bit
* division would be a waste of time, so limit it to 32 bits.
*/
return ((unsigned int)dbuf_hash(db->db_objset, db->db.db_object,
db->db_level, db->db_blkid) %
multilist_get_num_sublists(ml));
}
/*
* The target size of the dbuf cache can grow with the ARC target,
* unless limited by the tunable dbuf_cache_max_bytes.
*/
static inline unsigned long
dbuf_cache_target_bytes(void)
{
return (MIN(dbuf_cache_max_bytes,
arc_target_bytes() >> dbuf_cache_shift));
}
/*
* The target size of the dbuf metadata cache can grow with the ARC target,
* unless limited by the tunable dbuf_metadata_cache_max_bytes.
*/
static inline unsigned long
dbuf_metadata_cache_target_bytes(void)
{
return (MIN(dbuf_metadata_cache_max_bytes,
arc_target_bytes() >> dbuf_metadata_cache_shift));
}
static inline uint64_t
dbuf_cache_hiwater_bytes(void)
{
uint64_t dbuf_cache_target = dbuf_cache_target_bytes();
return (dbuf_cache_target +
(dbuf_cache_target * dbuf_cache_hiwater_pct) / 100);
}
static inline uint64_t
dbuf_cache_lowater_bytes(void)
{
uint64_t dbuf_cache_target = dbuf_cache_target_bytes();
return (dbuf_cache_target -
(dbuf_cache_target * dbuf_cache_lowater_pct) / 100);
}
static inline boolean_t
dbuf_cache_above_lowater(void)
{
return (zfs_refcount_count(&dbuf_caches[DB_DBUF_CACHE].size) >
dbuf_cache_lowater_bytes());
}
/*
* Evict the oldest eligible dbuf from the dbuf cache.
*/
static void
dbuf_evict_one(void)
{
int idx = multilist_get_random_index(&dbuf_caches[DB_DBUF_CACHE].cache);
multilist_sublist_t *mls = multilist_sublist_lock(
&dbuf_caches[DB_DBUF_CACHE].cache, idx);
ASSERT(!MUTEX_HELD(&dbuf_evict_lock));
dmu_buf_impl_t *db = multilist_sublist_tail(mls);
while (db != NULL && mutex_tryenter(&db->db_mtx) == 0) {
db = multilist_sublist_prev(mls, db);
}
DTRACE_PROBE2(dbuf__evict__one, dmu_buf_impl_t *, db,
multilist_sublist_t *, mls);
if (db != NULL) {
multilist_sublist_remove(mls, db);
multilist_sublist_unlock(mls);
(void) zfs_refcount_remove_many(
&dbuf_caches[DB_DBUF_CACHE].size, db->db.db_size, db);
DBUF_STAT_BUMPDOWN(cache_levels[db->db_level]);
DBUF_STAT_BUMPDOWN(cache_count);
DBUF_STAT_DECR(cache_levels_bytes[db->db_level],
db->db.db_size);
ASSERT3U(db->db_caching_status, ==, DB_DBUF_CACHE);
db->db_caching_status = DB_NO_CACHE;
dbuf_destroy(db);
DBUF_STAT_BUMP(cache_total_evicts);
} else {
multilist_sublist_unlock(mls);
}
}
/*
* The dbuf evict thread is responsible for aging out dbufs from the
* cache. Once the cache has reached it's maximum size, dbufs are removed
* and destroyed. The eviction thread will continue running until the size
* of the dbuf cache is at or below the maximum size. Once the dbuf is aged
* out of the cache it is destroyed and becomes eligible for arc eviction.
*/
static __attribute__((noreturn)) void
dbuf_evict_thread(void *unused)
{
(void) unused;
callb_cpr_t cpr;
CALLB_CPR_INIT(&cpr, &dbuf_evict_lock, callb_generic_cpr, FTAG);
mutex_enter(&dbuf_evict_lock);
while (!dbuf_evict_thread_exit) {
while (!dbuf_cache_above_lowater() && !dbuf_evict_thread_exit) {
CALLB_CPR_SAFE_BEGIN(&cpr);
(void) cv_timedwait_idle_hires(&dbuf_evict_cv,
&dbuf_evict_lock, SEC2NSEC(1), MSEC2NSEC(1), 0);
CALLB_CPR_SAFE_END(&cpr, &dbuf_evict_lock);
}
mutex_exit(&dbuf_evict_lock);
/*
* Keep evicting as long as we're above the low water mark
* for the cache. We do this without holding the locks to
* minimize lock contention.
*/
while (dbuf_cache_above_lowater() && !dbuf_evict_thread_exit) {
dbuf_evict_one();
}
mutex_enter(&dbuf_evict_lock);
}
dbuf_evict_thread_exit = B_FALSE;
cv_broadcast(&dbuf_evict_cv);
CALLB_CPR_EXIT(&cpr); /* drops dbuf_evict_lock */
thread_exit();
}
/*
* Wake up the dbuf eviction thread if the dbuf cache is at its max size.
* If the dbuf cache is at its high water mark, then evict a dbuf from the
* dbuf cache using the caller's context.
*/
static void
dbuf_evict_notify(uint64_t size)
{
/*
* We check if we should evict without holding the dbuf_evict_lock,
* because it's OK to occasionally make the wrong decision here,
* and grabbing the lock results in massive lock contention.
*/
if (size > dbuf_cache_target_bytes()) {
if (size > dbuf_cache_hiwater_bytes())
dbuf_evict_one();
cv_signal(&dbuf_evict_cv);
}
}
static int
dbuf_kstat_update(kstat_t *ksp, int rw)
{
dbuf_stats_t *ds = ksp->ks_data;
dbuf_hash_table_t *h = &dbuf_hash_table;
if (rw == KSTAT_WRITE)
return (SET_ERROR(EACCES));
ds->cache_count.value.ui64 =
wmsum_value(&dbuf_sums.cache_count);
ds->cache_size_bytes.value.ui64 =
zfs_refcount_count(&dbuf_caches[DB_DBUF_CACHE].size);
ds->cache_target_bytes.value.ui64 = dbuf_cache_target_bytes();
ds->cache_hiwater_bytes.value.ui64 = dbuf_cache_hiwater_bytes();
ds->cache_lowater_bytes.value.ui64 = dbuf_cache_lowater_bytes();
ds->cache_total_evicts.value.ui64 =
wmsum_value(&dbuf_sums.cache_total_evicts);
for (int i = 0; i < DN_MAX_LEVELS; i++) {
ds->cache_levels[i].value.ui64 =
wmsum_value(&dbuf_sums.cache_levels[i]);
ds->cache_levels_bytes[i].value.ui64 =
wmsum_value(&dbuf_sums.cache_levels_bytes[i]);
}
ds->hash_hits.value.ui64 =
wmsum_value(&dbuf_sums.hash_hits);
ds->hash_misses.value.ui64 =
wmsum_value(&dbuf_sums.hash_misses);
ds->hash_collisions.value.ui64 =
wmsum_value(&dbuf_sums.hash_collisions);
ds->hash_chains.value.ui64 =
wmsum_value(&dbuf_sums.hash_chains);
ds->hash_insert_race.value.ui64 =
wmsum_value(&dbuf_sums.hash_insert_race);
ds->hash_table_count.value.ui64 = h->hash_table_mask + 1;
ds->hash_mutex_count.value.ui64 = h->hash_mutex_mask + 1;
ds->metadata_cache_count.value.ui64 =
wmsum_value(&dbuf_sums.metadata_cache_count);
ds->metadata_cache_size_bytes.value.ui64 = zfs_refcount_count(
&dbuf_caches[DB_DBUF_METADATA_CACHE].size);
ds->metadata_cache_overflow.value.ui64 =
wmsum_value(&dbuf_sums.metadata_cache_overflow);
return (0);
}
void
dbuf_init(void)
{
uint64_t hmsize, hsize = 1ULL << 16;
dbuf_hash_table_t *h = &dbuf_hash_table;
/*
* The hash table is big enough to fill one eighth of physical memory
* with an average block size of zfs_arc_average_blocksize (default 8K).
* By default, the table will take up
* totalmem * sizeof(void*) / 8K (1MB per GB with 8-byte pointers).
*/
while (hsize * zfs_arc_average_blocksize < arc_all_memory() / 8)
hsize <<= 1;
h->hash_table = NULL;
while (h->hash_table == NULL) {
h->hash_table_mask = hsize - 1;
h->hash_table = vmem_zalloc(hsize * sizeof (void *), KM_SLEEP);
if (h->hash_table == NULL)
hsize >>= 1;
ASSERT3U(hsize, >=, 1ULL << 10);
}
/*
* The hash table buckets are protected by an array of mutexes where
* each mutex is reponsible for protecting 128 buckets. A minimum
* array size of 8192 is targeted to avoid contention.
*/
if (dbuf_mutex_cache_shift == 0)
hmsize = MAX(hsize >> 7, 1ULL << 13);
else
hmsize = 1ULL << MIN(dbuf_mutex_cache_shift, 24);
h->hash_mutexes = NULL;
while (h->hash_mutexes == NULL) {
h->hash_mutex_mask = hmsize - 1;
h->hash_mutexes = vmem_zalloc(hmsize * sizeof (kmutex_t),
KM_SLEEP);
if (h->hash_mutexes == NULL)
hmsize >>= 1;
}
dbuf_kmem_cache = kmem_cache_create("dmu_buf_impl_t",
sizeof (dmu_buf_impl_t),
0, dbuf_cons, dbuf_dest, NULL, NULL, NULL, 0);
for (int i = 0; i < hmsize; i++)
mutex_init(&h->hash_mutexes[i], NULL, MUTEX_DEFAULT, NULL);
dbuf_stats_init(h);
/*
* All entries are queued via taskq_dispatch_ent(), so min/maxalloc
* configuration is not required.
*/
dbu_evict_taskq = taskq_create("dbu_evict", 1, defclsyspri, 0, 0, 0);
for (dbuf_cached_state_t dcs = 0; dcs < DB_CACHE_MAX; dcs++) {
multilist_create(&dbuf_caches[dcs].cache,
sizeof (dmu_buf_impl_t),
offsetof(dmu_buf_impl_t, db_cache_link),
dbuf_cache_multilist_index_func);
zfs_refcount_create(&dbuf_caches[dcs].size);
}
dbuf_evict_thread_exit = B_FALSE;
mutex_init(&dbuf_evict_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&dbuf_evict_cv, NULL, CV_DEFAULT, NULL);
dbuf_cache_evict_thread = thread_create(NULL, 0, dbuf_evict_thread,
NULL, 0, &p0, TS_RUN, minclsyspri);
wmsum_init(&dbuf_sums.cache_count, 0);
wmsum_init(&dbuf_sums.cache_total_evicts, 0);
for (int i = 0; i < DN_MAX_LEVELS; i++) {
wmsum_init(&dbuf_sums.cache_levels[i], 0);
wmsum_init(&dbuf_sums.cache_levels_bytes[i], 0);
}
wmsum_init(&dbuf_sums.hash_hits, 0);
wmsum_init(&dbuf_sums.hash_misses, 0);
wmsum_init(&dbuf_sums.hash_collisions, 0);
wmsum_init(&dbuf_sums.hash_chains, 0);
wmsum_init(&dbuf_sums.hash_insert_race, 0);
wmsum_init(&dbuf_sums.metadata_cache_count, 0);
wmsum_init(&dbuf_sums.metadata_cache_overflow, 0);
dbuf_ksp = kstat_create("zfs", 0, "dbufstats", "misc",
KSTAT_TYPE_NAMED, sizeof (dbuf_stats) / sizeof (kstat_named_t),
KSTAT_FLAG_VIRTUAL);
if (dbuf_ksp != NULL) {
for (int i = 0; i < DN_MAX_LEVELS; i++) {
snprintf(dbuf_stats.cache_levels[i].name,
KSTAT_STRLEN, "cache_level_%d", i);
dbuf_stats.cache_levels[i].data_type =
KSTAT_DATA_UINT64;
snprintf(dbuf_stats.cache_levels_bytes[i].name,
KSTAT_STRLEN, "cache_level_%d_bytes", i);
dbuf_stats.cache_levels_bytes[i].data_type =
KSTAT_DATA_UINT64;
}
dbuf_ksp->ks_data = &dbuf_stats;
dbuf_ksp->ks_update = dbuf_kstat_update;
kstat_install(dbuf_ksp);
}
}
void
dbuf_fini(void)
{
dbuf_hash_table_t *h = &dbuf_hash_table;
dbuf_stats_destroy();
for (int i = 0; i < (h->hash_mutex_mask + 1); i++)
mutex_destroy(&h->hash_mutexes[i]);
vmem_free(h->hash_table, (h->hash_table_mask + 1) * sizeof (void *));
vmem_free(h->hash_mutexes, (h->hash_mutex_mask + 1) *
sizeof (kmutex_t));
kmem_cache_destroy(dbuf_kmem_cache);
taskq_destroy(dbu_evict_taskq);
mutex_enter(&dbuf_evict_lock);
dbuf_evict_thread_exit = B_TRUE;
while (dbuf_evict_thread_exit) {
cv_signal(&dbuf_evict_cv);
cv_wait(&dbuf_evict_cv, &dbuf_evict_lock);
}
mutex_exit(&dbuf_evict_lock);
mutex_destroy(&dbuf_evict_lock);
cv_destroy(&dbuf_evict_cv);
for (dbuf_cached_state_t dcs = 0; dcs < DB_CACHE_MAX; dcs++) {
zfs_refcount_destroy(&dbuf_caches[dcs].size);
multilist_destroy(&dbuf_caches[dcs].cache);
}
if (dbuf_ksp != NULL) {
kstat_delete(dbuf_ksp);
dbuf_ksp = NULL;
}
wmsum_fini(&dbuf_sums.cache_count);
wmsum_fini(&dbuf_sums.cache_total_evicts);
for (int i = 0; i < DN_MAX_LEVELS; i++) {
wmsum_fini(&dbuf_sums.cache_levels[i]);
wmsum_fini(&dbuf_sums.cache_levels_bytes[i]);
}
wmsum_fini(&dbuf_sums.hash_hits);
wmsum_fini(&dbuf_sums.hash_misses);
wmsum_fini(&dbuf_sums.hash_collisions);
wmsum_fini(&dbuf_sums.hash_chains);
wmsum_fini(&dbuf_sums.hash_insert_race);
wmsum_fini(&dbuf_sums.metadata_cache_count);
wmsum_fini(&dbuf_sums.metadata_cache_overflow);
}
/*
* Other stuff.
*/
#ifdef ZFS_DEBUG
static void
dbuf_verify(dmu_buf_impl_t *db)
{
dnode_t *dn;
dbuf_dirty_record_t *dr;
uint32_t txg_prev;
ASSERT(MUTEX_HELD(&db->db_mtx));
if (!(zfs_flags & ZFS_DEBUG_DBUF_VERIFY))
return;
ASSERT(db->db_objset != NULL);
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
if (dn == NULL) {
ASSERT(db->db_parent == NULL);
ASSERT(db->db_blkptr == NULL);
} else {
ASSERT3U(db->db.db_object, ==, dn->dn_object);
ASSERT3P(db->db_objset, ==, dn->dn_objset);
ASSERT3U(db->db_level, <, dn->dn_nlevels);
ASSERT(db->db_blkid == DMU_BONUS_BLKID ||
db->db_blkid == DMU_SPILL_BLKID ||
!avl_is_empty(&dn->dn_dbufs));
}
if (db->db_blkid == DMU_BONUS_BLKID) {
ASSERT(dn != NULL);
ASSERT3U(db->db.db_size, >=, dn->dn_bonuslen);
ASSERT3U(db->db.db_offset, ==, DMU_BONUS_BLKID);
} else if (db->db_blkid == DMU_SPILL_BLKID) {
ASSERT(dn != NULL);
ASSERT0(db->db.db_offset);
} else {
ASSERT3U(db->db.db_offset, ==, db->db_blkid * db->db.db_size);
}
if ((dr = list_head(&db->db_dirty_records)) != NULL) {
ASSERT(dr->dr_dbuf == db);
txg_prev = dr->dr_txg;
for (dr = list_next(&db->db_dirty_records, dr); dr != NULL;
dr = list_next(&db->db_dirty_records, dr)) {
ASSERT(dr->dr_dbuf == db);
ASSERT(txg_prev > dr->dr_txg);
txg_prev = dr->dr_txg;
}
}
/*
* We can't assert that db_size matches dn_datablksz because it
* can be momentarily different when another thread is doing
* dnode_set_blksz().
*/
if (db->db_level == 0 && db->db.db_object == DMU_META_DNODE_OBJECT) {
dr = db->db_data_pending;
/*
* It should only be modified in syncing context, so
* make sure we only have one copy of the data.
*/
ASSERT(dr == NULL || dr->dt.dl.dr_data == db->db_buf);
}
/* verify db->db_blkptr */
if (db->db_blkptr) {
if (db->db_parent == dn->dn_dbuf) {
/* db is pointed to by the dnode */
/* ASSERT3U(db->db_blkid, <, dn->dn_nblkptr); */
if (DMU_OBJECT_IS_SPECIAL(db->db.db_object))
ASSERT(db->db_parent == NULL);
else
ASSERT(db->db_parent != NULL);
if (db->db_blkid != DMU_SPILL_BLKID)
ASSERT3P(db->db_blkptr, ==,
&dn->dn_phys->dn_blkptr[db->db_blkid]);
} else {
/* db is pointed to by an indirect block */
int epb __maybe_unused = db->db_parent->db.db_size >>
SPA_BLKPTRSHIFT;
ASSERT3U(db->db_parent->db_level, ==, db->db_level+1);
ASSERT3U(db->db_parent->db.db_object, ==,
db->db.db_object);
/*
* dnode_grow_indblksz() can make this fail if we don't
* have the parent's rwlock. XXX indblksz no longer
* grows. safe to do this now?
*/
if (RW_LOCK_HELD(&db->db_parent->db_rwlock)) {
ASSERT3P(db->db_blkptr, ==,
((blkptr_t *)db->db_parent->db.db_data +
db->db_blkid % epb));
}
}
}
if ((db->db_blkptr == NULL || BP_IS_HOLE(db->db_blkptr)) &&
(db->db_buf == NULL || db->db_buf->b_data) &&
db->db.db_data && db->db_blkid != DMU_BONUS_BLKID &&
db->db_state != DB_FILL && (dn == NULL || !dn->dn_free_txg)) {
/*
* If the blkptr isn't set but they have nonzero data,
* it had better be dirty, otherwise we'll lose that
* data when we evict this buffer.
*
* There is an exception to this rule for indirect blocks; in
* this case, if the indirect block is a hole, we fill in a few
* fields on each of the child blocks (importantly, birth time)
* to prevent hole birth times from being lost when you
* partially fill in a hole.
*/
if (db->db_dirtycnt == 0) {
if (db->db_level == 0) {
uint64_t *buf = db->db.db_data;
int i;
for (i = 0; i < db->db.db_size >> 3; i++) {
ASSERT(buf[i] == 0);
}
} else {
blkptr_t *bps = db->db.db_data;
ASSERT3U(1 << DB_DNODE(db)->dn_indblkshift, ==,
db->db.db_size);
/*
* We want to verify that all the blkptrs in the
* indirect block are holes, but we may have
* automatically set up a few fields for them.
* We iterate through each blkptr and verify
* they only have those fields set.
*/
for (int i = 0;
i < db->db.db_size / sizeof (blkptr_t);
i++) {
blkptr_t *bp = &bps[i];
ASSERT(ZIO_CHECKSUM_IS_ZERO(
&bp->blk_cksum));
ASSERT(
DVA_IS_EMPTY(&bp->blk_dva[0]) &&
DVA_IS_EMPTY(&bp->blk_dva[1]) &&
DVA_IS_EMPTY(&bp->blk_dva[2]));
ASSERT0(bp->blk_fill);
ASSERT0(bp->blk_pad[0]);
ASSERT0(bp->blk_pad[1]);
ASSERT(!BP_IS_EMBEDDED(bp));
ASSERT(BP_IS_HOLE(bp));
ASSERT0(bp->blk_phys_birth);
}
}
}
}
DB_DNODE_EXIT(db);
}
#endif
static void
dbuf_clear_data(dmu_buf_impl_t *db)
{
ASSERT(MUTEX_HELD(&db->db_mtx));
dbuf_evict_user(db);
ASSERT3P(db->db_buf, ==, NULL);
db->db.db_data = NULL;
if (db->db_state != DB_NOFILL) {
db->db_state = DB_UNCACHED;
DTRACE_SET_STATE(db, "clear data");
}
}
static void
dbuf_set_data(dmu_buf_impl_t *db, arc_buf_t *buf)
{
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT(buf != NULL);
db->db_buf = buf;
ASSERT(buf->b_data != NULL);
db->db.db_data = buf->b_data;
}
static arc_buf_t *
dbuf_alloc_arcbuf(dmu_buf_impl_t *db)
{
spa_t *spa = db->db_objset->os_spa;
return (arc_alloc_buf(spa, db, DBUF_GET_BUFC_TYPE(db), db->db.db_size));
}
/*
* Loan out an arc_buf for read. Return the loaned arc_buf.
*/
arc_buf_t *
dbuf_loan_arcbuf(dmu_buf_impl_t *db)
{
arc_buf_t *abuf;
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
mutex_enter(&db->db_mtx);
if (arc_released(db->db_buf) || zfs_refcount_count(&db->db_holds) > 1) {
int blksz = db->db.db_size;
spa_t *spa = db->db_objset->os_spa;
mutex_exit(&db->db_mtx);
abuf = arc_loan_buf(spa, B_FALSE, blksz);
memcpy(abuf->b_data, db->db.db_data, blksz);
} else {
abuf = db->db_buf;
arc_loan_inuse_buf(abuf, db);
db->db_buf = NULL;
dbuf_clear_data(db);
mutex_exit(&db->db_mtx);
}
return (abuf);
}
/*
* Calculate which level n block references the data at the level 0 offset
* provided.
*/
uint64_t
dbuf_whichblock(const dnode_t *dn, const int64_t level, const uint64_t offset)
{
if (dn->dn_datablkshift != 0 && dn->dn_indblkshift != 0) {
/*
* The level n blkid is equal to the level 0 blkid divided by
* the number of level 0s in a level n block.
*
* The level 0 blkid is offset >> datablkshift =
* offset / 2^datablkshift.
*
* The number of level 0s in a level n is the number of block
* pointers in an indirect block, raised to the power of level.
* This is 2^(indblkshift - SPA_BLKPTRSHIFT)^level =
* 2^(level*(indblkshift - SPA_BLKPTRSHIFT)).
*
* Thus, the level n blkid is: offset /
* ((2^datablkshift)*(2^(level*(indblkshift-SPA_BLKPTRSHIFT))))
* = offset / 2^(datablkshift + level *
* (indblkshift - SPA_BLKPTRSHIFT))
* = offset >> (datablkshift + level *
* (indblkshift - SPA_BLKPTRSHIFT))
*/
const unsigned exp = dn->dn_datablkshift +
level * (dn->dn_indblkshift - SPA_BLKPTRSHIFT);
if (exp >= 8 * sizeof (offset)) {
/* This only happens on the highest indirection level */
ASSERT3U(level, ==, dn->dn_nlevels - 1);
return (0);
}
ASSERT3U(exp, <, 8 * sizeof (offset));
return (offset >> exp);
} else {
ASSERT3U(offset, <, dn->dn_datablksz);
return (0);
}
}
/*
* This function is used to lock the parent of the provided dbuf. This should be
* used when modifying or reading db_blkptr.
*/
db_lock_type_t
dmu_buf_lock_parent(dmu_buf_impl_t *db, krw_t rw, const void *tag)
{
enum db_lock_type ret = DLT_NONE;
if (db->db_parent != NULL) {
rw_enter(&db->db_parent->db_rwlock, rw);
ret = DLT_PARENT;
} else if (dmu_objset_ds(db->db_objset) != NULL) {
rrw_enter(&dmu_objset_ds(db->db_objset)->ds_bp_rwlock, rw,
tag);
ret = DLT_OBJSET;
}
/*
* We only return a DLT_NONE lock when it's the top-most indirect block
* of the meta-dnode of the MOS.
*/
return (ret);
}
/*
* We need to pass the lock type in because it's possible that the block will
* move from being the topmost indirect block in a dnode (and thus, have no
* parent) to not the top-most via an indirection increase. This would cause a
* panic if we didn't pass the lock type in.
*/
void
dmu_buf_unlock_parent(dmu_buf_impl_t *db, db_lock_type_t type, const void *tag)
{
if (type == DLT_PARENT)
rw_exit(&db->db_parent->db_rwlock);
else if (type == DLT_OBJSET)
rrw_exit(&dmu_objset_ds(db->db_objset)->ds_bp_rwlock, tag);
}
static void
dbuf_read_done(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp,
arc_buf_t *buf, void *vdb)
{
(void) zb, (void) bp;
dmu_buf_impl_t *db = vdb;
mutex_enter(&db->db_mtx);
ASSERT3U(db->db_state, ==, DB_READ);
/*
* All reads are synchronous, so we must have a hold on the dbuf
*/
ASSERT(zfs_refcount_count(&db->db_holds) > 0);
ASSERT(db->db_buf == NULL);
ASSERT(db->db.db_data == NULL);
if (buf == NULL) {
/* i/o error */
ASSERT(zio == NULL || zio->io_error != 0);
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
ASSERT3P(db->db_buf, ==, NULL);
db->db_state = DB_UNCACHED;
DTRACE_SET_STATE(db, "i/o error");
} else if (db->db_level == 0 && db->db_freed_in_flight) {
/* freed in flight */
ASSERT(zio == NULL || zio->io_error == 0);
arc_release(buf, db);
memset(buf->b_data, 0, db->db.db_size);
arc_buf_freeze(buf);
db->db_freed_in_flight = FALSE;
dbuf_set_data(db, buf);
db->db_state = DB_CACHED;
DTRACE_SET_STATE(db, "freed in flight");
} else {
/* success */
ASSERT(zio == NULL || zio->io_error == 0);
dbuf_set_data(db, buf);
db->db_state = DB_CACHED;
DTRACE_SET_STATE(db, "successful read");
}
cv_broadcast(&db->db_changed);
dbuf_rele_and_unlock(db, NULL, B_FALSE);
}
/*
* Shortcut for performing reads on bonus dbufs. Returns
* an error if we fail to verify the dnode associated with
* a decrypted block. Otherwise success.
*/
static int
dbuf_read_bonus(dmu_buf_impl_t *db, dnode_t *dn, uint32_t flags)
{
int bonuslen, max_bonuslen, err;
err = dbuf_read_verify_dnode_crypt(db, flags);
if (err)
return (err);
bonuslen = MIN(dn->dn_bonuslen, dn->dn_phys->dn_bonuslen);
max_bonuslen = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots);
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT(DB_DNODE_HELD(db));
ASSERT3U(bonuslen, <=, db->db.db_size);
db->db.db_data = kmem_alloc(max_bonuslen, KM_SLEEP);
arc_space_consume(max_bonuslen, ARC_SPACE_BONUS);
if (bonuslen < max_bonuslen)
memset(db->db.db_data, 0, max_bonuslen);
if (bonuslen)
memcpy(db->db.db_data, DN_BONUS(dn->dn_phys), bonuslen);
db->db_state = DB_CACHED;
DTRACE_SET_STATE(db, "bonus buffer filled");
return (0);
}
static void
dbuf_handle_indirect_hole(dmu_buf_impl_t *db, dnode_t *dn, blkptr_t *dbbp)
{
blkptr_t *bps = db->db.db_data;
uint32_t indbs = 1ULL << dn->dn_indblkshift;
int n_bps = indbs >> SPA_BLKPTRSHIFT;
for (int i = 0; i < n_bps; i++) {
blkptr_t *bp = &bps[i];
ASSERT3U(BP_GET_LSIZE(dbbp), ==, indbs);
BP_SET_LSIZE(bp, BP_GET_LEVEL(dbbp) == 1 ?
dn->dn_datablksz : BP_GET_LSIZE(dbbp));
BP_SET_TYPE(bp, BP_GET_TYPE(dbbp));
BP_SET_LEVEL(bp, BP_GET_LEVEL(dbbp) - 1);
BP_SET_BIRTH(bp, dbbp->blk_birth, 0);
}
}
/*
* Handle reads on dbufs that are holes, if necessary. This function
* requires that the dbuf's mutex is held. Returns success (0) if action
* was taken, ENOENT if no action was taken.
*/
static int
dbuf_read_hole(dmu_buf_impl_t *db, dnode_t *dn, blkptr_t *bp)
{
ASSERT(MUTEX_HELD(&db->db_mtx));
int is_hole = bp == NULL || BP_IS_HOLE(bp);
/*
* For level 0 blocks only, if the above check fails:
* Recheck BP_IS_HOLE() after dnode_block_freed() in case dnode_sync()
* processes the delete record and clears the bp while we are waiting
* for the dn_mtx (resulting in a "no" from block_freed).
*/
if (!is_hole && db->db_level == 0)
is_hole = dnode_block_freed(dn, db->db_blkid) || BP_IS_HOLE(bp);
if (is_hole) {
dbuf_set_data(db, dbuf_alloc_arcbuf(db));
memset(db->db.db_data, 0, db->db.db_size);
if (bp != NULL && db->db_level > 0 && BP_IS_HOLE(bp) &&
bp->blk_birth != 0) {
dbuf_handle_indirect_hole(db, dn, bp);
}
db->db_state = DB_CACHED;
DTRACE_SET_STATE(db, "hole read satisfied");
return (0);
}
return (ENOENT);
}
/*
* This function ensures that, when doing a decrypting read of a block,
* we make sure we have decrypted the dnode associated with it. We must do
* this so that we ensure we are fully authenticating the checksum-of-MACs
* tree from the root of the objset down to this block. Indirect blocks are
* always verified against their secure checksum-of-MACs assuming that the
* dnode containing them is correct. Now that we are doing a decrypting read,
* we can be sure that the key is loaded and verify that assumption. This is
* especially important considering that we always read encrypted dnode
* blocks as raw data (without verifying their MACs) to start, and
* decrypt / authenticate them when we need to read an encrypted bonus buffer.
*/
static int
dbuf_read_verify_dnode_crypt(dmu_buf_impl_t *db, uint32_t flags)
{
int err = 0;
objset_t *os = db->db_objset;
arc_buf_t *dnode_abuf;
dnode_t *dn;
zbookmark_phys_t zb;
ASSERT(MUTEX_HELD(&db->db_mtx));
if ((flags & DB_RF_NO_DECRYPT) != 0 ||
!os->os_encrypted || os->os_raw_receive)
return (0);
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
dnode_abuf = (dn->dn_dbuf != NULL) ? dn->dn_dbuf->db_buf : NULL;
if (dnode_abuf == NULL || !arc_is_encrypted(dnode_abuf)) {
DB_DNODE_EXIT(db);
return (0);
}
SET_BOOKMARK(&zb, dmu_objset_id(os),
DMU_META_DNODE_OBJECT, 0, dn->dn_dbuf->db_blkid);
err = arc_untransform(dnode_abuf, os->os_spa, &zb, B_TRUE);
/*
* An error code of EACCES tells us that the key is still not
* available. This is ok if we are only reading authenticated
* (and therefore non-encrypted) blocks.
*/
if (err == EACCES && ((db->db_blkid != DMU_BONUS_BLKID &&
!DMU_OT_IS_ENCRYPTED(dn->dn_type)) ||
(db->db_blkid == DMU_BONUS_BLKID &&
!DMU_OT_IS_ENCRYPTED(dn->dn_bonustype))))
err = 0;
DB_DNODE_EXIT(db);
return (err);
}
/*
* Drops db_mtx and the parent lock specified by dblt and tag before
* returning.
*/
static int
dbuf_read_impl(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags,
db_lock_type_t dblt, const void *tag)
{
dnode_t *dn;
zbookmark_phys_t zb;
uint32_t aflags = ARC_FLAG_NOWAIT;
int err, zio_flags;
blkptr_t bp, *bpp;
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT(db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL);
ASSERT(db->db_buf == NULL);
ASSERT(db->db_parent == NULL ||
RW_LOCK_HELD(&db->db_parent->db_rwlock));
if (db->db_blkid == DMU_BONUS_BLKID) {
err = dbuf_read_bonus(db, dn, flags);
goto early_unlock;
}
if (db->db_state == DB_UNCACHED) {
if (db->db_blkptr == NULL) {
bpp = NULL;
} else {
bp = *db->db_blkptr;
bpp = &bp;
}
} else {
struct dirty_leaf *dl;
dbuf_dirty_record_t *dr;
ASSERT3S(db->db_state, ==, DB_NOFILL);
dr = list_head(&db->db_dirty_records);
if (dr == NULL) {
err = EIO;
goto early_unlock;
} else {
dl = &dr->dt.dl;
if (!dl->dr_brtwrite) {
err = EIO;
goto early_unlock;
}
bp = dl->dr_overridden_by;
bpp = &bp;
}
}
err = dbuf_read_hole(db, dn, bpp);
if (err == 0)
goto early_unlock;
ASSERT(bpp != NULL);
/*
* Any attempt to read a redacted block should result in an error. This
* will never happen under normal conditions, but can be useful for
* debugging purposes.
*/
if (BP_IS_REDACTED(bpp)) {
ASSERT(dsl_dataset_feature_is_active(
db->db_objset->os_dsl_dataset,
SPA_FEATURE_REDACTED_DATASETS));
err = SET_ERROR(EIO);
goto early_unlock;
}
SET_BOOKMARK(&zb, dmu_objset_id(db->db_objset),
db->db.db_object, db->db_level, db->db_blkid);
/*
* All bps of an encrypted os should have the encryption bit set.
* If this is not true it indicates tampering and we report an error.
*/
if (db->db_objset->os_encrypted && !BP_USES_CRYPT(bpp)) {
- spa_log_error(db->db_objset->os_spa, &zb,
- &db->db_blkptr->blk_birth);
+ spa_log_error(db->db_objset->os_spa, &zb, &bpp->blk_birth);
zfs_panic_recover("unencrypted block in encrypted "
"object set %llu", dmu_objset_id(db->db_objset));
err = SET_ERROR(EIO);
goto early_unlock;
}
err = dbuf_read_verify_dnode_crypt(db, flags);
if (err != 0)
goto early_unlock;
DB_DNODE_EXIT(db);
db->db_state = DB_READ;
DTRACE_SET_STATE(db, "read issued");
mutex_exit(&db->db_mtx);
if (!DBUF_IS_CACHEABLE(db))
aflags |= ARC_FLAG_UNCACHED;
else if (dbuf_is_l2cacheable(db))
aflags |= ARC_FLAG_L2CACHE;
dbuf_add_ref(db, NULL);
zio_flags = (flags & DB_RF_CANFAIL) ?
ZIO_FLAG_CANFAIL : ZIO_FLAG_MUSTSUCCEED;
if ((flags & DB_RF_NO_DECRYPT) && BP_IS_PROTECTED(db->db_blkptr))
zio_flags |= ZIO_FLAG_RAW;
/*
* The zio layer will copy the provided blkptr later, but we have our
* own copy so that we can release the parent's rwlock. We have to
* do that so that if dbuf_read_done is called synchronously (on
* an l1 cache hit) we don't acquire the db_mtx while holding the
* parent's rwlock, which would be a lock ordering violation.
*/
dmu_buf_unlock_parent(db, dblt, tag);
(void) arc_read(zio, db->db_objset->os_spa, bpp,
dbuf_read_done, db, ZIO_PRIORITY_SYNC_READ, zio_flags,
&aflags, &zb);
return (err);
early_unlock:
DB_DNODE_EXIT(db);
mutex_exit(&db->db_mtx);
dmu_buf_unlock_parent(db, dblt, tag);
return (err);
}
/*
* This is our just-in-time copy function. It makes a copy of buffers that
* have been modified in a previous transaction group before we access them in
* the current active group.
*
* This function is used in three places: when we are dirtying a buffer for the
* first time in a txg, when we are freeing a range in a dnode that includes
* this buffer, and when we are accessing a buffer which was received compressed
* and later referenced in a WRITE_BYREF record.
*
* Note that when we are called from dbuf_free_range() we do not put a hold on
* the buffer, we just traverse the active dbuf list for the dnode.
*/
static void
dbuf_fix_old_data(dmu_buf_impl_t *db, uint64_t txg)
{
dbuf_dirty_record_t *dr = list_head(&db->db_dirty_records);
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT(db->db.db_data != NULL);
ASSERT(db->db_level == 0);
ASSERT(db->db.db_object != DMU_META_DNODE_OBJECT);
if (dr == NULL ||
(dr->dt.dl.dr_data !=
((db->db_blkid == DMU_BONUS_BLKID) ? db->db.db_data : db->db_buf)))
return;
/*
* If the last dirty record for this dbuf has not yet synced
* and its referencing the dbuf data, either:
* reset the reference to point to a new copy,
* or (if there a no active holders)
* just null out the current db_data pointer.
*/
ASSERT3U(dr->dr_txg, >=, txg - 2);
if (db->db_blkid == DMU_BONUS_BLKID) {
dnode_t *dn = DB_DNODE(db);
int bonuslen = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots);
dr->dt.dl.dr_data = kmem_alloc(bonuslen, KM_SLEEP);
arc_space_consume(bonuslen, ARC_SPACE_BONUS);
memcpy(dr->dt.dl.dr_data, db->db.db_data, bonuslen);
} else if (zfs_refcount_count(&db->db_holds) > db->db_dirtycnt) {
dnode_t *dn = DB_DNODE(db);
int size = arc_buf_size(db->db_buf);
arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db);
spa_t *spa = db->db_objset->os_spa;
enum zio_compress compress_type =
arc_get_compression(db->db_buf);
uint8_t complevel = arc_get_complevel(db->db_buf);
if (arc_is_encrypted(db->db_buf)) {
boolean_t byteorder;
uint8_t salt[ZIO_DATA_SALT_LEN];
uint8_t iv[ZIO_DATA_IV_LEN];
uint8_t mac[ZIO_DATA_MAC_LEN];
arc_get_raw_params(db->db_buf, &byteorder, salt,
iv, mac);
dr->dt.dl.dr_data = arc_alloc_raw_buf(spa, db,
dmu_objset_id(dn->dn_objset), byteorder, salt, iv,
mac, dn->dn_type, size, arc_buf_lsize(db->db_buf),
compress_type, complevel);
} else if (compress_type != ZIO_COMPRESS_OFF) {
ASSERT3U(type, ==, ARC_BUFC_DATA);
dr->dt.dl.dr_data = arc_alloc_compressed_buf(spa, db,
size, arc_buf_lsize(db->db_buf), compress_type,
complevel);
} else {
dr->dt.dl.dr_data = arc_alloc_buf(spa, db, type, size);
}
memcpy(dr->dt.dl.dr_data->b_data, db->db.db_data, size);
} else {
db->db_buf = NULL;
dbuf_clear_data(db);
}
}
int
dbuf_read(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags)
{
int err = 0;
boolean_t prefetch;
dnode_t *dn;
/*
* We don't have to hold the mutex to check db_state because it
* can't be freed while we have a hold on the buffer.
*/
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
prefetch = db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID &&
(flags & DB_RF_NOPREFETCH) == 0 && dn != NULL;
mutex_enter(&db->db_mtx);
if (flags & DB_RF_PARTIAL_FIRST)
db->db_partial_read = B_TRUE;
else if (!(flags & DB_RF_PARTIAL_MORE))
db->db_partial_read = B_FALSE;
if (db->db_state == DB_CACHED) {
/*
* Ensure that this block's dnode has been decrypted if
* the caller has requested decrypted data.
*/
err = dbuf_read_verify_dnode_crypt(db, flags);
/*
* If the arc buf is compressed or encrypted and the caller
* requested uncompressed data, we need to untransform it
* before returning. We also call arc_untransform() on any
* unauthenticated blocks, which will verify their MAC if
* the key is now available.
*/
if (err == 0 && db->db_buf != NULL &&
(flags & DB_RF_NO_DECRYPT) == 0 &&
(arc_is_encrypted(db->db_buf) ||
arc_is_unauthenticated(db->db_buf) ||
arc_get_compression(db->db_buf) != ZIO_COMPRESS_OFF)) {
spa_t *spa = dn->dn_objset->os_spa;
zbookmark_phys_t zb;
SET_BOOKMARK(&zb, dmu_objset_id(db->db_objset),
db->db.db_object, db->db_level, db->db_blkid);
dbuf_fix_old_data(db, spa_syncing_txg(spa));
err = arc_untransform(db->db_buf, spa, &zb, B_FALSE);
dbuf_set_data(db, db->db_buf);
}
mutex_exit(&db->db_mtx);
if (err == 0 && prefetch) {
dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1, B_TRUE,
B_FALSE, flags & DB_RF_HAVESTRUCT);
}
DB_DNODE_EXIT(db);
DBUF_STAT_BUMP(hash_hits);
} else if (db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL) {
boolean_t need_wait = B_FALSE;
db_lock_type_t dblt = dmu_buf_lock_parent(db, RW_READER, FTAG);
if (zio == NULL && (db->db_state == DB_NOFILL ||
(db->db_blkptr != NULL && !BP_IS_HOLE(db->db_blkptr)))) {
spa_t *spa = dn->dn_objset->os_spa;
zio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL);
need_wait = B_TRUE;
}
err = dbuf_read_impl(db, zio, flags, dblt, FTAG);
/*
* dbuf_read_impl has dropped db_mtx and our parent's rwlock
* for us
*/
if (!err && prefetch) {
dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1, B_TRUE,
db->db_state != DB_CACHED,
flags & DB_RF_HAVESTRUCT);
}
DB_DNODE_EXIT(db);
DBUF_STAT_BUMP(hash_misses);
/*
* If we created a zio_root we must execute it to avoid
* leaking it, even if it isn't attached to any work due
* to an error in dbuf_read_impl().
*/
if (need_wait) {
if (err == 0)
err = zio_wait(zio);
else
VERIFY0(zio_wait(zio));
}
} else {
/*
* Another reader came in while the dbuf was in flight
* between UNCACHED and CACHED. Either a writer will finish
* writing the buffer (sending the dbuf to CACHED) or the
* first reader's request will reach the read_done callback
* and send the dbuf to CACHED. Otherwise, a failure
* occurred and the dbuf went to UNCACHED.
*/
mutex_exit(&db->db_mtx);
if (prefetch) {
dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1, B_TRUE,
B_TRUE, flags & DB_RF_HAVESTRUCT);
}
DB_DNODE_EXIT(db);
DBUF_STAT_BUMP(hash_misses);
/* Skip the wait per the caller's request. */
if ((flags & DB_RF_NEVERWAIT) == 0) {
mutex_enter(&db->db_mtx);
while (db->db_state == DB_READ ||
db->db_state == DB_FILL) {
ASSERT(db->db_state == DB_READ ||
(flags & DB_RF_HAVESTRUCT) == 0);
DTRACE_PROBE2(blocked__read, dmu_buf_impl_t *,
db, zio_t *, zio);
cv_wait(&db->db_changed, &db->db_mtx);
}
if (db->db_state == DB_UNCACHED)
err = SET_ERROR(EIO);
mutex_exit(&db->db_mtx);
}
}
return (err);
}
static void
dbuf_noread(dmu_buf_impl_t *db)
{
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
mutex_enter(&db->db_mtx);
while (db->db_state == DB_READ || db->db_state == DB_FILL)
cv_wait(&db->db_changed, &db->db_mtx);
if (db->db_state == DB_UNCACHED) {
ASSERT(db->db_buf == NULL);
ASSERT(db->db.db_data == NULL);
dbuf_set_data(db, dbuf_alloc_arcbuf(db));
db->db_state = DB_FILL;
DTRACE_SET_STATE(db, "assigning filled buffer");
} else if (db->db_state == DB_NOFILL) {
dbuf_clear_data(db);
} else {
ASSERT3U(db->db_state, ==, DB_CACHED);
}
mutex_exit(&db->db_mtx);
}
void
dbuf_unoverride(dbuf_dirty_record_t *dr)
{
dmu_buf_impl_t *db = dr->dr_dbuf;
blkptr_t *bp = &dr->dt.dl.dr_overridden_by;
uint64_t txg = dr->dr_txg;
ASSERT(MUTEX_HELD(&db->db_mtx));
/*
* This assert is valid because dmu_sync() expects to be called by
* a zilog's get_data while holding a range lock. This call only
* comes from dbuf_dirty() callers who must also hold a range lock.
*/
ASSERT(dr->dt.dl.dr_override_state != DR_IN_DMU_SYNC);
ASSERT(db->db_level == 0);
if (db->db_blkid == DMU_BONUS_BLKID ||
dr->dt.dl.dr_override_state == DR_NOT_OVERRIDDEN)
return;
ASSERT(db->db_data_pending != dr);
/* free this block */
if (!BP_IS_HOLE(bp) && !dr->dt.dl.dr_nopwrite)
zio_free(db->db_objset->os_spa, txg, bp);
dr->dt.dl.dr_override_state = DR_NOT_OVERRIDDEN;
dr->dt.dl.dr_nopwrite = B_FALSE;
dr->dt.dl.dr_has_raw_params = B_FALSE;
/*
* Release the already-written buffer, so we leave it in
* a consistent dirty state. Note that all callers are
* modifying the buffer, so they will immediately do
* another (redundant) arc_release(). Therefore, leave
* the buf thawed to save the effort of freezing &
* immediately re-thawing it.
*/
if (!dr->dt.dl.dr_brtwrite)
arc_release(dr->dt.dl.dr_data, db);
}
/*
* Evict (if its unreferenced) or clear (if its referenced) any level-0
* data blocks in the free range, so that any future readers will find
* empty blocks.
*/
void
dbuf_free_range(dnode_t *dn, uint64_t start_blkid, uint64_t end_blkid,
dmu_tx_t *tx)
{
dmu_buf_impl_t *db_search;
dmu_buf_impl_t *db, *db_next;
uint64_t txg = tx->tx_txg;
avl_index_t where;
dbuf_dirty_record_t *dr;
if (end_blkid > dn->dn_maxblkid &&
!(start_blkid == DMU_SPILL_BLKID || end_blkid == DMU_SPILL_BLKID))
end_blkid = dn->dn_maxblkid;
dprintf_dnode(dn, "start=%llu end=%llu\n", (u_longlong_t)start_blkid,
(u_longlong_t)end_blkid);
db_search = kmem_alloc(sizeof (dmu_buf_impl_t), KM_SLEEP);
db_search->db_level = 0;
db_search->db_blkid = start_blkid;
db_search->db_state = DB_SEARCH;
mutex_enter(&dn->dn_dbufs_mtx);
db = avl_find(&dn->dn_dbufs, db_search, &where);
ASSERT3P(db, ==, NULL);
db = avl_nearest(&dn->dn_dbufs, where, AVL_AFTER);
for (; db != NULL; db = db_next) {
db_next = AVL_NEXT(&dn->dn_dbufs, db);
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
if (db->db_level != 0 || db->db_blkid > end_blkid) {
break;
}
ASSERT3U(db->db_blkid, >=, start_blkid);
/* found a level 0 buffer in the range */
mutex_enter(&db->db_mtx);
if (dbuf_undirty(db, tx)) {
/* mutex has been dropped and dbuf destroyed */
continue;
}
if (db->db_state == DB_UNCACHED ||
db->db_state == DB_NOFILL ||
db->db_state == DB_EVICTING) {
ASSERT(db->db.db_data == NULL);
mutex_exit(&db->db_mtx);
continue;
}
if (db->db_state == DB_READ || db->db_state == DB_FILL) {
/* will be handled in dbuf_read_done or dbuf_rele */
db->db_freed_in_flight = TRUE;
mutex_exit(&db->db_mtx);
continue;
}
if (zfs_refcount_count(&db->db_holds) == 0) {
ASSERT(db->db_buf);
dbuf_destroy(db);
continue;
}
/* The dbuf is referenced */
dr = list_head(&db->db_dirty_records);
if (dr != NULL) {
if (dr->dr_txg == txg) {
/*
* This buffer is "in-use", re-adjust the file
* size to reflect that this buffer may
* contain new data when we sync.
*/
if (db->db_blkid != DMU_SPILL_BLKID &&
db->db_blkid > dn->dn_maxblkid)
dn->dn_maxblkid = db->db_blkid;
dbuf_unoverride(dr);
if (dr->dt.dl.dr_brtwrite) {
ASSERT(db->db.db_data == NULL);
mutex_exit(&db->db_mtx);
continue;
}
} else {
/*
* This dbuf is not dirty in the open context.
* Either uncache it (if its not referenced in
* the open context) or reset its contents to
* empty.
*/
dbuf_fix_old_data(db, txg);
}
}
/* clear the contents if its cached */
if (db->db_state == DB_CACHED) {
ASSERT(db->db.db_data != NULL);
arc_release(db->db_buf, db);
rw_enter(&db->db_rwlock, RW_WRITER);
memset(db->db.db_data, 0, db->db.db_size);
rw_exit(&db->db_rwlock);
arc_buf_freeze(db->db_buf);
}
mutex_exit(&db->db_mtx);
}
mutex_exit(&dn->dn_dbufs_mtx);
kmem_free(db_search, sizeof (dmu_buf_impl_t));
}
void
dbuf_new_size(dmu_buf_impl_t *db, int size, dmu_tx_t *tx)
{
arc_buf_t *buf, *old_buf;
dbuf_dirty_record_t *dr;
int osize = db->db.db_size;
arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db);
dnode_t *dn;
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
/*
* XXX we should be doing a dbuf_read, checking the return
* value and returning that up to our callers
*/
dmu_buf_will_dirty(&db->db, tx);
/* create the data buffer for the new block */
buf = arc_alloc_buf(dn->dn_objset->os_spa, db, type, size);
/* copy old block data to the new block */
old_buf = db->db_buf;
memcpy(buf->b_data, old_buf->b_data, MIN(osize, size));
/* zero the remainder */
if (size > osize)
memset((uint8_t *)buf->b_data + osize, 0, size - osize);
mutex_enter(&db->db_mtx);
dbuf_set_data(db, buf);
arc_buf_destroy(old_buf, db);
db->db.db_size = size;
dr = list_head(&db->db_dirty_records);
/* dirty record added by dmu_buf_will_dirty() */
VERIFY(dr != NULL);
if (db->db_level == 0)
dr->dt.dl.dr_data = buf;
ASSERT3U(dr->dr_txg, ==, tx->tx_txg);
ASSERT3U(dr->dr_accounted, ==, osize);
dr->dr_accounted = size;
mutex_exit(&db->db_mtx);
dmu_objset_willuse_space(dn->dn_objset, size - osize, tx);
DB_DNODE_EXIT(db);
}
void
dbuf_release_bp(dmu_buf_impl_t *db)
{
objset_t *os __maybe_unused = db->db_objset;
ASSERT(dsl_pool_sync_context(dmu_objset_pool(os)));
ASSERT(arc_released(os->os_phys_buf) ||
list_link_active(&os->os_dsl_dataset->ds_synced_link));
ASSERT(db->db_parent == NULL || arc_released(db->db_parent->db_buf));
(void) arc_release(db->db_buf, db);
}
/*
* We already have a dirty record for this TXG, and we are being
* dirtied again.
*/
static void
dbuf_redirty(dbuf_dirty_record_t *dr)
{
dmu_buf_impl_t *db = dr->dr_dbuf;
ASSERT(MUTEX_HELD(&db->db_mtx));
if (db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID) {
/*
* If this buffer has already been written out,
* we now need to reset its state.
*/
dbuf_unoverride(dr);
if (db->db.db_object != DMU_META_DNODE_OBJECT &&
db->db_state != DB_NOFILL) {
/* Already released on initial dirty, so just thaw. */
ASSERT(arc_released(db->db_buf));
arc_buf_thaw(db->db_buf);
}
}
}
dbuf_dirty_record_t *
dbuf_dirty_lightweight(dnode_t *dn, uint64_t blkid, dmu_tx_t *tx)
{
rw_enter(&dn->dn_struct_rwlock, RW_READER);
IMPLY(dn->dn_objset->os_raw_receive, dn->dn_maxblkid >= blkid);
dnode_new_blkid(dn, blkid, tx, B_TRUE, B_FALSE);
ASSERT(dn->dn_maxblkid >= blkid);
dbuf_dirty_record_t *dr = kmem_zalloc(sizeof (*dr), KM_SLEEP);
list_link_init(&dr->dr_dirty_node);
list_link_init(&dr->dr_dbuf_node);
dr->dr_dnode = dn;
dr->dr_txg = tx->tx_txg;
dr->dt.dll.dr_blkid = blkid;
dr->dr_accounted = dn->dn_datablksz;
/*
* There should not be any dbuf for the block that we're dirtying.
* Otherwise the buffer contents could be inconsistent between the
* dbuf and the lightweight dirty record.
*/
ASSERT3P(NULL, ==, dbuf_find(dn->dn_objset, dn->dn_object, 0, blkid,
NULL));
mutex_enter(&dn->dn_mtx);
int txgoff = tx->tx_txg & TXG_MASK;
if (dn->dn_free_ranges[txgoff] != NULL) {
range_tree_clear(dn->dn_free_ranges[txgoff], blkid, 1);
}
if (dn->dn_nlevels == 1) {
ASSERT3U(blkid, <, dn->dn_nblkptr);
list_insert_tail(&dn->dn_dirty_records[txgoff], dr);
mutex_exit(&dn->dn_mtx);
rw_exit(&dn->dn_struct_rwlock);
dnode_setdirty(dn, tx);
} else {
mutex_exit(&dn->dn_mtx);
int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT;
dmu_buf_impl_t *parent_db = dbuf_hold_level(dn,
1, blkid >> epbs, FTAG);
rw_exit(&dn->dn_struct_rwlock);
if (parent_db == NULL) {
kmem_free(dr, sizeof (*dr));
return (NULL);
}
int err = dbuf_read(parent_db, NULL,
(DB_RF_NOPREFETCH | DB_RF_CANFAIL));
if (err != 0) {
dbuf_rele(parent_db, FTAG);
kmem_free(dr, sizeof (*dr));
return (NULL);
}
dbuf_dirty_record_t *parent_dr = dbuf_dirty(parent_db, tx);
dbuf_rele(parent_db, FTAG);
mutex_enter(&parent_dr->dt.di.dr_mtx);
ASSERT3U(parent_dr->dr_txg, ==, tx->tx_txg);
list_insert_tail(&parent_dr->dt.di.dr_children, dr);
mutex_exit(&parent_dr->dt.di.dr_mtx);
dr->dr_parent = parent_dr;
}
dmu_objset_willuse_space(dn->dn_objset, dr->dr_accounted, tx);
return (dr);
}
dbuf_dirty_record_t *
dbuf_dirty(dmu_buf_impl_t *db, dmu_tx_t *tx)
{
dnode_t *dn;
objset_t *os;
dbuf_dirty_record_t *dr, *dr_next, *dr_head;
int txgoff = tx->tx_txg & TXG_MASK;
boolean_t drop_struct_rwlock = B_FALSE;
ASSERT(tx->tx_txg != 0);
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
DMU_TX_DIRTY_BUF(tx, db);
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
/*
* Shouldn't dirty a regular buffer in syncing context. Private
* objects may be dirtied in syncing context, but only if they
* were already pre-dirtied in open context.
*/
#ifdef ZFS_DEBUG
if (dn->dn_objset->os_dsl_dataset != NULL) {
rrw_enter(&dn->dn_objset->os_dsl_dataset->ds_bp_rwlock,
RW_READER, FTAG);
}
ASSERT(!dmu_tx_is_syncing(tx) ||
BP_IS_HOLE(dn->dn_objset->os_rootbp) ||
DMU_OBJECT_IS_SPECIAL(dn->dn_object) ||
dn->dn_objset->os_dsl_dataset == NULL);
if (dn->dn_objset->os_dsl_dataset != NULL)
rrw_exit(&dn->dn_objset->os_dsl_dataset->ds_bp_rwlock, FTAG);
#endif
/*
* We make this assert for private objects as well, but after we
* check if we're already dirty. They are allowed to re-dirty
* in syncing context.
*/
ASSERT(dn->dn_object == DMU_META_DNODE_OBJECT ||
dn->dn_dirtyctx == DN_UNDIRTIED || dn->dn_dirtyctx ==
(dmu_tx_is_syncing(tx) ? DN_DIRTY_SYNC : DN_DIRTY_OPEN));
mutex_enter(&db->db_mtx);
/*
* XXX make this true for indirects too? The problem is that
* transactions created with dmu_tx_create_assigned() from
* syncing context don't bother holding ahead.
*/
ASSERT(db->db_level != 0 ||
db->db_state == DB_CACHED || db->db_state == DB_FILL ||
db->db_state == DB_NOFILL);
mutex_enter(&dn->dn_mtx);
dnode_set_dirtyctx(dn, tx, db);
if (tx->tx_txg > dn->dn_dirty_txg)
dn->dn_dirty_txg = tx->tx_txg;
mutex_exit(&dn->dn_mtx);
if (db->db_blkid == DMU_SPILL_BLKID)
dn->dn_have_spill = B_TRUE;
/*
* If this buffer is already dirty, we're done.
*/
dr_head = list_head(&db->db_dirty_records);
ASSERT(dr_head == NULL || dr_head->dr_txg <= tx->tx_txg ||
db->db.db_object == DMU_META_DNODE_OBJECT);
dr_next = dbuf_find_dirty_lte(db, tx->tx_txg);
if (dr_next && dr_next->dr_txg == tx->tx_txg) {
DB_DNODE_EXIT(db);
dbuf_redirty(dr_next);
mutex_exit(&db->db_mtx);
return (dr_next);
}
/*
* Only valid if not already dirty.
*/
ASSERT(dn->dn_object == 0 ||
dn->dn_dirtyctx == DN_UNDIRTIED || dn->dn_dirtyctx ==
(dmu_tx_is_syncing(tx) ? DN_DIRTY_SYNC : DN_DIRTY_OPEN));
ASSERT3U(dn->dn_nlevels, >, db->db_level);
/*
* We should only be dirtying in syncing context if it's the
* mos or we're initializing the os or it's a special object.
* However, we are allowed to dirty in syncing context provided
* we already dirtied it in open context. Hence we must make
* this assertion only if we're not already dirty.
*/
os = dn->dn_objset;
VERIFY3U(tx->tx_txg, <=, spa_final_dirty_txg(os->os_spa));
#ifdef ZFS_DEBUG
if (dn->dn_objset->os_dsl_dataset != NULL)
rrw_enter(&os->os_dsl_dataset->ds_bp_rwlock, RW_READER, FTAG);
ASSERT(!dmu_tx_is_syncing(tx) || DMU_OBJECT_IS_SPECIAL(dn->dn_object) ||
os->os_dsl_dataset == NULL || BP_IS_HOLE(os->os_rootbp));
if (dn->dn_objset->os_dsl_dataset != NULL)
rrw_exit(&os->os_dsl_dataset->ds_bp_rwlock, FTAG);
#endif
ASSERT(db->db.db_size != 0);
dprintf_dbuf(db, "size=%llx\n", (u_longlong_t)db->db.db_size);
if (db->db_blkid != DMU_BONUS_BLKID && db->db_state != DB_NOFILL) {
dmu_objset_willuse_space(os, db->db.db_size, tx);
}
/*
* If this buffer is dirty in an old transaction group we need
* to make a copy of it so that the changes we make in this
* transaction group won't leak out when we sync the older txg.
*/
dr = kmem_zalloc(sizeof (dbuf_dirty_record_t), KM_SLEEP);
list_link_init(&dr->dr_dirty_node);
list_link_init(&dr->dr_dbuf_node);
dr->dr_dnode = dn;
if (db->db_level == 0) {
void *data_old = db->db_buf;
if (db->db_state != DB_NOFILL) {
if (db->db_blkid == DMU_BONUS_BLKID) {
dbuf_fix_old_data(db, tx->tx_txg);
data_old = db->db.db_data;
} else if (db->db.db_object != DMU_META_DNODE_OBJECT) {
/*
* Release the data buffer from the cache so
* that we can modify it without impacting
* possible other users of this cached data
* block. Note that indirect blocks and
* private objects are not released until the
* syncing state (since they are only modified
* then).
*/
arc_release(db->db_buf, db);
dbuf_fix_old_data(db, tx->tx_txg);
data_old = db->db_buf;
}
ASSERT(data_old != NULL);
}
dr->dt.dl.dr_data = data_old;
} else {
mutex_init(&dr->dt.di.dr_mtx, NULL, MUTEX_NOLOCKDEP, NULL);
list_create(&dr->dt.di.dr_children,
sizeof (dbuf_dirty_record_t),
offsetof(dbuf_dirty_record_t, dr_dirty_node));
}
if (db->db_blkid != DMU_BONUS_BLKID && db->db_state != DB_NOFILL) {
dr->dr_accounted = db->db.db_size;
}
dr->dr_dbuf = db;
dr->dr_txg = tx->tx_txg;
list_insert_before(&db->db_dirty_records, dr_next, dr);
/*
* We could have been freed_in_flight between the dbuf_noread
* and dbuf_dirty. We win, as though the dbuf_noread() had
* happened after the free.
*/
if (db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID &&
db->db_blkid != DMU_SPILL_BLKID) {
mutex_enter(&dn->dn_mtx);
if (dn->dn_free_ranges[txgoff] != NULL) {
range_tree_clear(dn->dn_free_ranges[txgoff],
db->db_blkid, 1);
}
mutex_exit(&dn->dn_mtx);
db->db_freed_in_flight = FALSE;
}
/*
* This buffer is now part of this txg
*/
dbuf_add_ref(db, (void *)(uintptr_t)tx->tx_txg);
db->db_dirtycnt += 1;
ASSERT3U(db->db_dirtycnt, <=, 3);
mutex_exit(&db->db_mtx);
if (db->db_blkid == DMU_BONUS_BLKID ||
db->db_blkid == DMU_SPILL_BLKID) {
mutex_enter(&dn->dn_mtx);
ASSERT(!list_link_active(&dr->dr_dirty_node));
list_insert_tail(&dn->dn_dirty_records[txgoff], dr);
mutex_exit(&dn->dn_mtx);
dnode_setdirty(dn, tx);
DB_DNODE_EXIT(db);
return (dr);
}
if (!RW_WRITE_HELD(&dn->dn_struct_rwlock)) {
rw_enter(&dn->dn_struct_rwlock, RW_READER);
drop_struct_rwlock = B_TRUE;
}
/*
* If we are overwriting a dedup BP, then unless it is snapshotted,
* when we get to syncing context we will need to decrement its
* refcount in the DDT. Prefetch the relevant DDT block so that
* syncing context won't have to wait for the i/o.
*/
if (db->db_blkptr != NULL) {
db_lock_type_t dblt = dmu_buf_lock_parent(db, RW_READER, FTAG);
ddt_prefetch(os->os_spa, db->db_blkptr);
dmu_buf_unlock_parent(db, dblt, FTAG);
}
/*
* We need to hold the dn_struct_rwlock to make this assertion,
* because it protects dn_phys / dn_next_nlevels from changing.
*/
ASSERT((dn->dn_phys->dn_nlevels == 0 && db->db_level == 0) ||
dn->dn_phys->dn_nlevels > db->db_level ||
dn->dn_next_nlevels[txgoff] > db->db_level ||
dn->dn_next_nlevels[(tx->tx_txg-1) & TXG_MASK] > db->db_level ||
dn->dn_next_nlevels[(tx->tx_txg-2) & TXG_MASK] > db->db_level);
if (db->db_level == 0) {
ASSERT(!db->db_objset->os_raw_receive ||
dn->dn_maxblkid >= db->db_blkid);
dnode_new_blkid(dn, db->db_blkid, tx,
drop_struct_rwlock, B_FALSE);
ASSERT(dn->dn_maxblkid >= db->db_blkid);
}
if (db->db_level+1 < dn->dn_nlevels) {
dmu_buf_impl_t *parent = db->db_parent;
dbuf_dirty_record_t *di;
int parent_held = FALSE;
if (db->db_parent == NULL || db->db_parent == dn->dn_dbuf) {
int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT;
parent = dbuf_hold_level(dn, db->db_level + 1,
db->db_blkid >> epbs, FTAG);
ASSERT(parent != NULL);
parent_held = TRUE;
}
if (drop_struct_rwlock)
rw_exit(&dn->dn_struct_rwlock);
ASSERT3U(db->db_level + 1, ==, parent->db_level);
di = dbuf_dirty(parent, tx);
if (parent_held)
dbuf_rele(parent, FTAG);
mutex_enter(&db->db_mtx);
/*
* Since we've dropped the mutex, it's possible that
* dbuf_undirty() might have changed this out from under us.
*/
if (list_head(&db->db_dirty_records) == dr ||
dn->dn_object == DMU_META_DNODE_OBJECT) {
mutex_enter(&di->dt.di.dr_mtx);
ASSERT3U(di->dr_txg, ==, tx->tx_txg);
ASSERT(!list_link_active(&dr->dr_dirty_node));
list_insert_tail(&di->dt.di.dr_children, dr);
mutex_exit(&di->dt.di.dr_mtx);
dr->dr_parent = di;
}
mutex_exit(&db->db_mtx);
} else {
ASSERT(db->db_level + 1 == dn->dn_nlevels);
ASSERT(db->db_blkid < dn->dn_nblkptr);
ASSERT(db->db_parent == NULL || db->db_parent == dn->dn_dbuf);
mutex_enter(&dn->dn_mtx);
ASSERT(!list_link_active(&dr->dr_dirty_node));
list_insert_tail(&dn->dn_dirty_records[txgoff], dr);
mutex_exit(&dn->dn_mtx);
if (drop_struct_rwlock)
rw_exit(&dn->dn_struct_rwlock);
}
dnode_setdirty(dn, tx);
DB_DNODE_EXIT(db);
return (dr);
}
static void
dbuf_undirty_bonus(dbuf_dirty_record_t *dr)
{
dmu_buf_impl_t *db = dr->dr_dbuf;
if (dr->dt.dl.dr_data != db->db.db_data) {
struct dnode *dn = dr->dr_dnode;
int max_bonuslen = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots);
kmem_free(dr->dt.dl.dr_data, max_bonuslen);
arc_space_return(max_bonuslen, ARC_SPACE_BONUS);
}
db->db_data_pending = NULL;
ASSERT(list_next(&db->db_dirty_records, dr) == NULL);
list_remove(&db->db_dirty_records, dr);
if (dr->dr_dbuf->db_level != 0) {
mutex_destroy(&dr->dt.di.dr_mtx);
list_destroy(&dr->dt.di.dr_children);
}
kmem_free(dr, sizeof (dbuf_dirty_record_t));
ASSERT3U(db->db_dirtycnt, >, 0);
db->db_dirtycnt -= 1;
}
/*
* Undirty a buffer in the transaction group referenced by the given
* transaction. Return whether this evicted the dbuf.
*/
boolean_t
dbuf_undirty(dmu_buf_impl_t *db, dmu_tx_t *tx)
{
uint64_t txg = tx->tx_txg;
boolean_t brtwrite;
ASSERT(txg != 0);
/*
* Due to our use of dn_nlevels below, this can only be called
* in open context, unless we are operating on the MOS.
* From syncing context, dn_nlevels may be different from the
* dn_nlevels used when dbuf was dirtied.
*/
ASSERT(db->db_objset ==
dmu_objset_pool(db->db_objset)->dp_meta_objset ||
txg != spa_syncing_txg(dmu_objset_spa(db->db_objset)));
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
ASSERT0(db->db_level);
ASSERT(MUTEX_HELD(&db->db_mtx));
/*
* If this buffer is not dirty, we're done.
*/
dbuf_dirty_record_t *dr = dbuf_find_dirty_eq(db, txg);
if (dr == NULL)
return (B_FALSE);
ASSERT(dr->dr_dbuf == db);
brtwrite = dr->dt.dl.dr_brtwrite;
if (brtwrite) {
/*
* We are freeing a block that we cloned in the same
* transaction group.
*/
brt_pending_remove(dmu_objset_spa(db->db_objset),
&dr->dt.dl.dr_overridden_by, tx);
}
dnode_t *dn = dr->dr_dnode;
dprintf_dbuf(db, "size=%llx\n", (u_longlong_t)db->db.db_size);
ASSERT(db->db.db_size != 0);
dsl_pool_undirty_space(dmu_objset_pool(dn->dn_objset),
dr->dr_accounted, txg);
list_remove(&db->db_dirty_records, dr);
/*
* Note that there are three places in dbuf_dirty()
* where this dirty record may be put on a list.
* Make sure to do a list_remove corresponding to
* every one of those list_insert calls.
*/
if (dr->dr_parent) {
mutex_enter(&dr->dr_parent->dt.di.dr_mtx);
list_remove(&dr->dr_parent->dt.di.dr_children, dr);
mutex_exit(&dr->dr_parent->dt.di.dr_mtx);
} else if (db->db_blkid == DMU_SPILL_BLKID ||
db->db_level + 1 == dn->dn_nlevels) {
ASSERT(db->db_blkptr == NULL || db->db_parent == dn->dn_dbuf);
mutex_enter(&dn->dn_mtx);
list_remove(&dn->dn_dirty_records[txg & TXG_MASK], dr);
mutex_exit(&dn->dn_mtx);
}
if (db->db_state != DB_NOFILL && !brtwrite) {
dbuf_unoverride(dr);
ASSERT(db->db_buf != NULL);
ASSERT(dr->dt.dl.dr_data != NULL);
if (dr->dt.dl.dr_data != db->db_buf)
arc_buf_destroy(dr->dt.dl.dr_data, db);
}
kmem_free(dr, sizeof (dbuf_dirty_record_t));
ASSERT(db->db_dirtycnt > 0);
db->db_dirtycnt -= 1;
if (zfs_refcount_remove(&db->db_holds, (void *)(uintptr_t)txg) == 0) {
ASSERT(db->db_state == DB_NOFILL || brtwrite ||
arc_released(db->db_buf));
dbuf_destroy(db);
return (B_TRUE);
}
return (B_FALSE);
}
static void
dmu_buf_will_dirty_impl(dmu_buf_t *db_fake, int flags, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
ASSERT(tx->tx_txg != 0);
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
/*
* Quick check for dirtiness. For already dirty blocks, this
* reduces runtime of this function by >90%, and overall performance
* by 50% for some workloads (e.g. file deletion with indirect blocks
* cached).
*/
mutex_enter(&db->db_mtx);
if (db->db_state == DB_CACHED) {
dbuf_dirty_record_t *dr = dbuf_find_dirty_eq(db, tx->tx_txg);
/*
* It's possible that it is already dirty but not cached,
* because there are some calls to dbuf_dirty() that don't
* go through dmu_buf_will_dirty().
*/
if (dr != NULL) {
/* This dbuf is already dirty and cached. */
dbuf_redirty(dr);
mutex_exit(&db->db_mtx);
return;
}
}
mutex_exit(&db->db_mtx);
DB_DNODE_ENTER(db);
if (RW_WRITE_HELD(&DB_DNODE(db)->dn_struct_rwlock))
flags |= DB_RF_HAVESTRUCT;
DB_DNODE_EXIT(db);
(void) dbuf_read(db, NULL, flags);
(void) dbuf_dirty(db, tx);
}
void
dmu_buf_will_dirty(dmu_buf_t *db_fake, dmu_tx_t *tx)
{
dmu_buf_will_dirty_impl(db_fake,
DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH, tx);
}
boolean_t
dmu_buf_is_dirty(dmu_buf_t *db_fake, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
dbuf_dirty_record_t *dr;
mutex_enter(&db->db_mtx);
dr = dbuf_find_dirty_eq(db, tx->tx_txg);
mutex_exit(&db->db_mtx);
return (dr != NULL);
}
void
dmu_buf_will_not_fill(dmu_buf_t *db_fake, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
db->db_state = DB_NOFILL;
DTRACE_SET_STATE(db, "allocating NOFILL buffer");
dmu_buf_will_fill(db_fake, tx);
}
void
dmu_buf_will_fill(dmu_buf_t *db_fake, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
ASSERT(tx->tx_txg != 0);
ASSERT(db->db_level == 0);
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
ASSERT(db->db.db_object != DMU_META_DNODE_OBJECT ||
dmu_tx_private_ok(tx));
dbuf_noread(db);
(void) dbuf_dirty(db, tx);
}
/*
* This function is effectively the same as dmu_buf_will_dirty(), but
* indicates the caller expects raw encrypted data in the db, and provides
* the crypt params (byteorder, salt, iv, mac) which should be stored in the
* blkptr_t when this dbuf is written. This is only used for blocks of
* dnodes, during raw receive.
*/
void
dmu_buf_set_crypt_params(dmu_buf_t *db_fake, boolean_t byteorder,
const uint8_t *salt, const uint8_t *iv, const uint8_t *mac, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
dbuf_dirty_record_t *dr;
/*
* dr_has_raw_params is only processed for blocks of dnodes
* (see dbuf_sync_dnode_leaf_crypt()).
*/
ASSERT3U(db->db.db_object, ==, DMU_META_DNODE_OBJECT);
ASSERT3U(db->db_level, ==, 0);
ASSERT(db->db_objset->os_raw_receive);
dmu_buf_will_dirty_impl(db_fake,
DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH | DB_RF_NO_DECRYPT, tx);
dr = dbuf_find_dirty_eq(db, tx->tx_txg);
ASSERT3P(dr, !=, NULL);
dr->dt.dl.dr_has_raw_params = B_TRUE;
dr->dt.dl.dr_byteorder = byteorder;
memcpy(dr->dt.dl.dr_salt, salt, ZIO_DATA_SALT_LEN);
memcpy(dr->dt.dl.dr_iv, iv, ZIO_DATA_IV_LEN);
memcpy(dr->dt.dl.dr_mac, mac, ZIO_DATA_MAC_LEN);
}
static void
dbuf_override_impl(dmu_buf_impl_t *db, const blkptr_t *bp, dmu_tx_t *tx)
{
struct dirty_leaf *dl;
dbuf_dirty_record_t *dr;
dr = list_head(&db->db_dirty_records);
ASSERT3P(dr, !=, NULL);
ASSERT3U(dr->dr_txg, ==, tx->tx_txg);
dl = &dr->dt.dl;
dl->dr_overridden_by = *bp;
dl->dr_override_state = DR_OVERRIDDEN;
dl->dr_overridden_by.blk_birth = dr->dr_txg;
}
void
dmu_buf_fill_done(dmu_buf_t *dbuf, dmu_tx_t *tx)
{
(void) tx;
dmu_buf_impl_t *db = (dmu_buf_impl_t *)dbuf;
dbuf_states_t old_state;
mutex_enter(&db->db_mtx);
DBUF_VERIFY(db);
old_state = db->db_state;
db->db_state = DB_CACHED;
if (old_state == DB_FILL) {
if (db->db_level == 0 && db->db_freed_in_flight) {
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
/* we were freed while filling */
/* XXX dbuf_undirty? */
memset(db->db.db_data, 0, db->db.db_size);
db->db_freed_in_flight = FALSE;
DTRACE_SET_STATE(db,
"fill done handling freed in flight");
} else {
DTRACE_SET_STATE(db, "fill done");
}
cv_broadcast(&db->db_changed);
}
mutex_exit(&db->db_mtx);
}
void
dmu_buf_write_embedded(dmu_buf_t *dbuf, void *data,
bp_embedded_type_t etype, enum zio_compress comp,
int uncompressed_size, int compressed_size, int byteorder,
dmu_tx_t *tx)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)dbuf;
struct dirty_leaf *dl;
dmu_object_type_t type;
dbuf_dirty_record_t *dr;
if (etype == BP_EMBEDDED_TYPE_DATA) {
ASSERT(spa_feature_is_active(dmu_objset_spa(db->db_objset),
SPA_FEATURE_EMBEDDED_DATA));
}
DB_DNODE_ENTER(db);
type = DB_DNODE(db)->dn_type;
DB_DNODE_EXIT(db);
ASSERT0(db->db_level);
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
dmu_buf_will_not_fill(dbuf, tx);
dr = list_head(&db->db_dirty_records);
ASSERT3P(dr, !=, NULL);
ASSERT3U(dr->dr_txg, ==, tx->tx_txg);
dl = &dr->dt.dl;
encode_embedded_bp_compressed(&dl->dr_overridden_by,
data, comp, uncompressed_size, compressed_size);
BPE_SET_ETYPE(&dl->dr_overridden_by, etype);
BP_SET_TYPE(&dl->dr_overridden_by, type);
BP_SET_LEVEL(&dl->dr_overridden_by, 0);
BP_SET_BYTEORDER(&dl->dr_overridden_by, byteorder);
dl->dr_override_state = DR_OVERRIDDEN;
dl->dr_overridden_by.blk_birth = dr->dr_txg;
}
void
dmu_buf_redact(dmu_buf_t *dbuf, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)dbuf;
dmu_object_type_t type;
ASSERT(dsl_dataset_feature_is_active(db->db_objset->os_dsl_dataset,
SPA_FEATURE_REDACTED_DATASETS));
DB_DNODE_ENTER(db);
type = DB_DNODE(db)->dn_type;
DB_DNODE_EXIT(db);
ASSERT0(db->db_level);
dmu_buf_will_not_fill(dbuf, tx);
blkptr_t bp = { { { {0} } } };
BP_SET_TYPE(&bp, type);
BP_SET_LEVEL(&bp, 0);
BP_SET_BIRTH(&bp, tx->tx_txg, 0);
BP_SET_REDACTED(&bp);
BPE_SET_LSIZE(&bp, dbuf->db_size);
dbuf_override_impl(db, &bp, tx);
}
/*
* Directly assign a provided arc buf to a given dbuf if it's not referenced
* by anybody except our caller. Otherwise copy arcbuf's contents to dbuf.
*/
void
dbuf_assign_arcbuf(dmu_buf_impl_t *db, arc_buf_t *buf, dmu_tx_t *tx)
{
ASSERT(!zfs_refcount_is_zero(&db->db_holds));
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
ASSERT(db->db_level == 0);
ASSERT3U(dbuf_is_metadata(db), ==, arc_is_metadata(buf));
ASSERT(buf != NULL);
ASSERT3U(arc_buf_lsize(buf), ==, db->db.db_size);
ASSERT(tx->tx_txg != 0);
arc_return_buf(buf, db);
ASSERT(arc_released(buf));
mutex_enter(&db->db_mtx);
while (db->db_state == DB_READ || db->db_state == DB_FILL)
cv_wait(&db->db_changed, &db->db_mtx);
ASSERT(db->db_state == DB_CACHED || db->db_state == DB_UNCACHED);
if (db->db_state == DB_CACHED &&
zfs_refcount_count(&db->db_holds) - 1 > db->db_dirtycnt) {
/*
* In practice, we will never have a case where we have an
* encrypted arc buffer while additional holds exist on the
* dbuf. We don't handle this here so we simply assert that
* fact instead.
*/
ASSERT(!arc_is_encrypted(buf));
mutex_exit(&db->db_mtx);
(void) dbuf_dirty(db, tx);
memcpy(db->db.db_data, buf->b_data, db->db.db_size);
arc_buf_destroy(buf, db);
return;
}
if (db->db_state == DB_CACHED) {
dbuf_dirty_record_t *dr = list_head(&db->db_dirty_records);
ASSERT(db->db_buf != NULL);
if (dr != NULL && dr->dr_txg == tx->tx_txg) {
ASSERT(dr->dt.dl.dr_data == db->db_buf);
if (!arc_released(db->db_buf)) {
ASSERT(dr->dt.dl.dr_override_state ==
DR_OVERRIDDEN);
arc_release(db->db_buf, db);
}
dr->dt.dl.dr_data = buf;
arc_buf_destroy(db->db_buf, db);
} else if (dr == NULL || dr->dt.dl.dr_data != db->db_buf) {
arc_release(db->db_buf, db);
arc_buf_destroy(db->db_buf, db);
}
db->db_buf = NULL;
}
ASSERT(db->db_buf == NULL);
dbuf_set_data(db, buf);
db->db_state = DB_FILL;
DTRACE_SET_STATE(db, "filling assigned arcbuf");
mutex_exit(&db->db_mtx);
(void) dbuf_dirty(db, tx);
dmu_buf_fill_done(&db->db, tx);
}
void
dbuf_destroy(dmu_buf_impl_t *db)
{
dnode_t *dn;
dmu_buf_impl_t *parent = db->db_parent;
dmu_buf_impl_t *dndb;
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT(zfs_refcount_is_zero(&db->db_holds));
if (db->db_buf != NULL) {
arc_buf_destroy(db->db_buf, db);
db->db_buf = NULL;
}
if (db->db_blkid == DMU_BONUS_BLKID) {
int slots = DB_DNODE(db)->dn_num_slots;
int bonuslen = DN_SLOTS_TO_BONUSLEN(slots);
if (db->db.db_data != NULL) {
kmem_free(db->db.db_data, bonuslen);
arc_space_return(bonuslen, ARC_SPACE_BONUS);
db->db_state = DB_UNCACHED;
DTRACE_SET_STATE(db, "buffer cleared");
}
}
dbuf_clear_data(db);
if (multilist_link_active(&db->db_cache_link)) {
ASSERT(db->db_caching_status == DB_DBUF_CACHE ||
db->db_caching_status == DB_DBUF_METADATA_CACHE);
multilist_remove(&dbuf_caches[db->db_caching_status].cache, db);
(void) zfs_refcount_remove_many(
&dbuf_caches[db->db_caching_status].size,
db->db.db_size, db);
if (db->db_caching_status == DB_DBUF_METADATA_CACHE) {
DBUF_STAT_BUMPDOWN(metadata_cache_count);
} else {
DBUF_STAT_BUMPDOWN(cache_levels[db->db_level]);
DBUF_STAT_BUMPDOWN(cache_count);
DBUF_STAT_DECR(cache_levels_bytes[db->db_level],
db->db.db_size);
}
db->db_caching_status = DB_NO_CACHE;
}
ASSERT(db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL);
ASSERT(db->db_data_pending == NULL);
ASSERT(list_is_empty(&db->db_dirty_records));
db->db_state = DB_EVICTING;
DTRACE_SET_STATE(db, "buffer eviction started");
db->db_blkptr = NULL;
/*
* Now that db_state is DB_EVICTING, nobody else can find this via
* the hash table. We can now drop db_mtx, which allows us to
* acquire the dn_dbufs_mtx.
*/
mutex_exit(&db->db_mtx);
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
dndb = dn->dn_dbuf;
if (db->db_blkid != DMU_BONUS_BLKID) {
boolean_t needlock = !MUTEX_HELD(&dn->dn_dbufs_mtx);
if (needlock)
mutex_enter_nested(&dn->dn_dbufs_mtx,
NESTED_SINGLE);
avl_remove(&dn->dn_dbufs, db);
membar_producer();
DB_DNODE_EXIT(db);
if (needlock)
mutex_exit(&dn->dn_dbufs_mtx);
/*
* Decrementing the dbuf count means that the hold corresponding
* to the removed dbuf is no longer discounted in dnode_move(),
* so the dnode cannot be moved until after we release the hold.
* The membar_producer() ensures visibility of the decremented
* value in dnode_move(), since DB_DNODE_EXIT doesn't actually
* release any lock.
*/
mutex_enter(&dn->dn_mtx);
dnode_rele_and_unlock(dn, db, B_TRUE);
db->db_dnode_handle = NULL;
dbuf_hash_remove(db);
} else {
DB_DNODE_EXIT(db);
}
ASSERT(zfs_refcount_is_zero(&db->db_holds));
db->db_parent = NULL;
ASSERT(db->db_buf == NULL);
ASSERT(db->db.db_data == NULL);
ASSERT(db->db_hash_next == NULL);
ASSERT(db->db_blkptr == NULL);
ASSERT(db->db_data_pending == NULL);
ASSERT3U(db->db_caching_status, ==, DB_NO_CACHE);
ASSERT(!multilist_link_active(&db->db_cache_link));
/*
* If this dbuf is referenced from an indirect dbuf,
* decrement the ref count on the indirect dbuf.
*/
if (parent && parent != dndb) {
mutex_enter(&parent->db_mtx);
dbuf_rele_and_unlock(parent, db, B_TRUE);
}
kmem_cache_free(dbuf_kmem_cache, db);
arc_space_return(sizeof (dmu_buf_impl_t), ARC_SPACE_DBUF);
}
/*
* Note: While bpp will always be updated if the function returns success,
* parentp will not be updated if the dnode does not have dn_dbuf filled in;
* this happens when the dnode is the meta-dnode, or {user|group|project}used
* object.
*/
__attribute__((always_inline))
static inline int
dbuf_findbp(dnode_t *dn, int level, uint64_t blkid, int fail_sparse,
dmu_buf_impl_t **parentp, blkptr_t **bpp)
{
*parentp = NULL;
*bpp = NULL;
ASSERT(blkid != DMU_BONUS_BLKID);
if (blkid == DMU_SPILL_BLKID) {
mutex_enter(&dn->dn_mtx);
if (dn->dn_have_spill &&
(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR))
*bpp = DN_SPILL_BLKPTR(dn->dn_phys);
else
*bpp = NULL;
dbuf_add_ref(dn->dn_dbuf, NULL);
*parentp = dn->dn_dbuf;
mutex_exit(&dn->dn_mtx);
return (0);
}
int nlevels =
(dn->dn_phys->dn_nlevels == 0) ? 1 : dn->dn_phys->dn_nlevels;
int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT;
ASSERT3U(level * epbs, <, 64);
ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock));
/*
* This assertion shouldn't trip as long as the max indirect block size
* is less than 1M. The reason for this is that up to that point,
* the number of levels required to address an entire object with blocks
* of size SPA_MINBLOCKSIZE satisfies nlevels * epbs + 1 <= 64. In
* other words, if N * epbs + 1 > 64, then if (N-1) * epbs + 1 > 55
* (i.e. we can address the entire object), objects will all use at most
* N-1 levels and the assertion won't overflow. However, once epbs is
* 13, 4 * 13 + 1 = 53, but 5 * 13 + 1 = 66. Then, 4 levels will not be
* enough to address an entire object, so objects will have 5 levels,
* but then this assertion will overflow.
*
* All this is to say that if we ever increase DN_MAX_INDBLKSHIFT, we
* need to redo this logic to handle overflows.
*/
ASSERT(level >= nlevels ||
((nlevels - level - 1) * epbs) +
highbit64(dn->dn_phys->dn_nblkptr) <= 64);
if (level >= nlevels ||
blkid >= ((uint64_t)dn->dn_phys->dn_nblkptr <<
((nlevels - level - 1) * epbs)) ||
(fail_sparse &&
blkid > (dn->dn_phys->dn_maxblkid >> (level * epbs)))) {
/* the buffer has no parent yet */
return (SET_ERROR(ENOENT));
} else if (level < nlevels-1) {
/* this block is referenced from an indirect block */
int err;
err = dbuf_hold_impl(dn, level + 1,
blkid >> epbs, fail_sparse, FALSE, NULL, parentp);
if (err)
return (err);
err = dbuf_read(*parentp, NULL,
(DB_RF_HAVESTRUCT | DB_RF_NOPREFETCH | DB_RF_CANFAIL));
if (err) {
dbuf_rele(*parentp, NULL);
*parentp = NULL;
return (err);
}
rw_enter(&(*parentp)->db_rwlock, RW_READER);
*bpp = ((blkptr_t *)(*parentp)->db.db_data) +
(blkid & ((1ULL << epbs) - 1));
if (blkid > (dn->dn_phys->dn_maxblkid >> (level * epbs)))
ASSERT(BP_IS_HOLE(*bpp));
rw_exit(&(*parentp)->db_rwlock);
return (0);
} else {
/* the block is referenced from the dnode */
ASSERT3U(level, ==, nlevels-1);
ASSERT(dn->dn_phys->dn_nblkptr == 0 ||
blkid < dn->dn_phys->dn_nblkptr);
if (dn->dn_dbuf) {
dbuf_add_ref(dn->dn_dbuf, NULL);
*parentp = dn->dn_dbuf;
}
*bpp = &dn->dn_phys->dn_blkptr[blkid];
return (0);
}
}
static dmu_buf_impl_t *
dbuf_create(dnode_t *dn, uint8_t level, uint64_t blkid,
dmu_buf_impl_t *parent, blkptr_t *blkptr, uint64_t hash)
{
objset_t *os = dn->dn_objset;
dmu_buf_impl_t *db, *odb;
ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock));
ASSERT(dn->dn_type != DMU_OT_NONE);
db = kmem_cache_alloc(dbuf_kmem_cache, KM_SLEEP);
list_create(&db->db_dirty_records, sizeof (dbuf_dirty_record_t),
offsetof(dbuf_dirty_record_t, dr_dbuf_node));
db->db_objset = os;
db->db.db_object = dn->dn_object;
db->db_level = level;
db->db_blkid = blkid;
db->db_dirtycnt = 0;
db->db_dnode_handle = dn->dn_handle;
db->db_parent = parent;
db->db_blkptr = blkptr;
db->db_hash = hash;
db->db_user = NULL;
db->db_user_immediate_evict = FALSE;
db->db_freed_in_flight = FALSE;
db->db_pending_evict = FALSE;
if (blkid == DMU_BONUS_BLKID) {
ASSERT3P(parent, ==, dn->dn_dbuf);
db->db.db_size = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots) -
(dn->dn_nblkptr-1) * sizeof (blkptr_t);
ASSERT3U(db->db.db_size, >=, dn->dn_bonuslen);
db->db.db_offset = DMU_BONUS_BLKID;
db->db_state = DB_UNCACHED;
DTRACE_SET_STATE(db, "bonus buffer created");
db->db_caching_status = DB_NO_CACHE;
/* the bonus dbuf is not placed in the hash table */
arc_space_consume(sizeof (dmu_buf_impl_t), ARC_SPACE_DBUF);
return (db);
} else if (blkid == DMU_SPILL_BLKID) {
db->db.db_size = (blkptr != NULL) ?
BP_GET_LSIZE(blkptr) : SPA_MINBLOCKSIZE;
db->db.db_offset = 0;
} else {
int blocksize =
db->db_level ? 1 << dn->dn_indblkshift : dn->dn_datablksz;
db->db.db_size = blocksize;
db->db.db_offset = db->db_blkid * blocksize;
}
/*
* Hold the dn_dbufs_mtx while we get the new dbuf
* in the hash table *and* added to the dbufs list.
* This prevents a possible deadlock with someone
* trying to look up this dbuf before it's added to the
* dn_dbufs list.
*/
mutex_enter(&dn->dn_dbufs_mtx);
db->db_state = DB_EVICTING; /* not worth logging this state change */
if ((odb = dbuf_hash_insert(db)) != NULL) {
/* someone else inserted it first */
mutex_exit(&dn->dn_dbufs_mtx);
kmem_cache_free(dbuf_kmem_cache, db);
DBUF_STAT_BUMP(hash_insert_race);
return (odb);
}
avl_add(&dn->dn_dbufs, db);
db->db_state = DB_UNCACHED;
DTRACE_SET_STATE(db, "regular buffer created");
db->db_caching_status = DB_NO_CACHE;
mutex_exit(&dn->dn_dbufs_mtx);
arc_space_consume(sizeof (dmu_buf_impl_t), ARC_SPACE_DBUF);
if (parent && parent != dn->dn_dbuf)
dbuf_add_ref(parent, db);
ASSERT(dn->dn_object == DMU_META_DNODE_OBJECT ||
zfs_refcount_count(&dn->dn_holds) > 0);
(void) zfs_refcount_add(&dn->dn_holds, db);
dprintf_dbuf(db, "db=%p\n", db);
return (db);
}
/*
* This function returns a block pointer and information about the object,
* given a dnode and a block. This is a publicly accessible version of
* dbuf_findbp that only returns some information, rather than the
* dbuf. Note that the dnode passed in must be held, and the dn_struct_rwlock
* should be locked as (at least) a reader.
*/
int
dbuf_dnode_findbp(dnode_t *dn, uint64_t level, uint64_t blkid,
blkptr_t *bp, uint16_t *datablkszsec, uint8_t *indblkshift)
{
dmu_buf_impl_t *dbp = NULL;
blkptr_t *bp2;
int err = 0;
ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock));
err = dbuf_findbp(dn, level, blkid, B_FALSE, &dbp, &bp2);
if (err == 0) {
ASSERT3P(bp2, !=, NULL);
*bp = *bp2;
if (dbp != NULL)
dbuf_rele(dbp, NULL);
if (datablkszsec != NULL)
*datablkszsec = dn->dn_phys->dn_datablkszsec;
if (indblkshift != NULL)
*indblkshift = dn->dn_phys->dn_indblkshift;
}
return (err);
}
typedef struct dbuf_prefetch_arg {
spa_t *dpa_spa; /* The spa to issue the prefetch in. */
zbookmark_phys_t dpa_zb; /* The target block to prefetch. */
int dpa_epbs; /* Entries (blkptr_t's) Per Block Shift. */
int dpa_curlevel; /* The current level that we're reading */
dnode_t *dpa_dnode; /* The dnode associated with the prefetch */
zio_priority_t dpa_prio; /* The priority I/Os should be issued at. */
zio_t *dpa_zio; /* The parent zio_t for all prefetches. */
arc_flags_t dpa_aflags; /* Flags to pass to the final prefetch. */
dbuf_prefetch_fn dpa_cb; /* prefetch completion callback */
void *dpa_arg; /* prefetch completion arg */
} dbuf_prefetch_arg_t;
static void
dbuf_prefetch_fini(dbuf_prefetch_arg_t *dpa, boolean_t io_done)
{
if (dpa->dpa_cb != NULL) {
dpa->dpa_cb(dpa->dpa_arg, dpa->dpa_zb.zb_level,
dpa->dpa_zb.zb_blkid, io_done);
}
kmem_free(dpa, sizeof (*dpa));
}
static void
dbuf_issue_final_prefetch_done(zio_t *zio, const zbookmark_phys_t *zb,
const blkptr_t *iobp, arc_buf_t *abuf, void *private)
{
(void) zio, (void) zb, (void) iobp;
dbuf_prefetch_arg_t *dpa = private;
if (abuf != NULL)
arc_buf_destroy(abuf, private);
dbuf_prefetch_fini(dpa, B_TRUE);
}
/*
* Actually issue the prefetch read for the block given.
*/
static void
dbuf_issue_final_prefetch(dbuf_prefetch_arg_t *dpa, blkptr_t *bp)
{
ASSERT(!BP_IS_REDACTED(bp) ||
dsl_dataset_feature_is_active(
dpa->dpa_dnode->dn_objset->os_dsl_dataset,
SPA_FEATURE_REDACTED_DATASETS));
if (BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp) || BP_IS_REDACTED(bp))
return (dbuf_prefetch_fini(dpa, B_FALSE));
int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE;
arc_flags_t aflags =
dpa->dpa_aflags | ARC_FLAG_NOWAIT | ARC_FLAG_PREFETCH |
ARC_FLAG_NO_BUF;
/* dnodes are always read as raw and then converted later */
if (BP_GET_TYPE(bp) == DMU_OT_DNODE && BP_IS_PROTECTED(bp) &&
dpa->dpa_curlevel == 0)
zio_flags |= ZIO_FLAG_RAW;
ASSERT3U(dpa->dpa_curlevel, ==, BP_GET_LEVEL(bp));
ASSERT3U(dpa->dpa_curlevel, ==, dpa->dpa_zb.zb_level);
ASSERT(dpa->dpa_zio != NULL);
(void) arc_read(dpa->dpa_zio, dpa->dpa_spa, bp,
dbuf_issue_final_prefetch_done, dpa,
dpa->dpa_prio, zio_flags, &aflags, &dpa->dpa_zb);
}
/*
* Called when an indirect block above our prefetch target is read in. This
* will either read in the next indirect block down the tree or issue the actual
* prefetch if the next block down is our target.
*/
static void
dbuf_prefetch_indirect_done(zio_t *zio, const zbookmark_phys_t *zb,
const blkptr_t *iobp, arc_buf_t *abuf, void *private)
{
(void) zb, (void) iobp;
dbuf_prefetch_arg_t *dpa = private;
ASSERT3S(dpa->dpa_zb.zb_level, <, dpa->dpa_curlevel);
ASSERT3S(dpa->dpa_curlevel, >, 0);
if (abuf == NULL) {
ASSERT(zio == NULL || zio->io_error != 0);
dbuf_prefetch_fini(dpa, B_TRUE);
return;
}
ASSERT(zio == NULL || zio->io_error == 0);
/*
* The dpa_dnode is only valid if we are called with a NULL
* zio. This indicates that the arc_read() returned without
* first calling zio_read() to issue a physical read. Once
* a physical read is made the dpa_dnode must be invalidated
* as the locks guarding it may have been dropped. If the
* dpa_dnode is still valid, then we want to add it to the dbuf
* cache. To do so, we must hold the dbuf associated with the block
* we just prefetched, read its contents so that we associate it
* with an arc_buf_t, and then release it.
*/
if (zio != NULL) {
ASSERT3S(BP_GET_LEVEL(zio->io_bp), ==, dpa->dpa_curlevel);
if (zio->io_flags & ZIO_FLAG_RAW_COMPRESS) {
ASSERT3U(BP_GET_PSIZE(zio->io_bp), ==, zio->io_size);
} else {
ASSERT3U(BP_GET_LSIZE(zio->io_bp), ==, zio->io_size);
}
ASSERT3P(zio->io_spa, ==, dpa->dpa_spa);
dpa->dpa_dnode = NULL;
} else if (dpa->dpa_dnode != NULL) {
uint64_t curblkid = dpa->dpa_zb.zb_blkid >>
(dpa->dpa_epbs * (dpa->dpa_curlevel -
dpa->dpa_zb.zb_level));
dmu_buf_impl_t *db = dbuf_hold_level(dpa->dpa_dnode,
dpa->dpa_curlevel, curblkid, FTAG);
if (db == NULL) {
arc_buf_destroy(abuf, private);
dbuf_prefetch_fini(dpa, B_TRUE);
return;
}
(void) dbuf_read(db, NULL,
DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH | DB_RF_HAVESTRUCT);
dbuf_rele(db, FTAG);
}
dpa->dpa_curlevel--;
uint64_t nextblkid = dpa->dpa_zb.zb_blkid >>
(dpa->dpa_epbs * (dpa->dpa_curlevel - dpa->dpa_zb.zb_level));
blkptr_t *bp = ((blkptr_t *)abuf->b_data) +
P2PHASE(nextblkid, 1ULL << dpa->dpa_epbs);
ASSERT(!BP_IS_REDACTED(bp) || (dpa->dpa_dnode &&
dsl_dataset_feature_is_active(
dpa->dpa_dnode->dn_objset->os_dsl_dataset,
SPA_FEATURE_REDACTED_DATASETS)));
if (BP_IS_HOLE(bp) || BP_IS_REDACTED(bp)) {
arc_buf_destroy(abuf, private);
dbuf_prefetch_fini(dpa, B_TRUE);
return;
} else if (dpa->dpa_curlevel == dpa->dpa_zb.zb_level) {
ASSERT3U(nextblkid, ==, dpa->dpa_zb.zb_blkid);
dbuf_issue_final_prefetch(dpa, bp);
} else {
arc_flags_t iter_aflags = ARC_FLAG_NOWAIT;
zbookmark_phys_t zb;
/* flag if L2ARC eligible, l2arc_noprefetch then decides */
if (dpa->dpa_aflags & ARC_FLAG_L2CACHE)
iter_aflags |= ARC_FLAG_L2CACHE;
ASSERT3U(dpa->dpa_curlevel, ==, BP_GET_LEVEL(bp));
SET_BOOKMARK(&zb, dpa->dpa_zb.zb_objset,
dpa->dpa_zb.zb_object, dpa->dpa_curlevel, nextblkid);
(void) arc_read(dpa->dpa_zio, dpa->dpa_spa,
bp, dbuf_prefetch_indirect_done, dpa,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE,
&iter_aflags, &zb);
}
arc_buf_destroy(abuf, private);
}
/*
* Issue prefetch reads for the given block on the given level. If the indirect
* blocks above that block are not in memory, we will read them in
* asynchronously. As a result, this call never blocks waiting for a read to
* complete. Note that the prefetch might fail if the dataset is encrypted and
* the encryption key is unmapped before the IO completes.
*/
int
dbuf_prefetch_impl(dnode_t *dn, int64_t level, uint64_t blkid,
zio_priority_t prio, arc_flags_t aflags, dbuf_prefetch_fn cb,
void *arg)
{
blkptr_t bp;
int epbs, nlevels, curlevel;
uint64_t curblkid;
ASSERT(blkid != DMU_BONUS_BLKID);
ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock));
if (blkid > dn->dn_maxblkid)
goto no_issue;
if (level == 0 && dnode_block_freed(dn, blkid))
goto no_issue;
/*
* This dnode hasn't been written to disk yet, so there's nothing to
* prefetch.
*/
nlevels = dn->dn_phys->dn_nlevels;
if (level >= nlevels || dn->dn_phys->dn_nblkptr == 0)
goto no_issue;
epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT;
if (dn->dn_phys->dn_maxblkid < blkid << (epbs * level))
goto no_issue;
dmu_buf_impl_t *db = dbuf_find(dn->dn_objset, dn->dn_object,
level, blkid, NULL);
if (db != NULL) {
mutex_exit(&db->db_mtx);
/*
* This dbuf already exists. It is either CACHED, or
* (we assume) about to be read or filled.
*/
goto no_issue;
}
/*
* Find the closest ancestor (indirect block) of the target block
* that is present in the cache. In this indirect block, we will
* find the bp that is at curlevel, curblkid.
*/
curlevel = level;
curblkid = blkid;
while (curlevel < nlevels - 1) {
int parent_level = curlevel + 1;
uint64_t parent_blkid = curblkid >> epbs;
dmu_buf_impl_t *db;
if (dbuf_hold_impl(dn, parent_level, parent_blkid,
FALSE, TRUE, FTAG, &db) == 0) {
blkptr_t *bpp = db->db_buf->b_data;
bp = bpp[P2PHASE(curblkid, 1 << epbs)];
dbuf_rele(db, FTAG);
break;
}
curlevel = parent_level;
curblkid = parent_blkid;
}
if (curlevel == nlevels - 1) {
/* No cached indirect blocks found. */
ASSERT3U(curblkid, <, dn->dn_phys->dn_nblkptr);
bp = dn->dn_phys->dn_blkptr[curblkid];
}
ASSERT(!BP_IS_REDACTED(&bp) ||
dsl_dataset_feature_is_active(dn->dn_objset->os_dsl_dataset,
SPA_FEATURE_REDACTED_DATASETS));
if (BP_IS_HOLE(&bp) || BP_IS_REDACTED(&bp))
goto no_issue;
ASSERT3U(curlevel, ==, BP_GET_LEVEL(&bp));
zio_t *pio = zio_root(dmu_objset_spa(dn->dn_objset), NULL, NULL,
ZIO_FLAG_CANFAIL);
dbuf_prefetch_arg_t *dpa = kmem_zalloc(sizeof (*dpa), KM_SLEEP);
dsl_dataset_t *ds = dn->dn_objset->os_dsl_dataset;
SET_BOOKMARK(&dpa->dpa_zb, ds != NULL ? ds->ds_object : DMU_META_OBJSET,
dn->dn_object, level, blkid);
dpa->dpa_curlevel = curlevel;
dpa->dpa_prio = prio;
dpa->dpa_aflags = aflags;
dpa->dpa_spa = dn->dn_objset->os_spa;
dpa->dpa_dnode = dn;
dpa->dpa_epbs = epbs;
dpa->dpa_zio = pio;
dpa->dpa_cb = cb;
dpa->dpa_arg = arg;
if (!DNODE_LEVEL_IS_CACHEABLE(dn, level))
dpa->dpa_aflags |= ARC_FLAG_UNCACHED;
else if (dnode_level_is_l2cacheable(&bp, dn, level))
dpa->dpa_aflags |= ARC_FLAG_L2CACHE;
/*
* If we have the indirect just above us, no need to do the asynchronous
* prefetch chain; we'll just run the last step ourselves. If we're at
* a higher level, though, we want to issue the prefetches for all the
* indirect blocks asynchronously, so we can go on with whatever we were
* doing.
*/
if (curlevel == level) {
ASSERT3U(curblkid, ==, blkid);
dbuf_issue_final_prefetch(dpa, &bp);
} else {
arc_flags_t iter_aflags = ARC_FLAG_NOWAIT;
zbookmark_phys_t zb;
/* flag if L2ARC eligible, l2arc_noprefetch then decides */
if (dnode_level_is_l2cacheable(&bp, dn, level))
iter_aflags |= ARC_FLAG_L2CACHE;
SET_BOOKMARK(&zb, ds != NULL ? ds->ds_object : DMU_META_OBJSET,
dn->dn_object, curlevel, curblkid);
(void) arc_read(dpa->dpa_zio, dpa->dpa_spa,
&bp, dbuf_prefetch_indirect_done, dpa,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE,
&iter_aflags, &zb);
}
/*
* We use pio here instead of dpa_zio since it's possible that
* dpa may have already been freed.
*/
zio_nowait(pio);
return (1);
no_issue:
if (cb != NULL)
cb(arg, level, blkid, B_FALSE);
return (0);
}
int
dbuf_prefetch(dnode_t *dn, int64_t level, uint64_t blkid, zio_priority_t prio,
arc_flags_t aflags)
{
return (dbuf_prefetch_impl(dn, level, blkid, prio, aflags, NULL, NULL));
}
/*
* Helper function for dbuf_hold_impl() to copy a buffer. Handles
* the case of encrypted, compressed and uncompressed buffers by
* allocating the new buffer, respectively, with arc_alloc_raw_buf(),
* arc_alloc_compressed_buf() or arc_alloc_buf().*
*
* NOTE: Declared noinline to avoid stack bloat in dbuf_hold_impl().
*/
noinline static void
dbuf_hold_copy(dnode_t *dn, dmu_buf_impl_t *db)
{
dbuf_dirty_record_t *dr = db->db_data_pending;
arc_buf_t *data = dr->dt.dl.dr_data;
enum zio_compress compress_type = arc_get_compression(data);
uint8_t complevel = arc_get_complevel(data);
if (arc_is_encrypted(data)) {
boolean_t byteorder;
uint8_t salt[ZIO_DATA_SALT_LEN];
uint8_t iv[ZIO_DATA_IV_LEN];
uint8_t mac[ZIO_DATA_MAC_LEN];
arc_get_raw_params(data, &byteorder, salt, iv, mac);
dbuf_set_data(db, arc_alloc_raw_buf(dn->dn_objset->os_spa, db,
dmu_objset_id(dn->dn_objset), byteorder, salt, iv, mac,
dn->dn_type, arc_buf_size(data), arc_buf_lsize(data),
compress_type, complevel));
} else if (compress_type != ZIO_COMPRESS_OFF) {
dbuf_set_data(db, arc_alloc_compressed_buf(
dn->dn_objset->os_spa, db, arc_buf_size(data),
arc_buf_lsize(data), compress_type, complevel));
} else {
dbuf_set_data(db, arc_alloc_buf(dn->dn_objset->os_spa, db,
DBUF_GET_BUFC_TYPE(db), db->db.db_size));
}
rw_enter(&db->db_rwlock, RW_WRITER);
memcpy(db->db.db_data, data->b_data, arc_buf_size(data));
rw_exit(&db->db_rwlock);
}
/*
* Returns with db_holds incremented, and db_mtx not held.
* Note: dn_struct_rwlock must be held.
*/
int
dbuf_hold_impl(dnode_t *dn, uint8_t level, uint64_t blkid,
boolean_t fail_sparse, boolean_t fail_uncached,
const void *tag, dmu_buf_impl_t **dbp)
{
dmu_buf_impl_t *db, *parent = NULL;
uint64_t hv;
/* If the pool has been created, verify the tx_sync_lock is not held */
spa_t *spa = dn->dn_objset->os_spa;
dsl_pool_t *dp = spa->spa_dsl_pool;
if (dp != NULL) {
ASSERT(!MUTEX_HELD(&dp->dp_tx.tx_sync_lock));
}
ASSERT(blkid != DMU_BONUS_BLKID);
ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock));
ASSERT3U(dn->dn_nlevels, >, level);
*dbp = NULL;
/* dbuf_find() returns with db_mtx held */
db = dbuf_find(dn->dn_objset, dn->dn_object, level, blkid, &hv);
if (db == NULL) {
blkptr_t *bp = NULL;
int err;
if (fail_uncached)
return (SET_ERROR(ENOENT));
ASSERT3P(parent, ==, NULL);
err = dbuf_findbp(dn, level, blkid, fail_sparse, &parent, &bp);
if (fail_sparse) {
if (err == 0 && bp && BP_IS_HOLE(bp))
err = SET_ERROR(ENOENT);
if (err) {
if (parent)
dbuf_rele(parent, NULL);
return (err);
}
}
if (err && err != ENOENT)
return (err);
db = dbuf_create(dn, level, blkid, parent, bp, hv);
}
if (fail_uncached && db->db_state != DB_CACHED) {
mutex_exit(&db->db_mtx);
return (SET_ERROR(ENOENT));
}
if (db->db_buf != NULL) {
arc_buf_access(db->db_buf);
ASSERT3P(db->db.db_data, ==, db->db_buf->b_data);
}
ASSERT(db->db_buf == NULL || arc_referenced(db->db_buf));
/*
* If this buffer is currently syncing out, and we are
* still referencing it from db_data, we need to make a copy
* of it in case we decide we want to dirty it again in this txg.
*/
if (db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID &&
dn->dn_object != DMU_META_DNODE_OBJECT &&
db->db_state == DB_CACHED && db->db_data_pending) {
dbuf_dirty_record_t *dr = db->db_data_pending;
if (dr->dt.dl.dr_data == db->db_buf) {
ASSERT3P(db->db_buf, !=, NULL);
dbuf_hold_copy(dn, db);
}
}
if (multilist_link_active(&db->db_cache_link)) {
ASSERT(zfs_refcount_is_zero(&db->db_holds));
ASSERT(db->db_caching_status == DB_DBUF_CACHE ||
db->db_caching_status == DB_DBUF_METADATA_CACHE);
multilist_remove(&dbuf_caches[db->db_caching_status].cache, db);
(void) zfs_refcount_remove_many(
&dbuf_caches[db->db_caching_status].size,
db->db.db_size, db);
if (db->db_caching_status == DB_DBUF_METADATA_CACHE) {
DBUF_STAT_BUMPDOWN(metadata_cache_count);
} else {
DBUF_STAT_BUMPDOWN(cache_levels[db->db_level]);
DBUF_STAT_BUMPDOWN(cache_count);
DBUF_STAT_DECR(cache_levels_bytes[db->db_level],
db->db.db_size);
}
db->db_caching_status = DB_NO_CACHE;
}
(void) zfs_refcount_add(&db->db_holds, tag);
DBUF_VERIFY(db);
mutex_exit(&db->db_mtx);
/* NOTE: we can't rele the parent until after we drop the db_mtx */
if (parent)
dbuf_rele(parent, NULL);
ASSERT3P(DB_DNODE(db), ==, dn);
ASSERT3U(db->db_blkid, ==, blkid);
ASSERT3U(db->db_level, ==, level);
*dbp = db;
return (0);
}
dmu_buf_impl_t *
dbuf_hold(dnode_t *dn, uint64_t blkid, const void *tag)
{
return (dbuf_hold_level(dn, 0, blkid, tag));
}
dmu_buf_impl_t *
dbuf_hold_level(dnode_t *dn, int level, uint64_t blkid, const void *tag)
{
dmu_buf_impl_t *db;
int err = dbuf_hold_impl(dn, level, blkid, FALSE, FALSE, tag, &db);
return (err ? NULL : db);
}
void
dbuf_create_bonus(dnode_t *dn)
{
ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock));
ASSERT(dn->dn_bonus == NULL);
dn->dn_bonus = dbuf_create(dn, 0, DMU_BONUS_BLKID, dn->dn_dbuf, NULL,
dbuf_hash(dn->dn_objset, dn->dn_object, 0, DMU_BONUS_BLKID));
}
int
dbuf_spill_set_blksz(dmu_buf_t *db_fake, uint64_t blksz, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
if (db->db_blkid != DMU_SPILL_BLKID)
return (SET_ERROR(ENOTSUP));
if (blksz == 0)
blksz = SPA_MINBLOCKSIZE;
ASSERT3U(blksz, <=, spa_maxblocksize(dmu_objset_spa(db->db_objset)));
blksz = P2ROUNDUP(blksz, SPA_MINBLOCKSIZE);
dbuf_new_size(db, blksz, tx);
return (0);
}
void
dbuf_rm_spill(dnode_t *dn, dmu_tx_t *tx)
{
dbuf_free_range(dn, DMU_SPILL_BLKID, DMU_SPILL_BLKID, tx);
}
#pragma weak dmu_buf_add_ref = dbuf_add_ref
void
dbuf_add_ref(dmu_buf_impl_t *db, const void *tag)
{
int64_t holds = zfs_refcount_add(&db->db_holds, tag);
VERIFY3S(holds, >, 1);
}
#pragma weak dmu_buf_try_add_ref = dbuf_try_add_ref
boolean_t
dbuf_try_add_ref(dmu_buf_t *db_fake, objset_t *os, uint64_t obj, uint64_t blkid,
const void *tag)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
dmu_buf_impl_t *found_db;
boolean_t result = B_FALSE;
if (blkid == DMU_BONUS_BLKID)
found_db = dbuf_find_bonus(os, obj);
else
found_db = dbuf_find(os, obj, 0, blkid, NULL);
if (found_db != NULL) {
if (db == found_db && dbuf_refcount(db) > db->db_dirtycnt) {
(void) zfs_refcount_add(&db->db_holds, tag);
result = B_TRUE;
}
mutex_exit(&found_db->db_mtx);
}
return (result);
}
/*
* If you call dbuf_rele() you had better not be referencing the dnode handle
* unless you have some other direct or indirect hold on the dnode. (An indirect
* hold is a hold on one of the dnode's dbufs, including the bonus buffer.)
* Without that, the dbuf_rele() could lead to a dnode_rele() followed by the
* dnode's parent dbuf evicting its dnode handles.
*/
void
dbuf_rele(dmu_buf_impl_t *db, const void *tag)
{
mutex_enter(&db->db_mtx);
dbuf_rele_and_unlock(db, tag, B_FALSE);
}
void
dmu_buf_rele(dmu_buf_t *db, const void *tag)
{
dbuf_rele((dmu_buf_impl_t *)db, tag);
}
/*
* dbuf_rele() for an already-locked dbuf. This is necessary to allow
* db_dirtycnt and db_holds to be updated atomically. The 'evicting'
* argument should be set if we are already in the dbuf-evicting code
* path, in which case we don't want to recursively evict. This allows us to
* avoid deeply nested stacks that would have a call flow similar to this:
*
* dbuf_rele()-->dbuf_rele_and_unlock()-->dbuf_evict_notify()
* ^ |
* | |
* +-----dbuf_destroy()<--dbuf_evict_one()<--------+
*
*/
void
dbuf_rele_and_unlock(dmu_buf_impl_t *db, const void *tag, boolean_t evicting)
{
int64_t holds;
uint64_t size;
ASSERT(MUTEX_HELD(&db->db_mtx));
DBUF_VERIFY(db);
/*
* Remove the reference to the dbuf before removing its hold on the
* dnode so we can guarantee in dnode_move() that a referenced bonus
* buffer has a corresponding dnode hold.
*/
holds = zfs_refcount_remove(&db->db_holds, tag);
ASSERT(holds >= 0);
/*
* We can't freeze indirects if there is a possibility that they
* may be modified in the current syncing context.
*/
if (db->db_buf != NULL &&
holds == (db->db_level == 0 ? db->db_dirtycnt : 0)) {
arc_buf_freeze(db->db_buf);
}
if (holds == db->db_dirtycnt &&
db->db_level == 0 && db->db_user_immediate_evict)
dbuf_evict_user(db);
if (holds == 0) {
if (db->db_blkid == DMU_BONUS_BLKID) {
dnode_t *dn;
boolean_t evict_dbuf = db->db_pending_evict;
/*
* If the dnode moves here, we cannot cross this
* barrier until the move completes.
*/
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
atomic_dec_32(&dn->dn_dbufs_count);
/*
* Decrementing the dbuf count means that the bonus
* buffer's dnode hold is no longer discounted in
* dnode_move(). The dnode cannot move until after
* the dnode_rele() below.
*/
DB_DNODE_EXIT(db);
/*
* Do not reference db after its lock is dropped.
* Another thread may evict it.
*/
mutex_exit(&db->db_mtx);
if (evict_dbuf)
dnode_evict_bonus(dn);
dnode_rele(dn, db);
} else if (db->db_buf == NULL) {
/*
* This is a special case: we never associated this
* dbuf with any data allocated from the ARC.
*/
ASSERT(db->db_state == DB_UNCACHED ||
db->db_state == DB_NOFILL);
dbuf_destroy(db);
} else if (arc_released(db->db_buf)) {
/*
* This dbuf has anonymous data associated with it.
*/
dbuf_destroy(db);
} else if (!(DBUF_IS_CACHEABLE(db) || db->db_partial_read) ||
db->db_pending_evict) {
dbuf_destroy(db);
} else if (!multilist_link_active(&db->db_cache_link)) {
ASSERT3U(db->db_caching_status, ==, DB_NO_CACHE);
dbuf_cached_state_t dcs =
dbuf_include_in_metadata_cache(db) ?
DB_DBUF_METADATA_CACHE : DB_DBUF_CACHE;
db->db_caching_status = dcs;
multilist_insert(&dbuf_caches[dcs].cache, db);
uint64_t db_size = db->db.db_size;
size = zfs_refcount_add_many(
&dbuf_caches[dcs].size, db_size, db);
uint8_t db_level = db->db_level;
mutex_exit(&db->db_mtx);
if (dcs == DB_DBUF_METADATA_CACHE) {
DBUF_STAT_BUMP(metadata_cache_count);
DBUF_STAT_MAX(metadata_cache_size_bytes_max,
size);
} else {
DBUF_STAT_BUMP(cache_count);
DBUF_STAT_MAX(cache_size_bytes_max, size);
DBUF_STAT_BUMP(cache_levels[db_level]);
DBUF_STAT_INCR(cache_levels_bytes[db_level],
db_size);
}
if (dcs == DB_DBUF_CACHE && !evicting)
dbuf_evict_notify(size);
}
} else {
mutex_exit(&db->db_mtx);
}
}
#pragma weak dmu_buf_refcount = dbuf_refcount
uint64_t
dbuf_refcount(dmu_buf_impl_t *db)
{
return (zfs_refcount_count(&db->db_holds));
}
uint64_t
dmu_buf_user_refcount(dmu_buf_t *db_fake)
{
uint64_t holds;
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
mutex_enter(&db->db_mtx);
ASSERT3U(zfs_refcount_count(&db->db_holds), >=, db->db_dirtycnt);
holds = zfs_refcount_count(&db->db_holds) - db->db_dirtycnt;
mutex_exit(&db->db_mtx);
return (holds);
}
void *
dmu_buf_replace_user(dmu_buf_t *db_fake, dmu_buf_user_t *old_user,
dmu_buf_user_t *new_user)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
mutex_enter(&db->db_mtx);
dbuf_verify_user(db, DBVU_NOT_EVICTING);
if (db->db_user == old_user)
db->db_user = new_user;
else
old_user = db->db_user;
dbuf_verify_user(db, DBVU_NOT_EVICTING);
mutex_exit(&db->db_mtx);
return (old_user);
}
void *
dmu_buf_set_user(dmu_buf_t *db_fake, dmu_buf_user_t *user)
{
return (dmu_buf_replace_user(db_fake, NULL, user));
}
void *
dmu_buf_set_user_ie(dmu_buf_t *db_fake, dmu_buf_user_t *user)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
db->db_user_immediate_evict = TRUE;
return (dmu_buf_set_user(db_fake, user));
}
void *
dmu_buf_remove_user(dmu_buf_t *db_fake, dmu_buf_user_t *user)
{
return (dmu_buf_replace_user(db_fake, user, NULL));
}
void *
dmu_buf_get_user(dmu_buf_t *db_fake)
{
dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake;
dbuf_verify_user(db, DBVU_NOT_EVICTING);
return (db->db_user);
}
void
dmu_buf_user_evict_wait(void)
{
taskq_wait(dbu_evict_taskq);
}
blkptr_t *
dmu_buf_get_blkptr(dmu_buf_t *db)
{
dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db;
return (dbi->db_blkptr);
}
objset_t *
dmu_buf_get_objset(dmu_buf_t *db)
{
dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db;
return (dbi->db_objset);
}
dnode_t *
dmu_buf_dnode_enter(dmu_buf_t *db)
{
dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db;
DB_DNODE_ENTER(dbi);
return (DB_DNODE(dbi));
}
void
dmu_buf_dnode_exit(dmu_buf_t *db)
{
dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db;
DB_DNODE_EXIT(dbi);
}
static void
dbuf_check_blkptr(dnode_t *dn, dmu_buf_impl_t *db)
{
/* ASSERT(dmu_tx_is_syncing(tx) */
ASSERT(MUTEX_HELD(&db->db_mtx));
if (db->db_blkptr != NULL)
return;
if (db->db_blkid == DMU_SPILL_BLKID) {
db->db_blkptr = DN_SPILL_BLKPTR(dn->dn_phys);
BP_ZERO(db->db_blkptr);
return;
}
if (db->db_level == dn->dn_phys->dn_nlevels-1) {
/*
* This buffer was allocated at a time when there was
* no available blkptrs from the dnode, or it was
* inappropriate to hook it in (i.e., nlevels mismatch).
*/
ASSERT(db->db_blkid < dn->dn_phys->dn_nblkptr);
ASSERT(db->db_parent == NULL);
db->db_parent = dn->dn_dbuf;
db->db_blkptr = &dn->dn_phys->dn_blkptr[db->db_blkid];
DBUF_VERIFY(db);
} else {
dmu_buf_impl_t *parent = db->db_parent;
int epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT;
ASSERT(dn->dn_phys->dn_nlevels > 1);
if (parent == NULL) {
mutex_exit(&db->db_mtx);
rw_enter(&dn->dn_struct_rwlock, RW_READER);
parent = dbuf_hold_level(dn, db->db_level + 1,
db->db_blkid >> epbs, db);
rw_exit(&dn->dn_struct_rwlock);
mutex_enter(&db->db_mtx);
db->db_parent = parent;
}
db->db_blkptr = (blkptr_t *)parent->db.db_data +
(db->db_blkid & ((1ULL << epbs) - 1));
DBUF_VERIFY(db);
}
}
static void
dbuf_sync_bonus(dbuf_dirty_record_t *dr, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = dr->dr_dbuf;
void *data = dr->dt.dl.dr_data;
ASSERT0(db->db_level);
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT(db->db_blkid == DMU_BONUS_BLKID);
ASSERT(data != NULL);
dnode_t *dn = dr->dr_dnode;
ASSERT3U(DN_MAX_BONUS_LEN(dn->dn_phys), <=,
DN_SLOTS_TO_BONUSLEN(dn->dn_phys->dn_extra_slots + 1));
memcpy(DN_BONUS(dn->dn_phys), data, DN_MAX_BONUS_LEN(dn->dn_phys));
dbuf_sync_leaf_verify_bonus_dnode(dr);
dbuf_undirty_bonus(dr);
dbuf_rele_and_unlock(db, (void *)(uintptr_t)tx->tx_txg, B_FALSE);
}
/*
* When syncing out a blocks of dnodes, adjust the block to deal with
* encryption. Normally, we make sure the block is decrypted before writing
* it. If we have crypt params, then we are writing a raw (encrypted) block,
* from a raw receive. In this case, set the ARC buf's crypt params so
* that the BP will be filled with the correct byteorder, salt, iv, and mac.
*/
static void
dbuf_prepare_encrypted_dnode_leaf(dbuf_dirty_record_t *dr)
{
int err;
dmu_buf_impl_t *db = dr->dr_dbuf;
ASSERT(MUTEX_HELD(&db->db_mtx));
ASSERT3U(db->db.db_object, ==, DMU_META_DNODE_OBJECT);
ASSERT3U(db->db_level, ==, 0);
if (!db->db_objset->os_raw_receive && arc_is_encrypted(db->db_buf)) {
zbookmark_phys_t zb;
/*
* Unfortunately, there is currently no mechanism for
* syncing context to handle decryption errors. An error
* here is only possible if an attacker maliciously
* changed a dnode block and updated the associated
* checksums going up the block tree.
*/
SET_BOOKMARK(&zb, dmu_objset_id(db->db_objset),
db->db.db_object, db->db_level, db->db_blkid);
err = arc_untransform(db->db_buf, db->db_objset->os_spa,
&zb, B_TRUE);
if (err)
panic("Invalid dnode block MAC");
} else if (dr->dt.dl.dr_has_raw_params) {
(void) arc_release(dr->dt.dl.dr_data, db);
arc_convert_to_raw(dr->dt.dl.dr_data,
dmu_objset_id(db->db_objset),
dr->dt.dl.dr_byteorder, DMU_OT_DNODE,
dr->dt.dl.dr_salt, dr->dt.dl.dr_iv, dr->dt.dl.dr_mac);
}
}
/*
* dbuf_sync_indirect() is called recursively from dbuf_sync_list() so it
* is critical the we not allow the compiler to inline this function in to
* dbuf_sync_list() thereby drastically bloating the stack usage.
*/
noinline static void
dbuf_sync_indirect(dbuf_dirty_record_t *dr, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = dr->dr_dbuf;
dnode_t *dn = dr->dr_dnode;
ASSERT(dmu_tx_is_syncing(tx));
dprintf_dbuf_bp(db, db->db_blkptr, "blkptr=%p", db->db_blkptr);
mutex_enter(&db->db_mtx);
ASSERT(db->db_level > 0);
DBUF_VERIFY(db);
/* Read the block if it hasn't been read yet. */
if (db->db_buf == NULL) {
mutex_exit(&db->db_mtx);
(void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED);
mutex_enter(&db->db_mtx);
}
ASSERT3U(db->db_state, ==, DB_CACHED);
ASSERT(db->db_buf != NULL);
/* Indirect block size must match what the dnode thinks it is. */
ASSERT3U(db->db.db_size, ==, 1<<dn->dn_phys->dn_indblkshift);
dbuf_check_blkptr(dn, db);
/* Provide the pending dirty record to child dbufs */
db->db_data_pending = dr;
mutex_exit(&db->db_mtx);
dbuf_write(dr, db->db_buf, tx);
zio_t *zio = dr->dr_zio;
mutex_enter(&dr->dt.di.dr_mtx);
dbuf_sync_list(&dr->dt.di.dr_children, db->db_level - 1, tx);
ASSERT(list_head(&dr->dt.di.dr_children) == NULL);
mutex_exit(&dr->dt.di.dr_mtx);
zio_nowait(zio);
}
/*
* Verify that the size of the data in our bonus buffer does not exceed
* its recorded size.
*
* The purpose of this verification is to catch any cases in development
* where the size of a phys structure (i.e space_map_phys_t) grows and,
* due to incorrect feature management, older pools expect to read more
* data even though they didn't actually write it to begin with.
*
* For a example, this would catch an error in the feature logic where we
* open an older pool and we expect to write the space map histogram of
* a space map with size SPACE_MAP_SIZE_V0.
*/
static void
dbuf_sync_leaf_verify_bonus_dnode(dbuf_dirty_record_t *dr)
{
#ifdef ZFS_DEBUG
dnode_t *dn = dr->dr_dnode;
/*
* Encrypted bonus buffers can have data past their bonuslen.
* Skip the verification of these blocks.
*/
if (DMU_OT_IS_ENCRYPTED(dn->dn_bonustype))
return;
uint16_t bonuslen = dn->dn_phys->dn_bonuslen;
uint16_t maxbonuslen = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots);
ASSERT3U(bonuslen, <=, maxbonuslen);
arc_buf_t *datap = dr->dt.dl.dr_data;
char *datap_end = ((char *)datap) + bonuslen;
char *datap_max = ((char *)datap) + maxbonuslen;
/* ensure that everything is zero after our data */
for (; datap_end < datap_max; datap_end++)
ASSERT(*datap_end == 0);
#endif
}
static blkptr_t *
dbuf_lightweight_bp(dbuf_dirty_record_t *dr)
{
/* This must be a lightweight dirty record. */
ASSERT3P(dr->dr_dbuf, ==, NULL);
dnode_t *dn = dr->dr_dnode;
if (dn->dn_phys->dn_nlevels == 1) {
VERIFY3U(dr->dt.dll.dr_blkid, <, dn->dn_phys->dn_nblkptr);
return (&dn->dn_phys->dn_blkptr[dr->dt.dll.dr_blkid]);
} else {
dmu_buf_impl_t *parent_db = dr->dr_parent->dr_dbuf;
int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT;
VERIFY3U(parent_db->db_level, ==, 1);
VERIFY3P(parent_db->db_dnode_handle->dnh_dnode, ==, dn);
VERIFY3U(dr->dt.dll.dr_blkid >> epbs, ==, parent_db->db_blkid);
blkptr_t *bp = parent_db->db.db_data;
return (&bp[dr->dt.dll.dr_blkid & ((1 << epbs) - 1)]);
}
}
static void
dbuf_lightweight_ready(zio_t *zio)
{
dbuf_dirty_record_t *dr = zio->io_private;
blkptr_t *bp = zio->io_bp;
if (zio->io_error != 0)
return;
dnode_t *dn = dr->dr_dnode;
blkptr_t *bp_orig = dbuf_lightweight_bp(dr);
spa_t *spa = dmu_objset_spa(dn->dn_objset);
int64_t delta = bp_get_dsize_sync(spa, bp) -
bp_get_dsize_sync(spa, bp_orig);
dnode_diduse_space(dn, delta);
uint64_t blkid = dr->dt.dll.dr_blkid;
mutex_enter(&dn->dn_mtx);
if (blkid > dn->dn_phys->dn_maxblkid) {
ASSERT0(dn->dn_objset->os_raw_receive);
dn->dn_phys->dn_maxblkid = blkid;
}
mutex_exit(&dn->dn_mtx);
if (!BP_IS_EMBEDDED(bp)) {
uint64_t fill = BP_IS_HOLE(bp) ? 0 : 1;
BP_SET_FILL(bp, fill);
}
dmu_buf_impl_t *parent_db;
EQUIV(dr->dr_parent == NULL, dn->dn_phys->dn_nlevels == 1);
if (dr->dr_parent == NULL) {
parent_db = dn->dn_dbuf;
} else {
parent_db = dr->dr_parent->dr_dbuf;
}
rw_enter(&parent_db->db_rwlock, RW_WRITER);
*bp_orig = *bp;
rw_exit(&parent_db->db_rwlock);
}
static void
dbuf_lightweight_physdone(zio_t *zio)
{
dbuf_dirty_record_t *dr = zio->io_private;
dsl_pool_t *dp = spa_get_dsl(zio->io_spa);
ASSERT3U(dr->dr_txg, ==, zio->io_txg);
/*
* The callback will be called io_phys_children times. Retire one
* portion of our dirty space each time we are called. Any rounding
* error will be cleaned up by dbuf_lightweight_done().
*/
int delta = dr->dr_accounted / zio->io_phys_children;
dsl_pool_undirty_space(dp, delta, zio->io_txg);
}
static void
dbuf_lightweight_done(zio_t *zio)
{
dbuf_dirty_record_t *dr = zio->io_private;
VERIFY0(zio->io_error);
objset_t *os = dr->dr_dnode->dn_objset;
dmu_tx_t *tx = os->os_synctx;
if (zio->io_flags & (ZIO_FLAG_IO_REWRITE | ZIO_FLAG_NOPWRITE)) {
ASSERT(BP_EQUAL(zio->io_bp, &zio->io_bp_orig));
} else {
dsl_dataset_t *ds = os->os_dsl_dataset;
(void) dsl_dataset_block_kill(ds, &zio->io_bp_orig, tx, B_TRUE);
dsl_dataset_block_born(ds, zio->io_bp, tx);
}
/*
* See comment in dbuf_write_done().
*/
if (zio->io_phys_children == 0) {
dsl_pool_undirty_space(dmu_objset_pool(os),
dr->dr_accounted, zio->io_txg);
} else {
dsl_pool_undirty_space(dmu_objset_pool(os),
dr->dr_accounted % zio->io_phys_children, zio->io_txg);
}
abd_free(dr->dt.dll.dr_abd);
kmem_free(dr, sizeof (*dr));
}
noinline static void
dbuf_sync_lightweight(dbuf_dirty_record_t *dr, dmu_tx_t *tx)
{
dnode_t *dn = dr->dr_dnode;
zio_t *pio;
if (dn->dn_phys->dn_nlevels == 1) {
pio = dn->dn_zio;
} else {
pio = dr->dr_parent->dr_zio;
}
zbookmark_phys_t zb = {
.zb_objset = dmu_objset_id(dn->dn_objset),
.zb_object = dn->dn_object,
.zb_level = 0,
.zb_blkid = dr->dt.dll.dr_blkid,
};
/*
* See comment in dbuf_write(). This is so that zio->io_bp_orig
* will have the old BP in dbuf_lightweight_done().
*/
dr->dr_bp_copy = *dbuf_lightweight_bp(dr);
dr->dr_zio = zio_write(pio, dmu_objset_spa(dn->dn_objset),
dmu_tx_get_txg(tx), &dr->dr_bp_copy, dr->dt.dll.dr_abd,
dn->dn_datablksz, abd_get_size(dr->dt.dll.dr_abd),
&dr->dt.dll.dr_props, dbuf_lightweight_ready, NULL,
dbuf_lightweight_physdone, dbuf_lightweight_done, dr,
ZIO_PRIORITY_ASYNC_WRITE,
ZIO_FLAG_MUSTSUCCEED | dr->dt.dll.dr_flags, &zb);
zio_nowait(dr->dr_zio);
}
/*
* dbuf_sync_leaf() is called recursively from dbuf_sync_list() so it is
* critical the we not allow the compiler to inline this function in to
* dbuf_sync_list() thereby drastically bloating the stack usage.
*/
noinline static void
dbuf_sync_leaf(dbuf_dirty_record_t *dr, dmu_tx_t *tx)
{
arc_buf_t **datap = &dr->dt.dl.dr_data;
dmu_buf_impl_t *db = dr->dr_dbuf;
dnode_t *dn = dr->dr_dnode;
objset_t *os;
uint64_t txg = tx->tx_txg;
ASSERT(dmu_tx_is_syncing(tx));
dprintf_dbuf_bp(db, db->db_blkptr, "blkptr=%p", db->db_blkptr);
mutex_enter(&db->db_mtx);
/*
* To be synced, we must be dirtied. But we
* might have been freed after the dirty.
*/
if (db->db_state == DB_UNCACHED) {
/* This buffer has been freed since it was dirtied */
ASSERT(db->db.db_data == NULL);
} else if (db->db_state == DB_FILL) {
/* This buffer was freed and is now being re-filled */
ASSERT(db->db.db_data != dr->dt.dl.dr_data);
} else {
ASSERT(db->db_state == DB_CACHED || db->db_state == DB_NOFILL);
}
DBUF_VERIFY(db);
if (db->db_blkid == DMU_SPILL_BLKID) {
mutex_enter(&dn->dn_mtx);
if (!(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR)) {
/*
* In the previous transaction group, the bonus buffer
* was entirely used to store the attributes for the
* dnode which overrode the dn_spill field. However,
* when adding more attributes to the file a spill
* block was required to hold the extra attributes.
*
* Make sure to clear the garbage left in the dn_spill
* field from the previous attributes in the bonus
* buffer. Otherwise, after writing out the spill
* block to the new allocated dva, it will free
* the old block pointed to by the invalid dn_spill.
*/
db->db_blkptr = NULL;
}
dn->dn_phys->dn_flags |= DNODE_FLAG_SPILL_BLKPTR;
mutex_exit(&dn->dn_mtx);
}
/*
* If this is a bonus buffer, simply copy the bonus data into the
* dnode. It will be written out when the dnode is synced (and it
* will be synced, since it must have been dirty for dbuf_sync to
* be called).
*/
if (db->db_blkid == DMU_BONUS_BLKID) {
ASSERT(dr->dr_dbuf == db);
dbuf_sync_bonus(dr, tx);
return;
}
os = dn->dn_objset;
/*
* This function may have dropped the db_mtx lock allowing a dmu_sync
* operation to sneak in. As a result, we need to ensure that we
* don't check the dr_override_state until we have returned from
* dbuf_check_blkptr.
*/
dbuf_check_blkptr(dn, db);
/*
* If this buffer is in the middle of an immediate write,
* wait for the synchronous IO to complete.
*/
while (dr->dt.dl.dr_override_state == DR_IN_DMU_SYNC) {
ASSERT(dn->dn_object != DMU_META_DNODE_OBJECT);
cv_wait(&db->db_changed, &db->db_mtx);
}
/*
* If this is a dnode block, ensure it is appropriately encrypted
* or decrypted, depending on what we are writing to it this txg.
*/
if (os->os_encrypted && dn->dn_object == DMU_META_DNODE_OBJECT)
dbuf_prepare_encrypted_dnode_leaf(dr);
if (db->db_state != DB_NOFILL &&
dn->dn_object != DMU_META_DNODE_OBJECT &&
zfs_refcount_count(&db->db_holds) > 1 &&
dr->dt.dl.dr_override_state != DR_OVERRIDDEN &&
*datap == db->db_buf) {
/*
* If this buffer is currently "in use" (i.e., there
* are active holds and db_data still references it),
* then make a copy before we start the write so that
* any modifications from the open txg will not leak
* into this write.
*
* NOTE: this copy does not need to be made for
* objects only modified in the syncing context (e.g.
* DNONE_DNODE blocks).
*/
int psize = arc_buf_size(*datap);
int lsize = arc_buf_lsize(*datap);
arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db);
enum zio_compress compress_type = arc_get_compression(*datap);
uint8_t complevel = arc_get_complevel(*datap);
if (arc_is_encrypted(*datap)) {
boolean_t byteorder;
uint8_t salt[ZIO_DATA_SALT_LEN];
uint8_t iv[ZIO_DATA_IV_LEN];
uint8_t mac[ZIO_DATA_MAC_LEN];
arc_get_raw_params(*datap, &byteorder, salt, iv, mac);
*datap = arc_alloc_raw_buf(os->os_spa, db,
dmu_objset_id(os), byteorder, salt, iv, mac,
dn->dn_type, psize, lsize, compress_type,
complevel);
} else if (compress_type != ZIO_COMPRESS_OFF) {
ASSERT3U(type, ==, ARC_BUFC_DATA);
*datap = arc_alloc_compressed_buf(os->os_spa, db,
psize, lsize, compress_type, complevel);
} else {
*datap = arc_alloc_buf(os->os_spa, db, type, psize);
}
memcpy((*datap)->b_data, db->db.db_data, psize);
}
db->db_data_pending = dr;
mutex_exit(&db->db_mtx);
dbuf_write(dr, *datap, tx);
ASSERT(!list_link_active(&dr->dr_dirty_node));
if (dn->dn_object == DMU_META_DNODE_OBJECT) {
list_insert_tail(&dn->dn_dirty_records[txg & TXG_MASK], dr);
} else {
zio_nowait(dr->dr_zio);
}
}
void
dbuf_sync_list(list_t *list, int level, dmu_tx_t *tx)
{
dbuf_dirty_record_t *dr;
while ((dr = list_head(list))) {
if (dr->dr_zio != NULL) {
/*
* If we find an already initialized zio then we
* are processing the meta-dnode, and we have finished.
* The dbufs for all dnodes are put back on the list
* during processing, so that we can zio_wait()
* these IOs after initiating all child IOs.
*/
ASSERT3U(dr->dr_dbuf->db.db_object, ==,
DMU_META_DNODE_OBJECT);
break;
}
list_remove(list, dr);
if (dr->dr_dbuf == NULL) {
dbuf_sync_lightweight(dr, tx);
} else {
if (dr->dr_dbuf->db_blkid != DMU_BONUS_BLKID &&
dr->dr_dbuf->db_blkid != DMU_SPILL_BLKID) {
VERIFY3U(dr->dr_dbuf->db_level, ==, level);
}
if (dr->dr_dbuf->db_level > 0)
dbuf_sync_indirect(dr, tx);
else
dbuf_sync_leaf(dr, tx);
}
}
}
static void
dbuf_write_ready(zio_t *zio, arc_buf_t *buf, void *vdb)
{
(void) buf;
dmu_buf_impl_t *db = vdb;
dnode_t *dn;
blkptr_t *bp = zio->io_bp;
blkptr_t *bp_orig = &zio->io_bp_orig;
spa_t *spa = zio->io_spa;
int64_t delta;
uint64_t fill = 0;
int i;
ASSERT3P(db->db_blkptr, !=, NULL);
ASSERT3P(&db->db_data_pending->dr_bp_copy, ==, bp);
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
delta = bp_get_dsize_sync(spa, bp) - bp_get_dsize_sync(spa, bp_orig);
dnode_diduse_space(dn, delta - zio->io_prev_space_delta);
zio->io_prev_space_delta = delta;
if (bp->blk_birth != 0) {
ASSERT((db->db_blkid != DMU_SPILL_BLKID &&
BP_GET_TYPE(bp) == dn->dn_type) ||
(db->db_blkid == DMU_SPILL_BLKID &&
BP_GET_TYPE(bp) == dn->dn_bonustype) ||
BP_IS_EMBEDDED(bp));
ASSERT(BP_GET_LEVEL(bp) == db->db_level);
}
mutex_enter(&db->db_mtx);
#ifdef ZFS_DEBUG
if (db->db_blkid == DMU_SPILL_BLKID) {
ASSERT(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR);
ASSERT(!(BP_IS_HOLE(bp)) &&
db->db_blkptr == DN_SPILL_BLKPTR(dn->dn_phys));
}
#endif
if (db->db_level == 0) {
mutex_enter(&dn->dn_mtx);
if (db->db_blkid > dn->dn_phys->dn_maxblkid &&
db->db_blkid != DMU_SPILL_BLKID) {
ASSERT0(db->db_objset->os_raw_receive);
dn->dn_phys->dn_maxblkid = db->db_blkid;
}
mutex_exit(&dn->dn_mtx);
if (dn->dn_type == DMU_OT_DNODE) {
i = 0;
while (i < db->db.db_size) {
dnode_phys_t *dnp =
(void *)(((char *)db->db.db_data) + i);
i += DNODE_MIN_SIZE;
if (dnp->dn_type != DMU_OT_NONE) {
fill++;
i += dnp->dn_extra_slots *
DNODE_MIN_SIZE;
}
}
} else {
if (BP_IS_HOLE(bp)) {
fill = 0;
} else {
fill = 1;
}
}
} else {
blkptr_t *ibp = db->db.db_data;
ASSERT3U(db->db.db_size, ==, 1<<dn->dn_phys->dn_indblkshift);
for (i = db->db.db_size >> SPA_BLKPTRSHIFT; i > 0; i--, ibp++) {
if (BP_IS_HOLE(ibp))
continue;
fill += BP_GET_FILL(ibp);
}
}
DB_DNODE_EXIT(db);
if (!BP_IS_EMBEDDED(bp))
BP_SET_FILL(bp, fill);
mutex_exit(&db->db_mtx);
db_lock_type_t dblt = dmu_buf_lock_parent(db, RW_WRITER, FTAG);
*db->db_blkptr = *bp;
dmu_buf_unlock_parent(db, dblt, FTAG);
}
/*
* This function gets called just prior to running through the compression
* stage of the zio pipeline. If we're an indirect block comprised of only
* holes, then we want this indirect to be compressed away to a hole. In
* order to do that we must zero out any information about the holes that
* this indirect points to prior to before we try to compress it.
*/
static void
dbuf_write_children_ready(zio_t *zio, arc_buf_t *buf, void *vdb)
{
(void) zio, (void) buf;
dmu_buf_impl_t *db = vdb;
dnode_t *dn;
blkptr_t *bp;
unsigned int epbs, i;
ASSERT3U(db->db_level, >, 0);
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT;
ASSERT3U(epbs, <, 31);
/* Determine if all our children are holes */
for (i = 0, bp = db->db.db_data; i < 1ULL << epbs; i++, bp++) {
if (!BP_IS_HOLE(bp))
break;
}
/*
* If all the children are holes, then zero them all out so that
* we may get compressed away.
*/
if (i == 1ULL << epbs) {
/*
* We only found holes. Grab the rwlock to prevent
* anybody from reading the blocks we're about to
* zero out.
*/
rw_enter(&db->db_rwlock, RW_WRITER);
memset(db->db.db_data, 0, db->db.db_size);
rw_exit(&db->db_rwlock);
}
DB_DNODE_EXIT(db);
}
/*
* The SPA will call this callback several times for each zio - once
* for every physical child i/o (zio->io_phys_children times). This
* allows the DMU to monitor the progress of each logical i/o. For example,
* there may be 2 copies of an indirect block, or many fragments of a RAID-Z
* block. There may be a long delay before all copies/fragments are completed,
* so this callback allows us to retire dirty space gradually, as the physical
* i/os complete.
*/
static void
dbuf_write_physdone(zio_t *zio, arc_buf_t *buf, void *arg)
{
(void) buf;
dmu_buf_impl_t *db = arg;
objset_t *os = db->db_objset;
dsl_pool_t *dp = dmu_objset_pool(os);
dbuf_dirty_record_t *dr;
int delta = 0;
dr = db->db_data_pending;
ASSERT3U(dr->dr_txg, ==, zio->io_txg);
/*
* The callback will be called io_phys_children times. Retire one
* portion of our dirty space each time we are called. Any rounding
* error will be cleaned up by dbuf_write_done().
*/
delta = dr->dr_accounted / zio->io_phys_children;
dsl_pool_undirty_space(dp, delta, zio->io_txg);
}
static void
dbuf_write_done(zio_t *zio, arc_buf_t *buf, void *vdb)
{
(void) buf;
dmu_buf_impl_t *db = vdb;
blkptr_t *bp_orig = &zio->io_bp_orig;
blkptr_t *bp = db->db_blkptr;
objset_t *os = db->db_objset;
dmu_tx_t *tx = os->os_synctx;
ASSERT0(zio->io_error);
ASSERT(db->db_blkptr == bp);
/*
* For nopwrites and rewrites we ensure that the bp matches our
* original and bypass all the accounting.
*/
if (zio->io_flags & (ZIO_FLAG_IO_REWRITE | ZIO_FLAG_NOPWRITE)) {
ASSERT(BP_EQUAL(bp, bp_orig));
} else {
dsl_dataset_t *ds = os->os_dsl_dataset;
(void) dsl_dataset_block_kill(ds, bp_orig, tx, B_TRUE);
dsl_dataset_block_born(ds, bp, tx);
}
mutex_enter(&db->db_mtx);
DBUF_VERIFY(db);
dbuf_dirty_record_t *dr = db->db_data_pending;
dnode_t *dn = dr->dr_dnode;
ASSERT(!list_link_active(&dr->dr_dirty_node));
ASSERT(dr->dr_dbuf == db);
ASSERT(list_next(&db->db_dirty_records, dr) == NULL);
list_remove(&db->db_dirty_records, dr);
#ifdef ZFS_DEBUG
if (db->db_blkid == DMU_SPILL_BLKID) {
ASSERT(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR);
ASSERT(!(BP_IS_HOLE(db->db_blkptr)) &&
db->db_blkptr == DN_SPILL_BLKPTR(dn->dn_phys));
}
#endif
if (db->db_level == 0) {
ASSERT(db->db_blkid != DMU_BONUS_BLKID);
ASSERT(dr->dt.dl.dr_override_state == DR_NOT_OVERRIDDEN);
if (db->db_state != DB_NOFILL) {
if (dr->dt.dl.dr_data != NULL &&
dr->dt.dl.dr_data != db->db_buf) {
arc_buf_destroy(dr->dt.dl.dr_data, db);
}
}
} else {
ASSERT(list_head(&dr->dt.di.dr_children) == NULL);
ASSERT3U(db->db.db_size, ==, 1 << dn->dn_phys->dn_indblkshift);
if (!BP_IS_HOLE(db->db_blkptr)) {
int epbs __maybe_unused = dn->dn_phys->dn_indblkshift -
SPA_BLKPTRSHIFT;
ASSERT3U(db->db_blkid, <=,
dn->dn_phys->dn_maxblkid >> (db->db_level * epbs));
ASSERT3U(BP_GET_LSIZE(db->db_blkptr), ==,
db->db.db_size);
}
mutex_destroy(&dr->dt.di.dr_mtx);
list_destroy(&dr->dt.di.dr_children);
}
cv_broadcast(&db->db_changed);
ASSERT(db->db_dirtycnt > 0);
db->db_dirtycnt -= 1;
db->db_data_pending = NULL;
dbuf_rele_and_unlock(db, (void *)(uintptr_t)tx->tx_txg, B_FALSE);
/*
* If we didn't do a physical write in this ZIO and we
* still ended up here, it means that the space of the
* dbuf that we just released (and undirtied) above hasn't
* been marked as undirtied in the pool's accounting.
*
* Thus, we undirty that space in the pool's view of the
* world here. For physical writes this type of update
* happens in dbuf_write_physdone().
*
* If we did a physical write, cleanup any rounding errors
* that came up due to writing multiple copies of a block
* on disk [see dbuf_write_physdone()].
*/
if (zio->io_phys_children == 0) {
dsl_pool_undirty_space(dmu_objset_pool(os),
dr->dr_accounted, zio->io_txg);
} else {
dsl_pool_undirty_space(dmu_objset_pool(os),
dr->dr_accounted % zio->io_phys_children, zio->io_txg);
}
kmem_free(dr, sizeof (dbuf_dirty_record_t));
}
static void
dbuf_write_nofill_ready(zio_t *zio)
{
dbuf_write_ready(zio, NULL, zio->io_private);
}
static void
dbuf_write_nofill_done(zio_t *zio)
{
dbuf_write_done(zio, NULL, zio->io_private);
}
static void
dbuf_write_override_ready(zio_t *zio)
{
dbuf_dirty_record_t *dr = zio->io_private;
dmu_buf_impl_t *db = dr->dr_dbuf;
dbuf_write_ready(zio, NULL, db);
}
static void
dbuf_write_override_done(zio_t *zio)
{
dbuf_dirty_record_t *dr = zio->io_private;
dmu_buf_impl_t *db = dr->dr_dbuf;
blkptr_t *obp = &dr->dt.dl.dr_overridden_by;
mutex_enter(&db->db_mtx);
if (!BP_EQUAL(zio->io_bp, obp)) {
if (!BP_IS_HOLE(obp))
dsl_free(spa_get_dsl(zio->io_spa), zio->io_txg, obp);
arc_release(dr->dt.dl.dr_data, db);
}
mutex_exit(&db->db_mtx);
dbuf_write_done(zio, NULL, db);
if (zio->io_abd != NULL)
abd_free(zio->io_abd);
}
typedef struct dbuf_remap_impl_callback_arg {
objset_t *drica_os;
uint64_t drica_blk_birth;
dmu_tx_t *drica_tx;
} dbuf_remap_impl_callback_arg_t;
static void
dbuf_remap_impl_callback(uint64_t vdev, uint64_t offset, uint64_t size,
void *arg)
{
dbuf_remap_impl_callback_arg_t *drica = arg;
objset_t *os = drica->drica_os;
spa_t *spa = dmu_objset_spa(os);
dmu_tx_t *tx = drica->drica_tx;
ASSERT(dsl_pool_sync_context(spa_get_dsl(spa)));
if (os == spa_meta_objset(spa)) {
spa_vdev_indirect_mark_obsolete(spa, vdev, offset, size, tx);
} else {
dsl_dataset_block_remapped(dmu_objset_ds(os), vdev, offset,
size, drica->drica_blk_birth, tx);
}
}
static void
dbuf_remap_impl(dnode_t *dn, blkptr_t *bp, krwlock_t *rw, dmu_tx_t *tx)
{
blkptr_t bp_copy = *bp;
spa_t *spa = dmu_objset_spa(dn->dn_objset);
dbuf_remap_impl_callback_arg_t drica;
ASSERT(dsl_pool_sync_context(spa_get_dsl(spa)));
drica.drica_os = dn->dn_objset;
drica.drica_blk_birth = bp->blk_birth;
drica.drica_tx = tx;
if (spa_remap_blkptr(spa, &bp_copy, dbuf_remap_impl_callback,
&drica)) {
/*
* If the blkptr being remapped is tracked by a livelist,
* then we need to make sure the livelist reflects the update.
* First, cancel out the old blkptr by appending a 'FREE'
* entry. Next, add an 'ALLOC' to track the new version. This
* way we avoid trying to free an inaccurate blkptr at delete.
* Note that embedded blkptrs are not tracked in livelists.
*/
if (dn->dn_objset != spa_meta_objset(spa)) {
dsl_dataset_t *ds = dmu_objset_ds(dn->dn_objset);
if (dsl_deadlist_is_open(&ds->ds_dir->dd_livelist) &&
bp->blk_birth > ds->ds_dir->dd_origin_txg) {
ASSERT(!BP_IS_EMBEDDED(bp));
ASSERT(dsl_dir_is_clone(ds->ds_dir));
ASSERT(spa_feature_is_enabled(spa,
SPA_FEATURE_LIVELIST));
bplist_append(&ds->ds_dir->dd_pending_frees,
bp);
bplist_append(&ds->ds_dir->dd_pending_allocs,
&bp_copy);
}
}
/*
* The db_rwlock prevents dbuf_read_impl() from
* dereferencing the BP while we are changing it. To
* avoid lock contention, only grab it when we are actually
* changing the BP.
*/
if (rw != NULL)
rw_enter(rw, RW_WRITER);
*bp = bp_copy;
if (rw != NULL)
rw_exit(rw);
}
}
/*
* Remap any existing BP's to concrete vdevs, if possible.
*/
static void
dbuf_remap(dnode_t *dn, dmu_buf_impl_t *db, dmu_tx_t *tx)
{
spa_t *spa = dmu_objset_spa(db->db_objset);
ASSERT(dsl_pool_sync_context(spa_get_dsl(spa)));
if (!spa_feature_is_active(spa, SPA_FEATURE_DEVICE_REMOVAL))
return;
if (db->db_level > 0) {
blkptr_t *bp = db->db.db_data;
for (int i = 0; i < db->db.db_size >> SPA_BLKPTRSHIFT; i++) {
dbuf_remap_impl(dn, &bp[i], &db->db_rwlock, tx);
}
} else if (db->db.db_object == DMU_META_DNODE_OBJECT) {
dnode_phys_t *dnp = db->db.db_data;
ASSERT3U(db->db_dnode_handle->dnh_dnode->dn_type, ==,
DMU_OT_DNODE);
for (int i = 0; i < db->db.db_size >> DNODE_SHIFT;
i += dnp[i].dn_extra_slots + 1) {
for (int j = 0; j < dnp[i].dn_nblkptr; j++) {
krwlock_t *lock = (dn->dn_dbuf == NULL ? NULL :
&dn->dn_dbuf->db_rwlock);
dbuf_remap_impl(dn, &dnp[i].dn_blkptr[j], lock,
tx);
}
}
}
}
/* Issue I/O to commit a dirty buffer to disk. */
static void
dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx)
{
dmu_buf_impl_t *db = dr->dr_dbuf;
dnode_t *dn = dr->dr_dnode;
objset_t *os;
dmu_buf_impl_t *parent = db->db_parent;
uint64_t txg = tx->tx_txg;
zbookmark_phys_t zb;
zio_prop_t zp;
zio_t *pio; /* parent I/O */
int wp_flag = 0;
ASSERT(dmu_tx_is_syncing(tx));
os = dn->dn_objset;
if (db->db_state != DB_NOFILL) {
if (db->db_level > 0 || dn->dn_type == DMU_OT_DNODE) {
/*
* Private object buffers are released here rather
* than in dbuf_dirty() since they are only modified
* in the syncing context and we don't want the
* overhead of making multiple copies of the data.
*/
if (BP_IS_HOLE(db->db_blkptr)) {
arc_buf_thaw(data);
} else {
dbuf_release_bp(db);
}
dbuf_remap(dn, db, tx);
}
}
if (parent != dn->dn_dbuf) {
/* Our parent is an indirect block. */
/* We have a dirty parent that has been scheduled for write. */
ASSERT(parent && parent->db_data_pending);
/* Our parent's buffer is one level closer to the dnode. */
ASSERT(db->db_level == parent->db_level-1);
/*
* We're about to modify our parent's db_data by modifying
* our block pointer, so the parent must be released.
*/
ASSERT(arc_released(parent->db_buf));
pio = parent->db_data_pending->dr_zio;
} else {
/* Our parent is the dnode itself. */
ASSERT((db->db_level == dn->dn_phys->dn_nlevels-1 &&
db->db_blkid != DMU_SPILL_BLKID) ||
(db->db_blkid == DMU_SPILL_BLKID && db->db_level == 0));
if (db->db_blkid != DMU_SPILL_BLKID)
ASSERT3P(db->db_blkptr, ==,
&dn->dn_phys->dn_blkptr[db->db_blkid]);
pio = dn->dn_zio;
}
ASSERT(db->db_level == 0 || data == db->db_buf);
ASSERT3U(db->db_blkptr->blk_birth, <=, txg);
ASSERT(pio);
SET_BOOKMARK(&zb, os->os_dsl_dataset ?
os->os_dsl_dataset->ds_object : DMU_META_OBJSET,
db->db.db_object, db->db_level, db->db_blkid);
if (db->db_blkid == DMU_SPILL_BLKID)
wp_flag = WP_SPILL;
wp_flag |= (db->db_state == DB_NOFILL) ? WP_NOFILL : 0;
dmu_write_policy(os, dn, db->db_level, wp_flag, &zp);
/*
* We copy the blkptr now (rather than when we instantiate the dirty
* record), because its value can change between open context and
* syncing context. We do not need to hold dn_struct_rwlock to read
* db_blkptr because we are in syncing context.
*/
dr->dr_bp_copy = *db->db_blkptr;
if (db->db_level == 0 &&
dr->dt.dl.dr_override_state == DR_OVERRIDDEN) {
/*
* The BP for this block has been provided by open context
* (by dmu_sync() or dmu_buf_write_embedded()).
*/
abd_t *contents = (data != NULL) ?
abd_get_from_buf(data->b_data, arc_buf_size(data)) : NULL;
dr->dr_zio = zio_write(pio, os->os_spa, txg, &dr->dr_bp_copy,
contents, db->db.db_size, db->db.db_size, &zp,
dbuf_write_override_ready, NULL, NULL,
dbuf_write_override_done,
dr, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb);
mutex_enter(&db->db_mtx);
dr->dt.dl.dr_override_state = DR_NOT_OVERRIDDEN;
zio_write_override(dr->dr_zio, &dr->dt.dl.dr_overridden_by,
dr->dt.dl.dr_copies, dr->dt.dl.dr_nopwrite,
dr->dt.dl.dr_brtwrite);
mutex_exit(&db->db_mtx);
} else if (db->db_state == DB_NOFILL) {
ASSERT(zp.zp_checksum == ZIO_CHECKSUM_OFF ||
zp.zp_checksum == ZIO_CHECKSUM_NOPARITY);
dr->dr_zio = zio_write(pio, os->os_spa, txg,
&dr->dr_bp_copy, NULL, db->db.db_size, db->db.db_size, &zp,
dbuf_write_nofill_ready, NULL, NULL,
dbuf_write_nofill_done, db,
ZIO_PRIORITY_ASYNC_WRITE,
ZIO_FLAG_MUSTSUCCEED | ZIO_FLAG_NODATA, &zb);
} else {
ASSERT(arc_released(data));
/*
* For indirect blocks, we want to setup the children
* ready callback so that we can properly handle an indirect
* block that only contains holes.
*/
arc_write_done_func_t *children_ready_cb = NULL;
if (db->db_level != 0)
children_ready_cb = dbuf_write_children_ready;
dr->dr_zio = arc_write(pio, os->os_spa, txg,
&dr->dr_bp_copy, data, !DBUF_IS_CACHEABLE(db),
dbuf_is_l2cacheable(db), &zp, dbuf_write_ready,
children_ready_cb, dbuf_write_physdone,
dbuf_write_done, db, ZIO_PRIORITY_ASYNC_WRITE,
ZIO_FLAG_MUSTSUCCEED, &zb);
}
}
EXPORT_SYMBOL(dbuf_find);
EXPORT_SYMBOL(dbuf_is_metadata);
EXPORT_SYMBOL(dbuf_destroy);
EXPORT_SYMBOL(dbuf_loan_arcbuf);
EXPORT_SYMBOL(dbuf_whichblock);
EXPORT_SYMBOL(dbuf_read);
EXPORT_SYMBOL(dbuf_unoverride);
EXPORT_SYMBOL(dbuf_free_range);
EXPORT_SYMBOL(dbuf_new_size);
EXPORT_SYMBOL(dbuf_release_bp);
EXPORT_SYMBOL(dbuf_dirty);
EXPORT_SYMBOL(dmu_buf_set_crypt_params);
EXPORT_SYMBOL(dmu_buf_will_dirty);
EXPORT_SYMBOL(dmu_buf_is_dirty);
EXPORT_SYMBOL(dmu_buf_will_not_fill);
EXPORT_SYMBOL(dmu_buf_will_fill);
EXPORT_SYMBOL(dmu_buf_fill_done);
EXPORT_SYMBOL(dmu_buf_rele);
EXPORT_SYMBOL(dbuf_assign_arcbuf);
EXPORT_SYMBOL(dbuf_prefetch);
EXPORT_SYMBOL(dbuf_hold_impl);
EXPORT_SYMBOL(dbuf_hold);
EXPORT_SYMBOL(dbuf_hold_level);
EXPORT_SYMBOL(dbuf_create_bonus);
EXPORT_SYMBOL(dbuf_spill_set_blksz);
EXPORT_SYMBOL(dbuf_rm_spill);
EXPORT_SYMBOL(dbuf_add_ref);
EXPORT_SYMBOL(dbuf_rele);
EXPORT_SYMBOL(dbuf_rele_and_unlock);
EXPORT_SYMBOL(dbuf_refcount);
EXPORT_SYMBOL(dbuf_sync_list);
EXPORT_SYMBOL(dmu_buf_set_user);
EXPORT_SYMBOL(dmu_buf_set_user_ie);
EXPORT_SYMBOL(dmu_buf_get_user);
EXPORT_SYMBOL(dmu_buf_get_blkptr);
ZFS_MODULE_PARAM(zfs_dbuf_cache, dbuf_cache_, max_bytes, U64, ZMOD_RW,
"Maximum size in bytes of the dbuf cache.");
ZFS_MODULE_PARAM(zfs_dbuf_cache, dbuf_cache_, hiwater_pct, UINT, ZMOD_RW,
"Percentage over dbuf_cache_max_bytes for direct dbuf eviction.");
ZFS_MODULE_PARAM(zfs_dbuf_cache, dbuf_cache_, lowater_pct, UINT, ZMOD_RW,
"Percentage below dbuf_cache_max_bytes when dbuf eviction stops.");
ZFS_MODULE_PARAM(zfs_dbuf, dbuf_, metadata_cache_max_bytes, U64, ZMOD_RW,
"Maximum size in bytes of dbuf metadata cache.");
ZFS_MODULE_PARAM(zfs_dbuf, dbuf_, cache_shift, UINT, ZMOD_RW,
"Set size of dbuf cache to log2 fraction of arc size.");
ZFS_MODULE_PARAM(zfs_dbuf, dbuf_, metadata_cache_shift, UINT, ZMOD_RW,
"Set size of dbuf metadata cache to log2 fraction of arc size.");
ZFS_MODULE_PARAM(zfs_dbuf, dbuf_, mutex_cache_shift, UINT, ZMOD_RD,
"Set size of dbuf cache mutex array as log2 shift.");
diff --git a/sys/contrib/openzfs/module/zfs/mmp.c b/sys/contrib/openzfs/module/zfs/mmp.c
index ef0e01df390f..25eea0752941 100644
--- a/sys/contrib/openzfs/module/zfs/mmp.c
+++ b/sys/contrib/openzfs/module/zfs/mmp.c
@@ -1,747 +1,747 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2017 by Lawrence Livermore National Security, LLC.
*/
#include <sys/abd.h>
#include <sys/mmp.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/time.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
#include <sys/zfs_context.h>
#include <sys/callb.h>
/*
* Multi-Modifier Protection (MMP) attempts to prevent a user from importing
* or opening a pool on more than one host at a time. In particular, it
* prevents "zpool import -f" on a host from succeeding while the pool is
* already imported on another host. There are many other ways in which a
* device could be used by two hosts for different purposes at the same time
* resulting in pool damage. This implementation does not attempt to detect
* those cases.
*
* MMP operates by ensuring there are frequent visible changes on disk (a
* "heartbeat") at all times. And by altering the import process to check
* for these changes and failing the import when they are detected. This
* functionality is enabled by setting the 'multihost' pool property to on.
*
* Uberblocks written by the txg_sync thread always go into the first
* (N-MMP_BLOCKS_PER_LABEL) slots, the remaining slots are reserved for MMP.
* They are used to hold uberblocks which are exactly the same as the last
* synced uberblock except that the ub_timestamp and mmp_config are frequently
* updated. Like all other uberblocks, the slot is written with an embedded
* checksum, and slots with invalid checksums are ignored. This provides the
* "heartbeat", with no risk of overwriting good uberblocks that must be
* preserved, e.g. previous txgs and associated block pointers.
*
* Three optional fields are added to uberblock structure; ub_mmp_magic,
* ub_mmp_config, and ub_mmp_delay. The ub_mmp_magic value allows zfs to tell
* whether the other ub_mmp_* fields are valid. The ub_mmp_config field tells
* the importing host the settings of zfs_multihost_interval and
* zfs_multihost_fail_intervals on the host which last had (or currently has)
* the pool imported. These determine how long a host must wait to detect
* activity in the pool, before concluding the pool is not in use. The
* mmp_delay field is a decaying average of the amount of time between
* completion of successive MMP writes, in nanoseconds. It indicates whether
* MMP is enabled.
*
* During import an activity test may now be performed to determine if
* the pool is in use. The activity test is typically required if the
* ZPOOL_CONFIG_HOSTID does not match the system hostid, the pool state is
* POOL_STATE_ACTIVE, and the pool is not a root pool.
*
* The activity test finds the "best" uberblock (highest txg, timestamp, and, if
* ub_mmp_magic is valid, sequence number from ub_mmp_config). It then waits
* some time, and finds the "best" uberblock again. If any of the mentioned
* fields have different values in the newly read uberblock, the pool is in use
* by another host and the import fails. In order to assure the accuracy of the
* activity test, the default values result in an activity test duration of 20x
* the mmp write interval.
*
* The duration of the "zpool import" activity test depends on the information
* available in the "best" uberblock:
*
* 1) If uberblock was written by zfs-0.8 or newer and fail_intervals > 0:
* ub_mmp_config.fail_intervals * ub_mmp_config.multihost_interval * 2
*
* In this case, a weak guarantee is provided. Since the host which last had
* the pool imported will suspend the pool if no mmp writes land within
* fail_intervals * multihost_interval ms, the absence of writes during that
* time means either the pool is not imported, or it is imported but the pool
* is suspended and no further writes will occur.
*
* Note that resuming the suspended pool on the remote host would invalidate
* this guarantee, and so it is not allowed.
*
* The factor of 2 provides a conservative safety factor and derives from
* MMP_IMPORT_SAFETY_FACTOR;
*
* 2) If uberblock was written by zfs-0.8 or newer and fail_intervals == 0:
* (ub_mmp_config.multihost_interval + ub_mmp_delay) *
* zfs_multihost_import_intervals
*
* In this case no guarantee can provided. However, as long as some devices
* are healthy and connected, it is likely that at least one write will land
* within (multihost_interval + mmp_delay) because multihost_interval is
* enough time for a write to be attempted to each leaf vdev, and mmp_delay
* is enough for one to land, based on past delays. Multiplying by
* zfs_multihost_import_intervals provides a conservative safety factor.
*
* 3) If uberblock was written by zfs-0.7:
* (zfs_multihost_interval + ub_mmp_delay) * zfs_multihost_import_intervals
*
* The same logic as case #2 applies, but we do not know remote tunables.
*
* We use the local value for zfs_multihost_interval because the original MMP
* did not record this value in the uberblock.
*
* ub_mmp_delay >= (zfs_multihost_interval / leaves), so if the other host
* has a much larger zfs_multihost_interval set, ub_mmp_delay will reflect
* that. We will have waited enough time for zfs_multihost_import_intervals
* writes to be issued and all but one to land.
*
* single device pool example delays
*
* import_delay = (1 + 1) * 20 = 40s #defaults, no I/O delay
* import_delay = (1 + 10) * 20 = 220s #defaults, 10s I/O delay
* import_delay = (10 + 10) * 20 = 400s #10s multihost_interval,
* no I/O delay
* 100 device pool example delays
*
* import_delay = (1 + .01) * 20 = 20s #defaults, no I/O delay
* import_delay = (1 + 10) * 20 = 220s #defaults, 10s I/O delay
* import_delay = (10 + .1) * 20 = 202s #10s multihost_interval,
* no I/O delay
*
* 4) Otherwise, this uberblock was written by a pre-MMP zfs:
* zfs_multihost_import_intervals * zfs_multihost_interval
*
* In this case local tunables are used. By default this product = 10s, long
* enough for a pool with any activity at all to write at least one
* uberblock. No guarantee can be provided.
*
* Additionally, the duration is then extended by a random 25% to attempt to to
* detect simultaneous imports. For example, if both partner hosts are rebooted
* at the same time and automatically attempt to import the pool.
*/
/*
* Used to control the frequency of mmp writes which are performed when the
* 'multihost' pool property is on. This is one factor used to determine the
* length of the activity check during import.
*
* On average an mmp write will be issued for each leaf vdev every
* zfs_multihost_interval milliseconds. In practice, the observed period can
* vary with the I/O load and this observed value is the ub_mmp_delay which is
* stored in the uberblock. The minimum allowed value is 100 ms.
*/
uint64_t zfs_multihost_interval = MMP_DEFAULT_INTERVAL;
/*
* Used to control the duration of the activity test on import. Smaller values
* of zfs_multihost_import_intervals will reduce the import time but increase
* the risk of failing to detect an active pool. The total activity check time
* is never allowed to drop below one second. A value of 0 is ignored and
* treated as if it was set to 1.
*/
uint_t zfs_multihost_import_intervals = MMP_DEFAULT_IMPORT_INTERVALS;
/*
* Controls the behavior of the pool when mmp write failures or delays are
* detected.
*
* When zfs_multihost_fail_intervals = 0, mmp write failures or delays are
* ignored. The failures will still be reported to the ZED which depending on
* its configuration may take action such as suspending the pool or taking a
* device offline.
*
* When zfs_multihost_fail_intervals > 0, the pool will be suspended if
* zfs_multihost_fail_intervals * zfs_multihost_interval milliseconds pass
* without a successful mmp write. This guarantees the activity test will see
* mmp writes if the pool is imported. A value of 1 is ignored and treated as
* if it was set to 2, because a single leaf vdev pool will issue a write once
* per multihost_interval and thus any variation in latency would cause the
* pool to be suspended.
*/
uint_t zfs_multihost_fail_intervals = MMP_DEFAULT_FAIL_INTERVALS;
static const void *const mmp_tag = "mmp_write_uberblock";
static __attribute__((noreturn)) void mmp_thread(void *arg);
void
mmp_init(spa_t *spa)
{
mmp_thread_t *mmp = &spa->spa_mmp;
mutex_init(&mmp->mmp_thread_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&mmp->mmp_thread_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&mmp->mmp_io_lock, NULL, MUTEX_DEFAULT, NULL);
mmp->mmp_kstat_id = 1;
}
void
mmp_fini(spa_t *spa)
{
mmp_thread_t *mmp = &spa->spa_mmp;
mutex_destroy(&mmp->mmp_thread_lock);
cv_destroy(&mmp->mmp_thread_cv);
mutex_destroy(&mmp->mmp_io_lock);
}
static void
mmp_thread_enter(mmp_thread_t *mmp, callb_cpr_t *cpr)
{
CALLB_CPR_INIT(cpr, &mmp->mmp_thread_lock, callb_generic_cpr, FTAG);
mutex_enter(&mmp->mmp_thread_lock);
}
static void
mmp_thread_exit(mmp_thread_t *mmp, kthread_t **mpp, callb_cpr_t *cpr)
{
ASSERT(*mpp != NULL);
*mpp = NULL;
cv_broadcast(&mmp->mmp_thread_cv);
CALLB_CPR_EXIT(cpr); /* drops &mmp->mmp_thread_lock */
}
void
mmp_thread_start(spa_t *spa)
{
mmp_thread_t *mmp = &spa->spa_mmp;
if (spa_writeable(spa)) {
mutex_enter(&mmp->mmp_thread_lock);
if (!mmp->mmp_thread) {
mmp->mmp_thread = thread_create(NULL, 0, mmp_thread,
spa, 0, &p0, TS_RUN, defclsyspri);
zfs_dbgmsg("MMP thread started pool '%s' "
"gethrtime %llu", spa_name(spa), gethrtime());
}
mutex_exit(&mmp->mmp_thread_lock);
}
}
void
mmp_thread_stop(spa_t *spa)
{
mmp_thread_t *mmp = &spa->spa_mmp;
mutex_enter(&mmp->mmp_thread_lock);
mmp->mmp_thread_exiting = 1;
cv_broadcast(&mmp->mmp_thread_cv);
while (mmp->mmp_thread) {
cv_wait(&mmp->mmp_thread_cv, &mmp->mmp_thread_lock);
}
mutex_exit(&mmp->mmp_thread_lock);
zfs_dbgmsg("MMP thread stopped pool '%s' gethrtime %llu",
spa_name(spa), gethrtime());
ASSERT(mmp->mmp_thread == NULL);
mmp->mmp_thread_exiting = 0;
}
typedef enum mmp_vdev_state_flag {
MMP_FAIL_NOT_WRITABLE = (1 << 0),
MMP_FAIL_WRITE_PENDING = (1 << 1),
} mmp_vdev_state_flag_t;
/*
* Find a leaf vdev to write an MMP block to. It must not have an outstanding
* mmp write (if so a new write will also likely block). If there is no usable
* leaf, a nonzero error value is returned. The error value returned is a bit
* field.
*
* MMP_FAIL_WRITE_PENDING One or more leaf vdevs are writeable, but have an
* outstanding MMP write.
* MMP_FAIL_NOT_WRITABLE One or more leaf vdevs are not writeable.
*/
static int
mmp_next_leaf(spa_t *spa)
{
vdev_t *leaf;
vdev_t *starting_leaf;
int fail_mask = 0;
ASSERT(MUTEX_HELD(&spa->spa_mmp.mmp_io_lock));
ASSERT(spa_config_held(spa, SCL_STATE, RW_READER));
ASSERT(list_link_active(&spa->spa_leaf_list.list_head) == B_TRUE);
ASSERT(!list_is_empty(&spa->spa_leaf_list));
if (spa->spa_mmp.mmp_leaf_last_gen != spa->spa_leaf_list_gen) {
spa->spa_mmp.mmp_last_leaf = list_head(&spa->spa_leaf_list);
spa->spa_mmp.mmp_leaf_last_gen = spa->spa_leaf_list_gen;
}
leaf = spa->spa_mmp.mmp_last_leaf;
if (leaf == NULL)
leaf = list_head(&spa->spa_leaf_list);
starting_leaf = leaf;
do {
leaf = list_next(&spa->spa_leaf_list, leaf);
if (leaf == NULL) {
leaf = list_head(&spa->spa_leaf_list);
ASSERT3P(leaf, !=, NULL);
}
/*
* We skip unwritable, offline, detached, and dRAID spare
* devices as they are either not legal targets or the write
* may fail or not be seen by other hosts. Skipped dRAID
* spares can never be written so the fail mask is not set.
*/
if (!vdev_writeable(leaf) || leaf->vdev_offline ||
leaf->vdev_detached) {
fail_mask |= MMP_FAIL_NOT_WRITABLE;
} else if (leaf->vdev_ops == &vdev_draid_spare_ops) {
continue;
} else if (leaf->vdev_mmp_pending != 0) {
fail_mask |= MMP_FAIL_WRITE_PENDING;
} else {
spa->spa_mmp.mmp_last_leaf = leaf;
return (0);
}
} while (leaf != starting_leaf);
ASSERT(fail_mask);
return (fail_mask);
}
/*
* MMP writes are issued on a fixed schedule, but may complete at variable,
* much longer, intervals. The mmp_delay captures long periods between
* successful writes for any reason, including disk latency, scheduling delays,
* etc.
*
* The mmp_delay is usually calculated as a decaying average, but if the latest
* delay is higher we do not average it, so that we do not hide sudden spikes
* which the importing host must wait for.
*
* If writes are occurring frequently, such as due to a high rate of txg syncs,
* the mmp_delay could become very small. Since those short delays depend on
* activity we cannot count on, we never allow mmp_delay to get lower than rate
* expected if only mmp_thread writes occur.
*
* If an mmp write was skipped or fails, and we have already waited longer than
* mmp_delay, we need to update it so the next write reflects the longer delay.
*
* Do not set mmp_delay if the multihost property is not on, so as not to
* trigger an activity check on import.
*/
static void
mmp_delay_update(spa_t *spa, boolean_t write_completed)
{
mmp_thread_t *mts = &spa->spa_mmp;
hrtime_t delay = gethrtime() - mts->mmp_last_write;
ASSERT(MUTEX_HELD(&mts->mmp_io_lock));
if (spa_multihost(spa) == B_FALSE) {
mts->mmp_delay = 0;
return;
}
if (delay > mts->mmp_delay)
mts->mmp_delay = delay;
if (write_completed == B_FALSE)
return;
mts->mmp_last_write = gethrtime();
/*
* strictly less than, in case delay was changed above.
*/
if (delay < mts->mmp_delay) {
hrtime_t min_delay =
MSEC2NSEC(MMP_INTERVAL_OK(zfs_multihost_interval)) /
MAX(1, vdev_count_leaves(spa));
mts->mmp_delay = MAX(((delay + mts->mmp_delay * 127) / 128),
min_delay);
}
}
static void
mmp_write_done(zio_t *zio)
{
spa_t *spa = zio->io_spa;
vdev_t *vd = zio->io_vd;
mmp_thread_t *mts = zio->io_private;
mutex_enter(&mts->mmp_io_lock);
uint64_t mmp_kstat_id = vd->vdev_mmp_kstat_id;
hrtime_t mmp_write_duration = gethrtime() - vd->vdev_mmp_pending;
mmp_delay_update(spa, (zio->io_error == 0));
vd->vdev_mmp_pending = 0;
vd->vdev_mmp_kstat_id = 0;
mutex_exit(&mts->mmp_io_lock);
spa_config_exit(spa, SCL_STATE, mmp_tag);
spa_mmp_history_set(spa, mmp_kstat_id, zio->io_error,
mmp_write_duration);
abd_free(zio->io_abd);
}
/*
* When the uberblock on-disk is updated by a spa_sync,
* creating a new "best" uberblock, update the one stored
* in the mmp thread state, used for mmp writes.
*/
void
mmp_update_uberblock(spa_t *spa, uberblock_t *ub)
{
mmp_thread_t *mmp = &spa->spa_mmp;
mutex_enter(&mmp->mmp_io_lock);
mmp->mmp_ub = *ub;
mmp->mmp_seq = 1;
mmp->mmp_ub.ub_timestamp = gethrestime_sec();
mmp_delay_update(spa, B_TRUE);
mutex_exit(&mmp->mmp_io_lock);
}
/*
* Choose a random vdev, label, and MMP block, and write over it
* with a copy of the last-synced uberblock, whose timestamp
* has been updated to reflect that the pool is in use.
*/
static void
mmp_write_uberblock(spa_t *spa)
{
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
mmp_thread_t *mmp = &spa->spa_mmp;
uberblock_t *ub;
vdev_t *vd = NULL;
int label, error;
uint64_t offset;
hrtime_t lock_acquire_time = gethrtime();
- spa_config_enter(spa, SCL_STATE, mmp_tag, RW_READER);
+ spa_config_enter_mmp(spa, SCL_STATE, mmp_tag, RW_READER);
lock_acquire_time = gethrtime() - lock_acquire_time;
if (lock_acquire_time > (MSEC2NSEC(MMP_MIN_INTERVAL) / 10))
zfs_dbgmsg("MMP SCL_STATE acquisition pool '%s' took %llu ns "
"gethrtime %llu", spa_name(spa), lock_acquire_time,
gethrtime());
mutex_enter(&mmp->mmp_io_lock);
error = mmp_next_leaf(spa);
/*
* spa_mmp_history has two types of entries:
* Issued MMP write: records time issued, error status, etc.
* Skipped MMP write: an MMP write could not be issued because no
* suitable leaf vdev was available. See comment above struct
* spa_mmp_history for details.
*/
if (error) {
mmp_delay_update(spa, B_FALSE);
if (mmp->mmp_skip_error == error) {
spa_mmp_history_set_skip(spa, mmp->mmp_kstat_id - 1);
} else {
mmp->mmp_skip_error = error;
spa_mmp_history_add(spa, mmp->mmp_ub.ub_txg,
gethrestime_sec(), mmp->mmp_delay, NULL, 0,
mmp->mmp_kstat_id++, error);
zfs_dbgmsg("MMP error choosing leaf pool '%s' "
"gethrtime %llu fail_mask %#x", spa_name(spa),
gethrtime(), error);
}
mutex_exit(&mmp->mmp_io_lock);
spa_config_exit(spa, SCL_STATE, mmp_tag);
return;
}
vd = spa->spa_mmp.mmp_last_leaf;
if (mmp->mmp_skip_error != 0) {
mmp->mmp_skip_error = 0;
zfs_dbgmsg("MMP write after skipping due to unavailable "
"leaves, pool '%s' gethrtime %llu leaf %llu",
spa_name(spa), (u_longlong_t)gethrtime(),
(u_longlong_t)vd->vdev_guid);
}
if (mmp->mmp_zio_root == NULL)
mmp->mmp_zio_root = zio_root(spa, NULL, NULL,
flags | ZIO_FLAG_GODFATHER);
if (mmp->mmp_ub.ub_timestamp != gethrestime_sec()) {
/*
* Want to reset mmp_seq when timestamp advances because after
* an mmp_seq wrap new values will not be chosen by
* uberblock_compare() as the "best".
*/
mmp->mmp_ub.ub_timestamp = gethrestime_sec();
mmp->mmp_seq = 1;
}
ub = &mmp->mmp_ub;
ub->ub_mmp_magic = MMP_MAGIC;
ub->ub_mmp_delay = mmp->mmp_delay;
ub->ub_mmp_config = MMP_SEQ_SET(mmp->mmp_seq) |
MMP_INTERVAL_SET(MMP_INTERVAL_OK(zfs_multihost_interval)) |
MMP_FAIL_INT_SET(MMP_FAIL_INTVS_OK(
zfs_multihost_fail_intervals));
vd->vdev_mmp_pending = gethrtime();
vd->vdev_mmp_kstat_id = mmp->mmp_kstat_id;
zio_t *zio = zio_null(mmp->mmp_zio_root, spa, NULL, NULL, NULL, flags);
abd_t *ub_abd = abd_alloc_for_io(VDEV_UBERBLOCK_SIZE(vd), B_TRUE);
abd_zero(ub_abd, VDEV_UBERBLOCK_SIZE(vd));
abd_copy_from_buf(ub_abd, ub, sizeof (uberblock_t));
mmp->mmp_seq++;
mmp->mmp_kstat_id++;
mutex_exit(&mmp->mmp_io_lock);
offset = VDEV_UBERBLOCK_OFFSET(vd, VDEV_UBERBLOCK_COUNT(vd) -
MMP_BLOCKS_PER_LABEL + random_in_range(MMP_BLOCKS_PER_LABEL));
label = random_in_range(VDEV_LABELS);
vdev_label_write(zio, vd, label, ub_abd, offset,
VDEV_UBERBLOCK_SIZE(vd), mmp_write_done, mmp,
flags | ZIO_FLAG_DONT_PROPAGATE);
(void) spa_mmp_history_add(spa, ub->ub_txg, ub->ub_timestamp,
ub->ub_mmp_delay, vd, label, vd->vdev_mmp_kstat_id, 0);
zio_nowait(zio);
}
static __attribute__((noreturn)) void
mmp_thread(void *arg)
{
spa_t *spa = (spa_t *)arg;
mmp_thread_t *mmp = &spa->spa_mmp;
boolean_t suspended = spa_suspended(spa);
boolean_t multihost = spa_multihost(spa);
uint64_t mmp_interval = MSEC2NSEC(MMP_INTERVAL_OK(
zfs_multihost_interval));
uint32_t mmp_fail_intervals = MMP_FAIL_INTVS_OK(
zfs_multihost_fail_intervals);
hrtime_t mmp_fail_ns = mmp_fail_intervals * mmp_interval;
boolean_t last_spa_suspended;
boolean_t last_spa_multihost;
uint64_t last_mmp_interval;
uint32_t last_mmp_fail_intervals;
hrtime_t last_mmp_fail_ns;
callb_cpr_t cpr;
int skip_wait = 0;
mmp_thread_enter(mmp, &cpr);
/*
* There have been no MMP writes yet. Setting mmp_last_write here gives
* us one mmp_fail_ns period, which is consistent with the activity
* check duration, to try to land an MMP write before MMP suspends the
* pool (if so configured).
*/
mutex_enter(&mmp->mmp_io_lock);
mmp->mmp_last_write = gethrtime();
mmp->mmp_delay = MSEC2NSEC(MMP_INTERVAL_OK(zfs_multihost_interval));
mutex_exit(&mmp->mmp_io_lock);
while (!mmp->mmp_thread_exiting) {
hrtime_t next_time = gethrtime() +
MSEC2NSEC(MMP_DEFAULT_INTERVAL);
int leaves = MAX(vdev_count_leaves(spa), 1);
/* Detect changes in tunables or state */
last_spa_suspended = suspended;
last_spa_multihost = multihost;
suspended = spa_suspended(spa);
multihost = spa_multihost(spa);
last_mmp_interval = mmp_interval;
last_mmp_fail_intervals = mmp_fail_intervals;
last_mmp_fail_ns = mmp_fail_ns;
mmp_interval = MSEC2NSEC(MMP_INTERVAL_OK(
zfs_multihost_interval));
mmp_fail_intervals = MMP_FAIL_INTVS_OK(
zfs_multihost_fail_intervals);
/* Smooth so pool is not suspended when reducing tunables */
if (mmp_fail_intervals * mmp_interval < mmp_fail_ns) {
mmp_fail_ns = (mmp_fail_ns * 31 +
mmp_fail_intervals * mmp_interval) / 32;
} else {
mmp_fail_ns = mmp_fail_intervals *
mmp_interval;
}
if (mmp_interval != last_mmp_interval ||
mmp_fail_intervals != last_mmp_fail_intervals) {
/*
* We want other hosts to see new tunables as quickly as
* possible. Write out at higher frequency than usual.
*/
skip_wait += leaves;
}
if (multihost)
next_time = gethrtime() + mmp_interval / leaves;
if (mmp_fail_ns != last_mmp_fail_ns) {
zfs_dbgmsg("MMP interval change pool '%s' "
"gethrtime %llu last_mmp_interval %llu "
"mmp_interval %llu last_mmp_fail_intervals %u "
"mmp_fail_intervals %u mmp_fail_ns %llu "
"skip_wait %d leaves %d next_time %llu",
spa_name(spa), (u_longlong_t)gethrtime(),
(u_longlong_t)last_mmp_interval,
(u_longlong_t)mmp_interval, last_mmp_fail_intervals,
mmp_fail_intervals, (u_longlong_t)mmp_fail_ns,
skip_wait, leaves, (u_longlong_t)next_time);
}
/*
* MMP off => on, or suspended => !suspended:
* No writes occurred recently. Update mmp_last_write to give
* us some time to try.
*/
if ((!last_spa_multihost && multihost) ||
(last_spa_suspended && !suspended)) {
zfs_dbgmsg("MMP state change pool '%s': gethrtime %llu "
"last_spa_multihost %u multihost %u "
"last_spa_suspended %u suspended %u",
spa_name(spa), (u_longlong_t)gethrtime(),
last_spa_multihost, multihost, last_spa_suspended,
suspended);
mutex_enter(&mmp->mmp_io_lock);
mmp->mmp_last_write = gethrtime();
mmp->mmp_delay = mmp_interval;
mutex_exit(&mmp->mmp_io_lock);
}
/*
* MMP on => off:
* mmp_delay == 0 tells importing node to skip activity check.
*/
if (last_spa_multihost && !multihost) {
mutex_enter(&mmp->mmp_io_lock);
mmp->mmp_delay = 0;
mutex_exit(&mmp->mmp_io_lock);
}
/*
* Suspend the pool if no MMP write has succeeded in over
* mmp_interval * mmp_fail_intervals nanoseconds.
*/
if (multihost && !suspended && mmp_fail_intervals &&
(gethrtime() - mmp->mmp_last_write) > mmp_fail_ns) {
zfs_dbgmsg("MMP suspending pool '%s': gethrtime %llu "
"mmp_last_write %llu mmp_interval %llu "
"mmp_fail_intervals %llu mmp_fail_ns %llu",
spa_name(spa), (u_longlong_t)gethrtime(),
(u_longlong_t)mmp->mmp_last_write,
(u_longlong_t)mmp_interval,
(u_longlong_t)mmp_fail_intervals,
(u_longlong_t)mmp_fail_ns);
cmn_err(CE_WARN, "MMP writes to pool '%s' have not "
"succeeded in over %llu ms; suspending pool. "
"Hrtime %llu",
spa_name(spa),
NSEC2MSEC(gethrtime() - mmp->mmp_last_write),
gethrtime());
zio_suspend(spa, NULL, ZIO_SUSPEND_MMP);
}
if (multihost && !suspended)
mmp_write_uberblock(spa);
if (skip_wait > 0) {
next_time = gethrtime() + MSEC2NSEC(MMP_MIN_INTERVAL) /
leaves;
skip_wait--;
}
CALLB_CPR_SAFE_BEGIN(&cpr);
(void) cv_timedwait_idle_hires(&mmp->mmp_thread_cv,
&mmp->mmp_thread_lock, next_time, USEC2NSEC(100),
CALLOUT_FLAG_ABSOLUTE);
CALLB_CPR_SAFE_END(&cpr, &mmp->mmp_thread_lock);
}
/* Outstanding writes are allowed to complete. */
zio_wait(mmp->mmp_zio_root);
mmp->mmp_zio_root = NULL;
mmp_thread_exit(mmp, &mmp->mmp_thread, &cpr);
thread_exit();
}
/*
* Signal the MMP thread to wake it, when it is sleeping on
* its cv. Used when some module parameter has changed and
* we want the thread to know about it.
* Only signal if the pool is active and mmp thread is
* running, otherwise there is no thread to wake.
*/
static void
mmp_signal_thread(spa_t *spa)
{
mmp_thread_t *mmp = &spa->spa_mmp;
mutex_enter(&mmp->mmp_thread_lock);
if (mmp->mmp_thread)
cv_broadcast(&mmp->mmp_thread_cv);
mutex_exit(&mmp->mmp_thread_lock);
}
void
mmp_signal_all_threads(void)
{
spa_t *spa = NULL;
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(spa))) {
if (spa->spa_state == POOL_STATE_ACTIVE)
mmp_signal_thread(spa);
}
mutex_exit(&spa_namespace_lock);
}
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM_CALL(zfs_multihost, zfs_multihost_, interval,
param_set_multihost_interval, spl_param_get_u64, ZMOD_RW,
"Milliseconds between mmp writes to each leaf");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs_multihost, zfs_multihost_, fail_intervals, UINT, ZMOD_RW,
"Max allowed period without a successful mmp write");
ZFS_MODULE_PARAM(zfs_multihost, zfs_multihost_, import_intervals, UINT, ZMOD_RW,
"Number of zfs_multihost_interval periods to wait for activity");
diff --git a/sys/contrib/openzfs/module/zfs/spa.c b/sys/contrib/openzfs/module/zfs/spa.c
index 8c6afb3aac44..dd4a442d97a1 100644
--- a/sys/contrib/openzfs/module/zfs/spa.c
+++ b/sys/contrib/openzfs/module/zfs/spa.c
@@ -1,10143 +1,10155 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2020 by Delphix. All rights reserved.
* Copyright (c) 2018, Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
* Copyright 2013 Saso Kiselkov. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2016 Toomas Soome <tsoome@me.com>
* Copyright (c) 2016 Actifio, Inc. All rights reserved.
* Copyright 2018 Joyent, Inc.
* Copyright (c) 2017, 2019, Datto Inc. All rights reserved.
* Copyright 2017 Joyent, Inc.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2021, Colm Buckley <colm@tuatha.org>
*/
/*
* SPA: Storage Pool Allocator
*
* This file contains all the routines used when modifying on-disk SPA state.
* This includes opening, importing, destroying, exporting a pool, and syncing a
* pool.
*/
#include <sys/zfs_context.h>
#include <sys/fm/fs/zfs.h>
#include <sys/spa_impl.h>
#include <sys/zio.h>
#include <sys/zio_checksum.h>
#include <sys/dmu.h>
#include <sys/dmu_tx.h>
#include <sys/zap.h>
#include <sys/zil.h>
#include <sys/brt.h>
#include <sys/ddt.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_removal.h>
#include <sys/vdev_indirect_mapping.h>
#include <sys/vdev_indirect_births.h>
#include <sys/vdev_initialize.h>
#include <sys/vdev_rebuild.h>
#include <sys/vdev_trim.h>
#include <sys/vdev_disk.h>
#include <sys/vdev_draid.h>
#include <sys/metaslab.h>
#include <sys/metaslab_impl.h>
#include <sys/mmp.h>
#include <sys/uberblock_impl.h>
#include <sys/txg.h>
#include <sys/avl.h>
#include <sys/bpobj.h>
#include <sys/dmu_traverse.h>
#include <sys/dmu_objset.h>
#include <sys/unique.h>
#include <sys/dsl_pool.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_synctask.h>
#include <sys/fs/zfs.h>
#include <sys/arc.h>
#include <sys/callb.h>
#include <sys/systeminfo.h>
#include <sys/zfs_ioctl.h>
#include <sys/dsl_scan.h>
#include <sys/zfeature.h>
#include <sys/dsl_destroy.h>
#include <sys/zvol.h>
#ifdef _KERNEL
#include <sys/fm/protocol.h>
#include <sys/fm/util.h>
#include <sys/callb.h>
#include <sys/zone.h>
#include <sys/vmsystm.h>
#endif /* _KERNEL */
#include "zfs_prop.h"
#include "zfs_comutil.h"
/*
* The interval, in seconds, at which failed configuration cache file writes
* should be retried.
*/
int zfs_ccw_retry_interval = 300;
typedef enum zti_modes {
ZTI_MODE_FIXED, /* value is # of threads (min 1) */
ZTI_MODE_BATCH, /* cpu-intensive; value is ignored */
ZTI_MODE_SCALE, /* Taskqs scale with CPUs. */
ZTI_MODE_NULL, /* don't create a taskq */
ZTI_NMODES
} zti_modes_t;
#define ZTI_P(n, q) { ZTI_MODE_FIXED, (n), (q) }
#define ZTI_PCT(n) { ZTI_MODE_ONLINE_PERCENT, (n), 1 }
#define ZTI_BATCH { ZTI_MODE_BATCH, 0, 1 }
#define ZTI_SCALE { ZTI_MODE_SCALE, 0, 1 }
#define ZTI_NULL { ZTI_MODE_NULL, 0, 0 }
#define ZTI_N(n) ZTI_P(n, 1)
#define ZTI_ONE ZTI_N(1)
typedef struct zio_taskq_info {
zti_modes_t zti_mode;
uint_t zti_value;
uint_t zti_count;
} zio_taskq_info_t;
static const char *const zio_taskq_types[ZIO_TASKQ_TYPES] = {
"iss", "iss_h", "int", "int_h"
};
/*
* This table defines the taskq settings for each ZFS I/O type. When
* initializing a pool, we use this table to create an appropriately sized
* taskq. Some operations are low volume and therefore have a small, static
* number of threads assigned to their taskqs using the ZTI_N(#) or ZTI_ONE
* macros. Other operations process a large amount of data; the ZTI_BATCH
* macro causes us to create a taskq oriented for throughput. Some operations
* are so high frequency and short-lived that the taskq itself can become a
* point of lock contention. The ZTI_P(#, #) macro indicates that we need an
* additional degree of parallelism specified by the number of threads per-
* taskq and the number of taskqs; when dispatching an event in this case, the
* particular taskq is chosen at random. ZTI_SCALE is similar to ZTI_BATCH,
* but with number of taskqs also scaling with number of CPUs.
*
* The different taskq priorities are to handle the different contexts (issue
* and interrupt) and then to reserve threads for ZIO_PRIORITY_NOW I/Os that
* need to be handled with minimum delay.
*/
static const zio_taskq_info_t zio_taskqs[ZIO_TYPES][ZIO_TASKQ_TYPES] = {
/* ISSUE ISSUE_HIGH INTR INTR_HIGH */
{ ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* NULL */
{ ZTI_N(8), ZTI_NULL, ZTI_SCALE, ZTI_NULL }, /* READ */
{ ZTI_BATCH, ZTI_N(5), ZTI_SCALE, ZTI_N(5) }, /* WRITE */
{ ZTI_SCALE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* FREE */
{ ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* CLAIM */
{ ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* IOCTL */
{ ZTI_N(4), ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* TRIM */
};
static void spa_sync_version(void *arg, dmu_tx_t *tx);
static void spa_sync_props(void *arg, dmu_tx_t *tx);
static boolean_t spa_has_active_shared_spare(spa_t *spa);
static int spa_load_impl(spa_t *spa, spa_import_type_t type,
const char **ereport);
static void spa_vdev_resilver_done(spa_t *spa);
static uint_t zio_taskq_batch_pct = 80; /* 1 thread per cpu in pset */
static uint_t zio_taskq_batch_tpq; /* threads per taskq */
static const boolean_t zio_taskq_sysdc = B_TRUE; /* use SDC scheduling class */
static const uint_t zio_taskq_basedc = 80; /* base duty cycle */
static const boolean_t spa_create_process = B_TRUE; /* no process => no sysdc */
/*
* Report any spa_load_verify errors found, but do not fail spa_load.
* This is used by zdb to analyze non-idle pools.
*/
boolean_t spa_load_verify_dryrun = B_FALSE;
/*
* Allow read spacemaps in case of readonly import (spa_mode == SPA_MODE_READ).
* This is used by zdb for spacemaps verification.
*/
boolean_t spa_mode_readable_spacemaps = B_FALSE;
/*
* This (illegal) pool name is used when temporarily importing a spa_t in order
* to get the vdev stats associated with the imported devices.
*/
#define TRYIMPORT_NAME "$import"
/*
* For debugging purposes: print out vdev tree during pool import.
*/
static int spa_load_print_vdev_tree = B_FALSE;
/*
* A non-zero value for zfs_max_missing_tvds means that we allow importing
* pools with missing top-level vdevs. This is strictly intended for advanced
* pool recovery cases since missing data is almost inevitable. Pools with
* missing devices can only be imported read-only for safety reasons, and their
* fail-mode will be automatically set to "continue".
*
* With 1 missing vdev we should be able to import the pool and mount all
* datasets. User data that was not modified after the missing device has been
* added should be recoverable. This means that snapshots created prior to the
* addition of that device should be completely intact.
*
* With 2 missing vdevs, some datasets may fail to mount since there are
* dataset statistics that are stored as regular metadata. Some data might be
* recoverable if those vdevs were added recently.
*
* With 3 or more missing vdevs, the pool is severely damaged and MOS entries
* may be missing entirely. Chances of data recovery are very low. Note that
* there are also risks of performing an inadvertent rewind as we might be
* missing all the vdevs with the latest uberblocks.
*/
uint64_t zfs_max_missing_tvds = 0;
/*
* The parameters below are similar to zfs_max_missing_tvds but are only
* intended for a preliminary open of the pool with an untrusted config which
* might be incomplete or out-dated.
*
* We are more tolerant for pools opened from a cachefile since we could have
* an out-dated cachefile where a device removal was not registered.
* We could have set the limit arbitrarily high but in the case where devices
* are really missing we would want to return the proper error codes; we chose
* SPA_DVAS_PER_BP - 1 so that some copies of the MOS would still be available
* and we get a chance to retrieve the trusted config.
*/
uint64_t zfs_max_missing_tvds_cachefile = SPA_DVAS_PER_BP - 1;
/*
* In the case where config was assembled by scanning device paths (/dev/dsks
* by default) we are less tolerant since all the existing devices should have
* been detected and we want spa_load to return the right error codes.
*/
uint64_t zfs_max_missing_tvds_scan = 0;
/*
* Debugging aid that pauses spa_sync() towards the end.
*/
static const boolean_t zfs_pause_spa_sync = B_FALSE;
/*
* Variables to indicate the livelist condense zthr func should wait at certain
* points for the livelist to be removed - used to test condense/destroy races
*/
static int zfs_livelist_condense_zthr_pause = 0;
static int zfs_livelist_condense_sync_pause = 0;
/*
* Variables to track whether or not condense cancellation has been
* triggered in testing.
*/
static int zfs_livelist_condense_sync_cancel = 0;
static int zfs_livelist_condense_zthr_cancel = 0;
/*
* Variable to track whether or not extra ALLOC blkptrs were added to a
* livelist entry while it was being condensed (caused by the way we track
* remapped blkptrs in dbuf_remap_impl)
*/
static int zfs_livelist_condense_new_alloc = 0;
/*
* ==========================================================================
* SPA properties routines
* ==========================================================================
*/
/*
* Add a (source=src, propname=propval) list to an nvlist.
*/
static void
spa_prop_add_list(nvlist_t *nvl, zpool_prop_t prop, const char *strval,
uint64_t intval, zprop_source_t src)
{
const char *propname = zpool_prop_to_name(prop);
nvlist_t *propval;
propval = fnvlist_alloc();
fnvlist_add_uint64(propval, ZPROP_SOURCE, src);
if (strval != NULL)
fnvlist_add_string(propval, ZPROP_VALUE, strval);
else
fnvlist_add_uint64(propval, ZPROP_VALUE, intval);
fnvlist_add_nvlist(nvl, propname, propval);
nvlist_free(propval);
}
/*
* Add a user property (source=src, propname=propval) to an nvlist.
*/
static void
spa_prop_add_user(nvlist_t *nvl, const char *propname, char *strval,
zprop_source_t src)
{
nvlist_t *propval;
VERIFY(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP) == 0);
VERIFY(nvlist_add_uint64(propval, ZPROP_SOURCE, src) == 0);
VERIFY(nvlist_add_string(propval, ZPROP_VALUE, strval) == 0);
VERIFY(nvlist_add_nvlist(nvl, propname, propval) == 0);
nvlist_free(propval);
}
/*
* Get property values from the spa configuration.
*/
static void
spa_prop_get_config(spa_t *spa, nvlist_t **nvp)
{
vdev_t *rvd = spa->spa_root_vdev;
dsl_pool_t *pool = spa->spa_dsl_pool;
uint64_t size, alloc, cap, version;
const zprop_source_t src = ZPROP_SRC_NONE;
spa_config_dirent_t *dp;
metaslab_class_t *mc = spa_normal_class(spa);
ASSERT(MUTEX_HELD(&spa->spa_props_lock));
if (rvd != NULL) {
alloc = metaslab_class_get_alloc(mc);
alloc += metaslab_class_get_alloc(spa_special_class(spa));
alloc += metaslab_class_get_alloc(spa_dedup_class(spa));
alloc += metaslab_class_get_alloc(spa_embedded_log_class(spa));
size = metaslab_class_get_space(mc);
size += metaslab_class_get_space(spa_special_class(spa));
size += metaslab_class_get_space(spa_dedup_class(spa));
size += metaslab_class_get_space(spa_embedded_log_class(spa));
spa_prop_add_list(*nvp, ZPOOL_PROP_NAME, spa_name(spa), 0, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_SIZE, NULL, size, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_ALLOCATED, NULL, alloc, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_FREE, NULL,
size - alloc, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_CHECKPOINT, NULL,
spa->spa_checkpoint_info.sci_dspace, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_FRAGMENTATION, NULL,
metaslab_class_fragmentation(mc), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_EXPANDSZ, NULL,
metaslab_class_expandable_space(mc), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_READONLY, NULL,
(spa_mode(spa) == SPA_MODE_READ), src);
cap = (size == 0) ? 0 : (alloc * 100 / size);
spa_prop_add_list(*nvp, ZPOOL_PROP_CAPACITY, NULL, cap, src);
spa_prop_add_list(*nvp, ZPOOL_PROP_DEDUPRATIO, NULL,
ddt_get_pool_dedup_ratio(spa), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONEUSED, NULL,
brt_get_used(spa), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONESAVED, NULL,
brt_get_saved(spa), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONERATIO, NULL,
brt_get_ratio(spa), src);
spa_prop_add_list(*nvp, ZPOOL_PROP_HEALTH, NULL,
rvd->vdev_state, src);
version = spa_version(spa);
if (version == zpool_prop_default_numeric(ZPOOL_PROP_VERSION)) {
spa_prop_add_list(*nvp, ZPOOL_PROP_VERSION, NULL,
version, ZPROP_SRC_DEFAULT);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_VERSION, NULL,
version, ZPROP_SRC_LOCAL);
}
spa_prop_add_list(*nvp, ZPOOL_PROP_LOAD_GUID,
NULL, spa_load_guid(spa), src);
}
if (pool != NULL) {
/*
* The $FREE directory was introduced in SPA_VERSION_DEADLISTS,
* when opening pools before this version freedir will be NULL.
*/
if (pool->dp_free_dir != NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_FREEING, NULL,
dsl_dir_phys(pool->dp_free_dir)->dd_used_bytes,
src);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_FREEING,
NULL, 0, src);
}
if (pool->dp_leak_dir != NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_LEAKED, NULL,
dsl_dir_phys(pool->dp_leak_dir)->dd_used_bytes,
src);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_LEAKED,
NULL, 0, src);
}
}
spa_prop_add_list(*nvp, ZPOOL_PROP_GUID, NULL, spa_guid(spa), src);
if (spa->spa_comment != NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_COMMENT, spa->spa_comment,
0, ZPROP_SRC_LOCAL);
}
if (spa->spa_compatibility != NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_COMPATIBILITY,
spa->spa_compatibility, 0, ZPROP_SRC_LOCAL);
}
if (spa->spa_root != NULL)
spa_prop_add_list(*nvp, ZPOOL_PROP_ALTROOT, spa->spa_root,
0, ZPROP_SRC_LOCAL);
if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_BLOCKS)) {
spa_prop_add_list(*nvp, ZPOOL_PROP_MAXBLOCKSIZE, NULL,
MIN(zfs_max_recordsize, SPA_MAXBLOCKSIZE), ZPROP_SRC_NONE);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_MAXBLOCKSIZE, NULL,
SPA_OLD_MAXBLOCKSIZE, ZPROP_SRC_NONE);
}
if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_DNODE)) {
spa_prop_add_list(*nvp, ZPOOL_PROP_MAXDNODESIZE, NULL,
DNODE_MAX_SIZE, ZPROP_SRC_NONE);
} else {
spa_prop_add_list(*nvp, ZPOOL_PROP_MAXDNODESIZE, NULL,
DNODE_MIN_SIZE, ZPROP_SRC_NONE);
}
if ((dp = list_head(&spa->spa_config_list)) != NULL) {
if (dp->scd_path == NULL) {
spa_prop_add_list(*nvp, ZPOOL_PROP_CACHEFILE,
"none", 0, ZPROP_SRC_LOCAL);
} else if (strcmp(dp->scd_path, spa_config_path) != 0) {
spa_prop_add_list(*nvp, ZPOOL_PROP_CACHEFILE,
dp->scd_path, 0, ZPROP_SRC_LOCAL);
}
}
}
/*
* Get zpool property values.
*/
int
spa_prop_get(spa_t *spa, nvlist_t **nvp)
{
objset_t *mos = spa->spa_meta_objset;
zap_cursor_t zc;
zap_attribute_t za;
dsl_pool_t *dp;
int err;
err = nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP);
if (err)
return (err);
dp = spa_get_dsl(spa);
dsl_pool_config_enter(dp, FTAG);
mutex_enter(&spa->spa_props_lock);
/*
* Get properties from the spa config.
*/
spa_prop_get_config(spa, nvp);
/* If no pool property object, no more prop to get. */
if (mos == NULL || spa->spa_pool_props_object == 0)
goto out;
/*
* Get properties from the MOS pool property object.
*/
for (zap_cursor_init(&zc, mos, spa->spa_pool_props_object);
(err = zap_cursor_retrieve(&zc, &za)) == 0;
zap_cursor_advance(&zc)) {
uint64_t intval = 0;
char *strval = NULL;
zprop_source_t src = ZPROP_SRC_DEFAULT;
zpool_prop_t prop;
if ((prop = zpool_name_to_prop(za.za_name)) ==
ZPOOL_PROP_INVAL && !zfs_prop_user(za.za_name))
continue;
switch (za.za_integer_length) {
case 8:
/* integer property */
if (za.za_first_integer !=
zpool_prop_default_numeric(prop))
src = ZPROP_SRC_LOCAL;
if (prop == ZPOOL_PROP_BOOTFS) {
dsl_dataset_t *ds = NULL;
err = dsl_dataset_hold_obj(dp,
za.za_first_integer, FTAG, &ds);
if (err != 0)
break;
strval = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN,
KM_SLEEP);
dsl_dataset_name(ds, strval);
dsl_dataset_rele(ds, FTAG);
} else {
strval = NULL;
intval = za.za_first_integer;
}
spa_prop_add_list(*nvp, prop, strval, intval, src);
if (strval != NULL)
kmem_free(strval, ZFS_MAX_DATASET_NAME_LEN);
break;
case 1:
/* string property */
strval = kmem_alloc(za.za_num_integers, KM_SLEEP);
err = zap_lookup(mos, spa->spa_pool_props_object,
za.za_name, 1, za.za_num_integers, strval);
if (err) {
kmem_free(strval, za.za_num_integers);
break;
}
if (prop != ZPOOL_PROP_INVAL) {
spa_prop_add_list(*nvp, prop, strval, 0, src);
} else {
src = ZPROP_SRC_LOCAL;
spa_prop_add_user(*nvp, za.za_name, strval,
src);
}
kmem_free(strval, za.za_num_integers);
break;
default:
break;
}
}
zap_cursor_fini(&zc);
out:
mutex_exit(&spa->spa_props_lock);
dsl_pool_config_exit(dp, FTAG);
if (err && err != ENOENT) {
nvlist_free(*nvp);
*nvp = NULL;
return (err);
}
return (0);
}
/*
* Validate the given pool properties nvlist and modify the list
* for the property values to be set.
*/
static int
spa_prop_validate(spa_t *spa, nvlist_t *props)
{
nvpair_t *elem;
int error = 0, reset_bootfs = 0;
uint64_t objnum = 0;
boolean_t has_feature = B_FALSE;
elem = NULL;
while ((elem = nvlist_next_nvpair(props, elem)) != NULL) {
uint64_t intval;
const char *strval, *slash, *check, *fname;
const char *propname = nvpair_name(elem);
zpool_prop_t prop = zpool_name_to_prop(propname);
switch (prop) {
case ZPOOL_PROP_INVAL:
/*
* Sanitize the input.
*/
if (zfs_prop_user(propname)) {
if (strlen(propname) >= ZAP_MAXNAMELEN) {
error = SET_ERROR(ENAMETOOLONG);
break;
}
if (strlen(fnvpair_value_string(elem)) >=
ZAP_MAXVALUELEN) {
error = SET_ERROR(E2BIG);
break;
}
} else if (zpool_prop_feature(propname)) {
if (nvpair_type(elem) != DATA_TYPE_UINT64) {
error = SET_ERROR(EINVAL);
break;
}
if (nvpair_value_uint64(elem, &intval) != 0) {
error = SET_ERROR(EINVAL);
break;
}
if (intval != 0) {
error = SET_ERROR(EINVAL);
break;
}
fname = strchr(propname, '@') + 1;
if (zfeature_lookup_name(fname, NULL) != 0) {
error = SET_ERROR(EINVAL);
break;
}
has_feature = B_TRUE;
} else {
error = SET_ERROR(EINVAL);
break;
}
break;
case ZPOOL_PROP_VERSION:
error = nvpair_value_uint64(elem, &intval);
if (!error &&
(intval < spa_version(spa) ||
intval > SPA_VERSION_BEFORE_FEATURES ||
has_feature))
error = SET_ERROR(EINVAL);
break;
case ZPOOL_PROP_DELEGATION:
case ZPOOL_PROP_AUTOREPLACE:
case ZPOOL_PROP_LISTSNAPS:
case ZPOOL_PROP_AUTOEXPAND:
case ZPOOL_PROP_AUTOTRIM:
error = nvpair_value_uint64(elem, &intval);
if (!error && intval > 1)
error = SET_ERROR(EINVAL);
break;
case ZPOOL_PROP_MULTIHOST:
error = nvpair_value_uint64(elem, &intval);
if (!error && intval > 1)
error = SET_ERROR(EINVAL);
if (!error) {
uint32_t hostid = zone_get_hostid(NULL);
if (hostid)
spa->spa_hostid = hostid;
else
error = SET_ERROR(ENOTSUP);
}
break;
case ZPOOL_PROP_BOOTFS:
/*
* If the pool version is less than SPA_VERSION_BOOTFS,
* or the pool is still being created (version == 0),
* the bootfs property cannot be set.
*/
if (spa_version(spa) < SPA_VERSION_BOOTFS) {
error = SET_ERROR(ENOTSUP);
break;
}
/*
* Make sure the vdev config is bootable
*/
if (!vdev_is_bootable(spa->spa_root_vdev)) {
error = SET_ERROR(ENOTSUP);
break;
}
reset_bootfs = 1;
error = nvpair_value_string(elem, &strval);
if (!error) {
objset_t *os;
if (strval == NULL || strval[0] == '\0') {
objnum = zpool_prop_default_numeric(
ZPOOL_PROP_BOOTFS);
break;
}
error = dmu_objset_hold(strval, FTAG, &os);
if (error != 0)
break;
/* Must be ZPL. */
if (dmu_objset_type(os) != DMU_OST_ZFS) {
error = SET_ERROR(ENOTSUP);
} else {
objnum = dmu_objset_id(os);
}
dmu_objset_rele(os, FTAG);
}
break;
case ZPOOL_PROP_FAILUREMODE:
error = nvpair_value_uint64(elem, &intval);
if (!error && intval > ZIO_FAILURE_MODE_PANIC)
error = SET_ERROR(EINVAL);
/*
* This is a special case which only occurs when
* the pool has completely failed. This allows
* the user to change the in-core failmode property
* without syncing it out to disk (I/Os might
* currently be blocked). We do this by returning
* EIO to the caller (spa_prop_set) to trick it
* into thinking we encountered a property validation
* error.
*/
if (!error && spa_suspended(spa)) {
spa->spa_failmode = intval;
error = SET_ERROR(EIO);
}
break;
case ZPOOL_PROP_CACHEFILE:
if ((error = nvpair_value_string(elem, &strval)) != 0)
break;
if (strval[0] == '\0')
break;
if (strcmp(strval, "none") == 0)
break;
if (strval[0] != '/') {
error = SET_ERROR(EINVAL);
break;
}
slash = strrchr(strval, '/');
ASSERT(slash != NULL);
if (slash[1] == '\0' || strcmp(slash, "/.") == 0 ||
strcmp(slash, "/..") == 0)
error = SET_ERROR(EINVAL);
break;
case ZPOOL_PROP_COMMENT:
if ((error = nvpair_value_string(elem, &strval)) != 0)
break;
for (check = strval; *check != '\0'; check++) {
if (!isprint(*check)) {
error = SET_ERROR(EINVAL);
break;
}
}
if (strlen(strval) > ZPROP_MAX_COMMENT)
error = SET_ERROR(E2BIG);
break;
default:
break;
}
if (error)
break;
}
(void) nvlist_remove_all(props,
zpool_prop_to_name(ZPOOL_PROP_DEDUPDITTO));
if (!error && reset_bootfs) {
error = nvlist_remove(props,
zpool_prop_to_name(ZPOOL_PROP_BOOTFS), DATA_TYPE_STRING);
if (!error) {
error = nvlist_add_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_BOOTFS), objnum);
}
}
return (error);
}
void
spa_configfile_set(spa_t *spa, nvlist_t *nvp, boolean_t need_sync)
{
const char *cachefile;
spa_config_dirent_t *dp;
if (nvlist_lookup_string(nvp, zpool_prop_to_name(ZPOOL_PROP_CACHEFILE),
&cachefile) != 0)
return;
dp = kmem_alloc(sizeof (spa_config_dirent_t),
KM_SLEEP);
if (cachefile[0] == '\0')
dp->scd_path = spa_strdup(spa_config_path);
else if (strcmp(cachefile, "none") == 0)
dp->scd_path = NULL;
else
dp->scd_path = spa_strdup(cachefile);
list_insert_head(&spa->spa_config_list, dp);
if (need_sync)
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
int
spa_prop_set(spa_t *spa, nvlist_t *nvp)
{
int error;
nvpair_t *elem = NULL;
boolean_t need_sync = B_FALSE;
if ((error = spa_prop_validate(spa, nvp)) != 0)
return (error);
while ((elem = nvlist_next_nvpair(nvp, elem)) != NULL) {
zpool_prop_t prop = zpool_name_to_prop(nvpair_name(elem));
if (prop == ZPOOL_PROP_CACHEFILE ||
prop == ZPOOL_PROP_ALTROOT ||
prop == ZPOOL_PROP_READONLY)
continue;
if (prop == ZPOOL_PROP_INVAL &&
zfs_prop_user(nvpair_name(elem))) {
need_sync = B_TRUE;
break;
}
if (prop == ZPOOL_PROP_VERSION || prop == ZPOOL_PROP_INVAL) {
uint64_t ver = 0;
if (prop == ZPOOL_PROP_VERSION) {
VERIFY(nvpair_value_uint64(elem, &ver) == 0);
} else {
ASSERT(zpool_prop_feature(nvpair_name(elem)));
ver = SPA_VERSION_FEATURES;
need_sync = B_TRUE;
}
/* Save time if the version is already set. */
if (ver == spa_version(spa))
continue;
/*
* In addition to the pool directory object, we might
* create the pool properties object, the features for
* read object, the features for write object, or the
* feature descriptions object.
*/
error = dsl_sync_task(spa->spa_name, NULL,
spa_sync_version, &ver,
6, ZFS_SPACE_CHECK_RESERVED);
if (error)
return (error);
continue;
}
need_sync = B_TRUE;
break;
}
if (need_sync) {
return (dsl_sync_task(spa->spa_name, NULL, spa_sync_props,
nvp, 6, ZFS_SPACE_CHECK_RESERVED));
}
return (0);
}
/*
* If the bootfs property value is dsobj, clear it.
*/
void
spa_prop_clear_bootfs(spa_t *spa, uint64_t dsobj, dmu_tx_t *tx)
{
if (spa->spa_bootfs == dsobj && spa->spa_pool_props_object != 0) {
VERIFY(zap_remove(spa->spa_meta_objset,
spa->spa_pool_props_object,
zpool_prop_to_name(ZPOOL_PROP_BOOTFS), tx) == 0);
spa->spa_bootfs = 0;
}
}
static int
spa_change_guid_check(void *arg, dmu_tx_t *tx)
{
uint64_t *newguid __maybe_unused = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
vdev_t *rvd = spa->spa_root_vdev;
uint64_t vdev_state;
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) {
int error = (spa_has_checkpoint(spa)) ?
ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT;
return (SET_ERROR(error));
}
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
vdev_state = rvd->vdev_state;
spa_config_exit(spa, SCL_STATE, FTAG);
if (vdev_state != VDEV_STATE_HEALTHY)
return (SET_ERROR(ENXIO));
ASSERT3U(spa_guid(spa), !=, *newguid);
return (0);
}
static void
spa_change_guid_sync(void *arg, dmu_tx_t *tx)
{
uint64_t *newguid = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
uint64_t oldguid;
vdev_t *rvd = spa->spa_root_vdev;
oldguid = spa_guid(spa);
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
rvd->vdev_guid = *newguid;
rvd->vdev_guid_sum += (*newguid - oldguid);
vdev_config_dirty(rvd);
spa_config_exit(spa, SCL_STATE, FTAG);
spa_history_log_internal(spa, "guid change", tx, "old=%llu new=%llu",
(u_longlong_t)oldguid, (u_longlong_t)*newguid);
}
/*
* Change the GUID for the pool. This is done so that we can later
* re-import a pool built from a clone of our own vdevs. We will modify
* the root vdev's guid, our own pool guid, and then mark all of our
* vdevs dirty. Note that we must make sure that all our vdevs are
* online when we do this, or else any vdevs that weren't present
* would be orphaned from our pool. We are also going to issue a
* sysevent to update any watchers.
*/
int
spa_change_guid(spa_t *spa)
{
int error;
uint64_t guid;
mutex_enter(&spa->spa_vdev_top_lock);
mutex_enter(&spa_namespace_lock);
guid = spa_generate_guid(NULL);
error = dsl_sync_task(spa->spa_name, spa_change_guid_check,
spa_change_guid_sync, &guid, 5, ZFS_SPACE_CHECK_RESERVED);
if (error == 0) {
/*
* Clear the kobj flag from all the vdevs to allow
* vdev_cache_process_kobj_evt() to post events to all the
* vdevs since GUID is updated.
*/
vdev_clear_kobj_evt(spa->spa_root_vdev);
for (int i = 0; i < spa->spa_l2cache.sav_count; i++)
vdev_clear_kobj_evt(spa->spa_l2cache.sav_vdevs[i]);
spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_REGUID);
}
mutex_exit(&spa_namespace_lock);
mutex_exit(&spa->spa_vdev_top_lock);
return (error);
}
/*
* ==========================================================================
* SPA state manipulation (open/create/destroy/import/export)
* ==========================================================================
*/
static int
spa_error_entry_compare(const void *a, const void *b)
{
const spa_error_entry_t *sa = (const spa_error_entry_t *)a;
const spa_error_entry_t *sb = (const spa_error_entry_t *)b;
int ret;
ret = memcmp(&sa->se_bookmark, &sb->se_bookmark,
sizeof (zbookmark_phys_t));
return (TREE_ISIGN(ret));
}
/*
* Utility function which retrieves copies of the current logs and
* re-initializes them in the process.
*/
void
spa_get_errlists(spa_t *spa, avl_tree_t *last, avl_tree_t *scrub)
{
ASSERT(MUTEX_HELD(&spa->spa_errlist_lock));
memcpy(last, &spa->spa_errlist_last, sizeof (avl_tree_t));
memcpy(scrub, &spa->spa_errlist_scrub, sizeof (avl_tree_t));
avl_create(&spa->spa_errlist_scrub,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
avl_create(&spa->spa_errlist_last,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
}
static void
spa_taskqs_init(spa_t *spa, zio_type_t t, zio_taskq_type_t q)
{
const zio_taskq_info_t *ztip = &zio_taskqs[t][q];
enum zti_modes mode = ztip->zti_mode;
uint_t value = ztip->zti_value;
uint_t count = ztip->zti_count;
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
uint_t cpus, flags = TASKQ_DYNAMIC;
boolean_t batch = B_FALSE;
switch (mode) {
case ZTI_MODE_FIXED:
ASSERT3U(value, >, 0);
break;
case ZTI_MODE_BATCH:
batch = B_TRUE;
flags |= TASKQ_THREADS_CPU_PCT;
value = MIN(zio_taskq_batch_pct, 100);
break;
case ZTI_MODE_SCALE:
flags |= TASKQ_THREADS_CPU_PCT;
/*
* We want more taskqs to reduce lock contention, but we want
* less for better request ordering and CPU utilization.
*/
cpus = MAX(1, boot_ncpus * zio_taskq_batch_pct / 100);
if (zio_taskq_batch_tpq > 0) {
count = MAX(1, (cpus + zio_taskq_batch_tpq / 2) /
zio_taskq_batch_tpq);
} else {
/*
* Prefer 6 threads per taskq, but no more taskqs
* than threads in them on large systems. For 80%:
*
* taskq taskq total
* cpus taskqs percent threads threads
* ------- ------- ------- ------- -------
* 1 1 80% 1 1
* 2 1 80% 1 1
* 4 1 80% 3 3
* 8 2 40% 3 6
* 16 3 27% 4 12
* 32 5 16% 5 25
* 64 7 11% 7 49
* 128 10 8% 10 100
* 256 14 6% 15 210
*/
count = 1 + cpus / 6;
while (count * count > cpus)
count--;
}
/* Limit each taskq within 100% to not trigger assertion. */
count = MAX(count, (zio_taskq_batch_pct + 99) / 100);
value = (zio_taskq_batch_pct + count / 2) / count;
break;
case ZTI_MODE_NULL:
tqs->stqs_count = 0;
tqs->stqs_taskq = NULL;
return;
default:
panic("unrecognized mode for %s_%s taskq (%u:%u) in "
"spa_activate()",
zio_type_name[t], zio_taskq_types[q], mode, value);
break;
}
ASSERT3U(count, >, 0);
tqs->stqs_count = count;
tqs->stqs_taskq = kmem_alloc(count * sizeof (taskq_t *), KM_SLEEP);
for (uint_t i = 0; i < count; i++) {
taskq_t *tq;
char name[32];
if (count > 1)
(void) snprintf(name, sizeof (name), "%s_%s_%u",
zio_type_name[t], zio_taskq_types[q], i);
else
(void) snprintf(name, sizeof (name), "%s_%s",
zio_type_name[t], zio_taskq_types[q]);
if (zio_taskq_sysdc && spa->spa_proc != &p0) {
if (batch)
flags |= TASKQ_DC_BATCH;
(void) zio_taskq_basedc;
tq = taskq_create_sysdc(name, value, 50, INT_MAX,
spa->spa_proc, zio_taskq_basedc, flags);
} else {
pri_t pri = maxclsyspri;
/*
* The write issue taskq can be extremely CPU
* intensive. Run it at slightly less important
* priority than the other taskqs.
*
* Under Linux and FreeBSD this means incrementing
* the priority value as opposed to platforms like
* illumos where it should be decremented.
*
* On FreeBSD, if priorities divided by four (RQ_PPQ)
* are equal then a difference between them is
* insignificant.
*/
if (t == ZIO_TYPE_WRITE && q == ZIO_TASKQ_ISSUE) {
#if defined(__linux__)
pri++;
#elif defined(__FreeBSD__)
pri += 4;
#else
#error "unknown OS"
#endif
}
tq = taskq_create_proc(name, value, pri, 50,
INT_MAX, spa->spa_proc, flags);
}
tqs->stqs_taskq[i] = tq;
}
}
static void
spa_taskqs_fini(spa_t *spa, zio_type_t t, zio_taskq_type_t q)
{
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
if (tqs->stqs_taskq == NULL) {
ASSERT3U(tqs->stqs_count, ==, 0);
return;
}
for (uint_t i = 0; i < tqs->stqs_count; i++) {
ASSERT3P(tqs->stqs_taskq[i], !=, NULL);
taskq_destroy(tqs->stqs_taskq[i]);
}
kmem_free(tqs->stqs_taskq, tqs->stqs_count * sizeof (taskq_t *));
tqs->stqs_taskq = NULL;
}
/*
* Dispatch a task to the appropriate taskq for the ZFS I/O type and priority.
* Note that a type may have multiple discrete taskqs to avoid lock contention
* on the taskq itself. In that case we choose which taskq at random by using
* the low bits of gethrtime().
*/
void
spa_taskq_dispatch_ent(spa_t *spa, zio_type_t t, zio_taskq_type_t q,
task_func_t *func, void *arg, uint_t flags, taskq_ent_t *ent)
{
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
taskq_t *tq;
ASSERT3P(tqs->stqs_taskq, !=, NULL);
ASSERT3U(tqs->stqs_count, !=, 0);
if (tqs->stqs_count == 1) {
tq = tqs->stqs_taskq[0];
} else {
tq = tqs->stqs_taskq[((uint64_t)gethrtime()) % tqs->stqs_count];
}
taskq_dispatch_ent(tq, func, arg, flags, ent);
}
/*
* Same as spa_taskq_dispatch_ent() but block on the task until completion.
*/
void
spa_taskq_dispatch_sync(spa_t *spa, zio_type_t t, zio_taskq_type_t q,
task_func_t *func, void *arg, uint_t flags)
{
spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q];
taskq_t *tq;
taskqid_t id;
ASSERT3P(tqs->stqs_taskq, !=, NULL);
ASSERT3U(tqs->stqs_count, !=, 0);
if (tqs->stqs_count == 1) {
tq = tqs->stqs_taskq[0];
} else {
tq = tqs->stqs_taskq[((uint64_t)gethrtime()) % tqs->stqs_count];
}
id = taskq_dispatch(tq, func, arg, flags);
if (id)
taskq_wait_id(tq, id);
}
static void
spa_create_zio_taskqs(spa_t *spa)
{
for (int t = 0; t < ZIO_TYPES; t++) {
for (int q = 0; q < ZIO_TASKQ_TYPES; q++) {
spa_taskqs_init(spa, t, q);
}
}
}
/*
* Disabled until spa_thread() can be adapted for Linux.
*/
#undef HAVE_SPA_THREAD
#if defined(_KERNEL) && defined(HAVE_SPA_THREAD)
static void
spa_thread(void *arg)
{
psetid_t zio_taskq_psrset_bind = PS_NONE;
callb_cpr_t cprinfo;
spa_t *spa = arg;
user_t *pu = PTOU(curproc);
CALLB_CPR_INIT(&cprinfo, &spa->spa_proc_lock, callb_generic_cpr,
spa->spa_name);
ASSERT(curproc != &p0);
(void) snprintf(pu->u_psargs, sizeof (pu->u_psargs),
"zpool-%s", spa->spa_name);
(void) strlcpy(pu->u_comm, pu->u_psargs, sizeof (pu->u_comm));
/* bind this thread to the requested psrset */
if (zio_taskq_psrset_bind != PS_NONE) {
pool_lock();
mutex_enter(&cpu_lock);
mutex_enter(&pidlock);
mutex_enter(&curproc->p_lock);
if (cpupart_bind_thread(curthread, zio_taskq_psrset_bind,
0, NULL, NULL) == 0) {
curthread->t_bind_pset = zio_taskq_psrset_bind;
} else {
cmn_err(CE_WARN,
"Couldn't bind process for zfs pool \"%s\" to "
"pset %d\n", spa->spa_name, zio_taskq_psrset_bind);
}
mutex_exit(&curproc->p_lock);
mutex_exit(&pidlock);
mutex_exit(&cpu_lock);
pool_unlock();
}
if (zio_taskq_sysdc) {
sysdc_thread_enter(curthread, 100, 0);
}
spa->spa_proc = curproc;
spa->spa_did = curthread->t_did;
spa_create_zio_taskqs(spa);
mutex_enter(&spa->spa_proc_lock);
ASSERT(spa->spa_proc_state == SPA_PROC_CREATED);
spa->spa_proc_state = SPA_PROC_ACTIVE;
cv_broadcast(&spa->spa_proc_cv);
CALLB_CPR_SAFE_BEGIN(&cprinfo);
while (spa->spa_proc_state == SPA_PROC_ACTIVE)
cv_wait(&spa->spa_proc_cv, &spa->spa_proc_lock);
CALLB_CPR_SAFE_END(&cprinfo, &spa->spa_proc_lock);
ASSERT(spa->spa_proc_state == SPA_PROC_DEACTIVATE);
spa->spa_proc_state = SPA_PROC_GONE;
spa->spa_proc = &p0;
cv_broadcast(&spa->spa_proc_cv);
CALLB_CPR_EXIT(&cprinfo); /* drops spa_proc_lock */
mutex_enter(&curproc->p_lock);
lwp_exit();
}
#endif
/*
* Activate an uninitialized pool.
*/
static void
spa_activate(spa_t *spa, spa_mode_t mode)
{
ASSERT(spa->spa_state == POOL_STATE_UNINITIALIZED);
spa->spa_state = POOL_STATE_ACTIVE;
spa->spa_mode = mode;
spa->spa_read_spacemaps = spa_mode_readable_spacemaps;
spa->spa_normal_class = metaslab_class_create(spa, &zfs_metaslab_ops);
spa->spa_log_class = metaslab_class_create(spa, &zfs_metaslab_ops);
spa->spa_embedded_log_class =
metaslab_class_create(spa, &zfs_metaslab_ops);
spa->spa_special_class = metaslab_class_create(spa, &zfs_metaslab_ops);
spa->spa_dedup_class = metaslab_class_create(spa, &zfs_metaslab_ops);
/* Try to create a covering process */
mutex_enter(&spa->spa_proc_lock);
ASSERT(spa->spa_proc_state == SPA_PROC_NONE);
ASSERT(spa->spa_proc == &p0);
spa->spa_did = 0;
(void) spa_create_process;
#ifdef HAVE_SPA_THREAD
/* Only create a process if we're going to be around a while. */
if (spa_create_process && strcmp(spa->spa_name, TRYIMPORT_NAME) != 0) {
if (newproc(spa_thread, (caddr_t)spa, syscid, maxclsyspri,
NULL, 0) == 0) {
spa->spa_proc_state = SPA_PROC_CREATED;
while (spa->spa_proc_state == SPA_PROC_CREATED) {
cv_wait(&spa->spa_proc_cv,
&spa->spa_proc_lock);
}
ASSERT(spa->spa_proc_state == SPA_PROC_ACTIVE);
ASSERT(spa->spa_proc != &p0);
ASSERT(spa->spa_did != 0);
} else {
#ifdef _KERNEL
cmn_err(CE_WARN,
"Couldn't create process for zfs pool \"%s\"\n",
spa->spa_name);
#endif
}
}
#endif /* HAVE_SPA_THREAD */
mutex_exit(&spa->spa_proc_lock);
/* If we didn't create a process, we need to create our taskqs. */
if (spa->spa_proc == &p0) {
spa_create_zio_taskqs(spa);
}
for (size_t i = 0; i < TXG_SIZE; i++) {
spa->spa_txg_zio[i] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL);
}
list_create(&spa->spa_config_dirty_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_config_dirty_node));
list_create(&spa->spa_evicting_os_list, sizeof (objset_t),
offsetof(objset_t, os_evicting_node));
list_create(&spa->spa_state_dirty_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_state_dirty_node));
txg_list_create(&spa->spa_vdev_txg_list, spa,
offsetof(struct vdev, vdev_txg_node));
avl_create(&spa->spa_errlist_scrub,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
avl_create(&spa->spa_errlist_last,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
avl_create(&spa->spa_errlist_healed,
spa_error_entry_compare, sizeof (spa_error_entry_t),
offsetof(spa_error_entry_t, se_avl));
spa_activate_os(spa);
spa_keystore_init(&spa->spa_keystore);
/*
* This taskq is used to perform zvol-minor-related tasks
* asynchronously. This has several advantages, including easy
* resolution of various deadlocks.
*
* The taskq must be single threaded to ensure tasks are always
* processed in the order in which they were dispatched.
*
* A taskq per pool allows one to keep the pools independent.
* This way if one pool is suspended, it will not impact another.
*
* The preferred location to dispatch a zvol minor task is a sync
* task. In this context, there is easy access to the spa_t and minimal
* error handling is required because the sync task must succeed.
*/
spa->spa_zvol_taskq = taskq_create("z_zvol", 1, defclsyspri,
1, INT_MAX, 0);
/*
* Taskq dedicated to prefetcher threads: this is used to prevent the
* pool traverse code from monopolizing the global (and limited)
* system_taskq by inappropriately scheduling long running tasks on it.
*/
spa->spa_prefetch_taskq = taskq_create("z_prefetch", 100,
defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC | TASKQ_THREADS_CPU_PCT);
/*
* The taskq to upgrade datasets in this pool. Currently used by
* feature SPA_FEATURE_USEROBJ_ACCOUNTING/SPA_FEATURE_PROJECT_QUOTA.
*/
spa->spa_upgrade_taskq = taskq_create("z_upgrade", 100,
defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC | TASKQ_THREADS_CPU_PCT);
}
/*
* Opposite of spa_activate().
*/
static void
spa_deactivate(spa_t *spa)
{
ASSERT(spa->spa_sync_on == B_FALSE);
ASSERT(spa->spa_dsl_pool == NULL);
ASSERT(spa->spa_root_vdev == NULL);
ASSERT(spa->spa_async_zio_root == NULL);
ASSERT(spa->spa_state != POOL_STATE_UNINITIALIZED);
spa_evicting_os_wait(spa);
if (spa->spa_zvol_taskq) {
taskq_destroy(spa->spa_zvol_taskq);
spa->spa_zvol_taskq = NULL;
}
if (spa->spa_prefetch_taskq) {
taskq_destroy(spa->spa_prefetch_taskq);
spa->spa_prefetch_taskq = NULL;
}
if (spa->spa_upgrade_taskq) {
taskq_destroy(spa->spa_upgrade_taskq);
spa->spa_upgrade_taskq = NULL;
}
txg_list_destroy(&spa->spa_vdev_txg_list);
list_destroy(&spa->spa_config_dirty_list);
list_destroy(&spa->spa_evicting_os_list);
list_destroy(&spa->spa_state_dirty_list);
taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid);
for (int t = 0; t < ZIO_TYPES; t++) {
for (int q = 0; q < ZIO_TASKQ_TYPES; q++) {
spa_taskqs_fini(spa, t, q);
}
}
for (size_t i = 0; i < TXG_SIZE; i++) {
ASSERT3P(spa->spa_txg_zio[i], !=, NULL);
VERIFY0(zio_wait(spa->spa_txg_zio[i]));
spa->spa_txg_zio[i] = NULL;
}
metaslab_class_destroy(spa->spa_normal_class);
spa->spa_normal_class = NULL;
metaslab_class_destroy(spa->spa_log_class);
spa->spa_log_class = NULL;
metaslab_class_destroy(spa->spa_embedded_log_class);
spa->spa_embedded_log_class = NULL;
metaslab_class_destroy(spa->spa_special_class);
spa->spa_special_class = NULL;
metaslab_class_destroy(spa->spa_dedup_class);
spa->spa_dedup_class = NULL;
/*
* If this was part of an import or the open otherwise failed, we may
* still have errors left in the queues. Empty them just in case.
*/
spa_errlog_drain(spa);
avl_destroy(&spa->spa_errlist_scrub);
avl_destroy(&spa->spa_errlist_last);
avl_destroy(&spa->spa_errlist_healed);
spa_keystore_fini(&spa->spa_keystore);
spa->spa_state = POOL_STATE_UNINITIALIZED;
mutex_enter(&spa->spa_proc_lock);
if (spa->spa_proc_state != SPA_PROC_NONE) {
ASSERT(spa->spa_proc_state == SPA_PROC_ACTIVE);
spa->spa_proc_state = SPA_PROC_DEACTIVATE;
cv_broadcast(&spa->spa_proc_cv);
while (spa->spa_proc_state == SPA_PROC_DEACTIVATE) {
ASSERT(spa->spa_proc != &p0);
cv_wait(&spa->spa_proc_cv, &spa->spa_proc_lock);
}
ASSERT(spa->spa_proc_state == SPA_PROC_GONE);
spa->spa_proc_state = SPA_PROC_NONE;
}
ASSERT(spa->spa_proc == &p0);
mutex_exit(&spa->spa_proc_lock);
/*
* We want to make sure spa_thread() has actually exited the ZFS
* module, so that the module can't be unloaded out from underneath
* it.
*/
if (spa->spa_did != 0) {
thread_join(spa->spa_did);
spa->spa_did = 0;
}
spa_deactivate_os(spa);
}
/*
* Verify a pool configuration, and construct the vdev tree appropriately. This
* will create all the necessary vdevs in the appropriate layout, with each vdev
* in the CLOSED state. This will prep the pool before open/creation/import.
* All vdev validation is done by the vdev_alloc() routine.
*/
int
spa_config_parse(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent,
uint_t id, int atype)
{
nvlist_t **child;
uint_t children;
int error;
if ((error = vdev_alloc(spa, vdp, nv, parent, id, atype)) != 0)
return (error);
if ((*vdp)->vdev_ops->vdev_op_leaf)
return (0);
error = nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
&child, &children);
if (error == ENOENT)
return (0);
if (error) {
vdev_free(*vdp);
*vdp = NULL;
return (SET_ERROR(EINVAL));
}
for (int c = 0; c < children; c++) {
vdev_t *vd;
if ((error = spa_config_parse(spa, &vd, child[c], *vdp, c,
atype)) != 0) {
vdev_free(*vdp);
*vdp = NULL;
return (error);
}
}
ASSERT(*vdp != NULL);
return (0);
}
static boolean_t
spa_should_flush_logs_on_unload(spa_t *spa)
{
if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP))
return (B_FALSE);
if (!spa_writeable(spa))
return (B_FALSE);
if (!spa->spa_sync_on)
return (B_FALSE);
if (spa_state(spa) != POOL_STATE_EXPORTED)
return (B_FALSE);
if (zfs_keep_log_spacemaps_at_export)
return (B_FALSE);
return (B_TRUE);
}
/*
* Opens a transaction that will set the flag that will instruct
* spa_sync to attempt to flush all the metaslabs for that txg.
*/
static void
spa_unload_log_sm_flush_all(spa_t *spa)
{
dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
ASSERT3U(spa->spa_log_flushall_txg, ==, 0);
spa->spa_log_flushall_txg = dmu_tx_get_txg(tx);
dmu_tx_commit(tx);
txg_wait_synced(spa_get_dsl(spa), spa->spa_log_flushall_txg);
}
static void
spa_unload_log_sm_metadata(spa_t *spa)
{
void *cookie = NULL;
spa_log_sm_t *sls;
while ((sls = avl_destroy_nodes(&spa->spa_sm_logs_by_txg,
&cookie)) != NULL) {
VERIFY0(sls->sls_mscount);
kmem_free(sls, sizeof (spa_log_sm_t));
}
for (log_summary_entry_t *e = list_head(&spa->spa_log_summary);
e != NULL; e = list_head(&spa->spa_log_summary)) {
VERIFY0(e->lse_mscount);
list_remove(&spa->spa_log_summary, e);
kmem_free(e, sizeof (log_summary_entry_t));
}
spa->spa_unflushed_stats.sus_nblocks = 0;
spa->spa_unflushed_stats.sus_memused = 0;
spa->spa_unflushed_stats.sus_blocklimit = 0;
}
static void
spa_destroy_aux_threads(spa_t *spa)
{
if (spa->spa_condense_zthr != NULL) {
zthr_destroy(spa->spa_condense_zthr);
spa->spa_condense_zthr = NULL;
}
if (spa->spa_checkpoint_discard_zthr != NULL) {
zthr_destroy(spa->spa_checkpoint_discard_zthr);
spa->spa_checkpoint_discard_zthr = NULL;
}
if (spa->spa_livelist_delete_zthr != NULL) {
zthr_destroy(spa->spa_livelist_delete_zthr);
spa->spa_livelist_delete_zthr = NULL;
}
if (spa->spa_livelist_condense_zthr != NULL) {
zthr_destroy(spa->spa_livelist_condense_zthr);
spa->spa_livelist_condense_zthr = NULL;
}
}
/*
* Opposite of spa_load().
*/
static void
spa_unload(spa_t *spa)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa_state(spa) != POOL_STATE_UNINITIALIZED);
spa_import_progress_remove(spa_guid(spa));
spa_load_note(spa, "UNLOADING");
spa_wake_waiters(spa);
/*
* If we have set the spa_final_txg, we have already performed the
* tasks below in spa_export_common(). We should not redo it here since
* we delay the final TXGs beyond what spa_final_txg is set at.
*/
if (spa->spa_final_txg == UINT64_MAX) {
/*
* If the log space map feature is enabled and the pool is
* getting exported (but not destroyed), we want to spend some
* time flushing as many metaslabs as we can in an attempt to
* destroy log space maps and save import time.
*/
if (spa_should_flush_logs_on_unload(spa))
spa_unload_log_sm_flush_all(spa);
/*
* Stop async tasks.
*/
spa_async_suspend(spa);
if (spa->spa_root_vdev) {
vdev_t *root_vdev = spa->spa_root_vdev;
vdev_initialize_stop_all(root_vdev,
VDEV_INITIALIZE_ACTIVE);
vdev_trim_stop_all(root_vdev, VDEV_TRIM_ACTIVE);
vdev_autotrim_stop_all(spa);
vdev_rebuild_stop_all(spa);
}
}
/*
* Stop syncing.
*/
if (spa->spa_sync_on) {
txg_sync_stop(spa->spa_dsl_pool);
spa->spa_sync_on = B_FALSE;
}
/*
* This ensures that there is no async metaslab prefetching
* while we attempt to unload the spa.
*/
if (spa->spa_root_vdev != NULL) {
for (int c = 0; c < spa->spa_root_vdev->vdev_children; c++) {
vdev_t *vc = spa->spa_root_vdev->vdev_child[c];
if (vc->vdev_mg != NULL)
taskq_wait(vc->vdev_mg->mg_taskq);
}
}
if (spa->spa_mmp.mmp_thread)
mmp_thread_stop(spa);
/*
* Wait for any outstanding async I/O to complete.
*/
if (spa->spa_async_zio_root != NULL) {
for (int i = 0; i < max_ncpus; i++)
(void) zio_wait(spa->spa_async_zio_root[i]);
kmem_free(spa->spa_async_zio_root, max_ncpus * sizeof (void *));
spa->spa_async_zio_root = NULL;
}
if (spa->spa_vdev_removal != NULL) {
spa_vdev_removal_destroy(spa->spa_vdev_removal);
spa->spa_vdev_removal = NULL;
}
spa_destroy_aux_threads(spa);
spa_condense_fini(spa);
bpobj_close(&spa->spa_deferred_bpobj);
spa_config_enter(spa, SCL_ALL, spa, RW_WRITER);
/*
* Close all vdevs.
*/
if (spa->spa_root_vdev)
vdev_free(spa->spa_root_vdev);
ASSERT(spa->spa_root_vdev == NULL);
/*
* Close the dsl pool.
*/
if (spa->spa_dsl_pool) {
dsl_pool_close(spa->spa_dsl_pool);
spa->spa_dsl_pool = NULL;
spa->spa_meta_objset = NULL;
}
ddt_unload(spa);
brt_unload(spa);
spa_unload_log_sm_metadata(spa);
/*
* Drop and purge level 2 cache
*/
spa_l2cache_drop(spa);
if (spa->spa_spares.sav_vdevs) {
for (int i = 0; i < spa->spa_spares.sav_count; i++)
vdev_free(spa->spa_spares.sav_vdevs[i]);
kmem_free(spa->spa_spares.sav_vdevs,
spa->spa_spares.sav_count * sizeof (void *));
spa->spa_spares.sav_vdevs = NULL;
}
if (spa->spa_spares.sav_config) {
nvlist_free(spa->spa_spares.sav_config);
spa->spa_spares.sav_config = NULL;
}
spa->spa_spares.sav_count = 0;
if (spa->spa_l2cache.sav_vdevs) {
for (int i = 0; i < spa->spa_l2cache.sav_count; i++) {
vdev_clear_stats(spa->spa_l2cache.sav_vdevs[i]);
vdev_free(spa->spa_l2cache.sav_vdevs[i]);
}
kmem_free(spa->spa_l2cache.sav_vdevs,
spa->spa_l2cache.sav_count * sizeof (void *));
spa->spa_l2cache.sav_vdevs = NULL;
}
if (spa->spa_l2cache.sav_config) {
nvlist_free(spa->spa_l2cache.sav_config);
spa->spa_l2cache.sav_config = NULL;
}
spa->spa_l2cache.sav_count = 0;
spa->spa_async_suspended = 0;
spa->spa_indirect_vdevs_loaded = B_FALSE;
if (spa->spa_comment != NULL) {
spa_strfree(spa->spa_comment);
spa->spa_comment = NULL;
}
if (spa->spa_compatibility != NULL) {
spa_strfree(spa->spa_compatibility);
spa->spa_compatibility = NULL;
}
spa_config_exit(spa, SCL_ALL, spa);
}
/*
* Load (or re-load) the current list of vdevs describing the active spares for
* this pool. When this is called, we have some form of basic information in
* 'spa_spares.sav_config'. We parse this into vdevs, try to open them, and
* then re-generate a more complete list including status information.
*/
void
spa_load_spares(spa_t *spa)
{
nvlist_t **spares;
uint_t nspares;
int i;
vdev_t *vd, *tvd;
#ifndef _KERNEL
/*
* zdb opens both the current state of the pool and the
* checkpointed state (if present), with a different spa_t.
*
* As spare vdevs are shared among open pools, we skip loading
* them when we load the checkpointed state of the pool.
*/
if (!spa_writeable(spa))
return;
#endif
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
/*
* First, close and free any existing spare vdevs.
*/
if (spa->spa_spares.sav_vdevs) {
for (i = 0; i < spa->spa_spares.sav_count; i++) {
vd = spa->spa_spares.sav_vdevs[i];
/* Undo the call to spa_activate() below */
if ((tvd = spa_lookup_by_guid(spa, vd->vdev_guid,
B_FALSE)) != NULL && tvd->vdev_isspare)
spa_spare_remove(tvd);
vdev_close(vd);
vdev_free(vd);
}
kmem_free(spa->spa_spares.sav_vdevs,
spa->spa_spares.sav_count * sizeof (void *));
}
if (spa->spa_spares.sav_config == NULL)
nspares = 0;
else
VERIFY0(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, &spares, &nspares));
spa->spa_spares.sav_count = (int)nspares;
spa->spa_spares.sav_vdevs = NULL;
if (nspares == 0)
return;
/*
* Construct the array of vdevs, opening them to get status in the
* process. For each spare, there is potentially two different vdev_t
* structures associated with it: one in the list of spares (used only
* for basic validation purposes) and one in the active vdev
* configuration (if it's spared in). During this phase we open and
* validate each vdev on the spare list. If the vdev also exists in the
* active configuration, then we also mark this vdev as an active spare.
*/
spa->spa_spares.sav_vdevs = kmem_zalloc(nspares * sizeof (void *),
KM_SLEEP);
for (i = 0; i < spa->spa_spares.sav_count; i++) {
VERIFY(spa_config_parse(spa, &vd, spares[i], NULL, 0,
VDEV_ALLOC_SPARE) == 0);
ASSERT(vd != NULL);
spa->spa_spares.sav_vdevs[i] = vd;
if ((tvd = spa_lookup_by_guid(spa, vd->vdev_guid,
B_FALSE)) != NULL) {
if (!tvd->vdev_isspare)
spa_spare_add(tvd);
/*
* We only mark the spare active if we were successfully
* able to load the vdev. Otherwise, importing a pool
* with a bad active spare would result in strange
* behavior, because multiple pool would think the spare
* is actively in use.
*
* There is a vulnerability here to an equally bizarre
* circumstance, where a dead active spare is later
* brought back to life (onlined or otherwise). Given
* the rarity of this scenario, and the extra complexity
* it adds, we ignore the possibility.
*/
if (!vdev_is_dead(tvd))
spa_spare_activate(tvd);
}
vd->vdev_top = vd;
vd->vdev_aux = &spa->spa_spares;
if (vdev_open(vd) != 0)
continue;
if (vdev_validate_aux(vd) == 0)
spa_spare_add(vd);
}
/*
* Recompute the stashed list of spares, with status information
* this time.
*/
fnvlist_remove(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES);
spares = kmem_alloc(spa->spa_spares.sav_count * sizeof (void *),
KM_SLEEP);
for (i = 0; i < spa->spa_spares.sav_count; i++)
spares[i] = vdev_config_generate(spa,
spa->spa_spares.sav_vdevs[i], B_TRUE, VDEV_CONFIG_SPARE);
fnvlist_add_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares,
spa->spa_spares.sav_count);
for (i = 0; i < spa->spa_spares.sav_count; i++)
nvlist_free(spares[i]);
kmem_free(spares, spa->spa_spares.sav_count * sizeof (void *));
}
/*
* Load (or re-load) the current list of vdevs describing the active l2cache for
* this pool. When this is called, we have some form of basic information in
* 'spa_l2cache.sav_config'. We parse this into vdevs, try to open them, and
* then re-generate a more complete list including status information.
* Devices which are already active have their details maintained, and are
* not re-opened.
*/
void
spa_load_l2cache(spa_t *spa)
{
nvlist_t **l2cache = NULL;
uint_t nl2cache;
int i, j, oldnvdevs;
uint64_t guid;
vdev_t *vd, **oldvdevs, **newvdevs;
spa_aux_vdev_t *sav = &spa->spa_l2cache;
#ifndef _KERNEL
/*
* zdb opens both the current state of the pool and the
* checkpointed state (if present), with a different spa_t.
*
* As L2 caches are part of the ARC which is shared among open
* pools, we skip loading them when we load the checkpointed
* state of the pool.
*/
if (!spa_writeable(spa))
return;
#endif
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
oldvdevs = sav->sav_vdevs;
oldnvdevs = sav->sav_count;
sav->sav_vdevs = NULL;
sav->sav_count = 0;
if (sav->sav_config == NULL) {
nl2cache = 0;
newvdevs = NULL;
goto out;
}
VERIFY0(nvlist_lookup_nvlist_array(sav->sav_config,
ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache));
newvdevs = kmem_alloc(nl2cache * sizeof (void *), KM_SLEEP);
/*
* Process new nvlist of vdevs.
*/
for (i = 0; i < nl2cache; i++) {
guid = fnvlist_lookup_uint64(l2cache[i], ZPOOL_CONFIG_GUID);
newvdevs[i] = NULL;
for (j = 0; j < oldnvdevs; j++) {
vd = oldvdevs[j];
if (vd != NULL && guid == vd->vdev_guid) {
/*
* Retain previous vdev for add/remove ops.
*/
newvdevs[i] = vd;
oldvdevs[j] = NULL;
break;
}
}
if (newvdevs[i] == NULL) {
/*
* Create new vdev
*/
VERIFY(spa_config_parse(spa, &vd, l2cache[i], NULL, 0,
VDEV_ALLOC_L2CACHE) == 0);
ASSERT(vd != NULL);
newvdevs[i] = vd;
/*
* Commit this vdev as an l2cache device,
* even if it fails to open.
*/
spa_l2cache_add(vd);
vd->vdev_top = vd;
vd->vdev_aux = sav;
spa_l2cache_activate(vd);
if (vdev_open(vd) != 0)
continue;
(void) vdev_validate_aux(vd);
if (!vdev_is_dead(vd))
l2arc_add_vdev(spa, vd);
/*
* Upon cache device addition to a pool or pool
* creation with a cache device or if the header
* of the device is invalid we issue an async
* TRIM command for the whole device which will
* execute if l2arc_trim_ahead > 0.
*/
spa_async_request(spa, SPA_ASYNC_L2CACHE_TRIM);
}
}
sav->sav_vdevs = newvdevs;
sav->sav_count = (int)nl2cache;
/*
* Recompute the stashed list of l2cache devices, with status
* information this time.
*/
fnvlist_remove(sav->sav_config, ZPOOL_CONFIG_L2CACHE);
if (sav->sav_count > 0)
l2cache = kmem_alloc(sav->sav_count * sizeof (void *),
KM_SLEEP);
for (i = 0; i < sav->sav_count; i++)
l2cache[i] = vdev_config_generate(spa,
sav->sav_vdevs[i], B_TRUE, VDEV_CONFIG_L2CACHE);
fnvlist_add_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE,
(const nvlist_t * const *)l2cache, sav->sav_count);
out:
/*
* Purge vdevs that were dropped
*/
if (oldvdevs) {
for (i = 0; i < oldnvdevs; i++) {
uint64_t pool;
vd = oldvdevs[i];
if (vd != NULL) {
ASSERT(vd->vdev_isl2cache);
if (spa_l2cache_exists(vd->vdev_guid, &pool) &&
pool != 0ULL && l2arc_vdev_present(vd))
l2arc_remove_vdev(vd);
vdev_clear_stats(vd);
vdev_free(vd);
}
}
kmem_free(oldvdevs, oldnvdevs * sizeof (void *));
}
for (i = 0; i < sav->sav_count; i++)
nvlist_free(l2cache[i]);
if (sav->sav_count)
kmem_free(l2cache, sav->sav_count * sizeof (void *));
}
static int
load_nvlist(spa_t *spa, uint64_t obj, nvlist_t **value)
{
dmu_buf_t *db;
char *packed = NULL;
size_t nvsize = 0;
int error;
*value = NULL;
error = dmu_bonus_hold(spa->spa_meta_objset, obj, FTAG, &db);
if (error)
return (error);
nvsize = *(uint64_t *)db->db_data;
dmu_buf_rele(db, FTAG);
packed = vmem_alloc(nvsize, KM_SLEEP);
error = dmu_read(spa->spa_meta_objset, obj, 0, nvsize, packed,
DMU_READ_PREFETCH);
if (error == 0)
error = nvlist_unpack(packed, nvsize, value, 0);
vmem_free(packed, nvsize);
return (error);
}
/*
* Concrete top-level vdevs that are not missing and are not logs. At every
* spa_sync we write new uberblocks to at least SPA_SYNC_MIN_VDEVS core tvds.
*/
static uint64_t
spa_healthy_core_tvds(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t tvds = 0;
for (uint64_t i = 0; i < rvd->vdev_children; i++) {
vdev_t *vd = rvd->vdev_child[i];
if (vd->vdev_islog)
continue;
if (vdev_is_concrete(vd) && !vdev_is_dead(vd))
tvds++;
}
return (tvds);
}
/*
* Checks to see if the given vdev could not be opened, in which case we post a
* sysevent to notify the autoreplace code that the device has been removed.
*/
static void
spa_check_removed(vdev_t *vd)
{
for (uint64_t c = 0; c < vd->vdev_children; c++)
spa_check_removed(vd->vdev_child[c]);
if (vd->vdev_ops->vdev_op_leaf && vdev_is_dead(vd) &&
vdev_is_concrete(vd)) {
zfs_post_autoreplace(vd->vdev_spa, vd);
spa_event_notify(vd->vdev_spa, vd, NULL, ESC_ZFS_VDEV_CHECK);
}
}
static int
spa_check_for_missing_logs(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
/*
* If we're doing a normal import, then build up any additional
* diagnostic information about missing log devices.
* We'll pass this up to the user for further processing.
*/
if (!(spa->spa_import_flags & ZFS_IMPORT_MISSING_LOG)) {
nvlist_t **child, *nv;
uint64_t idx = 0;
child = kmem_alloc(rvd->vdev_children * sizeof (nvlist_t *),
KM_SLEEP);
nv = fnvlist_alloc();
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
/*
* We consider a device as missing only if it failed
* to open (i.e. offline or faulted is not considered
* as missing).
*/
if (tvd->vdev_islog &&
tvd->vdev_state == VDEV_STATE_CANT_OPEN) {
child[idx++] = vdev_config_generate(spa, tvd,
B_FALSE, VDEV_CONFIG_MISSING);
}
}
if (idx > 0) {
fnvlist_add_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t * const *)child, idx);
fnvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_MISSING_DEVICES, nv);
for (uint64_t i = 0; i < idx; i++)
nvlist_free(child[i]);
}
nvlist_free(nv);
kmem_free(child, rvd->vdev_children * sizeof (char **));
if (idx > 0) {
spa_load_failed(spa, "some log devices are missing");
vdev_dbgmsg_print_tree(rvd, 2);
return (SET_ERROR(ENXIO));
}
} else {
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_islog &&
tvd->vdev_state == VDEV_STATE_CANT_OPEN) {
spa_set_log_state(spa, SPA_LOG_CLEAR);
spa_load_note(spa, "some log devices are "
"missing, ZIL is dropped.");
vdev_dbgmsg_print_tree(rvd, 2);
break;
}
}
}
return (0);
}
/*
* Check for missing log devices
*/
static boolean_t
spa_check_logs(spa_t *spa)
{
boolean_t rv = B_FALSE;
dsl_pool_t *dp = spa_get_dsl(spa);
switch (spa->spa_log_state) {
default:
break;
case SPA_LOG_MISSING:
/* need to recheck in case slog has been restored */
case SPA_LOG_UNKNOWN:
rv = (dmu_objset_find_dp(dp, dp->dp_root_dir_obj,
zil_check_log_chain, NULL, DS_FIND_CHILDREN) != 0);
if (rv)
spa_set_log_state(spa, SPA_LOG_MISSING);
break;
}
return (rv);
}
/*
* Passivate any log vdevs (note, does not apply to embedded log metaslabs).
*/
static boolean_t
spa_passivate_log(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
boolean_t slog_found = B_FALSE;
ASSERT(spa_config_held(spa, SCL_ALLOC, RW_WRITER));
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_islog) {
ASSERT3P(tvd->vdev_log_mg, ==, NULL);
metaslab_group_passivate(tvd->vdev_mg);
slog_found = B_TRUE;
}
}
return (slog_found);
}
/*
* Activate any log vdevs (note, does not apply to embedded log metaslabs).
*/
static void
spa_activate_log(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(spa_config_held(spa, SCL_ALLOC, RW_WRITER));
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_islog) {
ASSERT3P(tvd->vdev_log_mg, ==, NULL);
metaslab_group_activate(tvd->vdev_mg);
}
}
}
int
spa_reset_logs(spa_t *spa)
{
int error;
error = dmu_objset_find(spa_name(spa), zil_reset,
NULL, DS_FIND_CHILDREN);
if (error == 0) {
/*
* We successfully offlined the log device, sync out the
* current txg so that the "stubby" block can be removed
* by zil_sync().
*/
txg_wait_synced(spa->spa_dsl_pool, 0);
}
return (error);
}
static void
spa_aux_check_removed(spa_aux_vdev_t *sav)
{
for (int i = 0; i < sav->sav_count; i++)
spa_check_removed(sav->sav_vdevs[i]);
}
void
spa_claim_notify(zio_t *zio)
{
spa_t *spa = zio->io_spa;
if (zio->io_error)
return;
mutex_enter(&spa->spa_props_lock); /* any mutex will do */
if (spa->spa_claim_max_txg < zio->io_bp->blk_birth)
spa->spa_claim_max_txg = zio->io_bp->blk_birth;
mutex_exit(&spa->spa_props_lock);
}
typedef struct spa_load_error {
boolean_t sle_verify_data;
uint64_t sle_meta_count;
uint64_t sle_data_count;
} spa_load_error_t;
static void
spa_load_verify_done(zio_t *zio)
{
blkptr_t *bp = zio->io_bp;
spa_load_error_t *sle = zio->io_private;
dmu_object_type_t type = BP_GET_TYPE(bp);
int error = zio->io_error;
spa_t *spa = zio->io_spa;
abd_free(zio->io_abd);
if (error) {
if ((BP_GET_LEVEL(bp) != 0 || DMU_OT_IS_METADATA(type)) &&
type != DMU_OT_INTENT_LOG)
atomic_inc_64(&sle->sle_meta_count);
else
atomic_inc_64(&sle->sle_data_count);
}
mutex_enter(&spa->spa_scrub_lock);
spa->spa_load_verify_bytes -= BP_GET_PSIZE(bp);
cv_broadcast(&spa->spa_scrub_io_cv);
mutex_exit(&spa->spa_scrub_lock);
}
/*
* Maximum number of inflight bytes is the log2 fraction of the arc size.
* By default, we set it to 1/16th of the arc.
*/
static uint_t spa_load_verify_shift = 4;
static int spa_load_verify_metadata = B_TRUE;
static int spa_load_verify_data = B_TRUE;
static int
spa_load_verify_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
const zbookmark_phys_t *zb, const dnode_phys_t *dnp, void *arg)
{
zio_t *rio = arg;
spa_load_error_t *sle = rio->io_private;
(void) zilog, (void) dnp;
/*
* Note: normally this routine will not be called if
* spa_load_verify_metadata is not set. However, it may be useful
* to manually set the flag after the traversal has begun.
*/
if (!spa_load_verify_metadata)
return (0);
/*
* Sanity check the block pointer in order to detect obvious damage
* before using the contents in subsequent checks or in zio_read().
* When damaged consider it to be a metadata error since we cannot
* trust the BP_GET_TYPE and BP_GET_LEVEL values.
*/
if (!zfs_blkptr_verify(spa, bp, B_FALSE, BLK_VERIFY_LOG)) {
atomic_inc_64(&sle->sle_meta_count);
return (0);
}
if (zb->zb_level == ZB_DNODE_LEVEL || BP_IS_HOLE(bp) ||
BP_IS_EMBEDDED(bp) || BP_IS_REDACTED(bp))
return (0);
if (!BP_IS_METADATA(bp) &&
(!spa_load_verify_data || !sle->sle_verify_data))
return (0);
uint64_t maxinflight_bytes =
arc_target_bytes() >> spa_load_verify_shift;
size_t size = BP_GET_PSIZE(bp);
mutex_enter(&spa->spa_scrub_lock);
while (spa->spa_load_verify_bytes >= maxinflight_bytes)
cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock);
spa->spa_load_verify_bytes += size;
mutex_exit(&spa->spa_scrub_lock);
zio_nowait(zio_read(rio, spa, bp, abd_alloc_for_io(size, B_FALSE), size,
spa_load_verify_done, rio->io_private, ZIO_PRIORITY_SCRUB,
ZIO_FLAG_SPECULATIVE | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SCRUB | ZIO_FLAG_RAW, zb));
return (0);
}
static int
verify_dataset_name_len(dsl_pool_t *dp, dsl_dataset_t *ds, void *arg)
{
(void) dp, (void) arg;
if (dsl_dataset_namelen(ds) >= ZFS_MAX_DATASET_NAME_LEN)
return (SET_ERROR(ENAMETOOLONG));
return (0);
}
static int
spa_load_verify(spa_t *spa)
{
zio_t *rio;
spa_load_error_t sle = { 0 };
zpool_load_policy_t policy;
boolean_t verify_ok = B_FALSE;
int error = 0;
zpool_get_load_policy(spa->spa_config, &policy);
if (policy.zlp_rewind & ZPOOL_NEVER_REWIND ||
policy.zlp_maxmeta == UINT64_MAX)
return (0);
dsl_pool_config_enter(spa->spa_dsl_pool, FTAG);
error = dmu_objset_find_dp(spa->spa_dsl_pool,
spa->spa_dsl_pool->dp_root_dir_obj, verify_dataset_name_len, NULL,
DS_FIND_CHILDREN);
dsl_pool_config_exit(spa->spa_dsl_pool, FTAG);
if (error != 0)
return (error);
/*
* Verify data only if we are rewinding or error limit was set.
* Otherwise nothing except dbgmsg care about it to waste time.
*/
sle.sle_verify_data = (policy.zlp_rewind & ZPOOL_REWIND_MASK) ||
(policy.zlp_maxdata < UINT64_MAX);
rio = zio_root(spa, NULL, &sle,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE);
if (spa_load_verify_metadata) {
if (spa->spa_extreme_rewind) {
spa_load_note(spa, "performing a complete scan of the "
"pool since extreme rewind is on. This may take "
"a very long time.\n (spa_load_verify_data=%u, "
"spa_load_verify_metadata=%u)",
spa_load_verify_data, spa_load_verify_metadata);
}
error = traverse_pool(spa, spa->spa_verify_min_txg,
TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA |
TRAVERSE_NO_DECRYPT, spa_load_verify_cb, rio);
}
(void) zio_wait(rio);
ASSERT0(spa->spa_load_verify_bytes);
spa->spa_load_meta_errors = sle.sle_meta_count;
spa->spa_load_data_errors = sle.sle_data_count;
if (sle.sle_meta_count != 0 || sle.sle_data_count != 0) {
spa_load_note(spa, "spa_load_verify found %llu metadata errors "
"and %llu data errors", (u_longlong_t)sle.sle_meta_count,
(u_longlong_t)sle.sle_data_count);
}
if (spa_load_verify_dryrun ||
(!error && sle.sle_meta_count <= policy.zlp_maxmeta &&
sle.sle_data_count <= policy.zlp_maxdata)) {
int64_t loss = 0;
verify_ok = B_TRUE;
spa->spa_load_txg = spa->spa_uberblock.ub_txg;
spa->spa_load_txg_ts = spa->spa_uberblock.ub_timestamp;
loss = spa->spa_last_ubsync_txg_ts - spa->spa_load_txg_ts;
fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_LOAD_TIME,
spa->spa_load_txg_ts);
fnvlist_add_int64(spa->spa_load_info, ZPOOL_CONFIG_REWIND_TIME,
loss);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_LOAD_META_ERRORS, sle.sle_meta_count);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_LOAD_DATA_ERRORS, sle.sle_data_count);
} else {
spa->spa_load_max_txg = spa->spa_uberblock.ub_txg;
}
if (spa_load_verify_dryrun)
return (0);
if (error) {
if (error != ENXIO && error != EIO)
error = SET_ERROR(EIO);
return (error);
}
return (verify_ok ? 0 : EIO);
}
/*
* Find a value in the pool props object.
*/
static void
spa_prop_find(spa_t *spa, zpool_prop_t prop, uint64_t *val)
{
(void) zap_lookup(spa->spa_meta_objset, spa->spa_pool_props_object,
zpool_prop_to_name(prop), sizeof (uint64_t), 1, val);
}
/*
* Find a value in the pool directory object.
*/
static int
spa_dir_prop(spa_t *spa, const char *name, uint64_t *val, boolean_t log_enoent)
{
int error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
name, sizeof (uint64_t), 1, val);
if (error != 0 && (error != ENOENT || log_enoent)) {
spa_load_failed(spa, "couldn't get '%s' value in MOS directory "
"[error=%d]", name, error);
}
return (error);
}
static int
spa_vdev_err(vdev_t *vdev, vdev_aux_t aux, int err)
{
vdev_set_state(vdev, B_TRUE, VDEV_STATE_CANT_OPEN, aux);
return (SET_ERROR(err));
}
boolean_t
spa_livelist_delete_check(spa_t *spa)
{
return (spa->spa_livelists_to_delete != 0);
}
static boolean_t
spa_livelist_delete_cb_check(void *arg, zthr_t *z)
{
(void) z;
spa_t *spa = arg;
return (spa_livelist_delete_check(spa));
}
static int
delete_blkptr_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
spa_t *spa = arg;
zio_free(spa, tx->tx_txg, bp);
dsl_dir_diduse_space(tx->tx_pool->dp_free_dir, DD_USED_HEAD,
-bp_get_dsize_sync(spa, bp),
-BP_GET_PSIZE(bp), -BP_GET_UCSIZE(bp), tx);
return (0);
}
static int
dsl_get_next_livelist_obj(objset_t *os, uint64_t zap_obj, uint64_t *llp)
{
int err;
zap_cursor_t zc;
zap_attribute_t za;
zap_cursor_init(&zc, os, zap_obj);
err = zap_cursor_retrieve(&zc, &za);
zap_cursor_fini(&zc);
if (err == 0)
*llp = za.za_first_integer;
return (err);
}
/*
* Components of livelist deletion that must be performed in syncing
* context: freeing block pointers and updating the pool-wide data
* structures to indicate how much work is left to do
*/
typedef struct sublist_delete_arg {
spa_t *spa;
dsl_deadlist_t *ll;
uint64_t key;
bplist_t *to_free;
} sublist_delete_arg_t;
static void
sublist_delete_sync(void *arg, dmu_tx_t *tx)
{
sublist_delete_arg_t *sda = arg;
spa_t *spa = sda->spa;
dsl_deadlist_t *ll = sda->ll;
uint64_t key = sda->key;
bplist_t *to_free = sda->to_free;
bplist_iterate(to_free, delete_blkptr_cb, spa, tx);
dsl_deadlist_remove_entry(ll, key, tx);
}
typedef struct livelist_delete_arg {
spa_t *spa;
uint64_t ll_obj;
uint64_t zap_obj;
} livelist_delete_arg_t;
static void
livelist_delete_sync(void *arg, dmu_tx_t *tx)
{
livelist_delete_arg_t *lda = arg;
spa_t *spa = lda->spa;
uint64_t ll_obj = lda->ll_obj;
uint64_t zap_obj = lda->zap_obj;
objset_t *mos = spa->spa_meta_objset;
uint64_t count;
/* free the livelist and decrement the feature count */
VERIFY0(zap_remove_int(mos, zap_obj, ll_obj, tx));
dsl_deadlist_free(mos, ll_obj, tx);
spa_feature_decr(spa, SPA_FEATURE_LIVELIST, tx);
VERIFY0(zap_count(mos, zap_obj, &count));
if (count == 0) {
/* no more livelists to delete */
VERIFY0(zap_remove(mos, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_DELETED_CLONES, tx));
VERIFY0(zap_destroy(mos, zap_obj, tx));
spa->spa_livelists_to_delete = 0;
spa_notify_waiters(spa);
}
}
/*
* Load in the value for the livelist to be removed and open it. Then,
* load its first sublist and determine which block pointers should actually
* be freed. Then, call a synctask which performs the actual frees and updates
* the pool-wide livelist data.
*/
static void
spa_livelist_delete_cb(void *arg, zthr_t *z)
{
spa_t *spa = arg;
uint64_t ll_obj = 0, count;
objset_t *mos = spa->spa_meta_objset;
uint64_t zap_obj = spa->spa_livelists_to_delete;
/*
* Determine the next livelist to delete. This function should only
* be called if there is at least one deleted clone.
*/
VERIFY0(dsl_get_next_livelist_obj(mos, zap_obj, &ll_obj));
VERIFY0(zap_count(mos, ll_obj, &count));
if (count > 0) {
dsl_deadlist_t *ll;
dsl_deadlist_entry_t *dle;
bplist_t to_free;
ll = kmem_zalloc(sizeof (dsl_deadlist_t), KM_SLEEP);
dsl_deadlist_open(ll, mos, ll_obj);
dle = dsl_deadlist_first(ll);
ASSERT3P(dle, !=, NULL);
bplist_create(&to_free);
int err = dsl_process_sub_livelist(&dle->dle_bpobj, &to_free,
z, NULL);
if (err == 0) {
sublist_delete_arg_t sync_arg = {
.spa = spa,
.ll = ll,
.key = dle->dle_mintxg,
.to_free = &to_free
};
zfs_dbgmsg("deleting sublist (id %llu) from"
" livelist %llu, %lld remaining",
(u_longlong_t)dle->dle_bpobj.bpo_object,
(u_longlong_t)ll_obj, (longlong_t)count - 1);
VERIFY0(dsl_sync_task(spa_name(spa), NULL,
sublist_delete_sync, &sync_arg, 0,
ZFS_SPACE_CHECK_DESTROY));
} else {
VERIFY3U(err, ==, EINTR);
}
bplist_clear(&to_free);
bplist_destroy(&to_free);
dsl_deadlist_close(ll);
kmem_free(ll, sizeof (dsl_deadlist_t));
} else {
livelist_delete_arg_t sync_arg = {
.spa = spa,
.ll_obj = ll_obj,
.zap_obj = zap_obj
};
zfs_dbgmsg("deletion of livelist %llu completed",
(u_longlong_t)ll_obj);
VERIFY0(dsl_sync_task(spa_name(spa), NULL, livelist_delete_sync,
&sync_arg, 0, ZFS_SPACE_CHECK_DESTROY));
}
}
static void
spa_start_livelist_destroy_thread(spa_t *spa)
{
ASSERT3P(spa->spa_livelist_delete_zthr, ==, NULL);
spa->spa_livelist_delete_zthr =
zthr_create("z_livelist_destroy",
spa_livelist_delete_cb_check, spa_livelist_delete_cb, spa,
minclsyspri);
}
typedef struct livelist_new_arg {
bplist_t *allocs;
bplist_t *frees;
} livelist_new_arg_t;
static int
livelist_track_new_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
ASSERT(tx == NULL);
livelist_new_arg_t *lna = arg;
if (bp_freed) {
bplist_append(lna->frees, bp);
} else {
bplist_append(lna->allocs, bp);
zfs_livelist_condense_new_alloc++;
}
return (0);
}
typedef struct livelist_condense_arg {
spa_t *spa;
bplist_t to_keep;
uint64_t first_size;
uint64_t next_size;
} livelist_condense_arg_t;
static void
spa_livelist_condense_sync(void *arg, dmu_tx_t *tx)
{
livelist_condense_arg_t *lca = arg;
spa_t *spa = lca->spa;
bplist_t new_frees;
dsl_dataset_t *ds = spa->spa_to_condense.ds;
/* Have we been cancelled? */
if (spa->spa_to_condense.cancelled) {
zfs_livelist_condense_sync_cancel++;
goto out;
}
dsl_deadlist_entry_t *first = spa->spa_to_condense.first;
dsl_deadlist_entry_t *next = spa->spa_to_condense.next;
dsl_deadlist_t *ll = &ds->ds_dir->dd_livelist;
/*
* It's possible that the livelist was changed while the zthr was
* running. Therefore, we need to check for new blkptrs in the two
* entries being condensed and continue to track them in the livelist.
* Because of the way we handle remapped blkptrs (see dbuf_remap_impl),
* it's possible that the newly added blkptrs are FREEs or ALLOCs so
* we need to sort them into two different bplists.
*/
uint64_t first_obj = first->dle_bpobj.bpo_object;
uint64_t next_obj = next->dle_bpobj.bpo_object;
uint64_t cur_first_size = first->dle_bpobj.bpo_phys->bpo_num_blkptrs;
uint64_t cur_next_size = next->dle_bpobj.bpo_phys->bpo_num_blkptrs;
bplist_create(&new_frees);
livelist_new_arg_t new_bps = {
.allocs = &lca->to_keep,
.frees = &new_frees,
};
if (cur_first_size > lca->first_size) {
VERIFY0(livelist_bpobj_iterate_from_nofree(&first->dle_bpobj,
livelist_track_new_cb, &new_bps, lca->first_size));
}
if (cur_next_size > lca->next_size) {
VERIFY0(livelist_bpobj_iterate_from_nofree(&next->dle_bpobj,
livelist_track_new_cb, &new_bps, lca->next_size));
}
dsl_deadlist_clear_entry(first, ll, tx);
ASSERT(bpobj_is_empty(&first->dle_bpobj));
dsl_deadlist_remove_entry(ll, next->dle_mintxg, tx);
bplist_iterate(&lca->to_keep, dsl_deadlist_insert_alloc_cb, ll, tx);
bplist_iterate(&new_frees, dsl_deadlist_insert_free_cb, ll, tx);
bplist_destroy(&new_frees);
char dsname[ZFS_MAX_DATASET_NAME_LEN];
dsl_dataset_name(ds, dsname);
zfs_dbgmsg("txg %llu condensing livelist of %s (id %llu), bpobj %llu "
"(%llu blkptrs) and bpobj %llu (%llu blkptrs) -> bpobj %llu "
"(%llu blkptrs)", (u_longlong_t)tx->tx_txg, dsname,
(u_longlong_t)ds->ds_object, (u_longlong_t)first_obj,
(u_longlong_t)cur_first_size, (u_longlong_t)next_obj,
(u_longlong_t)cur_next_size,
(u_longlong_t)first->dle_bpobj.bpo_object,
(u_longlong_t)first->dle_bpobj.bpo_phys->bpo_num_blkptrs);
out:
dmu_buf_rele(ds->ds_dbuf, spa);
spa->spa_to_condense.ds = NULL;
bplist_clear(&lca->to_keep);
bplist_destroy(&lca->to_keep);
kmem_free(lca, sizeof (livelist_condense_arg_t));
spa->spa_to_condense.syncing = B_FALSE;
}
static void
spa_livelist_condense_cb(void *arg, zthr_t *t)
{
while (zfs_livelist_condense_zthr_pause &&
!(zthr_has_waiters(t) || zthr_iscancelled(t)))
delay(1);
spa_t *spa = arg;
dsl_deadlist_entry_t *first = spa->spa_to_condense.first;
dsl_deadlist_entry_t *next = spa->spa_to_condense.next;
uint64_t first_size, next_size;
livelist_condense_arg_t *lca =
kmem_alloc(sizeof (livelist_condense_arg_t), KM_SLEEP);
bplist_create(&lca->to_keep);
/*
* Process the livelists (matching FREEs and ALLOCs) in open context
* so we have minimal work in syncing context to condense.
*
* We save bpobj sizes (first_size and next_size) to use later in
* syncing context to determine if entries were added to these sublists
* while in open context. This is possible because the clone is still
* active and open for normal writes and we want to make sure the new,
* unprocessed blockpointers are inserted into the livelist normally.
*
* Note that dsl_process_sub_livelist() both stores the size number of
* blockpointers and iterates over them while the bpobj's lock held, so
* the sizes returned to us are consistent which what was actually
* processed.
*/
int err = dsl_process_sub_livelist(&first->dle_bpobj, &lca->to_keep, t,
&first_size);
if (err == 0)
err = dsl_process_sub_livelist(&next->dle_bpobj, &lca->to_keep,
t, &next_size);
if (err == 0) {
while (zfs_livelist_condense_sync_pause &&
!(zthr_has_waiters(t) || zthr_iscancelled(t)))
delay(1);
dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
dmu_tx_mark_netfree(tx);
dmu_tx_hold_space(tx, 1);
err = dmu_tx_assign(tx, TXG_NOWAIT | TXG_NOTHROTTLE);
if (err == 0) {
/*
* Prevent the condense zthr restarting before
* the synctask completes.
*/
spa->spa_to_condense.syncing = B_TRUE;
lca->spa = spa;
lca->first_size = first_size;
lca->next_size = next_size;
dsl_sync_task_nowait(spa_get_dsl(spa),
spa_livelist_condense_sync, lca, tx);
dmu_tx_commit(tx);
return;
}
}
/*
* Condensing can not continue: either it was externally stopped or
* we were unable to assign to a tx because the pool has run out of
* space. In the second case, we'll just end up trying to condense
* again in a later txg.
*/
ASSERT(err != 0);
bplist_clear(&lca->to_keep);
bplist_destroy(&lca->to_keep);
kmem_free(lca, sizeof (livelist_condense_arg_t));
dmu_buf_rele(spa->spa_to_condense.ds->ds_dbuf, spa);
spa->spa_to_condense.ds = NULL;
if (err == EINTR)
zfs_livelist_condense_zthr_cancel++;
}
/*
* Check that there is something to condense but that a condense is not
* already in progress and that condensing has not been cancelled.
*/
static boolean_t
spa_livelist_condense_cb_check(void *arg, zthr_t *z)
{
(void) z;
spa_t *spa = arg;
if ((spa->spa_to_condense.ds != NULL) &&
(spa->spa_to_condense.syncing == B_FALSE) &&
(spa->spa_to_condense.cancelled == B_FALSE)) {
return (B_TRUE);
}
return (B_FALSE);
}
static void
spa_start_livelist_condensing_thread(spa_t *spa)
{
spa->spa_to_condense.ds = NULL;
spa->spa_to_condense.first = NULL;
spa->spa_to_condense.next = NULL;
spa->spa_to_condense.syncing = B_FALSE;
spa->spa_to_condense.cancelled = B_FALSE;
ASSERT3P(spa->spa_livelist_condense_zthr, ==, NULL);
spa->spa_livelist_condense_zthr =
zthr_create("z_livelist_condense",
spa_livelist_condense_cb_check,
spa_livelist_condense_cb, spa, minclsyspri);
}
static void
spa_spawn_aux_threads(spa_t *spa)
{
ASSERT(spa_writeable(spa));
ASSERT(MUTEX_HELD(&spa_namespace_lock));
spa_start_indirect_condensing_thread(spa);
spa_start_livelist_destroy_thread(spa);
spa_start_livelist_condensing_thread(spa);
ASSERT3P(spa->spa_checkpoint_discard_zthr, ==, NULL);
spa->spa_checkpoint_discard_zthr =
zthr_create("z_checkpoint_discard",
spa_checkpoint_discard_thread_check,
spa_checkpoint_discard_thread, spa, minclsyspri);
}
/*
* Fix up config after a partly-completed split. This is done with the
* ZPOOL_CONFIG_SPLIT nvlist. Both the splitting pool and the split-off
* pool have that entry in their config, but only the splitting one contains
* a list of all the guids of the vdevs that are being split off.
*
* This function determines what to do with that list: either rejoin
* all the disks to the pool, or complete the splitting process. To attempt
* the rejoin, each disk that is offlined is marked online again, and
* we do a reopen() call. If the vdev label for every disk that was
* marked online indicates it was successfully split off (VDEV_AUX_SPLIT_POOL)
* then we call vdev_split() on each disk, and complete the split.
*
* Otherwise we leave the config alone, with all the vdevs in place in
* the original pool.
*/
static void
spa_try_repair(spa_t *spa, nvlist_t *config)
{
uint_t extracted;
uint64_t *glist;
uint_t i, gcount;
nvlist_t *nvl;
vdev_t **vd;
boolean_t attempt_reopen;
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_SPLIT, &nvl) != 0)
return;
/* check that the config is complete */
if (nvlist_lookup_uint64_array(nvl, ZPOOL_CONFIG_SPLIT_LIST,
&glist, &gcount) != 0)
return;
vd = kmem_zalloc(gcount * sizeof (vdev_t *), KM_SLEEP);
/* attempt to online all the vdevs & validate */
attempt_reopen = B_TRUE;
for (i = 0; i < gcount; i++) {
if (glist[i] == 0) /* vdev is hole */
continue;
vd[i] = spa_lookup_by_guid(spa, glist[i], B_FALSE);
if (vd[i] == NULL) {
/*
* Don't bother attempting to reopen the disks;
* just do the split.
*/
attempt_reopen = B_FALSE;
} else {
/* attempt to re-online it */
vd[i]->vdev_offline = B_FALSE;
}
}
if (attempt_reopen) {
vdev_reopen(spa->spa_root_vdev);
/* check each device to see what state it's in */
for (extracted = 0, i = 0; i < gcount; i++) {
if (vd[i] != NULL &&
vd[i]->vdev_stat.vs_aux != VDEV_AUX_SPLIT_POOL)
break;
++extracted;
}
}
/*
* If every disk has been moved to the new pool, or if we never
* even attempted to look at them, then we split them off for
* good.
*/
if (!attempt_reopen || gcount == extracted) {
for (i = 0; i < gcount; i++)
if (vd[i] != NULL)
vdev_split(vd[i]);
vdev_reopen(spa->spa_root_vdev);
}
kmem_free(vd, gcount * sizeof (vdev_t *));
}
static int
spa_load(spa_t *spa, spa_load_state_t state, spa_import_type_t type)
{
const char *ereport = FM_EREPORT_ZFS_POOL;
int error;
spa->spa_load_state = state;
(void) spa_import_progress_set_state(spa_guid(spa),
spa_load_state(spa));
gethrestime(&spa->spa_loaded_ts);
error = spa_load_impl(spa, type, &ereport);
/*
* Don't count references from objsets that are already closed
* and are making their way through the eviction process.
*/
spa_evicting_os_wait(spa);
spa->spa_minref = zfs_refcount_count(&spa->spa_refcount);
if (error) {
if (error != EEXIST) {
spa->spa_loaded_ts.tv_sec = 0;
spa->spa_loaded_ts.tv_nsec = 0;
}
if (error != EBADF) {
(void) zfs_ereport_post(ereport, spa,
NULL, NULL, NULL, 0);
}
}
spa->spa_load_state = error ? SPA_LOAD_ERROR : SPA_LOAD_NONE;
spa->spa_ena = 0;
(void) spa_import_progress_set_state(spa_guid(spa),
spa_load_state(spa));
return (error);
}
#ifdef ZFS_DEBUG
/*
* Count the number of per-vdev ZAPs associated with all of the vdevs in the
* vdev tree rooted in the given vd, and ensure that each ZAP is present in the
* spa's per-vdev ZAP list.
*/
static uint64_t
vdev_count_verify_zaps(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
uint64_t total = 0;
+ if (spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_AVZ_V2) &&
+ vd->vdev_root_zap != 0) {
+ total++;
+ ASSERT0(zap_lookup_int(spa->spa_meta_objset,
+ spa->spa_all_vdev_zaps, vd->vdev_root_zap));
+ }
if (vd->vdev_top_zap != 0) {
total++;
ASSERT0(zap_lookup_int(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, vd->vdev_top_zap));
}
if (vd->vdev_leaf_zap != 0) {
total++;
ASSERT0(zap_lookup_int(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, vd->vdev_leaf_zap));
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
total += vdev_count_verify_zaps(vd->vdev_child[i]);
}
return (total);
}
#else
#define vdev_count_verify_zaps(vd) ((void) sizeof (vd), 0)
#endif
/*
* Determine whether the activity check is required.
*/
static boolean_t
spa_activity_check_required(spa_t *spa, uberblock_t *ub, nvlist_t *label,
nvlist_t *config)
{
uint64_t state = 0;
uint64_t hostid = 0;
uint64_t tryconfig_txg = 0;
uint64_t tryconfig_timestamp = 0;
uint16_t tryconfig_mmp_seq = 0;
nvlist_t *nvinfo;
if (nvlist_exists(config, ZPOOL_CONFIG_LOAD_INFO)) {
nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO);
(void) nvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_TXG,
&tryconfig_txg);
(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_TIMESTAMP,
&tryconfig_timestamp);
(void) nvlist_lookup_uint16(nvinfo, ZPOOL_CONFIG_MMP_SEQ,
&tryconfig_mmp_seq);
}
(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, &state);
/*
* Disable the MMP activity check - This is used by zdb which
* is intended to be used on potentially active pools.
*/
if (spa->spa_import_flags & ZFS_IMPORT_SKIP_MMP)
return (B_FALSE);
/*
* Skip the activity check when the MMP feature is disabled.
*/
if (ub->ub_mmp_magic == MMP_MAGIC && ub->ub_mmp_delay == 0)
return (B_FALSE);
/*
* If the tryconfig_ values are nonzero, they are the results of an
* earlier tryimport. If they all match the uberblock we just found,
* then the pool has not changed and we return false so we do not test
* a second time.
*/
if (tryconfig_txg && tryconfig_txg == ub->ub_txg &&
tryconfig_timestamp && tryconfig_timestamp == ub->ub_timestamp &&
tryconfig_mmp_seq && tryconfig_mmp_seq ==
(MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0))
return (B_FALSE);
/*
* Allow the activity check to be skipped when importing the pool
* on the same host which last imported it. Since the hostid from
* configuration may be stale use the one read from the label.
*/
if (nvlist_exists(label, ZPOOL_CONFIG_HOSTID))
hostid = fnvlist_lookup_uint64(label, ZPOOL_CONFIG_HOSTID);
if (hostid == spa_get_hostid(spa))
return (B_FALSE);
/*
* Skip the activity test when the pool was cleanly exported.
*/
if (state != POOL_STATE_ACTIVE)
return (B_FALSE);
return (B_TRUE);
}
/*
* Nanoseconds the activity check must watch for changes on-disk.
*/
static uint64_t
spa_activity_check_duration(spa_t *spa, uberblock_t *ub)
{
uint64_t import_intervals = MAX(zfs_multihost_import_intervals, 1);
uint64_t multihost_interval = MSEC2NSEC(
MMP_INTERVAL_OK(zfs_multihost_interval));
uint64_t import_delay = MAX(NANOSEC, import_intervals *
multihost_interval);
/*
* Local tunables determine a minimum duration except for the case
* where we know when the remote host will suspend the pool if MMP
* writes do not land.
*
* See Big Theory comment at the top of mmp.c for the reasoning behind
* these cases and times.
*/
ASSERT(MMP_IMPORT_SAFETY_FACTOR >= 100);
if (MMP_INTERVAL_VALID(ub) && MMP_FAIL_INT_VALID(ub) &&
MMP_FAIL_INT(ub) > 0) {
/* MMP on remote host will suspend pool after failed writes */
import_delay = MMP_FAIL_INT(ub) * MSEC2NSEC(MMP_INTERVAL(ub)) *
MMP_IMPORT_SAFETY_FACTOR / 100;
zfs_dbgmsg("fail_intvals>0 import_delay=%llu ub_mmp "
"mmp_fails=%llu ub_mmp mmp_interval=%llu "
"import_intervals=%llu", (u_longlong_t)import_delay,
(u_longlong_t)MMP_FAIL_INT(ub),
(u_longlong_t)MMP_INTERVAL(ub),
(u_longlong_t)import_intervals);
} else if (MMP_INTERVAL_VALID(ub) && MMP_FAIL_INT_VALID(ub) &&
MMP_FAIL_INT(ub) == 0) {
/* MMP on remote host will never suspend pool */
import_delay = MAX(import_delay, (MSEC2NSEC(MMP_INTERVAL(ub)) +
ub->ub_mmp_delay) * import_intervals);
zfs_dbgmsg("fail_intvals=0 import_delay=%llu ub_mmp "
"mmp_interval=%llu ub_mmp_delay=%llu "
"import_intervals=%llu", (u_longlong_t)import_delay,
(u_longlong_t)MMP_INTERVAL(ub),
(u_longlong_t)ub->ub_mmp_delay,
(u_longlong_t)import_intervals);
} else if (MMP_VALID(ub)) {
/*
* zfs-0.7 compatibility case
*/
import_delay = MAX(import_delay, (multihost_interval +
ub->ub_mmp_delay) * import_intervals);
zfs_dbgmsg("import_delay=%llu ub_mmp_delay=%llu "
"import_intervals=%llu leaves=%u",
(u_longlong_t)import_delay,
(u_longlong_t)ub->ub_mmp_delay,
(u_longlong_t)import_intervals,
vdev_count_leaves(spa));
} else {
/* Using local tunings is the only reasonable option */
zfs_dbgmsg("pool last imported on non-MMP aware "
"host using import_delay=%llu multihost_interval=%llu "
"import_intervals=%llu", (u_longlong_t)import_delay,
(u_longlong_t)multihost_interval,
(u_longlong_t)import_intervals);
}
return (import_delay);
}
/*
* Perform the import activity check. If the user canceled the import or
* we detected activity then fail.
*/
static int
spa_activity_check(spa_t *spa, uberblock_t *ub, nvlist_t *config)
{
uint64_t txg = ub->ub_txg;
uint64_t timestamp = ub->ub_timestamp;
uint64_t mmp_config = ub->ub_mmp_config;
uint16_t mmp_seq = MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0;
uint64_t import_delay;
hrtime_t import_expire;
nvlist_t *mmp_label = NULL;
vdev_t *rvd = spa->spa_root_vdev;
kcondvar_t cv;
kmutex_t mtx;
int error = 0;
cv_init(&cv, NULL, CV_DEFAULT, NULL);
mutex_init(&mtx, NULL, MUTEX_DEFAULT, NULL);
mutex_enter(&mtx);
/*
* If ZPOOL_CONFIG_MMP_TXG is present an activity check was performed
* during the earlier tryimport. If the txg recorded there is 0 then
* the pool is known to be active on another host.
*
* Otherwise, the pool might be in use on another host. Check for
* changes in the uberblocks on disk if necessary.
*/
if (nvlist_exists(config, ZPOOL_CONFIG_LOAD_INFO)) {
nvlist_t *nvinfo = fnvlist_lookup_nvlist(config,
ZPOOL_CONFIG_LOAD_INFO);
if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_TXG) &&
fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_TXG) == 0) {
vdev_uberblock_load(rvd, ub, &mmp_label);
error = SET_ERROR(EREMOTEIO);
goto out;
}
}
import_delay = spa_activity_check_duration(spa, ub);
/* Add a small random factor in case of simultaneous imports (0-25%) */
import_delay += import_delay * random_in_range(250) / 1000;
import_expire = gethrtime() + import_delay;
while (gethrtime() < import_expire) {
(void) spa_import_progress_set_mmp_check(spa_guid(spa),
NSEC2SEC(import_expire - gethrtime()));
vdev_uberblock_load(rvd, ub, &mmp_label);
if (txg != ub->ub_txg || timestamp != ub->ub_timestamp ||
mmp_seq != (MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0)) {
zfs_dbgmsg("multihost activity detected "
"txg %llu ub_txg %llu "
"timestamp %llu ub_timestamp %llu "
"mmp_config %#llx ub_mmp_config %#llx",
(u_longlong_t)txg, (u_longlong_t)ub->ub_txg,
(u_longlong_t)timestamp,
(u_longlong_t)ub->ub_timestamp,
(u_longlong_t)mmp_config,
(u_longlong_t)ub->ub_mmp_config);
error = SET_ERROR(EREMOTEIO);
break;
}
if (mmp_label) {
nvlist_free(mmp_label);
mmp_label = NULL;
}
error = cv_timedwait_sig(&cv, &mtx, ddi_get_lbolt() + hz);
if (error != -1) {
error = SET_ERROR(EINTR);
break;
}
error = 0;
}
out:
mutex_exit(&mtx);
mutex_destroy(&mtx);
cv_destroy(&cv);
/*
* If the pool is determined to be active store the status in the
* spa->spa_load_info nvlist. If the remote hostname or hostid are
* available from configuration read from disk store them as well.
* This allows 'zpool import' to generate a more useful message.
*
* ZPOOL_CONFIG_MMP_STATE - observed pool status (mandatory)
* ZPOOL_CONFIG_MMP_HOSTNAME - hostname from the active pool
* ZPOOL_CONFIG_MMP_HOSTID - hostid from the active pool
*/
if (error == EREMOTEIO) {
const char *hostname = "<unknown>";
uint64_t hostid = 0;
if (mmp_label) {
if (nvlist_exists(mmp_label, ZPOOL_CONFIG_HOSTNAME)) {
hostname = fnvlist_lookup_string(mmp_label,
ZPOOL_CONFIG_HOSTNAME);
fnvlist_add_string(spa->spa_load_info,
ZPOOL_CONFIG_MMP_HOSTNAME, hostname);
}
if (nvlist_exists(mmp_label, ZPOOL_CONFIG_HOSTID)) {
hostid = fnvlist_lookup_uint64(mmp_label,
ZPOOL_CONFIG_HOSTID);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_HOSTID, hostid);
}
}
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_STATE, MMP_STATE_ACTIVE);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_TXG, 0);
error = spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO);
}
if (mmp_label)
nvlist_free(mmp_label);
return (error);
}
static int
spa_verify_host(spa_t *spa, nvlist_t *mos_config)
{
uint64_t hostid;
const char *hostname;
uint64_t myhostid = 0;
if (!spa_is_root(spa) && nvlist_lookup_uint64(mos_config,
ZPOOL_CONFIG_HOSTID, &hostid) == 0) {
hostname = fnvlist_lookup_string(mos_config,
ZPOOL_CONFIG_HOSTNAME);
myhostid = zone_get_hostid(NULL);
if (hostid != 0 && myhostid != 0 && hostid != myhostid) {
cmn_err(CE_WARN, "pool '%s' could not be "
"loaded as it was last accessed by "
"another system (host: %s hostid: 0x%llx). "
"See: https://openzfs.github.io/openzfs-docs/msg/"
"ZFS-8000-EY",
spa_name(spa), hostname, (u_longlong_t)hostid);
spa_load_failed(spa, "hostid verification failed: pool "
"last accessed by host: %s (hostid: 0x%llx)",
hostname, (u_longlong_t)hostid);
return (SET_ERROR(EBADF));
}
}
return (0);
}
static int
spa_ld_parse_config(spa_t *spa, spa_import_type_t type)
{
int error = 0;
nvlist_t *nvtree, *nvl, *config = spa->spa_config;
int parse;
vdev_t *rvd;
uint64_t pool_guid;
const char *comment;
const char *compatibility;
/*
* Versioning wasn't explicitly added to the label until later, so if
* it's not present treat it as the initial version.
*/
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION,
&spa->spa_ubsync.ub_version) != 0)
spa->spa_ubsync.ub_version = SPA_VERSION_INITIAL;
if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &pool_guid)) {
spa_load_failed(spa, "invalid config provided: '%s' missing",
ZPOOL_CONFIG_POOL_GUID);
return (SET_ERROR(EINVAL));
}
/*
* If we are doing an import, ensure that the pool is not already
* imported by checking if its pool guid already exists in the
* spa namespace.
*
* The only case that we allow an already imported pool to be
* imported again, is when the pool is checkpointed and we want to
* look at its checkpointed state from userland tools like zdb.
*/
#ifdef _KERNEL
if ((spa->spa_load_state == SPA_LOAD_IMPORT ||
spa->spa_load_state == SPA_LOAD_TRYIMPORT) &&
spa_guid_exists(pool_guid, 0)) {
#else
if ((spa->spa_load_state == SPA_LOAD_IMPORT ||
spa->spa_load_state == SPA_LOAD_TRYIMPORT) &&
spa_guid_exists(pool_guid, 0) &&
!spa_importing_readonly_checkpoint(spa)) {
#endif
spa_load_failed(spa, "a pool with guid %llu is already open",
(u_longlong_t)pool_guid);
return (SET_ERROR(EEXIST));
}
spa->spa_config_guid = pool_guid;
nvlist_free(spa->spa_load_info);
spa->spa_load_info = fnvlist_alloc();
ASSERT(spa->spa_comment == NULL);
if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMMENT, &comment) == 0)
spa->spa_comment = spa_strdup(comment);
ASSERT(spa->spa_compatibility == NULL);
if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMPATIBILITY,
&compatibility) == 0)
spa->spa_compatibility = spa_strdup(compatibility);
(void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_TXG,
&spa->spa_config_txg);
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_SPLIT, &nvl) == 0)
spa->spa_config_splitting = fnvlist_dup(nvl);
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvtree)) {
spa_load_failed(spa, "invalid config provided: '%s' missing",
ZPOOL_CONFIG_VDEV_TREE);
return (SET_ERROR(EINVAL));
}
/*
* Create "The Godfather" zio to hold all async IOs
*/
spa->spa_async_zio_root = kmem_alloc(max_ncpus * sizeof (void *),
KM_SLEEP);
for (int i = 0; i < max_ncpus; i++) {
spa->spa_async_zio_root[i] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE |
ZIO_FLAG_GODFATHER);
}
/*
* Parse the configuration into a vdev tree. We explicitly set the
* value that will be returned by spa_version() since parsing the
* configuration requires knowing the version number.
*/
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
parse = (type == SPA_IMPORT_EXISTING ?
VDEV_ALLOC_LOAD : VDEV_ALLOC_SPLIT);
error = spa_config_parse(spa, &rvd, nvtree, NULL, 0, parse);
spa_config_exit(spa, SCL_ALL, FTAG);
if (error != 0) {
spa_load_failed(spa, "unable to parse config [error=%d]",
error);
return (error);
}
ASSERT(spa->spa_root_vdev == rvd);
ASSERT3U(spa->spa_min_ashift, >=, SPA_MINBLOCKSHIFT);
ASSERT3U(spa->spa_max_ashift, <=, SPA_MAXBLOCKSHIFT);
if (type != SPA_IMPORT_ASSEMBLE) {
ASSERT(spa_guid(spa) == pool_guid);
}
return (0);
}
/*
* Recursively open all vdevs in the vdev tree. This function is called twice:
* first with the untrusted config, then with the trusted config.
*/
static int
spa_ld_open_vdevs(spa_t *spa)
{
int error = 0;
/*
* spa_missing_tvds_allowed defines how many top-level vdevs can be
* missing/unopenable for the root vdev to be still considered openable.
*/
if (spa->spa_trust_config) {
spa->spa_missing_tvds_allowed = zfs_max_missing_tvds;
} else if (spa->spa_config_source == SPA_CONFIG_SRC_CACHEFILE) {
spa->spa_missing_tvds_allowed = zfs_max_missing_tvds_cachefile;
} else if (spa->spa_config_source == SPA_CONFIG_SRC_SCAN) {
spa->spa_missing_tvds_allowed = zfs_max_missing_tvds_scan;
} else {
spa->spa_missing_tvds_allowed = 0;
}
spa->spa_missing_tvds_allowed =
MAX(zfs_max_missing_tvds, spa->spa_missing_tvds_allowed);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
error = vdev_open(spa->spa_root_vdev);
spa_config_exit(spa, SCL_ALL, FTAG);
if (spa->spa_missing_tvds != 0) {
spa_load_note(spa, "vdev tree has %lld missing top-level "
"vdevs.", (u_longlong_t)spa->spa_missing_tvds);
if (spa->spa_trust_config && (spa->spa_mode & SPA_MODE_WRITE)) {
/*
* Although theoretically we could allow users to open
* incomplete pools in RW mode, we'd need to add a lot
* of extra logic (e.g. adjust pool space to account
* for missing vdevs).
* This limitation also prevents users from accidentally
* opening the pool in RW mode during data recovery and
* damaging it further.
*/
spa_load_note(spa, "pools with missing top-level "
"vdevs can only be opened in read-only mode.");
error = SET_ERROR(ENXIO);
} else {
spa_load_note(spa, "current settings allow for maximum "
"%lld missing top-level vdevs at this stage.",
(u_longlong_t)spa->spa_missing_tvds_allowed);
}
}
if (error != 0) {
spa_load_failed(spa, "unable to open vdev tree [error=%d]",
error);
}
if (spa->spa_missing_tvds != 0 || error != 0)
vdev_dbgmsg_print_tree(spa->spa_root_vdev, 2);
return (error);
}
/*
* We need to validate the vdev labels against the configuration that
* we have in hand. This function is called twice: first with an untrusted
* config, then with a trusted config. The validation is more strict when the
* config is trusted.
*/
static int
spa_ld_validate_vdevs(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
error = vdev_validate(rvd);
spa_config_exit(spa, SCL_ALL, FTAG);
if (error != 0) {
spa_load_failed(spa, "vdev_validate failed [error=%d]", error);
return (error);
}
if (rvd->vdev_state <= VDEV_STATE_CANT_OPEN) {
spa_load_failed(spa, "cannot open vdev tree after invalidating "
"some vdevs");
vdev_dbgmsg_print_tree(rvd, 2);
return (SET_ERROR(ENXIO));
}
return (0);
}
static void
spa_ld_select_uberblock_done(spa_t *spa, uberblock_t *ub)
{
spa->spa_state = POOL_STATE_ACTIVE;
spa->spa_ubsync = spa->spa_uberblock;
spa->spa_verify_min_txg = spa->spa_extreme_rewind ?
TXG_INITIAL - 1 : spa_last_synced_txg(spa) - TXG_DEFER_SIZE - 1;
spa->spa_first_txg = spa->spa_last_ubsync_txg ?
spa->spa_last_ubsync_txg : spa_last_synced_txg(spa) + 1;
spa->spa_claim_max_txg = spa->spa_first_txg;
spa->spa_prev_software_version = ub->ub_software_version;
}
static int
spa_ld_select_uberblock(spa_t *spa, spa_import_type_t type)
{
vdev_t *rvd = spa->spa_root_vdev;
nvlist_t *label;
uberblock_t *ub = &spa->spa_uberblock;
boolean_t activity_check = B_FALSE;
/*
* If we are opening the checkpointed state of the pool by
* rewinding to it, at this point we will have written the
* checkpointed uberblock to the vdev labels, so searching
* the labels will find the right uberblock. However, if
* we are opening the checkpointed state read-only, we have
* not modified the labels. Therefore, we must ignore the
* labels and continue using the spa_uberblock that was set
* by spa_ld_checkpoint_rewind.
*
* Note that it would be fine to ignore the labels when
* rewinding (opening writeable) as well. However, if we
* crash just after writing the labels, we will end up
* searching the labels. Doing so in the common case means
* that this code path gets exercised normally, rather than
* just in the edge case.
*/
if (ub->ub_checkpoint_txg != 0 &&
spa_importing_readonly_checkpoint(spa)) {
spa_ld_select_uberblock_done(spa, ub);
return (0);
}
/*
* Find the best uberblock.
*/
vdev_uberblock_load(rvd, ub, &label);
/*
* If we weren't able to find a single valid uberblock, return failure.
*/
if (ub->ub_txg == 0) {
nvlist_free(label);
spa_load_failed(spa, "no valid uberblock found");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, ENXIO));
}
if (spa->spa_load_max_txg != UINT64_MAX) {
(void) spa_import_progress_set_max_txg(spa_guid(spa),
(u_longlong_t)spa->spa_load_max_txg);
}
spa_load_note(spa, "using uberblock with txg=%llu",
(u_longlong_t)ub->ub_txg);
/*
* For pools which have the multihost property on determine if the
* pool is truly inactive and can be safely imported. Prevent
* hosts which don't have a hostid set from importing the pool.
*/
activity_check = spa_activity_check_required(spa, ub, label,
spa->spa_config);
if (activity_check) {
if (ub->ub_mmp_magic == MMP_MAGIC && ub->ub_mmp_delay &&
spa_get_hostid(spa) == 0) {
nvlist_free(label);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_STATE, MMP_STATE_NO_HOSTID);
return (spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO));
}
int error = spa_activity_check(spa, ub, spa->spa_config);
if (error) {
nvlist_free(label);
return (error);
}
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_STATE, MMP_STATE_INACTIVE);
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_TXG, ub->ub_txg);
fnvlist_add_uint16(spa->spa_load_info,
ZPOOL_CONFIG_MMP_SEQ,
(MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0));
}
/*
* If the pool has an unsupported version we can't open it.
*/
if (!SPA_VERSION_IS_SUPPORTED(ub->ub_version)) {
nvlist_free(label);
spa_load_failed(spa, "version %llu is not supported",
(u_longlong_t)ub->ub_version);
return (spa_vdev_err(rvd, VDEV_AUX_VERSION_NEWER, ENOTSUP));
}
if (ub->ub_version >= SPA_VERSION_FEATURES) {
nvlist_t *features;
/*
* If we weren't able to find what's necessary for reading the
* MOS in the label, return failure.
*/
if (label == NULL) {
spa_load_failed(spa, "label config unavailable");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA,
ENXIO));
}
if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_FEATURES_FOR_READ,
&features) != 0) {
nvlist_free(label);
spa_load_failed(spa, "invalid label: '%s' missing",
ZPOOL_CONFIG_FEATURES_FOR_READ);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA,
ENXIO));
}
/*
* Update our in-core representation with the definitive values
* from the label.
*/
nvlist_free(spa->spa_label_features);
spa->spa_label_features = fnvlist_dup(features);
}
nvlist_free(label);
/*
* Look through entries in the label nvlist's features_for_read. If
* there is a feature listed there which we don't understand then we
* cannot open a pool.
*/
if (ub->ub_version >= SPA_VERSION_FEATURES) {
nvlist_t *unsup_feat;
unsup_feat = fnvlist_alloc();
for (nvpair_t *nvp = nvlist_next_nvpair(spa->spa_label_features,
NULL); nvp != NULL;
nvp = nvlist_next_nvpair(spa->spa_label_features, nvp)) {
if (!zfeature_is_supported(nvpair_name(nvp))) {
fnvlist_add_string(unsup_feat,
nvpair_name(nvp), "");
}
}
if (!nvlist_empty(unsup_feat)) {
fnvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_UNSUP_FEAT, unsup_feat);
nvlist_free(unsup_feat);
spa_load_failed(spa, "some features are unsupported");
return (spa_vdev_err(rvd, VDEV_AUX_UNSUP_FEAT,
ENOTSUP));
}
nvlist_free(unsup_feat);
}
if (type != SPA_IMPORT_ASSEMBLE && spa->spa_config_splitting) {
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_try_repair(spa, spa->spa_config);
spa_config_exit(spa, SCL_ALL, FTAG);
nvlist_free(spa->spa_config_splitting);
spa->spa_config_splitting = NULL;
}
/*
* Initialize internal SPA structures.
*/
spa_ld_select_uberblock_done(spa, ub);
return (0);
}
static int
spa_ld_open_rootbp(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
error = dsl_pool_init(spa, spa->spa_first_txg, &spa->spa_dsl_pool);
if (error != 0) {
spa_load_failed(spa, "unable to open rootbp in dsl_pool_init "
"[error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
spa->spa_meta_objset = spa->spa_dsl_pool->dp_meta_objset;
return (0);
}
static int
spa_ld_trusted_config(spa_t *spa, spa_import_type_t type,
boolean_t reloading)
{
vdev_t *mrvd, *rvd = spa->spa_root_vdev;
nvlist_t *nv, *mos_config, *policy;
int error = 0, copy_error;
uint64_t healthy_tvds, healthy_tvds_mos;
uint64_t mos_config_txg;
if (spa_dir_prop(spa, DMU_POOL_CONFIG, &spa->spa_config_object, B_TRUE)
!= 0)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* If we're assembling a pool from a split, the config provided is
* already trusted so there is nothing to do.
*/
if (type == SPA_IMPORT_ASSEMBLE)
return (0);
healthy_tvds = spa_healthy_core_tvds(spa);
if (load_nvlist(spa, spa->spa_config_object, &mos_config)
!= 0) {
spa_load_failed(spa, "unable to retrieve MOS config");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
/*
* If we are doing an open, pool owner wasn't verified yet, thus do
* the verification here.
*/
if (spa->spa_load_state == SPA_LOAD_OPEN) {
error = spa_verify_host(spa, mos_config);
if (error != 0) {
nvlist_free(mos_config);
return (error);
}
}
nv = fnvlist_lookup_nvlist(mos_config, ZPOOL_CONFIG_VDEV_TREE);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
/*
* Build a new vdev tree from the trusted config
*/
error = spa_config_parse(spa, &mrvd, nv, NULL, 0, VDEV_ALLOC_LOAD);
if (error != 0) {
nvlist_free(mos_config);
spa_config_exit(spa, SCL_ALL, FTAG);
spa_load_failed(spa, "spa_config_parse failed [error=%d]",
error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
}
/*
* Vdev paths in the MOS may be obsolete. If the untrusted config was
* obtained by scanning /dev/dsk, then it will have the right vdev
* paths. We update the trusted MOS config with this information.
* We first try to copy the paths with vdev_copy_path_strict, which
* succeeds only when both configs have exactly the same vdev tree.
* If that fails, we fall back to a more flexible method that has a
* best effort policy.
*/
copy_error = vdev_copy_path_strict(rvd, mrvd);
if (copy_error != 0 || spa_load_print_vdev_tree) {
spa_load_note(spa, "provided vdev tree:");
vdev_dbgmsg_print_tree(rvd, 2);
spa_load_note(spa, "MOS vdev tree:");
vdev_dbgmsg_print_tree(mrvd, 2);
}
if (copy_error != 0) {
spa_load_note(spa, "vdev_copy_path_strict failed, falling "
"back to vdev_copy_path_relaxed");
vdev_copy_path_relaxed(rvd, mrvd);
}
vdev_close(rvd);
vdev_free(rvd);
spa->spa_root_vdev = mrvd;
rvd = mrvd;
spa_config_exit(spa, SCL_ALL, FTAG);
/*
* We will use spa_config if we decide to reload the spa or if spa_load
* fails and we rewind. We must thus regenerate the config using the
* MOS information with the updated paths. ZPOOL_LOAD_POLICY is used to
* pass settings on how to load the pool and is not stored in the MOS.
* We copy it over to our new, trusted config.
*/
mos_config_txg = fnvlist_lookup_uint64(mos_config,
ZPOOL_CONFIG_POOL_TXG);
nvlist_free(mos_config);
mos_config = spa_config_generate(spa, NULL, mos_config_txg, B_FALSE);
if (nvlist_lookup_nvlist(spa->spa_config, ZPOOL_LOAD_POLICY,
&policy) == 0)
fnvlist_add_nvlist(mos_config, ZPOOL_LOAD_POLICY, policy);
spa_config_set(spa, mos_config);
spa->spa_config_source = SPA_CONFIG_SRC_MOS;
/*
* Now that we got the config from the MOS, we should be more strict
* in checking blkptrs and can make assumptions about the consistency
* of the vdev tree. spa_trust_config must be set to true before opening
* vdevs in order for them to be writeable.
*/
spa->spa_trust_config = B_TRUE;
/*
* Open and validate the new vdev tree
*/
error = spa_ld_open_vdevs(spa);
if (error != 0)
return (error);
error = spa_ld_validate_vdevs(spa);
if (error != 0)
return (error);
if (copy_error != 0 || spa_load_print_vdev_tree) {
spa_load_note(spa, "final vdev tree:");
vdev_dbgmsg_print_tree(rvd, 2);
}
if (spa->spa_load_state != SPA_LOAD_TRYIMPORT &&
!spa->spa_extreme_rewind && zfs_max_missing_tvds == 0) {
/*
* Sanity check to make sure that we are indeed loading the
* latest uberblock. If we missed SPA_SYNC_MIN_VDEVS tvds
* in the config provided and they happened to be the only ones
* to have the latest uberblock, we could involuntarily perform
* an extreme rewind.
*/
healthy_tvds_mos = spa_healthy_core_tvds(spa);
if (healthy_tvds_mos - healthy_tvds >=
SPA_SYNC_MIN_VDEVS) {
spa_load_note(spa, "config provided misses too many "
"top-level vdevs compared to MOS (%lld vs %lld). ",
(u_longlong_t)healthy_tvds,
(u_longlong_t)healthy_tvds_mos);
spa_load_note(spa, "vdev tree:");
vdev_dbgmsg_print_tree(rvd, 2);
if (reloading) {
spa_load_failed(spa, "config was already "
"provided from MOS. Aborting.");
return (spa_vdev_err(rvd,
VDEV_AUX_CORRUPT_DATA, EIO));
}
spa_load_note(spa, "spa must be reloaded using MOS "
"config");
return (SET_ERROR(EAGAIN));
}
}
error = spa_check_for_missing_logs(spa);
if (error != 0)
return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM, ENXIO));
if (rvd->vdev_guid_sum != spa->spa_uberblock.ub_guid_sum) {
spa_load_failed(spa, "uberblock guid sum doesn't match MOS "
"guid sum (%llu != %llu)",
(u_longlong_t)spa->spa_uberblock.ub_guid_sum,
(u_longlong_t)rvd->vdev_guid_sum);
return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM,
ENXIO));
}
return (0);
}
static int
spa_ld_open_indirect_vdev_metadata(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
/*
* Everything that we read before spa_remove_init() must be stored
* on concreted vdevs. Therefore we do this as early as possible.
*/
error = spa_remove_init(spa);
if (error != 0) {
spa_load_failed(spa, "spa_remove_init failed [error=%d]",
error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
/*
* Retrieve information needed to condense indirect vdev mappings.
*/
error = spa_condense_init(spa);
if (error != 0) {
spa_load_failed(spa, "spa_condense_init failed [error=%d]",
error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
}
return (0);
}
static int
spa_ld_check_features(spa_t *spa, boolean_t *missing_feat_writep)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
if (spa_version(spa) >= SPA_VERSION_FEATURES) {
boolean_t missing_feat_read = B_FALSE;
nvlist_t *unsup_feat, *enabled_feat;
if (spa_dir_prop(spa, DMU_POOL_FEATURES_FOR_READ,
&spa->spa_feat_for_read_obj, B_TRUE) != 0) {
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
if (spa_dir_prop(spa, DMU_POOL_FEATURES_FOR_WRITE,
&spa->spa_feat_for_write_obj, B_TRUE) != 0) {
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
if (spa_dir_prop(spa, DMU_POOL_FEATURE_DESCRIPTIONS,
&spa->spa_feat_desc_obj, B_TRUE) != 0) {
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
enabled_feat = fnvlist_alloc();
unsup_feat = fnvlist_alloc();
if (!spa_features_check(spa, B_FALSE,
unsup_feat, enabled_feat))
missing_feat_read = B_TRUE;
if (spa_writeable(spa) ||
spa->spa_load_state == SPA_LOAD_TRYIMPORT) {
if (!spa_features_check(spa, B_TRUE,
unsup_feat, enabled_feat)) {
*missing_feat_writep = B_TRUE;
}
}
fnvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_ENABLED_FEAT, enabled_feat);
if (!nvlist_empty(unsup_feat)) {
fnvlist_add_nvlist(spa->spa_load_info,
ZPOOL_CONFIG_UNSUP_FEAT, unsup_feat);
}
fnvlist_free(enabled_feat);
fnvlist_free(unsup_feat);
if (!missing_feat_read) {
fnvlist_add_boolean(spa->spa_load_info,
ZPOOL_CONFIG_CAN_RDONLY);
}
/*
* If the state is SPA_LOAD_TRYIMPORT, our objective is
* twofold: to determine whether the pool is available for
* import in read-write mode and (if it is not) whether the
* pool is available for import in read-only mode. If the pool
* is available for import in read-write mode, it is displayed
* as available in userland; if it is not available for import
* in read-only mode, it is displayed as unavailable in
* userland. If the pool is available for import in read-only
* mode but not read-write mode, it is displayed as unavailable
* in userland with a special note that the pool is actually
* available for open in read-only mode.
*
* As a result, if the state is SPA_LOAD_TRYIMPORT and we are
* missing a feature for write, we must first determine whether
* the pool can be opened read-only before returning to
* userland in order to know whether to display the
* abovementioned note.
*/
if (missing_feat_read || (*missing_feat_writep &&
spa_writeable(spa))) {
spa_load_failed(spa, "pool uses unsupported features");
return (spa_vdev_err(rvd, VDEV_AUX_UNSUP_FEAT,
ENOTSUP));
}
/*
* Load refcounts for ZFS features from disk into an in-memory
* cache during SPA initialization.
*/
for (spa_feature_t i = 0; i < SPA_FEATURES; i++) {
uint64_t refcount;
error = feature_get_refcount_from_disk(spa,
&spa_feature_table[i], &refcount);
if (error == 0) {
spa->spa_feat_refcount_cache[i] = refcount;
} else if (error == ENOTSUP) {
spa->spa_feat_refcount_cache[i] =
SPA_FEATURE_DISABLED;
} else {
spa_load_failed(spa, "error getting refcount "
"for feature %s [error=%d]",
spa_feature_table[i].fi_guid, error);
return (spa_vdev_err(rvd,
VDEV_AUX_CORRUPT_DATA, EIO));
}
}
}
if (spa_feature_is_active(spa, SPA_FEATURE_ENABLED_TXG)) {
if (spa_dir_prop(spa, DMU_POOL_FEATURE_ENABLED_TXG,
&spa->spa_feat_enabled_txg_obj, B_TRUE) != 0)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
/*
* Encryption was added before bookmark_v2, even though bookmark_v2
* is now a dependency. If this pool has encryption enabled without
* bookmark_v2, trigger an errata message.
*/
if (spa_feature_is_enabled(spa, SPA_FEATURE_ENCRYPTION) &&
!spa_feature_is_enabled(spa, SPA_FEATURE_BOOKMARK_V2)) {
spa->spa_errata = ZPOOL_ERRATA_ZOL_8308_ENCRYPTION;
}
return (0);
}
static int
spa_ld_load_special_directories(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
spa->spa_is_initializing = B_TRUE;
error = dsl_pool_open(spa->spa_dsl_pool);
spa->spa_is_initializing = B_FALSE;
if (error != 0) {
spa_load_failed(spa, "dsl_pool_open failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
return (0);
}
static int
spa_ld_get_props(spa_t *spa)
{
int error = 0;
uint64_t obj;
vdev_t *rvd = spa->spa_root_vdev;
/* Grab the checksum salt from the MOS. */
error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_CHECKSUM_SALT, 1,
sizeof (spa->spa_cksum_salt.zcs_bytes),
spa->spa_cksum_salt.zcs_bytes);
if (error == ENOENT) {
/* Generate a new salt for subsequent use */
(void) random_get_pseudo_bytes(spa->spa_cksum_salt.zcs_bytes,
sizeof (spa->spa_cksum_salt.zcs_bytes));
} else if (error != 0) {
spa_load_failed(spa, "unable to retrieve checksum salt from "
"MOS [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
if (spa_dir_prop(spa, DMU_POOL_SYNC_BPOBJ, &obj, B_TRUE) != 0)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
error = bpobj_open(&spa->spa_deferred_bpobj, spa->spa_meta_objset, obj);
if (error != 0) {
spa_load_failed(spa, "error opening deferred-frees bpobj "
"[error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
/*
* Load the bit that tells us to use the new accounting function
* (raid-z deflation). If we have an older pool, this will not
* be present.
*/
error = spa_dir_prop(spa, DMU_POOL_DEFLATE, &spa->spa_deflate, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
error = spa_dir_prop(spa, DMU_POOL_CREATION_VERSION,
&spa->spa_creation_version, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* Load the persistent error log. If we have an older pool, this will
* not be present.
*/
error = spa_dir_prop(spa, DMU_POOL_ERRLOG_LAST, &spa->spa_errlog_last,
B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
error = spa_dir_prop(spa, DMU_POOL_ERRLOG_SCRUB,
&spa->spa_errlog_scrub, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* Load the livelist deletion field. If a livelist is queued for
* deletion, indicate that in the spa
*/
error = spa_dir_prop(spa, DMU_POOL_DELETED_CLONES,
&spa->spa_livelists_to_delete, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* Load the history object. If we have an older pool, this
* will not be present.
*/
error = spa_dir_prop(spa, DMU_POOL_HISTORY, &spa->spa_history, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
/*
* Load the per-vdev ZAP map. If we have an older pool, this will not
* be present; in this case, defer its creation to a later time to
* avoid dirtying the MOS this early / out of sync context. See
* spa_sync_config_object.
*/
/* The sentinel is only available in the MOS config. */
nvlist_t *mos_config;
if (load_nvlist(spa, spa->spa_config_object, &mos_config) != 0) {
spa_load_failed(spa, "unable to retrieve MOS config");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
error = spa_dir_prop(spa, DMU_POOL_VDEV_ZAP_MAP,
&spa->spa_all_vdev_zaps, B_FALSE);
if (error == ENOENT) {
VERIFY(!nvlist_exists(mos_config,
ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS));
spa->spa_avz_action = AVZ_ACTION_INITIALIZE;
ASSERT0(vdev_count_verify_zaps(spa->spa_root_vdev));
} else if (error != 0) {
nvlist_free(mos_config);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
} else if (!nvlist_exists(mos_config, ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS)) {
/*
* An older version of ZFS overwrote the sentinel value, so
* we have orphaned per-vdev ZAPs in the MOS. Defer their
* destruction to later; see spa_sync_config_object.
*/
spa->spa_avz_action = AVZ_ACTION_DESTROY;
/*
* We're assuming that no vdevs have had their ZAPs created
* before this. Better be sure of it.
*/
ASSERT0(vdev_count_verify_zaps(spa->spa_root_vdev));
}
nvlist_free(mos_config);
spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION);
error = spa_dir_prop(spa, DMU_POOL_PROPS, &spa->spa_pool_props_object,
B_FALSE);
if (error && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
if (error == 0) {
uint64_t autoreplace = 0;
spa_prop_find(spa, ZPOOL_PROP_BOOTFS, &spa->spa_bootfs);
spa_prop_find(spa, ZPOOL_PROP_AUTOREPLACE, &autoreplace);
spa_prop_find(spa, ZPOOL_PROP_DELEGATION, &spa->spa_delegation);
spa_prop_find(spa, ZPOOL_PROP_FAILUREMODE, &spa->spa_failmode);
spa_prop_find(spa, ZPOOL_PROP_AUTOEXPAND, &spa->spa_autoexpand);
spa_prop_find(spa, ZPOOL_PROP_MULTIHOST, &spa->spa_multihost);
spa_prop_find(spa, ZPOOL_PROP_AUTOTRIM, &spa->spa_autotrim);
spa->spa_autoreplace = (autoreplace != 0);
}
/*
* If we are importing a pool with missing top-level vdevs,
* we enforce that the pool doesn't panic or get suspended on
* error since the likelihood of missing data is extremely high.
*/
if (spa->spa_missing_tvds > 0 &&
spa->spa_failmode != ZIO_FAILURE_MODE_CONTINUE &&
spa->spa_load_state != SPA_LOAD_TRYIMPORT) {
spa_load_note(spa, "forcing failmode to 'continue' "
"as some top level vdevs are missing");
spa->spa_failmode = ZIO_FAILURE_MODE_CONTINUE;
}
return (0);
}
static int
spa_ld_open_aux_vdevs(spa_t *spa, spa_import_type_t type)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
/*
* If we're assembling the pool from the split-off vdevs of
* an existing pool, we don't want to attach the spares & cache
* devices.
*/
/*
* Load any hot spares for this pool.
*/
error = spa_dir_prop(spa, DMU_POOL_SPARES, &spa->spa_spares.sav_object,
B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
if (error == 0 && type != SPA_IMPORT_ASSEMBLE) {
ASSERT(spa_version(spa) >= SPA_VERSION_SPARES);
if (load_nvlist(spa, spa->spa_spares.sav_object,
&spa->spa_spares.sav_config) != 0) {
spa_load_failed(spa, "error loading spares nvlist");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_spares(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
} else if (error == 0) {
spa->spa_spares.sav_sync = B_TRUE;
}
/*
* Load any level 2 ARC devices for this pool.
*/
error = spa_dir_prop(spa, DMU_POOL_L2CACHE,
&spa->spa_l2cache.sav_object, B_FALSE);
if (error != 0 && error != ENOENT)
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
if (error == 0 && type != SPA_IMPORT_ASSEMBLE) {
ASSERT(spa_version(spa) >= SPA_VERSION_L2CACHE);
if (load_nvlist(spa, spa->spa_l2cache.sav_object,
&spa->spa_l2cache.sav_config) != 0) {
spa_load_failed(spa, "error loading l2cache nvlist");
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_l2cache(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
} else if (error == 0) {
spa->spa_l2cache.sav_sync = B_TRUE;
}
return (0);
}
static int
spa_ld_load_vdev_metadata(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
/*
* If the 'multihost' property is set, then never allow a pool to
* be imported when the system hostid is zero. The exception to
* this rule is zdb which is always allowed to access pools.
*/
if (spa_multihost(spa) && spa_get_hostid(spa) == 0 &&
(spa->spa_import_flags & ZFS_IMPORT_SKIP_MMP) == 0) {
fnvlist_add_uint64(spa->spa_load_info,
ZPOOL_CONFIG_MMP_STATE, MMP_STATE_NO_HOSTID);
return (spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO));
}
/*
* If the 'autoreplace' property is set, then post a resource notifying
* the ZFS DE that it should not issue any faults for unopenable
* devices. We also iterate over the vdevs, and post a sysevent for any
* unopenable vdevs so that the normal autoreplace handler can take
* over.
*/
if (spa->spa_autoreplace && spa->spa_load_state != SPA_LOAD_TRYIMPORT) {
spa_check_removed(spa->spa_root_vdev);
/*
* For the import case, this is done in spa_import(), because
* at this point we're using the spare definitions from
* the MOS config, not necessarily from the userland config.
*/
if (spa->spa_load_state != SPA_LOAD_IMPORT) {
spa_aux_check_removed(&spa->spa_spares);
spa_aux_check_removed(&spa->spa_l2cache);
}
}
/*
* Load the vdev metadata such as metaslabs, DTLs, spacemap object, etc.
*/
error = vdev_load(rvd);
if (error != 0) {
spa_load_failed(spa, "vdev_load failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
}
error = spa_ld_log_spacemaps(spa);
if (error != 0) {
spa_load_failed(spa, "spa_ld_log_spacemaps failed [error=%d]",
error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error));
}
/*
* Propagate the leaf DTLs we just loaded all the way up the vdev tree.
*/
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
vdev_dtl_reassess(rvd, 0, 0, B_FALSE, B_FALSE);
spa_config_exit(spa, SCL_ALL, FTAG);
return (0);
}
static int
spa_ld_load_dedup_tables(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
error = ddt_load(spa);
if (error != 0) {
spa_load_failed(spa, "ddt_load failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
return (0);
}
static int
spa_ld_load_brt(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
error = brt_load(spa);
if (error != 0) {
spa_load_failed(spa, "brt_load failed [error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO));
}
return (0);
}
static int
spa_ld_verify_logs(spa_t *spa, spa_import_type_t type, const char **ereport)
{
vdev_t *rvd = spa->spa_root_vdev;
if (type != SPA_IMPORT_ASSEMBLE && spa_writeable(spa)) {
boolean_t missing = spa_check_logs(spa);
if (missing) {
if (spa->spa_missing_tvds != 0) {
spa_load_note(spa, "spa_check_logs failed "
"so dropping the logs");
} else {
*ereport = FM_EREPORT_ZFS_LOG_REPLAY;
spa_load_failed(spa, "spa_check_logs failed");
return (spa_vdev_err(rvd, VDEV_AUX_BAD_LOG,
ENXIO));
}
}
}
return (0);
}
static int
spa_ld_verify_pool_data(spa_t *spa)
{
int error = 0;
vdev_t *rvd = spa->spa_root_vdev;
/*
* We've successfully opened the pool, verify that we're ready
* to start pushing transactions.
*/
if (spa->spa_load_state != SPA_LOAD_TRYIMPORT) {
error = spa_load_verify(spa);
if (error != 0) {
spa_load_failed(spa, "spa_load_verify failed "
"[error=%d]", error);
return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA,
error));
}
}
return (0);
}
static void
spa_ld_claim_log_blocks(spa_t *spa)
{
dmu_tx_t *tx;
dsl_pool_t *dp = spa_get_dsl(spa);
/*
* Claim log blocks that haven't been committed yet.
* This must all happen in a single txg.
* Note: spa_claim_max_txg is updated by spa_claim_notify(),
* invoked from zil_claim_log_block()'s i/o done callback.
* Price of rollback is that we abandon the log.
*/
spa->spa_claiming = B_TRUE;
tx = dmu_tx_create_assigned(dp, spa_first_txg(spa));
(void) dmu_objset_find_dp(dp, dp->dp_root_dir_obj,
zil_claim, tx, DS_FIND_CHILDREN);
dmu_tx_commit(tx);
spa->spa_claiming = B_FALSE;
spa_set_log_state(spa, SPA_LOG_GOOD);
}
static void
spa_ld_check_for_config_update(spa_t *spa, uint64_t config_cache_txg,
boolean_t update_config_cache)
{
vdev_t *rvd = spa->spa_root_vdev;
int need_update = B_FALSE;
/*
* If the config cache is stale, or we have uninitialized
* metaslabs (see spa_vdev_add()), then update the config.
*
* If this is a verbatim import, trust the current
* in-core spa_config and update the disk labels.
*/
if (update_config_cache || config_cache_txg != spa->spa_config_txg ||
spa->spa_load_state == SPA_LOAD_IMPORT ||
spa->spa_load_state == SPA_LOAD_RECOVER ||
(spa->spa_import_flags & ZFS_IMPORT_VERBATIM))
need_update = B_TRUE;
for (int c = 0; c < rvd->vdev_children; c++)
if (rvd->vdev_child[c]->vdev_ms_array == 0)
need_update = B_TRUE;
/*
* Update the config cache asynchronously in case we're the
* root pool, in which case the config cache isn't writable yet.
*/
if (need_update)
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
static void
spa_ld_prepare_for_reload(spa_t *spa)
{
spa_mode_t mode = spa->spa_mode;
int async_suspended = spa->spa_async_suspended;
spa_unload(spa);
spa_deactivate(spa);
spa_activate(spa, mode);
/*
* We save the value of spa_async_suspended as it gets reset to 0 by
* spa_unload(). We want to restore it back to the original value before
* returning as we might be calling spa_async_resume() later.
*/
spa->spa_async_suspended = async_suspended;
}
static int
spa_ld_read_checkpoint_txg(spa_t *spa)
{
uberblock_t checkpoint;
int error = 0;
ASSERT0(spa->spa_checkpoint_txg);
ASSERT(MUTEX_HELD(&spa_namespace_lock));
error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ZPOOL_CHECKPOINT, sizeof (uint64_t),
sizeof (uberblock_t) / sizeof (uint64_t), &checkpoint);
if (error == ENOENT)
return (0);
if (error != 0)
return (error);
ASSERT3U(checkpoint.ub_txg, !=, 0);
ASSERT3U(checkpoint.ub_checkpoint_txg, !=, 0);
ASSERT3U(checkpoint.ub_timestamp, !=, 0);
spa->spa_checkpoint_txg = checkpoint.ub_txg;
spa->spa_checkpoint_info.sci_timestamp = checkpoint.ub_timestamp;
return (0);
}
static int
spa_ld_mos_init(spa_t *spa, spa_import_type_t type)
{
int error = 0;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa->spa_config_source != SPA_CONFIG_SRC_NONE);
/*
* Never trust the config that is provided unless we are assembling
* a pool following a split.
* This means don't trust blkptrs and the vdev tree in general. This
* also effectively puts the spa in read-only mode since
* spa_writeable() checks for spa_trust_config to be true.
* We will later load a trusted config from the MOS.
*/
if (type != SPA_IMPORT_ASSEMBLE)
spa->spa_trust_config = B_FALSE;
/*
* Parse the config provided to create a vdev tree.
*/
error = spa_ld_parse_config(spa, type);
if (error != 0)
return (error);
spa_import_progress_add(spa);
/*
* Now that we have the vdev tree, try to open each vdev. This involves
* opening the underlying physical device, retrieving its geometry and
* probing the vdev with a dummy I/O. The state of each vdev will be set
* based on the success of those operations. After this we'll be ready
* to read from the vdevs.
*/
error = spa_ld_open_vdevs(spa);
if (error != 0)
return (error);
/*
* Read the label of each vdev and make sure that the GUIDs stored
* there match the GUIDs in the config provided.
* If we're assembling a new pool that's been split off from an
* existing pool, the labels haven't yet been updated so we skip
* validation for now.
*/
if (type != SPA_IMPORT_ASSEMBLE) {
error = spa_ld_validate_vdevs(spa);
if (error != 0)
return (error);
}
/*
* Read all vdev labels to find the best uberblock (i.e. latest,
* unless spa_load_max_txg is set) and store it in spa_uberblock. We
* get the list of features required to read blkptrs in the MOS from
* the vdev label with the best uberblock and verify that our version
* of zfs supports them all.
*/
error = spa_ld_select_uberblock(spa, type);
if (error != 0)
return (error);
/*
* Pass that uberblock to the dsl_pool layer which will open the root
* blkptr. This blkptr points to the latest version of the MOS and will
* allow us to read its contents.
*/
error = spa_ld_open_rootbp(spa);
if (error != 0)
return (error);
return (0);
}
static int
spa_ld_checkpoint_rewind(spa_t *spa)
{
uberblock_t checkpoint;
int error = 0;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT);
error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ZPOOL_CHECKPOINT, sizeof (uint64_t),
sizeof (uberblock_t) / sizeof (uint64_t), &checkpoint);
if (error != 0) {
spa_load_failed(spa, "unable to retrieve checkpointed "
"uberblock from the MOS config [error=%d]", error);
if (error == ENOENT)
error = ZFS_ERR_NO_CHECKPOINT;
return (error);
}
ASSERT3U(checkpoint.ub_txg, <, spa->spa_uberblock.ub_txg);
ASSERT3U(checkpoint.ub_txg, ==, checkpoint.ub_checkpoint_txg);
/*
* We need to update the txg and timestamp of the checkpointed
* uberblock to be higher than the latest one. This ensures that
* the checkpointed uberblock is selected if we were to close and
* reopen the pool right after we've written it in the vdev labels.
* (also see block comment in vdev_uberblock_compare)
*/
checkpoint.ub_txg = spa->spa_uberblock.ub_txg + 1;
checkpoint.ub_timestamp = gethrestime_sec();
/*
* Set current uberblock to be the checkpointed uberblock.
*/
spa->spa_uberblock = checkpoint;
/*
* If we are doing a normal rewind, then the pool is open for
* writing and we sync the "updated" checkpointed uberblock to
* disk. Once this is done, we've basically rewound the whole
* pool and there is no way back.
*
* There are cases when we don't want to attempt and sync the
* checkpointed uberblock to disk because we are opening a
* pool as read-only. Specifically, verifying the checkpointed
* state with zdb, and importing the checkpointed state to get
* a "preview" of its content.
*/
if (spa_writeable(spa)) {
vdev_t *rvd = spa->spa_root_vdev;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
vdev_t *svd[SPA_SYNC_MIN_VDEVS] = { NULL };
int svdcount = 0;
int children = rvd->vdev_children;
int c0 = random_in_range(children);
for (int c = 0; c < children; c++) {
vdev_t *vd = rvd->vdev_child[(c0 + c) % children];
/* Stop when revisiting the first vdev */
if (c > 0 && svd[0] == vd)
break;
if (vd->vdev_ms_array == 0 || vd->vdev_islog ||
!vdev_is_concrete(vd))
continue;
svd[svdcount++] = vd;
if (svdcount == SPA_SYNC_MIN_VDEVS)
break;
}
error = vdev_config_sync(svd, svdcount, spa->spa_first_txg);
if (error == 0)
spa->spa_last_synced_guid = rvd->vdev_guid;
spa_config_exit(spa, SCL_ALL, FTAG);
if (error != 0) {
spa_load_failed(spa, "failed to write checkpointed "
"uberblock to the vdev labels [error=%d]", error);
return (error);
}
}
return (0);
}
static int
spa_ld_mos_with_trusted_config(spa_t *spa, spa_import_type_t type,
boolean_t *update_config_cache)
{
int error;
/*
* Parse the config for pool, open and validate vdevs,
* select an uberblock, and use that uberblock to open
* the MOS.
*/
error = spa_ld_mos_init(spa, type);
if (error != 0)
return (error);
/*
* Retrieve the trusted config stored in the MOS and use it to create
* a new, exact version of the vdev tree, then reopen all vdevs.
*/
error = spa_ld_trusted_config(spa, type, B_FALSE);
if (error == EAGAIN) {
if (update_config_cache != NULL)
*update_config_cache = B_TRUE;
/*
* Redo the loading process with the trusted config if it is
* too different from the untrusted config.
*/
spa_ld_prepare_for_reload(spa);
spa_load_note(spa, "RELOADING");
error = spa_ld_mos_init(spa, type);
if (error != 0)
return (error);
error = spa_ld_trusted_config(spa, type, B_TRUE);
if (error != 0)
return (error);
} else if (error != 0) {
return (error);
}
return (0);
}
/*
* Load an existing storage pool, using the config provided. This config
* describes which vdevs are part of the pool and is later validated against
* partial configs present in each vdev's label and an entire copy of the
* config stored in the MOS.
*/
static int
spa_load_impl(spa_t *spa, spa_import_type_t type, const char **ereport)
{
int error = 0;
boolean_t missing_feat_write = B_FALSE;
boolean_t checkpoint_rewind =
(spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT);
boolean_t update_config_cache = B_FALSE;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa->spa_config_source != SPA_CONFIG_SRC_NONE);
spa_load_note(spa, "LOADING");
error = spa_ld_mos_with_trusted_config(spa, type, &update_config_cache);
if (error != 0)
return (error);
/*
* If we are rewinding to the checkpoint then we need to repeat
* everything we've done so far in this function but this time
* selecting the checkpointed uberblock and using that to open
* the MOS.
*/
if (checkpoint_rewind) {
/*
* If we are rewinding to the checkpoint update config cache
* anyway.
*/
update_config_cache = B_TRUE;
/*
* Extract the checkpointed uberblock from the current MOS
* and use this as the pool's uberblock from now on. If the
* pool is imported as writeable we also write the checkpoint
* uberblock to the labels, making the rewind permanent.
*/
error = spa_ld_checkpoint_rewind(spa);
if (error != 0)
return (error);
/*
* Redo the loading process again with the
* checkpointed uberblock.
*/
spa_ld_prepare_for_reload(spa);
spa_load_note(spa, "LOADING checkpointed uberblock");
error = spa_ld_mos_with_trusted_config(spa, type, NULL);
if (error != 0)
return (error);
}
/*
* Retrieve the checkpoint txg if the pool has a checkpoint.
*/
error = spa_ld_read_checkpoint_txg(spa);
if (error != 0)
return (error);
/*
* Retrieve the mapping of indirect vdevs. Those vdevs were removed
* from the pool and their contents were re-mapped to other vdevs. Note
* that everything that we read before this step must have been
* rewritten on concrete vdevs after the last device removal was
* initiated. Otherwise we could be reading from indirect vdevs before
* we have loaded their mappings.
*/
error = spa_ld_open_indirect_vdev_metadata(spa);
if (error != 0)
return (error);
/*
* Retrieve the full list of active features from the MOS and check if
* they are all supported.
*/
error = spa_ld_check_features(spa, &missing_feat_write);
if (error != 0)
return (error);
/*
* Load several special directories from the MOS needed by the dsl_pool
* layer.
*/
error = spa_ld_load_special_directories(spa);
if (error != 0)
return (error);
/*
* Retrieve pool properties from the MOS.
*/
error = spa_ld_get_props(spa);
if (error != 0)
return (error);
/*
* Retrieve the list of auxiliary devices - cache devices and spares -
* and open them.
*/
error = spa_ld_open_aux_vdevs(spa, type);
if (error != 0)
return (error);
/*
* Load the metadata for all vdevs. Also check if unopenable devices
* should be autoreplaced.
*/
error = spa_ld_load_vdev_metadata(spa);
if (error != 0)
return (error);
error = spa_ld_load_dedup_tables(spa);
if (error != 0)
return (error);
error = spa_ld_load_brt(spa);
if (error != 0)
return (error);
/*
* Verify the logs now to make sure we don't have any unexpected errors
* when we claim log blocks later.
*/
error = spa_ld_verify_logs(spa, type, ereport);
if (error != 0)
return (error);
if (missing_feat_write) {
ASSERT(spa->spa_load_state == SPA_LOAD_TRYIMPORT);
/*
* At this point, we know that we can open the pool in
* read-only mode but not read-write mode. We now have enough
* information and can return to userland.
*/
return (spa_vdev_err(spa->spa_root_vdev, VDEV_AUX_UNSUP_FEAT,
ENOTSUP));
}
/*
* Traverse the last txgs to make sure the pool was left off in a safe
* state. When performing an extreme rewind, we verify the whole pool,
* which can take a very long time.
*/
error = spa_ld_verify_pool_data(spa);
if (error != 0)
return (error);
/*
* Calculate the deflated space for the pool. This must be done before
* we write anything to the pool because we'd need to update the space
* accounting using the deflated sizes.
*/
spa_update_dspace(spa);
/*
* We have now retrieved all the information we needed to open the
* pool. If we are importing the pool in read-write mode, a few
* additional steps must be performed to finish the import.
*/
if (spa_writeable(spa) && (spa->spa_load_state == SPA_LOAD_RECOVER ||
spa->spa_load_max_txg == UINT64_MAX)) {
uint64_t config_cache_txg = spa->spa_config_txg;
ASSERT(spa->spa_load_state != SPA_LOAD_TRYIMPORT);
/*
* In case of a checkpoint rewind, log the original txg
* of the checkpointed uberblock.
*/
if (checkpoint_rewind) {
spa_history_log_internal(spa, "checkpoint rewind",
NULL, "rewound state to txg=%llu",
(u_longlong_t)spa->spa_uberblock.ub_checkpoint_txg);
}
/*
* Traverse the ZIL and claim all blocks.
*/
spa_ld_claim_log_blocks(spa);
/*
* Kick-off the syncing thread.
*/
spa->spa_sync_on = B_TRUE;
txg_sync_start(spa->spa_dsl_pool);
mmp_thread_start(spa);
/*
* Wait for all claims to sync. We sync up to the highest
* claimed log block birth time so that claimed log blocks
* don't appear to be from the future. spa_claim_max_txg
* will have been set for us by ZIL traversal operations
* performed above.
*/
txg_wait_synced(spa->spa_dsl_pool, spa->spa_claim_max_txg);
/*
* Check if we need to request an update of the config. On the
* next sync, we would update the config stored in vdev labels
* and the cachefile (by default /etc/zfs/zpool.cache).
*/
spa_ld_check_for_config_update(spa, config_cache_txg,
update_config_cache);
/*
* Check if a rebuild was in progress and if so resume it.
* Then check all DTLs to see if anything needs resilvering.
* The resilver will be deferred if a rebuild was started.
*/
if (vdev_rebuild_active(spa->spa_root_vdev)) {
vdev_rebuild_restart(spa);
} else if (!dsl_scan_resilvering(spa->spa_dsl_pool) &&
vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) {
spa_async_request(spa, SPA_ASYNC_RESILVER);
}
/*
* Log the fact that we booted up (so that we can detect if
* we rebooted in the middle of an operation).
*/
spa_history_log_version(spa, "open", NULL);
spa_restart_removal(spa);
spa_spawn_aux_threads(spa);
/*
* Delete any inconsistent datasets.
*
* Note:
* Since we may be issuing deletes for clones here,
* we make sure to do so after we've spawned all the
* auxiliary threads above (from which the livelist
* deletion zthr is part of).
*/
(void) dmu_objset_find(spa_name(spa),
dsl_destroy_inconsistent, NULL, DS_FIND_CHILDREN);
/*
* Clean up any stale temporary dataset userrefs.
*/
dsl_pool_clean_tmp_userrefs(spa->spa_dsl_pool);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_initialize_restart(spa->spa_root_vdev);
vdev_trim_restart(spa->spa_root_vdev);
vdev_autotrim_restart(spa);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
spa_import_progress_remove(spa_guid(spa));
spa_async_request(spa, SPA_ASYNC_L2CACHE_REBUILD);
spa_load_note(spa, "LOADED");
return (0);
}
static int
spa_load_retry(spa_t *spa, spa_load_state_t state)
{
spa_mode_t mode = spa->spa_mode;
spa_unload(spa);
spa_deactivate(spa);
spa->spa_load_max_txg = spa->spa_uberblock.ub_txg - 1;
spa_activate(spa, mode);
spa_async_suspend(spa);
spa_load_note(spa, "spa_load_retry: rewind, max txg: %llu",
(u_longlong_t)spa->spa_load_max_txg);
return (spa_load(spa, state, SPA_IMPORT_EXISTING));
}
/*
* If spa_load() fails this function will try loading prior txg's. If
* 'state' is SPA_LOAD_RECOVER and one of these loads succeeds the pool
* will be rewound to that txg. If 'state' is not SPA_LOAD_RECOVER this
* function will not rewind the pool and will return the same error as
* spa_load().
*/
static int
spa_load_best(spa_t *spa, spa_load_state_t state, uint64_t max_request,
int rewind_flags)
{
nvlist_t *loadinfo = NULL;
nvlist_t *config = NULL;
int load_error, rewind_error;
uint64_t safe_rewind_txg;
uint64_t min_txg;
if (spa->spa_load_txg && state == SPA_LOAD_RECOVER) {
spa->spa_load_max_txg = spa->spa_load_txg;
spa_set_log_state(spa, SPA_LOG_CLEAR);
} else {
spa->spa_load_max_txg = max_request;
if (max_request != UINT64_MAX)
spa->spa_extreme_rewind = B_TRUE;
}
load_error = rewind_error = spa_load(spa, state, SPA_IMPORT_EXISTING);
if (load_error == 0)
return (0);
if (load_error == ZFS_ERR_NO_CHECKPOINT) {
/*
* When attempting checkpoint-rewind on a pool with no
* checkpoint, we should not attempt to load uberblocks
* from previous txgs when spa_load fails.
*/
ASSERT(spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT);
spa_import_progress_remove(spa_guid(spa));
return (load_error);
}
if (spa->spa_root_vdev != NULL)
config = spa_config_generate(spa, NULL, -1ULL, B_TRUE);
spa->spa_last_ubsync_txg = spa->spa_uberblock.ub_txg;
spa->spa_last_ubsync_txg_ts = spa->spa_uberblock.ub_timestamp;
if (rewind_flags & ZPOOL_NEVER_REWIND) {
nvlist_free(config);
spa_import_progress_remove(spa_guid(spa));
return (load_error);
}
if (state == SPA_LOAD_RECOVER) {
/* Price of rolling back is discarding txgs, including log */
spa_set_log_state(spa, SPA_LOG_CLEAR);
} else {
/*
* If we aren't rolling back save the load info from our first
* import attempt so that we can restore it after attempting
* to rewind.
*/
loadinfo = spa->spa_load_info;
spa->spa_load_info = fnvlist_alloc();
}
spa->spa_load_max_txg = spa->spa_last_ubsync_txg;
safe_rewind_txg = spa->spa_last_ubsync_txg - TXG_DEFER_SIZE;
min_txg = (rewind_flags & ZPOOL_EXTREME_REWIND) ?
TXG_INITIAL : safe_rewind_txg;
/*
* Continue as long as we're finding errors, we're still within
* the acceptable rewind range, and we're still finding uberblocks
*/
while (rewind_error && spa->spa_uberblock.ub_txg >= min_txg &&
spa->spa_uberblock.ub_txg <= spa->spa_load_max_txg) {
if (spa->spa_load_max_txg < safe_rewind_txg)
spa->spa_extreme_rewind = B_TRUE;
rewind_error = spa_load_retry(spa, state);
}
spa->spa_extreme_rewind = B_FALSE;
spa->spa_load_max_txg = UINT64_MAX;
if (config && (rewind_error || state != SPA_LOAD_RECOVER))
spa_config_set(spa, config);
else
nvlist_free(config);
if (state == SPA_LOAD_RECOVER) {
ASSERT3P(loadinfo, ==, NULL);
spa_import_progress_remove(spa_guid(spa));
return (rewind_error);
} else {
/* Store the rewind info as part of the initial load info */
fnvlist_add_nvlist(loadinfo, ZPOOL_CONFIG_REWIND_INFO,
spa->spa_load_info);
/* Restore the initial load info */
fnvlist_free(spa->spa_load_info);
spa->spa_load_info = loadinfo;
spa_import_progress_remove(spa_guid(spa));
return (load_error);
}
}
/*
* Pool Open/Import
*
* The import case is identical to an open except that the configuration is sent
* down from userland, instead of grabbed from the configuration cache. For the
* case of an open, the pool configuration will exist in the
* POOL_STATE_UNINITIALIZED state.
*
* The stats information (gen/count/ustats) is used to gather vdev statistics at
* the same time open the pool, without having to keep around the spa_t in some
* ambiguous state.
*/
static int
spa_open_common(const char *pool, spa_t **spapp, const void *tag,
nvlist_t *nvpolicy, nvlist_t **config)
{
spa_t *spa;
spa_load_state_t state = SPA_LOAD_OPEN;
int error;
int locked = B_FALSE;
int firstopen = B_FALSE;
*spapp = NULL;
/*
* As disgusting as this is, we need to support recursive calls to this
* function because dsl_dir_open() is called during spa_load(), and ends
* up calling spa_open() again. The real fix is to figure out how to
* avoid dsl_dir_open() calling this in the first place.
*/
if (MUTEX_NOT_HELD(&spa_namespace_lock)) {
mutex_enter(&spa_namespace_lock);
locked = B_TRUE;
}
if ((spa = spa_lookup(pool)) == NULL) {
if (locked)
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(ENOENT));
}
if (spa->spa_state == POOL_STATE_UNINITIALIZED) {
zpool_load_policy_t policy;
firstopen = B_TRUE;
zpool_get_load_policy(nvpolicy ? nvpolicy : spa->spa_config,
&policy);
if (policy.zlp_rewind & ZPOOL_DO_REWIND)
state = SPA_LOAD_RECOVER;
spa_activate(spa, spa_mode_global);
if (state != SPA_LOAD_RECOVER)
spa->spa_last_ubsync_txg = spa->spa_load_txg = 0;
spa->spa_config_source = SPA_CONFIG_SRC_CACHEFILE;
zfs_dbgmsg("spa_open_common: opening %s", pool);
error = spa_load_best(spa, state, policy.zlp_txg,
policy.zlp_rewind);
if (error == EBADF) {
/*
* If vdev_validate() returns failure (indicated by
* EBADF), it indicates that one of the vdevs indicates
* that the pool has been exported or destroyed. If
* this is the case, the config cache is out of sync and
* we should remove the pool from the namespace.
*/
spa_unload(spa);
spa_deactivate(spa);
spa_write_cachefile(spa, B_TRUE, B_TRUE, B_FALSE);
spa_remove(spa);
if (locked)
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(ENOENT));
}
if (error) {
/*
* We can't open the pool, but we still have useful
* information: the state of each vdev after the
* attempted vdev_open(). Return this to the user.
*/
if (config != NULL && spa->spa_config) {
*config = fnvlist_dup(spa->spa_config);
fnvlist_add_nvlist(*config,
ZPOOL_CONFIG_LOAD_INFO,
spa->spa_load_info);
}
spa_unload(spa);
spa_deactivate(spa);
spa->spa_last_open_failed = error;
if (locked)
mutex_exit(&spa_namespace_lock);
*spapp = NULL;
return (error);
}
}
spa_open_ref(spa, tag);
if (config != NULL)
*config = spa_config_generate(spa, NULL, -1ULL, B_TRUE);
/*
* If we've recovered the pool, pass back any information we
* gathered while doing the load.
*/
if (state == SPA_LOAD_RECOVER && config != NULL) {
fnvlist_add_nvlist(*config, ZPOOL_CONFIG_LOAD_INFO,
spa->spa_load_info);
}
if (locked) {
spa->spa_last_open_failed = 0;
spa->spa_last_ubsync_txg = 0;
spa->spa_load_txg = 0;
mutex_exit(&spa_namespace_lock);
}
if (firstopen)
zvol_create_minors_recursive(spa_name(spa));
*spapp = spa;
return (0);
}
int
spa_open_rewind(const char *name, spa_t **spapp, const void *tag,
nvlist_t *policy, nvlist_t **config)
{
return (spa_open_common(name, spapp, tag, policy, config));
}
int
spa_open(const char *name, spa_t **spapp, const void *tag)
{
return (spa_open_common(name, spapp, tag, NULL, NULL));
}
/*
* Lookup the given spa_t, incrementing the inject count in the process,
* preventing it from being exported or destroyed.
*/
spa_t *
spa_inject_addref(char *name)
{
spa_t *spa;
mutex_enter(&spa_namespace_lock);
if ((spa = spa_lookup(name)) == NULL) {
mutex_exit(&spa_namespace_lock);
return (NULL);
}
spa->spa_inject_ref++;
mutex_exit(&spa_namespace_lock);
return (spa);
}
void
spa_inject_delref(spa_t *spa)
{
mutex_enter(&spa_namespace_lock);
spa->spa_inject_ref--;
mutex_exit(&spa_namespace_lock);
}
/*
* Add spares device information to the nvlist.
*/
static void
spa_add_spares(spa_t *spa, nvlist_t *config)
{
nvlist_t **spares;
uint_t i, nspares;
nvlist_t *nvroot;
uint64_t guid;
vdev_stat_t *vs;
uint_t vsc;
uint64_t pool;
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER));
if (spa->spa_spares.sav_count == 0)
return;
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
VERIFY0(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, &spares, &nspares));
if (nspares != 0) {
fnvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
(const nvlist_t * const *)spares, nspares);
VERIFY0(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares));
/*
* Go through and find any spares which have since been
* repurposed as an active spare. If this is the case, update
* their status appropriately.
*/
for (i = 0; i < nspares; i++) {
guid = fnvlist_lookup_uint64(spares[i],
ZPOOL_CONFIG_GUID);
VERIFY0(nvlist_lookup_uint64_array(spares[i],
ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc));
if (spa_spare_exists(guid, &pool, NULL) &&
pool != 0ULL) {
vs->vs_state = VDEV_STATE_CANT_OPEN;
vs->vs_aux = VDEV_AUX_SPARED;
} else {
vs->vs_state =
spa->spa_spares.sav_vdevs[i]->vdev_state;
}
}
}
}
/*
* Add l2cache device information to the nvlist, including vdev stats.
*/
static void
spa_add_l2cache(spa_t *spa, nvlist_t *config)
{
nvlist_t **l2cache;
uint_t i, j, nl2cache;
nvlist_t *nvroot;
uint64_t guid;
vdev_t *vd;
vdev_stat_t *vs;
uint_t vsc;
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER));
if (spa->spa_l2cache.sav_count == 0)
return;
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
VERIFY0(nvlist_lookup_nvlist_array(spa->spa_l2cache.sav_config,
ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache));
if (nl2cache != 0) {
fnvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
(const nvlist_t * const *)l2cache, nl2cache);
VERIFY0(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache));
/*
* Update level 2 cache device stats.
*/
for (i = 0; i < nl2cache; i++) {
guid = fnvlist_lookup_uint64(l2cache[i],
ZPOOL_CONFIG_GUID);
vd = NULL;
for (j = 0; j < spa->spa_l2cache.sav_count; j++) {
if (guid ==
spa->spa_l2cache.sav_vdevs[j]->vdev_guid) {
vd = spa->spa_l2cache.sav_vdevs[j];
break;
}
}
ASSERT(vd != NULL);
VERIFY0(nvlist_lookup_uint64_array(l2cache[i],
ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc));
vdev_get_stats(vd, vs);
vdev_config_generate_stats(vd, l2cache[i]);
}
}
}
static void
spa_feature_stats_from_disk(spa_t *spa, nvlist_t *features)
{
zap_cursor_t zc;
zap_attribute_t za;
if (spa->spa_feat_for_read_obj != 0) {
for (zap_cursor_init(&zc, spa->spa_meta_objset,
spa->spa_feat_for_read_obj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
ASSERT(za.za_integer_length == sizeof (uint64_t) &&
za.za_num_integers == 1);
VERIFY0(nvlist_add_uint64(features, za.za_name,
za.za_first_integer));
}
zap_cursor_fini(&zc);
}
if (spa->spa_feat_for_write_obj != 0) {
for (zap_cursor_init(&zc, spa->spa_meta_objset,
spa->spa_feat_for_write_obj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
ASSERT(za.za_integer_length == sizeof (uint64_t) &&
za.za_num_integers == 1);
VERIFY0(nvlist_add_uint64(features, za.za_name,
za.za_first_integer));
}
zap_cursor_fini(&zc);
}
}
static void
spa_feature_stats_from_cache(spa_t *spa, nvlist_t *features)
{
int i;
for (i = 0; i < SPA_FEATURES; i++) {
zfeature_info_t feature = spa_feature_table[i];
uint64_t refcount;
if (feature_get_refcount(spa, &feature, &refcount) != 0)
continue;
VERIFY0(nvlist_add_uint64(features, feature.fi_guid, refcount));
}
}
/*
* Store a list of pool features and their reference counts in the
* config.
*
* The first time this is called on a spa, allocate a new nvlist, fetch
* the pool features and reference counts from disk, then save the list
* in the spa. In subsequent calls on the same spa use the saved nvlist
* and refresh its values from the cached reference counts. This
* ensures we don't block here on I/O on a suspended pool so 'zpool
* clear' can resume the pool.
*/
static void
spa_add_feature_stats(spa_t *spa, nvlist_t *config)
{
nvlist_t *features;
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER));
mutex_enter(&spa->spa_feat_stats_lock);
features = spa->spa_feat_stats;
if (features != NULL) {
spa_feature_stats_from_cache(spa, features);
} else {
VERIFY0(nvlist_alloc(&features, NV_UNIQUE_NAME, KM_SLEEP));
spa->spa_feat_stats = features;
spa_feature_stats_from_disk(spa, features);
}
VERIFY0(nvlist_add_nvlist(config, ZPOOL_CONFIG_FEATURE_STATS,
features));
mutex_exit(&spa->spa_feat_stats_lock);
}
int
spa_get_stats(const char *name, nvlist_t **config,
char *altroot, size_t buflen)
{
int error;
spa_t *spa;
*config = NULL;
error = spa_open_common(name, &spa, FTAG, NULL, config);
if (spa != NULL) {
/*
* This still leaves a window of inconsistency where the spares
* or l2cache devices could change and the config would be
* self-inconsistent.
*/
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
if (*config != NULL) {
uint64_t loadtimes[2];
loadtimes[0] = spa->spa_loaded_ts.tv_sec;
loadtimes[1] = spa->spa_loaded_ts.tv_nsec;
fnvlist_add_uint64_array(*config,
ZPOOL_CONFIG_LOADED_TIME, loadtimes, 2);
fnvlist_add_uint64(*config,
ZPOOL_CONFIG_ERRCOUNT,
spa_approx_errlog_size(spa));
if (spa_suspended(spa)) {
fnvlist_add_uint64(*config,
ZPOOL_CONFIG_SUSPENDED,
spa->spa_failmode);
fnvlist_add_uint64(*config,
ZPOOL_CONFIG_SUSPENDED_REASON,
spa->spa_suspended);
}
spa_add_spares(spa, *config);
spa_add_l2cache(spa, *config);
spa_add_feature_stats(spa, *config);
}
}
/*
* We want to get the alternate root even for faulted pools, so we cheat
* and call spa_lookup() directly.
*/
if (altroot) {
if (spa == NULL) {
mutex_enter(&spa_namespace_lock);
spa = spa_lookup(name);
if (spa)
spa_altroot(spa, altroot, buflen);
else
altroot[0] = '\0';
spa = NULL;
mutex_exit(&spa_namespace_lock);
} else {
spa_altroot(spa, altroot, buflen);
}
}
if (spa != NULL) {
spa_config_exit(spa, SCL_CONFIG, FTAG);
spa_close(spa, FTAG);
}
return (error);
}
/*
* Validate that the auxiliary device array is well formed. We must have an
* array of nvlists, each which describes a valid leaf vdev. If this is an
* import (mode is VDEV_ALLOC_SPARE), then we allow corrupted spares to be
* specified, as long as they are well-formed.
*/
static int
spa_validate_aux_devs(spa_t *spa, nvlist_t *nvroot, uint64_t crtxg, int mode,
spa_aux_vdev_t *sav, const char *config, uint64_t version,
vdev_labeltype_t label)
{
nvlist_t **dev;
uint_t i, ndev;
vdev_t *vd;
int error;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
/*
* It's acceptable to have no devs specified.
*/
if (nvlist_lookup_nvlist_array(nvroot, config, &dev, &ndev) != 0)
return (0);
if (ndev == 0)
return (SET_ERROR(EINVAL));
/*
* Make sure the pool is formatted with a version that supports this
* device type.
*/
if (spa_version(spa) < version)
return (SET_ERROR(ENOTSUP));
/*
* Set the pending device list so we correctly handle device in-use
* checking.
*/
sav->sav_pending = dev;
sav->sav_npending = ndev;
for (i = 0; i < ndev; i++) {
if ((error = spa_config_parse(spa, &vd, dev[i], NULL, 0,
mode)) != 0)
goto out;
if (!vd->vdev_ops->vdev_op_leaf) {
vdev_free(vd);
error = SET_ERROR(EINVAL);
goto out;
}
vd->vdev_top = vd;
if ((error = vdev_open(vd)) == 0 &&
(error = vdev_label_init(vd, crtxg, label)) == 0) {
fnvlist_add_uint64(dev[i], ZPOOL_CONFIG_GUID,
vd->vdev_guid);
}
vdev_free(vd);
if (error &&
(mode != VDEV_ALLOC_SPARE && mode != VDEV_ALLOC_L2CACHE))
goto out;
else
error = 0;
}
out:
sav->sav_pending = NULL;
sav->sav_npending = 0;
return (error);
}
static int
spa_validate_aux(spa_t *spa, nvlist_t *nvroot, uint64_t crtxg, int mode)
{
int error;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
if ((error = spa_validate_aux_devs(spa, nvroot, crtxg, mode,
&spa->spa_spares, ZPOOL_CONFIG_SPARES, SPA_VERSION_SPARES,
VDEV_LABEL_SPARE)) != 0) {
return (error);
}
return (spa_validate_aux_devs(spa, nvroot, crtxg, mode,
&spa->spa_l2cache, ZPOOL_CONFIG_L2CACHE, SPA_VERSION_L2CACHE,
VDEV_LABEL_L2CACHE));
}
static void
spa_set_aux_vdevs(spa_aux_vdev_t *sav, nvlist_t **devs, int ndevs,
const char *config)
{
int i;
if (sav->sav_config != NULL) {
nvlist_t **olddevs;
uint_t oldndevs;
nvlist_t **newdevs;
/*
* Generate new dev list by concatenating with the
* current dev list.
*/
VERIFY0(nvlist_lookup_nvlist_array(sav->sav_config, config,
&olddevs, &oldndevs));
newdevs = kmem_alloc(sizeof (void *) *
(ndevs + oldndevs), KM_SLEEP);
for (i = 0; i < oldndevs; i++)
newdevs[i] = fnvlist_dup(olddevs[i]);
for (i = 0; i < ndevs; i++)
newdevs[i + oldndevs] = fnvlist_dup(devs[i]);
fnvlist_remove(sav->sav_config, config);
fnvlist_add_nvlist_array(sav->sav_config, config,
(const nvlist_t * const *)newdevs, ndevs + oldndevs);
for (i = 0; i < oldndevs + ndevs; i++)
nvlist_free(newdevs[i]);
kmem_free(newdevs, (oldndevs + ndevs) * sizeof (void *));
} else {
/*
* Generate a new dev list.
*/
sav->sav_config = fnvlist_alloc();
fnvlist_add_nvlist_array(sav->sav_config, config,
(const nvlist_t * const *)devs, ndevs);
}
}
/*
* Stop and drop level 2 ARC devices
*/
void
spa_l2cache_drop(spa_t *spa)
{
vdev_t *vd;
int i;
spa_aux_vdev_t *sav = &spa->spa_l2cache;
for (i = 0; i < sav->sav_count; i++) {
uint64_t pool;
vd = sav->sav_vdevs[i];
ASSERT(vd != NULL);
if (spa_l2cache_exists(vd->vdev_guid, &pool) &&
pool != 0ULL && l2arc_vdev_present(vd))
l2arc_remove_vdev(vd);
}
}
/*
* Verify encryption parameters for spa creation. If we are encrypting, we must
* have the encryption feature flag enabled.
*/
static int
spa_create_check_encryption_params(dsl_crypto_params_t *dcp,
boolean_t has_encryption)
{
if (dcp->cp_crypt != ZIO_CRYPT_OFF &&
dcp->cp_crypt != ZIO_CRYPT_INHERIT &&
!has_encryption)
return (SET_ERROR(ENOTSUP));
return (dmu_objset_create_crypt_check(NULL, dcp, NULL));
}
/*
* Pool Creation
*/
int
spa_create(const char *pool, nvlist_t *nvroot, nvlist_t *props,
nvlist_t *zplprops, dsl_crypto_params_t *dcp)
{
spa_t *spa;
const char *altroot = NULL;
vdev_t *rvd;
dsl_pool_t *dp;
dmu_tx_t *tx;
int error = 0;
uint64_t txg = TXG_INITIAL;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
uint64_t version, obj, ndraid = 0;
boolean_t has_features;
boolean_t has_encryption;
boolean_t has_allocclass;
spa_feature_t feat;
const char *feat_name;
const char *poolname;
nvlist_t *nvl;
if (props == NULL ||
nvlist_lookup_string(props, "tname", &poolname) != 0)
poolname = (char *)pool;
/*
* If this pool already exists, return failure.
*/
mutex_enter(&spa_namespace_lock);
if (spa_lookup(poolname) != NULL) {
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(EEXIST));
}
/*
* Allocate a new spa_t structure.
*/
nvl = fnvlist_alloc();
fnvlist_add_string(nvl, ZPOOL_CONFIG_POOL_NAME, pool);
(void) nvlist_lookup_string(props,
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot);
spa = spa_add(poolname, nvl, altroot);
fnvlist_free(nvl);
spa_activate(spa, spa_mode_global);
if (props && (error = spa_prop_validate(spa, props))) {
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
/*
* Temporary pool names should never be written to disk.
*/
if (poolname != pool)
spa->spa_import_flags |= ZFS_IMPORT_TEMP_NAME;
has_features = B_FALSE;
has_encryption = B_FALSE;
has_allocclass = B_FALSE;
for (nvpair_t *elem = nvlist_next_nvpair(props, NULL);
elem != NULL; elem = nvlist_next_nvpair(props, elem)) {
if (zpool_prop_feature(nvpair_name(elem))) {
has_features = B_TRUE;
feat_name = strchr(nvpair_name(elem), '@') + 1;
VERIFY0(zfeature_lookup_name(feat_name, &feat));
if (feat == SPA_FEATURE_ENCRYPTION)
has_encryption = B_TRUE;
if (feat == SPA_FEATURE_ALLOCATION_CLASSES)
has_allocclass = B_TRUE;
}
}
/* verify encryption params, if they were provided */
if (dcp != NULL) {
error = spa_create_check_encryption_params(dcp, has_encryption);
if (error != 0) {
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
}
if (!has_allocclass && zfs_special_devs(nvroot, NULL)) {
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (ENOTSUP);
}
if (has_features || nvlist_lookup_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_VERSION), &version) != 0) {
version = SPA_VERSION;
}
ASSERT(SPA_VERSION_IS_SUPPORTED(version));
spa->spa_first_txg = txg;
spa->spa_uberblock.ub_txg = txg - 1;
spa->spa_uberblock.ub_version = version;
spa->spa_ubsync = spa->spa_uberblock;
spa->spa_load_state = SPA_LOAD_CREATE;
spa->spa_removing_phys.sr_state = DSS_NONE;
spa->spa_removing_phys.sr_removing_vdev = -1;
spa->spa_removing_phys.sr_prev_indirect_vdev = -1;
spa->spa_indirect_vdevs_loaded = B_TRUE;
/*
* Create "The Godfather" zio to hold all async IOs
*/
spa->spa_async_zio_root = kmem_alloc(max_ncpus * sizeof (void *),
KM_SLEEP);
for (int i = 0; i < max_ncpus; i++) {
spa->spa_async_zio_root[i] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE |
ZIO_FLAG_GODFATHER);
}
/*
* Create the root vdev.
*/
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
error = spa_config_parse(spa, &rvd, nvroot, NULL, 0, VDEV_ALLOC_ADD);
ASSERT(error != 0 || rvd != NULL);
ASSERT(error != 0 || spa->spa_root_vdev == rvd);
if (error == 0 && !zfs_allocatable_devs(nvroot))
error = SET_ERROR(EINVAL);
if (error == 0 &&
(error = vdev_create(rvd, txg, B_FALSE)) == 0 &&
(error = vdev_draid_spare_create(nvroot, rvd, &ndraid, 0)) == 0 &&
(error = spa_validate_aux(spa, nvroot, txg, VDEV_ALLOC_ADD)) == 0) {
/*
* instantiate the metaslab groups (this will dirty the vdevs)
* we can no longer error exit past this point
*/
for (int c = 0; error == 0 && c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
vdev_metaslab_set_size(vd);
vdev_expand(vd, txg);
}
}
spa_config_exit(spa, SCL_ALL, FTAG);
if (error != 0) {
spa_unload(spa);
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
/*
* Get the list of spares, if specified.
*/
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0) {
spa->spa_spares.sav_config = fnvlist_alloc();
fnvlist_add_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares,
nspares);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_spares(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
spa->spa_spares.sav_sync = B_TRUE;
}
/*
* Get the list of level 2 cache devices, if specified.
*/
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0) {
VERIFY0(nvlist_alloc(&spa->spa_l2cache.sav_config,
NV_UNIQUE_NAME, KM_SLEEP));
fnvlist_add_nvlist_array(spa->spa_l2cache.sav_config,
ZPOOL_CONFIG_L2CACHE, (const nvlist_t * const *)l2cache,
nl2cache);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_l2cache(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
spa->spa_l2cache.sav_sync = B_TRUE;
}
spa->spa_is_initializing = B_TRUE;
spa->spa_dsl_pool = dp = dsl_pool_create(spa, zplprops, dcp, txg);
spa->spa_is_initializing = B_FALSE;
/*
* Create DDTs (dedup tables).
*/
ddt_create(spa);
/*
* Create BRT table and BRT table object.
*/
brt_create(spa);
spa_update_dspace(spa);
tx = dmu_tx_create_assigned(dp, txg);
/*
* Create the pool's history object.
*/
if (version >= SPA_VERSION_ZPOOL_HISTORY && !spa->spa_history)
spa_history_create_obj(spa, tx);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_CREATE);
spa_history_log_version(spa, "create", tx);
/*
* Create the pool config object.
*/
spa->spa_config_object = dmu_object_alloc(spa->spa_meta_objset,
DMU_OT_PACKED_NVLIST, SPA_CONFIG_BLOCKSIZE,
DMU_OT_PACKED_NVLIST_SIZE, sizeof (uint64_t), tx);
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CONFIG,
sizeof (uint64_t), 1, &spa->spa_config_object, tx) != 0) {
cmn_err(CE_PANIC, "failed to add pool config");
}
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CREATION_VERSION,
sizeof (uint64_t), 1, &version, tx) != 0) {
cmn_err(CE_PANIC, "failed to add pool version");
}
/* Newly created pools with the right version are always deflated. */
if (version >= SPA_VERSION_RAIDZ_DEFLATE) {
spa->spa_deflate = TRUE;
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_DEFLATE,
sizeof (uint64_t), 1, &spa->spa_deflate, tx) != 0) {
cmn_err(CE_PANIC, "failed to add deflate");
}
}
/*
* Create the deferred-free bpobj. Turn off compression
* because sync-to-convergence takes longer if the blocksize
* keeps changing.
*/
obj = bpobj_alloc(spa->spa_meta_objset, 1 << 14, tx);
dmu_object_set_compress(spa->spa_meta_objset, obj,
ZIO_COMPRESS_OFF, tx);
if (zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SYNC_BPOBJ,
sizeof (uint64_t), 1, &obj, tx) != 0) {
cmn_err(CE_PANIC, "failed to add bpobj");
}
VERIFY3U(0, ==, bpobj_open(&spa->spa_deferred_bpobj,
spa->spa_meta_objset, obj));
/*
* Generate some random noise for salted checksums to operate on.
*/
(void) random_get_pseudo_bytes(spa->spa_cksum_salt.zcs_bytes,
sizeof (spa->spa_cksum_salt.zcs_bytes));
/*
* Set pool properties.
*/
spa->spa_bootfs = zpool_prop_default_numeric(ZPOOL_PROP_BOOTFS);
spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION);
spa->spa_failmode = zpool_prop_default_numeric(ZPOOL_PROP_FAILUREMODE);
spa->spa_autoexpand = zpool_prop_default_numeric(ZPOOL_PROP_AUTOEXPAND);
spa->spa_multihost = zpool_prop_default_numeric(ZPOOL_PROP_MULTIHOST);
spa->spa_autotrim = zpool_prop_default_numeric(ZPOOL_PROP_AUTOTRIM);
if (props != NULL) {
spa_configfile_set(spa, props, B_FALSE);
spa_sync_props(props, tx);
}
for (int i = 0; i < ndraid; i++)
spa_feature_incr(spa, SPA_FEATURE_DRAID, tx);
dmu_tx_commit(tx);
spa->spa_sync_on = B_TRUE;
txg_sync_start(dp);
mmp_thread_start(spa);
txg_wait_synced(dp, txg);
spa_spawn_aux_threads(spa);
spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE);
/*
* Don't count references from objsets that are already closed
* and are making their way through the eviction process.
*/
spa_evicting_os_wait(spa);
spa->spa_minref = zfs_refcount_count(&spa->spa_refcount);
spa->spa_load_state = SPA_LOAD_NONE;
spa_import_os(spa);
mutex_exit(&spa_namespace_lock);
return (0);
}
/*
* Import a non-root pool into the system.
*/
int
spa_import(char *pool, nvlist_t *config, nvlist_t *props, uint64_t flags)
{
spa_t *spa;
const char *altroot = NULL;
spa_load_state_t state = SPA_LOAD_IMPORT;
zpool_load_policy_t policy;
spa_mode_t mode = spa_mode_global;
uint64_t readonly = B_FALSE;
int error;
nvlist_t *nvroot;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
/*
* If a pool with this name exists, return failure.
*/
mutex_enter(&spa_namespace_lock);
if (spa_lookup(pool) != NULL) {
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(EEXIST));
}
/*
* Create and initialize the spa structure.
*/
(void) nvlist_lookup_string(props,
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot);
(void) nvlist_lookup_uint64(props,
zpool_prop_to_name(ZPOOL_PROP_READONLY), &readonly);
if (readonly)
mode = SPA_MODE_READ;
spa = spa_add(pool, config, altroot);
spa->spa_import_flags = flags;
/*
* Verbatim import - Take a pool and insert it into the namespace
* as if it had been loaded at boot.
*/
if (spa->spa_import_flags & ZFS_IMPORT_VERBATIM) {
if (props != NULL)
spa_configfile_set(spa, props, B_FALSE);
spa_write_cachefile(spa, B_FALSE, B_TRUE, B_FALSE);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT);
zfs_dbgmsg("spa_import: verbatim import of %s", pool);
mutex_exit(&spa_namespace_lock);
return (0);
}
spa_activate(spa, mode);
/*
* Don't start async tasks until we know everything is healthy.
*/
spa_async_suspend(spa);
zpool_get_load_policy(config, &policy);
if (policy.zlp_rewind & ZPOOL_DO_REWIND)
state = SPA_LOAD_RECOVER;
spa->spa_config_source = SPA_CONFIG_SRC_TRYIMPORT;
if (state != SPA_LOAD_RECOVER) {
spa->spa_last_ubsync_txg = spa->spa_load_txg = 0;
zfs_dbgmsg("spa_import: importing %s", pool);
} else {
zfs_dbgmsg("spa_import: importing %s, max_txg=%lld "
"(RECOVERY MODE)", pool, (longlong_t)policy.zlp_txg);
}
error = spa_load_best(spa, state, policy.zlp_txg, policy.zlp_rewind);
/*
* Propagate anything learned while loading the pool and pass it
* back to caller (i.e. rewind info, missing devices, etc).
*/
fnvlist_add_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
/*
* Toss any existing sparelist, as it doesn't have any validity
* anymore, and conflicts with spa_has_spare().
*/
if (spa->spa_spares.sav_config) {
nvlist_free(spa->spa_spares.sav_config);
spa->spa_spares.sav_config = NULL;
spa_load_spares(spa);
}
if (spa->spa_l2cache.sav_config) {
nvlist_free(spa->spa_l2cache.sav_config);
spa->spa_l2cache.sav_config = NULL;
spa_load_l2cache(spa);
}
nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE);
spa_config_exit(spa, SCL_ALL, FTAG);
if (props != NULL)
spa_configfile_set(spa, props, B_FALSE);
if (error != 0 || (props && spa_writeable(spa) &&
(error = spa_prop_set(spa, props)))) {
spa_unload(spa);
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
spa_async_resume(spa);
/*
* Override any spares and level 2 cache devices as specified by
* the user, as these may have correct device names/devids, etc.
*/
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
&spares, &nspares) == 0) {
if (spa->spa_spares.sav_config)
fnvlist_remove(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES);
else
spa->spa_spares.sav_config = fnvlist_alloc();
fnvlist_add_nvlist_array(spa->spa_spares.sav_config,
ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares,
nspares);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_spares(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
spa->spa_spares.sav_sync = B_TRUE;
}
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
&l2cache, &nl2cache) == 0) {
if (spa->spa_l2cache.sav_config)
fnvlist_remove(spa->spa_l2cache.sav_config,
ZPOOL_CONFIG_L2CACHE);
else
spa->spa_l2cache.sav_config = fnvlist_alloc();
fnvlist_add_nvlist_array(spa->spa_l2cache.sav_config,
ZPOOL_CONFIG_L2CACHE, (const nvlist_t * const *)l2cache,
nl2cache);
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa_load_l2cache(spa);
spa_config_exit(spa, SCL_ALL, FTAG);
spa->spa_l2cache.sav_sync = B_TRUE;
}
/*
* Check for any removed devices.
*/
if (spa->spa_autoreplace) {
spa_aux_check_removed(&spa->spa_spares);
spa_aux_check_removed(&spa->spa_l2cache);
}
if (spa_writeable(spa)) {
/*
* Update the config cache to include the newly-imported pool.
*/
spa_config_update(spa, SPA_CONFIG_UPDATE_POOL);
}
/*
* It's possible that the pool was expanded while it was exported.
* We kick off an async task to handle this for us.
*/
spa_async_request(spa, SPA_ASYNC_AUTOEXPAND);
spa_history_log_version(spa, "import", NULL);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT);
mutex_exit(&spa_namespace_lock);
zvol_create_minors_recursive(pool);
spa_import_os(spa);
return (0);
}
nvlist_t *
spa_tryimport(nvlist_t *tryconfig)
{
nvlist_t *config = NULL;
const char *poolname, *cachefile;
spa_t *spa;
uint64_t state;
int error;
zpool_load_policy_t policy;
if (nvlist_lookup_string(tryconfig, ZPOOL_CONFIG_POOL_NAME, &poolname))
return (NULL);
if (nvlist_lookup_uint64(tryconfig, ZPOOL_CONFIG_POOL_STATE, &state))
return (NULL);
/*
* Create and initialize the spa structure.
*/
mutex_enter(&spa_namespace_lock);
spa = spa_add(TRYIMPORT_NAME, tryconfig, NULL);
spa_activate(spa, SPA_MODE_READ);
/*
* Rewind pool if a max txg was provided.
*/
zpool_get_load_policy(spa->spa_config, &policy);
if (policy.zlp_txg != UINT64_MAX) {
spa->spa_load_max_txg = policy.zlp_txg;
spa->spa_extreme_rewind = B_TRUE;
zfs_dbgmsg("spa_tryimport: importing %s, max_txg=%lld",
poolname, (longlong_t)policy.zlp_txg);
} else {
zfs_dbgmsg("spa_tryimport: importing %s", poolname);
}
if (nvlist_lookup_string(tryconfig, ZPOOL_CONFIG_CACHEFILE, &cachefile)
== 0) {
zfs_dbgmsg("spa_tryimport: using cachefile '%s'", cachefile);
spa->spa_config_source = SPA_CONFIG_SRC_CACHEFILE;
} else {
spa->spa_config_source = SPA_CONFIG_SRC_SCAN;
}
error = spa_load(spa, SPA_LOAD_TRYIMPORT, SPA_IMPORT_EXISTING);
/*
* If 'tryconfig' was at least parsable, return the current config.
*/
if (spa->spa_root_vdev != NULL) {
config = spa_config_generate(spa, NULL, -1ULL, B_TRUE);
fnvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME, poolname);
fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE, state);
fnvlist_add_uint64(config, ZPOOL_CONFIG_TIMESTAMP,
spa->spa_uberblock.ub_timestamp);
fnvlist_add_nvlist(config, ZPOOL_CONFIG_LOAD_INFO,
spa->spa_load_info);
fnvlist_add_uint64(config, ZPOOL_CONFIG_ERRATA,
spa->spa_errata);
/*
* If the bootfs property exists on this pool then we
* copy it out so that external consumers can tell which
* pools are bootable.
*/
if ((!error || error == EEXIST) && spa->spa_bootfs) {
char *tmpname = kmem_alloc(MAXPATHLEN, KM_SLEEP);
/*
* We have to play games with the name since the
* pool was opened as TRYIMPORT_NAME.
*/
if (dsl_dsobj_to_dsname(spa_name(spa),
spa->spa_bootfs, tmpname) == 0) {
char *cp;
char *dsname;
dsname = kmem_alloc(MAXPATHLEN, KM_SLEEP);
cp = strchr(tmpname, '/');
if (cp == NULL) {
(void) strlcpy(dsname, tmpname,
MAXPATHLEN);
} else {
(void) snprintf(dsname, MAXPATHLEN,
"%s/%s", poolname, ++cp);
}
fnvlist_add_string(config, ZPOOL_CONFIG_BOOTFS,
dsname);
kmem_free(dsname, MAXPATHLEN);
}
kmem_free(tmpname, MAXPATHLEN);
}
/*
* Add the list of hot spares and level 2 cache devices.
*/
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
spa_add_spares(spa, config);
spa_add_l2cache(spa, config);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
spa_unload(spa);
spa_deactivate(spa);
spa_remove(spa);
mutex_exit(&spa_namespace_lock);
return (config);
}
/*
* Pool export/destroy
*
* The act of destroying or exporting a pool is very simple. We make sure there
* is no more pending I/O and any references to the pool are gone. Then, we
* update the pool state and sync all the labels to disk, removing the
* configuration from the cache afterwards. If the 'hardforce' flag is set, then
* we don't sync the labels or remove the configuration cache.
*/
static int
spa_export_common(const char *pool, int new_state, nvlist_t **oldconfig,
boolean_t force, boolean_t hardforce)
{
int error;
spa_t *spa;
if (oldconfig)
*oldconfig = NULL;
if (!(spa_mode_global & SPA_MODE_WRITE))
return (SET_ERROR(EROFS));
mutex_enter(&spa_namespace_lock);
if ((spa = spa_lookup(pool)) == NULL) {
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(ENOENT));
}
if (spa->spa_is_exporting) {
/* the pool is being exported by another thread */
mutex_exit(&spa_namespace_lock);
return (SET_ERROR(ZFS_ERR_EXPORT_IN_PROGRESS));
}
spa->spa_is_exporting = B_TRUE;
/*
* Put a hold on the pool, drop the namespace lock, stop async tasks,
* reacquire the namespace lock, and see if we can export.
*/
spa_open_ref(spa, FTAG);
mutex_exit(&spa_namespace_lock);
spa_async_suspend(spa);
if (spa->spa_zvol_taskq) {
zvol_remove_minors(spa, spa_name(spa), B_TRUE);
taskq_wait(spa->spa_zvol_taskq);
}
mutex_enter(&spa_namespace_lock);
spa_close(spa, FTAG);
if (spa->spa_state == POOL_STATE_UNINITIALIZED)
goto export_spa;
/*
* The pool will be in core if it's openable, in which case we can
* modify its state. Objsets may be open only because they're dirty,
* so we have to force it to sync before checking spa_refcnt.
*/
if (spa->spa_sync_on) {
txg_wait_synced(spa->spa_dsl_pool, 0);
spa_evicting_os_wait(spa);
}
/*
* A pool cannot be exported or destroyed if there are active
* references. If we are resetting a pool, allow references by
* fault injection handlers.
*/
if (!spa_refcount_zero(spa) || (spa->spa_inject_ref != 0)) {
error = SET_ERROR(EBUSY);
goto fail;
}
if (spa->spa_sync_on) {
vdev_t *rvd = spa->spa_root_vdev;
/*
* A pool cannot be exported if it has an active shared spare.
* This is to prevent other pools stealing the active spare
* from an exported pool. At user's own will, such pool can
* be forcedly exported.
*/
if (!force && new_state == POOL_STATE_EXPORTED &&
spa_has_active_shared_spare(spa)) {
error = SET_ERROR(EXDEV);
goto fail;
}
/*
* We're about to export or destroy this pool. Make sure
* we stop all initialization and trim activity here before
* we set the spa_final_txg. This will ensure that all
* dirty data resulting from the initialization is
* committed to disk before we unload the pool.
*/
vdev_initialize_stop_all(rvd, VDEV_INITIALIZE_ACTIVE);
vdev_trim_stop_all(rvd, VDEV_TRIM_ACTIVE);
vdev_autotrim_stop_all(spa);
vdev_rebuild_stop_all(spa);
/*
* We want this to be reflected on every label,
* so mark them all dirty. spa_unload() will do the
* final sync that pushes these changes out.
*/
if (new_state != POOL_STATE_UNINITIALIZED && !hardforce) {
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa->spa_state = new_state;
vdev_config_dirty(rvd);
spa_config_exit(spa, SCL_ALL, FTAG);
}
/*
* If the log space map feature is enabled and the pool is
* getting exported (but not destroyed), we want to spend some
* time flushing as many metaslabs as we can in an attempt to
* destroy log space maps and save import time. This has to be
* done before we set the spa_final_txg, otherwise
* spa_sync() -> spa_flush_metaslabs() may dirty the final TXGs.
* spa_should_flush_logs_on_unload() should be called after
* spa_state has been set to the new_state.
*/
if (spa_should_flush_logs_on_unload(spa))
spa_unload_log_sm_flush_all(spa);
if (new_state != POOL_STATE_UNINITIALIZED && !hardforce) {
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
spa->spa_final_txg = spa_last_synced_txg(spa) +
TXG_DEFER_SIZE + 1;
spa_config_exit(spa, SCL_ALL, FTAG);
}
}
export_spa:
spa_export_os(spa);
if (new_state == POOL_STATE_DESTROYED)
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_DESTROY);
else if (new_state == POOL_STATE_EXPORTED)
spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_EXPORT);
if (spa->spa_state != POOL_STATE_UNINITIALIZED) {
spa_unload(spa);
spa_deactivate(spa);
}
if (oldconfig && spa->spa_config)
*oldconfig = fnvlist_dup(spa->spa_config);
if (new_state != POOL_STATE_UNINITIALIZED) {
if (!hardforce)
spa_write_cachefile(spa, B_TRUE, B_TRUE, B_FALSE);
spa_remove(spa);
} else {
/*
* If spa_remove() is not called for this spa_t and
* there is any possibility that it can be reused,
* we make sure to reset the exporting flag.
*/
spa->spa_is_exporting = B_FALSE;
}
mutex_exit(&spa_namespace_lock);
return (0);
fail:
spa->spa_is_exporting = B_FALSE;
spa_async_resume(spa);
mutex_exit(&spa_namespace_lock);
return (error);
}
/*
* Destroy a storage pool.
*/
int
spa_destroy(const char *pool)
{
return (spa_export_common(pool, POOL_STATE_DESTROYED, NULL,
B_FALSE, B_FALSE));
}
/*
* Export a storage pool.
*/
int
spa_export(const char *pool, nvlist_t **oldconfig, boolean_t force,
boolean_t hardforce)
{
return (spa_export_common(pool, POOL_STATE_EXPORTED, oldconfig,
force, hardforce));
}
/*
* Similar to spa_export(), this unloads the spa_t without actually removing it
* from the namespace in any way.
*/
int
spa_reset(const char *pool)
{
return (spa_export_common(pool, POOL_STATE_UNINITIALIZED, NULL,
B_FALSE, B_FALSE));
}
/*
* ==========================================================================
* Device manipulation
* ==========================================================================
*/
/*
* This is called as a synctask to increment the draid feature flag
*/
static void
spa_draid_feature_incr(void *arg, dmu_tx_t *tx)
{
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
int draid = (int)(uintptr_t)arg;
for (int c = 0; c < draid; c++)
spa_feature_incr(spa, SPA_FEATURE_DRAID, tx);
}
/*
* Add a device to a storage pool.
*/
int
spa_vdev_add(spa_t *spa, nvlist_t *nvroot)
{
uint64_t txg, ndraid = 0;
int error;
vdev_t *rvd = spa->spa_root_vdev;
vdev_t *vd, *tvd;
nvlist_t **spares, **l2cache;
uint_t nspares, nl2cache;
ASSERT(spa_writeable(spa));
txg = spa_vdev_enter(spa);
if ((error = spa_config_parse(spa, &vd, nvroot, NULL, 0,
VDEV_ALLOC_ADD)) != 0)
return (spa_vdev_exit(spa, NULL, txg, error));
spa->spa_pending_vdev = vd; /* spa_vdev_exit() will clear this */
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares,
&nspares) != 0)
nspares = 0;
if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache,
&nl2cache) != 0)
nl2cache = 0;
if (vd->vdev_children == 0 && nspares == 0 && nl2cache == 0)
return (spa_vdev_exit(spa, vd, txg, EINVAL));
if (vd->vdev_children != 0 &&
(error = vdev_create(vd, txg, B_FALSE)) != 0) {
return (spa_vdev_exit(spa, vd, txg, error));
}
/*
* The virtual dRAID spares must be added after vdev tree is created
* and the vdev guids are generated. The guid of their associated
* dRAID is stored in the config and used when opening the spare.
*/
if ((error = vdev_draid_spare_create(nvroot, vd, &ndraid,
rvd->vdev_children)) == 0) {
if (ndraid > 0 && nvlist_lookup_nvlist_array(nvroot,
ZPOOL_CONFIG_SPARES, &spares, &nspares) != 0)
nspares = 0;
} else {
return (spa_vdev_exit(spa, vd, txg, error));
}
/*
* We must validate the spares and l2cache devices after checking the
* children. Otherwise, vdev_inuse() will blindly overwrite the spare.
*/
if ((error = spa_validate_aux(spa, nvroot, txg, VDEV_ALLOC_ADD)) != 0)
return (spa_vdev_exit(spa, vd, txg, error));
/*
* If we are in the middle of a device removal, we can only add
* devices which match the existing devices in the pool.
* If we are in the middle of a removal, or have some indirect
* vdevs, we can not add raidz or dRAID top levels.
*/
if (spa->spa_vdev_removal != NULL ||
spa->spa_removing_phys.sr_prev_indirect_vdev != -1) {
for (int c = 0; c < vd->vdev_children; c++) {
tvd = vd->vdev_child[c];
if (spa->spa_vdev_removal != NULL &&
tvd->vdev_ashift != spa->spa_max_ashift) {
return (spa_vdev_exit(spa, vd, txg, EINVAL));
}
/* Fail if top level vdev is raidz or a dRAID */
if (vdev_get_nparity(tvd) != 0)
return (spa_vdev_exit(spa, vd, txg, EINVAL));
/*
* Need the top level mirror to be
* a mirror of leaf vdevs only
*/
if (tvd->vdev_ops == &vdev_mirror_ops) {
for (uint64_t cid = 0;
cid < tvd->vdev_children; cid++) {
vdev_t *cvd = tvd->vdev_child[cid];
if (!cvd->vdev_ops->vdev_op_leaf) {
return (spa_vdev_exit(spa, vd,
txg, EINVAL));
}
}
}
}
}
for (int c = 0; c < vd->vdev_children; c++) {
tvd = vd->vdev_child[c];
vdev_remove_child(vd, tvd);
tvd->vdev_id = rvd->vdev_children;
vdev_add_child(rvd, tvd);
vdev_config_dirty(tvd);
}
if (nspares != 0) {
spa_set_aux_vdevs(&spa->spa_spares, spares, nspares,
ZPOOL_CONFIG_SPARES);
spa_load_spares(spa);
spa->spa_spares.sav_sync = B_TRUE;
}
if (nl2cache != 0) {
spa_set_aux_vdevs(&spa->spa_l2cache, l2cache, nl2cache,
ZPOOL_CONFIG_L2CACHE);
spa_load_l2cache(spa);
spa->spa_l2cache.sav_sync = B_TRUE;
}
/*
* We can't increment a feature while holding spa_vdev so we
* have to do it in a synctask.
*/
if (ndraid != 0) {
dmu_tx_t *tx;
tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg);
dsl_sync_task_nowait(spa->spa_dsl_pool, spa_draid_feature_incr,
(void *)(uintptr_t)ndraid, tx);
dmu_tx_commit(tx);
}
/*
* We have to be careful when adding new vdevs to an existing pool.
* If other threads start allocating from these vdevs before we
* sync the config cache, and we lose power, then upon reboot we may
* fail to open the pool because there are DVAs that the config cache
* can't translate. Therefore, we first add the vdevs without
* initializing metaslabs; sync the config cache (via spa_vdev_exit());
* and then let spa_config_update() initialize the new metaslabs.
*
* spa_load() checks for added-but-not-initialized vdevs, so that
* if we lose power at any point in this sequence, the remaining
* steps will be completed the next time we load the pool.
*/
(void) spa_vdev_exit(spa, vd, txg, 0);
mutex_enter(&spa_namespace_lock);
spa_config_update(spa, SPA_CONFIG_UPDATE_POOL);
spa_event_notify(spa, NULL, NULL, ESC_ZFS_VDEV_ADD);
mutex_exit(&spa_namespace_lock);
return (0);
}
/*
* Attach a device to a mirror. The arguments are the path to any device
* in the mirror, and the nvroot for the new device. If the path specifies
* a device that is not mirrored, we automatically insert the mirror vdev.
*
* If 'replacing' is specified, the new device is intended to replace the
* existing device; in this case the two devices are made into their own
* mirror using the 'replacing' vdev, which is functionally identical to
* the mirror vdev (it actually reuses all the same ops) but has a few
* extra rules: you can't attach to it after it's been created, and upon
* completion of resilvering, the first disk (the one being replaced)
* is automatically detached.
*
* If 'rebuild' is specified, then sequential reconstruction (a.ka. rebuild)
* should be performed instead of traditional healing reconstruction. From
* an administrators perspective these are both resilver operations.
*/
int
spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot, int replacing,
int rebuild)
{
uint64_t txg, dtl_max_txg;
vdev_t *rvd = spa->spa_root_vdev;
vdev_t *oldvd, *newvd, *newrootvd, *pvd, *tvd;
vdev_ops_t *pvops;
char *oldvdpath, *newvdpath;
int newvd_isspare;
int error;
ASSERT(spa_writeable(spa));
txg = spa_vdev_enter(spa);
oldvd = spa_lookup_by_guid(spa, guid, B_FALSE);
ASSERT(MUTEX_HELD(&spa_namespace_lock));
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) {
error = (spa_has_checkpoint(spa)) ?
ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT;
return (spa_vdev_exit(spa, NULL, txg, error));
}
if (rebuild) {
if (!spa_feature_is_enabled(spa, SPA_FEATURE_DEVICE_REBUILD))
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
if (dsl_scan_resilvering(spa_get_dsl(spa)))
return (spa_vdev_exit(spa, NULL, txg,
ZFS_ERR_RESILVER_IN_PROGRESS));
} else {
if (vdev_rebuild_active(rvd))
return (spa_vdev_exit(spa, NULL, txg,
ZFS_ERR_REBUILD_IN_PROGRESS));
}
if (spa->spa_vdev_removal != NULL)
return (spa_vdev_exit(spa, NULL, txg, EBUSY));
if (oldvd == NULL)
return (spa_vdev_exit(spa, NULL, txg, ENODEV));
if (!oldvd->vdev_ops->vdev_op_leaf)
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
pvd = oldvd->vdev_parent;
if (spa_config_parse(spa, &newrootvd, nvroot, NULL, 0,
VDEV_ALLOC_ATTACH) != 0)
return (spa_vdev_exit(spa, NULL, txg, EINVAL));
if (newrootvd->vdev_children != 1)
return (spa_vdev_exit(spa, newrootvd, txg, EINVAL));
newvd = newrootvd->vdev_child[0];
if (!newvd->vdev_ops->vdev_op_leaf)
return (spa_vdev_exit(spa, newrootvd, txg, EINVAL));
if ((error = vdev_create(newrootvd, txg, replacing)) != 0)
return (spa_vdev_exit(spa, newrootvd, txg, error));
/*
* log, dedup and special vdevs should not be replaced by spares.
*/
if ((oldvd->vdev_top->vdev_alloc_bias != VDEV_BIAS_NONE ||
oldvd->vdev_top->vdev_islog) && newvd->vdev_isspare) {
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
}
/*
* A dRAID spare can only replace a child of its parent dRAID vdev.
*/
if (newvd->vdev_ops == &vdev_draid_spare_ops &&
oldvd->vdev_top != vdev_draid_spare_get_parent(newvd)) {
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
}
if (rebuild) {
/*
* For rebuilds, the top vdev must support reconstruction
* using only space maps. This means the only allowable
* vdevs types are the root vdev, a mirror, or dRAID.
*/
tvd = pvd;
if (pvd->vdev_top != NULL)
tvd = pvd->vdev_top;
if (tvd->vdev_ops != &vdev_mirror_ops &&
tvd->vdev_ops != &vdev_root_ops &&
tvd->vdev_ops != &vdev_draid_ops) {
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
}
}
if (!replacing) {
/*
* For attach, the only allowable parent is a mirror or the root
* vdev.
*/
if (pvd->vdev_ops != &vdev_mirror_ops &&
pvd->vdev_ops != &vdev_root_ops)
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
pvops = &vdev_mirror_ops;
} else {
/*
* Active hot spares can only be replaced by inactive hot
* spares.
*/
if (pvd->vdev_ops == &vdev_spare_ops &&
oldvd->vdev_isspare &&
!spa_has_spare(spa, newvd->vdev_guid))
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
/*
* If the source is a hot spare, and the parent isn't already a
* spare, then we want to create a new hot spare. Otherwise, we
* want to create a replacing vdev. The user is not allowed to
* attach to a spared vdev child unless the 'isspare' state is
* the same (spare replaces spare, non-spare replaces
* non-spare).
*/
if (pvd->vdev_ops == &vdev_replacing_ops &&
spa_version(spa) < SPA_VERSION_MULTI_REPLACE) {
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
} else if (pvd->vdev_ops == &vdev_spare_ops &&
newvd->vdev_isspare != oldvd->vdev_isspare) {
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
}
if (newvd->vdev_isspare)
pvops = &vdev_spare_ops;
else
pvops = &vdev_replacing_ops;
}
/*
* Make sure the new device is big enough.
*/
if (newvd->vdev_asize < vdev_get_min_asize(oldvd))
return (spa_vdev_exit(spa, newrootvd, txg, EOVERFLOW));
/*
* The new device cannot have a higher alignment requirement
* than the top-level vdev.
*/
if (newvd->vdev_ashift > oldvd->vdev_top->vdev_ashift)
return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP));
/*
* If this is an in-place replacement, update oldvd's path and devid
* to make it distinguishable from newvd, and unopenable from now on.
*/
if (strcmp(oldvd->vdev_path, newvd->vdev_path) == 0) {
spa_strfree(oldvd->vdev_path);
oldvd->vdev_path = kmem_alloc(strlen(newvd->vdev_path) + 5,
KM_SLEEP);
(void) snprintf(oldvd->vdev_path, strlen(newvd->vdev_path) + 5,
"%s/%s", newvd->vdev_path, "old");
if (oldvd->vdev_devid != NULL) {
spa_strfree(oldvd->vdev_devid);
oldvd->vdev_devid = NULL;
}
}
/*
* If the parent is not a mirror, or if we're replacing, insert the new
* mirror/replacing/spare vdev above oldvd.
*/
if (pvd->vdev_ops != pvops)
pvd = vdev_add_parent(oldvd, pvops);
ASSERT(pvd->vdev_top->vdev_parent == rvd);
ASSERT(pvd->vdev_ops == pvops);
ASSERT(oldvd->vdev_parent == pvd);
/*
* Extract the new device from its root and add it to pvd.
*/
vdev_remove_child(newrootvd, newvd);
newvd->vdev_id = pvd->vdev_children;
newvd->vdev_crtxg = oldvd->vdev_crtxg;
vdev_add_child(pvd, newvd);
/*
* Reevaluate the parent vdev state.
*/
vdev_propagate_state(pvd);
tvd = newvd->vdev_top;
ASSERT(pvd->vdev_top == tvd);
ASSERT(tvd->vdev_parent == rvd);
vdev_config_dirty(tvd);
/*
* Set newvd's DTL to [TXG_INITIAL, dtl_max_txg) so that we account
* for any dmu_sync-ed blocks. It will propagate upward when
* spa_vdev_exit() calls vdev_dtl_reassess().
*/
dtl_max_txg = txg + TXG_CONCURRENT_STATES;
vdev_dtl_dirty(newvd, DTL_MISSING,
TXG_INITIAL, dtl_max_txg - TXG_INITIAL);
if (newvd->vdev_isspare) {
spa_spare_activate(newvd);
spa_event_notify(spa, newvd, NULL, ESC_ZFS_VDEV_SPARE);
}
oldvdpath = spa_strdup(oldvd->vdev_path);
newvdpath = spa_strdup(newvd->vdev_path);
newvd_isspare = newvd->vdev_isspare;
/*
* Mark newvd's DTL dirty in this txg.
*/
vdev_dirty(tvd, VDD_DTL, newvd, txg);
/*
* Schedule the resilver or rebuild to restart in the future. We do
* this to ensure that dmu_sync-ed blocks have been stitched into the
* respective datasets.
*/
if (rebuild) {
newvd->vdev_rebuild_txg = txg;
vdev_rebuild(tvd);
} else {
newvd->vdev_resilver_txg = txg;
if (dsl_scan_resilvering(spa_get_dsl(spa)) &&
spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER)) {
vdev_defer_resilver(newvd);
} else {
dsl_scan_restart_resilver(spa->spa_dsl_pool,
dtl_max_txg);
}
}
if (spa->spa_bootfs)
spa_event_notify(spa, newvd, NULL, ESC_ZFS_BOOTFS_VDEV_ATTACH);
spa_event_notify(spa, newvd, NULL, ESC_ZFS_VDEV_ATTACH);
/*
* Commit the config
*/
(void) spa_vdev_exit(spa, newrootvd, dtl_max_txg, 0);
spa_history_log_internal(spa, "vdev attach", NULL,
"%s vdev=%s %s vdev=%s",
replacing && newvd_isspare ? "spare in" :
replacing ? "replace" : "attach", newvdpath,
replacing ? "for" : "to", oldvdpath);
spa_strfree(oldvdpath);
spa_strfree(newvdpath);
return (0);
}
/*
* Detach a device from a mirror or replacing vdev.
*
* If 'replace_done' is specified, only detach if the parent
- * is a replacing vdev.
+ * is a replacing or a spare vdev.
*/
int
spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid, int replace_done)
{
uint64_t txg;
int error;
vdev_t *rvd __maybe_unused = spa->spa_root_vdev;
vdev_t *vd, *pvd, *cvd, *tvd;
boolean_t unspare = B_FALSE;
uint64_t unspare_guid = 0;
char *vdpath;
ASSERT(spa_writeable(spa));
txg = spa_vdev_detach_enter(spa, guid);
vd = spa_lookup_by_guid(spa, guid, B_FALSE);
/*
* Besides being called directly from the userland through the
* ioctl interface, spa_vdev_detach() can be potentially called
* at the end of spa_vdev_resilver_done().
*
* In the regular case, when we have a checkpoint this shouldn't
* happen as we never empty the DTLs of a vdev during the scrub
* [see comment in dsl_scan_done()]. Thus spa_vdev_resilvering_done()
* should never get here when we have a checkpoint.
*
* That said, even in a case when we checkpoint the pool exactly
* as spa_vdev_resilver_done() calls this function everything
* should be fine as the resilver will return right away.
*/
ASSERT(MUTEX_HELD(&spa_namespace_lock));
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) {
error = (spa_has_checkpoint(spa)) ?
ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT;
return (spa_vdev_exit(spa, NULL, txg, error));
}
if (vd == NULL)
return (spa_vdev_exit(spa, NULL, txg, ENODEV));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
pvd = vd->vdev_parent;
/*
* If the parent/child relationship is not as expected, don't do it.
* Consider M(A,R(B,C)) -- that is, a mirror of A with a replacing
* vdev that's replacing B with C. The user's intent in replacing
* is to go from M(A,B) to M(A,C). If the user decides to cancel
* the replace by detaching C, the expected behavior is to end up
* M(A,B). But suppose that right after deciding to detach C,
* the replacement of B completes. We would have M(A,C), and then
* ask to detach C, which would leave us with just A -- not what
* the user wanted. To prevent this, we make sure that the
* parent/child relationship hasn't changed -- in this example,
* that C's parent is still the replacing vdev R.
*/
if (pvd->vdev_guid != pguid && pguid != 0)
return (spa_vdev_exit(spa, NULL, txg, EBUSY));
/*
* Only 'replacing' or 'spare' vdevs can be replaced.
*/
if (replace_done && pvd->vdev_ops != &vdev_replacing_ops &&
pvd->vdev_ops != &vdev_spare_ops)
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
ASSERT(pvd->vdev_ops != &vdev_spare_ops ||
spa_version(spa) >= SPA_VERSION_SPARES);
/*
* Only mirror, replacing, and spare vdevs support detach.
*/
if (pvd->vdev_ops != &vdev_replacing_ops &&
pvd->vdev_ops != &vdev_mirror_ops &&
pvd->vdev_ops != &vdev_spare_ops)
return (spa_vdev_exit(spa, NULL, txg, ENOTSUP));
/*
* If this device has the only valid copy of some data,
* we cannot safely detach it.
*/
if (vdev_dtl_required(vd))
return (spa_vdev_exit(spa, NULL, txg, EBUSY));
ASSERT(pvd->vdev_children >= 2);
/*
* If we are detaching the second disk from a replacing vdev, then
* check to see if we changed the original vdev's path to have "/old"
* at the end in spa_vdev_attach(). If so, undo that change now.
*/
if (pvd->vdev_ops == &vdev_replacing_ops && vd->vdev_id > 0 &&
vd->vdev_path != NULL) {
size_t len = strlen(vd->vdev_path);
for (int c = 0; c < pvd->vdev_children; c++) {
cvd = pvd->vdev_child[c];
if (cvd == vd || cvd->vdev_path == NULL)
continue;
if (strncmp(cvd->vdev_path, vd->vdev_path, len) == 0 &&
strcmp(cvd->vdev_path + len, "/old") == 0) {
spa_strfree(cvd->vdev_path);
cvd->vdev_path = spa_strdup(vd->vdev_path);
break;
}
}
}
/*
* If we are detaching the original disk from a normal spare, then it
* implies that the spare should become a real disk, and be removed
* from the active spare list for the pool. dRAID spares on the
* other hand are coupled to the pool and thus should never be removed
* from the spares list.
*/
if (pvd->vdev_ops == &vdev_spare_ops && vd->vdev_id == 0) {
vdev_t *last_cvd = pvd->vdev_child[pvd->vdev_children - 1];
if (last_cvd->vdev_isspare &&
last_cvd->vdev_ops != &vdev_draid_spare_ops) {
unspare = B_TRUE;
}
}
/*
* Erase the disk labels so the disk can be used for other things.
* This must be done after all other error cases are handled,
* but before we disembowel vd (so we can still do I/O to it).
* But if we can't do it, don't treat the error as fatal --
* it may be that the unwritability of the disk is the reason
* it's being detached!
*/
(void) vdev_label_init(vd, 0, VDEV_LABEL_REMOVE);
/*
* Remove vd from its parent and compact the parent's children.
*/
vdev_remove_child(pvd, vd);
vdev_compact_children(pvd);
/*
* Remember one of the remaining children so we can get tvd below.
*/
cvd = pvd->vdev_child[pvd->vdev_children - 1];
/*
* If we need to remove the remaining child from the list of hot spares,
* do it now, marking the vdev as no longer a spare in the process.
* We must do this before vdev_remove_parent(), because that can
* change the GUID if it creates a new toplevel GUID. For a similar
* reason, we must remove the spare now, in the same txg as the detach;
* otherwise someone could attach a new sibling, change the GUID, and
* the subsequent attempt to spa_vdev_remove(unspare_guid) would fail.
*/
if (unspare) {
ASSERT(cvd->vdev_isspare);
spa_spare_remove(cvd);
unspare_guid = cvd->vdev_guid;
(void) spa_vdev_remove(spa, unspare_guid, B_TRUE);
cvd->vdev_unspare = B_TRUE;
}
/*
* If the parent mirror/replacing vdev only has one child,
* the parent is no longer needed. Remove it from the tree.
*/
if (pvd->vdev_children == 1) {
if (pvd->vdev_ops == &vdev_spare_ops)
cvd->vdev_unspare = B_FALSE;
vdev_remove_parent(cvd);
}
/*
* We don't set tvd until now because the parent we just removed
* may have been the previous top-level vdev.
*/
tvd = cvd->vdev_top;
ASSERT(tvd->vdev_parent == rvd);
/*
* Reevaluate the parent vdev state.
*/
vdev_propagate_state(cvd);
/*
* If the 'autoexpand' property is set on the pool then automatically
* try to expand the size of the pool. For example if the device we
* just detached was smaller than the others, it may be possible to
* add metaslabs (i.e. grow the pool). We need to reopen the vdev
* first so that we can obtain the updated sizes of the leaf vdevs.
*/
if (spa->spa_autoexpand) {
vdev_reopen(tvd);
vdev_expand(tvd, txg);
}
vdev_config_dirty(tvd);
/*
* Mark vd's DTL as dirty in this txg. vdev_dtl_sync() will see that
* vd->vdev_detached is set and free vd's DTL object in syncing context.
* But first make sure we're not on any *other* txg's DTL list, to
* prevent vd from being accessed after it's freed.
*/
vdpath = spa_strdup(vd->vdev_path ? vd->vdev_path : "none");
for (int t = 0; t < TXG_SIZE; t++)
(void) txg_list_remove_this(&tvd->vdev_dtl_list, vd, t);
vd->vdev_detached = B_TRUE;
vdev_dirty(tvd, VDD_DTL, vd, txg);
spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_REMOVE);
spa_notify_waiters(spa);
/* hang on to the spa before we release the lock */
spa_open_ref(spa, FTAG);
error = spa_vdev_exit(spa, vd, txg, 0);
spa_history_log_internal(spa, "detach", NULL,
"vdev=%s", vdpath);
spa_strfree(vdpath);
/*
* If this was the removal of the original device in a hot spare vdev,
* then we want to go through and remove the device from the hot spare
* list of every other pool.
*/
if (unspare) {
spa_t *altspa = NULL;
mutex_enter(&spa_namespace_lock);
while ((altspa = spa_next(altspa)) != NULL) {
if (altspa->spa_state != POOL_STATE_ACTIVE ||
altspa == spa)
continue;
spa_open_ref(altspa, FTAG);
mutex_exit(&spa_namespace_lock);
(void) spa_vdev_remove(altspa, unspare_guid, B_TRUE);
mutex_enter(&spa_namespace_lock);
spa_close(altspa, FTAG);
}
mutex_exit(&spa_namespace_lock);
/* search the rest of the vdevs for spares to remove */
spa_vdev_resilver_done(spa);
}
/* all done with the spa; OK to release */
mutex_enter(&spa_namespace_lock);
spa_close(spa, FTAG);
mutex_exit(&spa_namespace_lock);
return (error);
}
static int
spa_vdev_initialize_impl(spa_t *spa, uint64_t guid, uint64_t cmd_type,
list_t *vd_list)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
/* Look up vdev and ensure it's a leaf. */
vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE);
if (vd == NULL || vd->vdev_detached) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(ENODEV));
} else if (!vd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(vd)) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EINVAL));
} else if (!vdev_writeable(vd)) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EROFS));
}
mutex_enter(&vd->vdev_initialize_lock);
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
/*
* When we activate an initialize action we check to see
* if the vdev_initialize_thread is NULL. We do this instead
* of using the vdev_initialize_state since there might be
* a previous initialization process which has completed but
* the thread is not exited.
*/
if (cmd_type == POOL_INITIALIZE_START &&
(vd->vdev_initialize_thread != NULL ||
vd->vdev_top->vdev_removing)) {
mutex_exit(&vd->vdev_initialize_lock);
return (SET_ERROR(EBUSY));
} else if (cmd_type == POOL_INITIALIZE_CANCEL &&
(vd->vdev_initialize_state != VDEV_INITIALIZE_ACTIVE &&
vd->vdev_initialize_state != VDEV_INITIALIZE_SUSPENDED)) {
mutex_exit(&vd->vdev_initialize_lock);
return (SET_ERROR(ESRCH));
} else if (cmd_type == POOL_INITIALIZE_SUSPEND &&
vd->vdev_initialize_state != VDEV_INITIALIZE_ACTIVE) {
mutex_exit(&vd->vdev_initialize_lock);
return (SET_ERROR(ESRCH));
}
switch (cmd_type) {
case POOL_INITIALIZE_START:
vdev_initialize(vd);
break;
case POOL_INITIALIZE_CANCEL:
vdev_initialize_stop(vd, VDEV_INITIALIZE_CANCELED, vd_list);
break;
case POOL_INITIALIZE_SUSPEND:
vdev_initialize_stop(vd, VDEV_INITIALIZE_SUSPENDED, vd_list);
break;
default:
panic("invalid cmd_type %llu", (unsigned long long)cmd_type);
}
mutex_exit(&vd->vdev_initialize_lock);
return (0);
}
int
spa_vdev_initialize(spa_t *spa, nvlist_t *nv, uint64_t cmd_type,
nvlist_t *vdev_errlist)
{
int total_errors = 0;
list_t vd_list;
list_create(&vd_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_initialize_node));
/*
* We hold the namespace lock through the whole function
* to prevent any changes to the pool while we're starting or
* stopping initialization. The config and state locks are held so that
* we can properly assess the vdev state before we commit to
* the initializing operation.
*/
mutex_enter(&spa_namespace_lock);
for (nvpair_t *pair = nvlist_next_nvpair(nv, NULL);
pair != NULL; pair = nvlist_next_nvpair(nv, pair)) {
uint64_t vdev_guid = fnvpair_value_uint64(pair);
int error = spa_vdev_initialize_impl(spa, vdev_guid, cmd_type,
&vd_list);
if (error != 0) {
char guid_as_str[MAXNAMELEN];
(void) snprintf(guid_as_str, sizeof (guid_as_str),
"%llu", (unsigned long long)vdev_guid);
fnvlist_add_int64(vdev_errlist, guid_as_str, error);
total_errors++;
}
}
/* Wait for all initialize threads to stop. */
vdev_initialize_stop_wait(spa, &vd_list);
/* Sync out the initializing state */
txg_wait_synced(spa->spa_dsl_pool, 0);
mutex_exit(&spa_namespace_lock);
list_destroy(&vd_list);
return (total_errors);
}
static int
spa_vdev_trim_impl(spa_t *spa, uint64_t guid, uint64_t cmd_type,
uint64_t rate, boolean_t partial, boolean_t secure, list_t *vd_list)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
/* Look up vdev and ensure it's a leaf. */
vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE);
if (vd == NULL || vd->vdev_detached) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(ENODEV));
} else if (!vd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(vd)) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EINVAL));
} else if (!vdev_writeable(vd)) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EROFS));
} else if (!vd->vdev_has_trim) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EOPNOTSUPP));
} else if (secure && !vd->vdev_has_securetrim) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (SET_ERROR(EOPNOTSUPP));
}
mutex_enter(&vd->vdev_trim_lock);
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
/*
* When we activate a TRIM action we check to see if the
* vdev_trim_thread is NULL. We do this instead of using the
* vdev_trim_state since there might be a previous TRIM process
* which has completed but the thread is not exited.
*/
if (cmd_type == POOL_TRIM_START &&
(vd->vdev_trim_thread != NULL || vd->vdev_top->vdev_removing)) {
mutex_exit(&vd->vdev_trim_lock);
return (SET_ERROR(EBUSY));
} else if (cmd_type == POOL_TRIM_CANCEL &&
(vd->vdev_trim_state != VDEV_TRIM_ACTIVE &&
vd->vdev_trim_state != VDEV_TRIM_SUSPENDED)) {
mutex_exit(&vd->vdev_trim_lock);
return (SET_ERROR(ESRCH));
} else if (cmd_type == POOL_TRIM_SUSPEND &&
vd->vdev_trim_state != VDEV_TRIM_ACTIVE) {
mutex_exit(&vd->vdev_trim_lock);
return (SET_ERROR(ESRCH));
}
switch (cmd_type) {
case POOL_TRIM_START:
vdev_trim(vd, rate, partial, secure);
break;
case POOL_TRIM_CANCEL:
vdev_trim_stop(vd, VDEV_TRIM_CANCELED, vd_list);
break;
case POOL_TRIM_SUSPEND:
vdev_trim_stop(vd, VDEV_TRIM_SUSPENDED, vd_list);
break;
default:
panic("invalid cmd_type %llu", (unsigned long long)cmd_type);
}
mutex_exit(&vd->vdev_trim_lock);
return (0);
}
/*
* Initiates a manual TRIM for the requested vdevs. This kicks off individual
* TRIM threads for each child vdev. These threads pass over all of the free
* space in the vdev's metaslabs and issues TRIM commands for that space.
*/
int
spa_vdev_trim(spa_t *spa, nvlist_t *nv, uint64_t cmd_type, uint64_t rate,
boolean_t partial, boolean_t secure, nvlist_t *vdev_errlist)
{
int total_errors = 0;
list_t vd_list;
list_create(&vd_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_trim_node));
/*
* We hold the namespace lock through the whole function
* to prevent any changes to the pool while we're starting or
* stopping TRIM. The config and state locks are held so that
* we can properly assess the vdev state before we commit to
* the TRIM operation.
*/
mutex_enter(&spa_namespace_lock);
for (nvpair_t *pair = nvlist_next_nvpair(nv, NULL);
pair != NULL; pair = nvlist_next_nvpair(nv, pair)) {
uint64_t vdev_guid = fnvpair_value_uint64(pair);
int error = spa_vdev_trim_impl(spa, vdev_guid, cmd_type,
rate, partial, secure, &vd_list);
if (error != 0) {
char guid_as_str[MAXNAMELEN];
(void) snprintf(guid_as_str, sizeof (guid_as_str),
"%llu", (unsigned long long)vdev_guid);
fnvlist_add_int64(vdev_errlist, guid_as_str, error);
total_errors++;
}
}
/* Wait for all TRIM threads to stop. */
vdev_trim_stop_wait(spa, &vd_list);
/* Sync out the TRIM state */
txg_wait_synced(spa->spa_dsl_pool, 0);
mutex_exit(&spa_namespace_lock);
list_destroy(&vd_list);
return (total_errors);
}
/*
* Split a set of devices from their mirrors, and create a new pool from them.
*/
int
spa_vdev_split_mirror(spa_t *spa, const char *newname, nvlist_t *config,
nvlist_t *props, boolean_t exp)
{
int error = 0;
uint64_t txg, *glist;
spa_t *newspa;
uint_t c, children, lastlog;
nvlist_t **child, *nvl, *tmp;
dmu_tx_t *tx;
const char *altroot = NULL;
vdev_t *rvd, **vml = NULL; /* vdev modify list */
boolean_t activate_slog;
ASSERT(spa_writeable(spa));
txg = spa_vdev_enter(spa);
ASSERT(MUTEX_HELD(&spa_namespace_lock));
if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) {
error = (spa_has_checkpoint(spa)) ?
ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT;
return (spa_vdev_exit(spa, NULL, txg, error));
}
/* clear the log and flush everything up to now */
activate_slog = spa_passivate_log(spa);
(void) spa_vdev_config_exit(spa, NULL, txg, 0, FTAG);
error = spa_reset_logs(spa);
txg = spa_vdev_config_enter(spa);
if (activate_slog)
spa_activate_log(spa);
if (error != 0)
return (spa_vdev_exit(spa, NULL, txg, error));
/* check new spa name before going any further */
if (spa_lookup(newname) != NULL)
return (spa_vdev_exit(spa, NULL, txg, EEXIST));
/*
* scan through all the children to ensure they're all mirrors
*/
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvl) != 0 ||
nvlist_lookup_nvlist_array(nvl, ZPOOL_CONFIG_CHILDREN, &child,
&children) != 0)
return (spa_vdev_exit(spa, NULL, txg, EINVAL));
/* first, check to ensure we've got the right child count */
rvd = spa->spa_root_vdev;
lastlog = 0;
for (c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
/* don't count the holes & logs as children */
if (vd->vdev_islog || (vd->vdev_ops != &vdev_indirect_ops &&
!vdev_is_concrete(vd))) {
if (lastlog == 0)
lastlog = c;
continue;
}
lastlog = 0;
}
if (children != (lastlog != 0 ? lastlog : rvd->vdev_children))
return (spa_vdev_exit(spa, NULL, txg, EINVAL));
/* next, ensure no spare or cache devices are part of the split */
if (nvlist_lookup_nvlist(nvl, ZPOOL_CONFIG_SPARES, &tmp) == 0 ||
nvlist_lookup_nvlist(nvl, ZPOOL_CONFIG_L2CACHE, &tmp) == 0)
return (spa_vdev_exit(spa, NULL, txg, EINVAL));
vml = kmem_zalloc(children * sizeof (vdev_t *), KM_SLEEP);
glist = kmem_zalloc(children * sizeof (uint64_t), KM_SLEEP);
/* then, loop over each vdev and validate it */
for (c = 0; c < children; c++) {
uint64_t is_hole = 0;
(void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE,
&is_hole);
if (is_hole != 0) {
if (spa->spa_root_vdev->vdev_child[c]->vdev_ishole ||
spa->spa_root_vdev->vdev_child[c]->vdev_islog) {
continue;
} else {
error = SET_ERROR(EINVAL);
break;
}
}
/* deal with indirect vdevs */
if (spa->spa_root_vdev->vdev_child[c]->vdev_ops ==
&vdev_indirect_ops)
continue;
/* which disk is going to be split? */
if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_GUID,
&glist[c]) != 0) {
error = SET_ERROR(EINVAL);
break;
}
/* look it up in the spa */
vml[c] = spa_lookup_by_guid(spa, glist[c], B_FALSE);
if (vml[c] == NULL) {
error = SET_ERROR(ENODEV);
break;
}
/* make sure there's nothing stopping the split */
if (vml[c]->vdev_parent->vdev_ops != &vdev_mirror_ops ||
vml[c]->vdev_islog ||
!vdev_is_concrete(vml[c]) ||
vml[c]->vdev_isspare ||
vml[c]->vdev_isl2cache ||
!vdev_writeable(vml[c]) ||
vml[c]->vdev_children != 0 ||
vml[c]->vdev_state != VDEV_STATE_HEALTHY ||
c != spa->spa_root_vdev->vdev_child[c]->vdev_id) {
error = SET_ERROR(EINVAL);
break;
}
if (vdev_dtl_required(vml[c]) ||
vdev_resilver_needed(vml[c], NULL, NULL)) {
error = SET_ERROR(EBUSY);
break;
}
/* we need certain info from the top level */
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_METASLAB_ARRAY,
vml[c]->vdev_top->vdev_ms_array);
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_METASLAB_SHIFT,
vml[c]->vdev_top->vdev_ms_shift);
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_ASIZE,
vml[c]->vdev_top->vdev_asize);
fnvlist_add_uint64(child[c], ZPOOL_CONFIG_ASHIFT,
vml[c]->vdev_top->vdev_ashift);
/* transfer per-vdev ZAPs */
ASSERT3U(vml[c]->vdev_leaf_zap, !=, 0);
VERIFY0(nvlist_add_uint64(child[c],
ZPOOL_CONFIG_VDEV_LEAF_ZAP, vml[c]->vdev_leaf_zap));
ASSERT3U(vml[c]->vdev_top->vdev_top_zap, !=, 0);
VERIFY0(nvlist_add_uint64(child[c],
ZPOOL_CONFIG_VDEV_TOP_ZAP,
vml[c]->vdev_parent->vdev_top_zap));
}
if (error != 0) {
kmem_free(vml, children * sizeof (vdev_t *));
kmem_free(glist, children * sizeof (uint64_t));
return (spa_vdev_exit(spa, NULL, txg, error));
}
/* stop writers from using the disks */
for (c = 0; c < children; c++) {
if (vml[c] != NULL)
vml[c]->vdev_offline = B_TRUE;
}
vdev_reopen(spa->spa_root_vdev);
/*
* Temporarily record the splitting vdevs in the spa config. This
* will disappear once the config is regenerated.
*/
nvl = fnvlist_alloc();
fnvlist_add_uint64_array(nvl, ZPOOL_CONFIG_SPLIT_LIST, glist, children);
kmem_free(glist, children * sizeof (uint64_t));
mutex_enter(&spa->spa_props_lock);
fnvlist_add_nvlist(spa->spa_config, ZPOOL_CONFIG_SPLIT, nvl);
mutex_exit(&spa->spa_props_lock);
spa->spa_config_splitting = nvl;
vdev_config_dirty(spa->spa_root_vdev);
/* configure and create the new pool */
fnvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME, newname);
fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE,
exp ? POOL_STATE_EXPORTED : POOL_STATE_ACTIVE);
fnvlist_add_uint64(config, ZPOOL_CONFIG_VERSION, spa_version(spa));
fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_TXG, spa->spa_config_txg);
fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_GUID,
spa_generate_guid(NULL));
VERIFY0(nvlist_add_boolean(config, ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS));
(void) nvlist_lookup_string(props,
zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot);
/* add the new pool to the namespace */
newspa = spa_add(newname, config, altroot);
newspa->spa_avz_action = AVZ_ACTION_REBUILD;
newspa->spa_config_txg = spa->spa_config_txg;
spa_set_log_state(newspa, SPA_LOG_CLEAR);
/* release the spa config lock, retaining the namespace lock */
spa_vdev_config_exit(spa, NULL, txg, 0, FTAG);
if (zio_injection_enabled)
zio_handle_panic_injection(spa, FTAG, 1);
spa_activate(newspa, spa_mode_global);
spa_async_suspend(newspa);
/*
* Temporarily stop the initializing and TRIM activity. We set the
* state to ACTIVE so that we know to resume initializing or TRIM
* once the split has completed.
*/
list_t vd_initialize_list;
list_create(&vd_initialize_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_initialize_node));
list_t vd_trim_list;
list_create(&vd_trim_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_trim_node));
for (c = 0; c < children; c++) {
if (vml[c] != NULL && vml[c]->vdev_ops != &vdev_indirect_ops) {
mutex_enter(&vml[c]->vdev_initialize_lock);
vdev_initialize_stop(vml[c],
VDEV_INITIALIZE_ACTIVE, &vd_initialize_list);
mutex_exit(&vml[c]->vdev_initialize_lock);
mutex_enter(&vml[c]->vdev_trim_lock);
vdev_trim_stop(vml[c], VDEV_TRIM_ACTIVE, &vd_trim_list);
mutex_exit(&vml[c]->vdev_trim_lock);
}
}
vdev_initialize_stop_wait(spa, &vd_initialize_list);
vdev_trim_stop_wait(spa, &vd_trim_list);
list_destroy(&vd_initialize_list);
list_destroy(&vd_trim_list);
newspa->spa_config_source = SPA_CONFIG_SRC_SPLIT;
newspa->spa_is_splitting = B_TRUE;
/* create the new pool from the disks of the original pool */
error = spa_load(newspa, SPA_LOAD_IMPORT, SPA_IMPORT_ASSEMBLE);
if (error)
goto out;
/* if that worked, generate a real config for the new pool */
if (newspa->spa_root_vdev != NULL) {
newspa->spa_config_splitting = fnvlist_alloc();
fnvlist_add_uint64(newspa->spa_config_splitting,
ZPOOL_CONFIG_SPLIT_GUID, spa_guid(spa));
spa_config_set(newspa, spa_config_generate(newspa, NULL, -1ULL,
B_TRUE));
}
/* set the props */
if (props != NULL) {
spa_configfile_set(newspa, props, B_FALSE);
error = spa_prop_set(newspa, props);
if (error)
goto out;
}
/* flush everything */
txg = spa_vdev_config_enter(newspa);
vdev_config_dirty(newspa->spa_root_vdev);
(void) spa_vdev_config_exit(newspa, NULL, txg, 0, FTAG);
if (zio_injection_enabled)
zio_handle_panic_injection(spa, FTAG, 2);
spa_async_resume(newspa);
/* finally, update the original pool's config */
txg = spa_vdev_config_enter(spa);
tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error != 0)
dmu_tx_abort(tx);
for (c = 0; c < children; c++) {
if (vml[c] != NULL && vml[c]->vdev_ops != &vdev_indirect_ops) {
vdev_t *tvd = vml[c]->vdev_top;
/*
* Need to be sure the detachable VDEV is not
* on any *other* txg's DTL list to prevent it
* from being accessed after it's freed.
*/
for (int t = 0; t < TXG_SIZE; t++) {
(void) txg_list_remove_this(
&tvd->vdev_dtl_list, vml[c], t);
}
vdev_split(vml[c]);
if (error == 0)
spa_history_log_internal(spa, "detach", tx,
"vdev=%s", vml[c]->vdev_path);
vdev_free(vml[c]);
}
}
spa->spa_avz_action = AVZ_ACTION_REBUILD;
vdev_config_dirty(spa->spa_root_vdev);
spa->spa_config_splitting = NULL;
nvlist_free(nvl);
if (error == 0)
dmu_tx_commit(tx);
(void) spa_vdev_exit(spa, NULL, txg, 0);
if (zio_injection_enabled)
zio_handle_panic_injection(spa, FTAG, 3);
/* split is complete; log a history record */
spa_history_log_internal(newspa, "split", NULL,
"from pool %s", spa_name(spa));
newspa->spa_is_splitting = B_FALSE;
kmem_free(vml, children * sizeof (vdev_t *));
/* if we're not going to mount the filesystems in userland, export */
if (exp)
error = spa_export_common(newname, POOL_STATE_EXPORTED, NULL,
B_FALSE, B_FALSE);
return (error);
out:
spa_unload(newspa);
spa_deactivate(newspa);
spa_remove(newspa);
txg = spa_vdev_config_enter(spa);
/* re-online all offlined disks */
for (c = 0; c < children; c++) {
if (vml[c] != NULL)
vml[c]->vdev_offline = B_FALSE;
}
/* restart initializing or trimming disks as necessary */
spa_async_request(spa, SPA_ASYNC_INITIALIZE_RESTART);
spa_async_request(spa, SPA_ASYNC_TRIM_RESTART);
spa_async_request(spa, SPA_ASYNC_AUTOTRIM_RESTART);
vdev_reopen(spa->spa_root_vdev);
nvlist_free(spa->spa_config_splitting);
spa->spa_config_splitting = NULL;
(void) spa_vdev_exit(spa, NULL, txg, error);
kmem_free(vml, children * sizeof (vdev_t *));
return (error);
}
/*
* Find any device that's done replacing, or a vdev marked 'unspare' that's
* currently spared, so we can detach it.
*/
static vdev_t *
spa_vdev_resilver_done_hunt(vdev_t *vd)
{
vdev_t *newvd, *oldvd;
for (int c = 0; c < vd->vdev_children; c++) {
oldvd = spa_vdev_resilver_done_hunt(vd->vdev_child[c]);
if (oldvd != NULL)
return (oldvd);
}
/*
* Check for a completed replacement. We always consider the first
* vdev in the list to be the oldest vdev, and the last one to be
* the newest (see spa_vdev_attach() for how that works). In
* the case where the newest vdev is faulted, we will not automatically
* remove it after a resilver completes. This is OK as it will require
* user intervention to determine which disk the admin wishes to keep.
*/
if (vd->vdev_ops == &vdev_replacing_ops) {
ASSERT(vd->vdev_children > 1);
newvd = vd->vdev_child[vd->vdev_children - 1];
oldvd = vd->vdev_child[0];
if (vdev_dtl_empty(newvd, DTL_MISSING) &&
vdev_dtl_empty(newvd, DTL_OUTAGE) &&
!vdev_dtl_required(oldvd))
return (oldvd);
}
/*
* Check for a completed resilver with the 'unspare' flag set.
* Also potentially update faulted state.
*/
if (vd->vdev_ops == &vdev_spare_ops) {
vdev_t *first = vd->vdev_child[0];
vdev_t *last = vd->vdev_child[vd->vdev_children - 1];
if (last->vdev_unspare) {
oldvd = first;
newvd = last;
} else if (first->vdev_unspare) {
oldvd = last;
newvd = first;
} else {
oldvd = NULL;
}
if (oldvd != NULL &&
vdev_dtl_empty(newvd, DTL_MISSING) &&
vdev_dtl_empty(newvd, DTL_OUTAGE) &&
!vdev_dtl_required(oldvd))
return (oldvd);
vdev_propagate_state(vd);
/*
* If there are more than two spares attached to a disk,
* and those spares are not required, then we want to
* attempt to free them up now so that they can be used
* by other pools. Once we're back down to a single
* disk+spare, we stop removing them.
*/
if (vd->vdev_children > 2) {
newvd = vd->vdev_child[1];
if (newvd->vdev_isspare && last->vdev_isspare &&
vdev_dtl_empty(last, DTL_MISSING) &&
vdev_dtl_empty(last, DTL_OUTAGE) &&
!vdev_dtl_required(newvd))
return (newvd);
}
}
return (NULL);
}
static void
spa_vdev_resilver_done(spa_t *spa)
{
vdev_t *vd, *pvd, *ppvd;
uint64_t guid, sguid, pguid, ppguid;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
while ((vd = spa_vdev_resilver_done_hunt(spa->spa_root_vdev)) != NULL) {
pvd = vd->vdev_parent;
ppvd = pvd->vdev_parent;
guid = vd->vdev_guid;
pguid = pvd->vdev_guid;
ppguid = ppvd->vdev_guid;
sguid = 0;
/*
* If we have just finished replacing a hot spared device, then
* we need to detach the parent's first child (the original hot
* spare) as well.
*/
if (ppvd->vdev_ops == &vdev_spare_ops && pvd->vdev_id == 0 &&
ppvd->vdev_children == 2) {
ASSERT(pvd->vdev_ops == &vdev_replacing_ops);
sguid = ppvd->vdev_child[1]->vdev_guid;
}
ASSERT(vd->vdev_resilver_txg == 0 || !vdev_dtl_required(vd));
spa_config_exit(spa, SCL_ALL, FTAG);
if (spa_vdev_detach(spa, guid, pguid, B_TRUE) != 0)
return;
if (sguid && spa_vdev_detach(spa, sguid, ppguid, B_TRUE) != 0)
return;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
}
spa_config_exit(spa, SCL_ALL, FTAG);
/*
* If a detach was not performed above replace waiters will not have
* been notified. In which case we must do so now.
*/
spa_notify_waiters(spa);
}
/*
* Update the stored path or FRU for this vdev.
*/
static int
spa_vdev_set_common(spa_t *spa, uint64_t guid, const char *value,
boolean_t ispath)
{
vdev_t *vd;
boolean_t sync = B_FALSE;
ASSERT(spa_writeable(spa));
spa_vdev_state_enter(spa, SCL_ALL);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, ENOENT));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, ENOTSUP));
if (ispath) {
if (strcmp(value, vd->vdev_path) != 0) {
spa_strfree(vd->vdev_path);
vd->vdev_path = spa_strdup(value);
sync = B_TRUE;
}
} else {
if (vd->vdev_fru == NULL) {
vd->vdev_fru = spa_strdup(value);
sync = B_TRUE;
} else if (strcmp(value, vd->vdev_fru) != 0) {
spa_strfree(vd->vdev_fru);
vd->vdev_fru = spa_strdup(value);
sync = B_TRUE;
}
}
return (spa_vdev_state_exit(spa, sync ? vd : NULL, 0));
}
int
spa_vdev_setpath(spa_t *spa, uint64_t guid, const char *newpath)
{
return (spa_vdev_set_common(spa, guid, newpath, B_TRUE));
}
int
spa_vdev_setfru(spa_t *spa, uint64_t guid, const char *newfru)
{
return (spa_vdev_set_common(spa, guid, newfru, B_FALSE));
}
/*
* ==========================================================================
* SPA Scanning
* ==========================================================================
*/
int
spa_scrub_pause_resume(spa_t *spa, pool_scrub_cmd_t cmd)
{
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
if (dsl_scan_resilvering(spa->spa_dsl_pool))
return (SET_ERROR(EBUSY));
return (dsl_scrub_set_pause_resume(spa->spa_dsl_pool, cmd));
}
int
spa_scan_stop(spa_t *spa)
{
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
if (dsl_scan_resilvering(spa->spa_dsl_pool))
return (SET_ERROR(EBUSY));
return (dsl_scan_cancel(spa->spa_dsl_pool));
}
int
spa_scan(spa_t *spa, pool_scan_func_t func)
{
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0);
if (func >= POOL_SCAN_FUNCS || func == POOL_SCAN_NONE)
return (SET_ERROR(ENOTSUP));
if (func == POOL_SCAN_RESILVER &&
!spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER))
return (SET_ERROR(ENOTSUP));
/*
* If a resilver was requested, but there is no DTL on a
* writeable leaf device, we have nothing to do.
*/
if (func == POOL_SCAN_RESILVER &&
!vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) {
spa_async_request(spa, SPA_ASYNC_RESILVER_DONE);
return (0);
}
return (dsl_scan(spa->spa_dsl_pool, func));
}
/*
* ==========================================================================
* SPA async task processing
* ==========================================================================
*/
static void
spa_async_remove(spa_t *spa, vdev_t *vd)
{
if (vd->vdev_remove_wanted) {
vd->vdev_remove_wanted = B_FALSE;
vd->vdev_delayed_close = B_FALSE;
vdev_set_state(vd, B_FALSE, VDEV_STATE_REMOVED, VDEV_AUX_NONE);
/*
* We want to clear the stats, but we don't want to do a full
* vdev_clear() as that will cause us to throw away
* degraded/faulted state as well as attempt to reopen the
* device, all of which is a waste.
*/
vd->vdev_stat.vs_read_errors = 0;
vd->vdev_stat.vs_write_errors = 0;
vd->vdev_stat.vs_checksum_errors = 0;
vdev_state_dirty(vd->vdev_top);
/* Tell userspace that the vdev is gone. */
zfs_post_remove(spa, vd);
}
for (int c = 0; c < vd->vdev_children; c++)
spa_async_remove(spa, vd->vdev_child[c]);
}
static void
spa_async_probe(spa_t *spa, vdev_t *vd)
{
if (vd->vdev_probe_wanted) {
vd->vdev_probe_wanted = B_FALSE;
vdev_reopen(vd); /* vdev_open() does the actual probe */
}
for (int c = 0; c < vd->vdev_children; c++)
spa_async_probe(spa, vd->vdev_child[c]);
}
static void
spa_async_autoexpand(spa_t *spa, vdev_t *vd)
{
if (!spa->spa_autoexpand)
return;
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
spa_async_autoexpand(spa, cvd);
}
if (!vd->vdev_ops->vdev_op_leaf || vd->vdev_physpath == NULL)
return;
spa_event_notify(vd->vdev_spa, vd, NULL, ESC_ZFS_VDEV_AUTOEXPAND);
}
static __attribute__((noreturn)) void
spa_async_thread(void *arg)
{
spa_t *spa = (spa_t *)arg;
dsl_pool_t *dp = spa->spa_dsl_pool;
int tasks;
ASSERT(spa->spa_sync_on);
mutex_enter(&spa->spa_async_lock);
tasks = spa->spa_async_tasks;
spa->spa_async_tasks = 0;
mutex_exit(&spa->spa_async_lock);
/*
* See if the config needs to be updated.
*/
if (tasks & SPA_ASYNC_CONFIG_UPDATE) {
uint64_t old_space, new_space;
mutex_enter(&spa_namespace_lock);
old_space = metaslab_class_get_space(spa_normal_class(spa));
old_space += metaslab_class_get_space(spa_special_class(spa));
old_space += metaslab_class_get_space(spa_dedup_class(spa));
old_space += metaslab_class_get_space(
spa_embedded_log_class(spa));
spa_config_update(spa, SPA_CONFIG_UPDATE_POOL);
new_space = metaslab_class_get_space(spa_normal_class(spa));
new_space += metaslab_class_get_space(spa_special_class(spa));
new_space += metaslab_class_get_space(spa_dedup_class(spa));
new_space += metaslab_class_get_space(
spa_embedded_log_class(spa));
mutex_exit(&spa_namespace_lock);
/*
* If the pool grew as a result of the config update,
* then log an internal history event.
*/
if (new_space != old_space) {
spa_history_log_internal(spa, "vdev online", NULL,
"pool '%s' size: %llu(+%llu)",
spa_name(spa), (u_longlong_t)new_space,
(u_longlong_t)(new_space - old_space));
}
}
/*
* See if any devices need to be marked REMOVED.
*/
if (tasks & SPA_ASYNC_REMOVE) {
spa_vdev_state_enter(spa, SCL_NONE);
spa_async_remove(spa, spa->spa_root_vdev);
for (int i = 0; i < spa->spa_l2cache.sav_count; i++)
spa_async_remove(spa, spa->spa_l2cache.sav_vdevs[i]);
for (int i = 0; i < spa->spa_spares.sav_count; i++)
spa_async_remove(spa, spa->spa_spares.sav_vdevs[i]);
(void) spa_vdev_state_exit(spa, NULL, 0);
}
if ((tasks & SPA_ASYNC_AUTOEXPAND) && !spa_suspended(spa)) {
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
spa_async_autoexpand(spa, spa->spa_root_vdev);
spa_config_exit(spa, SCL_CONFIG, FTAG);
}
/*
* See if any devices need to be probed.
*/
if (tasks & SPA_ASYNC_PROBE) {
spa_vdev_state_enter(spa, SCL_NONE);
spa_async_probe(spa, spa->spa_root_vdev);
(void) spa_vdev_state_exit(spa, NULL, 0);
}
/*
* If any devices are done replacing, detach them.
*/
if (tasks & SPA_ASYNC_RESILVER_DONE ||
- tasks & SPA_ASYNC_REBUILD_DONE) {
+ tasks & SPA_ASYNC_REBUILD_DONE ||
+ tasks & SPA_ASYNC_DETACH_SPARE) {
spa_vdev_resilver_done(spa);
}
/*
* Kick off a resilver.
*/
if (tasks & SPA_ASYNC_RESILVER &&
!vdev_rebuild_active(spa->spa_root_vdev) &&
(!dsl_scan_resilvering(dp) ||
!spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER)))
dsl_scan_restart_resilver(dp, 0);
if (tasks & SPA_ASYNC_INITIALIZE_RESTART) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_initialize_restart(spa->spa_root_vdev);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_exit(&spa_namespace_lock);
}
if (tasks & SPA_ASYNC_TRIM_RESTART) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_trim_restart(spa->spa_root_vdev);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_exit(&spa_namespace_lock);
}
if (tasks & SPA_ASYNC_AUTOTRIM_RESTART) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_autotrim_restart(spa);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_exit(&spa_namespace_lock);
}
/*
* Kick off L2 cache whole device TRIM.
*/
if (tasks & SPA_ASYNC_L2CACHE_TRIM) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
vdev_trim_l2arc(spa);
spa_config_exit(spa, SCL_CONFIG, FTAG);
mutex_exit(&spa_namespace_lock);
}
/*
* Kick off L2 cache rebuilding.
*/
if (tasks & SPA_ASYNC_L2CACHE_REBUILD) {
mutex_enter(&spa_namespace_lock);
spa_config_enter(spa, SCL_L2ARC, FTAG, RW_READER);
l2arc_spa_rebuild_start(spa);
spa_config_exit(spa, SCL_L2ARC, FTAG);
mutex_exit(&spa_namespace_lock);
}
/*
* Let the world know that we're done.
*/
mutex_enter(&spa->spa_async_lock);
spa->spa_async_thread = NULL;
cv_broadcast(&spa->spa_async_cv);
mutex_exit(&spa->spa_async_lock);
thread_exit();
}
void
spa_async_suspend(spa_t *spa)
{
mutex_enter(&spa->spa_async_lock);
spa->spa_async_suspended++;
while (spa->spa_async_thread != NULL)
cv_wait(&spa->spa_async_cv, &spa->spa_async_lock);
mutex_exit(&spa->spa_async_lock);
spa_vdev_remove_suspend(spa);
zthr_t *condense_thread = spa->spa_condense_zthr;
if (condense_thread != NULL)
zthr_cancel(condense_thread);
zthr_t *discard_thread = spa->spa_checkpoint_discard_zthr;
if (discard_thread != NULL)
zthr_cancel(discard_thread);
zthr_t *ll_delete_thread = spa->spa_livelist_delete_zthr;
if (ll_delete_thread != NULL)
zthr_cancel(ll_delete_thread);
zthr_t *ll_condense_thread = spa->spa_livelist_condense_zthr;
if (ll_condense_thread != NULL)
zthr_cancel(ll_condense_thread);
}
void
spa_async_resume(spa_t *spa)
{
mutex_enter(&spa->spa_async_lock);
ASSERT(spa->spa_async_suspended != 0);
spa->spa_async_suspended--;
mutex_exit(&spa->spa_async_lock);
spa_restart_removal(spa);
zthr_t *condense_thread = spa->spa_condense_zthr;
if (condense_thread != NULL)
zthr_resume(condense_thread);
zthr_t *discard_thread = spa->spa_checkpoint_discard_zthr;
if (discard_thread != NULL)
zthr_resume(discard_thread);
zthr_t *ll_delete_thread = spa->spa_livelist_delete_zthr;
if (ll_delete_thread != NULL)
zthr_resume(ll_delete_thread);
zthr_t *ll_condense_thread = spa->spa_livelist_condense_zthr;
if (ll_condense_thread != NULL)
zthr_resume(ll_condense_thread);
}
static boolean_t
spa_async_tasks_pending(spa_t *spa)
{
uint_t non_config_tasks;
uint_t config_task;
boolean_t config_task_suspended;
non_config_tasks = spa->spa_async_tasks & ~SPA_ASYNC_CONFIG_UPDATE;
config_task = spa->spa_async_tasks & SPA_ASYNC_CONFIG_UPDATE;
if (spa->spa_ccw_fail_time == 0) {
config_task_suspended = B_FALSE;
} else {
config_task_suspended =
(gethrtime() - spa->spa_ccw_fail_time) <
((hrtime_t)zfs_ccw_retry_interval * NANOSEC);
}
return (non_config_tasks || (config_task && !config_task_suspended));
}
static void
spa_async_dispatch(spa_t *spa)
{
mutex_enter(&spa->spa_async_lock);
if (spa_async_tasks_pending(spa) &&
!spa->spa_async_suspended &&
spa->spa_async_thread == NULL)
spa->spa_async_thread = thread_create(NULL, 0,
spa_async_thread, spa, 0, &p0, TS_RUN, maxclsyspri);
mutex_exit(&spa->spa_async_lock);
}
void
spa_async_request(spa_t *spa, int task)
{
zfs_dbgmsg("spa=%s async request task=%u", spa->spa_name, task);
mutex_enter(&spa->spa_async_lock);
spa->spa_async_tasks |= task;
mutex_exit(&spa->spa_async_lock);
}
int
spa_async_tasks(spa_t *spa)
{
return (spa->spa_async_tasks);
}
/*
* ==========================================================================
* SPA syncing routines
* ==========================================================================
*/
static int
bpobj_enqueue_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
bpobj_t *bpo = arg;
bpobj_enqueue(bpo, bp, bp_freed, tx);
return (0);
}
int
bpobj_enqueue_alloc_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
return (bpobj_enqueue_cb(arg, bp, B_FALSE, tx));
}
int
bpobj_enqueue_free_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
return (bpobj_enqueue_cb(arg, bp, B_TRUE, tx));
}
static int
spa_free_sync_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
zio_t *pio = arg;
zio_nowait(zio_free_sync(pio, pio->io_spa, dmu_tx_get_txg(tx), bp,
pio->io_flags));
return (0);
}
static int
bpobj_spa_free_sync_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed,
dmu_tx_t *tx)
{
ASSERT(!bp_freed);
return (spa_free_sync_cb(arg, bp, tx));
}
/*
* Note: this simple function is not inlined to make it easier to dtrace the
* amount of time spent syncing frees.
*/
static void
spa_sync_frees(spa_t *spa, bplist_t *bpl, dmu_tx_t *tx)
{
zio_t *zio = zio_root(spa, NULL, NULL, 0);
bplist_iterate(bpl, spa_free_sync_cb, zio, tx);
VERIFY(zio_wait(zio) == 0);
}
/*
* Note: this simple function is not inlined to make it easier to dtrace the
* amount of time spent syncing deferred frees.
*/
static void
spa_sync_deferred_frees(spa_t *spa, dmu_tx_t *tx)
{
if (spa_sync_pass(spa) != 1)
return;
/*
* Note:
* If the log space map feature is active, we stop deferring
* frees to the next TXG and therefore running this function
* would be considered a no-op as spa_deferred_bpobj should
* not have any entries.
*
* That said we run this function anyway (instead of returning
* immediately) for the edge-case scenario where we just
* activated the log space map feature in this TXG but we have
* deferred frees from the previous TXG.
*/
zio_t *zio = zio_root(spa, NULL, NULL, 0);
VERIFY3U(bpobj_iterate(&spa->spa_deferred_bpobj,
bpobj_spa_free_sync_cb, zio, tx), ==, 0);
VERIFY0(zio_wait(zio));
}
static void
spa_sync_nvlist(spa_t *spa, uint64_t obj, nvlist_t *nv, dmu_tx_t *tx)
{
char *packed = NULL;
size_t bufsize;
size_t nvsize = 0;
dmu_buf_t *db;
VERIFY(nvlist_size(nv, &nvsize, NV_ENCODE_XDR) == 0);
/*
* Write full (SPA_CONFIG_BLOCKSIZE) blocks of configuration
* information. This avoids the dmu_buf_will_dirty() path and
* saves us a pre-read to get data we don't actually care about.
*/
bufsize = P2ROUNDUP((uint64_t)nvsize, SPA_CONFIG_BLOCKSIZE);
packed = vmem_alloc(bufsize, KM_SLEEP);
VERIFY(nvlist_pack(nv, &packed, &nvsize, NV_ENCODE_XDR,
KM_SLEEP) == 0);
memset(packed + nvsize, 0, bufsize - nvsize);
dmu_write(spa->spa_meta_objset, obj, 0, bufsize, packed, tx);
vmem_free(packed, bufsize);
VERIFY(0 == dmu_bonus_hold(spa->spa_meta_objset, obj, FTAG, &db));
dmu_buf_will_dirty(db, tx);
*(uint64_t *)db->db_data = nvsize;
dmu_buf_rele(db, FTAG);
}
static void
spa_sync_aux_dev(spa_t *spa, spa_aux_vdev_t *sav, dmu_tx_t *tx,
const char *config, const char *entry)
{
nvlist_t *nvroot;
nvlist_t **list;
int i;
if (!sav->sav_sync)
return;
/*
* Update the MOS nvlist describing the list of available devices.
* spa_validate_aux() will have already made sure this nvlist is
* valid and the vdevs are labeled appropriately.
*/
if (sav->sav_object == 0) {
sav->sav_object = dmu_object_alloc(spa->spa_meta_objset,
DMU_OT_PACKED_NVLIST, 1 << 14, DMU_OT_PACKED_NVLIST_SIZE,
sizeof (uint64_t), tx);
VERIFY(zap_update(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, entry, sizeof (uint64_t), 1,
&sav->sav_object, tx) == 0);
}
nvroot = fnvlist_alloc();
if (sav->sav_count == 0) {
fnvlist_add_nvlist_array(nvroot, config,
(const nvlist_t * const *)NULL, 0);
} else {
list = kmem_alloc(sav->sav_count*sizeof (void *), KM_SLEEP);
for (i = 0; i < sav->sav_count; i++)
list[i] = vdev_config_generate(spa, sav->sav_vdevs[i],
B_FALSE, VDEV_CONFIG_L2CACHE);
fnvlist_add_nvlist_array(nvroot, config,
(const nvlist_t * const *)list, sav->sav_count);
for (i = 0; i < sav->sav_count; i++)
nvlist_free(list[i]);
kmem_free(list, sav->sav_count * sizeof (void *));
}
spa_sync_nvlist(spa, sav->sav_object, nvroot, tx);
nvlist_free(nvroot);
sav->sav_sync = B_FALSE;
}
/*
* Rebuild spa's all-vdev ZAP from the vdev ZAPs indicated in each vdev_t.
* The all-vdev ZAP must be empty.
*/
static void
spa_avz_build(vdev_t *vd, uint64_t avz, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
+ if (vd->vdev_root_zap != 0 &&
+ spa_feature_is_active(spa, SPA_FEATURE_AVZ_V2)) {
+ VERIFY0(zap_add_int(spa->spa_meta_objset, avz,
+ vd->vdev_root_zap, tx));
+ }
if (vd->vdev_top_zap != 0) {
VERIFY0(zap_add_int(spa->spa_meta_objset, avz,
vd->vdev_top_zap, tx));
}
if (vd->vdev_leaf_zap != 0) {
VERIFY0(zap_add_int(spa->spa_meta_objset, avz,
vd->vdev_leaf_zap, tx));
}
for (uint64_t i = 0; i < vd->vdev_children; i++) {
spa_avz_build(vd->vdev_child[i], avz, tx);
}
}
static void
spa_sync_config_object(spa_t *spa, dmu_tx_t *tx)
{
nvlist_t *config;
/*
* If the pool is being imported from a pre-per-vdev-ZAP version of ZFS,
* its config may not be dirty but we still need to build per-vdev ZAPs.
* Similarly, if the pool is being assembled (e.g. after a split), we
* need to rebuild the AVZ although the config may not be dirty.
*/
if (list_is_empty(&spa->spa_config_dirty_list) &&
spa->spa_avz_action == AVZ_ACTION_NONE)
return;
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
ASSERT(spa->spa_avz_action == AVZ_ACTION_NONE ||
spa->spa_avz_action == AVZ_ACTION_INITIALIZE ||
spa->spa_all_vdev_zaps != 0);
if (spa->spa_avz_action == AVZ_ACTION_REBUILD) {
/* Make and build the new AVZ */
uint64_t new_avz = zap_create(spa->spa_meta_objset,
DMU_OTN_ZAP_METADATA, DMU_OT_NONE, 0, tx);
spa_avz_build(spa->spa_root_vdev, new_avz, tx);
/* Diff old AVZ with new one */
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, spa->spa_meta_objset,
spa->spa_all_vdev_zaps);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
uint64_t vdzap = za.za_first_integer;
if (zap_lookup_int(spa->spa_meta_objset, new_avz,
vdzap) == ENOENT) {
/*
* ZAP is listed in old AVZ but not in new one;
* destroy it
*/
VERIFY0(zap_destroy(spa->spa_meta_objset, vdzap,
tx));
}
}
zap_cursor_fini(&zc);
/* Destroy the old AVZ */
VERIFY0(zap_destroy(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, tx));
/* Replace the old AVZ in the dir obj with the new one */
VERIFY0(zap_update(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_VDEV_ZAP_MAP,
sizeof (new_avz), 1, &new_avz, tx));
spa->spa_all_vdev_zaps = new_avz;
} else if (spa->spa_avz_action == AVZ_ACTION_DESTROY) {
zap_cursor_t zc;
zap_attribute_t za;
/* Walk through the AVZ and destroy all listed ZAPs */
for (zap_cursor_init(&zc, spa->spa_meta_objset,
spa->spa_all_vdev_zaps);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
uint64_t zap = za.za_first_integer;
VERIFY0(zap_destroy(spa->spa_meta_objset, zap, tx));
}
zap_cursor_fini(&zc);
/* Destroy and unlink the AVZ itself */
VERIFY0(zap_destroy(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, tx));
VERIFY0(zap_remove(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_VDEV_ZAP_MAP, tx));
spa->spa_all_vdev_zaps = 0;
}
if (spa->spa_all_vdev_zaps == 0) {
spa->spa_all_vdev_zaps = zap_create_link(spa->spa_meta_objset,
DMU_OTN_ZAP_METADATA, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_VDEV_ZAP_MAP, tx);
}
spa->spa_avz_action = AVZ_ACTION_NONE;
/* Create ZAPs for vdevs that don't have them. */
vdev_construct_zaps(spa->spa_root_vdev, tx);
config = spa_config_generate(spa, spa->spa_root_vdev,
dmu_tx_get_txg(tx), B_FALSE);
/*
* If we're upgrading the spa version then make sure that
* the config object gets updated with the correct version.
*/
if (spa->spa_ubsync.ub_version < spa->spa_uberblock.ub_version)
fnvlist_add_uint64(config, ZPOOL_CONFIG_VERSION,
spa->spa_uberblock.ub_version);
spa_config_exit(spa, SCL_STATE, FTAG);
nvlist_free(spa->spa_config_syncing);
spa->spa_config_syncing = config;
spa_sync_nvlist(spa, spa->spa_config_object, config, tx);
}
static void
spa_sync_version(void *arg, dmu_tx_t *tx)
{
uint64_t *versionp = arg;
uint64_t version = *versionp;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
/*
* Setting the version is special cased when first creating the pool.
*/
ASSERT(tx->tx_txg != TXG_INITIAL);
ASSERT(SPA_VERSION_IS_SUPPORTED(version));
ASSERT(version >= spa_version(spa));
spa->spa_uberblock.ub_version = version;
vdev_config_dirty(spa->spa_root_vdev);
spa_history_log_internal(spa, "set", tx, "version=%lld",
(longlong_t)version);
}
/*
* Set zpool properties.
*/
static void
spa_sync_props(void *arg, dmu_tx_t *tx)
{
nvlist_t *nvp = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
objset_t *mos = spa->spa_meta_objset;
nvpair_t *elem = NULL;
mutex_enter(&spa->spa_props_lock);
while ((elem = nvlist_next_nvpair(nvp, elem))) {
uint64_t intval;
const char *strval, *fname;
zpool_prop_t prop;
const char *propname;
const char *elemname = nvpair_name(elem);
zprop_type_t proptype;
spa_feature_t fid;
switch (prop = zpool_name_to_prop(elemname)) {
case ZPOOL_PROP_VERSION:
intval = fnvpair_value_uint64(elem);
/*
* The version is synced separately before other
* properties and should be correct by now.
*/
ASSERT3U(spa_version(spa), >=, intval);
break;
case ZPOOL_PROP_ALTROOT:
/*
* 'altroot' is a non-persistent property. It should
* have been set temporarily at creation or import time.
*/
ASSERT(spa->spa_root != NULL);
break;
case ZPOOL_PROP_READONLY:
case ZPOOL_PROP_CACHEFILE:
/*
* 'readonly' and 'cachefile' are also non-persistent
* properties.
*/
break;
case ZPOOL_PROP_COMMENT:
strval = fnvpair_value_string(elem);
if (spa->spa_comment != NULL)
spa_strfree(spa->spa_comment);
spa->spa_comment = spa_strdup(strval);
/*
* We need to dirty the configuration on all the vdevs
* so that their labels get updated. We also need to
* update the cache file to keep it in sync with the
* MOS version. It's unnecessary to do this for pool
* creation since the vdev's configuration has already
* been dirtied.
*/
if (tx->tx_txg != TXG_INITIAL) {
vdev_config_dirty(spa->spa_root_vdev);
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
spa_history_log_internal(spa, "set", tx,
"%s=%s", elemname, strval);
break;
case ZPOOL_PROP_COMPATIBILITY:
strval = fnvpair_value_string(elem);
if (spa->spa_compatibility != NULL)
spa_strfree(spa->spa_compatibility);
spa->spa_compatibility = spa_strdup(strval);
/*
* Dirty the configuration on vdevs as above.
*/
if (tx->tx_txg != TXG_INITIAL) {
vdev_config_dirty(spa->spa_root_vdev);
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
spa_history_log_internal(spa, "set", tx,
"%s=%s", nvpair_name(elem), strval);
break;
case ZPOOL_PROP_INVAL:
if (zpool_prop_feature(elemname)) {
fname = strchr(elemname, '@') + 1;
VERIFY0(zfeature_lookup_name(fname, &fid));
spa_feature_enable(spa, fid, tx);
spa_history_log_internal(spa, "set", tx,
"%s=enabled", elemname);
break;
} else if (!zfs_prop_user(elemname)) {
ASSERT(zpool_prop_feature(elemname));
break;
}
zfs_fallthrough;
default:
/*
* Set pool property values in the poolprops mos object.
*/
if (spa->spa_pool_props_object == 0) {
spa->spa_pool_props_object =
zap_create_link(mos, DMU_OT_POOL_PROPS,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_PROPS,
tx);
}
/* normalize the property name */
propname = zpool_prop_to_name(prop);
proptype = zpool_prop_get_type(prop);
if (prop == ZPOOL_PROP_INVAL &&
zfs_prop_user(elemname)) {
propname = elemname;
proptype = PROP_TYPE_STRING;
}
if (nvpair_type(elem) == DATA_TYPE_STRING) {
ASSERT(proptype == PROP_TYPE_STRING);
strval = fnvpair_value_string(elem);
VERIFY0(zap_update(mos,
spa->spa_pool_props_object, propname,
1, strlen(strval) + 1, strval, tx));
spa_history_log_internal(spa, "set", tx,
"%s=%s", elemname, strval);
} else if (nvpair_type(elem) == DATA_TYPE_UINT64) {
intval = fnvpair_value_uint64(elem);
if (proptype == PROP_TYPE_INDEX) {
const char *unused;
VERIFY0(zpool_prop_index_to_string(
prop, intval, &unused));
}
VERIFY0(zap_update(mos,
spa->spa_pool_props_object, propname,
8, 1, &intval, tx));
spa_history_log_internal(spa, "set", tx,
"%s=%lld", elemname,
(longlong_t)intval);
switch (prop) {
case ZPOOL_PROP_DELEGATION:
spa->spa_delegation = intval;
break;
case ZPOOL_PROP_BOOTFS:
spa->spa_bootfs = intval;
break;
case ZPOOL_PROP_FAILUREMODE:
spa->spa_failmode = intval;
break;
case ZPOOL_PROP_AUTOTRIM:
spa->spa_autotrim = intval;
spa_async_request(spa,
SPA_ASYNC_AUTOTRIM_RESTART);
break;
case ZPOOL_PROP_AUTOEXPAND:
spa->spa_autoexpand = intval;
if (tx->tx_txg != TXG_INITIAL)
spa_async_request(spa,
SPA_ASYNC_AUTOEXPAND);
break;
case ZPOOL_PROP_MULTIHOST:
spa->spa_multihost = intval;
break;
default:
break;
}
} else {
ASSERT(0); /* not allowed */
}
}
}
mutex_exit(&spa->spa_props_lock);
}
/*
* Perform one-time upgrade on-disk changes. spa_version() does not
* reflect the new version this txg, so there must be no changes this
* txg to anything that the upgrade code depends on after it executes.
* Therefore this must be called after dsl_pool_sync() does the sync
* tasks.
*/
static void
spa_sync_upgrades(spa_t *spa, dmu_tx_t *tx)
{
if (spa_sync_pass(spa) != 1)
return;
dsl_pool_t *dp = spa->spa_dsl_pool;
rrw_enter(&dp->dp_config_rwlock, RW_WRITER, FTAG);
if (spa->spa_ubsync.ub_version < SPA_VERSION_ORIGIN &&
spa->spa_uberblock.ub_version >= SPA_VERSION_ORIGIN) {
dsl_pool_create_origin(dp, tx);
/* Keeping the origin open increases spa_minref */
spa->spa_minref += 3;
}
if (spa->spa_ubsync.ub_version < SPA_VERSION_NEXT_CLONES &&
spa->spa_uberblock.ub_version >= SPA_VERSION_NEXT_CLONES) {
dsl_pool_upgrade_clones(dp, tx);
}
if (spa->spa_ubsync.ub_version < SPA_VERSION_DIR_CLONES &&
spa->spa_uberblock.ub_version >= SPA_VERSION_DIR_CLONES) {
dsl_pool_upgrade_dir_clones(dp, tx);
/* Keeping the freedir open increases spa_minref */
spa->spa_minref += 3;
}
if (spa->spa_ubsync.ub_version < SPA_VERSION_FEATURES &&
spa->spa_uberblock.ub_version >= SPA_VERSION_FEATURES) {
spa_feature_create_zap_objects(spa, tx);
}
/*
* LZ4_COMPRESS feature's behaviour was changed to activate_on_enable
* when possibility to use lz4 compression for metadata was added
* Old pools that have this feature enabled must be upgraded to have
* this feature active
*/
if (spa->spa_uberblock.ub_version >= SPA_VERSION_FEATURES) {
boolean_t lz4_en = spa_feature_is_enabled(spa,
SPA_FEATURE_LZ4_COMPRESS);
boolean_t lz4_ac = spa_feature_is_active(spa,
SPA_FEATURE_LZ4_COMPRESS);
if (lz4_en && !lz4_ac)
spa_feature_incr(spa, SPA_FEATURE_LZ4_COMPRESS, tx);
}
/*
* If we haven't written the salt, do so now. Note that the
* feature may not be activated yet, but that's fine since
* the presence of this ZAP entry is backwards compatible.
*/
if (zap_contains(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_CHECKSUM_SALT) == ENOENT) {
VERIFY0(zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CHECKSUM_SALT, 1,
sizeof (spa->spa_cksum_salt.zcs_bytes),
spa->spa_cksum_salt.zcs_bytes, tx));
}
rrw_exit(&dp->dp_config_rwlock, FTAG);
}
static void
vdev_indirect_state_sync_verify(vdev_t *vd)
{
vdev_indirect_mapping_t *vim __maybe_unused = vd->vdev_indirect_mapping;
vdev_indirect_births_t *vib __maybe_unused = vd->vdev_indirect_births;
if (vd->vdev_ops == &vdev_indirect_ops) {
ASSERT(vim != NULL);
ASSERT(vib != NULL);
}
uint64_t obsolete_sm_object = 0;
ASSERT0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (obsolete_sm_object != 0) {
ASSERT(vd->vdev_obsolete_sm != NULL);
ASSERT(vd->vdev_removing ||
vd->vdev_ops == &vdev_indirect_ops);
ASSERT(vdev_indirect_mapping_num_entries(vim) > 0);
ASSERT(vdev_indirect_mapping_bytes_mapped(vim) > 0);
ASSERT3U(obsolete_sm_object, ==,
space_map_object(vd->vdev_obsolete_sm));
ASSERT3U(vdev_indirect_mapping_bytes_mapped(vim), >=,
space_map_allocated(vd->vdev_obsolete_sm));
}
ASSERT(vd->vdev_obsolete_segments != NULL);
/*
* Since frees / remaps to an indirect vdev can only
* happen in syncing context, the obsolete segments
* tree must be empty when we start syncing.
*/
ASSERT0(range_tree_space(vd->vdev_obsolete_segments));
}
/*
* Set the top-level vdev's max queue depth. Evaluate each top-level's
* async write queue depth in case it changed. The max queue depth will
* not change in the middle of syncing out this txg.
*/
static void
spa_sync_adjust_vdev_max_queue_depth(spa_t *spa)
{
ASSERT(spa_writeable(spa));
vdev_t *rvd = spa->spa_root_vdev;
uint32_t max_queue_depth = zfs_vdev_async_write_max_active *
zfs_vdev_queue_depth_pct / 100;
metaslab_class_t *normal = spa_normal_class(spa);
metaslab_class_t *special = spa_special_class(spa);
metaslab_class_t *dedup = spa_dedup_class(spa);
uint64_t slots_per_allocator = 0;
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
metaslab_group_t *mg = tvd->vdev_mg;
if (mg == NULL || !metaslab_group_initialized(mg))
continue;
metaslab_class_t *mc = mg->mg_class;
if (mc != normal && mc != special && mc != dedup)
continue;
/*
* It is safe to do a lock-free check here because only async
* allocations look at mg_max_alloc_queue_depth, and async
* allocations all happen from spa_sync().
*/
for (int i = 0; i < mg->mg_allocators; i++) {
ASSERT0(zfs_refcount_count(
&(mg->mg_allocator[i].mga_alloc_queue_depth)));
}
mg->mg_max_alloc_queue_depth = max_queue_depth;
for (int i = 0; i < mg->mg_allocators; i++) {
mg->mg_allocator[i].mga_cur_max_alloc_queue_depth =
zfs_vdev_def_queue_depth;
}
slots_per_allocator += zfs_vdev_def_queue_depth;
}
for (int i = 0; i < spa->spa_alloc_count; i++) {
ASSERT0(zfs_refcount_count(&normal->mc_allocator[i].
mca_alloc_slots));
ASSERT0(zfs_refcount_count(&special->mc_allocator[i].
mca_alloc_slots));
ASSERT0(zfs_refcount_count(&dedup->mc_allocator[i].
mca_alloc_slots));
normal->mc_allocator[i].mca_alloc_max_slots =
slots_per_allocator;
special->mc_allocator[i].mca_alloc_max_slots =
slots_per_allocator;
dedup->mc_allocator[i].mca_alloc_max_slots =
slots_per_allocator;
}
normal->mc_alloc_throttle_enabled = zio_dva_throttle_enabled;
special->mc_alloc_throttle_enabled = zio_dva_throttle_enabled;
dedup->mc_alloc_throttle_enabled = zio_dva_throttle_enabled;
}
static void
spa_sync_condense_indirect(spa_t *spa, dmu_tx_t *tx)
{
ASSERT(spa_writeable(spa));
vdev_t *rvd = spa->spa_root_vdev;
for (int c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
vdev_indirect_state_sync_verify(vd);
if (vdev_indirect_should_condense(vd)) {
spa_condense_indirect_start_sync(vd, tx);
break;
}
}
}
static void
spa_sync_iterate_to_convergence(spa_t *spa, dmu_tx_t *tx)
{
objset_t *mos = spa->spa_meta_objset;
dsl_pool_t *dp = spa->spa_dsl_pool;
uint64_t txg = tx->tx_txg;
bplist_t *free_bpl = &spa->spa_free_bplist[txg & TXG_MASK];
do {
int pass = ++spa->spa_sync_pass;
spa_sync_config_object(spa, tx);
spa_sync_aux_dev(spa, &spa->spa_spares, tx,
ZPOOL_CONFIG_SPARES, DMU_POOL_SPARES);
spa_sync_aux_dev(spa, &spa->spa_l2cache, tx,
ZPOOL_CONFIG_L2CACHE, DMU_POOL_L2CACHE);
spa_errlog_sync(spa, txg);
dsl_pool_sync(dp, txg);
if (pass < zfs_sync_pass_deferred_free ||
spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP)) {
/*
* If the log space map feature is active we don't
* care about deferred frees and the deferred bpobj
* as the log space map should effectively have the
* same results (i.e. appending only to one object).
*/
spa_sync_frees(spa, free_bpl, tx);
} else {
/*
* We can not defer frees in pass 1, because
* we sync the deferred frees later in pass 1.
*/
ASSERT3U(pass, >, 1);
bplist_iterate(free_bpl, bpobj_enqueue_alloc_cb,
&spa->spa_deferred_bpobj, tx);
}
brt_sync(spa, txg);
ddt_sync(spa, txg);
dsl_scan_sync(dp, tx);
svr_sync(spa, tx);
spa_sync_upgrades(spa, tx);
spa_flush_metaslabs(spa, tx);
vdev_t *vd = NULL;
while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, txg))
!= NULL)
vdev_sync(vd, txg);
/*
* Note: We need to check if the MOS is dirty because we could
* have marked the MOS dirty without updating the uberblock
* (e.g. if we have sync tasks but no dirty user data). We need
* to check the uberblock's rootbp because it is updated if we
* have synced out dirty data (though in this case the MOS will
* most likely also be dirty due to second order effects, we
* don't want to rely on that here).
*/
if (pass == 1 &&
spa->spa_uberblock.ub_rootbp.blk_birth < txg &&
!dmu_objset_is_dirty(mos, txg)) {
/*
* Nothing changed on the first pass, therefore this
* TXG is a no-op. Avoid syncing deferred frees, so
* that we can keep this TXG as a no-op.
*/
ASSERT(txg_list_empty(&dp->dp_dirty_datasets, txg));
ASSERT(txg_list_empty(&dp->dp_dirty_dirs, txg));
ASSERT(txg_list_empty(&dp->dp_sync_tasks, txg));
ASSERT(txg_list_empty(&dp->dp_early_sync_tasks, txg));
break;
}
spa_sync_deferred_frees(spa, tx);
} while (dmu_objset_is_dirty(mos, txg));
}
/*
* Rewrite the vdev configuration (which includes the uberblock) to
* commit the transaction group.
*
* If there are no dirty vdevs, we sync the uberblock to a few random
* top-level vdevs that are known to be visible in the config cache
* (see spa_vdev_add() for a complete description). If there *are* dirty
* vdevs, sync the uberblock to all vdevs.
*/
static void
spa_sync_rewrite_vdev_config(spa_t *spa, dmu_tx_t *tx)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t txg = tx->tx_txg;
for (;;) {
int error = 0;
/*
* We hold SCL_STATE to prevent vdev open/close/etc.
* while we're attempting to write the vdev labels.
*/
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
if (list_is_empty(&spa->spa_config_dirty_list)) {
vdev_t *svd[SPA_SYNC_MIN_VDEVS] = { NULL };
int svdcount = 0;
int children = rvd->vdev_children;
int c0 = random_in_range(children);
for (int c = 0; c < children; c++) {
vdev_t *vd =
rvd->vdev_child[(c0 + c) % children];
/* Stop when revisiting the first vdev */
if (c > 0 && svd[0] == vd)
break;
if (vd->vdev_ms_array == 0 ||
vd->vdev_islog ||
!vdev_is_concrete(vd))
continue;
svd[svdcount++] = vd;
if (svdcount == SPA_SYNC_MIN_VDEVS)
break;
}
error = vdev_config_sync(svd, svdcount, txg);
} else {
error = vdev_config_sync(rvd->vdev_child,
rvd->vdev_children, txg);
}
if (error == 0)
spa->spa_last_synced_guid = rvd->vdev_guid;
spa_config_exit(spa, SCL_STATE, FTAG);
if (error == 0)
break;
zio_suspend(spa, NULL, ZIO_SUSPEND_IOERR);
zio_resume_wait(spa);
}
}
/*
* Sync the specified transaction group. New blocks may be dirtied as
* part of the process, so we iterate until it converges.
*/
void
spa_sync(spa_t *spa, uint64_t txg)
{
vdev_t *vd = NULL;
VERIFY(spa_writeable(spa));
/*
* Wait for i/os issued in open context that need to complete
* before this txg syncs.
*/
(void) zio_wait(spa->spa_txg_zio[txg & TXG_MASK]);
spa->spa_txg_zio[txg & TXG_MASK] = zio_root(spa, NULL, NULL,
ZIO_FLAG_CANFAIL);
/*
* Now that there can be no more cloning in this transaction group,
* but we are still before issuing frees, we can process pending BRT
* updates.
*/
brt_pending_apply(spa, txg);
/*
* Lock out configuration changes.
*/
spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER);
spa->spa_syncing_txg = txg;
spa->spa_sync_pass = 0;
for (int i = 0; i < spa->spa_alloc_count; i++) {
mutex_enter(&spa->spa_allocs[i].spaa_lock);
VERIFY0(avl_numnodes(&spa->spa_allocs[i].spaa_tree));
mutex_exit(&spa->spa_allocs[i].spaa_lock);
}
/*
* If there are any pending vdev state changes, convert them
* into config changes that go out with this transaction group.
*/
spa_config_enter(spa, SCL_STATE, FTAG, RW_READER);
while ((vd = list_head(&spa->spa_state_dirty_list)) != NULL) {
/* Avoid holding the write lock unless actually necessary */
if (vd->vdev_aux == NULL) {
vdev_state_clean(vd);
vdev_config_dirty(vd);
continue;
}
/*
* We need the write lock here because, for aux vdevs,
* calling vdev_config_dirty() modifies sav_config.
* This is ugly and will become unnecessary when we
* eliminate the aux vdev wart by integrating all vdevs
* into the root vdev tree.
*/
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_WRITER);
while ((vd = list_head(&spa->spa_state_dirty_list)) != NULL) {
vdev_state_clean(vd);
vdev_config_dirty(vd);
}
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
}
spa_config_exit(spa, SCL_STATE, FTAG);
dsl_pool_t *dp = spa->spa_dsl_pool;
dmu_tx_t *tx = dmu_tx_create_assigned(dp, txg);
spa->spa_sync_starttime = gethrtime();
taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid);
spa->spa_deadman_tqid = taskq_dispatch_delay(system_delay_taskq,
spa_deadman, spa, TQ_SLEEP, ddi_get_lbolt() +
NSEC_TO_TICK(spa->spa_deadman_synctime));
/*
* If we are upgrading to SPA_VERSION_RAIDZ_DEFLATE this txg,
* set spa_deflate if we have no raid-z vdevs.
*/
if (spa->spa_ubsync.ub_version < SPA_VERSION_RAIDZ_DEFLATE &&
spa->spa_uberblock.ub_version >= SPA_VERSION_RAIDZ_DEFLATE) {
vdev_t *rvd = spa->spa_root_vdev;
int i;
for (i = 0; i < rvd->vdev_children; i++) {
vd = rvd->vdev_child[i];
if (vd->vdev_deflate_ratio != SPA_MINBLOCKSIZE)
break;
}
if (i == rvd->vdev_children) {
spa->spa_deflate = TRUE;
VERIFY0(zap_add(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_DEFLATE,
sizeof (uint64_t), 1, &spa->spa_deflate, tx));
}
}
spa_sync_adjust_vdev_max_queue_depth(spa);
spa_sync_condense_indirect(spa, tx);
spa_sync_iterate_to_convergence(spa, tx);
#ifdef ZFS_DEBUG
if (!list_is_empty(&spa->spa_config_dirty_list)) {
/*
* Make sure that the number of ZAPs for all the vdevs matches
* the number of ZAPs in the per-vdev ZAP list. This only gets
* called if the config is dirty; otherwise there may be
* outstanding AVZ operations that weren't completed in
* spa_sync_config_object.
*/
uint64_t all_vdev_zap_entry_count;
ASSERT0(zap_count(spa->spa_meta_objset,
spa->spa_all_vdev_zaps, &all_vdev_zap_entry_count));
ASSERT3U(vdev_count_verify_zaps(spa->spa_root_vdev), ==,
all_vdev_zap_entry_count);
}
#endif
if (spa->spa_vdev_removal != NULL) {
ASSERT0(spa->spa_vdev_removal->svr_bytes_done[txg & TXG_MASK]);
}
spa_sync_rewrite_vdev_config(spa, tx);
dmu_tx_commit(tx);
taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid);
spa->spa_deadman_tqid = 0;
/*
* Clear the dirty config list.
*/
while ((vd = list_head(&spa->spa_config_dirty_list)) != NULL)
vdev_config_clean(vd);
/*
* Now that the new config has synced transactionally,
* let it become visible to the config cache.
*/
if (spa->spa_config_syncing != NULL) {
spa_config_set(spa, spa->spa_config_syncing);
spa->spa_config_txg = txg;
spa->spa_config_syncing = NULL;
}
dsl_pool_sync_done(dp, txg);
for (int i = 0; i < spa->spa_alloc_count; i++) {
mutex_enter(&spa->spa_allocs[i].spaa_lock);
VERIFY0(avl_numnodes(&spa->spa_allocs[i].spaa_tree));
mutex_exit(&spa->spa_allocs[i].spaa_lock);
}
/*
* Update usable space statistics.
*/
while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, TXG_CLEAN(txg)))
!= NULL)
vdev_sync_done(vd, txg);
metaslab_class_evict_old(spa->spa_normal_class, txg);
metaslab_class_evict_old(spa->spa_log_class, txg);
spa_sync_close_syncing_log_sm(spa);
spa_update_dspace(spa);
if (spa_get_autotrim(spa) == SPA_AUTOTRIM_ON)
vdev_autotrim_kick(spa);
/*
* It had better be the case that we didn't dirty anything
* since vdev_config_sync().
*/
ASSERT(txg_list_empty(&dp->dp_dirty_datasets, txg));
ASSERT(txg_list_empty(&dp->dp_dirty_dirs, txg));
ASSERT(txg_list_empty(&spa->spa_vdev_txg_list, txg));
while (zfs_pause_spa_sync)
delay(1);
spa->spa_sync_pass = 0;
/*
* Update the last synced uberblock here. We want to do this at
* the end of spa_sync() so that consumers of spa_last_synced_txg()
* will be guaranteed that all the processing associated with
* that txg has been completed.
*/
spa->spa_ubsync = spa->spa_uberblock;
spa_config_exit(spa, SCL_CONFIG, FTAG);
spa_handle_ignored_writes(spa);
/*
* If any async tasks have been requested, kick them off.
*/
spa_async_dispatch(spa);
}
/*
* Sync all pools. We don't want to hold the namespace lock across these
* operations, so we take a reference on the spa_t and drop the lock during the
* sync.
*/
void
spa_sync_allpools(void)
{
spa_t *spa = NULL;
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(spa)) != NULL) {
if (spa_state(spa) != POOL_STATE_ACTIVE ||
!spa_writeable(spa) || spa_suspended(spa))
continue;
spa_open_ref(spa, FTAG);
mutex_exit(&spa_namespace_lock);
txg_wait_synced(spa_get_dsl(spa), 0);
mutex_enter(&spa_namespace_lock);
spa_close(spa, FTAG);
}
mutex_exit(&spa_namespace_lock);
}
/*
* ==========================================================================
* Miscellaneous routines
* ==========================================================================
*/
/*
* Remove all pools in the system.
*/
void
spa_evict_all(void)
{
spa_t *spa;
/*
* Remove all cached state. All pools should be closed now,
* so every spa in the AVL tree should be unreferenced.
*/
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(NULL)) != NULL) {
/*
* Stop async tasks. The async thread may need to detach
* a device that's been replaced, which requires grabbing
* spa_namespace_lock, so we must drop it here.
*/
spa_open_ref(spa, FTAG);
mutex_exit(&spa_namespace_lock);
spa_async_suspend(spa);
mutex_enter(&spa_namespace_lock);
spa_close(spa, FTAG);
if (spa->spa_state != POOL_STATE_UNINITIALIZED) {
spa_unload(spa);
spa_deactivate(spa);
}
spa_remove(spa);
}
mutex_exit(&spa_namespace_lock);
}
vdev_t *
spa_lookup_by_guid(spa_t *spa, uint64_t guid, boolean_t aux)
{
vdev_t *vd;
int i;
if ((vd = vdev_lookup_by_guid(spa->spa_root_vdev, guid)) != NULL)
return (vd);
if (aux) {
for (i = 0; i < spa->spa_l2cache.sav_count; i++) {
vd = spa->spa_l2cache.sav_vdevs[i];
if (vd->vdev_guid == guid)
return (vd);
}
for (i = 0; i < spa->spa_spares.sav_count; i++) {
vd = spa->spa_spares.sav_vdevs[i];
if (vd->vdev_guid == guid)
return (vd);
}
}
return (NULL);
}
void
spa_upgrade(spa_t *spa, uint64_t version)
{
ASSERT(spa_writeable(spa));
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
/*
* This should only be called for a non-faulted pool, and since a
* future version would result in an unopenable pool, this shouldn't be
* possible.
*/
ASSERT(SPA_VERSION_IS_SUPPORTED(spa->spa_uberblock.ub_version));
ASSERT3U(version, >=, spa->spa_uberblock.ub_version);
spa->spa_uberblock.ub_version = version;
vdev_config_dirty(spa->spa_root_vdev);
spa_config_exit(spa, SCL_ALL, FTAG);
txg_wait_synced(spa_get_dsl(spa), 0);
}
static boolean_t
spa_has_aux_vdev(spa_t *spa, uint64_t guid, spa_aux_vdev_t *sav)
{
(void) spa;
int i;
uint64_t vdev_guid;
for (i = 0; i < sav->sav_count; i++)
if (sav->sav_vdevs[i]->vdev_guid == guid)
return (B_TRUE);
for (i = 0; i < sav->sav_npending; i++) {
if (nvlist_lookup_uint64(sav->sav_pending[i], ZPOOL_CONFIG_GUID,
&vdev_guid) == 0 && vdev_guid == guid)
return (B_TRUE);
}
return (B_FALSE);
}
boolean_t
spa_has_l2cache(spa_t *spa, uint64_t guid)
{
return (spa_has_aux_vdev(spa, guid, &spa->spa_l2cache));
}
boolean_t
spa_has_spare(spa_t *spa, uint64_t guid)
{
return (spa_has_aux_vdev(spa, guid, &spa->spa_spares));
}
/*
* Check if a pool has an active shared spare device.
* Note: reference count of an active spare is 2, as a spare and as a replace
*/
static boolean_t
spa_has_active_shared_spare(spa_t *spa)
{
int i, refcnt;
uint64_t pool;
spa_aux_vdev_t *sav = &spa->spa_spares;
for (i = 0; i < sav->sav_count; i++) {
if (spa_spare_exists(sav->sav_vdevs[i]->vdev_guid, &pool,
&refcnt) && pool != 0ULL && pool == spa_guid(spa) &&
refcnt > 2)
return (B_TRUE);
}
return (B_FALSE);
}
uint64_t
spa_total_metaslabs(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t m = 0;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
vdev_t *vd = rvd->vdev_child[c];
if (!vdev_is_concrete(vd))
continue;
m += vd->vdev_ms_count;
}
return (m);
}
/*
* Notify any waiting threads that some activity has switched from being in-
* progress to not-in-progress so that the thread can wake up and determine
* whether it is finished waiting.
*/
void
spa_notify_waiters(spa_t *spa)
{
/*
* Acquiring spa_activities_lock here prevents the cv_broadcast from
* happening between the waiting thread's check and cv_wait.
*/
mutex_enter(&spa->spa_activities_lock);
cv_broadcast(&spa->spa_activities_cv);
mutex_exit(&spa->spa_activities_lock);
}
/*
* Notify any waiting threads that the pool is exporting, and then block until
* they are finished using the spa_t.
*/
void
spa_wake_waiters(spa_t *spa)
{
mutex_enter(&spa->spa_activities_lock);
spa->spa_waiters_cancel = B_TRUE;
cv_broadcast(&spa->spa_activities_cv);
while (spa->spa_waiters != 0)
cv_wait(&spa->spa_waiters_cv, &spa->spa_activities_lock);
spa->spa_waiters_cancel = B_FALSE;
mutex_exit(&spa->spa_activities_lock);
}
/* Whether the vdev or any of its descendants are being initialized/trimmed. */
static boolean_t
spa_vdev_activity_in_progress_impl(vdev_t *vd, zpool_wait_activity_t activity)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_config_held(spa, SCL_CONFIG | SCL_STATE, RW_READER));
ASSERT(MUTEX_HELD(&spa->spa_activities_lock));
ASSERT(activity == ZPOOL_WAIT_INITIALIZE ||
activity == ZPOOL_WAIT_TRIM);
kmutex_t *lock = activity == ZPOOL_WAIT_INITIALIZE ?
&vd->vdev_initialize_lock : &vd->vdev_trim_lock;
mutex_exit(&spa->spa_activities_lock);
mutex_enter(lock);
mutex_enter(&spa->spa_activities_lock);
boolean_t in_progress = (activity == ZPOOL_WAIT_INITIALIZE) ?
(vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE) :
(vd->vdev_trim_state == VDEV_TRIM_ACTIVE);
mutex_exit(lock);
if (in_progress)
return (B_TRUE);
for (int i = 0; i < vd->vdev_children; i++) {
if (spa_vdev_activity_in_progress_impl(vd->vdev_child[i],
activity))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* If use_guid is true, this checks whether the vdev specified by guid is
* being initialized/trimmed. Otherwise, it checks whether any vdev in the pool
* is being initialized/trimmed. The caller must hold the config lock and
* spa_activities_lock.
*/
static int
spa_vdev_activity_in_progress(spa_t *spa, boolean_t use_guid, uint64_t guid,
zpool_wait_activity_t activity, boolean_t *in_progress)
{
mutex_exit(&spa->spa_activities_lock);
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
mutex_enter(&spa->spa_activities_lock);
vdev_t *vd;
if (use_guid) {
vd = spa_lookup_by_guid(spa, guid, B_FALSE);
if (vd == NULL || !vd->vdev_ops->vdev_op_leaf) {
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (EINVAL);
}
} else {
vd = spa->spa_root_vdev;
}
*in_progress = spa_vdev_activity_in_progress_impl(vd, activity);
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
return (0);
}
/*
* Locking for waiting threads
* ---------------------------
*
* Waiting threads need a way to check whether a given activity is in progress,
* and then, if it is, wait for it to complete. Each activity will have some
* in-memory representation of the relevant on-disk state which can be used to
* determine whether or not the activity is in progress. The in-memory state and
* the locking used to protect it will be different for each activity, and may
* not be suitable for use with a cvar (e.g., some state is protected by the
* config lock). To allow waiting threads to wait without any races, another
* lock, spa_activities_lock, is used.
*
* When the state is checked, both the activity-specific lock (if there is one)
* and spa_activities_lock are held. In some cases, the activity-specific lock
* is acquired explicitly (e.g. the config lock). In others, the locking is
* internal to some check (e.g. bpobj_is_empty). After checking, the waiting
* thread releases the activity-specific lock and, if the activity is in
* progress, then cv_waits using spa_activities_lock.
*
* The waiting thread is woken when another thread, one completing some
* activity, updates the state of the activity and then calls
* spa_notify_waiters, which will cv_broadcast. This 'completing' thread only
* needs to hold its activity-specific lock when updating the state, and this
* lock can (but doesn't have to) be dropped before calling spa_notify_waiters.
*
* Because spa_notify_waiters acquires spa_activities_lock before broadcasting,
* and because it is held when the waiting thread checks the state of the
* activity, it can never be the case that the completing thread both updates
* the activity state and cv_broadcasts in between the waiting thread's check
* and cv_wait. Thus, a waiting thread can never miss a wakeup.
*
* In order to prevent deadlock, when the waiting thread does its check, in some
* cases it will temporarily drop spa_activities_lock in order to acquire the
* activity-specific lock. The order in which spa_activities_lock and the
* activity specific lock are acquired in the waiting thread is determined by
* the order in which they are acquired in the completing thread; if the
* completing thread calls spa_notify_waiters with the activity-specific lock
* held, then the waiting thread must also acquire the activity-specific lock
* first.
*/
static int
spa_activity_in_progress(spa_t *spa, zpool_wait_activity_t activity,
boolean_t use_tag, uint64_t tag, boolean_t *in_progress)
{
int error = 0;
ASSERT(MUTEX_HELD(&spa->spa_activities_lock));
switch (activity) {
case ZPOOL_WAIT_CKPT_DISCARD:
*in_progress =
(spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT) &&
zap_contains(spa_meta_objset(spa),
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ZPOOL_CHECKPOINT) ==
ENOENT);
break;
case ZPOOL_WAIT_FREE:
*in_progress = ((spa_version(spa) >= SPA_VERSION_DEADLISTS &&
!bpobj_is_empty(&spa->spa_dsl_pool->dp_free_bpobj)) ||
spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY) ||
spa_livelist_delete_check(spa));
break;
case ZPOOL_WAIT_INITIALIZE:
case ZPOOL_WAIT_TRIM:
error = spa_vdev_activity_in_progress(spa, use_tag, tag,
activity, in_progress);
break;
case ZPOOL_WAIT_REPLACE:
mutex_exit(&spa->spa_activities_lock);
spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
mutex_enter(&spa->spa_activities_lock);
*in_progress = vdev_replace_in_progress(spa->spa_root_vdev);
spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);
break;
case ZPOOL_WAIT_REMOVE:
*in_progress = (spa->spa_removing_phys.sr_state ==
DSS_SCANNING);
break;
case ZPOOL_WAIT_RESILVER:
if ((*in_progress = vdev_rebuild_active(spa->spa_root_vdev)))
break;
zfs_fallthrough;
case ZPOOL_WAIT_SCRUB:
{
boolean_t scanning, paused, is_scrub;
dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan;
is_scrub = (scn->scn_phys.scn_func == POOL_SCAN_SCRUB);
scanning = (scn->scn_phys.scn_state == DSS_SCANNING);
paused = dsl_scan_is_paused_scrub(scn);
*in_progress = (scanning && !paused &&
is_scrub == (activity == ZPOOL_WAIT_SCRUB));
break;
}
default:
panic("unrecognized value for activity %d", activity);
}
return (error);
}
static int
spa_wait_common(const char *pool, zpool_wait_activity_t activity,
boolean_t use_tag, uint64_t tag, boolean_t *waited)
{
/*
* The tag is used to distinguish between instances of an activity.
* 'initialize' and 'trim' are the only activities that we use this for.
* The other activities can only have a single instance in progress in a
* pool at one time, making the tag unnecessary.
*
* There can be multiple devices being replaced at once, but since they
* all finish once resilvering finishes, we don't bother keeping track
* of them individually, we just wait for them all to finish.
*/
if (use_tag && activity != ZPOOL_WAIT_INITIALIZE &&
activity != ZPOOL_WAIT_TRIM)
return (EINVAL);
if (activity < 0 || activity >= ZPOOL_WAIT_NUM_ACTIVITIES)
return (EINVAL);
spa_t *spa;
int error = spa_open(pool, &spa, FTAG);
if (error != 0)
return (error);
/*
* Increment the spa's waiter count so that we can call spa_close and
* still ensure that the spa_t doesn't get freed before this thread is
* finished with it when the pool is exported. We want to call spa_close
* before we start waiting because otherwise the additional ref would
* prevent the pool from being exported or destroyed throughout the
* potentially long wait.
*/
mutex_enter(&spa->spa_activities_lock);
spa->spa_waiters++;
spa_close(spa, FTAG);
*waited = B_FALSE;
for (;;) {
boolean_t in_progress;
error = spa_activity_in_progress(spa, activity, use_tag, tag,
&in_progress);
if (error || !in_progress || spa->spa_waiters_cancel)
break;
*waited = B_TRUE;
if (cv_wait_sig(&spa->spa_activities_cv,
&spa->spa_activities_lock) == 0) {
error = EINTR;
break;
}
}
spa->spa_waiters--;
cv_signal(&spa->spa_waiters_cv);
mutex_exit(&spa->spa_activities_lock);
return (error);
}
/*
* Wait for a particular instance of the specified activity to complete, where
* the instance is identified by 'tag'
*/
int
spa_wait_tag(const char *pool, zpool_wait_activity_t activity, uint64_t tag,
boolean_t *waited)
{
return (spa_wait_common(pool, activity, B_TRUE, tag, waited));
}
/*
* Wait for all instances of the specified activity complete
*/
int
spa_wait(const char *pool, zpool_wait_activity_t activity, boolean_t *waited)
{
return (spa_wait_common(pool, activity, B_FALSE, 0, waited));
}
sysevent_t *
spa_event_create(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name)
{
sysevent_t *ev = NULL;
#ifdef _KERNEL
nvlist_t *resource;
resource = zfs_event_create(spa, vd, FM_SYSEVENT_CLASS, name, hist_nvl);
if (resource) {
ev = kmem_alloc(sizeof (sysevent_t), KM_SLEEP);
ev->resource = resource;
}
#else
(void) spa, (void) vd, (void) hist_nvl, (void) name;
#endif
return (ev);
}
void
spa_event_post(sysevent_t *ev)
{
#ifdef _KERNEL
if (ev) {
zfs_zevent_post(ev->resource, NULL, zfs_zevent_post_cb);
kmem_free(ev, sizeof (*ev));
}
#else
(void) ev;
#endif
}
/*
* Post a zevent corresponding to the given sysevent. The 'name' must be one
* of the event definitions in sys/sysevent/eventdefs.h. The payload will be
* filled in from the spa and (optionally) the vdev. This doesn't do anything
* in the userland libzpool, as we don't want consumers to misinterpret ztest
* or zdb as real changes.
*/
void
spa_event_notify(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name)
{
spa_event_post(spa_event_create(spa, vd, hist_nvl, name));
}
/* state manipulation functions */
EXPORT_SYMBOL(spa_open);
EXPORT_SYMBOL(spa_open_rewind);
EXPORT_SYMBOL(spa_get_stats);
EXPORT_SYMBOL(spa_create);
EXPORT_SYMBOL(spa_import);
EXPORT_SYMBOL(spa_tryimport);
EXPORT_SYMBOL(spa_destroy);
EXPORT_SYMBOL(spa_export);
EXPORT_SYMBOL(spa_reset);
EXPORT_SYMBOL(spa_async_request);
EXPORT_SYMBOL(spa_async_suspend);
EXPORT_SYMBOL(spa_async_resume);
EXPORT_SYMBOL(spa_inject_addref);
EXPORT_SYMBOL(spa_inject_delref);
EXPORT_SYMBOL(spa_scan_stat_init);
EXPORT_SYMBOL(spa_scan_get_stats);
/* device manipulation */
EXPORT_SYMBOL(spa_vdev_add);
EXPORT_SYMBOL(spa_vdev_attach);
EXPORT_SYMBOL(spa_vdev_detach);
EXPORT_SYMBOL(spa_vdev_setpath);
EXPORT_SYMBOL(spa_vdev_setfru);
EXPORT_SYMBOL(spa_vdev_split_mirror);
/* spare statech is global across all pools) */
EXPORT_SYMBOL(spa_spare_add);
EXPORT_SYMBOL(spa_spare_remove);
EXPORT_SYMBOL(spa_spare_exists);
EXPORT_SYMBOL(spa_spare_activate);
/* L2ARC statech is global across all pools) */
EXPORT_SYMBOL(spa_l2cache_add);
EXPORT_SYMBOL(spa_l2cache_remove);
EXPORT_SYMBOL(spa_l2cache_exists);
EXPORT_SYMBOL(spa_l2cache_activate);
EXPORT_SYMBOL(spa_l2cache_drop);
/* scanning */
EXPORT_SYMBOL(spa_scan);
EXPORT_SYMBOL(spa_scan_stop);
/* spa syncing */
EXPORT_SYMBOL(spa_sync); /* only for DMU use */
EXPORT_SYMBOL(spa_sync_allpools);
/* properties */
EXPORT_SYMBOL(spa_prop_set);
EXPORT_SYMBOL(spa_prop_get);
EXPORT_SYMBOL(spa_prop_clear_bootfs);
/* asynchronous event notification */
EXPORT_SYMBOL(spa_event_notify);
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_shift, UINT, ZMOD_RW,
"log2 fraction of arc that can be used by inflight I/Os when "
"verifying pool during import");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_metadata, INT, ZMOD_RW,
"Set to traverse metadata on pool import");
ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_data, INT, ZMOD_RW,
"Set to traverse data on pool import");
ZFS_MODULE_PARAM(zfs_spa, spa_, load_print_vdev_tree, INT, ZMOD_RW,
"Print vdev tree to zfs_dbgmsg during pool import");
ZFS_MODULE_PARAM(zfs_zio, zio_, taskq_batch_pct, UINT, ZMOD_RD,
"Percentage of CPUs to run an IO worker thread");
ZFS_MODULE_PARAM(zfs_zio, zio_, taskq_batch_tpq, UINT, ZMOD_RD,
"Number of threads per IO worker taskqueue");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs, zfs_, max_missing_tvds, U64, ZMOD_RW,
"Allow importing pool with up to this number of missing top-level "
"vdevs (in read-only mode)");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, zthr_pause, INT,
ZMOD_RW, "Set the livelist condense zthr to pause");
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, sync_pause, INT,
ZMOD_RW, "Set the livelist condense synctask to pause");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, sync_cancel,
INT, ZMOD_RW,
"Whether livelist condensing was canceled in the synctask");
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, zthr_cancel,
INT, ZMOD_RW,
"Whether livelist condensing was canceled in the zthr function");
ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, new_alloc, INT,
ZMOD_RW,
"Whether extra ALLOC blkptrs were added to a livelist entry while it "
"was being condensed");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/spa_errlog.c b/sys/contrib/openzfs/module/zfs/spa_errlog.c
index 41cb9d01273c..3bc8619b51a8 100644
--- a/sys/contrib/openzfs/module/zfs/spa_errlog.c
+++ b/sys/contrib/openzfs/module/zfs/spa_errlog.c
@@ -1,1342 +1,1342 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2006, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013, 2014, Delphix. All rights reserved.
* Copyright (c) 2019 Datto Inc.
* Copyright (c) 2021, 2022, George Amanakis. All rights reserved.
*/
/*
* Routines to manage the on-disk persistent error log.
*
* Each pool stores a log of all logical data errors seen during normal
* operation. This is actually the union of two distinct logs: the last log,
* and the current log. All errors seen are logged to the current log. When a
* scrub completes, the current log becomes the last log, the last log is thrown
* out, and the current log is reinitialized. This way, if an error is somehow
* corrected, a new scrub will show that it no longer exists, and will be
* deleted from the log when the scrub completes.
*
* The log is stored using a ZAP object whose key is a string form of the
* zbookmark_phys tuple (objset, object, level, blkid), and whose contents is an
* optional 'objset:object' human-readable string describing the data. When an
* error is first logged, this string will be empty, indicating that no name is
* known. This prevents us from having to issue a potentially large amount of
* I/O to discover the object name during an error path. Instead, we do the
* calculation when the data is requested, storing the result so future queries
* will be faster.
*
* If the head_errlog feature is enabled, a different on-disk format is used.
* The error log of each head dataset is stored separately in the zap object
* and keyed by the head id. This enables listing every dataset affected in
* userland. In order to be able to track whether an error block has been
* modified or added to snapshots since it was marked as an error, a new tuple
* is introduced: zbookmark_err_phys_t. It allows the storage of the birth
* transaction group of an error block on-disk. The birth transaction group is
* used by check_filesystem() to assess whether this block was freed,
* re-written or added to a snapshot since its marking as an error.
*
* This log is then shipped into an nvlist where the key is the dataset name and
* the value is the object name. Userland is then responsible for uniquifying
* this list and displaying it to the user.
*/
#include <sys/dmu_tx.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/zap.h>
#include <sys/zio.h>
#include <sys/dsl_dir.h>
#include <sys/dmu_objset.h>
#include <sys/dbuf.h>
#include <sys/zfs_znode.h>
#define NAME_MAX_LEN 64
/*
* spa_upgrade_errlog_limit : A zfs module parameter that controls the number
* of on-disk error log entries that will be converted to the new
* format when enabling head_errlog. Defaults to 0 which converts
* all log entries.
*/
static uint_t spa_upgrade_errlog_limit = 0;
/*
* Convert a bookmark to a string.
*/
static void
bookmark_to_name(zbookmark_phys_t *zb, char *buf, size_t len)
{
(void) snprintf(buf, len, "%llx:%llx:%llx:%llx",
(u_longlong_t)zb->zb_objset, (u_longlong_t)zb->zb_object,
(u_longlong_t)zb->zb_level, (u_longlong_t)zb->zb_blkid);
}
/*
* Convert an err_phys to a string.
*/
static void
errphys_to_name(zbookmark_err_phys_t *zep, char *buf, size_t len)
{
(void) snprintf(buf, len, "%llx:%llx:%llx:%llx",
(u_longlong_t)zep->zb_object, (u_longlong_t)zep->zb_level,
(u_longlong_t)zep->zb_blkid, (u_longlong_t)zep->zb_birth);
}
/*
* Convert a string to a err_phys.
*/
static void
name_to_errphys(char *buf, zbookmark_err_phys_t *zep)
{
zep->zb_object = zfs_strtonum(buf, &buf);
ASSERT(*buf == ':');
zep->zb_level = (int)zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == ':');
zep->zb_blkid = zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == ':');
zep->zb_birth = zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == '\0');
}
/*
* Convert a string to a bookmark.
*/
static void
name_to_bookmark(char *buf, zbookmark_phys_t *zb)
{
zb->zb_objset = zfs_strtonum(buf, &buf);
ASSERT(*buf == ':');
zb->zb_object = zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == ':');
zb->zb_level = (int)zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == ':');
zb->zb_blkid = zfs_strtonum(buf + 1, &buf);
ASSERT(*buf == '\0');
}
#ifdef _KERNEL
static int check_clones(spa_t *spa, uint64_t zap_clone, uint64_t snap_count,
uint64_t *snap_obj_array, zbookmark_err_phys_t *zep, void* uaddr,
uint64_t *count);
static void
zep_to_zb(uint64_t dataset, zbookmark_err_phys_t *zep, zbookmark_phys_t *zb)
{
zb->zb_objset = dataset;
zb->zb_object = zep->zb_object;
zb->zb_level = zep->zb_level;
zb->zb_blkid = zep->zb_blkid;
}
#endif
static void
name_to_object(char *buf, uint64_t *obj)
{
*obj = zfs_strtonum(buf, &buf);
ASSERT(*buf == '\0');
}
/*
* Retrieve the head filesystem.
*/
static int get_head_ds(spa_t *spa, uint64_t dsobj, uint64_t *head_ds)
{
dsl_dataset_t *ds;
int error = dsl_dataset_hold_obj(spa->spa_dsl_pool,
dsobj, FTAG, &ds);
if (error != 0)
return (error);
ASSERT(head_ds);
*head_ds = dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj;
dsl_dataset_rele(ds, FTAG);
return (error);
}
/*
* Log an uncorrectable error to the persistent error log. We add it to the
* spa's list of pending errors. The changes are actually synced out to disk
* during spa_errlog_sync().
*/
void
spa_log_error(spa_t *spa, const zbookmark_phys_t *zb, const uint64_t *birth)
{
spa_error_entry_t search;
spa_error_entry_t *new;
avl_tree_t *tree;
avl_index_t where;
/*
* If we are trying to import a pool, ignore any errors, as we won't be
* writing to the pool any time soon.
*/
if (spa_load_state(spa) == SPA_LOAD_TRYIMPORT)
return;
mutex_enter(&spa->spa_errlist_lock);
/*
* If we have had a request to rotate the log, log it to the next list
* instead of the current one.
*/
if (spa->spa_scrub_active || spa->spa_scrub_finished)
tree = &spa->spa_errlist_scrub;
else
tree = &spa->spa_errlist_last;
search.se_bookmark = *zb;
if (avl_find(tree, &search, &where) != NULL) {
mutex_exit(&spa->spa_errlist_lock);
return;
}
new = kmem_zalloc(sizeof (spa_error_entry_t), KM_SLEEP);
new->se_bookmark = *zb;
/*
* If the head_errlog feature is enabled, store the birth txg now. In
* case the file is deleted before spa_errlog_sync() runs, we will not
* be able to retrieve the birth txg.
*/
if (spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
new->se_zep.zb_object = zb->zb_object;
new->se_zep.zb_level = zb->zb_level;
new->se_zep.zb_blkid = zb->zb_blkid;
/*
* birth may end up being NULL, e.g. in zio_done(). We
* will handle this in process_error_block().
*/
if (birth != NULL)
new->se_zep.zb_birth = *birth;
}
avl_insert(tree, new, where);
mutex_exit(&spa->spa_errlist_lock);
}
#ifdef _KERNEL
static int
find_birth_txg(dsl_dataset_t *ds, zbookmark_err_phys_t *zep,
uint64_t *birth_txg)
{
objset_t *os;
int error = dmu_objset_from_ds(ds, &os);
if (error != 0)
return (error);
dnode_t *dn;
blkptr_t bp;
error = dnode_hold(os, zep->zb_object, FTAG, &dn);
if (error != 0)
return (error);
rw_enter(&dn->dn_struct_rwlock, RW_READER);
error = dbuf_dnode_findbp(dn, zep->zb_level, zep->zb_blkid, &bp, NULL,
NULL);
if (error == 0 && BP_IS_HOLE(&bp))
error = SET_ERROR(ENOENT);
*birth_txg = bp.blk_birth;
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
return (error);
}
/*
* Copy the bookmark to the end of the user-space buffer which starts at
* uaddr and has *count unused entries, and decrement *count by 1.
*/
static int
copyout_entry(const zbookmark_phys_t *zb, void *uaddr, uint64_t *count)
{
if (*count == 0)
return (SET_ERROR(ENOMEM));
*count -= 1;
if (copyout(zb, (char *)uaddr + (*count) * sizeof (zbookmark_phys_t),
sizeof (zbookmark_phys_t)) != 0)
return (SET_ERROR(EFAULT));
return (0);
}
/*
* Each time the error block is referenced by a snapshot or clone, add a
* zbookmark_phys_t entry to the userspace array at uaddr. The array is
* filled from the back and the in-out parameter *count is modified to be the
* number of unused entries at the beginning of the array.
*/
static int
check_filesystem(spa_t *spa, uint64_t head_ds, zbookmark_err_phys_t *zep,
void *uaddr, uint64_t *count)
{
dsl_dataset_t *ds;
dsl_pool_t *dp = spa->spa_dsl_pool;
int error = dsl_dataset_hold_obj(dp, head_ds, FTAG, &ds);
if (error != 0)
return (error);
uint64_t latest_txg;
uint64_t txg_to_consider = spa->spa_syncing_txg;
boolean_t check_snapshot = B_TRUE;
error = find_birth_txg(ds, zep, &latest_txg);
/*
* If the filesystem is encrypted and the key is not loaded
* or the encrypted filesystem is not mounted the error will be EACCES.
* In that case report an error in the head filesystem and return.
*/
if (error == EACCES) {
dsl_dataset_rele(ds, FTAG);
zbookmark_phys_t zb;
zep_to_zb(head_ds, zep, &zb);
error = copyout_entry(&zb, uaddr, count);
if (error != 0) {
dsl_dataset_rele(ds, FTAG);
return (error);
}
return (0);
}
/*
* If find_birth_txg() errors out otherwise, let txg_to_consider be
* equal to the spa's syncing txg: if check_filesystem() errors out
* then affected snapshots or clones will not be checked.
*/
if (error == 0 && zep->zb_birth == latest_txg) {
/* Block neither free nor rewritten. */
zbookmark_phys_t zb;
zep_to_zb(head_ds, zep, &zb);
error = copyout_entry(&zb, uaddr, count);
if (error != 0) {
dsl_dataset_rele(ds, FTAG);
return (error);
}
check_snapshot = B_FALSE;
} else if (error == 0) {
txg_to_consider = latest_txg;
}
/*
* Retrieve the number of snapshots if the dataset is not a snapshot.
*/
uint64_t snap_count = 0;
if (dsl_dataset_phys(ds)->ds_snapnames_zapobj != 0) {
error = zap_count(spa->spa_meta_objset,
dsl_dataset_phys(ds)->ds_snapnames_zapobj, &snap_count);
if (error != 0) {
dsl_dataset_rele(ds, FTAG);
return (error);
}
+ }
- if (snap_count == 0) {
- /* Filesystem without snapshots. */
- dsl_dataset_rele(ds, FTAG);
- return (0);
- }
+ if (snap_count == 0) {
+ /* Filesystem without snapshots. */
+ dsl_dataset_rele(ds, FTAG);
+ return (0);
}
uint64_t *snap_obj_array = kmem_zalloc(snap_count * sizeof (uint64_t),
KM_SLEEP);
int aff_snap_count = 0;
uint64_t snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj;
uint64_t snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg;
uint64_t zap_clone = dsl_dir_phys(ds->ds_dir)->dd_clones;
dsl_dataset_rele(ds, FTAG);
/* Check only snapshots created from this file system. */
while (snap_obj != 0 && zep->zb_birth < snap_obj_txg &&
snap_obj_txg <= txg_to_consider) {
error = dsl_dataset_hold_obj(dp, snap_obj, FTAG, &ds);
if (error != 0)
goto out;
if (dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj != head_ds) {
snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj;
snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg;
dsl_dataset_rele(ds, FTAG);
continue;
}
boolean_t affected = B_TRUE;
if (check_snapshot) {
uint64_t blk_txg;
error = find_birth_txg(ds, zep, &blk_txg);
affected = (error == 0 && zep->zb_birth == blk_txg);
}
/* Report errors in snapshots. */
if (affected) {
snap_obj_array[aff_snap_count] = snap_obj;
aff_snap_count++;
zbookmark_phys_t zb;
zep_to_zb(snap_obj, zep, &zb);
error = copyout_entry(&zb, uaddr, count);
if (error != 0) {
dsl_dataset_rele(ds, FTAG);
goto out;
}
}
snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj;
snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg;
dsl_dataset_rele(ds, FTAG);
}
if (zap_clone != 0 && aff_snap_count > 0) {
error = check_clones(spa, zap_clone, snap_count, snap_obj_array,
zep, uaddr, count);
}
out:
kmem_free(snap_obj_array, sizeof (*snap_obj_array));
return (error);
}
/*
* Clone checking.
*/
static int check_clones(spa_t *spa, uint64_t zap_clone, uint64_t snap_count,
uint64_t *snap_obj_array, zbookmark_err_phys_t *zep, void* uaddr,
uint64_t *count)
{
int error = 0;
zap_cursor_t *zc;
zap_attribute_t *za;
zc = kmem_zalloc(sizeof (zap_cursor_t), KM_SLEEP);
za = kmem_zalloc(sizeof (zap_attribute_t), KM_SLEEP);
for (zap_cursor_init(zc, spa->spa_meta_objset, zap_clone);
zap_cursor_retrieve(zc, za) == 0;
zap_cursor_advance(zc)) {
dsl_pool_t *dp = spa->spa_dsl_pool;
dsl_dataset_t *clone;
error = dsl_dataset_hold_obj(dp, za->za_first_integer,
FTAG, &clone);
if (error != 0)
break;
/*
* Only clones whose origins were affected could also
* have affected snapshots.
*/
boolean_t found = B_FALSE;
for (int i = 0; i < snap_count; i++) {
if (dsl_dir_phys(clone->ds_dir)->dd_origin_obj
== snap_obj_array[i])
found = B_TRUE;
}
dsl_dataset_rele(clone, FTAG);
if (!found)
continue;
error = check_filesystem(spa, za->za_first_integer, zep,
uaddr, count);
if (error != 0)
break;
}
+ zap_cursor_fini(zc);
kmem_free(za, sizeof (*za));
kmem_free(zc, sizeof (*zc));
- zap_cursor_fini(zc);
return (error);
}
static int
find_top_affected_fs(spa_t *spa, uint64_t head_ds, zbookmark_err_phys_t *zep,
uint64_t *top_affected_fs)
{
uint64_t oldest_dsobj;
int error = dsl_dataset_oldest_snapshot(spa, head_ds, zep->zb_birth,
&oldest_dsobj);
if (error != 0)
return (error);
dsl_dataset_t *ds;
error = dsl_dataset_hold_obj(spa->spa_dsl_pool, oldest_dsobj,
FTAG, &ds);
if (error != 0)
return (error);
*top_affected_fs =
dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj;
dsl_dataset_rele(ds, FTAG);
return (0);
}
static int
process_error_block(spa_t *spa, uint64_t head_ds, zbookmark_err_phys_t *zep,
void *uaddr, uint64_t *count)
{
/*
* If zb_birth == 0 or head_ds == 0 it means we failed to retrieve the
* birth txg or the head filesystem of the block pointer. This may
* happen e.g. when an encrypted filesystem is not mounted or when
* the key is not loaded. In this case do not proceed to
* check_filesystem(), instead do the accounting here.
*/
if (zep->zb_birth == 0 || head_ds == 0) {
zbookmark_phys_t zb;
zep_to_zb(head_ds, zep, &zb);
int error = copyout_entry(&zb, uaddr, count);
if (error != 0) {
return (error);
}
return (0);
}
uint64_t top_affected_fs;
int error = find_top_affected_fs(spa, head_ds, zep, &top_affected_fs);
if (error == 0) {
error = check_filesystem(spa, top_affected_fs, zep,
uaddr, count);
}
return (error);
}
#endif
/*
* If a healed bookmark matches an entry in the error log we stash it in a tree
* so that we can later remove the related log entries in sync context.
*/
static void
spa_add_healed_error(spa_t *spa, uint64_t obj, zbookmark_phys_t *healed_zb)
{
char name[NAME_MAX_LEN];
if (obj == 0)
return;
bookmark_to_name(healed_zb, name, sizeof (name));
mutex_enter(&spa->spa_errlog_lock);
if (zap_contains(spa->spa_meta_objset, obj, name) == 0) {
/*
* Found an error matching healed zb, add zb to our
* tree of healed errors
*/
avl_tree_t *tree = &spa->spa_errlist_healed;
spa_error_entry_t search;
spa_error_entry_t *new;
avl_index_t where;
search.se_bookmark = *healed_zb;
mutex_enter(&spa->spa_errlist_lock);
if (avl_find(tree, &search, &where) != NULL) {
mutex_exit(&spa->spa_errlist_lock);
mutex_exit(&spa->spa_errlog_lock);
return;
}
new = kmem_zalloc(sizeof (spa_error_entry_t), KM_SLEEP);
new->se_bookmark = *healed_zb;
avl_insert(tree, new, where);
mutex_exit(&spa->spa_errlist_lock);
}
mutex_exit(&spa->spa_errlog_lock);
}
/*
* If this error exists in the given tree remove it.
*/
static void
remove_error_from_list(spa_t *spa, avl_tree_t *t, const zbookmark_phys_t *zb)
{
spa_error_entry_t search, *found;
avl_index_t where;
mutex_enter(&spa->spa_errlist_lock);
search.se_bookmark = *zb;
if ((found = avl_find(t, &search, &where)) != NULL) {
avl_remove(t, found);
kmem_free(found, sizeof (spa_error_entry_t));
}
mutex_exit(&spa->spa_errlist_lock);
}
/*
* Removes all of the recv healed errors from both on-disk error logs
*/
static void
spa_remove_healed_errors(spa_t *spa, avl_tree_t *s, avl_tree_t *l, dmu_tx_t *tx)
{
char name[NAME_MAX_LEN];
spa_error_entry_t *se;
void *cookie = NULL;
ASSERT(MUTEX_HELD(&spa->spa_errlog_lock));
while ((se = avl_destroy_nodes(&spa->spa_errlist_healed,
&cookie)) != NULL) {
remove_error_from_list(spa, s, &se->se_bookmark);
remove_error_from_list(spa, l, &se->se_bookmark);
bookmark_to_name(&se->se_bookmark, name, sizeof (name));
kmem_free(se, sizeof (spa_error_entry_t));
(void) zap_remove(spa->spa_meta_objset,
spa->spa_errlog_last, name, tx);
(void) zap_remove(spa->spa_meta_objset,
spa->spa_errlog_scrub, name, tx);
}
}
/*
* Stash away healed bookmarks to remove them from the on-disk error logs
* later in spa_remove_healed_errors().
*/
void
spa_remove_error(spa_t *spa, zbookmark_phys_t *zb)
{
char name[NAME_MAX_LEN];
bookmark_to_name(zb, name, sizeof (name));
spa_add_healed_error(spa, spa->spa_errlog_last, zb);
spa_add_healed_error(spa, spa->spa_errlog_scrub, zb);
}
static uint64_t
approx_errlog_size_impl(spa_t *spa, uint64_t spa_err_obj)
{
if (spa_err_obj == 0)
return (0);
uint64_t total = 0;
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, spa->spa_meta_objset, spa_err_obj);
zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) {
uint64_t count;
if (zap_count(spa->spa_meta_objset, za.za_first_integer,
&count) == 0)
total += count;
}
zap_cursor_fini(&zc);
return (total);
}
/*
* Return the approximate number of errors currently in the error log. This
* will be nonzero if there are some errors, but otherwise it may be more
* or less than the number of entries returned by spa_get_errlog().
*/
uint64_t
spa_approx_errlog_size(spa_t *spa)
{
uint64_t total = 0;
if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
mutex_enter(&spa->spa_errlog_lock);
uint64_t count;
if (spa->spa_errlog_scrub != 0 &&
zap_count(spa->spa_meta_objset, spa->spa_errlog_scrub,
&count) == 0)
total += count;
if (spa->spa_errlog_last != 0 && !spa->spa_scrub_finished &&
zap_count(spa->spa_meta_objset, spa->spa_errlog_last,
&count) == 0)
total += count;
mutex_exit(&spa->spa_errlog_lock);
} else {
mutex_enter(&spa->spa_errlog_lock);
total += approx_errlog_size_impl(spa, spa->spa_errlog_last);
total += approx_errlog_size_impl(spa, spa->spa_errlog_scrub);
mutex_exit(&spa->spa_errlog_lock);
}
mutex_enter(&spa->spa_errlist_lock);
total += avl_numnodes(&spa->spa_errlist_last);
total += avl_numnodes(&spa->spa_errlist_scrub);
mutex_exit(&spa->spa_errlist_lock);
return (total);
}
/*
* This function sweeps through an on-disk error log and stores all bookmarks
* as error bookmarks in a new ZAP object. At the end we discard the old one,
* and spa_update_errlog() will set the spa's on-disk error log to new ZAP
* object.
*/
static void
sync_upgrade_errlog(spa_t *spa, uint64_t spa_err_obj, uint64_t *newobj,
dmu_tx_t *tx)
{
zap_cursor_t zc;
zap_attribute_t za;
zbookmark_phys_t zb;
uint64_t count;
*newobj = zap_create(spa->spa_meta_objset, DMU_OT_ERROR_LOG,
DMU_OT_NONE, 0, tx);
/*
* If we cannnot perform the upgrade we should clear the old on-disk
* error logs.
*/
if (zap_count(spa->spa_meta_objset, spa_err_obj, &count) != 0) {
VERIFY0(dmu_object_free(spa->spa_meta_objset, spa_err_obj, tx));
return;
}
for (zap_cursor_init(&zc, spa->spa_meta_objset, spa_err_obj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
if (spa_upgrade_errlog_limit != 0 &&
zc.zc_cd == spa_upgrade_errlog_limit)
break;
name_to_bookmark(za.za_name, &zb);
zbookmark_err_phys_t zep;
zep.zb_object = zb.zb_object;
zep.zb_level = zb.zb_level;
zep.zb_blkid = zb.zb_blkid;
zep.zb_birth = 0;
/*
* In case of an error we should simply continue instead of
* returning prematurely. See the next comment.
*/
uint64_t head_ds;
dsl_pool_t *dp = spa->spa_dsl_pool;
dsl_dataset_t *ds;
objset_t *os;
int error = dsl_dataset_hold_obj(dp, zb.zb_objset, FTAG, &ds);
if (error != 0)
continue;
head_ds = dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj;
/*
* The objset and the dnode are required for getting the block
* pointer, which is used to determine if BP_IS_HOLE(). If
* getting the objset or the dnode fails, do not create a
* zap entry (presuming we know the dataset) as this may create
* spurious errors that we cannot ever resolve. If an error is
* truly persistent, it should re-appear after a scan.
*/
if (dmu_objset_from_ds(ds, &os) != 0) {
dsl_dataset_rele(ds, FTAG);
continue;
}
dnode_t *dn;
blkptr_t bp;
if (dnode_hold(os, zep.zb_object, FTAG, &dn) != 0) {
dsl_dataset_rele(ds, FTAG);
continue;
}
rw_enter(&dn->dn_struct_rwlock, RW_READER);
error = dbuf_dnode_findbp(dn, zep.zb_level, zep.zb_blkid, &bp,
NULL, NULL);
if (error == EACCES)
error = 0;
else if (!error)
zep.zb_birth = bp.blk_birth;
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
dsl_dataset_rele(ds, FTAG);
if (error != 0 || BP_IS_HOLE(&bp))
continue;
uint64_t err_obj;
error = zap_lookup_int_key(spa->spa_meta_objset, *newobj,
head_ds, &err_obj);
if (error == ENOENT) {
err_obj = zap_create(spa->spa_meta_objset,
DMU_OT_ERROR_LOG, DMU_OT_NONE, 0, tx);
(void) zap_update_int_key(spa->spa_meta_objset,
*newobj, head_ds, err_obj, tx);
}
char buf[64];
errphys_to_name(&zep, buf, sizeof (buf));
const char *name = "";
(void) zap_update(spa->spa_meta_objset, err_obj,
buf, 1, strlen(name) + 1, name, tx);
}
zap_cursor_fini(&zc);
VERIFY0(dmu_object_free(spa->spa_meta_objset, spa_err_obj, tx));
}
void
spa_upgrade_errlog(spa_t *spa, dmu_tx_t *tx)
{
uint64_t newobj = 0;
mutex_enter(&spa->spa_errlog_lock);
if (spa->spa_errlog_last != 0) {
sync_upgrade_errlog(spa, spa->spa_errlog_last, &newobj, tx);
spa->spa_errlog_last = newobj;
}
if (spa->spa_errlog_scrub != 0) {
sync_upgrade_errlog(spa, spa->spa_errlog_scrub, &newobj, tx);
spa->spa_errlog_scrub = newobj;
}
mutex_exit(&spa->spa_errlog_lock);
}
#ifdef _KERNEL
/*
* If an error block is shared by two datasets it will be counted twice.
*/
static int
process_error_log(spa_t *spa, uint64_t obj, void *uaddr, uint64_t *count)
{
zap_cursor_t zc;
zap_attribute_t za;
if (obj == 0)
return (0);
if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
for (zap_cursor_init(&zc, spa->spa_meta_objset, obj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
if (*count == 0) {
zap_cursor_fini(&zc);
return (SET_ERROR(ENOMEM));
}
zbookmark_phys_t zb;
name_to_bookmark(za.za_name, &zb);
int error = copyout_entry(&zb, uaddr, count);
if (error != 0) {
zap_cursor_fini(&zc);
return (error);
}
}
zap_cursor_fini(&zc);
return (0);
}
for (zap_cursor_init(&zc, spa->spa_meta_objset, obj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
zap_cursor_t head_ds_cursor;
zap_attribute_t head_ds_attr;
uint64_t head_ds_err_obj = za.za_first_integer;
uint64_t head_ds;
name_to_object(za.za_name, &head_ds);
for (zap_cursor_init(&head_ds_cursor, spa->spa_meta_objset,
head_ds_err_obj); zap_cursor_retrieve(&head_ds_cursor,
&head_ds_attr) == 0; zap_cursor_advance(&head_ds_cursor)) {
zbookmark_err_phys_t head_ds_block;
name_to_errphys(head_ds_attr.za_name, &head_ds_block);
int error = process_error_block(spa, head_ds,
&head_ds_block, uaddr, count);
if (error != 0) {
zap_cursor_fini(&head_ds_cursor);
zap_cursor_fini(&zc);
return (error);
}
}
zap_cursor_fini(&head_ds_cursor);
}
zap_cursor_fini(&zc);
return (0);
}
static int
process_error_list(spa_t *spa, avl_tree_t *list, void *uaddr, uint64_t *count)
{
spa_error_entry_t *se;
if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
for (se = avl_first(list); se != NULL;
se = AVL_NEXT(list, se)) {
int error =
copyout_entry(&se->se_bookmark, uaddr, count);
if (error != 0) {
return (error);
}
}
return (0);
}
for (se = avl_first(list); se != NULL; se = AVL_NEXT(list, se)) {
uint64_t head_ds = 0;
int error = get_head_ds(spa, se->se_bookmark.zb_objset,
&head_ds);
/*
* If get_head_ds() errors out, set the head filesystem
* to the filesystem stored in the bookmark of the
* error block.
*/
if (error != 0)
head_ds = se->se_bookmark.zb_objset;
error = process_error_block(spa, head_ds,
&se->se_zep, uaddr, count);
if (error != 0)
return (error);
}
return (0);
}
#endif
/*
* Copy all known errors to userland as an array of bookmarks. This is
* actually a union of the on-disk last log and current log, as well as any
* pending error requests.
*
* Because the act of reading the on-disk log could cause errors to be
* generated, we have two separate locks: one for the error log and one for the
* in-core error lists. We only need the error list lock to log and error, so
* we grab the error log lock while we read the on-disk logs, and only pick up
* the error list lock when we are finished.
*/
int
spa_get_errlog(spa_t *spa, void *uaddr, uint64_t *count)
{
int ret = 0;
#ifdef _KERNEL
/*
* The pool config lock is needed to hold a dataset_t via (among other
* places) process_error_list() -> process_error_block()->
* find_top_affected_fs(), and lock ordering requires that we get it
* before the spa_errlog_lock.
*/
dsl_pool_config_enter(spa->spa_dsl_pool, FTAG);
mutex_enter(&spa->spa_errlog_lock);
ret = process_error_log(spa, spa->spa_errlog_scrub, uaddr, count);
if (!ret && !spa->spa_scrub_finished)
ret = process_error_log(spa, spa->spa_errlog_last, uaddr,
count);
mutex_enter(&spa->spa_errlist_lock);
if (!ret)
ret = process_error_list(spa, &spa->spa_errlist_scrub, uaddr,
count);
if (!ret)
ret = process_error_list(spa, &spa->spa_errlist_last, uaddr,
count);
mutex_exit(&spa->spa_errlist_lock);
mutex_exit(&spa->spa_errlog_lock);
dsl_pool_config_exit(spa->spa_dsl_pool, FTAG);
#else
(void) spa, (void) uaddr, (void) count;
#endif
return (ret);
}
/*
* Called when a scrub completes. This simply set a bit which tells which AVL
* tree to add new errors. spa_errlog_sync() is responsible for actually
* syncing the changes to the underlying objects.
*/
void
spa_errlog_rotate(spa_t *spa)
{
mutex_enter(&spa->spa_errlist_lock);
spa->spa_scrub_finished = B_TRUE;
mutex_exit(&spa->spa_errlist_lock);
}
/*
* Discard any pending errors from the spa_t. Called when unloading a faulted
* pool, as the errors encountered during the open cannot be synced to disk.
*/
void
spa_errlog_drain(spa_t *spa)
{
spa_error_entry_t *se;
void *cookie;
mutex_enter(&spa->spa_errlist_lock);
cookie = NULL;
while ((se = avl_destroy_nodes(&spa->spa_errlist_last,
&cookie)) != NULL)
kmem_free(se, sizeof (spa_error_entry_t));
cookie = NULL;
while ((se = avl_destroy_nodes(&spa->spa_errlist_scrub,
&cookie)) != NULL)
kmem_free(se, sizeof (spa_error_entry_t));
mutex_exit(&spa->spa_errlist_lock);
}
/*
* Process a list of errors into the current on-disk log.
*/
void
sync_error_list(spa_t *spa, avl_tree_t *t, uint64_t *obj, dmu_tx_t *tx)
{
spa_error_entry_t *se;
char buf[NAME_MAX_LEN];
void *cookie;
if (avl_numnodes(t) == 0)
return;
/* create log if necessary */
if (*obj == 0)
*obj = zap_create(spa->spa_meta_objset, DMU_OT_ERROR_LOG,
DMU_OT_NONE, 0, tx);
/* add errors to the current log */
if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
for (se = avl_first(t); se != NULL; se = AVL_NEXT(t, se)) {
bookmark_to_name(&se->se_bookmark, buf, sizeof (buf));
const char *name = se->se_name ? se->se_name : "";
(void) zap_update(spa->spa_meta_objset, *obj, buf, 1,
strlen(name) + 1, name, tx);
}
} else {
for (se = avl_first(t); se != NULL; se = AVL_NEXT(t, se)) {
zbookmark_err_phys_t zep;
zep.zb_object = se->se_zep.zb_object;
zep.zb_level = se->se_zep.zb_level;
zep.zb_blkid = se->se_zep.zb_blkid;
zep.zb_birth = se->se_zep.zb_birth;
uint64_t head_ds = 0;
int error = get_head_ds(spa, se->se_bookmark.zb_objset,
&head_ds);
/*
* If get_head_ds() errors out, set the head filesystem
* to the filesystem stored in the bookmark of the
* error block.
*/
if (error != 0)
head_ds = se->se_bookmark.zb_objset;
uint64_t err_obj;
error = zap_lookup_int_key(spa->spa_meta_objset,
*obj, head_ds, &err_obj);
if (error == ENOENT) {
err_obj = zap_create(spa->spa_meta_objset,
DMU_OT_ERROR_LOG, DMU_OT_NONE, 0, tx);
(void) zap_update_int_key(spa->spa_meta_objset,
*obj, head_ds, err_obj, tx);
}
errphys_to_name(&zep, buf, sizeof (buf));
const char *name = se->se_name ? se->se_name : "";
(void) zap_update(spa->spa_meta_objset,
err_obj, buf, 1, strlen(name) + 1, name, tx);
}
}
/* purge the error list */
cookie = NULL;
while ((se = avl_destroy_nodes(t, &cookie)) != NULL)
kmem_free(se, sizeof (spa_error_entry_t));
}
static void
delete_errlog(spa_t *spa, uint64_t spa_err_obj, dmu_tx_t *tx)
{
if (spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) {
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, spa->spa_meta_objset, spa_err_obj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
VERIFY0(dmu_object_free(spa->spa_meta_objset,
za.za_first_integer, tx));
}
zap_cursor_fini(&zc);
}
VERIFY0(dmu_object_free(spa->spa_meta_objset, spa_err_obj, tx));
}
/*
* Sync the error log out to disk. This is a little tricky because the act of
* writing the error log requires the spa_errlist_lock. So, we need to lock the
* error lists, take a copy of the lists, and then reinitialize them. Then, we
* drop the error list lock and take the error log lock, at which point we
* do the errlog processing. Then, if we encounter an I/O error during this
* process, we can successfully add the error to the list. Note that this will
* result in the perpetual recycling of errors, but it is an unlikely situation
* and not a performance critical operation.
*/
void
spa_errlog_sync(spa_t *spa, uint64_t txg)
{
dmu_tx_t *tx;
avl_tree_t scrub, last;
int scrub_finished;
mutex_enter(&spa->spa_errlist_lock);
/*
* Bail out early under normal circumstances.
*/
if (avl_numnodes(&spa->spa_errlist_scrub) == 0 &&
avl_numnodes(&spa->spa_errlist_last) == 0 &&
avl_numnodes(&spa->spa_errlist_healed) == 0 &&
!spa->spa_scrub_finished) {
mutex_exit(&spa->spa_errlist_lock);
return;
}
spa_get_errlists(spa, &last, &scrub);
scrub_finished = spa->spa_scrub_finished;
spa->spa_scrub_finished = B_FALSE;
mutex_exit(&spa->spa_errlist_lock);
/*
* The pool config lock is needed to hold a dataset_t via
* sync_error_list() -> get_head_ds(), and lock ordering
* requires that we get it before the spa_errlog_lock.
*/
dsl_pool_config_enter(spa->spa_dsl_pool, FTAG);
mutex_enter(&spa->spa_errlog_lock);
tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg);
/*
* Remove healed errors from errors.
*/
spa_remove_healed_errors(spa, &last, &scrub, tx);
/*
* Sync out the current list of errors.
*/
sync_error_list(spa, &last, &spa->spa_errlog_last, tx);
/*
* Rotate the log if necessary.
*/
if (scrub_finished) {
if (spa->spa_errlog_last != 0)
delete_errlog(spa, spa->spa_errlog_last, tx);
spa->spa_errlog_last = spa->spa_errlog_scrub;
spa->spa_errlog_scrub = 0;
sync_error_list(spa, &scrub, &spa->spa_errlog_last, tx);
}
/*
* Sync out any pending scrub errors.
*/
sync_error_list(spa, &scrub, &spa->spa_errlog_scrub, tx);
/*
* Update the MOS to reflect the new values.
*/
(void) zap_update(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ERRLOG_LAST, sizeof (uint64_t), 1,
&spa->spa_errlog_last, tx);
(void) zap_update(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ERRLOG_SCRUB, sizeof (uint64_t), 1,
&spa->spa_errlog_scrub, tx);
dmu_tx_commit(tx);
mutex_exit(&spa->spa_errlog_lock);
dsl_pool_config_exit(spa->spa_dsl_pool, FTAG);
}
static void
delete_dataset_errlog(spa_t *spa, uint64_t spa_err_obj, uint64_t ds,
dmu_tx_t *tx)
{
if (spa_err_obj == 0)
return;
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, spa->spa_meta_objset, spa_err_obj);
zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) {
uint64_t head_ds;
name_to_object(za.za_name, &head_ds);
if (head_ds == ds) {
(void) zap_remove(spa->spa_meta_objset, spa_err_obj,
za.za_name, tx);
VERIFY0(dmu_object_free(spa->spa_meta_objset,
za.za_first_integer, tx));
break;
}
}
zap_cursor_fini(&zc);
}
void
spa_delete_dataset_errlog(spa_t *spa, uint64_t ds, dmu_tx_t *tx)
{
mutex_enter(&spa->spa_errlog_lock);
delete_dataset_errlog(spa, spa->spa_errlog_scrub, ds, tx);
delete_dataset_errlog(spa, spa->spa_errlog_last, ds, tx);
mutex_exit(&spa->spa_errlog_lock);
}
static int
find_txg_ancestor_snapshot(spa_t *spa, uint64_t new_head, uint64_t old_head,
uint64_t *txg)
{
dsl_dataset_t *ds;
dsl_pool_t *dp = spa->spa_dsl_pool;
int error = dsl_dataset_hold_obj(dp, old_head, FTAG, &ds);
if (error != 0)
return (error);
uint64_t prev_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj;
uint64_t prev_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg;
while (prev_obj != 0) {
dsl_dataset_rele(ds, FTAG);
if ((error = dsl_dataset_hold_obj(dp, prev_obj,
FTAG, &ds)) == 0 &&
dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj == new_head)
break;
if (error != 0)
return (error);
prev_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg;
prev_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj;
}
dsl_dataset_rele(ds, FTAG);
ASSERT(prev_obj != 0);
*txg = prev_obj_txg;
return (0);
}
static void
swap_errlog(spa_t *spa, uint64_t spa_err_obj, uint64_t new_head, uint64_t
old_head, dmu_tx_t *tx)
{
if (spa_err_obj == 0)
return;
uint64_t old_head_errlog;
int error = zap_lookup_int_key(spa->spa_meta_objset, spa_err_obj,
old_head, &old_head_errlog);
/* If no error log, then there is nothing to do. */
if (error != 0)
return;
uint64_t txg;
error = find_txg_ancestor_snapshot(spa, new_head, old_head, &txg);
if (error != 0)
return;
/*
* Create an error log if the file system being promoted does not
* already have one.
*/
uint64_t new_head_errlog;
error = zap_lookup_int_key(spa->spa_meta_objset, spa_err_obj, new_head,
&new_head_errlog);
if (error != 0) {
new_head_errlog = zap_create(spa->spa_meta_objset,
DMU_OT_ERROR_LOG, DMU_OT_NONE, 0, tx);
(void) zap_update_int_key(spa->spa_meta_objset, spa_err_obj,
new_head, new_head_errlog, tx);
}
zap_cursor_t zc;
zap_attribute_t za;
zbookmark_err_phys_t err_block;
for (zap_cursor_init(&zc, spa->spa_meta_objset, old_head_errlog);
zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) {
const char *name = "";
name_to_errphys(za.za_name, &err_block);
if (err_block.zb_birth < txg) {
(void) zap_update(spa->spa_meta_objset, new_head_errlog,
za.za_name, 1, strlen(name) + 1, name, tx);
(void) zap_remove(spa->spa_meta_objset, old_head_errlog,
za.za_name, tx);
}
}
zap_cursor_fini(&zc);
}
void
spa_swap_errlog(spa_t *spa, uint64_t new_head_ds, uint64_t old_head_ds,
dmu_tx_t *tx)
{
mutex_enter(&spa->spa_errlog_lock);
swap_errlog(spa, spa->spa_errlog_scrub, new_head_ds, old_head_ds, tx);
swap_errlog(spa, spa->spa_errlog_last, new_head_ds, old_head_ds, tx);
mutex_exit(&spa->spa_errlog_lock);
}
#if defined(_KERNEL)
/* error handling */
EXPORT_SYMBOL(spa_log_error);
EXPORT_SYMBOL(spa_approx_errlog_size);
EXPORT_SYMBOL(spa_get_errlog);
EXPORT_SYMBOL(spa_errlog_rotate);
EXPORT_SYMBOL(spa_errlog_drain);
EXPORT_SYMBOL(spa_errlog_sync);
EXPORT_SYMBOL(spa_get_errlists);
EXPORT_SYMBOL(spa_delete_dataset_errlog);
EXPORT_SYMBOL(spa_swap_errlog);
EXPORT_SYMBOL(sync_error_list);
EXPORT_SYMBOL(spa_upgrade_errlog);
#endif
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs_spa, spa_, upgrade_errlog_limit, UINT, ZMOD_RW,
"Limit the number of errors which will be upgraded to the new "
"on-disk error log when enabling head_errlog");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/spa_misc.c b/sys/contrib/openzfs/module/zfs/spa_misc.c
index 1475a4a53f4a..54a0eeccf27b 100644
--- a/sys/contrib/openzfs/module/zfs/spa_misc.c
+++ b/sys/contrib/openzfs/module/zfs/spa_misc.c
@@ -1,2961 +1,2984 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2019 by Delphix. All rights reserved.
* Copyright 2015 Nexenta Systems, Inc. All rights reserved.
* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
* Copyright 2013 Saso Kiselkov. All rights reserved.
* Copyright (c) 2017 Datto Inc.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, loli10K <ezomori.nozomu@gmail.com>. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/zfs_chksum.h>
#include <sys/spa_impl.h>
#include <sys/zio.h>
#include <sys/zio_checksum.h>
#include <sys/zio_compress.h>
#include <sys/dmu.h>
#include <sys/dmu_tx.h>
#include <sys/zap.h>
#include <sys/zil.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_initialize.h>
#include <sys/vdev_trim.h>
#include <sys/vdev_file.h>
#include <sys/vdev_raidz.h>
#include <sys/metaslab.h>
#include <sys/uberblock_impl.h>
#include <sys/txg.h>
#include <sys/avl.h>
#include <sys/unique.h>
#include <sys/dsl_pool.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_prop.h>
#include <sys/fm/util.h>
#include <sys/dsl_scan.h>
#include <sys/fs/zfs.h>
#include <sys/metaslab_impl.h>
#include <sys/arc.h>
#include <sys/brt.h>
#include <sys/ddt.h>
#include <sys/kstat.h>
#include "zfs_prop.h"
#include <sys/btree.h>
#include <sys/zfeature.h>
#include <sys/qat.h>
#include <sys/zstd/zstd.h>
/*
* SPA locking
*
* There are three basic locks for managing spa_t structures:
*
* spa_namespace_lock (global mutex)
*
* This lock must be acquired to do any of the following:
*
* - Lookup a spa_t by name
* - Add or remove a spa_t from the namespace
* - Increase spa_refcount from non-zero
* - Check if spa_refcount is zero
* - Rename a spa_t
* - add/remove/attach/detach devices
* - Held for the duration of create/destroy/import/export
*
* It does not need to handle recursion. A create or destroy may
* reference objects (files or zvols) in other pools, but by
* definition they must have an existing reference, and will never need
* to lookup a spa_t by name.
*
* spa_refcount (per-spa zfs_refcount_t protected by mutex)
*
* This reference count keep track of any active users of the spa_t. The
* spa_t cannot be destroyed or freed while this is non-zero. Internally,
* the refcount is never really 'zero' - opening a pool implicitly keeps
* some references in the DMU. Internally we check against spa_minref, but
* present the image of a zero/non-zero value to consumers.
*
* spa_config_lock[] (per-spa array of rwlocks)
*
* This protects the spa_t from config changes, and must be held in
* the following circumstances:
*
* - RW_READER to perform I/O to the spa
* - RW_WRITER to change the vdev config
*
* The locking order is fairly straightforward:
*
* spa_namespace_lock -> spa_refcount
*
* The namespace lock must be acquired to increase the refcount from 0
* or to check if it is zero.
*
* spa_refcount -> spa_config_lock[]
*
* There must be at least one valid reference on the spa_t to acquire
* the config lock.
*
* spa_namespace_lock -> spa_config_lock[]
*
* The namespace lock must always be taken before the config lock.
*
*
* The spa_namespace_lock can be acquired directly and is globally visible.
*
* The namespace is manipulated using the following functions, all of which
* require the spa_namespace_lock to be held.
*
* spa_lookup() Lookup a spa_t by name.
*
* spa_add() Create a new spa_t in the namespace.
*
* spa_remove() Remove a spa_t from the namespace. This also
* frees up any memory associated with the spa_t.
*
* spa_next() Returns the next spa_t in the system, or the
* first if NULL is passed.
*
* spa_evict_all() Shutdown and remove all spa_t structures in
* the system.
*
* spa_guid_exists() Determine whether a pool/device guid exists.
*
* The spa_refcount is manipulated using the following functions:
*
* spa_open_ref() Adds a reference to the given spa_t. Must be
* called with spa_namespace_lock held if the
* refcount is currently zero.
*
* spa_close() Remove a reference from the spa_t. This will
* not free the spa_t or remove it from the
* namespace. No locking is required.
*
* spa_refcount_zero() Returns true if the refcount is currently
* zero. Must be called with spa_namespace_lock
* held.
*
* The spa_config_lock[] is an array of rwlocks, ordered as follows:
* SCL_CONFIG > SCL_STATE > SCL_ALLOC > SCL_ZIO > SCL_FREE > SCL_VDEV.
* spa_config_lock[] is manipulated with spa_config_{enter,exit,held}().
*
* To read the configuration, it suffices to hold one of these locks as reader.
* To modify the configuration, you must hold all locks as writer. To modify
* vdev state without altering the vdev tree's topology (e.g. online/offline),
* you must hold SCL_STATE and SCL_ZIO as writer.
*
* We use these distinct config locks to avoid recursive lock entry.
* For example, spa_sync() (which holds SCL_CONFIG as reader) induces
* block allocations (SCL_ALLOC), which may require reading space maps
* from disk (dmu_read() -> zio_read() -> SCL_ZIO).
*
* The spa config locks cannot be normal rwlocks because we need the
* ability to hand off ownership. For example, SCL_ZIO is acquired
* by the issuing thread and later released by an interrupt thread.
* They do, however, obey the usual write-wanted semantics to prevent
* writer (i.e. system administrator) starvation.
*
* The lock acquisition rules are as follows:
*
* SCL_CONFIG
* Protects changes to the vdev tree topology, such as vdev
* add/remove/attach/detach. Protects the dirty config list
* (spa_config_dirty_list) and the set of spares and l2arc devices.
*
* SCL_STATE
* Protects changes to pool state and vdev state, such as vdev
* online/offline/fault/degrade/clear. Protects the dirty state list
* (spa_state_dirty_list) and global pool state (spa_state).
*
* SCL_ALLOC
* Protects changes to metaslab groups and classes.
* Held as reader by metaslab_alloc() and metaslab_claim().
*
* SCL_ZIO
* Held by bp-level zios (those which have no io_vd upon entry)
* to prevent changes to the vdev tree. The bp-level zio implicitly
* protects all of its vdev child zios, which do not hold SCL_ZIO.
*
* SCL_FREE
* Protects changes to metaslab groups and classes.
* Held as reader by metaslab_free(). SCL_FREE is distinct from
* SCL_ALLOC, and lower than SCL_ZIO, so that we can safely free
* blocks in zio_done() while another i/o that holds either
* SCL_ALLOC or SCL_ZIO is waiting for this i/o to complete.
*
* SCL_VDEV
* Held as reader to prevent changes to the vdev tree during trivial
* inquiries such as bp_get_dsize(). SCL_VDEV is distinct from the
* other locks, and lower than all of them, to ensure that it's safe
* to acquire regardless of caller context.
*
* In addition, the following rules apply:
*
* (a) spa_props_lock protects pool properties, spa_config and spa_config_list.
* The lock ordering is SCL_CONFIG > spa_props_lock.
*
* (b) I/O operations on leaf vdevs. For any zio operation that takes
* an explicit vdev_t argument -- such as zio_ioctl(), zio_read_phys(),
* or zio_write_phys() -- the caller must ensure that the config cannot
* cannot change in the interim, and that the vdev cannot be reopened.
* SCL_STATE as reader suffices for both.
*
* The vdev configuration is protected by spa_vdev_enter() / spa_vdev_exit().
*
* spa_vdev_enter() Acquire the namespace lock and the config lock
* for writing.
*
* spa_vdev_exit() Release the config lock, wait for all I/O
* to complete, sync the updated configs to the
* cache, and release the namespace lock.
*
* vdev state is protected by spa_vdev_state_enter() / spa_vdev_state_exit().
* Like spa_vdev_enter/exit, these are convenience wrappers -- the actual
* locking is, always, based on spa_namespace_lock and spa_config_lock[].
*/
static avl_tree_t spa_namespace_avl;
kmutex_t spa_namespace_lock;
static kcondvar_t spa_namespace_cv;
static const int spa_max_replication_override = SPA_DVAS_PER_BP;
static kmutex_t spa_spare_lock;
static avl_tree_t spa_spare_avl;
static kmutex_t spa_l2cache_lock;
static avl_tree_t spa_l2cache_avl;
spa_mode_t spa_mode_global = SPA_MODE_UNINIT;
#ifdef ZFS_DEBUG
/*
* Everything except dprintf, set_error, spa, and indirect_remap is on
* by default in debug builds.
*/
int zfs_flags = ~(ZFS_DEBUG_DPRINTF | ZFS_DEBUG_SET_ERROR |
ZFS_DEBUG_INDIRECT_REMAP);
#else
int zfs_flags = 0;
#endif
/*
* zfs_recover can be set to nonzero to attempt to recover from
* otherwise-fatal errors, typically caused by on-disk corruption. When
* set, calls to zfs_panic_recover() will turn into warning messages.
* This should only be used as a last resort, as it typically results
* in leaked space, or worse.
*/
int zfs_recover = B_FALSE;
/*
* If destroy encounters an EIO while reading metadata (e.g. indirect
* blocks), space referenced by the missing metadata can not be freed.
* Normally this causes the background destroy to become "stalled", as
* it is unable to make forward progress. While in this stalled state,
* all remaining space to free from the error-encountering filesystem is
* "temporarily leaked". Set this flag to cause it to ignore the EIO,
* permanently leak the space from indirect blocks that can not be read,
* and continue to free everything else that it can.
*
* The default, "stalling" behavior is useful if the storage partially
* fails (i.e. some but not all i/os fail), and then later recovers. In
* this case, we will be able to continue pool operations while it is
* partially failed, and when it recovers, we can continue to free the
* space, with no leaks. However, note that this case is actually
* fairly rare.
*
* Typically pools either (a) fail completely (but perhaps temporarily,
* e.g. a top-level vdev going offline), or (b) have localized,
* permanent errors (e.g. disk returns the wrong data due to bit flip or
* firmware bug). In case (a), this setting does not matter because the
* pool will be suspended and the sync thread will not be able to make
* forward progress regardless. In case (b), because the error is
* permanent, the best we can do is leak the minimum amount of space,
* which is what setting this flag will do. Therefore, it is reasonable
* for this flag to normally be set, but we chose the more conservative
* approach of not setting it, so that there is no possibility of
* leaking space in the "partial temporary" failure case.
*/
int zfs_free_leak_on_eio = B_FALSE;
/*
* Expiration time in milliseconds. This value has two meanings. First it is
* used to determine when the spa_deadman() logic should fire. By default the
* spa_deadman() will fire if spa_sync() has not completed in 600 seconds.
* Secondly, the value determines if an I/O is considered "hung". Any I/O that
* has not completed in zfs_deadman_synctime_ms is considered "hung" resulting
* in one of three behaviors controlled by zfs_deadman_failmode.
*/
uint64_t zfs_deadman_synctime_ms = 600000UL; /* 10 min. */
/*
* This value controls the maximum amount of time zio_wait() will block for an
* outstanding IO. By default this is 300 seconds at which point the "hung"
* behavior will be applied as described for zfs_deadman_synctime_ms.
*/
uint64_t zfs_deadman_ziotime_ms = 300000UL; /* 5 min. */
/*
* Check time in milliseconds. This defines the frequency at which we check
* for hung I/O.
*/
uint64_t zfs_deadman_checktime_ms = 60000UL; /* 1 min. */
/*
* By default the deadman is enabled.
*/
int zfs_deadman_enabled = B_TRUE;
/*
* Controls the behavior of the deadman when it detects a "hung" I/O.
* Valid values are zfs_deadman_failmode=<wait|continue|panic>.
*
* wait - Wait for the "hung" I/O (default)
* continue - Attempt to recover from a "hung" I/O
* panic - Panic the system
*/
const char *zfs_deadman_failmode = "wait";
/*
* The worst case is single-sector max-parity RAID-Z blocks, in which
* case the space requirement is exactly (VDEV_RAIDZ_MAXPARITY + 1)
* times the size; so just assume that. Add to this the fact that
* we can have up to 3 DVAs per bp, and one more factor of 2 because
* the block may be dittoed with up to 3 DVAs by ddt_sync(). All together,
* the worst case is:
* (VDEV_RAIDZ_MAXPARITY + 1) * SPA_DVAS_PER_BP * 2 == 24
*/
uint_t spa_asize_inflation = 24;
/*
* Normally, we don't allow the last 3.2% (1/(2^spa_slop_shift)) of space in
* the pool to be consumed (bounded by spa_max_slop). This ensures that we
* don't run the pool completely out of space, due to unaccounted changes (e.g.
* to the MOS). It also limits the worst-case time to allocate space. If we
* have less than this amount of free space, most ZPL operations (e.g. write,
* create) will return ENOSPC. The ZIL metaslabs (spa_embedded_log_class) are
* also part of this 3.2% of space which can't be consumed by normal writes;
* the slop space "proper" (spa_get_slop_space()) is decreased by the embedded
* log space.
*
* Certain operations (e.g. file removal, most administrative actions) can
* use half the slop space. They will only return ENOSPC if less than half
* the slop space is free. Typically, once the pool has less than the slop
* space free, the user will use these operations to free up space in the pool.
* These are the operations that call dsl_pool_adjustedsize() with the netfree
* argument set to TRUE.
*
* Operations that are almost guaranteed to free up space in the absence of
* a pool checkpoint can use up to three quarters of the slop space
* (e.g zfs destroy).
*
* A very restricted set of operations are always permitted, regardless of
* the amount of free space. These are the operations that call
* dsl_sync_task(ZFS_SPACE_CHECK_NONE). If these operations result in a net
* increase in the amount of space used, it is possible to run the pool
* completely out of space, causing it to be permanently read-only.
*
* Note that on very small pools, the slop space will be larger than
* 3.2%, in an effort to have it be at least spa_min_slop (128MB),
* but we never allow it to be more than half the pool size.
*
* Further, on very large pools, the slop space will be smaller than
* 3.2%, to avoid reserving much more space than we actually need; bounded
* by spa_max_slop (128GB).
*
* See also the comments in zfs_space_check_t.
*/
uint_t spa_slop_shift = 5;
static const uint64_t spa_min_slop = 128ULL * 1024 * 1024;
static const uint64_t spa_max_slop = 128ULL * 1024 * 1024 * 1024;
static const int spa_allocators = 4;
void
spa_load_failed(spa_t *spa, const char *fmt, ...)
{
va_list adx;
char buf[256];
va_start(adx, fmt);
(void) vsnprintf(buf, sizeof (buf), fmt, adx);
va_end(adx);
zfs_dbgmsg("spa_load(%s, config %s): FAILED: %s", spa->spa_name,
spa->spa_trust_config ? "trusted" : "untrusted", buf);
}
void
spa_load_note(spa_t *spa, const char *fmt, ...)
{
va_list adx;
char buf[256];
va_start(adx, fmt);
(void) vsnprintf(buf, sizeof (buf), fmt, adx);
va_end(adx);
zfs_dbgmsg("spa_load(%s, config %s): %s", spa->spa_name,
spa->spa_trust_config ? "trusted" : "untrusted", buf);
}
/*
* By default dedup and user data indirects land in the special class
*/
static int zfs_ddt_data_is_special = B_TRUE;
static int zfs_user_indirect_is_special = B_TRUE;
/*
* The percentage of special class final space reserved for metadata only.
* Once we allocate 100 - zfs_special_class_metadata_reserve_pct we only
* let metadata into the class.
*/
static uint_t zfs_special_class_metadata_reserve_pct = 25;
/*
* ==========================================================================
* SPA config locking
* ==========================================================================
*/
static void
spa_config_lock_init(spa_t *spa)
{
for (int i = 0; i < SCL_LOCKS; i++) {
spa_config_lock_t *scl = &spa->spa_config_lock[i];
mutex_init(&scl->scl_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&scl->scl_cv, NULL, CV_DEFAULT, NULL);
scl->scl_writer = NULL;
scl->scl_write_wanted = 0;
scl->scl_count = 0;
}
}
static void
spa_config_lock_destroy(spa_t *spa)
{
for (int i = 0; i < SCL_LOCKS; i++) {
spa_config_lock_t *scl = &spa->spa_config_lock[i];
mutex_destroy(&scl->scl_lock);
cv_destroy(&scl->scl_cv);
ASSERT(scl->scl_writer == NULL);
ASSERT(scl->scl_write_wanted == 0);
ASSERT(scl->scl_count == 0);
}
}
int
spa_config_tryenter(spa_t *spa, int locks, const void *tag, krw_t rw)
{
for (int i = 0; i < SCL_LOCKS; i++) {
spa_config_lock_t *scl = &spa->spa_config_lock[i];
if (!(locks & (1 << i)))
continue;
mutex_enter(&scl->scl_lock);
if (rw == RW_READER) {
if (scl->scl_writer || scl->scl_write_wanted) {
mutex_exit(&scl->scl_lock);
spa_config_exit(spa, locks & ((1 << i) - 1),
tag);
return (0);
}
} else {
ASSERT(scl->scl_writer != curthread);
if (scl->scl_count != 0) {
mutex_exit(&scl->scl_lock);
spa_config_exit(spa, locks & ((1 << i) - 1),
tag);
return (0);
}
scl->scl_writer = curthread;
}
scl->scl_count++;
mutex_exit(&scl->scl_lock);
}
return (1);
}
-void
-spa_config_enter(spa_t *spa, int locks, const void *tag, krw_t rw)
+static void
+spa_config_enter_impl(spa_t *spa, int locks, const void *tag, krw_t rw,
+ int mmp_flag)
{
(void) tag;
int wlocks_held = 0;
ASSERT3U(SCL_LOCKS, <, sizeof (wlocks_held) * NBBY);
for (int i = 0; i < SCL_LOCKS; i++) {
spa_config_lock_t *scl = &spa->spa_config_lock[i];
if (scl->scl_writer == curthread)
wlocks_held |= (1 << i);
if (!(locks & (1 << i)))
continue;
mutex_enter(&scl->scl_lock);
if (rw == RW_READER) {
- while (scl->scl_writer || scl->scl_write_wanted) {
+ while (scl->scl_writer ||
+ (!mmp_flag && scl->scl_write_wanted)) {
cv_wait(&scl->scl_cv, &scl->scl_lock);
}
} else {
ASSERT(scl->scl_writer != curthread);
while (scl->scl_count != 0) {
scl->scl_write_wanted++;
cv_wait(&scl->scl_cv, &scl->scl_lock);
scl->scl_write_wanted--;
}
scl->scl_writer = curthread;
}
scl->scl_count++;
mutex_exit(&scl->scl_lock);
}
ASSERT3U(wlocks_held, <=, locks);
}
+void
+spa_config_enter(spa_t *spa, int locks, const void *tag, krw_t rw)
+{
+ spa_config_enter_impl(spa, locks, tag, rw, 0);
+}
+
+/*
+ * The spa_config_enter_mmp() allows the mmp thread to cut in front of
+ * outstanding write lock requests. This is needed since the mmp updates are
+ * time sensitive and failure to service them promptly will result in a
+ * suspended pool. This pool suspension has been seen in practice when there is
+ * a single disk in a pool that is responding slowly and presumably about to
+ * fail.
+ */
+
+void
+spa_config_enter_mmp(spa_t *spa, int locks, const void *tag, krw_t rw)
+{
+ spa_config_enter_impl(spa, locks, tag, rw, 1);
+}
+
void
spa_config_exit(spa_t *spa, int locks, const void *tag)
{
(void) tag;
for (int i = SCL_LOCKS - 1; i >= 0; i--) {
spa_config_lock_t *scl = &spa->spa_config_lock[i];
if (!(locks & (1 << i)))
continue;
mutex_enter(&scl->scl_lock);
ASSERT(scl->scl_count > 0);
if (--scl->scl_count == 0) {
ASSERT(scl->scl_writer == NULL ||
scl->scl_writer == curthread);
scl->scl_writer = NULL; /* OK in either case */
cv_broadcast(&scl->scl_cv);
}
mutex_exit(&scl->scl_lock);
}
}
int
spa_config_held(spa_t *spa, int locks, krw_t rw)
{
int locks_held = 0;
for (int i = 0; i < SCL_LOCKS; i++) {
spa_config_lock_t *scl = &spa->spa_config_lock[i];
if (!(locks & (1 << i)))
continue;
if ((rw == RW_READER && scl->scl_count != 0) ||
(rw == RW_WRITER && scl->scl_writer == curthread))
locks_held |= 1 << i;
}
return (locks_held);
}
/*
* ==========================================================================
* SPA namespace functions
* ==========================================================================
*/
/*
* Lookup the named spa_t in the AVL tree. The spa_namespace_lock must be held.
* Returns NULL if no matching spa_t is found.
*/
spa_t *
spa_lookup(const char *name)
{
static spa_t search; /* spa_t is large; don't allocate on stack */
spa_t *spa;
avl_index_t where;
char *cp;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
(void) strlcpy(search.spa_name, name, sizeof (search.spa_name));
/*
* If it's a full dataset name, figure out the pool name and
* just use that.
*/
cp = strpbrk(search.spa_name, "/@#");
if (cp != NULL)
*cp = '\0';
spa = avl_find(&spa_namespace_avl, &search, &where);
return (spa);
}
/*
* Fires when spa_sync has not completed within zfs_deadman_synctime_ms.
* If the zfs_deadman_enabled flag is set then it inspects all vdev queues
* looking for potentially hung I/Os.
*/
void
spa_deadman(void *arg)
{
spa_t *spa = arg;
/* Disable the deadman if the pool is suspended. */
if (spa_suspended(spa))
return;
zfs_dbgmsg("slow spa_sync: started %llu seconds ago, calls %llu",
(gethrtime() - spa->spa_sync_starttime) / NANOSEC,
(u_longlong_t)++spa->spa_deadman_calls);
if (zfs_deadman_enabled)
vdev_deadman(spa->spa_root_vdev, FTAG);
spa->spa_deadman_tqid = taskq_dispatch_delay(system_delay_taskq,
spa_deadman, spa, TQ_SLEEP, ddi_get_lbolt() +
MSEC_TO_TICK(zfs_deadman_checktime_ms));
}
static int
spa_log_sm_sort_by_txg(const void *va, const void *vb)
{
const spa_log_sm_t *a = va;
const spa_log_sm_t *b = vb;
return (TREE_CMP(a->sls_txg, b->sls_txg));
}
/*
* Create an uninitialized spa_t with the given name. Requires
* spa_namespace_lock. The caller must ensure that the spa_t doesn't already
* exist by calling spa_lookup() first.
*/
spa_t *
spa_add(const char *name, nvlist_t *config, const char *altroot)
{
spa_t *spa;
spa_config_dirent_t *dp;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
spa = kmem_zalloc(sizeof (spa_t), KM_SLEEP);
mutex_init(&spa->spa_async_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_errlist_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_errlog_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_evicting_os_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_history_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_proc_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_props_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_cksum_tmpls_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_scrub_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_suspend_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_vdev_top_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_feat_stats_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_flushed_ms_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa->spa_activities_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&spa->spa_async_cv, NULL, CV_DEFAULT, NULL);
cv_init(&spa->spa_evicting_os_cv, NULL, CV_DEFAULT, NULL);
cv_init(&spa->spa_proc_cv, NULL, CV_DEFAULT, NULL);
cv_init(&spa->spa_scrub_io_cv, NULL, CV_DEFAULT, NULL);
cv_init(&spa->spa_suspend_cv, NULL, CV_DEFAULT, NULL);
cv_init(&spa->spa_activities_cv, NULL, CV_DEFAULT, NULL);
cv_init(&spa->spa_waiters_cv, NULL, CV_DEFAULT, NULL);
for (int t = 0; t < TXG_SIZE; t++)
bplist_create(&spa->spa_free_bplist[t]);
(void) strlcpy(spa->spa_name, name, sizeof (spa->spa_name));
spa->spa_state = POOL_STATE_UNINITIALIZED;
spa->spa_freeze_txg = UINT64_MAX;
spa->spa_final_txg = UINT64_MAX;
spa->spa_load_max_txg = UINT64_MAX;
spa->spa_proc = &p0;
spa->spa_proc_state = SPA_PROC_NONE;
spa->spa_trust_config = B_TRUE;
spa->spa_hostid = zone_get_hostid(NULL);
spa->spa_deadman_synctime = MSEC2NSEC(zfs_deadman_synctime_ms);
spa->spa_deadman_ziotime = MSEC2NSEC(zfs_deadman_ziotime_ms);
spa_set_deadman_failmode(spa, zfs_deadman_failmode);
zfs_refcount_create(&spa->spa_refcount);
spa_config_lock_init(spa);
spa_stats_init(spa);
avl_add(&spa_namespace_avl, spa);
/*
* Set the alternate root, if there is one.
*/
if (altroot)
spa->spa_root = spa_strdup(altroot);
spa->spa_alloc_count = spa_allocators;
spa->spa_allocs = kmem_zalloc(spa->spa_alloc_count *
sizeof (spa_alloc_t), KM_SLEEP);
for (int i = 0; i < spa->spa_alloc_count; i++) {
mutex_init(&spa->spa_allocs[i].spaa_lock, NULL, MUTEX_DEFAULT,
NULL);
avl_create(&spa->spa_allocs[i].spaa_tree, zio_bookmark_compare,
sizeof (zio_t), offsetof(zio_t, io_alloc_node));
}
avl_create(&spa->spa_metaslabs_by_flushed, metaslab_sort_by_flushed,
sizeof (metaslab_t), offsetof(metaslab_t, ms_spa_txg_node));
avl_create(&spa->spa_sm_logs_by_txg, spa_log_sm_sort_by_txg,
sizeof (spa_log_sm_t), offsetof(spa_log_sm_t, sls_node));
list_create(&spa->spa_log_summary, sizeof (log_summary_entry_t),
offsetof(log_summary_entry_t, lse_node));
/*
* Every pool starts with the default cachefile
*/
list_create(&spa->spa_config_list, sizeof (spa_config_dirent_t),
offsetof(spa_config_dirent_t, scd_link));
dp = kmem_zalloc(sizeof (spa_config_dirent_t), KM_SLEEP);
dp->scd_path = altroot ? NULL : spa_strdup(spa_config_path);
list_insert_head(&spa->spa_config_list, dp);
VERIFY(nvlist_alloc(&spa->spa_load_info, NV_UNIQUE_NAME,
KM_SLEEP) == 0);
if (config != NULL) {
nvlist_t *features;
if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_FEATURES_FOR_READ,
&features) == 0) {
VERIFY(nvlist_dup(features, &spa->spa_label_features,
0) == 0);
}
VERIFY(nvlist_dup(config, &spa->spa_config, 0) == 0);
}
if (spa->spa_label_features == NULL) {
VERIFY(nvlist_alloc(&spa->spa_label_features, NV_UNIQUE_NAME,
KM_SLEEP) == 0);
}
spa->spa_min_ashift = INT_MAX;
spa->spa_max_ashift = 0;
spa->spa_min_alloc = INT_MAX;
/* Reset cached value */
spa->spa_dedup_dspace = ~0ULL;
/*
* As a pool is being created, treat all features as disabled by
* setting SPA_FEATURE_DISABLED for all entries in the feature
* refcount cache.
*/
for (int i = 0; i < SPA_FEATURES; i++) {
spa->spa_feat_refcount_cache[i] = SPA_FEATURE_DISABLED;
}
list_create(&spa->spa_leaf_list, sizeof (vdev_t),
offsetof(vdev_t, vdev_leaf_node));
return (spa);
}
/*
* Removes a spa_t from the namespace, freeing up any memory used. Requires
* spa_namespace_lock. This is called only after the spa_t has been closed and
* deactivated.
*/
void
spa_remove(spa_t *spa)
{
spa_config_dirent_t *dp;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
ASSERT(spa_state(spa) == POOL_STATE_UNINITIALIZED);
ASSERT3U(zfs_refcount_count(&spa->spa_refcount), ==, 0);
ASSERT0(spa->spa_waiters);
nvlist_free(spa->spa_config_splitting);
avl_remove(&spa_namespace_avl, spa);
cv_broadcast(&spa_namespace_cv);
if (spa->spa_root)
spa_strfree(spa->spa_root);
while ((dp = list_head(&spa->spa_config_list)) != NULL) {
list_remove(&spa->spa_config_list, dp);
if (dp->scd_path != NULL)
spa_strfree(dp->scd_path);
kmem_free(dp, sizeof (spa_config_dirent_t));
}
for (int i = 0; i < spa->spa_alloc_count; i++) {
avl_destroy(&spa->spa_allocs[i].spaa_tree);
mutex_destroy(&spa->spa_allocs[i].spaa_lock);
}
kmem_free(spa->spa_allocs, spa->spa_alloc_count *
sizeof (spa_alloc_t));
avl_destroy(&spa->spa_metaslabs_by_flushed);
avl_destroy(&spa->spa_sm_logs_by_txg);
list_destroy(&spa->spa_log_summary);
list_destroy(&spa->spa_config_list);
list_destroy(&spa->spa_leaf_list);
nvlist_free(spa->spa_label_features);
nvlist_free(spa->spa_load_info);
nvlist_free(spa->spa_feat_stats);
spa_config_set(spa, NULL);
zfs_refcount_destroy(&spa->spa_refcount);
spa_stats_destroy(spa);
spa_config_lock_destroy(spa);
for (int t = 0; t < TXG_SIZE; t++)
bplist_destroy(&spa->spa_free_bplist[t]);
zio_checksum_templates_free(spa);
cv_destroy(&spa->spa_async_cv);
cv_destroy(&spa->spa_evicting_os_cv);
cv_destroy(&spa->spa_proc_cv);
cv_destroy(&spa->spa_scrub_io_cv);
cv_destroy(&spa->spa_suspend_cv);
cv_destroy(&spa->spa_activities_cv);
cv_destroy(&spa->spa_waiters_cv);
mutex_destroy(&spa->spa_flushed_ms_lock);
mutex_destroy(&spa->spa_async_lock);
mutex_destroy(&spa->spa_errlist_lock);
mutex_destroy(&spa->spa_errlog_lock);
mutex_destroy(&spa->spa_evicting_os_lock);
mutex_destroy(&spa->spa_history_lock);
mutex_destroy(&spa->spa_proc_lock);
mutex_destroy(&spa->spa_props_lock);
mutex_destroy(&spa->spa_cksum_tmpls_lock);
mutex_destroy(&spa->spa_scrub_lock);
mutex_destroy(&spa->spa_suspend_lock);
mutex_destroy(&spa->spa_vdev_top_lock);
mutex_destroy(&spa->spa_feat_stats_lock);
mutex_destroy(&spa->spa_activities_lock);
kmem_free(spa, sizeof (spa_t));
}
/*
* Given a pool, return the next pool in the namespace, or NULL if there is
* none. If 'prev' is NULL, return the first pool.
*/
spa_t *
spa_next(spa_t *prev)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
if (prev)
return (AVL_NEXT(&spa_namespace_avl, prev));
else
return (avl_first(&spa_namespace_avl));
}
/*
* ==========================================================================
* SPA refcount functions
* ==========================================================================
*/
/*
* Add a reference to the given spa_t. Must have at least one reference, or
* have the namespace lock held.
*/
void
spa_open_ref(spa_t *spa, const void *tag)
{
ASSERT(zfs_refcount_count(&spa->spa_refcount) >= spa->spa_minref ||
MUTEX_HELD(&spa_namespace_lock));
(void) zfs_refcount_add(&spa->spa_refcount, tag);
}
/*
* Remove a reference to the given spa_t. Must have at least one reference, or
* have the namespace lock held.
*/
void
spa_close(spa_t *spa, const void *tag)
{
ASSERT(zfs_refcount_count(&spa->spa_refcount) > spa->spa_minref ||
MUTEX_HELD(&spa_namespace_lock));
(void) zfs_refcount_remove(&spa->spa_refcount, tag);
}
/*
* Remove a reference to the given spa_t held by a dsl dir that is
* being asynchronously released. Async releases occur from a taskq
* performing eviction of dsl datasets and dirs. The namespace lock
* isn't held and the hold by the object being evicted may contribute to
* spa_minref (e.g. dataset or directory released during pool export),
* so the asserts in spa_close() do not apply.
*/
void
spa_async_close(spa_t *spa, const void *tag)
{
(void) zfs_refcount_remove(&spa->spa_refcount, tag);
}
/*
* Check to see if the spa refcount is zero. Must be called with
* spa_namespace_lock held. We really compare against spa_minref, which is the
* number of references acquired when opening a pool
*/
boolean_t
spa_refcount_zero(spa_t *spa)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
return (zfs_refcount_count(&spa->spa_refcount) == spa->spa_minref);
}
/*
* ==========================================================================
* SPA spare and l2cache tracking
* ==========================================================================
*/
/*
* Hot spares and cache devices are tracked using the same code below,
* for 'auxiliary' devices.
*/
typedef struct spa_aux {
uint64_t aux_guid;
uint64_t aux_pool;
avl_node_t aux_avl;
int aux_count;
} spa_aux_t;
static inline int
spa_aux_compare(const void *a, const void *b)
{
const spa_aux_t *sa = (const spa_aux_t *)a;
const spa_aux_t *sb = (const spa_aux_t *)b;
return (TREE_CMP(sa->aux_guid, sb->aux_guid));
}
static void
spa_aux_add(vdev_t *vd, avl_tree_t *avl)
{
avl_index_t where;
spa_aux_t search;
spa_aux_t *aux;
search.aux_guid = vd->vdev_guid;
if ((aux = avl_find(avl, &search, &where)) != NULL) {
aux->aux_count++;
} else {
aux = kmem_zalloc(sizeof (spa_aux_t), KM_SLEEP);
aux->aux_guid = vd->vdev_guid;
aux->aux_count = 1;
avl_insert(avl, aux, where);
}
}
static void
spa_aux_remove(vdev_t *vd, avl_tree_t *avl)
{
spa_aux_t search;
spa_aux_t *aux;
avl_index_t where;
search.aux_guid = vd->vdev_guid;
aux = avl_find(avl, &search, &where);
ASSERT(aux != NULL);
if (--aux->aux_count == 0) {
avl_remove(avl, aux);
kmem_free(aux, sizeof (spa_aux_t));
} else if (aux->aux_pool == spa_guid(vd->vdev_spa)) {
aux->aux_pool = 0ULL;
}
}
static boolean_t
spa_aux_exists(uint64_t guid, uint64_t *pool, int *refcnt, avl_tree_t *avl)
{
spa_aux_t search, *found;
search.aux_guid = guid;
found = avl_find(avl, &search, NULL);
if (pool) {
if (found)
*pool = found->aux_pool;
else
*pool = 0ULL;
}
if (refcnt) {
if (found)
*refcnt = found->aux_count;
else
*refcnt = 0;
}
return (found != NULL);
}
static void
spa_aux_activate(vdev_t *vd, avl_tree_t *avl)
{
spa_aux_t search, *found;
avl_index_t where;
search.aux_guid = vd->vdev_guid;
found = avl_find(avl, &search, &where);
ASSERT(found != NULL);
ASSERT(found->aux_pool == 0ULL);
found->aux_pool = spa_guid(vd->vdev_spa);
}
/*
* Spares are tracked globally due to the following constraints:
*
* - A spare may be part of multiple pools.
* - A spare may be added to a pool even if it's actively in use within
* another pool.
* - A spare in use in any pool can only be the source of a replacement if
* the target is a spare in the same pool.
*
* We keep track of all spares on the system through the use of a reference
* counted AVL tree. When a vdev is added as a spare, or used as a replacement
* spare, then we bump the reference count in the AVL tree. In addition, we set
* the 'vdev_isspare' member to indicate that the device is a spare (active or
* inactive). When a spare is made active (used to replace a device in the
* pool), we also keep track of which pool its been made a part of.
*
* The 'spa_spare_lock' protects the AVL tree. These functions are normally
* called under the spa_namespace lock as part of vdev reconfiguration. The
* separate spare lock exists for the status query path, which does not need to
* be completely consistent with respect to other vdev configuration changes.
*/
static int
spa_spare_compare(const void *a, const void *b)
{
return (spa_aux_compare(a, b));
}
void
spa_spare_add(vdev_t *vd)
{
mutex_enter(&spa_spare_lock);
ASSERT(!vd->vdev_isspare);
spa_aux_add(vd, &spa_spare_avl);
vd->vdev_isspare = B_TRUE;
mutex_exit(&spa_spare_lock);
}
void
spa_spare_remove(vdev_t *vd)
{
mutex_enter(&spa_spare_lock);
ASSERT(vd->vdev_isspare);
spa_aux_remove(vd, &spa_spare_avl);
vd->vdev_isspare = B_FALSE;
mutex_exit(&spa_spare_lock);
}
boolean_t
spa_spare_exists(uint64_t guid, uint64_t *pool, int *refcnt)
{
boolean_t found;
mutex_enter(&spa_spare_lock);
found = spa_aux_exists(guid, pool, refcnt, &spa_spare_avl);
mutex_exit(&spa_spare_lock);
return (found);
}
void
spa_spare_activate(vdev_t *vd)
{
mutex_enter(&spa_spare_lock);
ASSERT(vd->vdev_isspare);
spa_aux_activate(vd, &spa_spare_avl);
mutex_exit(&spa_spare_lock);
}
/*
* Level 2 ARC devices are tracked globally for the same reasons as spares.
* Cache devices currently only support one pool per cache device, and so
* for these devices the aux reference count is currently unused beyond 1.
*/
static int
spa_l2cache_compare(const void *a, const void *b)
{
return (spa_aux_compare(a, b));
}
void
spa_l2cache_add(vdev_t *vd)
{
mutex_enter(&spa_l2cache_lock);
ASSERT(!vd->vdev_isl2cache);
spa_aux_add(vd, &spa_l2cache_avl);
vd->vdev_isl2cache = B_TRUE;
mutex_exit(&spa_l2cache_lock);
}
void
spa_l2cache_remove(vdev_t *vd)
{
mutex_enter(&spa_l2cache_lock);
ASSERT(vd->vdev_isl2cache);
spa_aux_remove(vd, &spa_l2cache_avl);
vd->vdev_isl2cache = B_FALSE;
mutex_exit(&spa_l2cache_lock);
}
boolean_t
spa_l2cache_exists(uint64_t guid, uint64_t *pool)
{
boolean_t found;
mutex_enter(&spa_l2cache_lock);
found = spa_aux_exists(guid, pool, NULL, &spa_l2cache_avl);
mutex_exit(&spa_l2cache_lock);
return (found);
}
void
spa_l2cache_activate(vdev_t *vd)
{
mutex_enter(&spa_l2cache_lock);
ASSERT(vd->vdev_isl2cache);
spa_aux_activate(vd, &spa_l2cache_avl);
mutex_exit(&spa_l2cache_lock);
}
/*
* ==========================================================================
* SPA vdev locking
* ==========================================================================
*/
/*
* Lock the given spa_t for the purpose of adding or removing a vdev.
* Grabs the global spa_namespace_lock plus the spa config lock for writing.
* It returns the next transaction group for the spa_t.
*/
uint64_t
spa_vdev_enter(spa_t *spa)
{
mutex_enter(&spa->spa_vdev_top_lock);
mutex_enter(&spa_namespace_lock);
vdev_autotrim_stop_all(spa);
return (spa_vdev_config_enter(spa));
}
/*
* The same as spa_vdev_enter() above but additionally takes the guid of
* the vdev being detached. When there is a rebuild in process it will be
* suspended while the vdev tree is modified then resumed by spa_vdev_exit().
* The rebuild is canceled if only a single child remains after the detach.
*/
uint64_t
spa_vdev_detach_enter(spa_t *spa, uint64_t guid)
{
mutex_enter(&spa->spa_vdev_top_lock);
mutex_enter(&spa_namespace_lock);
vdev_autotrim_stop_all(spa);
if (guid != 0) {
vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE);
if (vd) {
vdev_rebuild_stop_wait(vd->vdev_top);
}
}
return (spa_vdev_config_enter(spa));
}
/*
* Internal implementation for spa_vdev_enter(). Used when a vdev
* operation requires multiple syncs (i.e. removing a device) while
* keeping the spa_namespace_lock held.
*/
uint64_t
spa_vdev_config_enter(spa_t *spa)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
spa_config_enter(spa, SCL_ALL, spa, RW_WRITER);
return (spa_last_synced_txg(spa) + 1);
}
/*
* Used in combination with spa_vdev_config_enter() to allow the syncing
* of multiple transactions without releasing the spa_namespace_lock.
*/
void
spa_vdev_config_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error,
const char *tag)
{
ASSERT(MUTEX_HELD(&spa_namespace_lock));
int config_changed = B_FALSE;
ASSERT(txg > spa_last_synced_txg(spa));
spa->spa_pending_vdev = NULL;
/*
* Reassess the DTLs.
*/
vdev_dtl_reassess(spa->spa_root_vdev, 0, 0, B_FALSE, B_FALSE);
if (error == 0 && !list_is_empty(&spa->spa_config_dirty_list)) {
config_changed = B_TRUE;
spa->spa_config_generation++;
}
/*
* Verify the metaslab classes.
*/
ASSERT(metaslab_class_validate(spa_normal_class(spa)) == 0);
ASSERT(metaslab_class_validate(spa_log_class(spa)) == 0);
ASSERT(metaslab_class_validate(spa_embedded_log_class(spa)) == 0);
ASSERT(metaslab_class_validate(spa_special_class(spa)) == 0);
ASSERT(metaslab_class_validate(spa_dedup_class(spa)) == 0);
spa_config_exit(spa, SCL_ALL, spa);
/*
* Panic the system if the specified tag requires it. This
* is useful for ensuring that configurations are updated
* transactionally.
*/
if (zio_injection_enabled)
zio_handle_panic_injection(spa, tag, 0);
/*
* Note: this txg_wait_synced() is important because it ensures
* that there won't be more than one config change per txg.
* This allows us to use the txg as the generation number.
*/
if (error == 0)
txg_wait_synced(spa->spa_dsl_pool, txg);
if (vd != NULL) {
ASSERT(!vd->vdev_detached || vd->vdev_dtl_sm == NULL);
if (vd->vdev_ops->vdev_op_leaf) {
mutex_enter(&vd->vdev_initialize_lock);
vdev_initialize_stop(vd, VDEV_INITIALIZE_CANCELED,
NULL);
mutex_exit(&vd->vdev_initialize_lock);
mutex_enter(&vd->vdev_trim_lock);
vdev_trim_stop(vd, VDEV_TRIM_CANCELED, NULL);
mutex_exit(&vd->vdev_trim_lock);
}
/*
* The vdev may be both a leaf and top-level device.
*/
vdev_autotrim_stop_wait(vd);
spa_config_enter(spa, SCL_STATE_ALL, spa, RW_WRITER);
vdev_free(vd);
spa_config_exit(spa, SCL_STATE_ALL, spa);
}
/*
* If the config changed, update the config cache.
*/
if (config_changed)
spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE);
}
/*
* Unlock the spa_t after adding or removing a vdev. Besides undoing the
* locking of spa_vdev_enter(), we also want make sure the transactions have
* synced to disk, and then update the global configuration cache with the new
* information.
*/
int
spa_vdev_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error)
{
vdev_autotrim_restart(spa);
vdev_rebuild_restart(spa);
spa_vdev_config_exit(spa, vd, txg, error, FTAG);
mutex_exit(&spa_namespace_lock);
mutex_exit(&spa->spa_vdev_top_lock);
return (error);
}
/*
* Lock the given spa_t for the purpose of changing vdev state.
*/
void
spa_vdev_state_enter(spa_t *spa, int oplocks)
{
int locks = SCL_STATE_ALL | oplocks;
/*
* Root pools may need to read of the underlying devfs filesystem
* when opening up a vdev. Unfortunately if we're holding the
* SCL_ZIO lock it will result in a deadlock when we try to issue
* the read from the root filesystem. Instead we "prefetch"
* the associated vnodes that we need prior to opening the
* underlying devices and cache them so that we can prevent
* any I/O when we are doing the actual open.
*/
if (spa_is_root(spa)) {
int low = locks & ~(SCL_ZIO - 1);
int high = locks & ~low;
spa_config_enter(spa, high, spa, RW_WRITER);
vdev_hold(spa->spa_root_vdev);
spa_config_enter(spa, low, spa, RW_WRITER);
} else {
spa_config_enter(spa, locks, spa, RW_WRITER);
}
spa->spa_vdev_locks = locks;
}
int
spa_vdev_state_exit(spa_t *spa, vdev_t *vd, int error)
{
boolean_t config_changed = B_FALSE;
vdev_t *vdev_top;
if (vd == NULL || vd == spa->spa_root_vdev) {
vdev_top = spa->spa_root_vdev;
} else {
vdev_top = vd->vdev_top;
}
if (vd != NULL || error == 0)
vdev_dtl_reassess(vdev_top, 0, 0, B_FALSE, B_FALSE);
if (vd != NULL) {
if (vd != spa->spa_root_vdev)
vdev_state_dirty(vdev_top);
config_changed = B_TRUE;
spa->spa_config_generation++;
}
if (spa_is_root(spa))
vdev_rele(spa->spa_root_vdev);
ASSERT3U(spa->spa_vdev_locks, >=, SCL_STATE_ALL);
spa_config_exit(spa, spa->spa_vdev_locks, spa);
/*
* If anything changed, wait for it to sync. This ensures that,
* from the system administrator's perspective, zpool(8) commands
* are synchronous. This is important for things like zpool offline:
* when the command completes, you expect no further I/O from ZFS.
*/
if (vd != NULL)
txg_wait_synced(spa->spa_dsl_pool, 0);
/*
* If the config changed, update the config cache.
*/
if (config_changed) {
mutex_enter(&spa_namespace_lock);
spa_write_cachefile(spa, B_FALSE, B_TRUE, B_FALSE);
mutex_exit(&spa_namespace_lock);
}
return (error);
}
/*
* ==========================================================================
* Miscellaneous functions
* ==========================================================================
*/
void
spa_activate_mos_feature(spa_t *spa, const char *feature, dmu_tx_t *tx)
{
if (!nvlist_exists(spa->spa_label_features, feature)) {
fnvlist_add_boolean(spa->spa_label_features, feature);
/*
* When we are creating the pool (tx_txg==TXG_INITIAL), we can't
* dirty the vdev config because lock SCL_CONFIG is not held.
* Thankfully, in this case we don't need to dirty the config
* because it will be written out anyway when we finish
* creating the pool.
*/
if (tx->tx_txg != TXG_INITIAL)
vdev_config_dirty(spa->spa_root_vdev);
}
}
void
spa_deactivate_mos_feature(spa_t *spa, const char *feature)
{
if (nvlist_remove_all(spa->spa_label_features, feature) == 0)
vdev_config_dirty(spa->spa_root_vdev);
}
/*
* Return the spa_t associated with given pool_guid, if it exists. If
* device_guid is non-zero, determine whether the pool exists *and* contains
* a device with the specified device_guid.
*/
spa_t *
spa_by_guid(uint64_t pool_guid, uint64_t device_guid)
{
spa_t *spa;
avl_tree_t *t = &spa_namespace_avl;
ASSERT(MUTEX_HELD(&spa_namespace_lock));
for (spa = avl_first(t); spa != NULL; spa = AVL_NEXT(t, spa)) {
if (spa->spa_state == POOL_STATE_UNINITIALIZED)
continue;
if (spa->spa_root_vdev == NULL)
continue;
if (spa_guid(spa) == pool_guid) {
if (device_guid == 0)
break;
if (vdev_lookup_by_guid(spa->spa_root_vdev,
device_guid) != NULL)
break;
/*
* Check any devices we may be in the process of adding.
*/
if (spa->spa_pending_vdev) {
if (vdev_lookup_by_guid(spa->spa_pending_vdev,
device_guid) != NULL)
break;
}
}
}
return (spa);
}
/*
* Determine whether a pool with the given pool_guid exists.
*/
boolean_t
spa_guid_exists(uint64_t pool_guid, uint64_t device_guid)
{
return (spa_by_guid(pool_guid, device_guid) != NULL);
}
char *
spa_strdup(const char *s)
{
size_t len;
char *new;
len = strlen(s);
new = kmem_alloc(len + 1, KM_SLEEP);
memcpy(new, s, len + 1);
return (new);
}
void
spa_strfree(char *s)
{
kmem_free(s, strlen(s) + 1);
}
uint64_t
spa_generate_guid(spa_t *spa)
{
uint64_t guid;
if (spa != NULL) {
do {
(void) random_get_pseudo_bytes((void *)&guid,
sizeof (guid));
} while (guid == 0 || spa_guid_exists(spa_guid(spa), guid));
} else {
do {
(void) random_get_pseudo_bytes((void *)&guid,
sizeof (guid));
} while (guid == 0 || spa_guid_exists(guid, 0));
}
return (guid);
}
void
snprintf_blkptr(char *buf, size_t buflen, const blkptr_t *bp)
{
char type[256];
const char *checksum = NULL;
const char *compress = NULL;
if (bp != NULL) {
if (BP_GET_TYPE(bp) & DMU_OT_NEWTYPE) {
dmu_object_byteswap_t bswap =
DMU_OT_BYTESWAP(BP_GET_TYPE(bp));
(void) snprintf(type, sizeof (type), "bswap %s %s",
DMU_OT_IS_METADATA(BP_GET_TYPE(bp)) ?
"metadata" : "data",
dmu_ot_byteswap[bswap].ob_name);
} else {
(void) strlcpy(type, dmu_ot[BP_GET_TYPE(bp)].ot_name,
sizeof (type));
}
if (!BP_IS_EMBEDDED(bp)) {
checksum =
zio_checksum_table[BP_GET_CHECKSUM(bp)].ci_name;
}
compress = zio_compress_table[BP_GET_COMPRESS(bp)].ci_name;
}
SNPRINTF_BLKPTR(kmem_scnprintf, ' ', buf, buflen, bp, type, checksum,
compress);
}
void
spa_freeze(spa_t *spa)
{
uint64_t freeze_txg = 0;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
if (spa->spa_freeze_txg == UINT64_MAX) {
freeze_txg = spa_last_synced_txg(spa) + TXG_SIZE;
spa->spa_freeze_txg = freeze_txg;
}
spa_config_exit(spa, SCL_ALL, FTAG);
if (freeze_txg != 0)
txg_wait_synced(spa_get_dsl(spa), freeze_txg);
}
void
zfs_panic_recover(const char *fmt, ...)
{
va_list adx;
va_start(adx, fmt);
vcmn_err(zfs_recover ? CE_WARN : CE_PANIC, fmt, adx);
va_end(adx);
}
/*
* This is a stripped-down version of strtoull, suitable only for converting
* lowercase hexadecimal numbers that don't overflow.
*/
uint64_t
zfs_strtonum(const char *str, char **nptr)
{
uint64_t val = 0;
char c;
int digit;
while ((c = *str) != '\0') {
if (c >= '0' && c <= '9')
digit = c - '0';
else if (c >= 'a' && c <= 'f')
digit = 10 + c - 'a';
else
break;
val *= 16;
val += digit;
str++;
}
if (nptr)
*nptr = (char *)str;
return (val);
}
void
spa_activate_allocation_classes(spa_t *spa, dmu_tx_t *tx)
{
/*
* We bump the feature refcount for each special vdev added to the pool
*/
ASSERT(spa_feature_is_enabled(spa, SPA_FEATURE_ALLOCATION_CLASSES));
spa_feature_incr(spa, SPA_FEATURE_ALLOCATION_CLASSES, tx);
}
/*
* ==========================================================================
* Accessor functions
* ==========================================================================
*/
boolean_t
spa_shutting_down(spa_t *spa)
{
return (spa->spa_async_suspended);
}
dsl_pool_t *
spa_get_dsl(spa_t *spa)
{
return (spa->spa_dsl_pool);
}
boolean_t
spa_is_initializing(spa_t *spa)
{
return (spa->spa_is_initializing);
}
boolean_t
spa_indirect_vdevs_loaded(spa_t *spa)
{
return (spa->spa_indirect_vdevs_loaded);
}
blkptr_t *
spa_get_rootblkptr(spa_t *spa)
{
return (&spa->spa_ubsync.ub_rootbp);
}
void
spa_set_rootblkptr(spa_t *spa, const blkptr_t *bp)
{
spa->spa_uberblock.ub_rootbp = *bp;
}
void
spa_altroot(spa_t *spa, char *buf, size_t buflen)
{
if (spa->spa_root == NULL)
buf[0] = '\0';
else
(void) strlcpy(buf, spa->spa_root, buflen);
}
uint32_t
spa_sync_pass(spa_t *spa)
{
return (spa->spa_sync_pass);
}
char *
spa_name(spa_t *spa)
{
return (spa->spa_name);
}
uint64_t
spa_guid(spa_t *spa)
{
dsl_pool_t *dp = spa_get_dsl(spa);
uint64_t guid;
/*
* If we fail to parse the config during spa_load(), we can go through
* the error path (which posts an ereport) and end up here with no root
* vdev. We stash the original pool guid in 'spa_config_guid' to handle
* this case.
*/
if (spa->spa_root_vdev == NULL)
return (spa->spa_config_guid);
guid = spa->spa_last_synced_guid != 0 ?
spa->spa_last_synced_guid : spa->spa_root_vdev->vdev_guid;
/*
* Return the most recently synced out guid unless we're
* in syncing context.
*/
if (dp && dsl_pool_sync_context(dp))
return (spa->spa_root_vdev->vdev_guid);
else
return (guid);
}
uint64_t
spa_load_guid(spa_t *spa)
{
/*
* This is a GUID that exists solely as a reference for the
* purposes of the arc. It is generated at load time, and
* is never written to persistent storage.
*/
return (spa->spa_load_guid);
}
uint64_t
spa_last_synced_txg(spa_t *spa)
{
return (spa->spa_ubsync.ub_txg);
}
uint64_t
spa_first_txg(spa_t *spa)
{
return (spa->spa_first_txg);
}
uint64_t
spa_syncing_txg(spa_t *spa)
{
return (spa->spa_syncing_txg);
}
/*
* Return the last txg where data can be dirtied. The final txgs
* will be used to just clear out any deferred frees that remain.
*/
uint64_t
spa_final_dirty_txg(spa_t *spa)
{
return (spa->spa_final_txg - TXG_DEFER_SIZE);
}
pool_state_t
spa_state(spa_t *spa)
{
return (spa->spa_state);
}
spa_load_state_t
spa_load_state(spa_t *spa)
{
return (spa->spa_load_state);
}
uint64_t
spa_freeze_txg(spa_t *spa)
{
return (spa->spa_freeze_txg);
}
/*
* Return the inflated asize for a logical write in bytes. This is used by the
* DMU to calculate the space a logical write will require on disk.
* If lsize is smaller than the largest physical block size allocatable on this
* pool we use its value instead, since the write will end up using the whole
* block anyway.
*/
uint64_t
spa_get_worst_case_asize(spa_t *spa, uint64_t lsize)
{
if (lsize == 0)
return (0); /* No inflation needed */
return (MAX(lsize, 1 << spa->spa_max_ashift) * spa_asize_inflation);
}
/*
* Return the amount of slop space in bytes. It is typically 1/32 of the pool
* (3.2%), minus the embedded log space. On very small pools, it may be
* slightly larger than this. On very large pools, it will be capped to
* the value of spa_max_slop. The embedded log space is not included in
* spa_dspace. By subtracting it, the usable space (per "zfs list") is a
* constant 97% of the total space, regardless of metaslab size (assuming the
* default spa_slop_shift=5 and a non-tiny pool).
*
* See the comment above spa_slop_shift for more details.
*/
uint64_t
spa_get_slop_space(spa_t *spa)
{
uint64_t space = 0;
uint64_t slop = 0;
/*
* Make sure spa_dedup_dspace has been set.
*/
if (spa->spa_dedup_dspace == ~0ULL)
spa_update_dspace(spa);
/*
* spa_get_dspace() includes the space only logically "used" by
* deduplicated data, so since it's not useful to reserve more
* space with more deduplicated data, we subtract that out here.
*/
space = spa_get_dspace(spa) - spa->spa_dedup_dspace;
slop = MIN(space >> spa_slop_shift, spa_max_slop);
/*
* Subtract the embedded log space, but no more than half the (3.2%)
* unusable space. Note, the "no more than half" is only relevant if
* zfs_embedded_slog_min_ms >> spa_slop_shift < 2, which is not true by
* default.
*/
uint64_t embedded_log =
metaslab_class_get_dspace(spa_embedded_log_class(spa));
slop -= MIN(embedded_log, slop >> 1);
/*
* Slop space should be at least spa_min_slop, but no more than half
* the entire pool.
*/
slop = MAX(slop, MIN(space >> 1, spa_min_slop));
return (slop);
}
uint64_t
spa_get_dspace(spa_t *spa)
{
return (spa->spa_dspace);
}
uint64_t
spa_get_checkpoint_space(spa_t *spa)
{
return (spa->spa_checkpoint_info.sci_dspace);
}
void
spa_update_dspace(spa_t *spa)
{
spa->spa_dspace = metaslab_class_get_dspace(spa_normal_class(spa)) +
ddt_get_dedup_dspace(spa) + brt_get_dspace(spa);
if (spa->spa_nonallocating_dspace > 0) {
/*
* Subtract the space provided by all non-allocating vdevs that
* contribute to dspace. If a file is overwritten, its old
* blocks are freed and new blocks are allocated. If there are
* no snapshots of the file, the available space should remain
* the same. The old blocks could be freed from the
* non-allocating vdev, but the new blocks must be allocated on
* other (allocating) vdevs. By reserving the entire size of
* the non-allocating vdevs (including allocated space), we
* ensure that there will be enough space on the allocating
* vdevs for this file overwrite to succeed.
*
* Note that the DMU/DSL doesn't actually know or care
* how much space is allocated (it does its own tracking
* of how much space has been logically used). So it
* doesn't matter that the data we are moving may be
* allocated twice (on the old device and the new device).
*/
ASSERT3U(spa->spa_dspace, >=, spa->spa_nonallocating_dspace);
spa->spa_dspace -= spa->spa_nonallocating_dspace;
}
}
/*
* Return the failure mode that has been set to this pool. The default
* behavior will be to block all I/Os when a complete failure occurs.
*/
uint64_t
spa_get_failmode(spa_t *spa)
{
return (spa->spa_failmode);
}
boolean_t
spa_suspended(spa_t *spa)
{
return (spa->spa_suspended != ZIO_SUSPEND_NONE);
}
uint64_t
spa_version(spa_t *spa)
{
return (spa->spa_ubsync.ub_version);
}
boolean_t
spa_deflate(spa_t *spa)
{
return (spa->spa_deflate);
}
metaslab_class_t *
spa_normal_class(spa_t *spa)
{
return (spa->spa_normal_class);
}
metaslab_class_t *
spa_log_class(spa_t *spa)
{
return (spa->spa_log_class);
}
metaslab_class_t *
spa_embedded_log_class(spa_t *spa)
{
return (spa->spa_embedded_log_class);
}
metaslab_class_t *
spa_special_class(spa_t *spa)
{
return (spa->spa_special_class);
}
metaslab_class_t *
spa_dedup_class(spa_t *spa)
{
return (spa->spa_dedup_class);
}
/*
* Locate an appropriate allocation class
*/
metaslab_class_t *
spa_preferred_class(spa_t *spa, uint64_t size, dmu_object_type_t objtype,
uint_t level, uint_t special_smallblk)
{
/*
* ZIL allocations determine their class in zio_alloc_zil().
*/
ASSERT(objtype != DMU_OT_INTENT_LOG);
boolean_t has_special_class = spa->spa_special_class->mc_groups != 0;
if (DMU_OT_IS_DDT(objtype)) {
if (spa->spa_dedup_class->mc_groups != 0)
return (spa_dedup_class(spa));
else if (has_special_class && zfs_ddt_data_is_special)
return (spa_special_class(spa));
else
return (spa_normal_class(spa));
}
/* Indirect blocks for user data can land in special if allowed */
if (level > 0 && (DMU_OT_IS_FILE(objtype) || objtype == DMU_OT_ZVOL)) {
if (has_special_class && zfs_user_indirect_is_special)
return (spa_special_class(spa));
else
return (spa_normal_class(spa));
}
if (DMU_OT_IS_METADATA(objtype) || level > 0) {
if (has_special_class)
return (spa_special_class(spa));
else
return (spa_normal_class(spa));
}
/*
* Allow small file blocks in special class in some cases (like
* for the dRAID vdev feature). But always leave a reserve of
* zfs_special_class_metadata_reserve_pct exclusively for metadata.
*/
if (DMU_OT_IS_FILE(objtype) &&
has_special_class && size <= special_smallblk) {
metaslab_class_t *special = spa_special_class(spa);
uint64_t alloc = metaslab_class_get_alloc(special);
uint64_t space = metaslab_class_get_space(special);
uint64_t limit =
(space * (100 - zfs_special_class_metadata_reserve_pct))
/ 100;
if (alloc < limit)
return (special);
}
return (spa_normal_class(spa));
}
void
spa_evicting_os_register(spa_t *spa, objset_t *os)
{
mutex_enter(&spa->spa_evicting_os_lock);
list_insert_head(&spa->spa_evicting_os_list, os);
mutex_exit(&spa->spa_evicting_os_lock);
}
void
spa_evicting_os_deregister(spa_t *spa, objset_t *os)
{
mutex_enter(&spa->spa_evicting_os_lock);
list_remove(&spa->spa_evicting_os_list, os);
cv_broadcast(&spa->spa_evicting_os_cv);
mutex_exit(&spa->spa_evicting_os_lock);
}
void
spa_evicting_os_wait(spa_t *spa)
{
mutex_enter(&spa->spa_evicting_os_lock);
while (!list_is_empty(&spa->spa_evicting_os_list))
cv_wait(&spa->spa_evicting_os_cv, &spa->spa_evicting_os_lock);
mutex_exit(&spa->spa_evicting_os_lock);
dmu_buf_user_evict_wait();
}
int
spa_max_replication(spa_t *spa)
{
/*
* As of SPA_VERSION == SPA_VERSION_DITTO_BLOCKS, we are able to
* handle BPs with more than one DVA allocated. Set our max
* replication level accordingly.
*/
if (spa_version(spa) < SPA_VERSION_DITTO_BLOCKS)
return (1);
return (MIN(SPA_DVAS_PER_BP, spa_max_replication_override));
}
int
spa_prev_software_version(spa_t *spa)
{
return (spa->spa_prev_software_version);
}
uint64_t
spa_deadman_synctime(spa_t *spa)
{
return (spa->spa_deadman_synctime);
}
spa_autotrim_t
spa_get_autotrim(spa_t *spa)
{
return (spa->spa_autotrim);
}
uint64_t
spa_deadman_ziotime(spa_t *spa)
{
return (spa->spa_deadman_ziotime);
}
uint64_t
spa_get_deadman_failmode(spa_t *spa)
{
return (spa->spa_deadman_failmode);
}
void
spa_set_deadman_failmode(spa_t *spa, const char *failmode)
{
if (strcmp(failmode, "wait") == 0)
spa->spa_deadman_failmode = ZIO_FAILURE_MODE_WAIT;
else if (strcmp(failmode, "continue") == 0)
spa->spa_deadman_failmode = ZIO_FAILURE_MODE_CONTINUE;
else if (strcmp(failmode, "panic") == 0)
spa->spa_deadman_failmode = ZIO_FAILURE_MODE_PANIC;
else
spa->spa_deadman_failmode = ZIO_FAILURE_MODE_WAIT;
}
void
spa_set_deadman_ziotime(hrtime_t ns)
{
spa_t *spa = NULL;
if (spa_mode_global != SPA_MODE_UNINIT) {
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(spa)) != NULL)
spa->spa_deadman_ziotime = ns;
mutex_exit(&spa_namespace_lock);
}
}
void
spa_set_deadman_synctime(hrtime_t ns)
{
spa_t *spa = NULL;
if (spa_mode_global != SPA_MODE_UNINIT) {
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(spa)) != NULL)
spa->spa_deadman_synctime = ns;
mutex_exit(&spa_namespace_lock);
}
}
uint64_t
dva_get_dsize_sync(spa_t *spa, const dva_t *dva)
{
uint64_t asize = DVA_GET_ASIZE(dva);
uint64_t dsize = asize;
ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
if (asize != 0 && spa->spa_deflate) {
vdev_t *vd = vdev_lookup_top(spa, DVA_GET_VDEV(dva));
if (vd != NULL)
dsize = (asize >> SPA_MINBLOCKSHIFT) *
vd->vdev_deflate_ratio;
}
return (dsize);
}
uint64_t
bp_get_dsize_sync(spa_t *spa, const blkptr_t *bp)
{
uint64_t dsize = 0;
for (int d = 0; d < BP_GET_NDVAS(bp); d++)
dsize += dva_get_dsize_sync(spa, &bp->blk_dva[d]);
return (dsize);
}
uint64_t
bp_get_dsize(spa_t *spa, const blkptr_t *bp)
{
uint64_t dsize = 0;
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
for (int d = 0; d < BP_GET_NDVAS(bp); d++)
dsize += dva_get_dsize_sync(spa, &bp->blk_dva[d]);
spa_config_exit(spa, SCL_VDEV, FTAG);
return (dsize);
}
uint64_t
spa_dirty_data(spa_t *spa)
{
return (spa->spa_dsl_pool->dp_dirty_total);
}
/*
* ==========================================================================
* SPA Import Progress Routines
* ==========================================================================
*/
typedef struct spa_import_progress {
uint64_t pool_guid; /* unique id for updates */
char *pool_name;
spa_load_state_t spa_load_state;
uint64_t mmp_sec_remaining; /* MMP activity check */
uint64_t spa_load_max_txg; /* rewind txg */
procfs_list_node_t smh_node;
} spa_import_progress_t;
spa_history_list_t *spa_import_progress_list = NULL;
static int
spa_import_progress_show_header(struct seq_file *f)
{
seq_printf(f, "%-20s %-14s %-14s %-12s %s\n", "pool_guid",
"load_state", "multihost_secs", "max_txg",
"pool_name");
return (0);
}
static int
spa_import_progress_show(struct seq_file *f, void *data)
{
spa_import_progress_t *sip = (spa_import_progress_t *)data;
seq_printf(f, "%-20llu %-14llu %-14llu %-12llu %s\n",
(u_longlong_t)sip->pool_guid, (u_longlong_t)sip->spa_load_state,
(u_longlong_t)sip->mmp_sec_remaining,
(u_longlong_t)sip->spa_load_max_txg,
(sip->pool_name ? sip->pool_name : "-"));
return (0);
}
/* Remove oldest elements from list until there are no more than 'size' left */
static void
spa_import_progress_truncate(spa_history_list_t *shl, unsigned int size)
{
spa_import_progress_t *sip;
while (shl->size > size) {
sip = list_remove_head(&shl->procfs_list.pl_list);
if (sip->pool_name)
spa_strfree(sip->pool_name);
kmem_free(sip, sizeof (spa_import_progress_t));
shl->size--;
}
IMPLY(size == 0, list_is_empty(&shl->procfs_list.pl_list));
}
static void
spa_import_progress_init(void)
{
spa_import_progress_list = kmem_zalloc(sizeof (spa_history_list_t),
KM_SLEEP);
spa_import_progress_list->size = 0;
spa_import_progress_list->procfs_list.pl_private =
spa_import_progress_list;
procfs_list_install("zfs",
NULL,
"import_progress",
0644,
&spa_import_progress_list->procfs_list,
spa_import_progress_show,
spa_import_progress_show_header,
NULL,
offsetof(spa_import_progress_t, smh_node));
}
static void
spa_import_progress_destroy(void)
{
spa_history_list_t *shl = spa_import_progress_list;
procfs_list_uninstall(&shl->procfs_list);
spa_import_progress_truncate(shl, 0);
procfs_list_destroy(&shl->procfs_list);
kmem_free(shl, sizeof (spa_history_list_t));
}
int
spa_import_progress_set_state(uint64_t pool_guid,
spa_load_state_t load_state)
{
spa_history_list_t *shl = spa_import_progress_list;
spa_import_progress_t *sip;
int error = ENOENT;
if (shl->size == 0)
return (0);
mutex_enter(&shl->procfs_list.pl_lock);
for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL;
sip = list_prev(&shl->procfs_list.pl_list, sip)) {
if (sip->pool_guid == pool_guid) {
sip->spa_load_state = load_state;
error = 0;
break;
}
}
mutex_exit(&shl->procfs_list.pl_lock);
return (error);
}
int
spa_import_progress_set_max_txg(uint64_t pool_guid, uint64_t load_max_txg)
{
spa_history_list_t *shl = spa_import_progress_list;
spa_import_progress_t *sip;
int error = ENOENT;
if (shl->size == 0)
return (0);
mutex_enter(&shl->procfs_list.pl_lock);
for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL;
sip = list_prev(&shl->procfs_list.pl_list, sip)) {
if (sip->pool_guid == pool_guid) {
sip->spa_load_max_txg = load_max_txg;
error = 0;
break;
}
}
mutex_exit(&shl->procfs_list.pl_lock);
return (error);
}
int
spa_import_progress_set_mmp_check(uint64_t pool_guid,
uint64_t mmp_sec_remaining)
{
spa_history_list_t *shl = spa_import_progress_list;
spa_import_progress_t *sip;
int error = ENOENT;
if (shl->size == 0)
return (0);
mutex_enter(&shl->procfs_list.pl_lock);
for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL;
sip = list_prev(&shl->procfs_list.pl_list, sip)) {
if (sip->pool_guid == pool_guid) {
sip->mmp_sec_remaining = mmp_sec_remaining;
error = 0;
break;
}
}
mutex_exit(&shl->procfs_list.pl_lock);
return (error);
}
/*
* A new import is in progress, add an entry.
*/
void
spa_import_progress_add(spa_t *spa)
{
spa_history_list_t *shl = spa_import_progress_list;
spa_import_progress_t *sip;
const char *poolname = NULL;
sip = kmem_zalloc(sizeof (spa_import_progress_t), KM_SLEEP);
sip->pool_guid = spa_guid(spa);
(void) nvlist_lookup_string(spa->spa_config, ZPOOL_CONFIG_POOL_NAME,
&poolname);
if (poolname == NULL)
poolname = spa_name(spa);
sip->pool_name = spa_strdup(poolname);
sip->spa_load_state = spa_load_state(spa);
mutex_enter(&shl->procfs_list.pl_lock);
procfs_list_add(&shl->procfs_list, sip);
shl->size++;
mutex_exit(&shl->procfs_list.pl_lock);
}
void
spa_import_progress_remove(uint64_t pool_guid)
{
spa_history_list_t *shl = spa_import_progress_list;
spa_import_progress_t *sip;
mutex_enter(&shl->procfs_list.pl_lock);
for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL;
sip = list_prev(&shl->procfs_list.pl_list, sip)) {
if (sip->pool_guid == pool_guid) {
if (sip->pool_name)
spa_strfree(sip->pool_name);
list_remove(&shl->procfs_list.pl_list, sip);
shl->size--;
kmem_free(sip, sizeof (spa_import_progress_t));
break;
}
}
mutex_exit(&shl->procfs_list.pl_lock);
}
/*
* ==========================================================================
* Initialization and Termination
* ==========================================================================
*/
static int
spa_name_compare(const void *a1, const void *a2)
{
const spa_t *s1 = a1;
const spa_t *s2 = a2;
int s;
s = strcmp(s1->spa_name, s2->spa_name);
return (TREE_ISIGN(s));
}
void
spa_boot_init(void)
{
spa_config_load();
}
void
spa_init(spa_mode_t mode)
{
mutex_init(&spa_namespace_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa_spare_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&spa_l2cache_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&spa_namespace_cv, NULL, CV_DEFAULT, NULL);
avl_create(&spa_namespace_avl, spa_name_compare, sizeof (spa_t),
offsetof(spa_t, spa_avl));
avl_create(&spa_spare_avl, spa_spare_compare, sizeof (spa_aux_t),
offsetof(spa_aux_t, aux_avl));
avl_create(&spa_l2cache_avl, spa_l2cache_compare, sizeof (spa_aux_t),
offsetof(spa_aux_t, aux_avl));
spa_mode_global = mode;
#ifndef _KERNEL
if (spa_mode_global != SPA_MODE_READ && dprintf_find_string("watch")) {
struct sigaction sa;
sa.sa_flags = SA_SIGINFO;
sigemptyset(&sa.sa_mask);
sa.sa_sigaction = arc_buf_sigsegv;
if (sigaction(SIGSEGV, &sa, NULL) == -1) {
perror("could not enable watchpoints: "
"sigaction(SIGSEGV, ...) = ");
} else {
arc_watch = B_TRUE;
}
}
#endif
fm_init();
zfs_refcount_init();
unique_init();
zfs_btree_init();
metaslab_stat_init();
brt_init();
ddt_init();
zio_init();
dmu_init();
zil_init();
vdev_cache_stat_init();
vdev_mirror_stat_init();
vdev_raidz_math_init();
vdev_file_init();
zfs_prop_init();
chksum_init();
zpool_prop_init();
zpool_feature_init();
spa_config_load();
vdev_prop_init();
l2arc_start();
scan_init();
qat_init();
spa_import_progress_init();
}
void
spa_fini(void)
{
l2arc_stop();
spa_evict_all();
vdev_file_fini();
vdev_cache_stat_fini();
vdev_mirror_stat_fini();
vdev_raidz_math_fini();
chksum_fini();
zil_fini();
dmu_fini();
zio_fini();
ddt_fini();
brt_fini();
metaslab_stat_fini();
zfs_btree_fini();
unique_fini();
zfs_refcount_fini();
fm_fini();
scan_fini();
qat_fini();
spa_import_progress_destroy();
avl_destroy(&spa_namespace_avl);
avl_destroy(&spa_spare_avl);
avl_destroy(&spa_l2cache_avl);
cv_destroy(&spa_namespace_cv);
mutex_destroy(&spa_namespace_lock);
mutex_destroy(&spa_spare_lock);
mutex_destroy(&spa_l2cache_lock);
}
/*
* Return whether this pool has a dedicated slog device. No locking needed.
* It's not a problem if the wrong answer is returned as it's only for
* performance and not correctness.
*/
boolean_t
spa_has_slogs(spa_t *spa)
{
return (spa->spa_log_class->mc_groups != 0);
}
spa_log_state_t
spa_get_log_state(spa_t *spa)
{
return (spa->spa_log_state);
}
void
spa_set_log_state(spa_t *spa, spa_log_state_t state)
{
spa->spa_log_state = state;
}
boolean_t
spa_is_root(spa_t *spa)
{
return (spa->spa_is_root);
}
boolean_t
spa_writeable(spa_t *spa)
{
return (!!(spa->spa_mode & SPA_MODE_WRITE) && spa->spa_trust_config);
}
/*
* Returns true if there is a pending sync task in any of the current
* syncing txg, the current quiescing txg, or the current open txg.
*/
boolean_t
spa_has_pending_synctask(spa_t *spa)
{
return (!txg_all_lists_empty(&spa->spa_dsl_pool->dp_sync_tasks) ||
!txg_all_lists_empty(&spa->spa_dsl_pool->dp_early_sync_tasks));
}
spa_mode_t
spa_mode(spa_t *spa)
{
return (spa->spa_mode);
}
uint64_t
spa_bootfs(spa_t *spa)
{
return (spa->spa_bootfs);
}
uint64_t
spa_delegation(spa_t *spa)
{
return (spa->spa_delegation);
}
objset_t *
spa_meta_objset(spa_t *spa)
{
return (spa->spa_meta_objset);
}
enum zio_checksum
spa_dedup_checksum(spa_t *spa)
{
return (spa->spa_dedup_checksum);
}
/*
* Reset pool scan stat per scan pass (or reboot).
*/
void
spa_scan_stat_init(spa_t *spa)
{
/* data not stored on disk */
spa->spa_scan_pass_start = gethrestime_sec();
if (dsl_scan_is_paused_scrub(spa->spa_dsl_pool->dp_scan))
spa->spa_scan_pass_scrub_pause = spa->spa_scan_pass_start;
else
spa->spa_scan_pass_scrub_pause = 0;
spa->spa_scan_pass_scrub_spent_paused = 0;
spa->spa_scan_pass_exam = 0;
spa->spa_scan_pass_issued = 0;
}
/*
* Get scan stats for zpool status reports
*/
int
spa_scan_get_stats(spa_t *spa, pool_scan_stat_t *ps)
{
dsl_scan_t *scn = spa->spa_dsl_pool ? spa->spa_dsl_pool->dp_scan : NULL;
if (scn == NULL || scn->scn_phys.scn_func == POOL_SCAN_NONE)
return (SET_ERROR(ENOENT));
memset(ps, 0, sizeof (pool_scan_stat_t));
/* data stored on disk */
ps->pss_func = scn->scn_phys.scn_func;
ps->pss_state = scn->scn_phys.scn_state;
ps->pss_start_time = scn->scn_phys.scn_start_time;
ps->pss_end_time = scn->scn_phys.scn_end_time;
ps->pss_to_examine = scn->scn_phys.scn_to_examine;
ps->pss_examined = scn->scn_phys.scn_examined;
ps->pss_to_process = scn->scn_phys.scn_to_process;
ps->pss_processed = scn->scn_phys.scn_processed;
ps->pss_errors = scn->scn_phys.scn_errors;
/* data not stored on disk */
ps->pss_pass_exam = spa->spa_scan_pass_exam;
ps->pss_pass_start = spa->spa_scan_pass_start;
ps->pss_pass_scrub_pause = spa->spa_scan_pass_scrub_pause;
ps->pss_pass_scrub_spent_paused = spa->spa_scan_pass_scrub_spent_paused;
ps->pss_pass_issued = spa->spa_scan_pass_issued;
ps->pss_issued =
scn->scn_issued_before_pass + spa->spa_scan_pass_issued;
return (0);
}
int
spa_maxblocksize(spa_t *spa)
{
if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_BLOCKS))
return (SPA_MAXBLOCKSIZE);
else
return (SPA_OLD_MAXBLOCKSIZE);
}
/*
* Returns the txg that the last device removal completed. No indirect mappings
* have been added since this txg.
*/
uint64_t
spa_get_last_removal_txg(spa_t *spa)
{
uint64_t vdevid;
uint64_t ret = -1ULL;
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
/*
* sr_prev_indirect_vdev is only modified while holding all the
* config locks, so it is sufficient to hold SCL_VDEV as reader when
* examining it.
*/
vdevid = spa->spa_removing_phys.sr_prev_indirect_vdev;
while (vdevid != -1ULL) {
vdev_t *vd = vdev_lookup_top(spa, vdevid);
vdev_indirect_births_t *vib = vd->vdev_indirect_births;
ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops);
/*
* If the removal did not remap any data, we don't care.
*/
if (vdev_indirect_births_count(vib) != 0) {
ret = vdev_indirect_births_last_entry_txg(vib);
break;
}
vdevid = vd->vdev_indirect_config.vic_prev_indirect_vdev;
}
spa_config_exit(spa, SCL_VDEV, FTAG);
IMPLY(ret != -1ULL,
spa_feature_is_active(spa, SPA_FEATURE_DEVICE_REMOVAL));
return (ret);
}
int
spa_maxdnodesize(spa_t *spa)
{
if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_DNODE))
return (DNODE_MAX_SIZE);
else
return (DNODE_MIN_SIZE);
}
boolean_t
spa_multihost(spa_t *spa)
{
return (spa->spa_multihost ? B_TRUE : B_FALSE);
}
uint32_t
spa_get_hostid(spa_t *spa)
{
return (spa->spa_hostid);
}
boolean_t
spa_trust_config(spa_t *spa)
{
return (spa->spa_trust_config);
}
uint64_t
spa_missing_tvds_allowed(spa_t *spa)
{
return (spa->spa_missing_tvds_allowed);
}
space_map_t *
spa_syncing_log_sm(spa_t *spa)
{
return (spa->spa_syncing_log_sm);
}
void
spa_set_missing_tvds(spa_t *spa, uint64_t missing)
{
spa->spa_missing_tvds = missing;
}
/*
* Return the pool state string ("ONLINE", "DEGRADED", "SUSPENDED", etc).
*/
const char *
spa_state_to_name(spa_t *spa)
{
ASSERT3P(spa, !=, NULL);
/*
* it is possible for the spa to exist, without root vdev
* as the spa transitions during import/export
*/
vdev_t *rvd = spa->spa_root_vdev;
if (rvd == NULL) {
return ("TRANSITIONING");
}
vdev_state_t state = rvd->vdev_state;
vdev_aux_t aux = rvd->vdev_stat.vs_aux;
if (spa_suspended(spa) &&
(spa_get_failmode(spa) != ZIO_FAILURE_MODE_CONTINUE))
return ("SUSPENDED");
switch (state) {
case VDEV_STATE_CLOSED:
case VDEV_STATE_OFFLINE:
return ("OFFLINE");
case VDEV_STATE_REMOVED:
return ("REMOVED");
case VDEV_STATE_CANT_OPEN:
if (aux == VDEV_AUX_CORRUPT_DATA || aux == VDEV_AUX_BAD_LOG)
return ("FAULTED");
else if (aux == VDEV_AUX_SPLIT_POOL)
return ("SPLIT");
else
return ("UNAVAIL");
case VDEV_STATE_FAULTED:
return ("FAULTED");
case VDEV_STATE_DEGRADED:
return ("DEGRADED");
case VDEV_STATE_HEALTHY:
return ("ONLINE");
default:
break;
}
return ("UNKNOWN");
}
boolean_t
spa_top_vdevs_spacemap_addressable(spa_t *spa)
{
vdev_t *rvd = spa->spa_root_vdev;
for (uint64_t c = 0; c < rvd->vdev_children; c++) {
if (!vdev_is_spacemap_addressable(rvd->vdev_child[c]))
return (B_FALSE);
}
return (B_TRUE);
}
boolean_t
spa_has_checkpoint(spa_t *spa)
{
return (spa->spa_checkpoint_txg != 0);
}
boolean_t
spa_importing_readonly_checkpoint(spa_t *spa)
{
return ((spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT) &&
spa->spa_mode == SPA_MODE_READ);
}
uint64_t
spa_min_claim_txg(spa_t *spa)
{
uint64_t checkpoint_txg = spa->spa_uberblock.ub_checkpoint_txg;
if (checkpoint_txg != 0)
return (checkpoint_txg + 1);
return (spa->spa_first_txg);
}
/*
* If there is a checkpoint, async destroys may consume more space from
* the pool instead of freeing it. In an attempt to save the pool from
* getting suspended when it is about to run out of space, we stop
* processing async destroys.
*/
boolean_t
spa_suspend_async_destroy(spa_t *spa)
{
dsl_pool_t *dp = spa_get_dsl(spa);
uint64_t unreserved = dsl_pool_unreserved_space(dp,
ZFS_SPACE_CHECK_EXTRA_RESERVED);
uint64_t used = dsl_dir_phys(dp->dp_root_dir)->dd_used_bytes;
uint64_t avail = (unreserved > used) ? (unreserved - used) : 0;
if (spa_has_checkpoint(spa) && avail == 0)
return (B_TRUE);
return (B_FALSE);
}
#if defined(_KERNEL)
int
param_set_deadman_failmode_common(const char *val)
{
spa_t *spa = NULL;
char *p;
if (val == NULL)
return (SET_ERROR(EINVAL));
if ((p = strchr(val, '\n')) != NULL)
*p = '\0';
if (strcmp(val, "wait") != 0 && strcmp(val, "continue") != 0 &&
strcmp(val, "panic"))
return (SET_ERROR(EINVAL));
if (spa_mode_global != SPA_MODE_UNINIT) {
mutex_enter(&spa_namespace_lock);
while ((spa = spa_next(spa)) != NULL)
spa_set_deadman_failmode(spa, val);
mutex_exit(&spa_namespace_lock);
}
return (0);
}
#endif
/* Namespace manipulation */
EXPORT_SYMBOL(spa_lookup);
EXPORT_SYMBOL(spa_add);
EXPORT_SYMBOL(spa_remove);
EXPORT_SYMBOL(spa_next);
/* Refcount functions */
EXPORT_SYMBOL(spa_open_ref);
EXPORT_SYMBOL(spa_close);
EXPORT_SYMBOL(spa_refcount_zero);
/* Pool configuration lock */
EXPORT_SYMBOL(spa_config_tryenter);
EXPORT_SYMBOL(spa_config_enter);
EXPORT_SYMBOL(spa_config_exit);
EXPORT_SYMBOL(spa_config_held);
/* Pool vdev add/remove lock */
EXPORT_SYMBOL(spa_vdev_enter);
EXPORT_SYMBOL(spa_vdev_exit);
/* Pool vdev state change lock */
EXPORT_SYMBOL(spa_vdev_state_enter);
EXPORT_SYMBOL(spa_vdev_state_exit);
/* Accessor functions */
EXPORT_SYMBOL(spa_shutting_down);
EXPORT_SYMBOL(spa_get_dsl);
EXPORT_SYMBOL(spa_get_rootblkptr);
EXPORT_SYMBOL(spa_set_rootblkptr);
EXPORT_SYMBOL(spa_altroot);
EXPORT_SYMBOL(spa_sync_pass);
EXPORT_SYMBOL(spa_name);
EXPORT_SYMBOL(spa_guid);
EXPORT_SYMBOL(spa_last_synced_txg);
EXPORT_SYMBOL(spa_first_txg);
EXPORT_SYMBOL(spa_syncing_txg);
EXPORT_SYMBOL(spa_version);
EXPORT_SYMBOL(spa_state);
EXPORT_SYMBOL(spa_load_state);
EXPORT_SYMBOL(spa_freeze_txg);
EXPORT_SYMBOL(spa_get_dspace);
EXPORT_SYMBOL(spa_update_dspace);
EXPORT_SYMBOL(spa_deflate);
EXPORT_SYMBOL(spa_normal_class);
EXPORT_SYMBOL(spa_log_class);
EXPORT_SYMBOL(spa_special_class);
EXPORT_SYMBOL(spa_preferred_class);
EXPORT_SYMBOL(spa_max_replication);
EXPORT_SYMBOL(spa_prev_software_version);
EXPORT_SYMBOL(spa_get_failmode);
EXPORT_SYMBOL(spa_suspended);
EXPORT_SYMBOL(spa_bootfs);
EXPORT_SYMBOL(spa_delegation);
EXPORT_SYMBOL(spa_meta_objset);
EXPORT_SYMBOL(spa_maxblocksize);
EXPORT_SYMBOL(spa_maxdnodesize);
/* Miscellaneous support routines */
EXPORT_SYMBOL(spa_guid_exists);
EXPORT_SYMBOL(spa_strdup);
EXPORT_SYMBOL(spa_strfree);
EXPORT_SYMBOL(spa_generate_guid);
EXPORT_SYMBOL(snprintf_blkptr);
EXPORT_SYMBOL(spa_freeze);
EXPORT_SYMBOL(spa_upgrade);
EXPORT_SYMBOL(spa_evict_all);
EXPORT_SYMBOL(spa_lookup_by_guid);
EXPORT_SYMBOL(spa_has_spare);
EXPORT_SYMBOL(dva_get_dsize_sync);
EXPORT_SYMBOL(bp_get_dsize_sync);
EXPORT_SYMBOL(bp_get_dsize);
EXPORT_SYMBOL(spa_has_slogs);
EXPORT_SYMBOL(spa_is_root);
EXPORT_SYMBOL(spa_writeable);
EXPORT_SYMBOL(spa_mode);
EXPORT_SYMBOL(spa_namespace_lock);
EXPORT_SYMBOL(spa_trust_config);
EXPORT_SYMBOL(spa_missing_tvds_allowed);
EXPORT_SYMBOL(spa_set_missing_tvds);
EXPORT_SYMBOL(spa_state_to_name);
EXPORT_SYMBOL(spa_importing_readonly_checkpoint);
EXPORT_SYMBOL(spa_min_claim_txg);
EXPORT_SYMBOL(spa_suspend_async_destroy);
EXPORT_SYMBOL(spa_has_checkpoint);
EXPORT_SYMBOL(spa_top_vdevs_spacemap_addressable);
ZFS_MODULE_PARAM(zfs, zfs_, flags, UINT, ZMOD_RW,
"Set additional debugging flags");
ZFS_MODULE_PARAM(zfs, zfs_, recover, INT, ZMOD_RW,
"Set to attempt to recover from fatal errors");
ZFS_MODULE_PARAM(zfs, zfs_, free_leak_on_eio, INT, ZMOD_RW,
"Set to ignore IO errors during free and permanently leak the space");
ZFS_MODULE_PARAM(zfs_deadman, zfs_deadman_, checktime_ms, U64, ZMOD_RW,
"Dead I/O check interval in milliseconds");
ZFS_MODULE_PARAM(zfs_deadman, zfs_deadman_, enabled, INT, ZMOD_RW,
"Enable deadman timer");
ZFS_MODULE_PARAM(zfs_spa, spa_, asize_inflation, UINT, ZMOD_RW,
"SPA size estimate multiplication factor");
ZFS_MODULE_PARAM(zfs, zfs_, ddt_data_is_special, INT, ZMOD_RW,
"Place DDT data into the special class");
ZFS_MODULE_PARAM(zfs, zfs_, user_indirect_is_special, INT, ZMOD_RW,
"Place user data indirect blocks into the special class");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, failmode,
param_set_deadman_failmode, param_get_charp, ZMOD_RW,
"Failmode for deadman timer");
ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, synctime_ms,
param_set_deadman_synctime, spl_param_get_u64, ZMOD_RW,
"Pool sync expiration time in milliseconds");
ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, ziotime_ms,
param_set_deadman_ziotime, spl_param_get_u64, ZMOD_RW,
"IO expiration time in milliseconds");
ZFS_MODULE_PARAM(zfs, zfs_, special_class_metadata_reserve_pct, UINT, ZMOD_RW,
"Small file blocks in special vdevs depends on this much "
"free space available");
/* END CSTYLED */
ZFS_MODULE_PARAM_CALL(zfs_spa, spa_, slop_shift, param_set_slop_shift,
param_get_uint, ZMOD_RW, "Reserved free space in pool");
diff --git a/sys/contrib/openzfs/module/zfs/vdev.c b/sys/contrib/openzfs/module/zfs/vdev.c
index ad932a7ba764..4bfd95861e02 100644
--- a/sys/contrib/openzfs/module/zfs/vdev.c
+++ b/sys/contrib/openzfs/module/zfs/vdev.c
@@ -1,6329 +1,6365 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2021 by Delphix. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc.
* Copyright (c) 2014 Integros [integros.com]
* Copyright 2016 Toomas Soome <tsoome@me.com>
* Copyright 2017 Joyent, Inc.
* Copyright (c) 2017, Intel Corporation.
* Copyright (c) 2019, Datto Inc. All rights reserved.
* Copyright (c) 2021, Klara Inc.
* Copyright [2021] Hewlett Packard Enterprise Development LP
*/
#include <sys/zfs_context.h>
#include <sys/fm/fs/zfs.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/bpobj.h>
#include <sys/dmu.h>
#include <sys/dmu_tx.h>
#include <sys/dsl_dir.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_rebuild.h>
#include <sys/vdev_draid.h>
#include <sys/uberblock_impl.h>
#include <sys/metaslab.h>
#include <sys/metaslab_impl.h>
#include <sys/space_map.h>
#include <sys/space_reftree.h>
#include <sys/zio.h>
#include <sys/zap.h>
#include <sys/fs/zfs.h>
#include <sys/arc.h>
#include <sys/zil.h>
#include <sys/dsl_scan.h>
#include <sys/vdev_raidz.h>
#include <sys/abd.h>
#include <sys/vdev_initialize.h>
#include <sys/vdev_trim.h>
#include <sys/zvol.h>
#include <sys/zfs_ratelimit.h>
#include "zfs_prop.h"
/*
* One metaslab from each (normal-class) vdev is used by the ZIL. These are
* called "embedded slog metaslabs", are referenced by vdev_log_mg, and are
* part of the spa_embedded_log_class. The metaslab with the most free space
* in each vdev is selected for this purpose when the pool is opened (or a
* vdev is added). See vdev_metaslab_init().
*
* Log blocks can be allocated from the following locations. Each one is tried
* in order until the allocation succeeds:
* 1. dedicated log vdevs, aka "slog" (spa_log_class)
* 2. embedded slog metaslabs (spa_embedded_log_class)
* 3. other metaslabs in normal vdevs (spa_normal_class)
*
* zfs_embedded_slog_min_ms disables the embedded slog if there are fewer
* than this number of metaslabs in the vdev. This ensures that we don't set
* aside an unreasonable amount of space for the ZIL. If set to less than
* 1 << (spa_slop_shift + 1), on small pools the usable space may be reduced
* (by more than 1<<spa_slop_shift) due to the embedded slog metaslab.
*/
static uint_t zfs_embedded_slog_min_ms = 64;
/* default target for number of metaslabs per top-level vdev */
static uint_t zfs_vdev_default_ms_count = 200;
/* minimum number of metaslabs per top-level vdev */
static uint_t zfs_vdev_min_ms_count = 16;
/* practical upper limit of total metaslabs per top-level vdev */
static uint_t zfs_vdev_ms_count_limit = 1ULL << 17;
/* lower limit for metaslab size (512M) */
static uint_t zfs_vdev_default_ms_shift = 29;
/* upper limit for metaslab size (16G) */
-static const uint_t zfs_vdev_max_ms_shift = 34;
+static uint_t zfs_vdev_max_ms_shift = 34;
int vdev_validate_skip = B_FALSE;
/*
* Since the DTL space map of a vdev is not expected to have a lot of
* entries, we default its block size to 4K.
*/
int zfs_vdev_dtl_sm_blksz = (1 << 12);
/*
* Rate limit slow IO (delay) events to this many per second.
*/
static unsigned int zfs_slow_io_events_per_second = 20;
/*
* Rate limit checksum events after this many checksum errors per second.
*/
static unsigned int zfs_checksum_events_per_second = 20;
/*
* Ignore errors during scrub/resilver. Allows to work around resilver
* upon import when there are pool errors.
*/
static int zfs_scan_ignore_errors = 0;
/*
* vdev-wide space maps that have lots of entries written to them at
* the end of each transaction can benefit from a higher I/O bandwidth
* (e.g. vdev_obsolete_sm), thus we default their block size to 128K.
*/
int zfs_vdev_standard_sm_blksz = (1 << 17);
/*
* Tunable parameter for debugging or performance analysis. Setting this
* will cause pool corruption on power loss if a volatile out-of-order
* write cache is enabled.
*/
int zfs_nocacheflush = 0;
/*
* Maximum and minimum ashift values that can be automatically set based on
* vdev's physical ashift (disk's physical sector size). While ASHIFT_MAX
* is higher than the maximum value, it is intentionally limited here to not
* excessively impact pool space efficiency. Higher ashift values may still
* be forced by vdev logical ashift or by user via ashift property, but won't
* be set automatically as a performance optimization.
*/
uint_t zfs_vdev_max_auto_ashift = 14;
uint_t zfs_vdev_min_auto_ashift = ASHIFT_MIN;
void
vdev_dbgmsg(vdev_t *vd, const char *fmt, ...)
{
va_list adx;
char buf[256];
va_start(adx, fmt);
(void) vsnprintf(buf, sizeof (buf), fmt, adx);
va_end(adx);
if (vd->vdev_path != NULL) {
zfs_dbgmsg("%s vdev '%s': %s", vd->vdev_ops->vdev_op_type,
vd->vdev_path, buf);
} else {
zfs_dbgmsg("%s-%llu vdev (guid %llu): %s",
vd->vdev_ops->vdev_op_type,
(u_longlong_t)vd->vdev_id,
(u_longlong_t)vd->vdev_guid, buf);
}
}
void
vdev_dbgmsg_print_tree(vdev_t *vd, int indent)
{
char state[20];
if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops) {
zfs_dbgmsg("%*svdev %llu: %s", indent, "",
(u_longlong_t)vd->vdev_id,
vd->vdev_ops->vdev_op_type);
return;
}
switch (vd->vdev_state) {
case VDEV_STATE_UNKNOWN:
(void) snprintf(state, sizeof (state), "unknown");
break;
case VDEV_STATE_CLOSED:
(void) snprintf(state, sizeof (state), "closed");
break;
case VDEV_STATE_OFFLINE:
(void) snprintf(state, sizeof (state), "offline");
break;
case VDEV_STATE_REMOVED:
(void) snprintf(state, sizeof (state), "removed");
break;
case VDEV_STATE_CANT_OPEN:
(void) snprintf(state, sizeof (state), "can't open");
break;
case VDEV_STATE_FAULTED:
(void) snprintf(state, sizeof (state), "faulted");
break;
case VDEV_STATE_DEGRADED:
(void) snprintf(state, sizeof (state), "degraded");
break;
case VDEV_STATE_HEALTHY:
(void) snprintf(state, sizeof (state), "healthy");
break;
default:
(void) snprintf(state, sizeof (state), "<state %u>",
(uint_t)vd->vdev_state);
}
zfs_dbgmsg("%*svdev %u: %s%s, guid: %llu, path: %s, %s", indent,
"", (int)vd->vdev_id, vd->vdev_ops->vdev_op_type,
vd->vdev_islog ? " (log)" : "",
(u_longlong_t)vd->vdev_guid,
vd->vdev_path ? vd->vdev_path : "N/A", state);
for (uint64_t i = 0; i < vd->vdev_children; i++)
vdev_dbgmsg_print_tree(vd->vdev_child[i], indent + 2);
}
/*
* Virtual device management.
*/
static vdev_ops_t *const vdev_ops_table[] = {
&vdev_root_ops,
&vdev_raidz_ops,
&vdev_draid_ops,
&vdev_draid_spare_ops,
&vdev_mirror_ops,
&vdev_replacing_ops,
&vdev_spare_ops,
&vdev_disk_ops,
&vdev_file_ops,
&vdev_missing_ops,
&vdev_hole_ops,
&vdev_indirect_ops,
NULL
};
/*
* Given a vdev type, return the appropriate ops vector.
*/
static vdev_ops_t *
vdev_getops(const char *type)
{
vdev_ops_t *ops, *const *opspp;
for (opspp = vdev_ops_table; (ops = *opspp) != NULL; opspp++)
if (strcmp(ops->vdev_op_type, type) == 0)
break;
return (ops);
}
/*
* Given a vdev and a metaslab class, find which metaslab group we're
* interested in. All vdevs may belong to two different metaslab classes.
* Dedicated slog devices use only the primary metaslab group, rather than a
* separate log group. For embedded slogs, the vdev_log_mg will be non-NULL.
*/
metaslab_group_t *
vdev_get_mg(vdev_t *vd, metaslab_class_t *mc)
{
if (mc == spa_embedded_log_class(vd->vdev_spa) &&
vd->vdev_log_mg != NULL)
return (vd->vdev_log_mg);
else
return (vd->vdev_mg);
}
void
vdev_default_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs)
{
(void) vd, (void) remain_rs;
physical_rs->rs_start = logical_rs->rs_start;
physical_rs->rs_end = logical_rs->rs_end;
}
/*
* Derive the enumerated allocation bias from string input.
* String origin is either the per-vdev zap or zpool(8).
*/
static vdev_alloc_bias_t
vdev_derive_alloc_bias(const char *bias)
{
vdev_alloc_bias_t alloc_bias = VDEV_BIAS_NONE;
if (strcmp(bias, VDEV_ALLOC_BIAS_LOG) == 0)
alloc_bias = VDEV_BIAS_LOG;
else if (strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0)
alloc_bias = VDEV_BIAS_SPECIAL;
else if (strcmp(bias, VDEV_ALLOC_BIAS_DEDUP) == 0)
alloc_bias = VDEV_BIAS_DEDUP;
return (alloc_bias);
}
/*
* Default asize function: return the MAX of psize with the asize of
* all children. This is what's used by anything other than RAID-Z.
*/
uint64_t
vdev_default_asize(vdev_t *vd, uint64_t psize)
{
uint64_t asize = P2ROUNDUP(psize, 1ULL << vd->vdev_top->vdev_ashift);
uint64_t csize;
for (int c = 0; c < vd->vdev_children; c++) {
csize = vdev_psize_to_asize(vd->vdev_child[c], psize);
asize = MAX(asize, csize);
}
return (asize);
}
uint64_t
vdev_default_min_asize(vdev_t *vd)
{
return (vd->vdev_min_asize);
}
/*
* Get the minimum allocatable size. We define the allocatable size as
* the vdev's asize rounded to the nearest metaslab. This allows us to
* replace or attach devices which don't have the same physical size but
* can still satisfy the same number of allocations.
*/
uint64_t
vdev_get_min_asize(vdev_t *vd)
{
vdev_t *pvd = vd->vdev_parent;
/*
* If our parent is NULL (inactive spare or cache) or is the root,
* just return our own asize.
*/
if (pvd == NULL)
return (vd->vdev_asize);
/*
* The top-level vdev just returns the allocatable size rounded
* to the nearest metaslab.
*/
if (vd == vd->vdev_top)
return (P2ALIGN(vd->vdev_asize, 1ULL << vd->vdev_ms_shift));
return (pvd->vdev_ops->vdev_op_min_asize(pvd));
}
void
vdev_set_min_asize(vdev_t *vd)
{
vd->vdev_min_asize = vdev_get_min_asize(vd);
for (int c = 0; c < vd->vdev_children; c++)
vdev_set_min_asize(vd->vdev_child[c]);
}
/*
* Get the minimal allocation size for the top-level vdev.
*/
uint64_t
vdev_get_min_alloc(vdev_t *vd)
{
uint64_t min_alloc = 1ULL << vd->vdev_ashift;
if (vd->vdev_ops->vdev_op_min_alloc != NULL)
min_alloc = vd->vdev_ops->vdev_op_min_alloc(vd);
return (min_alloc);
}
/*
* Get the parity level for a top-level vdev.
*/
uint64_t
vdev_get_nparity(vdev_t *vd)
{
uint64_t nparity = 0;
if (vd->vdev_ops->vdev_op_nparity != NULL)
nparity = vd->vdev_ops->vdev_op_nparity(vd);
return (nparity);
}
static int
vdev_prop_get_int(vdev_t *vd, vdev_prop_t prop, uint64_t *value)
{
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
uint64_t objid;
int err;
- if (vd->vdev_top_zap != 0) {
+ if (vd->vdev_root_zap != 0) {
+ objid = vd->vdev_root_zap;
+ } else if (vd->vdev_top_zap != 0) {
objid = vd->vdev_top_zap;
} else if (vd->vdev_leaf_zap != 0) {
objid = vd->vdev_leaf_zap;
} else {
return (EINVAL);
}
err = zap_lookup(mos, objid, vdev_prop_to_name(prop),
sizeof (uint64_t), 1, value);
if (err == ENOENT)
*value = vdev_prop_default_numeric(prop);
return (err);
}
/*
* Get the number of data disks for a top-level vdev.
*/
uint64_t
vdev_get_ndisks(vdev_t *vd)
{
uint64_t ndisks = 1;
if (vd->vdev_ops->vdev_op_ndisks != NULL)
ndisks = vd->vdev_ops->vdev_op_ndisks(vd);
return (ndisks);
}
vdev_t *
vdev_lookup_top(spa_t *spa, uint64_t vdev)
{
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
if (vdev < rvd->vdev_children) {
ASSERT(rvd->vdev_child[vdev] != NULL);
return (rvd->vdev_child[vdev]);
}
return (NULL);
}
vdev_t *
vdev_lookup_by_guid(vdev_t *vd, uint64_t guid)
{
vdev_t *mvd;
if (vd->vdev_guid == guid)
return (vd);
for (int c = 0; c < vd->vdev_children; c++)
if ((mvd = vdev_lookup_by_guid(vd->vdev_child[c], guid)) !=
NULL)
return (mvd);
return (NULL);
}
static int
vdev_count_leaves_impl(vdev_t *vd)
{
int n = 0;
if (vd->vdev_ops->vdev_op_leaf)
return (1);
for (int c = 0; c < vd->vdev_children; c++)
n += vdev_count_leaves_impl(vd->vdev_child[c]);
return (n);
}
int
vdev_count_leaves(spa_t *spa)
{
int rc;
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
rc = vdev_count_leaves_impl(spa->spa_root_vdev);
spa_config_exit(spa, SCL_VDEV, FTAG);
return (rc);
}
void
vdev_add_child(vdev_t *pvd, vdev_t *cvd)
{
size_t oldsize, newsize;
uint64_t id = cvd->vdev_id;
vdev_t **newchild;
ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
ASSERT(cvd->vdev_parent == NULL);
cvd->vdev_parent = pvd;
if (pvd == NULL)
return;
ASSERT(id >= pvd->vdev_children || pvd->vdev_child[id] == NULL);
oldsize = pvd->vdev_children * sizeof (vdev_t *);
pvd->vdev_children = MAX(pvd->vdev_children, id + 1);
newsize = pvd->vdev_children * sizeof (vdev_t *);
newchild = kmem_alloc(newsize, KM_SLEEP);
if (pvd->vdev_child != NULL) {
memcpy(newchild, pvd->vdev_child, oldsize);
kmem_free(pvd->vdev_child, oldsize);
}
pvd->vdev_child = newchild;
pvd->vdev_child[id] = cvd;
cvd->vdev_top = (pvd->vdev_top ? pvd->vdev_top: cvd);
ASSERT(cvd->vdev_top->vdev_parent->vdev_parent == NULL);
/*
* Walk up all ancestors to update guid sum.
*/
for (; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += cvd->vdev_guid_sum;
if (cvd->vdev_ops->vdev_op_leaf) {
list_insert_head(&cvd->vdev_spa->spa_leaf_list, cvd);
cvd->vdev_spa->spa_leaf_list_gen++;
}
}
void
vdev_remove_child(vdev_t *pvd, vdev_t *cvd)
{
int c;
uint_t id = cvd->vdev_id;
ASSERT(cvd->vdev_parent == pvd);
if (pvd == NULL)
return;
ASSERT(id < pvd->vdev_children);
ASSERT(pvd->vdev_child[id] == cvd);
pvd->vdev_child[id] = NULL;
cvd->vdev_parent = NULL;
for (c = 0; c < pvd->vdev_children; c++)
if (pvd->vdev_child[c])
break;
if (c == pvd->vdev_children) {
kmem_free(pvd->vdev_child, c * sizeof (vdev_t *));
pvd->vdev_child = NULL;
pvd->vdev_children = 0;
}
if (cvd->vdev_ops->vdev_op_leaf) {
spa_t *spa = cvd->vdev_spa;
list_remove(&spa->spa_leaf_list, cvd);
spa->spa_leaf_list_gen++;
}
/*
* Walk up all ancestors to update guid sum.
*/
for (; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum -= cvd->vdev_guid_sum;
}
/*
* Remove any holes in the child array.
*/
void
vdev_compact_children(vdev_t *pvd)
{
vdev_t **newchild, *cvd;
int oldc = pvd->vdev_children;
int newc;
ASSERT(spa_config_held(pvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
if (oldc == 0)
return;
for (int c = newc = 0; c < oldc; c++)
if (pvd->vdev_child[c])
newc++;
if (newc > 0) {
newchild = kmem_zalloc(newc * sizeof (vdev_t *), KM_SLEEP);
for (int c = newc = 0; c < oldc; c++) {
if ((cvd = pvd->vdev_child[c]) != NULL) {
newchild[newc] = cvd;
cvd->vdev_id = newc++;
}
}
} else {
newchild = NULL;
}
kmem_free(pvd->vdev_child, oldc * sizeof (vdev_t *));
pvd->vdev_child = newchild;
pvd->vdev_children = newc;
}
/*
* Allocate and minimally initialize a vdev_t.
*/
vdev_t *
vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid, vdev_ops_t *ops)
{
vdev_t *vd;
vdev_indirect_config_t *vic;
vd = kmem_zalloc(sizeof (vdev_t), KM_SLEEP);
vic = &vd->vdev_indirect_config;
if (spa->spa_root_vdev == NULL) {
ASSERT(ops == &vdev_root_ops);
spa->spa_root_vdev = vd;
spa->spa_load_guid = spa_generate_guid(NULL);
}
if (guid == 0 && ops != &vdev_hole_ops) {
if (spa->spa_root_vdev == vd) {
/*
* The root vdev's guid will also be the pool guid,
* which must be unique among all pools.
*/
guid = spa_generate_guid(NULL);
} else {
/*
* Any other vdev's guid must be unique within the pool.
*/
guid = spa_generate_guid(spa);
}
ASSERT(!spa_guid_exists(spa_guid(spa), guid));
}
vd->vdev_spa = spa;
vd->vdev_id = id;
vd->vdev_guid = guid;
vd->vdev_guid_sum = guid;
vd->vdev_ops = ops;
vd->vdev_state = VDEV_STATE_CLOSED;
vd->vdev_ishole = (ops == &vdev_hole_ops);
vic->vic_prev_indirect_vdev = UINT64_MAX;
rw_init(&vd->vdev_indirect_rwlock, NULL, RW_DEFAULT, NULL);
mutex_init(&vd->vdev_obsolete_lock, NULL, MUTEX_DEFAULT, NULL);
vd->vdev_obsolete_segments = range_tree_create(NULL, RANGE_SEG64, NULL,
0, 0);
/*
* Initialize rate limit structs for events. We rate limit ZIO delay
* and checksum events so that we don't overwhelm ZED with thousands
* of events when a disk is acting up.
*/
zfs_ratelimit_init(&vd->vdev_delay_rl, &zfs_slow_io_events_per_second,
1);
zfs_ratelimit_init(&vd->vdev_deadman_rl, &zfs_slow_io_events_per_second,
1);
zfs_ratelimit_init(&vd->vdev_checksum_rl,
&zfs_checksum_events_per_second, 1);
/*
* Default Thresholds for tuning ZED
*/
vd->vdev_checksum_n = vdev_prop_default_numeric(VDEV_PROP_CHECKSUM_N);
vd->vdev_checksum_t = vdev_prop_default_numeric(VDEV_PROP_CHECKSUM_T);
vd->vdev_io_n = vdev_prop_default_numeric(VDEV_PROP_IO_N);
vd->vdev_io_t = vdev_prop_default_numeric(VDEV_PROP_IO_T);
list_link_init(&vd->vdev_config_dirty_node);
list_link_init(&vd->vdev_state_dirty_node);
list_link_init(&vd->vdev_initialize_node);
list_link_init(&vd->vdev_leaf_node);
list_link_init(&vd->vdev_trim_node);
mutex_init(&vd->vdev_dtl_lock, NULL, MUTEX_NOLOCKDEP, NULL);
mutex_init(&vd->vdev_stat_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_probe_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_scan_io_queue_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_initialize_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_initialize_io_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&vd->vdev_initialize_cv, NULL, CV_DEFAULT, NULL);
cv_init(&vd->vdev_initialize_io_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&vd->vdev_trim_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_autotrim_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&vd->vdev_trim_io_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&vd->vdev_trim_cv, NULL, CV_DEFAULT, NULL);
cv_init(&vd->vdev_autotrim_cv, NULL, CV_DEFAULT, NULL);
cv_init(&vd->vdev_autotrim_kick_cv, NULL, CV_DEFAULT, NULL);
cv_init(&vd->vdev_trim_io_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&vd->vdev_rebuild_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&vd->vdev_rebuild_cv, NULL, CV_DEFAULT, NULL);
for (int t = 0; t < DTL_TYPES; t++) {
vd->vdev_dtl[t] = range_tree_create(NULL, RANGE_SEG64, NULL, 0,
0);
}
txg_list_create(&vd->vdev_ms_list, spa,
offsetof(struct metaslab, ms_txg_node));
txg_list_create(&vd->vdev_dtl_list, spa,
offsetof(struct vdev, vdev_dtl_node));
vd->vdev_stat.vs_timestamp = gethrtime();
vdev_queue_init(vd);
vdev_cache_init(vd);
return (vd);
}
/*
* Allocate a new vdev. The 'alloctype' is used to control whether we are
* creating a new vdev or loading an existing one - the behavior is slightly
* different for each case.
*/
int
vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent, uint_t id,
int alloctype)
{
vdev_ops_t *ops;
const char *type;
uint64_t guid = 0, islog;
vdev_t *vd;
vdev_indirect_config_t *vic;
const char *tmp = NULL;
int rc;
vdev_alloc_bias_t alloc_bias = VDEV_BIAS_NONE;
boolean_t top_level = (parent && !parent->vdev_parent);
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) != 0)
return (SET_ERROR(EINVAL));
if ((ops = vdev_getops(type)) == NULL)
return (SET_ERROR(EINVAL));
/*
* If this is a load, get the vdev guid from the nvlist.
* Otherwise, vdev_alloc_common() will generate one for us.
*/
if (alloctype == VDEV_ALLOC_LOAD) {
uint64_t label_id;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID, &label_id) ||
label_id != id)
return (SET_ERROR(EINVAL));
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (SET_ERROR(EINVAL));
} else if (alloctype == VDEV_ALLOC_SPARE) {
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (SET_ERROR(EINVAL));
} else if (alloctype == VDEV_ALLOC_L2CACHE) {
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (SET_ERROR(EINVAL));
} else if (alloctype == VDEV_ALLOC_ROOTPOOL) {
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0)
return (SET_ERROR(EINVAL));
}
/*
* The first allocated vdev must be of type 'root'.
*/
if (ops != &vdev_root_ops && spa->spa_root_vdev == NULL)
return (SET_ERROR(EINVAL));
/*
* Determine whether we're a log vdev.
*/
islog = 0;
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG, &islog);
if (islog && spa_version(spa) < SPA_VERSION_SLOGS)
return (SET_ERROR(ENOTSUP));
if (ops == &vdev_hole_ops && spa_version(spa) < SPA_VERSION_HOLES)
return (SET_ERROR(ENOTSUP));
if (top_level && alloctype == VDEV_ALLOC_ADD) {
const char *bias;
/*
* If creating a top-level vdev, check for allocation
* classes input.
*/
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_ALLOCATION_BIAS,
&bias) == 0) {
alloc_bias = vdev_derive_alloc_bias(bias);
/* spa_vdev_add() expects feature to be enabled */
if (spa->spa_load_state != SPA_LOAD_CREATE &&
!spa_feature_is_enabled(spa,
SPA_FEATURE_ALLOCATION_CLASSES)) {
return (SET_ERROR(ENOTSUP));
}
}
/* spa_vdev_add() expects feature to be enabled */
if (ops == &vdev_draid_ops &&
spa->spa_load_state != SPA_LOAD_CREATE &&
!spa_feature_is_enabled(spa, SPA_FEATURE_DRAID)) {
return (SET_ERROR(ENOTSUP));
}
}
/*
* Initialize the vdev specific data. This is done before calling
* vdev_alloc_common() since it may fail and this simplifies the
* error reporting and cleanup code paths.
*/
void *tsd = NULL;
if (ops->vdev_op_init != NULL) {
rc = ops->vdev_op_init(spa, nv, &tsd);
if (rc != 0) {
return (rc);
}
}
vd = vdev_alloc_common(spa, id, guid, ops);
vd->vdev_tsd = tsd;
vd->vdev_islog = islog;
if (top_level && alloc_bias != VDEV_BIAS_NONE)
vd->vdev_alloc_bias = alloc_bias;
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &tmp) == 0)
vd->vdev_path = spa_strdup(tmp);
/*
* ZPOOL_CONFIG_AUX_STATE = "external" means we previously forced a
* fault on a vdev and want it to persist across imports (like with
* zpool offline -f).
*/
rc = nvlist_lookup_string(nv, ZPOOL_CONFIG_AUX_STATE, &tmp);
if (rc == 0 && tmp != NULL && strcmp(tmp, "external") == 0) {
vd->vdev_stat.vs_aux = VDEV_AUX_EXTERNAL;
vd->vdev_faulted = 1;
vd->vdev_label_aux = VDEV_AUX_EXTERNAL;
}
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &tmp) == 0)
vd->vdev_devid = spa_strdup(tmp);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PHYS_PATH, &tmp) == 0)
vd->vdev_physpath = spa_strdup(tmp);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
&tmp) == 0)
vd->vdev_enc_sysfs_path = spa_strdup(tmp);
if (nvlist_lookup_string(nv, ZPOOL_CONFIG_FRU, &tmp) == 0)
vd->vdev_fru = spa_strdup(tmp);
/*
* Set the whole_disk property. If it's not specified, leave the value
* as -1.
*/
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
&vd->vdev_wholedisk) != 0)
vd->vdev_wholedisk = -1ULL;
vic = &vd->vdev_indirect_config;
ASSERT0(vic->vic_mapping_object);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_INDIRECT_OBJECT,
&vic->vic_mapping_object);
ASSERT0(vic->vic_births_object);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_INDIRECT_BIRTHS,
&vic->vic_births_object);
ASSERT3U(vic->vic_prev_indirect_vdev, ==, UINT64_MAX);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_PREV_INDIRECT_VDEV,
&vic->vic_prev_indirect_vdev);
/*
* Look for the 'not present' flag. This will only be set if the device
* was not present at the time of import.
*/
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT,
&vd->vdev_not_present);
/*
* Get the alignment requirement.
*/
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASHIFT, &vd->vdev_ashift);
/*
* Retrieve the vdev creation time.
*/
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_CREATE_TXG,
&vd->vdev_crtxg);
+ if (vd->vdev_ops == &vdev_root_ops &&
+ (alloctype == VDEV_ALLOC_LOAD ||
+ alloctype == VDEV_ALLOC_SPLIT ||
+ alloctype == VDEV_ALLOC_ROOTPOOL)) {
+ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_VDEV_ROOT_ZAP,
+ &vd->vdev_root_zap);
+ }
+
/*
* If we're a top-level vdev, try to load the allocation parameters.
*/
if (top_level &&
(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT)) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_ARRAY,
&vd->vdev_ms_array);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_SHIFT,
&vd->vdev_ms_shift);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASIZE,
&vd->vdev_asize);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NONALLOCATING,
&vd->vdev_noalloc);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVING,
&vd->vdev_removing);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_VDEV_TOP_ZAP,
&vd->vdev_top_zap);
} else {
ASSERT0(vd->vdev_top_zap);
}
if (top_level && alloctype != VDEV_ALLOC_ATTACH) {
ASSERT(alloctype == VDEV_ALLOC_LOAD ||
alloctype == VDEV_ALLOC_ADD ||
alloctype == VDEV_ALLOC_SPLIT ||
alloctype == VDEV_ALLOC_ROOTPOOL);
/* Note: metaslab_group_create() is now deferred */
}
if (vd->vdev_ops->vdev_op_leaf &&
(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT)) {
(void) nvlist_lookup_uint64(nv,
ZPOOL_CONFIG_VDEV_LEAF_ZAP, &vd->vdev_leaf_zap);
} else {
ASSERT0(vd->vdev_leaf_zap);
}
/*
* If we're a leaf vdev, try to load the DTL object and other state.
*/
if (vd->vdev_ops->vdev_op_leaf &&
(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_L2CACHE ||
alloctype == VDEV_ALLOC_ROOTPOOL)) {
if (alloctype == VDEV_ALLOC_LOAD) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DTL,
&vd->vdev_dtl_object);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_UNSPARE,
&vd->vdev_unspare);
}
if (alloctype == VDEV_ALLOC_ROOTPOOL) {
uint64_t spare = 0;
if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_SPARE,
&spare) == 0 && spare)
spa_spare_add(vd);
}
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_OFFLINE,
&vd->vdev_offline);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_RESILVER_TXG,
&vd->vdev_resilver_txg);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REBUILD_TXG,
&vd->vdev_rebuild_txg);
if (nvlist_exists(nv, ZPOOL_CONFIG_RESILVER_DEFER))
vdev_defer_resilver(vd);
/*
* In general, when importing a pool we want to ignore the
* persistent fault state, as the diagnosis made on another
* system may not be valid in the current context. The only
* exception is if we forced a vdev to a persistently faulted
* state with 'zpool offline -f'. The persistent fault will
* remain across imports until cleared.
*
* Local vdevs will remain in the faulted state.
*/
if (spa_load_state(spa) == SPA_LOAD_OPEN ||
spa_load_state(spa) == SPA_LOAD_IMPORT) {
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_FAULTED,
&vd->vdev_faulted);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DEGRADED,
&vd->vdev_degraded);
(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVED,
&vd->vdev_removed);
if (vd->vdev_faulted || vd->vdev_degraded) {
const char *aux;
vd->vdev_label_aux =
VDEV_AUX_ERR_EXCEEDED;
if (nvlist_lookup_string(nv,
ZPOOL_CONFIG_AUX_STATE, &aux) == 0 &&
strcmp(aux, "external") == 0)
vd->vdev_label_aux = VDEV_AUX_EXTERNAL;
else
vd->vdev_faulted = 0ULL;
}
}
}
/*
* Add ourselves to the parent's list of children.
*/
vdev_add_child(parent, vd);
*vdp = vd;
return (0);
}
void
vdev_free(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT3P(vd->vdev_initialize_thread, ==, NULL);
ASSERT3P(vd->vdev_trim_thread, ==, NULL);
ASSERT3P(vd->vdev_autotrim_thread, ==, NULL);
ASSERT3P(vd->vdev_rebuild_thread, ==, NULL);
/*
* Scan queues are normally destroyed at the end of a scan. If the
* queue exists here, that implies the vdev is being removed while
* the scan is still running.
*/
if (vd->vdev_scan_io_queue != NULL) {
mutex_enter(&vd->vdev_scan_io_queue_lock);
dsl_scan_io_queue_destroy(vd->vdev_scan_io_queue);
vd->vdev_scan_io_queue = NULL;
mutex_exit(&vd->vdev_scan_io_queue_lock);
}
/*
* vdev_free() implies closing the vdev first. This is simpler than
* trying to ensure complicated semantics for all callers.
*/
vdev_close(vd);
ASSERT(!list_link_active(&vd->vdev_config_dirty_node));
ASSERT(!list_link_active(&vd->vdev_state_dirty_node));
/*
* Free all children.
*/
for (int c = 0; c < vd->vdev_children; c++)
vdev_free(vd->vdev_child[c]);
ASSERT(vd->vdev_child == NULL);
ASSERT(vd->vdev_guid_sum == vd->vdev_guid);
if (vd->vdev_ops->vdev_op_fini != NULL)
vd->vdev_ops->vdev_op_fini(vd);
/*
* Discard allocation state.
*/
if (vd->vdev_mg != NULL) {
vdev_metaslab_fini(vd);
metaslab_group_destroy(vd->vdev_mg);
vd->vdev_mg = NULL;
}
if (vd->vdev_log_mg != NULL) {
ASSERT0(vd->vdev_ms_count);
metaslab_group_destroy(vd->vdev_log_mg);
vd->vdev_log_mg = NULL;
}
ASSERT0(vd->vdev_stat.vs_space);
ASSERT0(vd->vdev_stat.vs_dspace);
ASSERT0(vd->vdev_stat.vs_alloc);
/*
* Remove this vdev from its parent's child list.
*/
vdev_remove_child(vd->vdev_parent, vd);
ASSERT(vd->vdev_parent == NULL);
ASSERT(!list_link_active(&vd->vdev_leaf_node));
/*
* Clean up vdev structure.
*/
vdev_queue_fini(vd);
vdev_cache_fini(vd);
if (vd->vdev_path)
spa_strfree(vd->vdev_path);
if (vd->vdev_devid)
spa_strfree(vd->vdev_devid);
if (vd->vdev_physpath)
spa_strfree(vd->vdev_physpath);
if (vd->vdev_enc_sysfs_path)
spa_strfree(vd->vdev_enc_sysfs_path);
if (vd->vdev_fru)
spa_strfree(vd->vdev_fru);
if (vd->vdev_isspare)
spa_spare_remove(vd);
if (vd->vdev_isl2cache)
spa_l2cache_remove(vd);
txg_list_destroy(&vd->vdev_ms_list);
txg_list_destroy(&vd->vdev_dtl_list);
mutex_enter(&vd->vdev_dtl_lock);
space_map_close(vd->vdev_dtl_sm);
for (int t = 0; t < DTL_TYPES; t++) {
range_tree_vacate(vd->vdev_dtl[t], NULL, NULL);
range_tree_destroy(vd->vdev_dtl[t]);
}
mutex_exit(&vd->vdev_dtl_lock);
EQUIV(vd->vdev_indirect_births != NULL,
vd->vdev_indirect_mapping != NULL);
if (vd->vdev_indirect_births != NULL) {
vdev_indirect_mapping_close(vd->vdev_indirect_mapping);
vdev_indirect_births_close(vd->vdev_indirect_births);
}
if (vd->vdev_obsolete_sm != NULL) {
ASSERT(vd->vdev_removing ||
vd->vdev_ops == &vdev_indirect_ops);
space_map_close(vd->vdev_obsolete_sm);
vd->vdev_obsolete_sm = NULL;
}
range_tree_destroy(vd->vdev_obsolete_segments);
rw_destroy(&vd->vdev_indirect_rwlock);
mutex_destroy(&vd->vdev_obsolete_lock);
mutex_destroy(&vd->vdev_dtl_lock);
mutex_destroy(&vd->vdev_stat_lock);
mutex_destroy(&vd->vdev_probe_lock);
mutex_destroy(&vd->vdev_scan_io_queue_lock);
mutex_destroy(&vd->vdev_initialize_lock);
mutex_destroy(&vd->vdev_initialize_io_lock);
cv_destroy(&vd->vdev_initialize_io_cv);
cv_destroy(&vd->vdev_initialize_cv);
mutex_destroy(&vd->vdev_trim_lock);
mutex_destroy(&vd->vdev_autotrim_lock);
mutex_destroy(&vd->vdev_trim_io_lock);
cv_destroy(&vd->vdev_trim_cv);
cv_destroy(&vd->vdev_autotrim_cv);
cv_destroy(&vd->vdev_autotrim_kick_cv);
cv_destroy(&vd->vdev_trim_io_cv);
mutex_destroy(&vd->vdev_rebuild_lock);
cv_destroy(&vd->vdev_rebuild_cv);
zfs_ratelimit_fini(&vd->vdev_delay_rl);
zfs_ratelimit_fini(&vd->vdev_deadman_rl);
zfs_ratelimit_fini(&vd->vdev_checksum_rl);
if (vd == spa->spa_root_vdev)
spa->spa_root_vdev = NULL;
kmem_free(vd, sizeof (vdev_t));
}
/*
* Transfer top-level vdev state from svd to tvd.
*/
static void
vdev_top_transfer(vdev_t *svd, vdev_t *tvd)
{
spa_t *spa = svd->vdev_spa;
metaslab_t *msp;
vdev_t *vd;
int t;
ASSERT(tvd == tvd->vdev_top);
tvd->vdev_pending_fastwrite = svd->vdev_pending_fastwrite;
tvd->vdev_ms_array = svd->vdev_ms_array;
tvd->vdev_ms_shift = svd->vdev_ms_shift;
tvd->vdev_ms_count = svd->vdev_ms_count;
tvd->vdev_top_zap = svd->vdev_top_zap;
svd->vdev_ms_array = 0;
svd->vdev_ms_shift = 0;
svd->vdev_ms_count = 0;
svd->vdev_top_zap = 0;
if (tvd->vdev_mg)
ASSERT3P(tvd->vdev_mg, ==, svd->vdev_mg);
if (tvd->vdev_log_mg)
ASSERT3P(tvd->vdev_log_mg, ==, svd->vdev_log_mg);
tvd->vdev_mg = svd->vdev_mg;
tvd->vdev_log_mg = svd->vdev_log_mg;
tvd->vdev_ms = svd->vdev_ms;
svd->vdev_mg = NULL;
svd->vdev_log_mg = NULL;
svd->vdev_ms = NULL;
if (tvd->vdev_mg != NULL)
tvd->vdev_mg->mg_vd = tvd;
if (tvd->vdev_log_mg != NULL)
tvd->vdev_log_mg->mg_vd = tvd;
tvd->vdev_checkpoint_sm = svd->vdev_checkpoint_sm;
svd->vdev_checkpoint_sm = NULL;
tvd->vdev_alloc_bias = svd->vdev_alloc_bias;
svd->vdev_alloc_bias = VDEV_BIAS_NONE;
tvd->vdev_stat.vs_alloc = svd->vdev_stat.vs_alloc;
tvd->vdev_stat.vs_space = svd->vdev_stat.vs_space;
tvd->vdev_stat.vs_dspace = svd->vdev_stat.vs_dspace;
svd->vdev_stat.vs_alloc = 0;
svd->vdev_stat.vs_space = 0;
svd->vdev_stat.vs_dspace = 0;
/*
* State which may be set on a top-level vdev that's in the
* process of being removed.
*/
ASSERT0(tvd->vdev_indirect_config.vic_births_object);
ASSERT0(tvd->vdev_indirect_config.vic_mapping_object);
ASSERT3U(tvd->vdev_indirect_config.vic_prev_indirect_vdev, ==, -1ULL);
ASSERT3P(tvd->vdev_indirect_mapping, ==, NULL);
ASSERT3P(tvd->vdev_indirect_births, ==, NULL);
ASSERT3P(tvd->vdev_obsolete_sm, ==, NULL);
ASSERT0(tvd->vdev_noalloc);
ASSERT0(tvd->vdev_removing);
ASSERT0(tvd->vdev_rebuilding);
tvd->vdev_noalloc = svd->vdev_noalloc;
tvd->vdev_removing = svd->vdev_removing;
tvd->vdev_rebuilding = svd->vdev_rebuilding;
tvd->vdev_rebuild_config = svd->vdev_rebuild_config;
tvd->vdev_indirect_config = svd->vdev_indirect_config;
tvd->vdev_indirect_mapping = svd->vdev_indirect_mapping;
tvd->vdev_indirect_births = svd->vdev_indirect_births;
range_tree_swap(&svd->vdev_obsolete_segments,
&tvd->vdev_obsolete_segments);
tvd->vdev_obsolete_sm = svd->vdev_obsolete_sm;
svd->vdev_indirect_config.vic_mapping_object = 0;
svd->vdev_indirect_config.vic_births_object = 0;
svd->vdev_indirect_config.vic_prev_indirect_vdev = -1ULL;
svd->vdev_indirect_mapping = NULL;
svd->vdev_indirect_births = NULL;
svd->vdev_obsolete_sm = NULL;
svd->vdev_noalloc = 0;
svd->vdev_removing = 0;
svd->vdev_rebuilding = 0;
for (t = 0; t < TXG_SIZE; t++) {
while ((msp = txg_list_remove(&svd->vdev_ms_list, t)) != NULL)
(void) txg_list_add(&tvd->vdev_ms_list, msp, t);
while ((vd = txg_list_remove(&svd->vdev_dtl_list, t)) != NULL)
(void) txg_list_add(&tvd->vdev_dtl_list, vd, t);
if (txg_list_remove_this(&spa->spa_vdev_txg_list, svd, t))
(void) txg_list_add(&spa->spa_vdev_txg_list, tvd, t);
}
if (list_link_active(&svd->vdev_config_dirty_node)) {
vdev_config_clean(svd);
vdev_config_dirty(tvd);
}
if (list_link_active(&svd->vdev_state_dirty_node)) {
vdev_state_clean(svd);
vdev_state_dirty(tvd);
}
tvd->vdev_deflate_ratio = svd->vdev_deflate_ratio;
svd->vdev_deflate_ratio = 0;
tvd->vdev_islog = svd->vdev_islog;
svd->vdev_islog = 0;
dsl_scan_io_queue_vdev_xfer(svd, tvd);
}
static void
vdev_top_update(vdev_t *tvd, vdev_t *vd)
{
if (vd == NULL)
return;
vd->vdev_top = tvd;
for (int c = 0; c < vd->vdev_children; c++)
vdev_top_update(tvd, vd->vdev_child[c]);
}
/*
* Add a mirror/replacing vdev above an existing vdev. There is no need to
* call .vdev_op_init() since mirror/replacing vdevs do not have private state.
*/
vdev_t *
vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops)
{
spa_t *spa = cvd->vdev_spa;
vdev_t *pvd = cvd->vdev_parent;
vdev_t *mvd;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
mvd = vdev_alloc_common(spa, cvd->vdev_id, 0, ops);
mvd->vdev_asize = cvd->vdev_asize;
mvd->vdev_min_asize = cvd->vdev_min_asize;
mvd->vdev_max_asize = cvd->vdev_max_asize;
mvd->vdev_psize = cvd->vdev_psize;
mvd->vdev_ashift = cvd->vdev_ashift;
mvd->vdev_logical_ashift = cvd->vdev_logical_ashift;
mvd->vdev_physical_ashift = cvd->vdev_physical_ashift;
mvd->vdev_state = cvd->vdev_state;
mvd->vdev_crtxg = cvd->vdev_crtxg;
vdev_remove_child(pvd, cvd);
vdev_add_child(pvd, mvd);
cvd->vdev_id = mvd->vdev_children;
vdev_add_child(mvd, cvd);
vdev_top_update(cvd->vdev_top, cvd->vdev_top);
if (mvd == mvd->vdev_top)
vdev_top_transfer(cvd, mvd);
return (mvd);
}
/*
* Remove a 1-way mirror/replacing vdev from the tree.
*/
void
vdev_remove_parent(vdev_t *cvd)
{
vdev_t *mvd = cvd->vdev_parent;
vdev_t *pvd = mvd->vdev_parent;
ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
ASSERT(mvd->vdev_children == 1);
ASSERT(mvd->vdev_ops == &vdev_mirror_ops ||
mvd->vdev_ops == &vdev_replacing_ops ||
mvd->vdev_ops == &vdev_spare_ops);
cvd->vdev_ashift = mvd->vdev_ashift;
cvd->vdev_logical_ashift = mvd->vdev_logical_ashift;
cvd->vdev_physical_ashift = mvd->vdev_physical_ashift;
vdev_remove_child(mvd, cvd);
vdev_remove_child(pvd, mvd);
/*
* If cvd will replace mvd as a top-level vdev, preserve mvd's guid.
* Otherwise, we could have detached an offline device, and when we
* go to import the pool we'll think we have two top-level vdevs,
* instead of a different version of the same top-level vdev.
*/
if (mvd->vdev_top == mvd) {
uint64_t guid_delta = mvd->vdev_guid - cvd->vdev_guid;
cvd->vdev_orig_guid = cvd->vdev_guid;
cvd->vdev_guid += guid_delta;
cvd->vdev_guid_sum += guid_delta;
/*
* If pool not set for autoexpand, we need to also preserve
* mvd's asize to prevent automatic expansion of cvd.
* Otherwise if we are adjusting the mirror by attaching and
* detaching children of non-uniform sizes, the mirror could
* autoexpand, unexpectedly requiring larger devices to
* re-establish the mirror.
*/
if (!cvd->vdev_spa->spa_autoexpand)
cvd->vdev_asize = mvd->vdev_asize;
}
cvd->vdev_id = mvd->vdev_id;
vdev_add_child(pvd, cvd);
vdev_top_update(cvd->vdev_top, cvd->vdev_top);
if (cvd == cvd->vdev_top)
vdev_top_transfer(mvd, cvd);
ASSERT(mvd->vdev_children == 0);
vdev_free(mvd);
}
void
vdev_metaslab_group_create(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
/*
* metaslab_group_create was delayed until allocation bias was available
*/
if (vd->vdev_mg == NULL) {
metaslab_class_t *mc;
if (vd->vdev_islog && vd->vdev_alloc_bias == VDEV_BIAS_NONE)
vd->vdev_alloc_bias = VDEV_BIAS_LOG;
ASSERT3U(vd->vdev_islog, ==,
(vd->vdev_alloc_bias == VDEV_BIAS_LOG));
switch (vd->vdev_alloc_bias) {
case VDEV_BIAS_LOG:
mc = spa_log_class(spa);
break;
case VDEV_BIAS_SPECIAL:
mc = spa_special_class(spa);
break;
case VDEV_BIAS_DEDUP:
mc = spa_dedup_class(spa);
break;
default:
mc = spa_normal_class(spa);
}
vd->vdev_mg = metaslab_group_create(mc, vd,
spa->spa_alloc_count);
if (!vd->vdev_islog) {
vd->vdev_log_mg = metaslab_group_create(
spa_embedded_log_class(spa), vd, 1);
}
/*
* The spa ashift min/max only apply for the normal metaslab
* class. Class destination is late binding so ashift boundary
* setting had to wait until now.
*/
if (vd->vdev_top == vd && vd->vdev_ashift != 0 &&
mc == spa_normal_class(spa) && vd->vdev_aux == NULL) {
if (vd->vdev_ashift > spa->spa_max_ashift)
spa->spa_max_ashift = vd->vdev_ashift;
if (vd->vdev_ashift < spa->spa_min_ashift)
spa->spa_min_ashift = vd->vdev_ashift;
uint64_t min_alloc = vdev_get_min_alloc(vd);
if (min_alloc < spa->spa_min_alloc)
spa->spa_min_alloc = min_alloc;
}
}
}
int
vdev_metaslab_init(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
uint64_t oldc = vd->vdev_ms_count;
uint64_t newc = vd->vdev_asize >> vd->vdev_ms_shift;
metaslab_t **mspp;
int error;
boolean_t expanding = (oldc != 0);
ASSERT(txg == 0 || spa_config_held(spa, SCL_ALLOC, RW_WRITER));
/*
* This vdev is not being allocated from yet or is a hole.
*/
if (vd->vdev_ms_shift == 0)
return (0);
ASSERT(!vd->vdev_ishole);
ASSERT(oldc <= newc);
mspp = vmem_zalloc(newc * sizeof (*mspp), KM_SLEEP);
if (expanding) {
memcpy(mspp, vd->vdev_ms, oldc * sizeof (*mspp));
vmem_free(vd->vdev_ms, oldc * sizeof (*mspp));
}
vd->vdev_ms = mspp;
vd->vdev_ms_count = newc;
for (uint64_t m = oldc; m < newc; m++) {
uint64_t object = 0;
/*
* vdev_ms_array may be 0 if we are creating the "fake"
* metaslabs for an indirect vdev for zdb's leak detection.
* See zdb_leak_init().
*/
if (txg == 0 && vd->vdev_ms_array != 0) {
error = dmu_read(spa->spa_meta_objset,
vd->vdev_ms_array,
m * sizeof (uint64_t), sizeof (uint64_t), &object,
DMU_READ_PREFETCH);
if (error != 0) {
vdev_dbgmsg(vd, "unable to read the metaslab "
"array [error=%d]", error);
return (error);
}
}
error = metaslab_init(vd->vdev_mg, m, object, txg,
&(vd->vdev_ms[m]));
if (error != 0) {
vdev_dbgmsg(vd, "metaslab_init failed [error=%d]",
error);
return (error);
}
}
/*
* Find the emptiest metaslab on the vdev and mark it for use for
* embedded slog by moving it from the regular to the log metaslab
* group.
*/
if (vd->vdev_mg->mg_class == spa_normal_class(spa) &&
vd->vdev_ms_count > zfs_embedded_slog_min_ms &&
avl_is_empty(&vd->vdev_log_mg->mg_metaslab_tree)) {
uint64_t slog_msid = 0;
uint64_t smallest = UINT64_MAX;
/*
* Note, we only search the new metaslabs, because the old
* (pre-existing) ones may be active (e.g. have non-empty
* range_tree's), and we don't move them to the new
* metaslab_t.
*/
for (uint64_t m = oldc; m < newc; m++) {
uint64_t alloc =
space_map_allocated(vd->vdev_ms[m]->ms_sm);
if (alloc < smallest) {
slog_msid = m;
smallest = alloc;
}
}
metaslab_t *slog_ms = vd->vdev_ms[slog_msid];
/*
* The metaslab was marked as dirty at the end of
* metaslab_init(). Remove it from the dirty list so that we
* can uninitialize and reinitialize it to the new class.
*/
if (txg != 0) {
(void) txg_list_remove_this(&vd->vdev_ms_list,
slog_ms, txg);
}
uint64_t sm_obj = space_map_object(slog_ms->ms_sm);
metaslab_fini(slog_ms);
VERIFY0(metaslab_init(vd->vdev_log_mg, slog_msid, sm_obj, txg,
&vd->vdev_ms[slog_msid]));
}
if (txg == 0)
spa_config_enter(spa, SCL_ALLOC, FTAG, RW_WRITER);
/*
* If the vdev is marked as non-allocating then don't
* activate the metaslabs since we want to ensure that
* no allocations are performed on this device.
*/
if (vd->vdev_noalloc) {
/* track non-allocating vdev space */
spa->spa_nonallocating_dspace += spa_deflate(spa) ?
vd->vdev_stat.vs_dspace : vd->vdev_stat.vs_space;
} else if (!expanding) {
metaslab_group_activate(vd->vdev_mg);
if (vd->vdev_log_mg != NULL)
metaslab_group_activate(vd->vdev_log_mg);
}
if (txg == 0)
spa_config_exit(spa, SCL_ALLOC, FTAG);
return (0);
}
void
vdev_metaslab_fini(vdev_t *vd)
{
if (vd->vdev_checkpoint_sm != NULL) {
ASSERT(spa_feature_is_active(vd->vdev_spa,
SPA_FEATURE_POOL_CHECKPOINT));
space_map_close(vd->vdev_checkpoint_sm);
/*
* Even though we close the space map, we need to set its
* pointer to NULL. The reason is that vdev_metaslab_fini()
* may be called multiple times for certain operations
* (i.e. when destroying a pool) so we need to ensure that
* this clause never executes twice. This logic is similar
* to the one used for the vdev_ms clause below.
*/
vd->vdev_checkpoint_sm = NULL;
}
if (vd->vdev_ms != NULL) {
metaslab_group_t *mg = vd->vdev_mg;
metaslab_group_passivate(mg);
if (vd->vdev_log_mg != NULL) {
ASSERT(!vd->vdev_islog);
metaslab_group_passivate(vd->vdev_log_mg);
}
uint64_t count = vd->vdev_ms_count;
for (uint64_t m = 0; m < count; m++) {
metaslab_t *msp = vd->vdev_ms[m];
if (msp != NULL)
metaslab_fini(msp);
}
vmem_free(vd->vdev_ms, count * sizeof (metaslab_t *));
vd->vdev_ms = NULL;
vd->vdev_ms_count = 0;
for (int i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) {
ASSERT0(mg->mg_histogram[i]);
if (vd->vdev_log_mg != NULL)
ASSERT0(vd->vdev_log_mg->mg_histogram[i]);
}
}
ASSERT0(vd->vdev_ms_count);
ASSERT3U(vd->vdev_pending_fastwrite, ==, 0);
}
typedef struct vdev_probe_stats {
boolean_t vps_readable;
boolean_t vps_writeable;
int vps_flags;
} vdev_probe_stats_t;
static void
vdev_probe_done(zio_t *zio)
{
spa_t *spa = zio->io_spa;
vdev_t *vd = zio->io_vd;
vdev_probe_stats_t *vps = zio->io_private;
ASSERT(vd->vdev_probe_zio != NULL);
if (zio->io_type == ZIO_TYPE_READ) {
if (zio->io_error == 0)
vps->vps_readable = 1;
if (zio->io_error == 0 && spa_writeable(spa)) {
zio_nowait(zio_write_phys(vd->vdev_probe_zio, vd,
zio->io_offset, zio->io_size, zio->io_abd,
ZIO_CHECKSUM_OFF, vdev_probe_done, vps,
ZIO_PRIORITY_SYNC_WRITE, vps->vps_flags, B_TRUE));
} else {
abd_free(zio->io_abd);
}
} else if (zio->io_type == ZIO_TYPE_WRITE) {
if (zio->io_error == 0)
vps->vps_writeable = 1;
abd_free(zio->io_abd);
} else if (zio->io_type == ZIO_TYPE_NULL) {
zio_t *pio;
zio_link_t *zl;
vd->vdev_cant_read |= !vps->vps_readable;
vd->vdev_cant_write |= !vps->vps_writeable;
if (vdev_readable(vd) &&
(vdev_writeable(vd) || !spa_writeable(spa))) {
zio->io_error = 0;
} else {
ASSERT(zio->io_error != 0);
vdev_dbgmsg(vd, "failed probe");
(void) zfs_ereport_post(FM_EREPORT_ZFS_PROBE_FAILURE,
spa, vd, NULL, NULL, 0);
zio->io_error = SET_ERROR(ENXIO);
}
mutex_enter(&vd->vdev_probe_lock);
ASSERT(vd->vdev_probe_zio == zio);
vd->vdev_probe_zio = NULL;
mutex_exit(&vd->vdev_probe_lock);
zl = NULL;
while ((pio = zio_walk_parents(zio, &zl)) != NULL)
if (!vdev_accessible(vd, pio))
pio->io_error = SET_ERROR(ENXIO);
kmem_free(vps, sizeof (*vps));
}
}
/*
* Determine whether this device is accessible.
*
* Read and write to several known locations: the pad regions of each
* vdev label but the first, which we leave alone in case it contains
* a VTOC.
*/
zio_t *
vdev_probe(vdev_t *vd, zio_t *zio)
{
spa_t *spa = vd->vdev_spa;
vdev_probe_stats_t *vps = NULL;
zio_t *pio;
ASSERT(vd->vdev_ops->vdev_op_leaf);
/*
* Don't probe the probe.
*/
if (zio && (zio->io_flags & ZIO_FLAG_PROBE))
return (NULL);
/*
* To prevent 'probe storms' when a device fails, we create
* just one probe i/o at a time. All zios that want to probe
* this vdev will become parents of the probe io.
*/
mutex_enter(&vd->vdev_probe_lock);
if ((pio = vd->vdev_probe_zio) == NULL) {
vps = kmem_zalloc(sizeof (*vps), KM_SLEEP);
vps->vps_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_PROBE |
ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_AGGREGATE |
ZIO_FLAG_TRYHARD;
if (spa_config_held(spa, SCL_ZIO, RW_WRITER)) {
/*
* vdev_cant_read and vdev_cant_write can only
* transition from TRUE to FALSE when we have the
* SCL_ZIO lock as writer; otherwise they can only
* transition from FALSE to TRUE. This ensures that
* any zio looking at these values can assume that
* failures persist for the life of the I/O. That's
* important because when a device has intermittent
* connectivity problems, we want to ensure that
* they're ascribed to the device (ENXIO) and not
* the zio (EIO).
*
* Since we hold SCL_ZIO as writer here, clear both
* values so the probe can reevaluate from first
* principles.
*/
vps->vps_flags |= ZIO_FLAG_CONFIG_WRITER;
vd->vdev_cant_read = B_FALSE;
vd->vdev_cant_write = B_FALSE;
}
vd->vdev_probe_zio = pio = zio_null(NULL, spa, vd,
vdev_probe_done, vps,
vps->vps_flags | ZIO_FLAG_DONT_PROPAGATE);
/*
* We can't change the vdev state in this context, so we
* kick off an async task to do it on our behalf.
*/
if (zio != NULL) {
vd->vdev_probe_wanted = B_TRUE;
spa_async_request(spa, SPA_ASYNC_PROBE);
}
}
if (zio != NULL)
zio_add_child(zio, pio);
mutex_exit(&vd->vdev_probe_lock);
if (vps == NULL) {
ASSERT(zio != NULL);
return (NULL);
}
for (int l = 1; l < VDEV_LABELS; l++) {
zio_nowait(zio_read_phys(pio, vd,
vdev_label_offset(vd->vdev_psize, l,
offsetof(vdev_label_t, vl_be)), VDEV_PAD_SIZE,
abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE),
ZIO_CHECKSUM_OFF, vdev_probe_done, vps,
ZIO_PRIORITY_SYNC_READ, vps->vps_flags, B_TRUE));
}
if (zio == NULL)
return (pio);
zio_nowait(pio);
return (NULL);
}
static void
vdev_load_child(void *arg)
{
vdev_t *vd = arg;
vd->vdev_load_error = vdev_load(vd);
}
static void
vdev_open_child(void *arg)
{
vdev_t *vd = arg;
vd->vdev_open_thread = curthread;
vd->vdev_open_error = vdev_open(vd);
vd->vdev_open_thread = NULL;
}
static boolean_t
vdev_uses_zvols(vdev_t *vd)
{
#ifdef _KERNEL
if (zvol_is_zvol(vd->vdev_path))
return (B_TRUE);
#endif
for (int c = 0; c < vd->vdev_children; c++)
if (vdev_uses_zvols(vd->vdev_child[c]))
return (B_TRUE);
return (B_FALSE);
}
/*
* Returns B_TRUE if the passed child should be opened.
*/
static boolean_t
vdev_default_open_children_func(vdev_t *vd)
{
(void) vd;
return (B_TRUE);
}
/*
* Open the requested child vdevs. If any of the leaf vdevs are using
* a ZFS volume then do the opens in a single thread. This avoids a
* deadlock when the current thread is holding the spa_namespace_lock.
*/
static void
vdev_open_children_impl(vdev_t *vd, vdev_open_children_func_t *open_func)
{
int children = vd->vdev_children;
taskq_t *tq = taskq_create("vdev_open", children, minclsyspri,
children, children, TASKQ_PREPOPULATE);
vd->vdev_nonrot = B_TRUE;
for (int c = 0; c < children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (open_func(cvd) == B_FALSE)
continue;
if (tq == NULL || vdev_uses_zvols(vd)) {
cvd->vdev_open_error = vdev_open(cvd);
} else {
VERIFY(taskq_dispatch(tq, vdev_open_child,
cvd, TQ_SLEEP) != TASKQID_INVALID);
}
vd->vdev_nonrot &= cvd->vdev_nonrot;
}
if (tq != NULL) {
taskq_wait(tq);
taskq_destroy(tq);
}
}
/*
* Open all child vdevs.
*/
void
vdev_open_children(vdev_t *vd)
{
vdev_open_children_impl(vd, vdev_default_open_children_func);
}
/*
* Conditionally open a subset of child vdevs.
*/
void
vdev_open_children_subset(vdev_t *vd, vdev_open_children_func_t *open_func)
{
vdev_open_children_impl(vd, open_func);
}
/*
* Compute the raidz-deflation ratio. Note, we hard-code
* in 128k (1 << 17) because it is the "typical" blocksize.
* Even though SPA_MAXBLOCKSIZE changed, this algorithm can not change,
* otherwise it would inconsistently account for existing bp's.
*/
static void
vdev_set_deflate_ratio(vdev_t *vd)
{
if (vd == vd->vdev_top && !vd->vdev_ishole && vd->vdev_ashift != 0) {
vd->vdev_deflate_ratio = (1 << 17) /
(vdev_psize_to_asize(vd, 1 << 17) >> SPA_MINBLOCKSHIFT);
}
}
/*
* Choose the best of two ashifts, preferring one between logical ashift
* (absolute minimum) and administrator defined maximum, otherwise take
* the biggest of the two.
*/
uint64_t
vdev_best_ashift(uint64_t logical, uint64_t a, uint64_t b)
{
if (a > logical && a <= zfs_vdev_max_auto_ashift) {
if (b <= logical || b > zfs_vdev_max_auto_ashift)
return (a);
else
return (MAX(a, b));
} else if (b <= logical || b > zfs_vdev_max_auto_ashift)
return (MAX(a, b));
return (b);
}
/*
* Maximize performance by inflating the configured ashift for top level
* vdevs to be as close to the physical ashift as possible while maintaining
* administrator defined limits and ensuring it doesn't go below the
* logical ashift.
*/
static void
vdev_ashift_optimize(vdev_t *vd)
{
ASSERT(vd == vd->vdev_top);
if (vd->vdev_ashift < vd->vdev_physical_ashift &&
vd->vdev_physical_ashift <= zfs_vdev_max_auto_ashift) {
vd->vdev_ashift = MIN(
MAX(zfs_vdev_max_auto_ashift, vd->vdev_ashift),
MAX(zfs_vdev_min_auto_ashift,
vd->vdev_physical_ashift));
} else {
/*
* If the logical and physical ashifts are the same, then
* we ensure that the top-level vdev's ashift is not smaller
* than our minimum ashift value. For the unusual case
* where logical ashift > physical ashift, we can't cap
* the calculated ashift based on max ashift as that
* would cause failures.
* We still check if we need to increase it to match
* the min ashift.
*/
vd->vdev_ashift = MAX(zfs_vdev_min_auto_ashift,
vd->vdev_ashift);
}
}
/*
* Prepare a virtual device for access.
*/
int
vdev_open(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
int error;
uint64_t osize = 0;
uint64_t max_osize = 0;
uint64_t asize, max_asize, psize;
uint64_t logical_ashift = 0;
uint64_t physical_ashift = 0;
ASSERT(vd->vdev_open_thread == curthread ||
spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
ASSERT(vd->vdev_state == VDEV_STATE_CLOSED ||
vd->vdev_state == VDEV_STATE_CANT_OPEN ||
vd->vdev_state == VDEV_STATE_OFFLINE);
vd->vdev_stat.vs_aux = VDEV_AUX_NONE;
vd->vdev_cant_read = B_FALSE;
vd->vdev_cant_write = B_FALSE;
vd->vdev_min_asize = vdev_get_min_asize(vd);
/*
* If this vdev is not removed, check its fault status. If it's
* faulted, bail out of the open.
*/
if (!vd->vdev_removed && vd->vdev_faulted) {
ASSERT(vd->vdev_children == 0);
ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED ||
vd->vdev_label_aux == VDEV_AUX_EXTERNAL);
vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED,
vd->vdev_label_aux);
return (SET_ERROR(ENXIO));
} else if (vd->vdev_offline) {
ASSERT(vd->vdev_children == 0);
vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE, VDEV_AUX_NONE);
return (SET_ERROR(ENXIO));
}
error = vd->vdev_ops->vdev_op_open(vd, &osize, &max_osize,
&logical_ashift, &physical_ashift);
/* Keep the device in removed state if unplugged */
if (error == ENOENT && vd->vdev_removed) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_REMOVED,
VDEV_AUX_NONE);
return (error);
}
/*
* Physical volume size should never be larger than its max size, unless
* the disk has shrunk while we were reading it or the device is buggy
* or damaged: either way it's not safe for use, bail out of the open.
*/
if (osize > max_osize) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_OPEN_FAILED);
return (SET_ERROR(ENXIO));
}
/*
* Reset the vdev_reopening flag so that we actually close
* the vdev on error.
*/
vd->vdev_reopening = B_FALSE;
if (zio_injection_enabled && error == 0)
error = zio_handle_device_injection(vd, NULL, SET_ERROR(ENXIO));
if (error) {
if (vd->vdev_removed &&
vd->vdev_stat.vs_aux != VDEV_AUX_OPEN_FAILED)
vd->vdev_removed = B_FALSE;
if (vd->vdev_stat.vs_aux == VDEV_AUX_CHILDREN_OFFLINE) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE,
vd->vdev_stat.vs_aux);
} else {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
vd->vdev_stat.vs_aux);
}
return (error);
}
vd->vdev_removed = B_FALSE;
/*
* Recheck the faulted flag now that we have confirmed that
* the vdev is accessible. If we're faulted, bail.
*/
if (vd->vdev_faulted) {
ASSERT(vd->vdev_children == 0);
ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED ||
vd->vdev_label_aux == VDEV_AUX_EXTERNAL);
vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED,
vd->vdev_label_aux);
return (SET_ERROR(ENXIO));
}
if (vd->vdev_degraded) {
ASSERT(vd->vdev_children == 0);
vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED,
VDEV_AUX_ERR_EXCEEDED);
} else {
vdev_set_state(vd, B_TRUE, VDEV_STATE_HEALTHY, 0);
}
/*
* For hole or missing vdevs we just return success.
*/
if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops)
return (0);
for (int c = 0; c < vd->vdev_children; c++) {
if (vd->vdev_child[c]->vdev_state != VDEV_STATE_HEALTHY) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED,
VDEV_AUX_NONE);
break;
}
}
osize = P2ALIGN(osize, (uint64_t)sizeof (vdev_label_t));
max_osize = P2ALIGN(max_osize, (uint64_t)sizeof (vdev_label_t));
if (vd->vdev_children == 0) {
if (osize < SPA_MINDEVSIZE) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_TOO_SMALL);
return (SET_ERROR(EOVERFLOW));
}
psize = osize;
asize = osize - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE);
max_asize = max_osize - (VDEV_LABEL_START_SIZE +
VDEV_LABEL_END_SIZE);
} else {
if (vd->vdev_parent != NULL && osize < SPA_MINDEVSIZE -
(VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE)) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_TOO_SMALL);
return (SET_ERROR(EOVERFLOW));
}
psize = 0;
asize = osize;
max_asize = max_osize;
}
/*
* If the vdev was expanded, record this so that we can re-create the
* uberblock rings in labels {2,3}, during the next sync.
*/
if ((psize > vd->vdev_psize) && (vd->vdev_psize != 0))
vd->vdev_copy_uberblocks = B_TRUE;
vd->vdev_psize = psize;
/*
* Make sure the allocatable size hasn't shrunk too much.
*/
if (asize < vd->vdev_min_asize) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_LABEL);
return (SET_ERROR(EINVAL));
}
/*
* We can always set the logical/physical ashift members since
* their values are only used to calculate the vdev_ashift when
* the device is first added to the config. These values should
* not be used for anything else since they may change whenever
* the device is reopened and we don't store them in the label.
*/
vd->vdev_physical_ashift =
MAX(physical_ashift, vd->vdev_physical_ashift);
vd->vdev_logical_ashift = MAX(logical_ashift,
vd->vdev_logical_ashift);
if (vd->vdev_asize == 0) {
/*
* This is the first-ever open, so use the computed values.
* For compatibility, a different ashift can be requested.
*/
vd->vdev_asize = asize;
vd->vdev_max_asize = max_asize;
/*
* If the vdev_ashift was not overridden at creation time,
* then set it the logical ashift and optimize the ashift.
*/
if (vd->vdev_ashift == 0) {
vd->vdev_ashift = vd->vdev_logical_ashift;
if (vd->vdev_logical_ashift > ASHIFT_MAX) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_ASHIFT_TOO_BIG);
return (SET_ERROR(EDOM));
}
if (vd->vdev_top == vd) {
vdev_ashift_optimize(vd);
}
}
if (vd->vdev_ashift != 0 && (vd->vdev_ashift < ASHIFT_MIN ||
vd->vdev_ashift > ASHIFT_MAX)) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_ASHIFT);
return (SET_ERROR(EDOM));
}
} else {
/*
* Make sure the alignment required hasn't increased.
*/
if (vd->vdev_ashift > vd->vdev_top->vdev_ashift &&
vd->vdev_ops->vdev_op_leaf) {
(void) zfs_ereport_post(
FM_EREPORT_ZFS_DEVICE_BAD_ASHIFT,
spa, vd, NULL, NULL, 0);
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_LABEL);
return (SET_ERROR(EDOM));
}
vd->vdev_max_asize = max_asize;
}
/*
* If all children are healthy we update asize if either:
* The asize has increased, due to a device expansion caused by dynamic
* LUN growth or vdev replacement, and automatic expansion is enabled;
* making the additional space available.
*
* The asize has decreased, due to a device shrink usually caused by a
* vdev replace with a smaller device. This ensures that calculations
* based of max_asize and asize e.g. esize are always valid. It's safe
* to do this as we've already validated that asize is greater than
* vdev_min_asize.
*/
if (vd->vdev_state == VDEV_STATE_HEALTHY &&
((asize > vd->vdev_asize &&
(vd->vdev_expanding || spa->spa_autoexpand)) ||
(asize < vd->vdev_asize)))
vd->vdev_asize = asize;
vdev_set_min_asize(vd);
/*
* Ensure we can issue some IO before declaring the
* vdev open for business.
*/
if (vd->vdev_ops->vdev_op_leaf &&
(error = zio_wait(vdev_probe(vd, NULL))) != 0) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED,
VDEV_AUX_ERR_EXCEEDED);
return (error);
}
/*
* Track the minimum allocation size.
*/
if (vd->vdev_top == vd && vd->vdev_ashift != 0 &&
vd->vdev_islog == 0 && vd->vdev_aux == NULL) {
uint64_t min_alloc = vdev_get_min_alloc(vd);
if (min_alloc < spa->spa_min_alloc)
spa->spa_min_alloc = min_alloc;
}
/*
* If this is a leaf vdev, assess whether a resilver is needed.
* But don't do this if we are doing a reopen for a scrub, since
* this would just restart the scrub we are already doing.
*/
if (vd->vdev_ops->vdev_op_leaf && !spa->spa_scrub_reopen)
dsl_scan_assess_vdev(spa->spa_dsl_pool, vd);
return (0);
}
static void
vdev_validate_child(void *arg)
{
vdev_t *vd = arg;
vd->vdev_validate_thread = curthread;
vd->vdev_validate_error = vdev_validate(vd);
vd->vdev_validate_thread = NULL;
}
/*
* Called once the vdevs are all opened, this routine validates the label
* contents. This needs to be done before vdev_load() so that we don't
* inadvertently do repair I/Os to the wrong device.
*
* This function will only return failure if one of the vdevs indicates that it
* has since been destroyed or exported. This is only possible if
* /etc/zfs/zpool.cache was readonly at the time. Otherwise, the vdev state
* will be updated but the function will return 0.
*/
int
vdev_validate(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
taskq_t *tq = NULL;
nvlist_t *label;
uint64_t guid = 0, aux_guid = 0, top_guid;
uint64_t state;
nvlist_t *nvl;
uint64_t txg;
int children = vd->vdev_children;
if (vdev_validate_skip)
return (0);
if (children > 0) {
tq = taskq_create("vdev_validate", children, minclsyspri,
children, children, TASKQ_PREPOPULATE);
}
for (uint64_t c = 0; c < children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (tq == NULL || vdev_uses_zvols(cvd)) {
vdev_validate_child(cvd);
} else {
VERIFY(taskq_dispatch(tq, vdev_validate_child, cvd,
TQ_SLEEP) != TASKQID_INVALID);
}
}
if (tq != NULL) {
taskq_wait(tq);
taskq_destroy(tq);
}
for (int c = 0; c < children; c++) {
int error = vd->vdev_child[c]->vdev_validate_error;
if (error != 0)
return (SET_ERROR(EBADF));
}
/*
* If the device has already failed, or was marked offline, don't do
* any further validation. Otherwise, label I/O will fail and we will
* overwrite the previous state.
*/
if (!vd->vdev_ops->vdev_op_leaf || !vdev_readable(vd))
return (0);
/*
* If we are performing an extreme rewind, we allow for a label that
* was modified at a point after the current txg.
* If config lock is not held do not check for the txg. spa_sync could
* be updating the vdev's label before updating spa_last_synced_txg.
*/
if (spa->spa_extreme_rewind || spa_last_synced_txg(spa) == 0 ||
spa_config_held(spa, SCL_CONFIG, RW_WRITER) != SCL_CONFIG)
txg = UINT64_MAX;
else
txg = spa_last_synced_txg(spa);
if ((label = vdev_label_read_config(vd, txg)) == NULL) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_BAD_LABEL);
vdev_dbgmsg(vd, "vdev_validate: failed reading config for "
"txg %llu", (u_longlong_t)txg);
return (0);
}
/*
* Determine if this vdev has been split off into another
* pool. If so, then refuse to open it.
*/
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_SPLIT_GUID,
&aux_guid) == 0 && aux_guid == spa_guid(spa)) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_SPLIT_POOL);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: vdev split into other pool");
return (0);
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID, &guid) != 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label",
ZPOOL_CONFIG_POOL_GUID);
return (0);
}
/*
* If config is not trusted then ignore the spa guid check. This is
* necessary because if the machine crashed during a re-guid the new
* guid might have been written to all of the vdev labels, but not the
* cached config. The check will be performed again once we have the
* trusted config from the MOS.
*/
if (spa->spa_trust_config && guid != spa_guid(spa)) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: vdev label pool_guid doesn't "
"match config (%llu != %llu)", (u_longlong_t)guid,
(u_longlong_t)spa_guid(spa));
return (0);
}
if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_VDEV_TREE, &nvl)
!= 0 || nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_ORIG_GUID,
&aux_guid) != 0)
aux_guid = 0;
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label",
ZPOOL_CONFIG_GUID);
return (0);
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_TOP_GUID, &top_guid)
!= 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label",
ZPOOL_CONFIG_TOP_GUID);
return (0);
}
/*
* If this vdev just became a top-level vdev because its sibling was
* detached, it will have adopted the parent's vdev guid -- but the
* label may or may not be on disk yet. Fortunately, either version
* of the label will have the same top guid, so if we're a top-level
* vdev, we can safely compare to that instead.
* However, if the config comes from a cachefile that failed to update
* after the detach, a top-level vdev will appear as a non top-level
* vdev in the config. Also relax the constraints if we perform an
* extreme rewind.
*
* If we split this vdev off instead, then we also check the
* original pool's guid. We don't want to consider the vdev
* corrupt if it is partway through a split operation.
*/
if (vd->vdev_guid != guid && vd->vdev_guid != aux_guid) {
boolean_t mismatch = B_FALSE;
if (spa->spa_trust_config && !spa->spa_extreme_rewind) {
if (vd != vd->vdev_top || vd->vdev_guid != top_guid)
mismatch = B_TRUE;
} else {
if (vd->vdev_guid != top_guid &&
vd->vdev_top->vdev_guid != guid)
mismatch = B_TRUE;
}
if (mismatch) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: config guid "
"doesn't match label guid");
vdev_dbgmsg(vd, "CONFIG: guid %llu, top_guid %llu",
(u_longlong_t)vd->vdev_guid,
(u_longlong_t)vd->vdev_top->vdev_guid);
vdev_dbgmsg(vd, "LABEL: guid %llu, top_guid %llu, "
"aux_guid %llu", (u_longlong_t)guid,
(u_longlong_t)top_guid, (u_longlong_t)aux_guid);
return (0);
}
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE,
&state) != 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label",
ZPOOL_CONFIG_POOL_STATE);
return (0);
}
nvlist_free(label);
/*
* If this is a verbatim import, no need to check the
* state of the pool.
*/
if (!(spa->spa_import_flags & ZFS_IMPORT_VERBATIM) &&
spa_load_state(spa) == SPA_LOAD_OPEN &&
state != POOL_STATE_ACTIVE) {
vdev_dbgmsg(vd, "vdev_validate: invalid pool state (%llu) "
"for spa %s", (u_longlong_t)state, spa->spa_name);
return (SET_ERROR(EBADF));
}
/*
* If we were able to open and validate a vdev that was
* previously marked permanently unavailable, clear that state
* now.
*/
if (vd->vdev_not_present)
vd->vdev_not_present = 0;
return (0);
}
static void
vdev_copy_path_impl(vdev_t *svd, vdev_t *dvd)
{
char *old, *new;
if (svd->vdev_path != NULL && dvd->vdev_path != NULL) {
if (strcmp(svd->vdev_path, dvd->vdev_path) != 0) {
zfs_dbgmsg("vdev_copy_path: vdev %llu: path changed "
"from '%s' to '%s'", (u_longlong_t)dvd->vdev_guid,
dvd->vdev_path, svd->vdev_path);
spa_strfree(dvd->vdev_path);
dvd->vdev_path = spa_strdup(svd->vdev_path);
}
} else if (svd->vdev_path != NULL) {
dvd->vdev_path = spa_strdup(svd->vdev_path);
zfs_dbgmsg("vdev_copy_path: vdev %llu: path set to '%s'",
(u_longlong_t)dvd->vdev_guid, dvd->vdev_path);
}
/*
* Our enclosure sysfs path may have changed between imports
*/
old = dvd->vdev_enc_sysfs_path;
new = svd->vdev_enc_sysfs_path;
if ((old != NULL && new == NULL) ||
(old == NULL && new != NULL) ||
((old != NULL && new != NULL) && strcmp(new, old) != 0)) {
zfs_dbgmsg("vdev_copy_path: vdev %llu: vdev_enc_sysfs_path "
"changed from '%s' to '%s'", (u_longlong_t)dvd->vdev_guid,
old, new);
if (dvd->vdev_enc_sysfs_path)
spa_strfree(dvd->vdev_enc_sysfs_path);
if (svd->vdev_enc_sysfs_path) {
dvd->vdev_enc_sysfs_path = spa_strdup(
svd->vdev_enc_sysfs_path);
} else {
dvd->vdev_enc_sysfs_path = NULL;
}
}
}
/*
* Recursively copy vdev paths from one vdev to another. Source and destination
* vdev trees must have same geometry otherwise return error. Intended to copy
* paths from userland config into MOS config.
*/
int
vdev_copy_path_strict(vdev_t *svd, vdev_t *dvd)
{
if ((svd->vdev_ops == &vdev_missing_ops) ||
(svd->vdev_ishole && dvd->vdev_ishole) ||
(dvd->vdev_ops == &vdev_indirect_ops))
return (0);
if (svd->vdev_ops != dvd->vdev_ops) {
vdev_dbgmsg(svd, "vdev_copy_path: vdev type mismatch: %s != %s",
svd->vdev_ops->vdev_op_type, dvd->vdev_ops->vdev_op_type);
return (SET_ERROR(EINVAL));
}
if (svd->vdev_guid != dvd->vdev_guid) {
vdev_dbgmsg(svd, "vdev_copy_path: guids mismatch (%llu != "
"%llu)", (u_longlong_t)svd->vdev_guid,
(u_longlong_t)dvd->vdev_guid);
return (SET_ERROR(EINVAL));
}
if (svd->vdev_children != dvd->vdev_children) {
vdev_dbgmsg(svd, "vdev_copy_path: children count mismatch: "
"%llu != %llu", (u_longlong_t)svd->vdev_children,
(u_longlong_t)dvd->vdev_children);
return (SET_ERROR(EINVAL));
}
for (uint64_t i = 0; i < svd->vdev_children; i++) {
int error = vdev_copy_path_strict(svd->vdev_child[i],
dvd->vdev_child[i]);
if (error != 0)
return (error);
}
if (svd->vdev_ops->vdev_op_leaf)
vdev_copy_path_impl(svd, dvd);
return (0);
}
static void
vdev_copy_path_search(vdev_t *stvd, vdev_t *dvd)
{
ASSERT(stvd->vdev_top == stvd);
ASSERT3U(stvd->vdev_id, ==, dvd->vdev_top->vdev_id);
for (uint64_t i = 0; i < dvd->vdev_children; i++) {
vdev_copy_path_search(stvd, dvd->vdev_child[i]);
}
if (!dvd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(dvd))
return;
/*
* The idea here is that while a vdev can shift positions within
* a top vdev (when replacing, attaching mirror, etc.) it cannot
* step outside of it.
*/
vdev_t *vd = vdev_lookup_by_guid(stvd, dvd->vdev_guid);
if (vd == NULL || vd->vdev_ops != dvd->vdev_ops)
return;
ASSERT(vd->vdev_ops->vdev_op_leaf);
vdev_copy_path_impl(vd, dvd);
}
/*
* Recursively copy vdev paths from one root vdev to another. Source and
* destination vdev trees may differ in geometry. For each destination leaf
* vdev, search a vdev with the same guid and top vdev id in the source.
* Intended to copy paths from userland config into MOS config.
*/
void
vdev_copy_path_relaxed(vdev_t *srvd, vdev_t *drvd)
{
uint64_t children = MIN(srvd->vdev_children, drvd->vdev_children);
ASSERT(srvd->vdev_ops == &vdev_root_ops);
ASSERT(drvd->vdev_ops == &vdev_root_ops);
for (uint64_t i = 0; i < children; i++) {
vdev_copy_path_search(srvd->vdev_child[i],
drvd->vdev_child[i]);
}
}
/*
* Close a virtual device.
*/
void
vdev_close(vdev_t *vd)
{
vdev_t *pvd = vd->vdev_parent;
spa_t *spa __maybe_unused = vd->vdev_spa;
ASSERT(vd != NULL);
ASSERT(vd->vdev_open_thread == curthread ||
spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
/*
* If our parent is reopening, then we are as well, unless we are
* going offline.
*/
if (pvd != NULL && pvd->vdev_reopening)
vd->vdev_reopening = (pvd->vdev_reopening && !vd->vdev_offline);
vd->vdev_ops->vdev_op_close(vd);
vdev_cache_purge(vd);
/*
* We record the previous state before we close it, so that if we are
* doing a reopen(), we don't generate FMA ereports if we notice that
* it's still faulted.
*/
vd->vdev_prevstate = vd->vdev_state;
if (vd->vdev_offline)
vd->vdev_state = VDEV_STATE_OFFLINE;
else
vd->vdev_state = VDEV_STATE_CLOSED;
vd->vdev_stat.vs_aux = VDEV_AUX_NONE;
}
void
vdev_hold(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_is_root(spa));
if (spa->spa_state == POOL_STATE_UNINITIALIZED)
return;
for (int c = 0; c < vd->vdev_children; c++)
vdev_hold(vd->vdev_child[c]);
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_ops->vdev_op_hold != NULL)
vd->vdev_ops->vdev_op_hold(vd);
}
void
vdev_rele(vdev_t *vd)
{
ASSERT(spa_is_root(vd->vdev_spa));
for (int c = 0; c < vd->vdev_children; c++)
vdev_rele(vd->vdev_child[c]);
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_ops->vdev_op_rele != NULL)
vd->vdev_ops->vdev_op_rele(vd);
}
/*
* Reopen all interior vdevs and any unopened leaves. We don't actually
* reopen leaf vdevs which had previously been opened as they might deadlock
* on the spa_config_lock. Instead we only obtain the leaf's physical size.
* If the leaf has never been opened then open it, as usual.
*/
void
vdev_reopen(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
/* set the reopening flag unless we're taking the vdev offline */
vd->vdev_reopening = !vd->vdev_offline;
vdev_close(vd);
(void) vdev_open(vd);
/*
* Call vdev_validate() here to make sure we have the same device.
* Otherwise, a device with an invalid label could be successfully
* opened in response to vdev_reopen().
*/
if (vd->vdev_aux) {
(void) vdev_validate_aux(vd);
if (vdev_readable(vd) && vdev_writeable(vd) &&
vd->vdev_aux == &spa->spa_l2cache) {
/*
* In case the vdev is present we should evict all ARC
* buffers and pointers to log blocks and reclaim their
* space before restoring its contents to L2ARC.
*/
if (l2arc_vdev_present(vd)) {
l2arc_rebuild_vdev(vd, B_TRUE);
} else {
l2arc_add_vdev(spa, vd);
}
spa_async_request(spa, SPA_ASYNC_L2CACHE_REBUILD);
spa_async_request(spa, SPA_ASYNC_L2CACHE_TRIM);
}
} else {
(void) vdev_validate(vd);
}
/*
* Reassess parent vdev's health.
*/
vdev_propagate_state(vd);
}
int
vdev_create(vdev_t *vd, uint64_t txg, boolean_t isreplacing)
{
int error;
/*
* Normally, partial opens (e.g. of a mirror) are allowed.
* For a create, however, we want to fail the request if
* there are any components we can't open.
*/
error = vdev_open(vd);
if (error || vd->vdev_state != VDEV_STATE_HEALTHY) {
vdev_close(vd);
return (error ? error : SET_ERROR(ENXIO));
}
/*
* Recursively load DTLs and initialize all labels.
*/
if ((error = vdev_dtl_load(vd)) != 0 ||
(error = vdev_label_init(vd, txg, isreplacing ?
VDEV_LABEL_REPLACE : VDEV_LABEL_CREATE)) != 0) {
vdev_close(vd);
return (error);
}
return (0);
}
void
vdev_metaslab_set_size(vdev_t *vd)
{
uint64_t asize = vd->vdev_asize;
uint64_t ms_count = asize >> zfs_vdev_default_ms_shift;
uint64_t ms_shift;
/*
* There are two dimensions to the metaslab sizing calculation:
* the size of the metaslab and the count of metaslabs per vdev.
*
* The default values used below are a good balance between memory
* usage (larger metaslab size means more memory needed for loaded
* metaslabs; more metaslabs means more memory needed for the
* metaslab_t structs), metaslab load time (larger metaslabs take
* longer to load), and metaslab sync time (more metaslabs means
* more time spent syncing all of them).
*
* In general, we aim for zfs_vdev_default_ms_count (200) metaslabs.
* The range of the dimensions are as follows:
*
* 2^29 <= ms_size <= 2^34
* 16 <= ms_count <= 131,072
*
* On the lower end of vdev sizes, we aim for metaslabs sizes of
* at least 512MB (2^29) to minimize fragmentation effects when
* testing with smaller devices. However, the count constraint
* of at least 16 metaslabs will override this minimum size goal.
*
* On the upper end of vdev sizes, we aim for a maximum metaslab
* size of 16GB. However, we will cap the total count to 2^17
* metaslabs to keep our memory footprint in check and let the
* metaslab size grow from there if that limit is hit.
*
* The net effect of applying above constrains is summarized below.
*
* vdev size metaslab count
* --------------|-----------------
* < 8GB ~16
* 8GB - 100GB one per 512MB
* 100GB - 3TB ~200
* 3TB - 2PB one per 16GB
* > 2PB ~131,072
* --------------------------------
*
* Finally, note that all of the above calculate the initial
* number of metaslabs. Expanding a top-level vdev will result
* in additional metaslabs being allocated making it possible
* to exceed the zfs_vdev_ms_count_limit.
*/
if (ms_count < zfs_vdev_min_ms_count)
ms_shift = highbit64(asize / zfs_vdev_min_ms_count);
else if (ms_count > zfs_vdev_default_ms_count)
ms_shift = highbit64(asize / zfs_vdev_default_ms_count);
else
ms_shift = zfs_vdev_default_ms_shift;
if (ms_shift < SPA_MAXBLOCKSHIFT) {
ms_shift = SPA_MAXBLOCKSHIFT;
} else if (ms_shift > zfs_vdev_max_ms_shift) {
ms_shift = zfs_vdev_max_ms_shift;
/* cap the total count to constrain memory footprint */
if ((asize >> ms_shift) > zfs_vdev_ms_count_limit)
ms_shift = highbit64(asize / zfs_vdev_ms_count_limit);
}
vd->vdev_ms_shift = ms_shift;
ASSERT3U(vd->vdev_ms_shift, >=, SPA_MAXBLOCKSHIFT);
}
void
vdev_dirty(vdev_t *vd, int flags, void *arg, uint64_t txg)
{
ASSERT(vd == vd->vdev_top);
/* indirect vdevs don't have metaslabs or dtls */
ASSERT(vdev_is_concrete(vd) || flags == 0);
ASSERT(ISP2(flags));
ASSERT(spa_writeable(vd->vdev_spa));
if (flags & VDD_METASLAB)
(void) txg_list_add(&vd->vdev_ms_list, arg, txg);
if (flags & VDD_DTL)
(void) txg_list_add(&vd->vdev_dtl_list, arg, txg);
(void) txg_list_add(&vd->vdev_spa->spa_vdev_txg_list, vd, txg);
}
void
vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg)
{
for (int c = 0; c < vd->vdev_children; c++)
vdev_dirty_leaves(vd->vdev_child[c], flags, txg);
if (vd->vdev_ops->vdev_op_leaf)
vdev_dirty(vd->vdev_top, flags, vd, txg);
}
/*
* DTLs.
*
* A vdev's DTL (dirty time log) is the set of transaction groups for which
* the vdev has less than perfect replication. There are four kinds of DTL:
*
* DTL_MISSING: txgs for which the vdev has no valid copies of the data
*
* DTL_PARTIAL: txgs for which data is available, but not fully replicated
*
* DTL_SCRUB: the txgs that could not be repaired by the last scrub; upon
* scrub completion, DTL_SCRUB replaces DTL_MISSING in the range of
* txgs that was scrubbed.
*
* DTL_OUTAGE: txgs which cannot currently be read, whether due to
* persistent errors or just some device being offline.
* Unlike the other three, the DTL_OUTAGE map is not generally
* maintained; it's only computed when needed, typically to
* determine whether a device can be detached.
*
* For leaf vdevs, DTL_MISSING and DTL_PARTIAL are identical: the device
* either has the data or it doesn't.
*
* For interior vdevs such as mirror and RAID-Z the picture is more complex.
* A vdev's DTL_PARTIAL is the union of its children's DTL_PARTIALs, because
* if any child is less than fully replicated, then so is its parent.
* A vdev's DTL_MISSING is a modified union of its children's DTL_MISSINGs,
* comprising only those txgs which appear in 'maxfaults' or more children;
* those are the txgs we don't have enough replication to read. For example,
* double-parity RAID-Z can tolerate up to two missing devices (maxfaults == 2);
* thus, its DTL_MISSING consists of the set of txgs that appear in more than
* two child DTL_MISSING maps.
*
* It should be clear from the above that to compute the DTLs and outage maps
* for all vdevs, it suffices to know just the leaf vdevs' DTL_MISSING maps.
* Therefore, that is all we keep on disk. When loading the pool, or after
* a configuration change, we generate all other DTLs from first principles.
*/
void
vdev_dtl_dirty(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size)
{
range_tree_t *rt = vd->vdev_dtl[t];
ASSERT(t < DTL_TYPES);
ASSERT(vd != vd->vdev_spa->spa_root_vdev);
ASSERT(spa_writeable(vd->vdev_spa));
mutex_enter(&vd->vdev_dtl_lock);
if (!range_tree_contains(rt, txg, size))
range_tree_add(rt, txg, size);
mutex_exit(&vd->vdev_dtl_lock);
}
boolean_t
vdev_dtl_contains(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size)
{
range_tree_t *rt = vd->vdev_dtl[t];
boolean_t dirty = B_FALSE;
ASSERT(t < DTL_TYPES);
ASSERT(vd != vd->vdev_spa->spa_root_vdev);
/*
* While we are loading the pool, the DTLs have not been loaded yet.
* This isn't a problem but it can result in devices being tried
* which are known to not have the data. In which case, the import
* is relying on the checksum to ensure that we get the right data.
* Note that while importing we are only reading the MOS, which is
* always checksummed.
*/
mutex_enter(&vd->vdev_dtl_lock);
if (!range_tree_is_empty(rt))
dirty = range_tree_contains(rt, txg, size);
mutex_exit(&vd->vdev_dtl_lock);
return (dirty);
}
boolean_t
vdev_dtl_empty(vdev_t *vd, vdev_dtl_type_t t)
{
range_tree_t *rt = vd->vdev_dtl[t];
boolean_t empty;
mutex_enter(&vd->vdev_dtl_lock);
empty = range_tree_is_empty(rt);
mutex_exit(&vd->vdev_dtl_lock);
return (empty);
}
/*
* Check if the txg falls within the range which must be
* resilvered. DVAs outside this range can always be skipped.
*/
boolean_t
vdev_default_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize,
uint64_t phys_birth)
{
(void) dva, (void) psize;
/* Set by sequential resilver. */
if (phys_birth == TXG_UNKNOWN)
return (B_TRUE);
return (vdev_dtl_contains(vd, DTL_PARTIAL, phys_birth, 1));
}
/*
* Returns B_TRUE if the vdev determines the DVA needs to be resilvered.
*/
boolean_t
vdev_dtl_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize,
uint64_t phys_birth)
{
ASSERT(vd != vd->vdev_spa->spa_root_vdev);
if (vd->vdev_ops->vdev_op_need_resilver == NULL ||
vd->vdev_ops->vdev_op_leaf)
return (B_TRUE);
return (vd->vdev_ops->vdev_op_need_resilver(vd, dva, psize,
phys_birth));
}
/*
* Returns the lowest txg in the DTL range.
*/
static uint64_t
vdev_dtl_min(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock));
ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0);
ASSERT0(vd->vdev_children);
return (range_tree_min(vd->vdev_dtl[DTL_MISSING]) - 1);
}
/*
* Returns the highest txg in the DTL.
*/
static uint64_t
vdev_dtl_max(vdev_t *vd)
{
ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock));
ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0);
ASSERT0(vd->vdev_children);
return (range_tree_max(vd->vdev_dtl[DTL_MISSING]));
}
/*
* Determine if a resilvering vdev should remove any DTL entries from
* its range. If the vdev was resilvering for the entire duration of the
* scan then it should excise that range from its DTLs. Otherwise, this
* vdev is considered partially resilvered and should leave its DTL
* entries intact. The comment in vdev_dtl_reassess() describes how we
* excise the DTLs.
*/
static boolean_t
vdev_dtl_should_excise(vdev_t *vd, boolean_t rebuild_done)
{
ASSERT0(vd->vdev_children);
if (vd->vdev_state < VDEV_STATE_DEGRADED)
return (B_FALSE);
if (vd->vdev_resilver_deferred)
return (B_FALSE);
if (range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]))
return (B_TRUE);
if (rebuild_done) {
vdev_rebuild_t *vr = &vd->vdev_top->vdev_rebuild_config;
vdev_rebuild_phys_t *vrp = &vr->vr_rebuild_phys;
/* Rebuild not initiated by attach */
if (vd->vdev_rebuild_txg == 0)
return (B_TRUE);
/*
* When a rebuild completes without error then all missing data
* up to the rebuild max txg has been reconstructed and the DTL
* is eligible for excision.
*/
if (vrp->vrp_rebuild_state == VDEV_REBUILD_COMPLETE &&
vdev_dtl_max(vd) <= vrp->vrp_max_txg) {
ASSERT3U(vrp->vrp_min_txg, <=, vdev_dtl_min(vd));
ASSERT3U(vrp->vrp_min_txg, <, vd->vdev_rebuild_txg);
ASSERT3U(vd->vdev_rebuild_txg, <=, vrp->vrp_max_txg);
return (B_TRUE);
}
} else {
dsl_scan_t *scn = vd->vdev_spa->spa_dsl_pool->dp_scan;
dsl_scan_phys_t *scnp __maybe_unused = &scn->scn_phys;
/* Resilver not initiated by attach */
if (vd->vdev_resilver_txg == 0)
return (B_TRUE);
/*
* When a resilver is initiated the scan will assign the
* scn_max_txg value to the highest txg value that exists
* in all DTLs. If this device's max DTL is not part of this
* scan (i.e. it is not in the range (scn_min_txg, scn_max_txg]
* then it is not eligible for excision.
*/
if (vdev_dtl_max(vd) <= scn->scn_phys.scn_max_txg) {
ASSERT3U(scnp->scn_min_txg, <=, vdev_dtl_min(vd));
ASSERT3U(scnp->scn_min_txg, <, vd->vdev_resilver_txg);
ASSERT3U(vd->vdev_resilver_txg, <=, scnp->scn_max_txg);
return (B_TRUE);
}
}
return (B_FALSE);
}
/*
* Reassess DTLs after a config change or scrub completion. If txg == 0 no
* write operations will be issued to the pool.
*/
void
vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg,
boolean_t scrub_done, boolean_t rebuild_done)
{
spa_t *spa = vd->vdev_spa;
avl_tree_t reftree;
int minref;
ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
for (int c = 0; c < vd->vdev_children; c++)
vdev_dtl_reassess(vd->vdev_child[c], txg,
scrub_txg, scrub_done, rebuild_done);
if (vd == spa->spa_root_vdev || !vdev_is_concrete(vd) || vd->vdev_aux)
return;
if (vd->vdev_ops->vdev_op_leaf) {
dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan;
vdev_rebuild_t *vr = &vd->vdev_top->vdev_rebuild_config;
boolean_t check_excise = B_FALSE;
boolean_t wasempty = B_TRUE;
mutex_enter(&vd->vdev_dtl_lock);
/*
* If requested, pretend the scan or rebuild completed cleanly.
*/
if (zfs_scan_ignore_errors) {
if (scn != NULL)
scn->scn_phys.scn_errors = 0;
if (vr != NULL)
vr->vr_rebuild_phys.vrp_errors = 0;
}
if (scrub_txg != 0 &&
!range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) {
wasempty = B_FALSE;
zfs_dbgmsg("guid:%llu txg:%llu scrub:%llu started:%d "
"dtl:%llu/%llu errors:%llu",
(u_longlong_t)vd->vdev_guid, (u_longlong_t)txg,
(u_longlong_t)scrub_txg, spa->spa_scrub_started,
(u_longlong_t)vdev_dtl_min(vd),
(u_longlong_t)vdev_dtl_max(vd),
(u_longlong_t)(scn ? scn->scn_phys.scn_errors : 0));
}
/*
* If we've completed a scrub/resilver or a rebuild cleanly
* then determine if this vdev should remove any DTLs. We
* only want to excise regions on vdevs that were available
* during the entire duration of this scan.
*/
if (rebuild_done &&
vr != NULL && vr->vr_rebuild_phys.vrp_errors == 0) {
check_excise = B_TRUE;
} else {
if (spa->spa_scrub_started ||
(scn != NULL && scn->scn_phys.scn_errors == 0)) {
check_excise = B_TRUE;
}
}
if (scrub_txg && check_excise &&
vdev_dtl_should_excise(vd, rebuild_done)) {
/*
* We completed a scrub, resilver or rebuild up to
* scrub_txg. If we did it without rebooting, then
* the scrub dtl will be valid, so excise the old
* region and fold in the scrub dtl. Otherwise,
* leave the dtl as-is if there was an error.
*
* There's little trick here: to excise the beginning
* of the DTL_MISSING map, we put it into a reference
* tree and then add a segment with refcnt -1 that
* covers the range [0, scrub_txg). This means
* that each txg in that range has refcnt -1 or 0.
* We then add DTL_SCRUB with a refcnt of 2, so that
* entries in the range [0, scrub_txg) will have a
* positive refcnt -- either 1 or 2. We then convert
* the reference tree into the new DTL_MISSING map.
*/
space_reftree_create(&reftree);
space_reftree_add_map(&reftree,
vd->vdev_dtl[DTL_MISSING], 1);
space_reftree_add_seg(&reftree, 0, scrub_txg, -1);
space_reftree_add_map(&reftree,
vd->vdev_dtl[DTL_SCRUB], 2);
space_reftree_generate_map(&reftree,
vd->vdev_dtl[DTL_MISSING], 1);
space_reftree_destroy(&reftree);
if (!range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) {
zfs_dbgmsg("update DTL_MISSING:%llu/%llu",
(u_longlong_t)vdev_dtl_min(vd),
(u_longlong_t)vdev_dtl_max(vd));
} else if (!wasempty) {
zfs_dbgmsg("DTL_MISSING is now empty");
}
}
range_tree_vacate(vd->vdev_dtl[DTL_PARTIAL], NULL, NULL);
range_tree_walk(vd->vdev_dtl[DTL_MISSING],
range_tree_add, vd->vdev_dtl[DTL_PARTIAL]);
if (scrub_done)
range_tree_vacate(vd->vdev_dtl[DTL_SCRUB], NULL, NULL);
range_tree_vacate(vd->vdev_dtl[DTL_OUTAGE], NULL, NULL);
if (!vdev_readable(vd))
range_tree_add(vd->vdev_dtl[DTL_OUTAGE], 0, -1ULL);
else
range_tree_walk(vd->vdev_dtl[DTL_MISSING],
range_tree_add, vd->vdev_dtl[DTL_OUTAGE]);
/*
* If the vdev was resilvering or rebuilding and no longer
* has any DTLs then reset the appropriate flag and dirty
* the top level so that we persist the change.
*/
if (txg != 0 &&
range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]) &&
range_tree_is_empty(vd->vdev_dtl[DTL_OUTAGE])) {
if (vd->vdev_rebuild_txg != 0) {
vd->vdev_rebuild_txg = 0;
vdev_config_dirty(vd->vdev_top);
} else if (vd->vdev_resilver_txg != 0) {
vd->vdev_resilver_txg = 0;
vdev_config_dirty(vd->vdev_top);
}
}
mutex_exit(&vd->vdev_dtl_lock);
if (txg != 0)
vdev_dirty(vd->vdev_top, VDD_DTL, vd, txg);
return;
}
mutex_enter(&vd->vdev_dtl_lock);
for (int t = 0; t < DTL_TYPES; t++) {
/* account for child's outage in parent's missing map */
int s = (t == DTL_MISSING) ? DTL_OUTAGE: t;
if (t == DTL_SCRUB)
continue; /* leaf vdevs only */
if (t == DTL_PARTIAL)
minref = 1; /* i.e. non-zero */
else if (vdev_get_nparity(vd) != 0)
minref = vdev_get_nparity(vd) + 1; /* RAID-Z, dRAID */
else
minref = vd->vdev_children; /* any kind of mirror */
space_reftree_create(&reftree);
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
mutex_enter(&cvd->vdev_dtl_lock);
space_reftree_add_map(&reftree, cvd->vdev_dtl[s], 1);
mutex_exit(&cvd->vdev_dtl_lock);
}
space_reftree_generate_map(&reftree, vd->vdev_dtl[t], minref);
space_reftree_destroy(&reftree);
}
mutex_exit(&vd->vdev_dtl_lock);
}
/*
* Iterate over all the vdevs except spare, and post kobj events
*/
void
vdev_post_kobj_evt(vdev_t *vd)
{
if (vd->vdev_ops->vdev_op_kobj_evt_post &&
vd->vdev_kobj_flag == B_FALSE) {
vd->vdev_kobj_flag = B_TRUE;
vd->vdev_ops->vdev_op_kobj_evt_post(vd);
}
for (int c = 0; c < vd->vdev_children; c++)
vdev_post_kobj_evt(vd->vdev_child[c]);
}
/*
* Iterate over all the vdevs except spare, and clear kobj events
*/
void
vdev_clear_kobj_evt(vdev_t *vd)
{
vd->vdev_kobj_flag = B_FALSE;
for (int c = 0; c < vd->vdev_children; c++)
vdev_clear_kobj_evt(vd->vdev_child[c]);
}
int
vdev_dtl_load(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
range_tree_t *rt;
int error = 0;
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_dtl_object != 0) {
ASSERT(vdev_is_concrete(vd));
/*
* If the dtl cannot be sync'd there is no need to open it.
*/
if (spa->spa_mode == SPA_MODE_READ && !spa->spa_read_spacemaps)
return (0);
error = space_map_open(&vd->vdev_dtl_sm, mos,
vd->vdev_dtl_object, 0, -1ULL, 0);
if (error)
return (error);
ASSERT(vd->vdev_dtl_sm != NULL);
rt = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
error = space_map_load(vd->vdev_dtl_sm, rt, SM_ALLOC);
if (error == 0) {
mutex_enter(&vd->vdev_dtl_lock);
range_tree_walk(rt, range_tree_add,
vd->vdev_dtl[DTL_MISSING]);
mutex_exit(&vd->vdev_dtl_lock);
}
range_tree_vacate(rt, NULL, NULL);
range_tree_destroy(rt);
return (error);
}
for (int c = 0; c < vd->vdev_children; c++) {
error = vdev_dtl_load(vd->vdev_child[c]);
if (error != 0)
break;
}
return (error);
}
static void
vdev_zap_allocation_data(vdev_t *vd, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
vdev_alloc_bias_t alloc_bias = vd->vdev_alloc_bias;
const char *string;
ASSERT(alloc_bias != VDEV_BIAS_NONE);
string =
(alloc_bias == VDEV_BIAS_LOG) ? VDEV_ALLOC_BIAS_LOG :
(alloc_bias == VDEV_BIAS_SPECIAL) ? VDEV_ALLOC_BIAS_SPECIAL :
(alloc_bias == VDEV_BIAS_DEDUP) ? VDEV_ALLOC_BIAS_DEDUP : NULL;
ASSERT(string != NULL);
VERIFY0(zap_add(mos, vd->vdev_top_zap, VDEV_TOP_ZAP_ALLOCATION_BIAS,
1, strlen(string) + 1, string, tx));
if (alloc_bias == VDEV_BIAS_SPECIAL || alloc_bias == VDEV_BIAS_DEDUP) {
spa_activate_allocation_classes(spa, tx);
}
}
void
vdev_destroy_unlink_zap(vdev_t *vd, uint64_t zapobj, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
VERIFY0(zap_destroy(spa->spa_meta_objset, zapobj, tx));
VERIFY0(zap_remove_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps,
zapobj, tx));
}
uint64_t
vdev_create_link_zap(vdev_t *vd, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
uint64_t zap = zap_create(spa->spa_meta_objset, DMU_OTN_ZAP_METADATA,
DMU_OT_NONE, 0, tx);
ASSERT(zap != 0);
VERIFY0(zap_add_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps,
zap, tx));
return (zap);
}
void
vdev_construct_zaps(vdev_t *vd, dmu_tx_t *tx)
{
if (vd->vdev_ops != &vdev_hole_ops &&
vd->vdev_ops != &vdev_missing_ops &&
vd->vdev_ops != &vdev_root_ops &&
!vd->vdev_top->vdev_removing) {
if (vd->vdev_ops->vdev_op_leaf && vd->vdev_leaf_zap == 0) {
vd->vdev_leaf_zap = vdev_create_link_zap(vd, tx);
}
if (vd == vd->vdev_top && vd->vdev_top_zap == 0) {
vd->vdev_top_zap = vdev_create_link_zap(vd, tx);
if (vd->vdev_alloc_bias != VDEV_BIAS_NONE)
vdev_zap_allocation_data(vd, tx);
}
}
+ if (vd->vdev_ops == &vdev_root_ops && vd->vdev_root_zap == 0 &&
+ spa_feature_is_enabled(vd->vdev_spa, SPA_FEATURE_AVZ_V2)) {
+ if (!spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_AVZ_V2))
+ spa_feature_incr(vd->vdev_spa, SPA_FEATURE_AVZ_V2, tx);
+ vd->vdev_root_zap = vdev_create_link_zap(vd, tx);
+ }
for (uint64_t i = 0; i < vd->vdev_children; i++) {
vdev_construct_zaps(vd->vdev_child[i], tx);
}
}
static void
vdev_dtl_sync(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
range_tree_t *rt = vd->vdev_dtl[DTL_MISSING];
objset_t *mos = spa->spa_meta_objset;
range_tree_t *rtsync;
dmu_tx_t *tx;
uint64_t object = space_map_object(vd->vdev_dtl_sm);
ASSERT(vdev_is_concrete(vd));
ASSERT(vd->vdev_ops->vdev_op_leaf);
tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg);
if (vd->vdev_detached || vd->vdev_top->vdev_removing) {
mutex_enter(&vd->vdev_dtl_lock);
space_map_free(vd->vdev_dtl_sm, tx);
space_map_close(vd->vdev_dtl_sm);
vd->vdev_dtl_sm = NULL;
mutex_exit(&vd->vdev_dtl_lock);
/*
* We only destroy the leaf ZAP for detached leaves or for
* removed log devices. Removed data devices handle leaf ZAP
* cleanup later, once cancellation is no longer possible.
*/
if (vd->vdev_leaf_zap != 0 && (vd->vdev_detached ||
vd->vdev_top->vdev_islog)) {
vdev_destroy_unlink_zap(vd, vd->vdev_leaf_zap, tx);
vd->vdev_leaf_zap = 0;
}
dmu_tx_commit(tx);
return;
}
if (vd->vdev_dtl_sm == NULL) {
uint64_t new_object;
new_object = space_map_alloc(mos, zfs_vdev_dtl_sm_blksz, tx);
VERIFY3U(new_object, !=, 0);
VERIFY0(space_map_open(&vd->vdev_dtl_sm, mos, new_object,
0, -1ULL, 0));
ASSERT(vd->vdev_dtl_sm != NULL);
}
rtsync = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0);
mutex_enter(&vd->vdev_dtl_lock);
range_tree_walk(rt, range_tree_add, rtsync);
mutex_exit(&vd->vdev_dtl_lock);
space_map_truncate(vd->vdev_dtl_sm, zfs_vdev_dtl_sm_blksz, tx);
space_map_write(vd->vdev_dtl_sm, rtsync, SM_ALLOC, SM_NO_VDEVID, tx);
range_tree_vacate(rtsync, NULL, NULL);
range_tree_destroy(rtsync);
/*
* If the object for the space map has changed then dirty
* the top level so that we update the config.
*/
if (object != space_map_object(vd->vdev_dtl_sm)) {
vdev_dbgmsg(vd, "txg %llu, spa %s, DTL old object %llu, "
"new object %llu", (u_longlong_t)txg, spa_name(spa),
(u_longlong_t)object,
(u_longlong_t)space_map_object(vd->vdev_dtl_sm));
vdev_config_dirty(vd->vdev_top);
}
dmu_tx_commit(tx);
}
/*
* Determine whether the specified vdev can be offlined/detached/removed
* without losing data.
*/
boolean_t
vdev_dtl_required(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
vdev_t *tvd = vd->vdev_top;
uint8_t cant_read = vd->vdev_cant_read;
boolean_t required;
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
if (vd == spa->spa_root_vdev || vd == tvd)
return (B_TRUE);
/*
* Temporarily mark the device as unreadable, and then determine
* whether this results in any DTL outages in the top-level vdev.
* If not, we can safely offline/detach/remove the device.
*/
vd->vdev_cant_read = B_TRUE;
vdev_dtl_reassess(tvd, 0, 0, B_FALSE, B_FALSE);
required = !vdev_dtl_empty(tvd, DTL_OUTAGE);
vd->vdev_cant_read = cant_read;
vdev_dtl_reassess(tvd, 0, 0, B_FALSE, B_FALSE);
if (!required && zio_injection_enabled) {
required = !!zio_handle_device_injection(vd, NULL,
SET_ERROR(ECHILD));
}
return (required);
}
/*
* Determine if resilver is needed, and if so the txg range.
*/
boolean_t
vdev_resilver_needed(vdev_t *vd, uint64_t *minp, uint64_t *maxp)
{
boolean_t needed = B_FALSE;
uint64_t thismin = UINT64_MAX;
uint64_t thismax = 0;
if (vd->vdev_children == 0) {
mutex_enter(&vd->vdev_dtl_lock);
if (!range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]) &&
vdev_writeable(vd)) {
thismin = vdev_dtl_min(vd);
thismax = vdev_dtl_max(vd);
needed = B_TRUE;
}
mutex_exit(&vd->vdev_dtl_lock);
} else {
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
uint64_t cmin, cmax;
if (vdev_resilver_needed(cvd, &cmin, &cmax)) {
thismin = MIN(thismin, cmin);
thismax = MAX(thismax, cmax);
needed = B_TRUE;
}
}
}
if (needed && minp) {
*minp = thismin;
*maxp = thismax;
}
return (needed);
}
/*
* Gets the checkpoint space map object from the vdev's ZAP. On success sm_obj
* will contain either the checkpoint spacemap object or zero if none exists.
* All other errors are returned to the caller.
*/
int
vdev_checkpoint_sm_object(vdev_t *vd, uint64_t *sm_obj)
{
ASSERT0(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER));
if (vd->vdev_top_zap == 0) {
*sm_obj = 0;
return (0);
}
int error = zap_lookup(spa_meta_objset(vd->vdev_spa), vd->vdev_top_zap,
VDEV_TOP_ZAP_POOL_CHECKPOINT_SM, sizeof (uint64_t), 1, sm_obj);
if (error == ENOENT) {
*sm_obj = 0;
error = 0;
}
return (error);
}
int
vdev_load(vdev_t *vd)
{
int children = vd->vdev_children;
int error = 0;
taskq_t *tq = NULL;
/*
* It's only worthwhile to use the taskq for the root vdev, because the
* slow part is metaslab_init, and that only happens for top-level
* vdevs.
*/
if (vd->vdev_ops == &vdev_root_ops && vd->vdev_children > 0) {
tq = taskq_create("vdev_load", children, minclsyspri,
children, children, TASKQ_PREPOPULATE);
}
/*
* Recursively load all children.
*/
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
if (tq == NULL || vdev_uses_zvols(cvd)) {
cvd->vdev_load_error = vdev_load(cvd);
} else {
VERIFY(taskq_dispatch(tq, vdev_load_child,
cvd, TQ_SLEEP) != TASKQID_INVALID);
}
}
if (tq != NULL) {
taskq_wait(tq);
taskq_destroy(tq);
}
for (int c = 0; c < vd->vdev_children; c++) {
int error = vd->vdev_child[c]->vdev_load_error;
if (error != 0)
return (error);
}
vdev_set_deflate_ratio(vd);
/*
* On spa_load path, grab the allocation bias from our zap
*/
if (vd == vd->vdev_top && vd->vdev_top_zap != 0) {
spa_t *spa = vd->vdev_spa;
char bias_str[64];
error = zap_lookup(spa->spa_meta_objset, vd->vdev_top_zap,
VDEV_TOP_ZAP_ALLOCATION_BIAS, 1, sizeof (bias_str),
bias_str);
if (error == 0) {
ASSERT(vd->vdev_alloc_bias == VDEV_BIAS_NONE);
vd->vdev_alloc_bias = vdev_derive_alloc_bias(bias_str);
} else if (error != ENOENT) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: zap_lookup(top_zap=%llu) "
"failed [error=%d]",
(u_longlong_t)vd->vdev_top_zap, error);
return (error);
}
}
if (vd == vd->vdev_top && vd->vdev_top_zap != 0) {
spa_t *spa = vd->vdev_spa;
uint64_t failfast;
error = zap_lookup(spa->spa_meta_objset, vd->vdev_top_zap,
vdev_prop_to_name(VDEV_PROP_FAILFAST), sizeof (failfast),
1, &failfast);
if (error == 0) {
vd->vdev_failfast = failfast & 1;
} else if (error == ENOENT) {
vd->vdev_failfast = vdev_prop_default_numeric(
VDEV_PROP_FAILFAST);
} else {
vdev_dbgmsg(vd,
"vdev_load: zap_lookup(top_zap=%llu) "
"failed [error=%d]",
(u_longlong_t)vd->vdev_top_zap, error);
}
}
/*
* Load any rebuild state from the top-level vdev zap.
*/
if (vd == vd->vdev_top && vd->vdev_top_zap != 0) {
error = vdev_rebuild_load(vd);
if (error && error != ENOTSUP) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: vdev_rebuild_load "
"failed [error=%d]", error);
return (error);
}
}
if (vd->vdev_top_zap != 0 || vd->vdev_leaf_zap != 0) {
uint64_t zapobj;
if (vd->vdev_top_zap != 0)
zapobj = vd->vdev_top_zap;
else
zapobj = vd->vdev_leaf_zap;
error = vdev_prop_get_int(vd, VDEV_PROP_CHECKSUM_N,
&vd->vdev_checksum_n);
if (error && error != ENOENT)
vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) "
"failed [error=%d]", (u_longlong_t)zapobj, error);
error = vdev_prop_get_int(vd, VDEV_PROP_CHECKSUM_T,
&vd->vdev_checksum_t);
if (error && error != ENOENT)
vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) "
"failed [error=%d]", (u_longlong_t)zapobj, error);
error = vdev_prop_get_int(vd, VDEV_PROP_IO_N,
&vd->vdev_io_n);
if (error && error != ENOENT)
vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) "
"failed [error=%d]", (u_longlong_t)zapobj, error);
error = vdev_prop_get_int(vd, VDEV_PROP_IO_T,
&vd->vdev_io_t);
if (error && error != ENOENT)
vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) "
"failed [error=%d]", (u_longlong_t)zapobj, error);
}
/*
* If this is a top-level vdev, initialize its metaslabs.
*/
if (vd == vd->vdev_top && vdev_is_concrete(vd)) {
vdev_metaslab_group_create(vd);
if (vd->vdev_ashift == 0 || vd->vdev_asize == 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: invalid size. ashift=%llu, "
"asize=%llu", (u_longlong_t)vd->vdev_ashift,
(u_longlong_t)vd->vdev_asize);
return (SET_ERROR(ENXIO));
}
error = vdev_metaslab_init(vd, 0);
if (error != 0) {
vdev_dbgmsg(vd, "vdev_load: metaslab_init failed "
"[error=%d]", error);
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
return (error);
}
uint64_t checkpoint_sm_obj;
error = vdev_checkpoint_sm_object(vd, &checkpoint_sm_obj);
if (error == 0 && checkpoint_sm_obj != 0) {
objset_t *mos = spa_meta_objset(vd->vdev_spa);
ASSERT(vd->vdev_asize != 0);
ASSERT3P(vd->vdev_checkpoint_sm, ==, NULL);
error = space_map_open(&vd->vdev_checkpoint_sm,
mos, checkpoint_sm_obj, 0, vd->vdev_asize,
vd->vdev_ashift);
if (error != 0) {
vdev_dbgmsg(vd, "vdev_load: space_map_open "
"failed for checkpoint spacemap (obj %llu) "
"[error=%d]",
(u_longlong_t)checkpoint_sm_obj, error);
return (error);
}
ASSERT3P(vd->vdev_checkpoint_sm, !=, NULL);
/*
* Since the checkpoint_sm contains free entries
* exclusively we can use space_map_allocated() to
* indicate the cumulative checkpointed space that
* has been freed.
*/
vd->vdev_stat.vs_checkpoint_space =
-space_map_allocated(vd->vdev_checkpoint_sm);
vd->vdev_spa->spa_checkpoint_info.sci_dspace +=
vd->vdev_stat.vs_checkpoint_space;
} else if (error != 0) {
vdev_dbgmsg(vd, "vdev_load: failed to retrieve "
"checkpoint space map object from vdev ZAP "
"[error=%d]", error);
return (error);
}
}
/*
* If this is a leaf vdev, load its DTL.
*/
if (vd->vdev_ops->vdev_op_leaf && (error = vdev_dtl_load(vd)) != 0) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: vdev_dtl_load failed "
"[error=%d]", error);
return (error);
}
uint64_t obsolete_sm_object;
error = vdev_obsolete_sm_object(vd, &obsolete_sm_object);
if (error == 0 && obsolete_sm_object != 0) {
objset_t *mos = vd->vdev_spa->spa_meta_objset;
ASSERT(vd->vdev_asize != 0);
ASSERT3P(vd->vdev_obsolete_sm, ==, NULL);
if ((error = space_map_open(&vd->vdev_obsolete_sm, mos,
obsolete_sm_object, 0, vd->vdev_asize, 0))) {
vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
vdev_dbgmsg(vd, "vdev_load: space_map_open failed for "
"obsolete spacemap (obj %llu) [error=%d]",
(u_longlong_t)obsolete_sm_object, error);
return (error);
}
} else if (error != 0) {
vdev_dbgmsg(vd, "vdev_load: failed to retrieve obsolete "
"space map object from vdev ZAP [error=%d]", error);
return (error);
}
return (0);
}
/*
* The special vdev case is used for hot spares and l2cache devices. Its
* sole purpose it to set the vdev state for the associated vdev. To do this,
* we make sure that we can open the underlying device, then try to read the
* label, and make sure that the label is sane and that it hasn't been
* repurposed to another pool.
*/
int
vdev_validate_aux(vdev_t *vd)
{
nvlist_t *label;
uint64_t guid, version;
uint64_t state;
if (!vdev_readable(vd))
return (0);
if ((label = vdev_label_read_config(vd, -1ULL)) == NULL) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
return (-1);
}
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_VERSION, &version) != 0 ||
!SPA_VERSION_IS_SUPPORTED(version) ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0 ||
guid != vd->vdev_guid ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0) {
vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
nvlist_free(label);
return (-1);
}
/*
* We don't actually check the pool state here. If it's in fact in
* use by another pool, we update this fact on the fly when requested.
*/
nvlist_free(label);
return (0);
}
static void
vdev_destroy_ms_flush_data(vdev_t *vd, dmu_tx_t *tx)
{
objset_t *mos = spa_meta_objset(vd->vdev_spa);
if (vd->vdev_top_zap == 0)
return;
uint64_t object = 0;
int err = zap_lookup(mos, vd->vdev_top_zap,
VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS, sizeof (uint64_t), 1, &object);
if (err == ENOENT)
return;
VERIFY0(err);
VERIFY0(dmu_object_free(mos, object, tx));
VERIFY0(zap_remove(mos, vd->vdev_top_zap,
VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS, tx));
}
/*
* Free the objects used to store this vdev's spacemaps, and the array
* that points to them.
*/
void
vdev_destroy_spacemaps(vdev_t *vd, dmu_tx_t *tx)
{
if (vd->vdev_ms_array == 0)
return;
objset_t *mos = vd->vdev_spa->spa_meta_objset;
uint64_t array_count = vd->vdev_asize >> vd->vdev_ms_shift;
size_t array_bytes = array_count * sizeof (uint64_t);
uint64_t *smobj_array = kmem_alloc(array_bytes, KM_SLEEP);
VERIFY0(dmu_read(mos, vd->vdev_ms_array, 0,
array_bytes, smobj_array, 0));
for (uint64_t i = 0; i < array_count; i++) {
uint64_t smobj = smobj_array[i];
if (smobj == 0)
continue;
space_map_free_obj(mos, smobj, tx);
}
kmem_free(smobj_array, array_bytes);
VERIFY0(dmu_object_free(mos, vd->vdev_ms_array, tx));
vdev_destroy_ms_flush_data(vd, tx);
vd->vdev_ms_array = 0;
}
static void
vdev_remove_empty_log(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
ASSERT(vd->vdev_islog);
ASSERT(vd == vd->vdev_top);
ASSERT3U(txg, ==, spa_syncing_txg(spa));
dmu_tx_t *tx = dmu_tx_create_assigned(spa_get_dsl(spa), txg);
vdev_destroy_spacemaps(vd, tx);
if (vd->vdev_top_zap != 0) {
vdev_destroy_unlink_zap(vd, vd->vdev_top_zap, tx);
vd->vdev_top_zap = 0;
}
dmu_tx_commit(tx);
}
void
vdev_sync_done(vdev_t *vd, uint64_t txg)
{
metaslab_t *msp;
boolean_t reassess = !txg_list_empty(&vd->vdev_ms_list, TXG_CLEAN(txg));
ASSERT(vdev_is_concrete(vd));
while ((msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg)))
!= NULL)
metaslab_sync_done(msp, txg);
if (reassess) {
metaslab_sync_reassess(vd->vdev_mg);
if (vd->vdev_log_mg != NULL)
metaslab_sync_reassess(vd->vdev_log_mg);
}
}
void
vdev_sync(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
vdev_t *lvd;
metaslab_t *msp;
ASSERT3U(txg, ==, spa->spa_syncing_txg);
dmu_tx_t *tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg);
if (range_tree_space(vd->vdev_obsolete_segments) > 0) {
ASSERT(vd->vdev_removing ||
vd->vdev_ops == &vdev_indirect_ops);
vdev_indirect_sync_obsolete(vd, tx);
/*
* If the vdev is indirect, it can't have dirty
* metaslabs or DTLs.
*/
if (vd->vdev_ops == &vdev_indirect_ops) {
ASSERT(txg_list_empty(&vd->vdev_ms_list, txg));
ASSERT(txg_list_empty(&vd->vdev_dtl_list, txg));
dmu_tx_commit(tx);
return;
}
}
ASSERT(vdev_is_concrete(vd));
if (vd->vdev_ms_array == 0 && vd->vdev_ms_shift != 0 &&
!vd->vdev_removing) {
ASSERT(vd == vd->vdev_top);
ASSERT0(vd->vdev_indirect_config.vic_mapping_object);
vd->vdev_ms_array = dmu_object_alloc(spa->spa_meta_objset,
DMU_OT_OBJECT_ARRAY, 0, DMU_OT_NONE, 0, tx);
ASSERT(vd->vdev_ms_array != 0);
vdev_config_dirty(vd);
}
while ((msp = txg_list_remove(&vd->vdev_ms_list, txg)) != NULL) {
metaslab_sync(msp, txg);
(void) txg_list_add(&vd->vdev_ms_list, msp, TXG_CLEAN(txg));
}
while ((lvd = txg_list_remove(&vd->vdev_dtl_list, txg)) != NULL)
vdev_dtl_sync(lvd, txg);
/*
* If this is an empty log device being removed, destroy the
* metadata associated with it.
*/
if (vd->vdev_islog && vd->vdev_stat.vs_alloc == 0 && vd->vdev_removing)
vdev_remove_empty_log(vd, txg);
(void) txg_list_add(&spa->spa_vdev_txg_list, vd, TXG_CLEAN(txg));
dmu_tx_commit(tx);
}
uint64_t
vdev_psize_to_asize(vdev_t *vd, uint64_t psize)
{
return (vd->vdev_ops->vdev_op_asize(vd, psize));
}
/*
* Mark the given vdev faulted. A faulted vdev behaves as if the device could
* not be opened, and no I/O is attempted.
*/
int
vdev_fault(spa_t *spa, uint64_t guid, vdev_aux_t aux)
{
vdev_t *vd, *tvd;
spa_vdev_state_enter(spa, SCL_NONE);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP)));
tvd = vd->vdev_top;
/*
* If user did a 'zpool offline -f' then make the fault persist across
* reboots.
*/
if (aux == VDEV_AUX_EXTERNAL_PERSIST) {
/*
* There are two kinds of forced faults: temporary and
* persistent. Temporary faults go away at pool import, while
* persistent faults stay set. Both types of faults can be
* cleared with a zpool clear.
*
* We tell if a vdev is persistently faulted by looking at the
* ZPOOL_CONFIG_AUX_STATE nvpair. If it's set to "external" at
* import then it's a persistent fault. Otherwise, it's
* temporary. We get ZPOOL_CONFIG_AUX_STATE set to "external"
* by setting vd.vdev_stat.vs_aux to VDEV_AUX_EXTERNAL. This
* tells vdev_config_generate() (which gets run later) to set
* ZPOOL_CONFIG_AUX_STATE to "external" in the nvlist.
*/
vd->vdev_stat.vs_aux = VDEV_AUX_EXTERNAL;
vd->vdev_tmpoffline = B_FALSE;
aux = VDEV_AUX_EXTERNAL;
} else {
vd->vdev_tmpoffline = B_TRUE;
}
/*
* We don't directly use the aux state here, but if we do a
* vdev_reopen(), we need this value to be present to remember why we
* were faulted.
*/
vd->vdev_label_aux = aux;
/*
* Faulted state takes precedence over degraded.
*/
vd->vdev_delayed_close = B_FALSE;
vd->vdev_faulted = 1ULL;
vd->vdev_degraded = 0ULL;
vdev_set_state(vd, B_FALSE, VDEV_STATE_FAULTED, aux);
/*
* If this device has the only valid copy of the data, then
* back off and simply mark the vdev as degraded instead.
*/
if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) {
vd->vdev_degraded = 1ULL;
vd->vdev_faulted = 0ULL;
/*
* If we reopen the device and it's not dead, only then do we
* mark it degraded.
*/
vdev_reopen(tvd);
if (vdev_readable(vd))
vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux);
}
return (spa_vdev_state_exit(spa, vd, 0));
}
/*
* Mark the given vdev degraded. A degraded vdev is purely an indication to the
* user that something is wrong. The vdev continues to operate as normal as far
* as I/O is concerned.
*/
int
vdev_degrade(spa_t *spa, uint64_t guid, vdev_aux_t aux)
{
vdev_t *vd;
spa_vdev_state_enter(spa, SCL_NONE);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP)));
/*
* If the vdev is already faulted, then don't do anything.
*/
if (vd->vdev_faulted || vd->vdev_degraded)
return (spa_vdev_state_exit(spa, NULL, 0));
vd->vdev_degraded = 1ULL;
if (!vdev_is_dead(vd))
vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED,
aux);
return (spa_vdev_state_exit(spa, vd, 0));
}
int
vdev_remove_wanted(spa_t *spa, uint64_t guid)
{
vdev_t *vd;
spa_vdev_state_enter(spa, SCL_NONE);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
/*
* If the vdev is already removed, then don't do anything.
*/
if (vd->vdev_removed)
return (spa_vdev_state_exit(spa, NULL, 0));
vd->vdev_remove_wanted = B_TRUE;
spa_async_request(spa, SPA_ASYNC_REMOVE);
return (spa_vdev_state_exit(spa, vd, 0));
}
/*
* Online the given vdev.
*
* If 'ZFS_ONLINE_UNSPARE' is set, it implies two things. First, any attached
* spare device should be detached when the device finishes resilvering.
* Second, the online should be treated like a 'test' online case, so no FMA
* events are generated if the device fails to open.
*/
int
vdev_online(spa_t *spa, uint64_t guid, uint64_t flags, vdev_state_t *newstate)
{
vdev_t *vd, *tvd, *pvd, *rvd = spa->spa_root_vdev;
boolean_t wasoffline;
vdev_state_t oldstate;
spa_vdev_state_enter(spa, SCL_NONE);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP)));
wasoffline = (vd->vdev_offline || vd->vdev_tmpoffline);
oldstate = vd->vdev_state;
tvd = vd->vdev_top;
vd->vdev_offline = B_FALSE;
vd->vdev_tmpoffline = B_FALSE;
vd->vdev_checkremove = !!(flags & ZFS_ONLINE_CHECKREMOVE);
vd->vdev_forcefault = !!(flags & ZFS_ONLINE_FORCEFAULT);
/* XXX - L2ARC 1.0 does not support expansion */
if (!vd->vdev_aux) {
for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent)
pvd->vdev_expanding = !!((flags & ZFS_ONLINE_EXPAND) ||
spa->spa_autoexpand);
vd->vdev_expansion_time = gethrestime_sec();
}
vdev_reopen(tvd);
vd->vdev_checkremove = vd->vdev_forcefault = B_FALSE;
if (!vd->vdev_aux) {
for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent)
pvd->vdev_expanding = B_FALSE;
}
if (newstate)
*newstate = vd->vdev_state;
if ((flags & ZFS_ONLINE_UNSPARE) &&
!vdev_is_dead(vd) && vd->vdev_parent &&
vd->vdev_parent->vdev_ops == &vdev_spare_ops &&
vd->vdev_parent->vdev_child[0] == vd)
vd->vdev_unspare = B_TRUE;
if ((flags & ZFS_ONLINE_EXPAND) || spa->spa_autoexpand) {
/* XXX - L2ARC 1.0 does not support expansion */
if (vd->vdev_aux)
return (spa_vdev_state_exit(spa, vd, ENOTSUP));
spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE);
}
/* Restart initializing if necessary */
mutex_enter(&vd->vdev_initialize_lock);
if (vdev_writeable(vd) &&
vd->vdev_initialize_thread == NULL &&
vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE) {
(void) vdev_initialize(vd);
}
mutex_exit(&vd->vdev_initialize_lock);
/*
* Restart trimming if necessary. We do not restart trimming for cache
* devices here. This is triggered by l2arc_rebuild_vdev()
* asynchronously for the whole device or in l2arc_evict() as it evicts
* space for upcoming writes.
*/
mutex_enter(&vd->vdev_trim_lock);
if (vdev_writeable(vd) && !vd->vdev_isl2cache &&
vd->vdev_trim_thread == NULL &&
vd->vdev_trim_state == VDEV_TRIM_ACTIVE) {
(void) vdev_trim(vd, vd->vdev_trim_rate, vd->vdev_trim_partial,
vd->vdev_trim_secure);
}
mutex_exit(&vd->vdev_trim_lock);
if (wasoffline ||
(oldstate < VDEV_STATE_DEGRADED &&
- vd->vdev_state >= VDEV_STATE_DEGRADED))
+ vd->vdev_state >= VDEV_STATE_DEGRADED)) {
spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_ONLINE);
+ /*
+ * Asynchronously detach spare vdev if resilver or
+ * rebuild is not required
+ */
+ if (vd->vdev_unspare &&
+ !dsl_scan_resilvering(spa->spa_dsl_pool) &&
+ !dsl_scan_resilver_scheduled(spa->spa_dsl_pool) &&
+ !vdev_rebuild_active(tvd))
+ spa_async_request(spa, SPA_ASYNC_DETACH_SPARE);
+ }
return (spa_vdev_state_exit(spa, vd, 0));
}
static int
vdev_offline_locked(spa_t *spa, uint64_t guid, uint64_t flags)
{
vdev_t *vd, *tvd;
int error = 0;
uint64_t generation;
metaslab_group_t *mg;
top:
spa_vdev_state_enter(spa, SCL_ALLOC);
if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV)));
if (!vd->vdev_ops->vdev_op_leaf)
return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP)));
if (vd->vdev_ops == &vdev_draid_spare_ops)
return (spa_vdev_state_exit(spa, NULL, ENOTSUP));
tvd = vd->vdev_top;
mg = tvd->vdev_mg;
generation = spa->spa_config_generation + 1;
/*
* If the device isn't already offline, try to offline it.
*/
if (!vd->vdev_offline) {
/*
* If this device has the only valid copy of some data,
* don't allow it to be offlined. Log devices are always
* expendable.
*/
if (!tvd->vdev_islog && vd->vdev_aux == NULL &&
vdev_dtl_required(vd))
return (spa_vdev_state_exit(spa, NULL,
SET_ERROR(EBUSY)));
/*
* If the top-level is a slog and it has had allocations
* then proceed. We check that the vdev's metaslab group
* is not NULL since it's possible that we may have just
* added this vdev but not yet initialized its metaslabs.
*/
if (tvd->vdev_islog && mg != NULL) {
/*
* Prevent any future allocations.
*/
ASSERT3P(tvd->vdev_log_mg, ==, NULL);
metaslab_group_passivate(mg);
(void) spa_vdev_state_exit(spa, vd, 0);
error = spa_reset_logs(spa);
/*
* If the log device was successfully reset but has
* checkpointed data, do not offline it.
*/
if (error == 0 &&
tvd->vdev_checkpoint_sm != NULL) {
ASSERT3U(space_map_allocated(
tvd->vdev_checkpoint_sm), !=, 0);
error = ZFS_ERR_CHECKPOINT_EXISTS;
}
spa_vdev_state_enter(spa, SCL_ALLOC);
/*
* Check to see if the config has changed.
*/
if (error || generation != spa->spa_config_generation) {
metaslab_group_activate(mg);
if (error)
return (spa_vdev_state_exit(spa,
vd, error));
(void) spa_vdev_state_exit(spa, vd, 0);
goto top;
}
ASSERT0(tvd->vdev_stat.vs_alloc);
}
/*
* Offline this device and reopen its top-level vdev.
* If the top-level vdev is a log device then just offline
* it. Otherwise, if this action results in the top-level
* vdev becoming unusable, undo it and fail the request.
*/
vd->vdev_offline = B_TRUE;
vdev_reopen(tvd);
if (!tvd->vdev_islog && vd->vdev_aux == NULL &&
vdev_is_dead(tvd)) {
vd->vdev_offline = B_FALSE;
vdev_reopen(tvd);
return (spa_vdev_state_exit(spa, NULL,
SET_ERROR(EBUSY)));
}
/*
* Add the device back into the metaslab rotor so that
* once we online the device it's open for business.
*/
if (tvd->vdev_islog && mg != NULL)
metaslab_group_activate(mg);
}
vd->vdev_tmpoffline = !!(flags & ZFS_OFFLINE_TEMPORARY);
return (spa_vdev_state_exit(spa, vd, 0));
}
int
vdev_offline(spa_t *spa, uint64_t guid, uint64_t flags)
{
int error;
mutex_enter(&spa->spa_vdev_top_lock);
error = vdev_offline_locked(spa, guid, flags);
mutex_exit(&spa->spa_vdev_top_lock);
return (error);
}
/*
* Clear the error counts associated with this vdev. Unlike vdev_online() and
* vdev_offline(), we assume the spa config is locked. We also clear all
* children. If 'vd' is NULL, then the user wants to clear all vdevs.
*/
void
vdev_clear(spa_t *spa, vdev_t *vd)
{
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
if (vd == NULL)
vd = rvd;
vd->vdev_stat.vs_read_errors = 0;
vd->vdev_stat.vs_write_errors = 0;
vd->vdev_stat.vs_checksum_errors = 0;
vd->vdev_stat.vs_slow_ios = 0;
for (int c = 0; c < vd->vdev_children; c++)
vdev_clear(spa, vd->vdev_child[c]);
/*
* It makes no sense to "clear" an indirect or removed vdev.
*/
if (!vdev_is_concrete(vd) || vd->vdev_removed)
return;
/*
* If we're in the FAULTED state or have experienced failed I/O, then
* clear the persistent state and attempt to reopen the device. We
* also mark the vdev config dirty, so that the new faulted state is
* written out to disk.
*/
if (vd->vdev_faulted || vd->vdev_degraded ||
!vdev_readable(vd) || !vdev_writeable(vd)) {
/*
* When reopening in response to a clear event, it may be due to
* a fmadm repair request. In this case, if the device is
* still broken, we want to still post the ereport again.
*/
vd->vdev_forcefault = B_TRUE;
vd->vdev_faulted = vd->vdev_degraded = 0ULL;
vd->vdev_cant_read = B_FALSE;
vd->vdev_cant_write = B_FALSE;
vd->vdev_stat.vs_aux = 0;
vdev_reopen(vd == rvd ? rvd : vd->vdev_top);
vd->vdev_forcefault = B_FALSE;
if (vd != rvd && vdev_writeable(vd->vdev_top))
vdev_state_dirty(vd->vdev_top);
/* If a resilver isn't required, check if vdevs can be culled */
if (vd->vdev_aux == NULL && !vdev_is_dead(vd) &&
!dsl_scan_resilvering(spa->spa_dsl_pool) &&
!dsl_scan_resilver_scheduled(spa->spa_dsl_pool))
spa_async_request(spa, SPA_ASYNC_RESILVER_DONE);
spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_CLEAR);
}
/*
* When clearing a FMA-diagnosed fault, we always want to
* unspare the device, as we assume that the original spare was
* done in response to the FMA fault.
*/
if (!vdev_is_dead(vd) && vd->vdev_parent != NULL &&
vd->vdev_parent->vdev_ops == &vdev_spare_ops &&
vd->vdev_parent->vdev_child[0] == vd)
vd->vdev_unspare = B_TRUE;
/* Clear recent error events cache (i.e. duplicate events tracking) */
zfs_ereport_clear(spa, vd);
}
boolean_t
vdev_is_dead(vdev_t *vd)
{
/*
* Holes and missing devices are always considered "dead".
* This simplifies the code since we don't have to check for
* these types of devices in the various code paths.
* Instead we rely on the fact that we skip over dead devices
* before issuing I/O to them.
*/
return (vd->vdev_state < VDEV_STATE_DEGRADED ||
vd->vdev_ops == &vdev_hole_ops ||
vd->vdev_ops == &vdev_missing_ops);
}
boolean_t
vdev_readable(vdev_t *vd)
{
return (!vdev_is_dead(vd) && !vd->vdev_cant_read);
}
boolean_t
vdev_writeable(vdev_t *vd)
{
return (!vdev_is_dead(vd) && !vd->vdev_cant_write &&
vdev_is_concrete(vd));
}
boolean_t
vdev_allocatable(vdev_t *vd)
{
uint64_t state = vd->vdev_state;
/*
* We currently allow allocations from vdevs which may be in the
* process of reopening (i.e. VDEV_STATE_CLOSED). If the device
* fails to reopen then we'll catch it later when we're holding
* the proper locks. Note that we have to get the vdev state
* in a local variable because although it changes atomically,
* we're asking two separate questions about it.
*/
return (!(state < VDEV_STATE_DEGRADED && state != VDEV_STATE_CLOSED) &&
!vd->vdev_cant_write && vdev_is_concrete(vd) &&
vd->vdev_mg->mg_initialized);
}
boolean_t
vdev_accessible(vdev_t *vd, zio_t *zio)
{
ASSERT(zio->io_vd == vd);
if (vdev_is_dead(vd) || vd->vdev_remove_wanted)
return (B_FALSE);
if (zio->io_type == ZIO_TYPE_READ)
return (!vd->vdev_cant_read);
if (zio->io_type == ZIO_TYPE_WRITE)
return (!vd->vdev_cant_write);
return (B_TRUE);
}
static void
vdev_get_child_stat(vdev_t *cvd, vdev_stat_t *vs, vdev_stat_t *cvs)
{
/*
* Exclude the dRAID spare when aggregating to avoid double counting
* the ops and bytes. These IOs are counted by the physical leaves.
*/
if (cvd->vdev_ops == &vdev_draid_spare_ops)
return;
for (int t = 0; t < VS_ZIO_TYPES; t++) {
vs->vs_ops[t] += cvs->vs_ops[t];
vs->vs_bytes[t] += cvs->vs_bytes[t];
}
cvs->vs_scan_removing = cvd->vdev_removing;
}
/*
* Get extended stats
*/
static void
vdev_get_child_stat_ex(vdev_t *cvd, vdev_stat_ex_t *vsx, vdev_stat_ex_t *cvsx)
{
(void) cvd;
int t, b;
for (t = 0; t < ZIO_TYPES; t++) {
for (b = 0; b < ARRAY_SIZE(vsx->vsx_disk_histo[0]); b++)
vsx->vsx_disk_histo[t][b] += cvsx->vsx_disk_histo[t][b];
for (b = 0; b < ARRAY_SIZE(vsx->vsx_total_histo[0]); b++) {
vsx->vsx_total_histo[t][b] +=
cvsx->vsx_total_histo[t][b];
}
}
for (t = 0; t < ZIO_PRIORITY_NUM_QUEUEABLE; t++) {
for (b = 0; b < ARRAY_SIZE(vsx->vsx_queue_histo[0]); b++) {
vsx->vsx_queue_histo[t][b] +=
cvsx->vsx_queue_histo[t][b];
}
vsx->vsx_active_queue[t] += cvsx->vsx_active_queue[t];
vsx->vsx_pend_queue[t] += cvsx->vsx_pend_queue[t];
for (b = 0; b < ARRAY_SIZE(vsx->vsx_ind_histo[0]); b++)
vsx->vsx_ind_histo[t][b] += cvsx->vsx_ind_histo[t][b];
for (b = 0; b < ARRAY_SIZE(vsx->vsx_agg_histo[0]); b++)
vsx->vsx_agg_histo[t][b] += cvsx->vsx_agg_histo[t][b];
}
}
boolean_t
vdev_is_spacemap_addressable(vdev_t *vd)
{
if (spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_SPACEMAP_V2))
return (B_TRUE);
/*
* If double-word space map entries are not enabled we assume
* 47 bits of the space map entry are dedicated to the entry's
* offset (see SM_OFFSET_BITS in space_map.h). We then use that
* to calculate the maximum address that can be described by a
* space map entry for the given device.
*/
uint64_t shift = vd->vdev_ashift + SM_OFFSET_BITS;
if (shift >= 63) /* detect potential overflow */
return (B_TRUE);
return (vd->vdev_asize < (1ULL << shift));
}
/*
* Get statistics for the given vdev.
*/
static void
vdev_get_stats_ex_impl(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx)
{
int t;
/*
* If we're getting stats on the root vdev, aggregate the I/O counts
* over all top-level vdevs (i.e. the direct children of the root).
*/
if (!vd->vdev_ops->vdev_op_leaf) {
if (vs) {
memset(vs->vs_ops, 0, sizeof (vs->vs_ops));
memset(vs->vs_bytes, 0, sizeof (vs->vs_bytes));
}
if (vsx)
memset(vsx, 0, sizeof (*vsx));
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
vdev_stat_t *cvs = &cvd->vdev_stat;
vdev_stat_ex_t *cvsx = &cvd->vdev_stat_ex;
vdev_get_stats_ex_impl(cvd, cvs, cvsx);
if (vs)
vdev_get_child_stat(cvd, vs, cvs);
if (vsx)
vdev_get_child_stat_ex(cvd, vsx, cvsx);
}
} else {
/*
* We're a leaf. Just copy our ZIO active queue stats in. The
* other leaf stats are updated in vdev_stat_update().
*/
if (!vsx)
return;
memcpy(vsx, &vd->vdev_stat_ex, sizeof (vd->vdev_stat_ex));
for (t = 0; t < ARRAY_SIZE(vd->vdev_queue.vq_class); t++) {
vsx->vsx_active_queue[t] =
vd->vdev_queue.vq_class[t].vqc_active;
vsx->vsx_pend_queue[t] = avl_numnodes(
&vd->vdev_queue.vq_class[t].vqc_queued_tree);
}
}
}
void
vdev_get_stats_ex(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx)
{
vdev_t *tvd = vd->vdev_top;
mutex_enter(&vd->vdev_stat_lock);
if (vs) {
memcpy(vs, &vd->vdev_stat, sizeof (*vs));
vs->vs_timestamp = gethrtime() - vs->vs_timestamp;
vs->vs_state = vd->vdev_state;
vs->vs_rsize = vdev_get_min_asize(vd);
if (vd->vdev_ops->vdev_op_leaf) {
vs->vs_pspace = vd->vdev_psize;
vs->vs_rsize += VDEV_LABEL_START_SIZE +
VDEV_LABEL_END_SIZE;
/*
* Report initializing progress. Since we don't
* have the initializing locks held, this is only
* an estimate (although a fairly accurate one).
*/
vs->vs_initialize_bytes_done =
vd->vdev_initialize_bytes_done;
vs->vs_initialize_bytes_est =
vd->vdev_initialize_bytes_est;
vs->vs_initialize_state = vd->vdev_initialize_state;
vs->vs_initialize_action_time =
vd->vdev_initialize_action_time;
/*
* Report manual TRIM progress. Since we don't have
* the manual TRIM locks held, this is only an
* estimate (although fairly accurate one).
*/
vs->vs_trim_notsup = !vd->vdev_has_trim;
vs->vs_trim_bytes_done = vd->vdev_trim_bytes_done;
vs->vs_trim_bytes_est = vd->vdev_trim_bytes_est;
vs->vs_trim_state = vd->vdev_trim_state;
vs->vs_trim_action_time = vd->vdev_trim_action_time;
/* Set when there is a deferred resilver. */
vs->vs_resilver_deferred = vd->vdev_resilver_deferred;
}
/*
* Report expandable space on top-level, non-auxiliary devices
* only. The expandable space is reported in terms of metaslab
* sized units since that determines how much space the pool
* can expand.
*/
if (vd->vdev_aux == NULL && tvd != NULL) {
vs->vs_esize = P2ALIGN(
vd->vdev_max_asize - vd->vdev_asize,
1ULL << tvd->vdev_ms_shift);
}
vs->vs_configured_ashift = vd->vdev_top != NULL
? vd->vdev_top->vdev_ashift : vd->vdev_ashift;
vs->vs_logical_ashift = vd->vdev_logical_ashift;
if (vd->vdev_physical_ashift <= ASHIFT_MAX)
vs->vs_physical_ashift = vd->vdev_physical_ashift;
else
vs->vs_physical_ashift = 0;
/*
* Report fragmentation and rebuild progress for top-level,
* non-auxiliary, concrete devices.
*/
if (vd->vdev_aux == NULL && vd == vd->vdev_top &&
vdev_is_concrete(vd)) {
/*
* The vdev fragmentation rating doesn't take into
* account the embedded slog metaslab (vdev_log_mg).
* Since it's only one metaslab, it would have a tiny
* impact on the overall fragmentation.
*/
vs->vs_fragmentation = (vd->vdev_mg != NULL) ?
vd->vdev_mg->mg_fragmentation : 0;
}
vs->vs_noalloc = MAX(vd->vdev_noalloc,
tvd ? tvd->vdev_noalloc : 0);
}
vdev_get_stats_ex_impl(vd, vs, vsx);
mutex_exit(&vd->vdev_stat_lock);
}
void
vdev_get_stats(vdev_t *vd, vdev_stat_t *vs)
{
return (vdev_get_stats_ex(vd, vs, NULL));
}
void
vdev_clear_stats(vdev_t *vd)
{
mutex_enter(&vd->vdev_stat_lock);
vd->vdev_stat.vs_space = 0;
vd->vdev_stat.vs_dspace = 0;
vd->vdev_stat.vs_alloc = 0;
mutex_exit(&vd->vdev_stat_lock);
}
void
vdev_scan_stat_init(vdev_t *vd)
{
vdev_stat_t *vs = &vd->vdev_stat;
for (int c = 0; c < vd->vdev_children; c++)
vdev_scan_stat_init(vd->vdev_child[c]);
mutex_enter(&vd->vdev_stat_lock);
vs->vs_scan_processed = 0;
mutex_exit(&vd->vdev_stat_lock);
}
void
vdev_stat_update(zio_t *zio, uint64_t psize)
{
spa_t *spa = zio->io_spa;
vdev_t *rvd = spa->spa_root_vdev;
vdev_t *vd = zio->io_vd ? zio->io_vd : rvd;
vdev_t *pvd;
uint64_t txg = zio->io_txg;
/* Suppress ASAN false positive */
#ifdef __SANITIZE_ADDRESS__
vdev_stat_t *vs = vd ? &vd->vdev_stat : NULL;
vdev_stat_ex_t *vsx = vd ? &vd->vdev_stat_ex : NULL;
#else
vdev_stat_t *vs = &vd->vdev_stat;
vdev_stat_ex_t *vsx = &vd->vdev_stat_ex;
#endif
zio_type_t type = zio->io_type;
int flags = zio->io_flags;
/*
* If this i/o is a gang leader, it didn't do any actual work.
*/
if (zio->io_gang_tree)
return;
if (zio->io_error == 0) {
/*
* If this is a root i/o, don't count it -- we've already
* counted the top-level vdevs, and vdev_get_stats() will
* aggregate them when asked. This reduces contention on
* the root vdev_stat_lock and implicitly handles blocks
* that compress away to holes, for which there is no i/o.
* (Holes never create vdev children, so all the counters
* remain zero, which is what we want.)
*
* Note: this only applies to successful i/o (io_error == 0)
* because unlike i/o counts, errors are not additive.
* When reading a ditto block, for example, failure of
* one top-level vdev does not imply a root-level error.
*/
if (vd == rvd)
return;
ASSERT(vd == zio->io_vd);
if (flags & ZIO_FLAG_IO_BYPASS)
return;
mutex_enter(&vd->vdev_stat_lock);
if (flags & ZIO_FLAG_IO_REPAIR) {
/*
* Repair is the result of a resilver issued by the
* scan thread (spa_sync).
*/
if (flags & ZIO_FLAG_SCAN_THREAD) {
dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan;
dsl_scan_phys_t *scn_phys = &scn->scn_phys;
uint64_t *processed = &scn_phys->scn_processed;
if (vd->vdev_ops->vdev_op_leaf)
atomic_add_64(processed, psize);
vs->vs_scan_processed += psize;
}
/*
* Repair is the result of a rebuild issued by the
* rebuild thread (vdev_rebuild_thread). To avoid
* double counting repaired bytes the virtual dRAID
* spare vdev is excluded from the processed bytes.
*/
if (zio->io_priority == ZIO_PRIORITY_REBUILD) {
vdev_t *tvd = vd->vdev_top;
vdev_rebuild_t *vr = &tvd->vdev_rebuild_config;
vdev_rebuild_phys_t *vrp = &vr->vr_rebuild_phys;
uint64_t *rebuilt = &vrp->vrp_bytes_rebuilt;
if (vd->vdev_ops->vdev_op_leaf &&
vd->vdev_ops != &vdev_draid_spare_ops) {
atomic_add_64(rebuilt, psize);
}
vs->vs_rebuild_processed += psize;
}
if (flags & ZIO_FLAG_SELF_HEAL)
vs->vs_self_healed += psize;
}
/*
* The bytes/ops/histograms are recorded at the leaf level and
* aggregated into the higher level vdevs in vdev_get_stats().
*/
if (vd->vdev_ops->vdev_op_leaf &&
(zio->io_priority < ZIO_PRIORITY_NUM_QUEUEABLE)) {
zio_type_t vs_type = type;
zio_priority_t priority = zio->io_priority;
/*
* TRIM ops and bytes are reported to user space as
* ZIO_TYPE_IOCTL. This is done to preserve the
* vdev_stat_t structure layout for user space.
*/
if (type == ZIO_TYPE_TRIM)
vs_type = ZIO_TYPE_IOCTL;
/*
* Solely for the purposes of 'zpool iostat -lqrw'
* reporting use the priority to categorize the IO.
* Only the following are reported to user space:
*
* ZIO_PRIORITY_SYNC_READ,
* ZIO_PRIORITY_SYNC_WRITE,
* ZIO_PRIORITY_ASYNC_READ,
* ZIO_PRIORITY_ASYNC_WRITE,
* ZIO_PRIORITY_SCRUB,
* ZIO_PRIORITY_TRIM,
* ZIO_PRIORITY_REBUILD.
*/
if (priority == ZIO_PRIORITY_INITIALIZING) {
ASSERT3U(type, ==, ZIO_TYPE_WRITE);
priority = ZIO_PRIORITY_ASYNC_WRITE;
} else if (priority == ZIO_PRIORITY_REMOVAL) {
priority = ((type == ZIO_TYPE_WRITE) ?
ZIO_PRIORITY_ASYNC_WRITE :
ZIO_PRIORITY_ASYNC_READ);
}
vs->vs_ops[vs_type]++;
vs->vs_bytes[vs_type] += psize;
if (flags & ZIO_FLAG_DELEGATED) {
vsx->vsx_agg_histo[priority]
[RQ_HISTO(zio->io_size)]++;
} else {
vsx->vsx_ind_histo[priority]
[RQ_HISTO(zio->io_size)]++;
}
if (zio->io_delta && zio->io_delay) {
vsx->vsx_queue_histo[priority]
[L_HISTO(zio->io_delta - zio->io_delay)]++;
vsx->vsx_disk_histo[type]
[L_HISTO(zio->io_delay)]++;
vsx->vsx_total_histo[type]
[L_HISTO(zio->io_delta)]++;
}
}
mutex_exit(&vd->vdev_stat_lock);
return;
}
if (flags & ZIO_FLAG_SPECULATIVE)
return;
/*
* If this is an I/O error that is going to be retried, then ignore the
* error. Otherwise, the user may interpret B_FAILFAST I/O errors as
* hard errors, when in reality they can happen for any number of
* innocuous reasons (bus resets, MPxIO link failure, etc).
*/
if (zio->io_error == EIO &&
!(zio->io_flags & ZIO_FLAG_IO_RETRY))
return;
/*
* Intent logs writes won't propagate their error to the root
* I/O so don't mark these types of failures as pool-level
* errors.
*/
if (zio->io_vd == NULL && (zio->io_flags & ZIO_FLAG_DONT_PROPAGATE))
return;
if (type == ZIO_TYPE_WRITE && txg != 0 &&
(!(flags & ZIO_FLAG_IO_REPAIR) ||
(flags & ZIO_FLAG_SCAN_THREAD) ||
spa->spa_claiming)) {
/*
* This is either a normal write (not a repair), or it's
* a repair induced by the scrub thread, or it's a repair
* made by zil_claim() during spa_load() in the first txg.
* In the normal case, we commit the DTL change in the same
* txg as the block was born. In the scrub-induced repair
* case, we know that scrubs run in first-pass syncing context,
* so we commit the DTL change in spa_syncing_txg(spa).
* In the zil_claim() case, we commit in spa_first_txg(spa).
*
* We currently do not make DTL entries for failed spontaneous
* self-healing writes triggered by normal (non-scrubbing)
* reads, because we have no transactional context in which to
* do so -- and it's not clear that it'd be desirable anyway.
*/
if (vd->vdev_ops->vdev_op_leaf) {
uint64_t commit_txg = txg;
if (flags & ZIO_FLAG_SCAN_THREAD) {
ASSERT(flags & ZIO_FLAG_IO_REPAIR);
ASSERT(spa_sync_pass(spa) == 1);
vdev_dtl_dirty(vd, DTL_SCRUB, txg, 1);
commit_txg = spa_syncing_txg(spa);
} else if (spa->spa_claiming) {
ASSERT(flags & ZIO_FLAG_IO_REPAIR);
commit_txg = spa_first_txg(spa);
}
ASSERT(commit_txg >= spa_syncing_txg(spa));
if (vdev_dtl_contains(vd, DTL_MISSING, txg, 1))
return;
for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent)
vdev_dtl_dirty(pvd, DTL_PARTIAL, txg, 1);
vdev_dirty(vd->vdev_top, VDD_DTL, vd, commit_txg);
}
if (vd != rvd)
vdev_dtl_dirty(vd, DTL_MISSING, txg, 1);
}
}
int64_t
vdev_deflated_space(vdev_t *vd, int64_t space)
{
ASSERT((space & (SPA_MINBLOCKSIZE-1)) == 0);
ASSERT(vd->vdev_deflate_ratio != 0 || vd->vdev_isl2cache);
return ((space >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio);
}
/*
* Update the in-core space usage stats for this vdev, its metaslab class,
* and the root vdev.
*/
void
vdev_space_update(vdev_t *vd, int64_t alloc_delta, int64_t defer_delta,
int64_t space_delta)
{
(void) defer_delta;
int64_t dspace_delta;
spa_t *spa = vd->vdev_spa;
vdev_t *rvd = spa->spa_root_vdev;
ASSERT(vd == vd->vdev_top);
/*
* Apply the inverse of the psize-to-asize (ie. RAID-Z) space-expansion
* factor. We must calculate this here and not at the root vdev
* because the root vdev's psize-to-asize is simply the max of its
* children's, thus not accurate enough for us.
*/
dspace_delta = vdev_deflated_space(vd, space_delta);
mutex_enter(&vd->vdev_stat_lock);
/* ensure we won't underflow */
if (alloc_delta < 0) {
ASSERT3U(vd->vdev_stat.vs_alloc, >=, -alloc_delta);
}
vd->vdev_stat.vs_alloc += alloc_delta;
vd->vdev_stat.vs_space += space_delta;
vd->vdev_stat.vs_dspace += dspace_delta;
mutex_exit(&vd->vdev_stat_lock);
/* every class but log contributes to root space stats */
if (vd->vdev_mg != NULL && !vd->vdev_islog) {
ASSERT(!vd->vdev_isl2cache);
mutex_enter(&rvd->vdev_stat_lock);
rvd->vdev_stat.vs_alloc += alloc_delta;
rvd->vdev_stat.vs_space += space_delta;
rvd->vdev_stat.vs_dspace += dspace_delta;
mutex_exit(&rvd->vdev_stat_lock);
}
/* Note: metaslab_class_space_update moved to metaslab_space_update */
}
/*
* Mark a top-level vdev's config as dirty, placing it on the dirty list
* so that it will be written out next time the vdev configuration is synced.
* If the root vdev is specified (vdev_top == NULL), dirty all top-level vdevs.
*/
void
vdev_config_dirty(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
vdev_t *rvd = spa->spa_root_vdev;
int c;
ASSERT(spa_writeable(spa));
/*
* If this is an aux vdev (as with l2cache and spare devices), then we
* update the vdev config manually and set the sync flag.
*/
if (vd->vdev_aux != NULL) {
spa_aux_vdev_t *sav = vd->vdev_aux;
nvlist_t **aux;
uint_t naux;
for (c = 0; c < sav->sav_count; c++) {
if (sav->sav_vdevs[c] == vd)
break;
}
if (c == sav->sav_count) {
/*
* We're being removed. There's nothing more to do.
*/
ASSERT(sav->sav_sync == B_TRUE);
return;
}
sav->sav_sync = B_TRUE;
if (nvlist_lookup_nvlist_array(sav->sav_config,
ZPOOL_CONFIG_L2CACHE, &aux, &naux) != 0) {
VERIFY(nvlist_lookup_nvlist_array(sav->sav_config,
ZPOOL_CONFIG_SPARES, &aux, &naux) == 0);
}
ASSERT(c < naux);
/*
* Setting the nvlist in the middle if the array is a little
* sketchy, but it will work.
*/
nvlist_free(aux[c]);
aux[c] = vdev_config_generate(spa, vd, B_TRUE, 0);
return;
}
/*
* The dirty list is protected by the SCL_CONFIG lock. The caller
* must either hold SCL_CONFIG as writer, or must be the sync thread
* (which holds SCL_CONFIG as reader). There's only one sync thread,
* so this is sufficient to ensure mutual exclusion.
*/
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) ||
(dsl_pool_sync_context(spa_get_dsl(spa)) &&
spa_config_held(spa, SCL_CONFIG, RW_READER)));
if (vd == rvd) {
for (c = 0; c < rvd->vdev_children; c++)
vdev_config_dirty(rvd->vdev_child[c]);
} else {
ASSERT(vd == vd->vdev_top);
if (!list_link_active(&vd->vdev_config_dirty_node) &&
vdev_is_concrete(vd)) {
list_insert_head(&spa->spa_config_dirty_list, vd);
}
}
}
void
vdev_config_clean(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) ||
(dsl_pool_sync_context(spa_get_dsl(spa)) &&
spa_config_held(spa, SCL_CONFIG, RW_READER)));
ASSERT(list_link_active(&vd->vdev_config_dirty_node));
list_remove(&spa->spa_config_dirty_list, vd);
}
/*
* Mark a top-level vdev's state as dirty, so that the next pass of
* spa_sync() can convert this into vdev_config_dirty(). We distinguish
* the state changes from larger config changes because they require
* much less locking, and are often needed for administrative actions.
*/
void
vdev_state_dirty(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_writeable(spa));
ASSERT(vd == vd->vdev_top);
/*
* The state list is protected by the SCL_STATE lock. The caller
* must either hold SCL_STATE as writer, or must be the sync thread
* (which holds SCL_STATE as reader). There's only one sync thread,
* so this is sufficient to ensure mutual exclusion.
*/
ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) ||
(dsl_pool_sync_context(spa_get_dsl(spa)) &&
spa_config_held(spa, SCL_STATE, RW_READER)));
if (!list_link_active(&vd->vdev_state_dirty_node) &&
vdev_is_concrete(vd))
list_insert_head(&spa->spa_state_dirty_list, vd);
}
void
vdev_state_clean(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) ||
(dsl_pool_sync_context(spa_get_dsl(spa)) &&
spa_config_held(spa, SCL_STATE, RW_READER)));
ASSERT(list_link_active(&vd->vdev_state_dirty_node));
list_remove(&spa->spa_state_dirty_list, vd);
}
/*
* Propagate vdev state up from children to parent.
*/
void
vdev_propagate_state(vdev_t *vd)
{
spa_t *spa = vd->vdev_spa;
vdev_t *rvd = spa->spa_root_vdev;
int degraded = 0, faulted = 0;
int corrupted = 0;
vdev_t *child;
if (vd->vdev_children > 0) {
for (int c = 0; c < vd->vdev_children; c++) {
child = vd->vdev_child[c];
/*
* Don't factor holes or indirect vdevs into the
* decision.
*/
if (!vdev_is_concrete(child))
continue;
if (!vdev_readable(child) ||
(!vdev_writeable(child) && spa_writeable(spa))) {
/*
* Root special: if there is a top-level log
* device, treat the root vdev as if it were
* degraded.
*/
if (child->vdev_islog && vd == rvd)
degraded++;
else
faulted++;
} else if (child->vdev_state <= VDEV_STATE_DEGRADED) {
degraded++;
}
if (child->vdev_stat.vs_aux == VDEV_AUX_CORRUPT_DATA)
corrupted++;
}
vd->vdev_ops->vdev_op_state_change(vd, faulted, degraded);
/*
* Root special: if there is a top-level vdev that cannot be
* opened due to corrupted metadata, then propagate the root
* vdev's aux state as 'corrupt' rather than 'insufficient
* replicas'.
*/
if (corrupted && vd == rvd &&
rvd->vdev_state == VDEV_STATE_CANT_OPEN)
vdev_set_state(rvd, B_FALSE, VDEV_STATE_CANT_OPEN,
VDEV_AUX_CORRUPT_DATA);
}
if (vd->vdev_parent)
vdev_propagate_state(vd->vdev_parent);
}
/*
* Set a vdev's state. If this is during an open, we don't update the parent
* state, because we're in the process of opening children depth-first.
* Otherwise, we propagate the change to the parent.
*
* If this routine places a device in a faulted state, an appropriate ereport is
* generated.
*/
void
vdev_set_state(vdev_t *vd, boolean_t isopen, vdev_state_t state, vdev_aux_t aux)
{
uint64_t save_state;
spa_t *spa = vd->vdev_spa;
if (state == vd->vdev_state) {
/*
* Since vdev_offline() code path is already in an offline
* state we can miss a statechange event to OFFLINE. Check
* the previous state to catch this condition.
*/
if (vd->vdev_ops->vdev_op_leaf &&
(state == VDEV_STATE_OFFLINE) &&
(vd->vdev_prevstate >= VDEV_STATE_FAULTED)) {
/* post an offline state change */
zfs_post_state_change(spa, vd, vd->vdev_prevstate);
}
vd->vdev_stat.vs_aux = aux;
return;
}
save_state = vd->vdev_state;
vd->vdev_state = state;
vd->vdev_stat.vs_aux = aux;
/*
* If we are setting the vdev state to anything but an open state, then
* always close the underlying device unless the device has requested
* a delayed close (i.e. we're about to remove or fault the device).
* Otherwise, we keep accessible but invalid devices open forever.
* We don't call vdev_close() itself, because that implies some extra
* checks (offline, etc) that we don't want here. This is limited to
* leaf devices, because otherwise closing the device will affect other
* children.
*/
if (!vd->vdev_delayed_close && vdev_is_dead(vd) &&
vd->vdev_ops->vdev_op_leaf)
vd->vdev_ops->vdev_op_close(vd);
if (vd->vdev_removed &&
state == VDEV_STATE_CANT_OPEN &&
(aux == VDEV_AUX_OPEN_FAILED || vd->vdev_checkremove)) {
/*
* If the previous state is set to VDEV_STATE_REMOVED, then this
* device was previously marked removed and someone attempted to
* reopen it. If this failed due to a nonexistent device, then
* keep the device in the REMOVED state. We also let this be if
* it is one of our special test online cases, which is only
* attempting to online the device and shouldn't generate an FMA
* fault.
*/
vd->vdev_state = VDEV_STATE_REMOVED;
vd->vdev_stat.vs_aux = VDEV_AUX_NONE;
} else if (state == VDEV_STATE_REMOVED) {
vd->vdev_removed = B_TRUE;
} else if (state == VDEV_STATE_CANT_OPEN) {
/*
* If we fail to open a vdev during an import or recovery, we
* mark it as "not available", which signifies that it was
* never there to begin with. Failure to open such a device
* is not considered an error.
*/
if ((spa_load_state(spa) == SPA_LOAD_IMPORT ||
spa_load_state(spa) == SPA_LOAD_RECOVER) &&
vd->vdev_ops->vdev_op_leaf)
vd->vdev_not_present = 1;
/*
* Post the appropriate ereport. If the 'prevstate' field is
* set to something other than VDEV_STATE_UNKNOWN, it indicates
* that this is part of a vdev_reopen(). In this case, we don't
* want to post the ereport if the device was already in the
* CANT_OPEN state beforehand.
*
* If the 'checkremove' flag is set, then this is an attempt to
* online the device in response to an insertion event. If we
* hit this case, then we have detected an insertion event for a
* faulted or offline device that wasn't in the removed state.
* In this scenario, we don't post an ereport because we are
* about to replace the device, or attempt an online with
* vdev_forcefault, which will generate the fault for us.
*/
if ((vd->vdev_prevstate != state || vd->vdev_forcefault) &&
!vd->vdev_not_present && !vd->vdev_checkremove &&
vd != spa->spa_root_vdev) {
const char *class;
switch (aux) {
case VDEV_AUX_OPEN_FAILED:
class = FM_EREPORT_ZFS_DEVICE_OPEN_FAILED;
break;
case VDEV_AUX_CORRUPT_DATA:
class = FM_EREPORT_ZFS_DEVICE_CORRUPT_DATA;
break;
case VDEV_AUX_NO_REPLICAS:
class = FM_EREPORT_ZFS_DEVICE_NO_REPLICAS;
break;
case VDEV_AUX_BAD_GUID_SUM:
class = FM_EREPORT_ZFS_DEVICE_BAD_GUID_SUM;
break;
case VDEV_AUX_TOO_SMALL:
class = FM_EREPORT_ZFS_DEVICE_TOO_SMALL;
break;
case VDEV_AUX_BAD_LABEL:
class = FM_EREPORT_ZFS_DEVICE_BAD_LABEL;
break;
case VDEV_AUX_BAD_ASHIFT:
class = FM_EREPORT_ZFS_DEVICE_BAD_ASHIFT;
break;
default:
class = FM_EREPORT_ZFS_DEVICE_UNKNOWN;
}
(void) zfs_ereport_post(class, spa, vd, NULL, NULL,
save_state);
}
/* Erase any notion of persistent removed state */
vd->vdev_removed = B_FALSE;
} else {
vd->vdev_removed = B_FALSE;
}
/*
* Notify ZED of any significant state-change on a leaf vdev.
*
*/
if (vd->vdev_ops->vdev_op_leaf) {
/* preserve original state from a vdev_reopen() */
if ((vd->vdev_prevstate != VDEV_STATE_UNKNOWN) &&
(vd->vdev_prevstate != vd->vdev_state) &&
(save_state <= VDEV_STATE_CLOSED))
save_state = vd->vdev_prevstate;
/* filter out state change due to initial vdev_open */
if (save_state > VDEV_STATE_CLOSED)
zfs_post_state_change(spa, vd, save_state);
}
if (!isopen && vd->vdev_parent)
vdev_propagate_state(vd->vdev_parent);
}
boolean_t
vdev_children_are_offline(vdev_t *vd)
{
ASSERT(!vd->vdev_ops->vdev_op_leaf);
for (uint64_t i = 0; i < vd->vdev_children; i++) {
if (vd->vdev_child[i]->vdev_state != VDEV_STATE_OFFLINE)
return (B_FALSE);
}
return (B_TRUE);
}
/*
* Check the vdev configuration to ensure that it's capable of supporting
* a root pool. We do not support partial configuration.
*/
boolean_t
vdev_is_bootable(vdev_t *vd)
{
if (!vd->vdev_ops->vdev_op_leaf) {
const char *vdev_type = vd->vdev_ops->vdev_op_type;
if (strcmp(vdev_type, VDEV_TYPE_MISSING) == 0)
return (B_FALSE);
}
for (int c = 0; c < vd->vdev_children; c++) {
if (!vdev_is_bootable(vd->vdev_child[c]))
return (B_FALSE);
}
return (B_TRUE);
}
boolean_t
vdev_is_concrete(vdev_t *vd)
{
vdev_ops_t *ops = vd->vdev_ops;
if (ops == &vdev_indirect_ops || ops == &vdev_hole_ops ||
ops == &vdev_missing_ops || ops == &vdev_root_ops) {
return (B_FALSE);
} else {
return (B_TRUE);
}
}
/*
* Determine if a log device has valid content. If the vdev was
* removed or faulted in the MOS config then we know that
* the content on the log device has already been written to the pool.
*/
boolean_t
vdev_log_state_valid(vdev_t *vd)
{
if (vd->vdev_ops->vdev_op_leaf && !vd->vdev_faulted &&
!vd->vdev_removed)
return (B_TRUE);
for (int c = 0; c < vd->vdev_children; c++)
if (vdev_log_state_valid(vd->vdev_child[c]))
return (B_TRUE);
return (B_FALSE);
}
/*
* Expand a vdev if possible.
*/
void
vdev_expand(vdev_t *vd, uint64_t txg)
{
ASSERT(vd->vdev_top == vd);
ASSERT(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL);
ASSERT(vdev_is_concrete(vd));
vdev_set_deflate_ratio(vd);
if ((vd->vdev_asize >> vd->vdev_ms_shift) > vd->vdev_ms_count &&
vdev_is_concrete(vd)) {
vdev_metaslab_group_create(vd);
VERIFY(vdev_metaslab_init(vd, txg) == 0);
vdev_config_dirty(vd);
}
}
/*
* Split a vdev.
*/
void
vdev_split(vdev_t *vd)
{
vdev_t *cvd, *pvd = vd->vdev_parent;
VERIFY3U(pvd->vdev_children, >, 1);
vdev_remove_child(pvd, vd);
vdev_compact_children(pvd);
ASSERT3P(pvd->vdev_child, !=, NULL);
cvd = pvd->vdev_child[0];
if (pvd->vdev_children == 1) {
vdev_remove_parent(cvd);
cvd->vdev_splitting = B_TRUE;
}
vdev_propagate_state(cvd);
}
void
vdev_deadman(vdev_t *vd, const char *tag)
{
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
vdev_deadman(cvd, tag);
}
if (vd->vdev_ops->vdev_op_leaf) {
vdev_queue_t *vq = &vd->vdev_queue;
mutex_enter(&vq->vq_lock);
if (avl_numnodes(&vq->vq_active_tree) > 0) {
spa_t *spa = vd->vdev_spa;
zio_t *fio;
uint64_t delta;
zfs_dbgmsg("slow vdev: %s has %lu active IOs",
vd->vdev_path, avl_numnodes(&vq->vq_active_tree));
/*
* Look at the head of all the pending queues,
* if any I/O has been outstanding for longer than
* the spa_deadman_synctime invoke the deadman logic.
*/
fio = avl_first(&vq->vq_active_tree);
delta = gethrtime() - fio->io_timestamp;
if (delta > spa_deadman_synctime(spa))
zio_deadman(fio, tag);
}
mutex_exit(&vq->vq_lock);
}
}
void
vdev_defer_resilver(vdev_t *vd)
{
ASSERT(vd->vdev_ops->vdev_op_leaf);
vd->vdev_resilver_deferred = B_TRUE;
vd->vdev_spa->spa_resilver_deferred = B_TRUE;
}
/*
* Clears the resilver deferred flag on all leaf devs under vd. Returns
* B_TRUE if we have devices that need to be resilvered and are available to
* accept resilver I/Os.
*/
boolean_t
vdev_clear_resilver_deferred(vdev_t *vd, dmu_tx_t *tx)
{
boolean_t resilver_needed = B_FALSE;
spa_t *spa = vd->vdev_spa;
for (int c = 0; c < vd->vdev_children; c++) {
vdev_t *cvd = vd->vdev_child[c];
resilver_needed |= vdev_clear_resilver_deferred(cvd, tx);
}
if (vd == spa->spa_root_vdev &&
spa_feature_is_active(spa, SPA_FEATURE_RESILVER_DEFER)) {
spa_feature_decr(spa, SPA_FEATURE_RESILVER_DEFER, tx);
vdev_config_dirty(vd);
spa->spa_resilver_deferred = B_FALSE;
return (resilver_needed);
}
if (!vdev_is_concrete(vd) || vd->vdev_aux ||
!vd->vdev_ops->vdev_op_leaf)
return (resilver_needed);
vd->vdev_resilver_deferred = B_FALSE;
return (!vdev_is_dead(vd) && !vd->vdev_offline &&
vdev_resilver_needed(vd, NULL, NULL));
}
boolean_t
vdev_xlate_is_empty(range_seg64_t *rs)
{
return (rs->rs_start == rs->rs_end);
}
/*
* Translate a logical range to the first contiguous physical range for the
* specified vdev_t. This function is initially called with a leaf vdev and
* will walk each parent vdev until it reaches a top-level vdev. Once the
* top-level is reached the physical range is initialized and the recursive
* function begins to unwind. As it unwinds it calls the parent's vdev
* specific translation function to do the real conversion.
*/
void
vdev_xlate(vdev_t *vd, const range_seg64_t *logical_rs,
range_seg64_t *physical_rs, range_seg64_t *remain_rs)
{
/*
* Walk up the vdev tree
*/
if (vd != vd->vdev_top) {
vdev_xlate(vd->vdev_parent, logical_rs, physical_rs,
remain_rs);
} else {
/*
* We've reached the top-level vdev, initialize the physical
* range to the logical range and set an empty remaining
* range then start to unwind.
*/
physical_rs->rs_start = logical_rs->rs_start;
physical_rs->rs_end = logical_rs->rs_end;
remain_rs->rs_start = logical_rs->rs_start;
remain_rs->rs_end = logical_rs->rs_start;
return;
}
vdev_t *pvd = vd->vdev_parent;
ASSERT3P(pvd, !=, NULL);
ASSERT3P(pvd->vdev_ops->vdev_op_xlate, !=, NULL);
/*
* As this recursive function unwinds, translate the logical
* range into its physical and any remaining components by calling
* the vdev specific translate function.
*/
range_seg64_t intermediate = { 0 };
pvd->vdev_ops->vdev_op_xlate(vd, physical_rs, &intermediate, remain_rs);
physical_rs->rs_start = intermediate.rs_start;
physical_rs->rs_end = intermediate.rs_end;
}
void
vdev_xlate_walk(vdev_t *vd, const range_seg64_t *logical_rs,
vdev_xlate_func_t *func, void *arg)
{
range_seg64_t iter_rs = *logical_rs;
range_seg64_t physical_rs;
range_seg64_t remain_rs;
while (!vdev_xlate_is_empty(&iter_rs)) {
vdev_xlate(vd, &iter_rs, &physical_rs, &remain_rs);
/*
* With raidz and dRAID, it's possible that the logical range
* does not live on this leaf vdev. Only when there is a non-
* zero physical size call the provided function.
*/
if (!vdev_xlate_is_empty(&physical_rs))
func(arg, &physical_rs);
iter_rs = remain_rs;
}
}
static char *
vdev_name(vdev_t *vd, char *buf, int buflen)
{
if (vd->vdev_path == NULL) {
if (strcmp(vd->vdev_ops->vdev_op_type, "root") == 0) {
strlcpy(buf, vd->vdev_spa->spa_name, buflen);
} else if (!vd->vdev_ops->vdev_op_leaf) {
snprintf(buf, buflen, "%s-%llu",
vd->vdev_ops->vdev_op_type,
(u_longlong_t)vd->vdev_id);
}
} else {
strlcpy(buf, vd->vdev_path, buflen);
}
return (buf);
}
/*
* Look at the vdev tree and determine whether any devices are currently being
* replaced.
*/
boolean_t
vdev_replace_in_progress(vdev_t *vdev)
{
ASSERT(spa_config_held(vdev->vdev_spa, SCL_ALL, RW_READER) != 0);
if (vdev->vdev_ops == &vdev_replacing_ops)
return (B_TRUE);
/*
* A 'spare' vdev indicates that we have a replace in progress, unless
* it has exactly two children, and the second, the hot spare, has
* finished being resilvered.
*/
if (vdev->vdev_ops == &vdev_spare_ops && (vdev->vdev_children > 2 ||
!vdev_dtl_empty(vdev->vdev_child[1], DTL_MISSING)))
return (B_TRUE);
for (int i = 0; i < vdev->vdev_children; i++) {
if (vdev_replace_in_progress(vdev->vdev_child[i]))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Add a (source=src, propname=propval) list to an nvlist.
*/
static void
vdev_prop_add_list(nvlist_t *nvl, const char *propname, const char *strval,
uint64_t intval, zprop_source_t src)
{
nvlist_t *propval;
propval = fnvlist_alloc();
fnvlist_add_uint64(propval, ZPROP_SOURCE, src);
if (strval != NULL)
fnvlist_add_string(propval, ZPROP_VALUE, strval);
else
fnvlist_add_uint64(propval, ZPROP_VALUE, intval);
fnvlist_add_nvlist(nvl, propname, propval);
nvlist_free(propval);
}
static void
vdev_props_set_sync(void *arg, dmu_tx_t *tx)
{
vdev_t *vd;
nvlist_t *nvp = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
objset_t *mos = spa->spa_meta_objset;
nvpair_t *elem = NULL;
uint64_t vdev_guid;
nvlist_t *nvprops;
vdev_guid = fnvlist_lookup_uint64(nvp, ZPOOL_VDEV_PROPS_SET_VDEV);
nvprops = fnvlist_lookup_nvlist(nvp, ZPOOL_VDEV_PROPS_SET_PROPS);
vd = spa_lookup_by_guid(spa, vdev_guid, B_TRUE);
/* this vdev could get removed while waiting for this sync task */
if (vd == NULL)
return;
mutex_enter(&spa->spa_props_lock);
while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) {
uint64_t intval, objid = 0;
const char *strval;
vdev_prop_t prop;
const char *propname = nvpair_name(elem);
zprop_type_t proptype;
/*
* Set vdev property values in the vdev props mos object.
*/
- if (vd->vdev_top_zap != 0) {
+ if (vd->vdev_root_zap != 0) {
+ objid = vd->vdev_root_zap;
+ } else if (vd->vdev_top_zap != 0) {
objid = vd->vdev_top_zap;
} else if (vd->vdev_leaf_zap != 0) {
objid = vd->vdev_leaf_zap;
} else {
- panic("vdev not top or leaf");
+ /*
+ * XXX: implement vdev_props_set_check()
+ */
+ panic("vdev not root/top/leaf");
}
switch (prop = vdev_name_to_prop(propname)) {
case VDEV_PROP_USERPROP:
if (vdev_prop_user(propname)) {
strval = fnvpair_value_string(elem);
if (strlen(strval) == 0) {
/* remove the property if value == "" */
(void) zap_remove(mos, objid, propname,
tx);
} else {
VERIFY0(zap_update(mos, objid, propname,
1, strlen(strval) + 1, strval, tx));
}
spa_history_log_internal(spa, "vdev set", tx,
"vdev_guid=%llu: %s=%s",
(u_longlong_t)vdev_guid, nvpair_name(elem),
strval);
}
break;
default:
/* normalize the property name */
propname = vdev_prop_to_name(prop);
proptype = vdev_prop_get_type(prop);
if (nvpair_type(elem) == DATA_TYPE_STRING) {
ASSERT(proptype == PROP_TYPE_STRING);
strval = fnvpair_value_string(elem);
VERIFY0(zap_update(mos, objid, propname,
1, strlen(strval) + 1, strval, tx));
spa_history_log_internal(spa, "vdev set", tx,
"vdev_guid=%llu: %s=%s",
(u_longlong_t)vdev_guid, nvpair_name(elem),
strval);
} else if (nvpair_type(elem) == DATA_TYPE_UINT64) {
intval = fnvpair_value_uint64(elem);
if (proptype == PROP_TYPE_INDEX) {
const char *unused;
VERIFY0(vdev_prop_index_to_string(
prop, intval, &unused));
}
VERIFY0(zap_update(mos, objid, propname,
sizeof (uint64_t), 1, &intval, tx));
spa_history_log_internal(spa, "vdev set", tx,
"vdev_guid=%llu: %s=%lld",
(u_longlong_t)vdev_guid,
nvpair_name(elem), (longlong_t)intval);
} else {
panic("invalid vdev property type %u",
nvpair_type(elem));
}
}
}
mutex_exit(&spa->spa_props_lock);
}
int
vdev_prop_set(vdev_t *vd, nvlist_t *innvl, nvlist_t *outnvl)
{
spa_t *spa = vd->vdev_spa;
nvpair_t *elem = NULL;
uint64_t vdev_guid;
nvlist_t *nvprops;
int error = 0;
ASSERT(vd != NULL);
if (nvlist_lookup_uint64(innvl, ZPOOL_VDEV_PROPS_SET_VDEV,
&vdev_guid) != 0)
return (SET_ERROR(EINVAL));
if (nvlist_lookup_nvlist(innvl, ZPOOL_VDEV_PROPS_SET_PROPS,
&nvprops) != 0)
return (SET_ERROR(EINVAL));
if ((vd = spa_lookup_by_guid(spa, vdev_guid, B_TRUE)) == NULL)
return (SET_ERROR(EINVAL));
while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) {
const char *propname = nvpair_name(elem);
vdev_prop_t prop = vdev_name_to_prop(propname);
uint64_t intval = 0;
const char *strval = NULL;
if (prop == VDEV_PROP_USERPROP && !vdev_prop_user(propname)) {
error = EINVAL;
goto end;
}
if (vdev_prop_readonly(prop)) {
error = EROFS;
goto end;
}
/* Special Processing */
switch (prop) {
case VDEV_PROP_PATH:
if (vd->vdev_path == NULL) {
error = EROFS;
break;
}
if (nvpair_value_string(elem, &strval) != 0) {
error = EINVAL;
break;
}
/* New path must start with /dev/ */
if (strncmp(strval, "/dev/", 5)) {
error = EINVAL;
break;
}
error = spa_vdev_setpath(spa, vdev_guid, strval);
break;
case VDEV_PROP_ALLOCATING:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
if (intval != vd->vdev_noalloc)
break;
if (intval == 0)
error = spa_vdev_noalloc(spa, vdev_guid);
else
error = spa_vdev_alloc(spa, vdev_guid);
break;
case VDEV_PROP_FAILFAST:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_failfast = intval & 1;
break;
case VDEV_PROP_CHECKSUM_N:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_checksum_n = intval;
break;
case VDEV_PROP_CHECKSUM_T:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_checksum_t = intval;
break;
case VDEV_PROP_IO_N:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_io_n = intval;
break;
case VDEV_PROP_IO_T:
if (nvpair_value_uint64(elem, &intval) != 0) {
error = EINVAL;
break;
}
vd->vdev_io_t = intval;
break;
default:
/* Most processing is done in vdev_props_set_sync */
break;
}
end:
if (error != 0) {
intval = error;
vdev_prop_add_list(outnvl, propname, strval, intval, 0);
return (error);
}
}
return (dsl_sync_task(spa->spa_name, NULL, vdev_props_set_sync,
innvl, 6, ZFS_SPACE_CHECK_EXTRA_RESERVED));
}
int
vdev_prop_get(vdev_t *vd, nvlist_t *innvl, nvlist_t *outnvl)
{
spa_t *spa = vd->vdev_spa;
objset_t *mos = spa->spa_meta_objset;
int err = 0;
uint64_t objid;
uint64_t vdev_guid;
nvpair_t *elem = NULL;
nvlist_t *nvprops = NULL;
uint64_t intval = 0;
char *strval = NULL;
const char *propname = NULL;
vdev_prop_t prop;
ASSERT(vd != NULL);
ASSERT(mos != NULL);
if (nvlist_lookup_uint64(innvl, ZPOOL_VDEV_PROPS_GET_VDEV,
&vdev_guid) != 0)
return (SET_ERROR(EINVAL));
nvlist_lookup_nvlist(innvl, ZPOOL_VDEV_PROPS_GET_PROPS, &nvprops);
- if (vd->vdev_top_zap != 0) {
+ if (vd->vdev_root_zap != 0) {
+ objid = vd->vdev_root_zap;
+ } else if (vd->vdev_top_zap != 0) {
objid = vd->vdev_top_zap;
} else if (vd->vdev_leaf_zap != 0) {
objid = vd->vdev_leaf_zap;
} else {
return (SET_ERROR(EINVAL));
}
ASSERT(objid != 0);
mutex_enter(&spa->spa_props_lock);
if (nvprops != NULL) {
char namebuf[64] = { 0 };
while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) {
intval = 0;
strval = NULL;
propname = nvpair_name(elem);
prop = vdev_name_to_prop(propname);
zprop_source_t src = ZPROP_SRC_DEFAULT;
uint64_t integer_size, num_integers;
switch (prop) {
/* Special Read-only Properties */
case VDEV_PROP_NAME:
strval = vdev_name(vd, namebuf,
sizeof (namebuf));
if (strval == NULL)
continue;
vdev_prop_add_list(outnvl, propname, strval, 0,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_CAPACITY:
/* percent used */
intval = (vd->vdev_stat.vs_dspace == 0) ? 0 :
(vd->vdev_stat.vs_alloc * 100 /
vd->vdev_stat.vs_dspace);
vdev_prop_add_list(outnvl, propname, NULL,
intval, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_STATE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_state, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_GUID:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_guid, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_ASIZE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_asize, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PSIZE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_psize, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_ASHIFT:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_ashift, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_SIZE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_dspace, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_FREE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_dspace -
vd->vdev_stat.vs_alloc, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_ALLOCATED:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_alloc, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_EXPANDSZ:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_esize, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_FRAGMENTATION:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_fragmentation,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PARITY:
vdev_prop_add_list(outnvl, propname, NULL,
vdev_get_nparity(vd), ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PATH:
if (vd->vdev_path == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_path, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_DEVID:
if (vd->vdev_devid == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_devid, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PHYS_PATH:
if (vd->vdev_physpath == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_physpath, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_ENC_PATH:
if (vd->vdev_enc_sysfs_path == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_enc_sysfs_path, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_FRU:
if (vd->vdev_fru == NULL)
continue;
vdev_prop_add_list(outnvl, propname,
vd->vdev_fru, 0, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_PARENT:
if (vd->vdev_parent != NULL) {
strval = vdev_name(vd->vdev_parent,
namebuf, sizeof (namebuf));
vdev_prop_add_list(outnvl, propname,
strval, 0, ZPROP_SRC_NONE);
}
continue;
case VDEV_PROP_CHILDREN:
if (vd->vdev_children > 0)
strval = kmem_zalloc(ZAP_MAXVALUELEN,
KM_SLEEP);
for (uint64_t i = 0; i < vd->vdev_children;
i++) {
const char *vname;
vname = vdev_name(vd->vdev_child[i],
namebuf, sizeof (namebuf));
if (vname == NULL)
vname = "(unknown)";
if (strlen(strval) > 0)
strlcat(strval, ",",
ZAP_MAXVALUELEN);
strlcat(strval, vname, ZAP_MAXVALUELEN);
}
if (strval != NULL) {
vdev_prop_add_list(outnvl, propname,
strval, 0, ZPROP_SRC_NONE);
kmem_free(strval, ZAP_MAXVALUELEN);
}
continue;
case VDEV_PROP_NUMCHILDREN:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_children, ZPROP_SRC_NONE);
continue;
case VDEV_PROP_READ_ERRORS:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_read_errors,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_WRITE_ERRORS:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_write_errors,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_CHECKSUM_ERRORS:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_checksum_errors,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_INITIALIZE_ERRORS:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_initialize_errors,
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_NULL:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_NULL],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_READ:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_READ],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_WRITE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_WRITE],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_FREE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_FREE],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_CLAIM:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_CLAIM],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_OPS_TRIM:
/*
* TRIM ops and bytes are reported to user
* space as ZIO_TYPE_IOCTL. This is done to
* preserve the vdev_stat_t structure layout
* for user space.
*/
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_ops[ZIO_TYPE_IOCTL],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_NULL:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_NULL],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_READ:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_READ],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_WRITE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_WRITE],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_FREE:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_FREE],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_CLAIM:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_CLAIM],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_BYTES_TRIM:
/*
* TRIM ops and bytes are reported to user
* space as ZIO_TYPE_IOCTL. This is done to
* preserve the vdev_stat_t structure layout
* for user space.
*/
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_stat.vs_bytes[ZIO_TYPE_IOCTL],
ZPROP_SRC_NONE);
continue;
case VDEV_PROP_REMOVING:
vdev_prop_add_list(outnvl, propname, NULL,
vd->vdev_removing, ZPROP_SRC_NONE);
continue;
/* Numeric Properites */
case VDEV_PROP_ALLOCATING:
/* Leaf vdevs cannot have this property */
if (vd->vdev_mg == NULL &&
vd->vdev_top != NULL) {
src = ZPROP_SRC_NONE;
intval = ZPROP_BOOLEAN_NA;
} else {
err = vdev_prop_get_int(vd, prop,
&intval);
if (err && err != ENOENT)
break;
if (intval ==
vdev_prop_default_numeric(prop))
src = ZPROP_SRC_DEFAULT;
else
src = ZPROP_SRC_LOCAL;
}
vdev_prop_add_list(outnvl, propname, NULL,
intval, src);
break;
case VDEV_PROP_FAILFAST:
src = ZPROP_SRC_LOCAL;
strval = NULL;
err = zap_lookup(mos, objid, nvpair_name(elem),
sizeof (uint64_t), 1, &intval);
if (err == ENOENT) {
intval = vdev_prop_default_numeric(
prop);
err = 0;
} else if (err) {
break;
}
if (intval == vdev_prop_default_numeric(prop))
src = ZPROP_SRC_DEFAULT;
vdev_prop_add_list(outnvl, propname, strval,
intval, src);
break;
case VDEV_PROP_CHECKSUM_N:
case VDEV_PROP_CHECKSUM_T:
case VDEV_PROP_IO_N:
case VDEV_PROP_IO_T:
err = vdev_prop_get_int(vd, prop, &intval);
if (err && err != ENOENT)
break;
if (intval == vdev_prop_default_numeric(prop))
src = ZPROP_SRC_DEFAULT;
else
src = ZPROP_SRC_LOCAL;
vdev_prop_add_list(outnvl, propname, NULL,
intval, src);
break;
/* Text Properties */
case VDEV_PROP_COMMENT:
/* Exists in the ZAP below */
/* FALLTHRU */
case VDEV_PROP_USERPROP:
/* User Properites */
src = ZPROP_SRC_LOCAL;
err = zap_length(mos, objid, nvpair_name(elem),
&integer_size, &num_integers);
if (err)
break;
switch (integer_size) {
case 8:
/* User properties cannot be integers */
err = EINVAL;
break;
case 1:
/* string property */
strval = kmem_alloc(num_integers,
KM_SLEEP);
err = zap_lookup(mos, objid,
nvpair_name(elem), 1,
num_integers, strval);
if (err) {
kmem_free(strval,
num_integers);
break;
}
vdev_prop_add_list(outnvl, propname,
strval, 0, src);
kmem_free(strval, num_integers);
break;
}
break;
default:
err = ENOENT;
break;
}
if (err)
break;
}
} else {
/*
* Get all properties from the MOS vdev property object.
*/
zap_cursor_t zc;
zap_attribute_t za;
for (zap_cursor_init(&zc, mos, objid);
(err = zap_cursor_retrieve(&zc, &za)) == 0;
zap_cursor_advance(&zc)) {
intval = 0;
strval = NULL;
zprop_source_t src = ZPROP_SRC_DEFAULT;
propname = za.za_name;
switch (za.za_integer_length) {
case 8:
/* We do not allow integer user properties */
/* This is likely an internal value */
break;
case 1:
/* string property */
strval = kmem_alloc(za.za_num_integers,
KM_SLEEP);
err = zap_lookup(mos, objid, za.za_name, 1,
za.za_num_integers, strval);
if (err) {
kmem_free(strval, za.za_num_integers);
break;
}
vdev_prop_add_list(outnvl, propname, strval, 0,
src);
kmem_free(strval, za.za_num_integers);
break;
default:
break;
}
}
zap_cursor_fini(&zc);
}
mutex_exit(&spa->spa_props_lock);
if (err && err != ENOENT) {
return (err);
}
return (0);
}
EXPORT_SYMBOL(vdev_fault);
EXPORT_SYMBOL(vdev_degrade);
EXPORT_SYMBOL(vdev_online);
EXPORT_SYMBOL(vdev_offline);
EXPORT_SYMBOL(vdev_clear);
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, default_ms_count, UINT, ZMOD_RW,
"Target number of metaslabs per top-level vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, default_ms_shift, UINT, ZMOD_RW,
- "Default limit for metaslab size");
+ "Default lower limit for metaslab size");
+
+ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, max_ms_shift, UINT, ZMOD_RW,
+ "Default upper limit for metaslab size");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, min_ms_count, UINT, ZMOD_RW,
"Minimum number of metaslabs per top-level vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, ms_count_limit, UINT, ZMOD_RW,
"Practical upper limit of total metaslabs per top-level vdev");
ZFS_MODULE_PARAM(zfs, zfs_, slow_io_events_per_second, UINT, ZMOD_RW,
"Rate limit slow IO (delay) events to this many per second");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs, zfs_, checksum_events_per_second, UINT, ZMOD_RW,
"Rate limit checksum events to this many checksum errors per second "
"(do not set below ZED threshold).");
/* END CSTYLED */
ZFS_MODULE_PARAM(zfs, zfs_, scan_ignore_errors, INT, ZMOD_RW,
"Ignore errors during resilver/scrub");
ZFS_MODULE_PARAM(zfs_vdev, vdev_, validate_skip, INT, ZMOD_RW,
"Bypass vdev_validate()");
ZFS_MODULE_PARAM(zfs, zfs_, nocacheflush, INT, ZMOD_RW,
"Disable cache flushes");
ZFS_MODULE_PARAM(zfs, zfs_, embedded_slog_min_ms, UINT, ZMOD_RW,
"Minimum number of metaslabs required to dedicate one for log blocks");
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM_CALL(zfs_vdev, zfs_vdev_, min_auto_ashift,
param_set_min_auto_ashift, param_get_uint, ZMOD_RW,
"Minimum ashift used when creating new top-level vdevs");
ZFS_MODULE_PARAM_CALL(zfs_vdev, zfs_vdev_, max_auto_ashift,
param_set_max_auto_ashift, param_get_uint, ZMOD_RW,
"Maximum ashift used when optimizing for logical -> physical sector "
"size on new top-level vdevs");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/vdev_indirect.c b/sys/contrib/openzfs/module/zfs/vdev_indirect.c
index 8c11a574ae86..a16ad2f4e7cf 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_indirect.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_indirect.c
@@ -1,1911 +1,1912 @@
/*
* CDDL HEADER START
*
* This file and its contents are supplied under the terms of the
* Common Development and Distribution License ("CDDL"), version 1.0.
* You may only use this file in accordance with the terms of version
* 1.0 of the CDDL.
*
* A full copy of the text of the CDDL should have accompanied this
* source. A copy of the CDDL is also available via the Internet at
* http://www.illumos.org/license/CDDL.
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2014, 2017 by Delphix. All rights reserved.
* Copyright (c) 2019, loli10K <ezomori.nozomu@gmail.com>. All rights reserved.
* Copyright (c) 2014, 2020 by Delphix. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/vdev_impl.h>
#include <sys/fs/zfs.h>
#include <sys/zio.h>
#include <sys/zio_checksum.h>
#include <sys/metaslab.h>
#include <sys/dmu.h>
#include <sys/vdev_indirect_mapping.h>
#include <sys/dmu_tx.h>
#include <sys/dsl_synctask.h>
#include <sys/zap.h>
#include <sys/abd.h>
#include <sys/zthr.h>
/*
* An indirect vdev corresponds to a vdev that has been removed. Since
* we cannot rewrite block pointers of snapshots, etc., we keep a
* mapping from old location on the removed device to the new location
* on another device in the pool and use this mapping whenever we need
* to access the DVA. Unfortunately, this mapping did not respect
* logical block boundaries when it was first created, and so a DVA on
* this indirect vdev may be "split" into multiple sections that each
* map to a different location. As a consequence, not all DVAs can be
* translated to an equivalent new DVA. Instead we must provide a
* "vdev_remap" operation that executes a callback on each contiguous
* segment of the new location. This function is used in multiple ways:
*
* - I/Os to this vdev use the callback to determine where the
* data is now located, and issue child I/Os for each segment's new
* location.
*
* - frees and claims to this vdev use the callback to free or claim
* each mapped segment. (Note that we don't actually need to claim
* log blocks on indirect vdevs, because we don't allocate to
* removing vdevs. However, zdb uses zio_claim() for its leak
* detection.)
*/
/*
* "Big theory statement" for how we mark blocks obsolete.
*
* When a block on an indirect vdev is freed or remapped, a section of
* that vdev's mapping may no longer be referenced (aka "obsolete"). We
* keep track of how much of each mapping entry is obsolete. When
* an entry becomes completely obsolete, we can remove it, thus reducing
* the memory used by the mapping. The complete picture of obsolescence
* is given by the following data structures, described below:
* - the entry-specific obsolete count
* - the vdev-specific obsolete spacemap
* - the pool-specific obsolete bpobj
*
* == On disk data structures used ==
*
* We track the obsolete space for the pool using several objects. Each
* of these objects is created on demand and freed when no longer
* needed, and is assumed to be empty if it does not exist.
* SPA_FEATURE_OBSOLETE_COUNTS includes the count of these objects.
*
* - Each vic_mapping_object (associated with an indirect vdev) can
* have a vimp_counts_object. This is an array of uint32_t's
* with the same number of entries as the vic_mapping_object. When
* the mapping is condensed, entries from the vic_obsolete_sm_object
* (see below) are folded into the counts. Therefore, each
* obsolete_counts entry tells us the number of bytes in the
* corresponding mapping entry that were not referenced when the
* mapping was last condensed.
*
* - Each indirect or removing vdev can have a vic_obsolete_sm_object.
* This is a space map containing an alloc entry for every DVA that
* has been obsoleted since the last time this indirect vdev was
* condensed. We use this object in order to improve performance
* when marking a DVA as obsolete. Instead of modifying an arbitrary
* offset of the vimp_counts_object, we only need to append an entry
* to the end of this object. When a DVA becomes obsolete, it is
* added to the obsolete space map. This happens when the DVA is
* freed, remapped and not referenced by a snapshot, or the last
* snapshot referencing it is destroyed.
*
* - Each dataset can have a ds_remap_deadlist object. This is a
* deadlist object containing all blocks that were remapped in this
* dataset but referenced in a previous snapshot. Blocks can *only*
* appear on this list if they were remapped (dsl_dataset_block_remapped);
* blocks that were killed in a head dataset are put on the normal
* ds_deadlist and marked obsolete when they are freed.
*
* - The pool can have a dp_obsolete_bpobj. This is a list of blocks
* in the pool that need to be marked obsolete. When a snapshot is
* destroyed, we move some of the ds_remap_deadlist to the obsolete
* bpobj (see dsl_destroy_snapshot_handle_remaps()). We then
* asynchronously process the obsolete bpobj, moving its entries to
* the specific vdevs' obsolete space maps.
*
* == Summary of how we mark blocks as obsolete ==
*
* - When freeing a block: if any DVA is on an indirect vdev, append to
* vic_obsolete_sm_object.
* - When remapping a block, add dva to ds_remap_deadlist (if prev snap
* references; otherwise append to vic_obsolete_sm_object).
* - When freeing a snapshot: move parts of ds_remap_deadlist to
* dp_obsolete_bpobj (same algorithm as ds_deadlist).
* - When syncing the spa: process dp_obsolete_bpobj, moving ranges to
* individual vdev's vic_obsolete_sm_object.
*/
/*
* "Big theory statement" for how we condense indirect vdevs.
*
* Condensing an indirect vdev's mapping is the process of determining
* the precise counts of obsolete space for each mapping entry (by
* integrating the obsolete spacemap into the obsolete counts) and
* writing out a new mapping that contains only referenced entries.
*
* We condense a vdev when we expect the mapping to shrink (see
* vdev_indirect_should_condense()), but only perform one condense at a
* time to limit the memory usage. In addition, we use a separate
* open-context thread (spa_condense_indirect_thread) to incrementally
* create the new mapping object in a way that minimizes the impact on
* the rest of the system.
*
* == Generating a new mapping ==
*
* To generate a new mapping, we follow these steps:
*
* 1. Save the old obsolete space map and create a new mapping object
* (see spa_condense_indirect_start_sync()). This initializes the
* spa_condensing_indirect_phys with the "previous obsolete space map",
* which is now read only. Newly obsolete DVAs will be added to a
* new (initially empty) obsolete space map, and will not be
* considered as part of this condense operation.
*
* 2. Construct in memory the precise counts of obsolete space for each
* mapping entry, by incorporating the obsolete space map into the
* counts. (See vdev_indirect_mapping_load_obsolete_{counts,spacemap}().)
*
* 3. Iterate through each mapping entry, writing to the new mapping any
* entries that are not completely obsolete (i.e. which don't have
* obsolete count == mapping length). (See
* spa_condense_indirect_generate_new_mapping().)
*
* 4. Destroy the old mapping object and switch over to the new one
* (spa_condense_indirect_complete_sync).
*
* == Restarting from failure ==
*
* To restart the condense when we import/open the pool, we must start
* at the 2nd step above: reconstruct the precise counts in memory,
* based on the space map + counts. Then in the 3rd step, we start
* iterating where we left off: at vimp_max_offset of the new mapping
* object.
*/
static int zfs_condense_indirect_vdevs_enable = B_TRUE;
/*
* Condense if at least this percent of the bytes in the mapping is
* obsolete. With the default of 25%, the amount of space mapped
* will be reduced to 1% of its original size after at most 16
* condenses. Higher values will condense less often (causing less
* i/o); lower values will reduce the mapping size more quickly.
*/
static uint_t zfs_condense_indirect_obsolete_pct = 25;
/*
* Condense if the obsolete space map takes up more than this amount of
* space on disk (logically). This limits the amount of disk space
* consumed by the obsolete space map; the default of 1GB is small enough
* that we typically don't mind "wasting" it.
*/
static uint64_t zfs_condense_max_obsolete_bytes = 1024 * 1024 * 1024;
/*
* Don't bother condensing if the mapping uses less than this amount of
* memory. The default of 128KB is considered a "trivial" amount of
* memory and not worth reducing.
*/
static uint64_t zfs_condense_min_mapping_bytes = 128 * 1024;
/*
* This is used by the test suite so that it can ensure that certain
* actions happen while in the middle of a condense (which might otherwise
* complete too quickly). If used to reduce the performance impact of
* condensing in production, a maximum value of 1 should be sufficient.
*/
static uint_t zfs_condense_indirect_commit_entry_delay_ms = 0;
/*
* If an indirect split block contains more than this many possible unique
* combinations when being reconstructed, consider it too computationally
* expensive to check them all. Instead, try at most 100 randomly-selected
* combinations each time the block is accessed. This allows all segment
* copies to participate fairly in the reconstruction when all combinations
* cannot be checked and prevents repeated use of one bad copy.
*/
uint_t zfs_reconstruct_indirect_combinations_max = 4096;
/*
* Enable to simulate damaged segments and validate reconstruction. This
* is intentionally not exposed as a module parameter.
*/
unsigned long zfs_reconstruct_indirect_damage_fraction = 0;
/*
* The indirect_child_t represents the vdev that we will read from, when we
* need to read all copies of the data (e.g. for scrub or reconstruction).
* For plain (non-mirror) top-level vdevs (i.e. is_vdev is not a mirror),
* ic_vdev is the same as is_vdev. However, for mirror top-level vdevs,
* ic_vdev is a child of the mirror.
*/
typedef struct indirect_child {
abd_t *ic_data;
vdev_t *ic_vdev;
/*
* ic_duplicate is NULL when the ic_data contents are unique, when it
* is determined to be a duplicate it references the primary child.
*/
struct indirect_child *ic_duplicate;
list_node_t ic_node; /* node on is_unique_child */
int ic_error; /* set when a child does not contain the data */
} indirect_child_t;
/*
* The indirect_split_t represents one mapped segment of an i/o to the
* indirect vdev. For non-split (contiguously-mapped) blocks, there will be
* only one indirect_split_t, with is_split_offset==0 and is_size==io_size.
* For split blocks, there will be several of these.
*/
typedef struct indirect_split {
list_node_t is_node; /* link on iv_splits */
/*
* is_split_offset is the offset into the i/o.
* This is the sum of the previous splits' is_size's.
*/
uint64_t is_split_offset;
vdev_t *is_vdev; /* top-level vdev */
uint64_t is_target_offset; /* offset on is_vdev */
uint64_t is_size;
int is_children; /* number of entries in is_child[] */
int is_unique_children; /* number of entries in is_unique_child */
list_t is_unique_child;
/*
* is_good_child is the child that we are currently using to
* attempt reconstruction.
*/
indirect_child_t *is_good_child;
indirect_child_t is_child[];
} indirect_split_t;
/*
* The indirect_vsd_t is associated with each i/o to the indirect vdev.
* It is the "Vdev-Specific Data" in the zio_t's io_vsd.
*/
typedef struct indirect_vsd {
boolean_t iv_split_block;
boolean_t iv_reconstruct;
uint64_t iv_unique_combinations;
uint64_t iv_attempts;
uint64_t iv_attempts_max;
list_t iv_splits; /* list of indirect_split_t's */
} indirect_vsd_t;
static void
vdev_indirect_map_free(zio_t *zio)
{
indirect_vsd_t *iv = zio->io_vsd;
indirect_split_t *is;
while ((is = list_head(&iv->iv_splits)) != NULL) {
for (int c = 0; c < is->is_children; c++) {
indirect_child_t *ic = &is->is_child[c];
if (ic->ic_data != NULL)
abd_free(ic->ic_data);
}
list_remove(&iv->iv_splits, is);
indirect_child_t *ic;
while ((ic = list_head(&is->is_unique_child)) != NULL)
list_remove(&is->is_unique_child, ic);
list_destroy(&is->is_unique_child);
kmem_free(is,
offsetof(indirect_split_t, is_child[is->is_children]));
}
kmem_free(iv, sizeof (*iv));
}
static const zio_vsd_ops_t vdev_indirect_vsd_ops = {
.vsd_free = vdev_indirect_map_free,
};
/*
* Mark the given offset and size as being obsolete.
*/
void
vdev_indirect_mark_obsolete(vdev_t *vd, uint64_t offset, uint64_t size)
{
spa_t *spa = vd->vdev_spa;
ASSERT3U(vd->vdev_indirect_config.vic_mapping_object, !=, 0);
ASSERT(vd->vdev_removing || vd->vdev_ops == &vdev_indirect_ops);
ASSERT(size > 0);
VERIFY(vdev_indirect_mapping_entry_for_offset(
vd->vdev_indirect_mapping, offset) != NULL);
if (spa_feature_is_enabled(spa, SPA_FEATURE_OBSOLETE_COUNTS)) {
mutex_enter(&vd->vdev_obsolete_lock);
range_tree_add(vd->vdev_obsolete_segments, offset, size);
mutex_exit(&vd->vdev_obsolete_lock);
vdev_dirty(vd, 0, NULL, spa_syncing_txg(spa));
}
}
/*
* Mark the DVA vdev_id:offset:size as being obsolete in the given tx. This
* wrapper is provided because the DMU does not know about vdev_t's and
* cannot directly call vdev_indirect_mark_obsolete.
*/
void
spa_vdev_indirect_mark_obsolete(spa_t *spa, uint64_t vdev_id, uint64_t offset,
uint64_t size, dmu_tx_t *tx)
{
vdev_t *vd = vdev_lookup_top(spa, vdev_id);
ASSERT(dmu_tx_is_syncing(tx));
/* The DMU can only remap indirect vdevs. */
ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops);
vdev_indirect_mark_obsolete(vd, offset, size);
}
static spa_condensing_indirect_t *
spa_condensing_indirect_create(spa_t *spa)
{
spa_condensing_indirect_phys_t *scip =
&spa->spa_condensing_indirect_phys;
spa_condensing_indirect_t *sci = kmem_zalloc(sizeof (*sci), KM_SLEEP);
objset_t *mos = spa->spa_meta_objset;
for (int i = 0; i < TXG_SIZE; i++) {
list_create(&sci->sci_new_mapping_entries[i],
sizeof (vdev_indirect_mapping_entry_t),
offsetof(vdev_indirect_mapping_entry_t, vime_node));
}
sci->sci_new_mapping =
vdev_indirect_mapping_open(mos, scip->scip_next_mapping_object);
return (sci);
}
static void
spa_condensing_indirect_destroy(spa_condensing_indirect_t *sci)
{
for (int i = 0; i < TXG_SIZE; i++)
list_destroy(&sci->sci_new_mapping_entries[i]);
if (sci->sci_new_mapping != NULL)
vdev_indirect_mapping_close(sci->sci_new_mapping);
kmem_free(sci, sizeof (*sci));
}
boolean_t
vdev_indirect_should_condense(vdev_t *vd)
{
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
spa_t *spa = vd->vdev_spa;
ASSERT(dsl_pool_sync_context(spa->spa_dsl_pool));
if (!zfs_condense_indirect_vdevs_enable)
return (B_FALSE);
/*
* We can only condense one indirect vdev at a time.
*/
if (spa->spa_condensing_indirect != NULL)
return (B_FALSE);
if (spa_shutting_down(spa))
return (B_FALSE);
/*
* The mapping object size must not change while we are
* condensing, so we can only condense indirect vdevs
* (not vdevs that are still in the middle of being removed).
*/
if (vd->vdev_ops != &vdev_indirect_ops)
return (B_FALSE);
/*
* If nothing new has been marked obsolete, there is no
* point in condensing.
*/
uint64_t obsolete_sm_obj __maybe_unused;
ASSERT0(vdev_obsolete_sm_object(vd, &obsolete_sm_obj));
if (vd->vdev_obsolete_sm == NULL) {
ASSERT0(obsolete_sm_obj);
return (B_FALSE);
}
ASSERT(vd->vdev_obsolete_sm != NULL);
ASSERT3U(obsolete_sm_obj, ==, space_map_object(vd->vdev_obsolete_sm));
uint64_t bytes_mapped = vdev_indirect_mapping_bytes_mapped(vim);
uint64_t bytes_obsolete = space_map_allocated(vd->vdev_obsolete_sm);
uint64_t mapping_size = vdev_indirect_mapping_size(vim);
uint64_t obsolete_sm_size = space_map_length(vd->vdev_obsolete_sm);
ASSERT3U(bytes_obsolete, <=, bytes_mapped);
/*
* If a high percentage of the bytes that are mapped have become
* obsolete, condense (unless the mapping is already small enough).
* This has a good chance of reducing the amount of memory used
* by the mapping.
*/
if (bytes_obsolete * 100 / bytes_mapped >=
zfs_condense_indirect_obsolete_pct &&
mapping_size > zfs_condense_min_mapping_bytes) {
zfs_dbgmsg("should condense vdev %llu because obsolete "
"spacemap covers %d%% of %lluMB mapping",
(u_longlong_t)vd->vdev_id,
(int)(bytes_obsolete * 100 / bytes_mapped),
(u_longlong_t)bytes_mapped / 1024 / 1024);
return (B_TRUE);
}
/*
* If the obsolete space map takes up too much space on disk,
* condense in order to free up this disk space.
*/
if (obsolete_sm_size >= zfs_condense_max_obsolete_bytes) {
zfs_dbgmsg("should condense vdev %llu because obsolete sm "
"length %lluMB >= max size %lluMB",
(u_longlong_t)vd->vdev_id,
(u_longlong_t)obsolete_sm_size / 1024 / 1024,
(u_longlong_t)zfs_condense_max_obsolete_bytes /
1024 / 1024);
return (B_TRUE);
}
return (B_FALSE);
}
/*
* This sync task completes (finishes) a condense, deleting the old
* mapping and replacing it with the new one.
*/
static void
spa_condense_indirect_complete_sync(void *arg, dmu_tx_t *tx)
{
spa_condensing_indirect_t *sci = arg;
spa_t *spa = dmu_tx_pool(tx)->dp_spa;
spa_condensing_indirect_phys_t *scip =
&spa->spa_condensing_indirect_phys;
vdev_t *vd = vdev_lookup_top(spa, scip->scip_vdev);
vdev_indirect_config_t *vic = &vd->vdev_indirect_config;
objset_t *mos = spa->spa_meta_objset;
vdev_indirect_mapping_t *old_mapping = vd->vdev_indirect_mapping;
uint64_t old_count = vdev_indirect_mapping_num_entries(old_mapping);
uint64_t new_count =
vdev_indirect_mapping_num_entries(sci->sci_new_mapping);
ASSERT(dmu_tx_is_syncing(tx));
ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops);
ASSERT3P(sci, ==, spa->spa_condensing_indirect);
for (int i = 0; i < TXG_SIZE; i++) {
ASSERT(list_is_empty(&sci->sci_new_mapping_entries[i]));
}
ASSERT(vic->vic_mapping_object != 0);
ASSERT3U(vd->vdev_id, ==, scip->scip_vdev);
ASSERT(scip->scip_next_mapping_object != 0);
ASSERT(scip->scip_prev_obsolete_sm_object != 0);
/*
* Reset vdev_indirect_mapping to refer to the new object.
*/
rw_enter(&vd->vdev_indirect_rwlock, RW_WRITER);
vdev_indirect_mapping_close(vd->vdev_indirect_mapping);
vd->vdev_indirect_mapping = sci->sci_new_mapping;
rw_exit(&vd->vdev_indirect_rwlock);
sci->sci_new_mapping = NULL;
vdev_indirect_mapping_free(mos, vic->vic_mapping_object, tx);
vic->vic_mapping_object = scip->scip_next_mapping_object;
scip->scip_next_mapping_object = 0;
space_map_free_obj(mos, scip->scip_prev_obsolete_sm_object, tx);
spa_feature_decr(spa, SPA_FEATURE_OBSOLETE_COUNTS, tx);
scip->scip_prev_obsolete_sm_object = 0;
scip->scip_vdev = 0;
VERIFY0(zap_remove(mos, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_CONDENSING_INDIRECT, tx));
spa_condensing_indirect_destroy(spa->spa_condensing_indirect);
spa->spa_condensing_indirect = NULL;
zfs_dbgmsg("finished condense of vdev %llu in txg %llu: "
"new mapping object %llu has %llu entries "
"(was %llu entries)",
(u_longlong_t)vd->vdev_id, (u_longlong_t)dmu_tx_get_txg(tx),
(u_longlong_t)vic->vic_mapping_object,
(u_longlong_t)new_count, (u_longlong_t)old_count);
vdev_config_dirty(spa->spa_root_vdev);
}
/*
* This sync task appends entries to the new mapping object.
*/
static void
spa_condense_indirect_commit_sync(void *arg, dmu_tx_t *tx)
{
spa_condensing_indirect_t *sci = arg;
uint64_t txg = dmu_tx_get_txg(tx);
spa_t *spa __maybe_unused = dmu_tx_pool(tx)->dp_spa;
ASSERT(dmu_tx_is_syncing(tx));
ASSERT3P(sci, ==, spa->spa_condensing_indirect);
vdev_indirect_mapping_add_entries(sci->sci_new_mapping,
&sci->sci_new_mapping_entries[txg & TXG_MASK], tx);
ASSERT(list_is_empty(&sci->sci_new_mapping_entries[txg & TXG_MASK]));
}
/*
* Open-context function to add one entry to the new mapping. The new
* entry will be remembered and written from syncing context.
*/
static void
spa_condense_indirect_commit_entry(spa_t *spa,
vdev_indirect_mapping_entry_phys_t *vimep, uint32_t count)
{
spa_condensing_indirect_t *sci = spa->spa_condensing_indirect;
ASSERT3U(count, <, DVA_GET_ASIZE(&vimep->vimep_dst));
dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir);
dmu_tx_hold_space(tx, sizeof (*vimep) + sizeof (count));
VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
int txgoff = dmu_tx_get_txg(tx) & TXG_MASK;
/*
* If we are the first entry committed this txg, kick off the sync
* task to write to the MOS on our behalf.
*/
if (list_is_empty(&sci->sci_new_mapping_entries[txgoff])) {
dsl_sync_task_nowait(dmu_tx_pool(tx),
spa_condense_indirect_commit_sync, sci, tx);
}
vdev_indirect_mapping_entry_t *vime =
kmem_alloc(sizeof (*vime), KM_SLEEP);
vime->vime_mapping = *vimep;
vime->vime_obsolete_count = count;
list_insert_tail(&sci->sci_new_mapping_entries[txgoff], vime);
dmu_tx_commit(tx);
}
static void
spa_condense_indirect_generate_new_mapping(vdev_t *vd,
uint32_t *obsolete_counts, uint64_t start_index, zthr_t *zthr)
{
spa_t *spa = vd->vdev_spa;
uint64_t mapi = start_index;
vdev_indirect_mapping_t *old_mapping = vd->vdev_indirect_mapping;
uint64_t old_num_entries =
vdev_indirect_mapping_num_entries(old_mapping);
ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops);
ASSERT3U(vd->vdev_id, ==, spa->spa_condensing_indirect_phys.scip_vdev);
zfs_dbgmsg("starting condense of vdev %llu from index %llu",
(u_longlong_t)vd->vdev_id,
(u_longlong_t)mapi);
while (mapi < old_num_entries) {
if (zthr_iscancelled(zthr)) {
zfs_dbgmsg("pausing condense of vdev %llu "
"at index %llu", (u_longlong_t)vd->vdev_id,
(u_longlong_t)mapi);
break;
}
vdev_indirect_mapping_entry_phys_t *entry =
&old_mapping->vim_entries[mapi];
uint64_t entry_size = DVA_GET_ASIZE(&entry->vimep_dst);
ASSERT3U(obsolete_counts[mapi], <=, entry_size);
if (obsolete_counts[mapi] < entry_size) {
spa_condense_indirect_commit_entry(spa, entry,
obsolete_counts[mapi]);
/*
* This delay may be requested for testing, debugging,
* or performance reasons.
*/
hrtime_t now = gethrtime();
hrtime_t sleep_until = now + MSEC2NSEC(
zfs_condense_indirect_commit_entry_delay_ms);
zfs_sleep_until(sleep_until);
}
mapi++;
}
}
static boolean_t
spa_condense_indirect_thread_check(void *arg, zthr_t *zthr)
{
(void) zthr;
spa_t *spa = arg;
return (spa->spa_condensing_indirect != NULL);
}
static void
spa_condense_indirect_thread(void *arg, zthr_t *zthr)
{
spa_t *spa = arg;
vdev_t *vd;
ASSERT3P(spa->spa_condensing_indirect, !=, NULL);
spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER);
vd = vdev_lookup_top(spa, spa->spa_condensing_indirect_phys.scip_vdev);
ASSERT3P(vd, !=, NULL);
spa_config_exit(spa, SCL_VDEV, FTAG);
spa_condensing_indirect_t *sci = spa->spa_condensing_indirect;
spa_condensing_indirect_phys_t *scip =
&spa->spa_condensing_indirect_phys;
uint32_t *counts;
uint64_t start_index;
vdev_indirect_mapping_t *old_mapping = vd->vdev_indirect_mapping;
space_map_t *prev_obsolete_sm = NULL;
ASSERT3U(vd->vdev_id, ==, scip->scip_vdev);
ASSERT(scip->scip_next_mapping_object != 0);
ASSERT(scip->scip_prev_obsolete_sm_object != 0);
ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops);
for (int i = 0; i < TXG_SIZE; i++) {
/*
* The list must start out empty in order for the
* _commit_sync() sync task to be properly registered
* on the first call to _commit_entry(); so it's wise
* to double check and ensure we actually are starting
* with empty lists.
*/
ASSERT(list_is_empty(&sci->sci_new_mapping_entries[i]));
}
VERIFY0(space_map_open(&prev_obsolete_sm, spa->spa_meta_objset,
scip->scip_prev_obsolete_sm_object, 0, vd->vdev_asize, 0));
counts = vdev_indirect_mapping_load_obsolete_counts(old_mapping);
if (prev_obsolete_sm != NULL) {
vdev_indirect_mapping_load_obsolete_spacemap(old_mapping,
counts, prev_obsolete_sm);
}
space_map_close(prev_obsolete_sm);
/*
* Generate new mapping. Determine what index to continue from
* based on the max offset that we've already written in the
* new mapping.
*/
uint64_t max_offset =
vdev_indirect_mapping_max_offset(sci->sci_new_mapping);
if (max_offset == 0) {
/* We haven't written anything to the new mapping yet. */
start_index = 0;
} else {
/*
* Pick up from where we left off. _entry_for_offset()
* returns a pointer into the vim_entries array. If
* max_offset is greater than any of the mappings
* contained in the table NULL will be returned and
* that indicates we've exhausted our iteration of the
* old_mapping.
*/
vdev_indirect_mapping_entry_phys_t *entry =
vdev_indirect_mapping_entry_for_offset_or_next(old_mapping,
max_offset);
if (entry == NULL) {
/*
* We've already written the whole new mapping.
* This special value will cause us to skip the
* generate_new_mapping step and just do the sync
* task to complete the condense.
*/
start_index = UINT64_MAX;
} else {
start_index = entry - old_mapping->vim_entries;
ASSERT3U(start_index, <,
vdev_indirect_mapping_num_entries(old_mapping));
}
}
spa_condense_indirect_generate_new_mapping(vd, counts,
start_index, zthr);
vdev_indirect_mapping_free_obsolete_counts(old_mapping, counts);
/*
* If the zthr has received a cancellation signal while running
* in generate_new_mapping() or at any point after that, then bail
* early. We don't want to complete the condense if the spa is
* shutting down.
*/
if (zthr_iscancelled(zthr))
return;
VERIFY0(dsl_sync_task(spa_name(spa), NULL,
spa_condense_indirect_complete_sync, sci, 0,
ZFS_SPACE_CHECK_EXTRA_RESERVED));
}
/*
* Sync task to begin the condensing process.
*/
void
spa_condense_indirect_start_sync(vdev_t *vd, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
spa_condensing_indirect_phys_t *scip =
&spa->spa_condensing_indirect_phys;
ASSERT0(scip->scip_next_mapping_object);
ASSERT0(scip->scip_prev_obsolete_sm_object);
ASSERT0(scip->scip_vdev);
ASSERT(dmu_tx_is_syncing(tx));
ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops);
ASSERT(spa_feature_is_active(spa, SPA_FEATURE_OBSOLETE_COUNTS));
ASSERT(vdev_indirect_mapping_num_entries(vd->vdev_indirect_mapping));
uint64_t obsolete_sm_obj;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_obj));
ASSERT3U(obsolete_sm_obj, !=, 0);
scip->scip_vdev = vd->vdev_id;
scip->scip_next_mapping_object =
vdev_indirect_mapping_alloc(spa->spa_meta_objset, tx);
scip->scip_prev_obsolete_sm_object = obsolete_sm_obj;
/*
* We don't need to allocate a new space map object, since
* vdev_indirect_sync_obsolete will allocate one when needed.
*/
space_map_close(vd->vdev_obsolete_sm);
vd->vdev_obsolete_sm = NULL;
VERIFY0(zap_remove(spa->spa_meta_objset, vd->vdev_top_zap,
VDEV_TOP_ZAP_INDIRECT_OBSOLETE_SM, tx));
VERIFY0(zap_add(spa->spa_dsl_pool->dp_meta_objset,
DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_CONDENSING_INDIRECT, sizeof (uint64_t),
sizeof (*scip) / sizeof (uint64_t), scip, tx));
ASSERT3P(spa->spa_condensing_indirect, ==, NULL);
spa->spa_condensing_indirect = spa_condensing_indirect_create(spa);
zfs_dbgmsg("starting condense of vdev %llu in txg %llu: "
"posm=%llu nm=%llu",
(u_longlong_t)vd->vdev_id, (u_longlong_t)dmu_tx_get_txg(tx),
(u_longlong_t)scip->scip_prev_obsolete_sm_object,
(u_longlong_t)scip->scip_next_mapping_object);
zthr_wakeup(spa->spa_condense_zthr);
}
/*
* Sync to the given vdev's obsolete space map any segments that are no longer
* referenced as of the given txg.
*
* If the obsolete space map doesn't exist yet, create and open it.
*/
void
vdev_indirect_sync_obsolete(vdev_t *vd, dmu_tx_t *tx)
{
spa_t *spa = vd->vdev_spa;
vdev_indirect_config_t *vic __maybe_unused = &vd->vdev_indirect_config;
ASSERT3U(vic->vic_mapping_object, !=, 0);
ASSERT(range_tree_space(vd->vdev_obsolete_segments) > 0);
ASSERT(vd->vdev_removing || vd->vdev_ops == &vdev_indirect_ops);
ASSERT(spa_feature_is_enabled(spa, SPA_FEATURE_OBSOLETE_COUNTS));
uint64_t obsolete_sm_object;
VERIFY0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
if (obsolete_sm_object == 0) {
obsolete_sm_object = space_map_alloc(spa->spa_meta_objset,
zfs_vdev_standard_sm_blksz, tx);
ASSERT(vd->vdev_top_zap != 0);
VERIFY0(zap_add(vd->vdev_spa->spa_meta_objset, vd->vdev_top_zap,
VDEV_TOP_ZAP_INDIRECT_OBSOLETE_SM,
sizeof (obsolete_sm_object), 1, &obsolete_sm_object, tx));
ASSERT0(vdev_obsolete_sm_object(vd, &obsolete_sm_object));
ASSERT3U(obsolete_sm_object, !=, 0);
spa_feature_incr(spa, SPA_FEATURE_OBSOLETE_COUNTS, tx);
VERIFY0(space_map_open(&vd->vdev_obsolete_sm,
spa->spa_meta_objset, obsolete_sm_object,
0, vd->vdev_asize, 0));
}
ASSERT(vd->vdev_obsolete_sm != NULL);
ASSERT3U(obsolete_sm_object, ==,
space_map_object(vd->vdev_obsolete_sm));
space_map_write(vd->vdev_obsolete_sm,
vd->vdev_obsolete_segments, SM_ALLOC, SM_NO_VDEVID, tx);
range_tree_vacate(vd->vdev_obsolete_segments, NULL, NULL);
}
int
spa_condense_init(spa_t *spa)
{
int error = zap_lookup(spa->spa_meta_objset,
DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_CONDENSING_INDIRECT, sizeof (uint64_t),
sizeof (spa->spa_condensing_indirect_phys) / sizeof (uint64_t),
&spa->spa_condensing_indirect_phys);
if (error == 0) {
if (spa_writeable(spa)) {
spa->spa_condensing_indirect =
spa_condensing_indirect_create(spa);
}
return (0);
} else if (error == ENOENT) {
return (0);
} else {
return (error);
}
}
void
spa_condense_fini(spa_t *spa)
{
if (spa->spa_condensing_indirect != NULL) {
spa_condensing_indirect_destroy(spa->spa_condensing_indirect);
spa->spa_condensing_indirect = NULL;
}
}
void
spa_start_indirect_condensing_thread(spa_t *spa)
{
ASSERT3P(spa->spa_condense_zthr, ==, NULL);
spa->spa_condense_zthr = zthr_create("z_indirect_condense",
spa_condense_indirect_thread_check,
spa_condense_indirect_thread, spa, minclsyspri);
}
/*
* Gets the obsolete spacemap object from the vdev's ZAP. On success sm_obj
* will contain either the obsolete spacemap object or zero if none exists.
* All other errors are returned to the caller.
*/
int
vdev_obsolete_sm_object(vdev_t *vd, uint64_t *sm_obj)
{
ASSERT0(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER));
if (vd->vdev_top_zap == 0) {
*sm_obj = 0;
return (0);
}
int error = zap_lookup(vd->vdev_spa->spa_meta_objset, vd->vdev_top_zap,
VDEV_TOP_ZAP_INDIRECT_OBSOLETE_SM, sizeof (uint64_t), 1, sm_obj);
if (error == ENOENT) {
*sm_obj = 0;
error = 0;
}
return (error);
}
/*
* Gets the obsolete count are precise spacemap object from the vdev's ZAP.
* On success are_precise will be set to reflect if the counts are precise.
* All other errors are returned to the caller.
*/
int
vdev_obsolete_counts_are_precise(vdev_t *vd, boolean_t *are_precise)
{
ASSERT0(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER));
if (vd->vdev_top_zap == 0) {
*are_precise = B_FALSE;
return (0);
}
uint64_t val = 0;
int error = zap_lookup(vd->vdev_spa->spa_meta_objset, vd->vdev_top_zap,
VDEV_TOP_ZAP_OBSOLETE_COUNTS_ARE_PRECISE, sizeof (val), 1, &val);
if (error == 0) {
*are_precise = (val != 0);
} else if (error == ENOENT) {
*are_precise = B_FALSE;
error = 0;
}
return (error);
}
static void
vdev_indirect_close(vdev_t *vd)
{
(void) vd;
}
static int
vdev_indirect_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize,
uint64_t *logical_ashift, uint64_t *physical_ashift)
{
*psize = *max_psize = vd->vdev_asize +
VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE;
*logical_ashift = vd->vdev_ashift;
*physical_ashift = vd->vdev_physical_ashift;
return (0);
}
typedef struct remap_segment {
vdev_t *rs_vd;
uint64_t rs_offset;
uint64_t rs_asize;
uint64_t rs_split_offset;
list_node_t rs_node;
} remap_segment_t;
static remap_segment_t *
rs_alloc(vdev_t *vd, uint64_t offset, uint64_t asize, uint64_t split_offset)
{
remap_segment_t *rs = kmem_alloc(sizeof (remap_segment_t), KM_SLEEP);
rs->rs_vd = vd;
rs->rs_offset = offset;
rs->rs_asize = asize;
rs->rs_split_offset = split_offset;
return (rs);
}
/*
* Given an indirect vdev and an extent on that vdev, it duplicates the
* physical entries of the indirect mapping that correspond to the extent
* to a new array and returns a pointer to it. In addition, copied_entries
* is populated with the number of mapping entries that were duplicated.
*
* Note that the function assumes that the caller holds vdev_indirect_rwlock.
* This ensures that the mapping won't change due to condensing as we
* copy over its contents.
*
* Finally, since we are doing an allocation, it is up to the caller to
* free the array allocated in this function.
*/
static vdev_indirect_mapping_entry_phys_t *
vdev_indirect_mapping_duplicate_adjacent_entries(vdev_t *vd, uint64_t offset,
uint64_t asize, uint64_t *copied_entries)
{
vdev_indirect_mapping_entry_phys_t *duplicate_mappings = NULL;
vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping;
uint64_t entries = 0;
ASSERT(RW_READ_HELD(&vd->vdev_indirect_rwlock));
vdev_indirect_mapping_entry_phys_t *first_mapping =
vdev_indirect_mapping_entry_for_offset(vim, offset);
ASSERT3P(first_mapping, !=, NULL);
vdev_indirect_mapping_entry_phys_t *m = first_mapping;
while (asize > 0) {
uint64_t size = DVA_GET_ASIZE(&m->vimep_dst);
ASSERT3U(offset, >=, DVA_MAPPING_GET_SRC_OFFSET(m));
ASSERT3U(offset, <, DVA_MAPPING_GET_SRC_OFFSET(m) + size);
uint64_t inner_offset = offset - DVA_MAPPING_GET_SRC_OFFSET(m);
uint64_t inner_size = MIN(asize, size - inner_offset);
offset += inner_size;
asize -= inner_size;
entries++;
m++;
}
size_t copy_length = entries * sizeof (*first_mapping);
duplicate_mappings = kmem_alloc(copy_length, KM_SLEEP);
memcpy(duplicate_mappings, first_mapping, copy_length);
*copied_entries = entries;
return (duplicate_mappings);
}
/*
* Goes through the relevant indirect mappings until it hits a concrete vdev
* and issues the callback. On the way to the concrete vdev, if any other
* indirect vdevs are encountered, then the callback will also be called on
* each of those indirect vdevs. For example, if the segment is mapped to
* segment A on indirect vdev 1, and then segment A on indirect vdev 1 is
* mapped to segment B on concrete vdev 2, then the callback will be called on
* both vdev 1 and vdev 2.
*
* While the callback passed to vdev_indirect_remap() is called on every vdev
* the function encounters, certain callbacks only care about concrete vdevs.
* These types of callbacks should return immediately and explicitly when they
* are called on an indirect vdev.
*
* Because there is a possibility that a DVA section in the indirect device
* has been split into multiple sections in our mapping, we keep track
* of the relevant contiguous segments of the new location (remap_segment_t)
* in a stack. This way we can call the callback for each of the new sections
* created by a single section of the indirect device. Note though, that in
* this scenario the callbacks in each split block won't occur in-order in
* terms of offset, so callers should not make any assumptions about that.
*
* For callbacks that don't handle split blocks and immediately return when
* they encounter them (as is the case for remap_blkptr_cb), the caller can
* assume that its callback will be applied from the first indirect vdev
* encountered to the last one and then the concrete vdev, in that order.
*/
static void
vdev_indirect_remap(vdev_t *vd, uint64_t offset, uint64_t asize,
void (*func)(uint64_t, vdev_t *, uint64_t, uint64_t, void *), void *arg)
{
list_t stack;
spa_t *spa = vd->vdev_spa;
list_create(&stack, sizeof (remap_segment_t),
offsetof(remap_segment_t, rs_node));
for (remap_segment_t *rs = rs_alloc(vd, offset, asize, 0);
rs != NULL; rs = list_remove_head(&stack)) {
vdev_t *v = rs->rs_vd;
uint64_t num_entries = 0;
ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
ASSERT(rs->rs_asize > 0);
/*
* Note: As this function can be called from open context
* (e.g. zio_read()), we need the following rwlock to
* prevent the mapping from being changed by condensing.
*
* So we grab the lock and we make a copy of the entries
* that are relevant to the extent that we are working on.
* Once that is done, we drop the lock and iterate over
* our copy of the mapping. Once we are done with the with
* the remap segment and we free it, we also free our copy
* of the indirect mapping entries that are relevant to it.
*
* This way we don't need to wait until the function is
* finished with a segment, to condense it. In addition, we
* don't need a recursive rwlock for the case that a call to
* vdev_indirect_remap() needs to call itself (through the
* codepath of its callback) for the same vdev in the middle
* of its execution.
*/
rw_enter(&v->vdev_indirect_rwlock, RW_READER);
ASSERT3P(v->vdev_indirect_mapping, !=, NULL);
vdev_indirect_mapping_entry_phys_t *mapping =
vdev_indirect_mapping_duplicate_adjacent_entries(v,
rs->rs_offset, rs->rs_asize, &num_entries);
ASSERT3P(mapping, !=, NULL);
ASSERT3U(num_entries, >, 0);
rw_exit(&v->vdev_indirect_rwlock);
for (uint64_t i = 0; i < num_entries; i++) {
/*
* Note: the vdev_indirect_mapping can not change
* while we are running. It only changes while the
* removal is in progress, and then only from syncing
* context. While a removal is in progress, this
* function is only called for frees, which also only
* happen from syncing context.
*/
vdev_indirect_mapping_entry_phys_t *m = &mapping[i];
ASSERT3P(m, !=, NULL);
ASSERT3U(rs->rs_asize, >, 0);
uint64_t size = DVA_GET_ASIZE(&m->vimep_dst);
uint64_t dst_offset = DVA_GET_OFFSET(&m->vimep_dst);
uint64_t dst_vdev = DVA_GET_VDEV(&m->vimep_dst);
ASSERT3U(rs->rs_offset, >=,
DVA_MAPPING_GET_SRC_OFFSET(m));
ASSERT3U(rs->rs_offset, <,
DVA_MAPPING_GET_SRC_OFFSET(m) + size);
ASSERT3U(dst_vdev, !=, v->vdev_id);
uint64_t inner_offset = rs->rs_offset -
DVA_MAPPING_GET_SRC_OFFSET(m);
uint64_t inner_size =
MIN(rs->rs_asize, size - inner_offset);
vdev_t *dst_v = vdev_lookup_top(spa, dst_vdev);
ASSERT3P(dst_v, !=, NULL);
if (dst_v->vdev_ops == &vdev_indirect_ops) {
list_insert_head(&stack,
rs_alloc(dst_v, dst_offset + inner_offset,
inner_size, rs->rs_split_offset));
}
if ((zfs_flags & ZFS_DEBUG_INDIRECT_REMAP) &&
IS_P2ALIGNED(inner_size, 2 * SPA_MINBLOCKSIZE)) {
/*
* Note: This clause exists only solely for
* testing purposes. We use it to ensure that
* split blocks work and that the callbacks
* using them yield the same result if issued
* in reverse order.
*/
uint64_t inner_half = inner_size / 2;
func(rs->rs_split_offset + inner_half, dst_v,
dst_offset + inner_offset + inner_half,
inner_half, arg);
func(rs->rs_split_offset, dst_v,
dst_offset + inner_offset,
inner_half, arg);
} else {
func(rs->rs_split_offset, dst_v,
dst_offset + inner_offset,
inner_size, arg);
}
rs->rs_offset += inner_size;
rs->rs_asize -= inner_size;
rs->rs_split_offset += inner_size;
}
VERIFY0(rs->rs_asize);
kmem_free(mapping, num_entries * sizeof (*mapping));
kmem_free(rs, sizeof (remap_segment_t));
}
list_destroy(&stack);
}
static void
vdev_indirect_child_io_done(zio_t *zio)
{
zio_t *pio = zio->io_private;
mutex_enter(&pio->io_lock);
pio->io_error = zio_worst_error(pio->io_error, zio->io_error);
mutex_exit(&pio->io_lock);
abd_free(zio->io_abd);
}
/*
* This is a callback for vdev_indirect_remap() which allocates an
* indirect_split_t for each split segment and adds it to iv_splits.
*/
static void
vdev_indirect_gather_splits(uint64_t split_offset, vdev_t *vd, uint64_t offset,
uint64_t size, void *arg)
{
zio_t *zio = arg;
indirect_vsd_t *iv = zio->io_vsd;
ASSERT3P(vd, !=, NULL);
if (vd->vdev_ops == &vdev_indirect_ops)
return;
int n = 1;
if (vd->vdev_ops == &vdev_mirror_ops)
n = vd->vdev_children;
indirect_split_t *is =
kmem_zalloc(offsetof(indirect_split_t, is_child[n]), KM_SLEEP);
is->is_children = n;
is->is_size = size;
is->is_split_offset = split_offset;
is->is_target_offset = offset;
is->is_vdev = vd;
list_create(&is->is_unique_child, sizeof (indirect_child_t),
offsetof(indirect_child_t, ic_node));
/*
* Note that we only consider multiple copies of the data for
* *mirror* vdevs. We don't for "replacing" or "spare" vdevs, even
* though they use the same ops as mirror, because there's only one
* "good" copy under the replacing/spare.
*/
if (vd->vdev_ops == &vdev_mirror_ops) {
for (int i = 0; i < n; i++) {
is->is_child[i].ic_vdev = vd->vdev_child[i];
list_link_init(&is->is_child[i].ic_node);
}
} else {
is->is_child[0].ic_vdev = vd;
}
list_insert_tail(&iv->iv_splits, is);
}
static void
vdev_indirect_read_split_done(zio_t *zio)
{
indirect_child_t *ic = zio->io_private;
if (zio->io_error != 0) {
/*
* Clear ic_data to indicate that we do not have data for this
* child.
*/
abd_free(ic->ic_data);
ic->ic_data = NULL;
}
}
/*
* Issue reads for all copies (mirror children) of all splits.
*/
static void
vdev_indirect_read_all(zio_t *zio)
{
indirect_vsd_t *iv = zio->io_vsd;
ASSERT3U(zio->io_type, ==, ZIO_TYPE_READ);
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
for (int i = 0; i < is->is_children; i++) {
indirect_child_t *ic = &is->is_child[i];
if (!vdev_readable(ic->ic_vdev))
continue;
/*
* If a child is missing the data, set ic_error. Used
* in vdev_indirect_repair(). We perform the read
* nevertheless which provides the opportunity to
* reconstruct the split block if at all possible.
*/
if (vdev_dtl_contains(ic->ic_vdev, DTL_MISSING,
zio->io_txg, 1))
ic->ic_error = SET_ERROR(ESTALE);
ic->ic_data = abd_alloc_sametype(zio->io_abd,
is->is_size);
ic->ic_duplicate = NULL;
zio_nowait(zio_vdev_child_io(zio, NULL,
ic->ic_vdev, is->is_target_offset, ic->ic_data,
is->is_size, zio->io_type, zio->io_priority, 0,
vdev_indirect_read_split_done, ic));
}
}
iv->iv_reconstruct = B_TRUE;
}
static void
vdev_indirect_io_start(zio_t *zio)
{
spa_t *spa __maybe_unused = zio->io_spa;
indirect_vsd_t *iv = kmem_zalloc(sizeof (*iv), KM_SLEEP);
list_create(&iv->iv_splits,
sizeof (indirect_split_t), offsetof(indirect_split_t, is_node));
zio->io_vsd = iv;
zio->io_vsd_ops = &vdev_indirect_vsd_ops;
ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0);
if (zio->io_type != ZIO_TYPE_READ) {
ASSERT3U(zio->io_type, ==, ZIO_TYPE_WRITE);
/*
* Note: this code can handle other kinds of writes,
* but we don't expect them.
*/
ASSERT((zio->io_flags & (ZIO_FLAG_SELF_HEAL |
ZIO_FLAG_RESILVER | ZIO_FLAG_INDUCE_DAMAGE)) != 0);
}
vdev_indirect_remap(zio->io_vd, zio->io_offset, zio->io_size,
vdev_indirect_gather_splits, zio);
indirect_split_t *first = list_head(&iv->iv_splits);
ASSERT3P(first, !=, NULL);
if (first->is_size == zio->io_size) {
/*
* This is not a split block; we are pointing to the entire
* data, which will checksum the same as the original data.
* Pass the BP down so that the child i/o can verify the
* checksum, and try a different location if available
* (e.g. on a mirror).
*
* While this special case could be handled the same as the
* general (split block) case, doing it this way ensures
* that the vast majority of blocks on indirect vdevs
* (which are not split) are handled identically to blocks
* on non-indirect vdevs. This allows us to be less strict
* about performance in the general (but rare) case.
*/
ASSERT0(first->is_split_offset);
ASSERT3P(list_next(&iv->iv_splits, first), ==, NULL);
zio_nowait(zio_vdev_child_io(zio, zio->io_bp,
first->is_vdev, first->is_target_offset,
abd_get_offset(zio->io_abd, 0),
zio->io_size, zio->io_type, zio->io_priority, 0,
vdev_indirect_child_io_done, zio));
} else {
iv->iv_split_block = B_TRUE;
if (zio->io_type == ZIO_TYPE_READ &&
zio->io_flags & (ZIO_FLAG_SCRUB | ZIO_FLAG_RESILVER)) {
/*
* Read all copies. Note that for simplicity,
* we don't bother consulting the DTL in the
* resilver case.
*/
vdev_indirect_read_all(zio);
} else {
/*
* If this is a read zio, we read one copy of each
* split segment, from the top-level vdev. Since
* we don't know the checksum of each split
* individually, the child zio can't ensure that
* we get the right data. E.g. if it's a mirror,
* it will just read from a random (healthy) leaf
* vdev. We have to verify the checksum in
* vdev_indirect_io_done().
*
* For write zios, the vdev code will ensure we write
* to all children.
*/
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
zio_nowait(zio_vdev_child_io(zio, NULL,
is->is_vdev, is->is_target_offset,
- abd_get_offset(zio->io_abd,
- is->is_split_offset), is->is_size,
- zio->io_type, zio->io_priority, 0,
+ abd_get_offset_size(zio->io_abd,
+ is->is_split_offset, is->is_size),
+ is->is_size, zio->io_type,
+ zio->io_priority, 0,
vdev_indirect_child_io_done, zio));
}
}
}
zio_execute(zio);
}
/*
* Report a checksum error for a child.
*/
static void
vdev_indirect_checksum_error(zio_t *zio,
indirect_split_t *is, indirect_child_t *ic)
{
vdev_t *vd = ic->ic_vdev;
if (zio->io_flags & ZIO_FLAG_SPECULATIVE)
return;
mutex_enter(&vd->vdev_stat_lock);
vd->vdev_stat.vs_checksum_errors++;
mutex_exit(&vd->vdev_stat_lock);
zio_bad_cksum_t zbc = {{{ 0 }}};
abd_t *bad_abd = ic->ic_data;
abd_t *good_abd = is->is_good_child->ic_data;
(void) zfs_ereport_post_checksum(zio->io_spa, vd, NULL, zio,
is->is_target_offset, is->is_size, good_abd, bad_abd, &zbc);
}
/*
* Issue repair i/os for any incorrect copies. We do this by comparing
* each split segment's correct data (is_good_child's ic_data) with each
* other copy of the data. If they differ, then we overwrite the bad data
* with the good copy. The DTL is checked in vdev_indirect_read_all() and
* if a vdev is missing a copy of the data we set ic_error and the read is
* performed. This provides the opportunity to reconstruct the split block
* if at all possible. ic_error is checked here and if set it suppresses
* incrementing the checksum counter. Aside from this DTLs are not checked,
* which simplifies this code and also issues the optimal number of writes
* (based on which copies actually read bad data, as opposed to which we
* think might be wrong). For the same reason, we always use
* ZIO_FLAG_SELF_HEAL, to bypass the DTL check in zio_vdev_io_start().
*/
static void
vdev_indirect_repair(zio_t *zio)
{
indirect_vsd_t *iv = zio->io_vsd;
if (!spa_writeable(zio->io_spa))
return;
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
for (int c = 0; c < is->is_children; c++) {
indirect_child_t *ic = &is->is_child[c];
if (ic == is->is_good_child)
continue;
if (ic->ic_data == NULL)
continue;
if (ic->ic_duplicate == is->is_good_child)
continue;
zio_nowait(zio_vdev_child_io(zio, NULL,
ic->ic_vdev, is->is_target_offset,
is->is_good_child->ic_data, is->is_size,
ZIO_TYPE_WRITE, ZIO_PRIORITY_ASYNC_WRITE,
ZIO_FLAG_IO_REPAIR | ZIO_FLAG_SELF_HEAL,
NULL, NULL));
/*
* If ic_error is set the current child does not have
* a copy of the data, so suppress incrementing the
* checksum counter.
*/
if (ic->ic_error == ESTALE)
continue;
vdev_indirect_checksum_error(zio, is, ic);
}
}
}
/*
* Report checksum errors on all children that we read from.
*/
static void
vdev_indirect_all_checksum_errors(zio_t *zio)
{
indirect_vsd_t *iv = zio->io_vsd;
if (zio->io_flags & ZIO_FLAG_SPECULATIVE)
return;
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
for (int c = 0; c < is->is_children; c++) {
indirect_child_t *ic = &is->is_child[c];
if (ic->ic_data == NULL)
continue;
vdev_t *vd = ic->ic_vdev;
mutex_enter(&vd->vdev_stat_lock);
vd->vdev_stat.vs_checksum_errors++;
mutex_exit(&vd->vdev_stat_lock);
(void) zfs_ereport_post_checksum(zio->io_spa, vd,
NULL, zio, is->is_target_offset, is->is_size,
NULL, NULL, NULL);
}
}
}
/*
* Copy data from all the splits to a main zio then validate the checksum.
* If then checksum is successfully validated return success.
*/
static int
vdev_indirect_splits_checksum_validate(indirect_vsd_t *iv, zio_t *zio)
{
zio_bad_cksum_t zbc;
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
ASSERT3P(is->is_good_child->ic_data, !=, NULL);
ASSERT3P(is->is_good_child->ic_duplicate, ==, NULL);
abd_copy_off(zio->io_abd, is->is_good_child->ic_data,
is->is_split_offset, 0, is->is_size);
}
return (zio_checksum_error(zio, &zbc));
}
/*
* There are relatively few possible combinations making it feasible to
* deterministically check them all. We do this by setting the good_child
* to the next unique split version. If we reach the end of the list then
* "carry over" to the next unique split version (like counting in base
* is_unique_children, but each digit can have a different base).
*/
static int
vdev_indirect_splits_enumerate_all(indirect_vsd_t *iv, zio_t *zio)
{
boolean_t more = B_TRUE;
iv->iv_attempts = 0;
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is))
is->is_good_child = list_head(&is->is_unique_child);
while (more == B_TRUE) {
iv->iv_attempts++;
more = B_FALSE;
if (vdev_indirect_splits_checksum_validate(iv, zio) == 0)
return (0);
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
is->is_good_child = list_next(&is->is_unique_child,
is->is_good_child);
if (is->is_good_child != NULL) {
more = B_TRUE;
break;
}
is->is_good_child = list_head(&is->is_unique_child);
}
}
ASSERT3S(iv->iv_attempts, <=, iv->iv_unique_combinations);
return (SET_ERROR(ECKSUM));
}
/*
* There are too many combinations to try all of them in a reasonable amount
* of time. So try a fixed number of random combinations from the unique
* split versions, after which we'll consider the block unrecoverable.
*/
static int
vdev_indirect_splits_enumerate_randomly(indirect_vsd_t *iv, zio_t *zio)
{
iv->iv_attempts = 0;
while (iv->iv_attempts < iv->iv_attempts_max) {
iv->iv_attempts++;
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
indirect_child_t *ic = list_head(&is->is_unique_child);
int children = is->is_unique_children;
for (int i = random_in_range(children); i > 0; i--)
ic = list_next(&is->is_unique_child, ic);
ASSERT3P(ic, !=, NULL);
is->is_good_child = ic;
}
if (vdev_indirect_splits_checksum_validate(iv, zio) == 0)
return (0);
}
return (SET_ERROR(ECKSUM));
}
/*
* This is a validation function for reconstruction. It randomly selects
* a good combination, if one can be found, and then it intentionally
* damages all other segment copes by zeroing them. This forces the
* reconstruction algorithm to locate the one remaining known good copy.
*/
static int
vdev_indirect_splits_damage(indirect_vsd_t *iv, zio_t *zio)
{
int error;
/* Presume all the copies are unique for initial selection. */
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
is->is_unique_children = 0;
for (int i = 0; i < is->is_children; i++) {
indirect_child_t *ic = &is->is_child[i];
if (ic->ic_data != NULL) {
is->is_unique_children++;
list_insert_tail(&is->is_unique_child, ic);
}
}
if (list_is_empty(&is->is_unique_child)) {
error = SET_ERROR(EIO);
goto out;
}
}
/*
* Set each is_good_child to a randomly-selected child which
* is known to contain validated data.
*/
error = vdev_indirect_splits_enumerate_randomly(iv, zio);
if (error)
goto out;
/*
* Damage all but the known good copy by zeroing it. This will
* result in two or less unique copies per indirect_child_t.
* Both may need to be checked in order to reconstruct the block.
* Set iv->iv_attempts_max such that all unique combinations will
* enumerated, but limit the damage to at most 12 indirect splits.
*/
iv->iv_attempts_max = 1;
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
for (int c = 0; c < is->is_children; c++) {
indirect_child_t *ic = &is->is_child[c];
if (ic == is->is_good_child)
continue;
if (ic->ic_data == NULL)
continue;
abd_zero(ic->ic_data, abd_get_size(ic->ic_data));
}
iv->iv_attempts_max *= 2;
if (iv->iv_attempts_max >= (1ULL << 12)) {
iv->iv_attempts_max = UINT64_MAX;
break;
}
}
out:
/* Empty the unique children lists so they can be reconstructed. */
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
indirect_child_t *ic;
while ((ic = list_head(&is->is_unique_child)) != NULL)
list_remove(&is->is_unique_child, ic);
is->is_unique_children = 0;
}
return (error);
}
/*
* This function is called when we have read all copies of the data and need
* to try to find a combination of copies that gives us the right checksum.
*
* If we pointed to any mirror vdevs, this effectively does the job of the
* mirror. The mirror vdev code can't do its own job because we don't know
* the checksum of each split segment individually.
*
* We have to try every unique combination of copies of split segments, until
* we find one that checksums correctly. Duplicate segment copies are first
* identified and latter skipped during reconstruction. This optimization
* reduces the search space and ensures that of the remaining combinations
* at most one is correct.
*
* When the total number of combinations is small they can all be checked.
* For example, if we have 3 segments in the split, and each points to a
* 2-way mirror with unique copies, we will have the following pieces of data:
*
* | mirror child
* split | [0] [1]
* ======|=====================
* A | data_A_0 data_A_1
* B | data_B_0 data_B_1
* C | data_C_0 data_C_1
*
* We will try the following (mirror children)^(number of splits) (2^3=8)
* combinations, which is similar to bitwise-little-endian counting in
* binary. In general each "digit" corresponds to a split segment, and the
* base of each digit is is_children, which can be different for each
* digit.
*
* "low bit" "high bit"
* v v
* data_A_0 data_B_0 data_C_0
* data_A_1 data_B_0 data_C_0
* data_A_0 data_B_1 data_C_0
* data_A_1 data_B_1 data_C_0
* data_A_0 data_B_0 data_C_1
* data_A_1 data_B_0 data_C_1
* data_A_0 data_B_1 data_C_1
* data_A_1 data_B_1 data_C_1
*
* Note that the split segments may be on the same or different top-level
* vdevs. In either case, we may need to try lots of combinations (see
* zfs_reconstruct_indirect_combinations_max). This ensures that if a mirror
* has small silent errors on all of its children, we can still reconstruct
* the correct data, as long as those errors are at sufficiently-separated
* offsets (specifically, separated by the largest block size - default of
* 128KB, but up to 16MB).
*/
static void
vdev_indirect_reconstruct_io_done(zio_t *zio)
{
indirect_vsd_t *iv = zio->io_vsd;
boolean_t known_good = B_FALSE;
int error;
iv->iv_unique_combinations = 1;
iv->iv_attempts_max = UINT64_MAX;
if (zfs_reconstruct_indirect_combinations_max > 0)
iv->iv_attempts_max = zfs_reconstruct_indirect_combinations_max;
/*
* If nonzero, every 1/x blocks will be damaged, in order to validate
* reconstruction when there are split segments with damaged copies.
* Known_good will be TRUE when reconstruction is known to be possible.
*/
if (zfs_reconstruct_indirect_damage_fraction != 0 &&
random_in_range(zfs_reconstruct_indirect_damage_fraction) == 0)
known_good = (vdev_indirect_splits_damage(iv, zio) == 0);
/*
* Determine the unique children for a split segment and add them
* to the is_unique_child list. By restricting reconstruction
* to these children, only unique combinations will be considered.
* This can vastly reduce the search space when there are a large
* number of indirect splits.
*/
for (indirect_split_t *is = list_head(&iv->iv_splits);
is != NULL; is = list_next(&iv->iv_splits, is)) {
is->is_unique_children = 0;
for (int i = 0; i < is->is_children; i++) {
indirect_child_t *ic_i = &is->is_child[i];
if (ic_i->ic_data == NULL ||
ic_i->ic_duplicate != NULL)
continue;
for (int j = i + 1; j < is->is_children; j++) {
indirect_child_t *ic_j = &is->is_child[j];
if (ic_j->ic_data == NULL ||
ic_j->ic_duplicate != NULL)
continue;
if (abd_cmp(ic_i->ic_data, ic_j->ic_data) == 0)
ic_j->ic_duplicate = ic_i;
}
is->is_unique_children++;
list_insert_tail(&is->is_unique_child, ic_i);
}
/* Reconstruction is impossible, no valid children */
EQUIV(list_is_empty(&is->is_unique_child),
is->is_unique_children == 0);
if (list_is_empty(&is->is_unique_child)) {
zio->io_error = EIO;
vdev_indirect_all_checksum_errors(zio);
zio_checksum_verified(zio);
return;
}
iv->iv_unique_combinations *= is->is_unique_children;
}
if (iv->iv_unique_combinations <= iv->iv_attempts_max)
error = vdev_indirect_splits_enumerate_all(iv, zio);
else
error = vdev_indirect_splits_enumerate_randomly(iv, zio);
if (error != 0) {
/* All attempted combinations failed. */
ASSERT3B(known_good, ==, B_FALSE);
zio->io_error = error;
vdev_indirect_all_checksum_errors(zio);
} else {
/*
* The checksum has been successfully validated. Issue
* repair I/Os to any copies of splits which don't match
* the validated version.
*/
ASSERT0(vdev_indirect_splits_checksum_validate(iv, zio));
vdev_indirect_repair(zio);
zio_checksum_verified(zio);
}
}
static void
vdev_indirect_io_done(zio_t *zio)
{
indirect_vsd_t *iv = zio->io_vsd;
if (iv->iv_reconstruct) {
/*
* We have read all copies of the data (e.g. from mirrors),
* either because this was a scrub/resilver, or because the
* one-copy read didn't checksum correctly.
*/
vdev_indirect_reconstruct_io_done(zio);
return;
}
if (!iv->iv_split_block) {
/*
* This was not a split block, so we passed the BP down,
* and the checksum was handled by the (one) child zio.
*/
return;
}
zio_bad_cksum_t zbc;
int ret = zio_checksum_error(zio, &zbc);
if (ret == 0) {
zio_checksum_verified(zio);
return;
}
/*
* The checksum didn't match. Read all copies of all splits, and
* then we will try to reconstruct. The next time
* vdev_indirect_io_done() is called, iv_reconstruct will be set.
*/
vdev_indirect_read_all(zio);
zio_vdev_io_redone(zio);
}
vdev_ops_t vdev_indirect_ops = {
.vdev_op_init = NULL,
.vdev_op_fini = NULL,
.vdev_op_open = vdev_indirect_open,
.vdev_op_close = vdev_indirect_close,
.vdev_op_asize = vdev_default_asize,
.vdev_op_min_asize = vdev_default_min_asize,
.vdev_op_min_alloc = NULL,
.vdev_op_io_start = vdev_indirect_io_start,
.vdev_op_io_done = vdev_indirect_io_done,
.vdev_op_state_change = NULL,
.vdev_op_need_resilver = NULL,
.vdev_op_hold = NULL,
.vdev_op_rele = NULL,
.vdev_op_remap = vdev_indirect_remap,
.vdev_op_xlate = NULL,
.vdev_op_rebuild_asize = NULL,
.vdev_op_metaslab_init = NULL,
.vdev_op_config_generate = NULL,
.vdev_op_nparity = NULL,
.vdev_op_ndisks = NULL,
.vdev_op_type = VDEV_TYPE_INDIRECT, /* name of this vdev type */
.vdev_op_leaf = B_FALSE /* leaf vdev */
};
EXPORT_SYMBOL(spa_condense_fini);
EXPORT_SYMBOL(spa_start_indirect_condensing_thread);
EXPORT_SYMBOL(spa_condense_indirect_start_sync);
EXPORT_SYMBOL(spa_condense_init);
EXPORT_SYMBOL(spa_vdev_indirect_mark_obsolete);
EXPORT_SYMBOL(vdev_indirect_mark_obsolete);
EXPORT_SYMBOL(vdev_indirect_should_condense);
EXPORT_SYMBOL(vdev_indirect_sync_obsolete);
EXPORT_SYMBOL(vdev_obsolete_counts_are_precise);
EXPORT_SYMBOL(vdev_obsolete_sm_object);
/* BEGIN CSTYLED */
ZFS_MODULE_PARAM(zfs_condense, zfs_condense_, indirect_vdevs_enable, INT,
ZMOD_RW, "Whether to attempt condensing indirect vdev mappings");
ZFS_MODULE_PARAM(zfs_condense, zfs_condense_, indirect_obsolete_pct, UINT,
ZMOD_RW,
"Minimum obsolete percent of bytes in the mapping "
"to attempt condensing");
ZFS_MODULE_PARAM(zfs_condense, zfs_condense_, min_mapping_bytes, U64, ZMOD_RW,
"Don't bother condensing if the mapping uses less than this amount of "
"memory");
ZFS_MODULE_PARAM(zfs_condense, zfs_condense_, max_obsolete_bytes, U64,
ZMOD_RW,
"Minimum size obsolete spacemap to attempt condensing");
ZFS_MODULE_PARAM(zfs_condense, zfs_condense_, indirect_commit_entry_delay_ms,
UINT, ZMOD_RW,
"Used by tests to ensure certain actions happen in the middle of a "
"condense. A maximum value of 1 should be sufficient.");
ZFS_MODULE_PARAM(zfs_reconstruct, zfs_reconstruct_, indirect_combinations_max,
UINT, ZMOD_RW,
"Maximum number of combinations when reconstructing split segments");
/* END CSTYLED */
diff --git a/sys/contrib/openzfs/module/zfs/vdev_label.c b/sys/contrib/openzfs/module/zfs/vdev_label.c
index f61be65a2c72..85c7134ca4c4 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_label.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_label.c
@@ -1,2022 +1,2028 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2020 by Delphix. All rights reserved.
* Copyright (c) 2017, Intel Corporation.
*/
/*
* Virtual Device Labels
* ---------------------
*
* The vdev label serves several distinct purposes:
*
* 1. Uniquely identify this device as part of a ZFS pool and confirm its
* identity within the pool.
*
* 2. Verify that all the devices given in a configuration are present
* within the pool.
*
* 3. Determine the uberblock for the pool.
*
* 4. In case of an import operation, determine the configuration of the
* toplevel vdev of which it is a part.
*
* 5. If an import operation cannot find all the devices in the pool,
* provide enough information to the administrator to determine which
* devices are missing.
*
* It is important to note that while the kernel is responsible for writing the
* label, it only consumes the information in the first three cases. The
* latter information is only consumed in userland when determining the
* configuration to import a pool.
*
*
* Label Organization
* ------------------
*
* Before describing the contents of the label, it's important to understand how
* the labels are written and updated with respect to the uberblock.
*
* When the pool configuration is altered, either because it was newly created
* or a device was added, we want to update all the labels such that we can deal
* with fatal failure at any point. To this end, each disk has two labels which
* are updated before and after the uberblock is synced. Assuming we have
* labels and an uberblock with the following transaction groups:
*
* L1 UB L2
* +------+ +------+ +------+
* | | | | | |
* | t10 | | t10 | | t10 |
* | | | | | |
* +------+ +------+ +------+
*
* In this stable state, the labels and the uberblock were all updated within
* the same transaction group (10). Each label is mirrored and checksummed, so
* that we can detect when we fail partway through writing the label.
*
* In order to identify which labels are valid, the labels are written in the
* following manner:
*
* 1. For each vdev, update 'L1' to the new label
* 2. Update the uberblock
* 3. For each vdev, update 'L2' to the new label
*
* Given arbitrary failure, we can determine the correct label to use based on
* the transaction group. If we fail after updating L1 but before updating the
* UB, we will notice that L1's transaction group is greater than the uberblock,
* so L2 must be valid. If we fail after writing the uberblock but before
* writing L2, we will notice that L2's transaction group is less than L1, and
* therefore L1 is valid.
*
* Another added complexity is that not every label is updated when the config
* is synced. If we add a single device, we do not want to have to re-write
* every label for every device in the pool. This means that both L1 and L2 may
* be older than the pool uberblock, because the necessary information is stored
* on another vdev.
*
*
* On-disk Format
* --------------
*
* The vdev label consists of two distinct parts, and is wrapped within the
* vdev_label_t structure. The label includes 8k of padding to permit legacy
* VTOC disk labels, but is otherwise ignored.
*
* The first half of the label is a packed nvlist which contains pool wide
* properties, per-vdev properties, and configuration information. It is
* described in more detail below.
*
* The latter half of the label consists of a redundant array of uberblocks.
* These uberblocks are updated whenever a transaction group is committed,
* or when the configuration is updated. When a pool is loaded, we scan each
* vdev for the 'best' uberblock.
*
*
* Configuration Information
* -------------------------
*
* The nvlist describing the pool and vdev contains the following elements:
*
* version ZFS on-disk version
* name Pool name
* state Pool state
* txg Transaction group in which this label was written
* pool_guid Unique identifier for this pool
* vdev_tree An nvlist describing vdev tree.
* features_for_read
* An nvlist of the features necessary for reading the MOS.
*
* Each leaf device label also contains the following:
*
* top_guid Unique ID for top-level vdev in which this is contained
* guid Unique ID for the leaf vdev
*
* The 'vs' configuration follows the format described in 'spa_config.c'.
*/
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/dmu.h>
#include <sys/zap.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
#include <sys/vdev_draid.h>
#include <sys/uberblock_impl.h>
#include <sys/metaslab.h>
#include <sys/metaslab_impl.h>
#include <sys/zio.h>
#include <sys/dsl_scan.h>
#include <sys/abd.h>
#include <sys/fs/zfs.h>
#include <sys/byteorder.h>
#include <sys/zfs_bootenv.h>
/*
* Basic routines to read and write from a vdev label.
* Used throughout the rest of this file.
*/
uint64_t
vdev_label_offset(uint64_t psize, int l, uint64_t offset)
{
ASSERT(offset < sizeof (vdev_label_t));
ASSERT(P2PHASE_TYPED(psize, sizeof (vdev_label_t), uint64_t) == 0);
return (offset + l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
0 : psize - VDEV_LABELS * sizeof (vdev_label_t)));
}
/*
* Returns back the vdev label associated with the passed in offset.
*/
int
vdev_label_number(uint64_t psize, uint64_t offset)
{
int l;
if (offset >= psize - VDEV_LABEL_END_SIZE) {
offset -= psize - VDEV_LABEL_END_SIZE;
offset += (VDEV_LABELS / 2) * sizeof (vdev_label_t);
}
l = offset / sizeof (vdev_label_t);
return (l < VDEV_LABELS ? l : -1);
}
static void
vdev_label_read(zio_t *zio, vdev_t *vd, int l, abd_t *buf, uint64_t offset,
uint64_t size, zio_done_func_t *done, void *private, int flags)
{
ASSERT(
spa_config_held(zio->io_spa, SCL_STATE, RW_READER) == SCL_STATE ||
spa_config_held(zio->io_spa, SCL_STATE, RW_WRITER) == SCL_STATE);
ASSERT(flags & ZIO_FLAG_CONFIG_WRITER);
zio_nowait(zio_read_phys(zio, vd,
vdev_label_offset(vd->vdev_psize, l, offset),
size, buf, ZIO_CHECKSUM_LABEL, done, private,
ZIO_PRIORITY_SYNC_READ, flags, B_TRUE));
}
void
vdev_label_write(zio_t *zio, vdev_t *vd, int l, abd_t *buf, uint64_t offset,
uint64_t size, zio_done_func_t *done, void *private, int flags)
{
ASSERT(
spa_config_held(zio->io_spa, SCL_STATE, RW_READER) == SCL_STATE ||
spa_config_held(zio->io_spa, SCL_STATE, RW_WRITER) == SCL_STATE);
ASSERT(flags & ZIO_FLAG_CONFIG_WRITER);
zio_nowait(zio_write_phys(zio, vd,
vdev_label_offset(vd->vdev_psize, l, offset),
size, buf, ZIO_CHECKSUM_LABEL, done, private,
ZIO_PRIORITY_SYNC_WRITE, flags, B_TRUE));
}
/*
* Generate the nvlist representing this vdev's stats
*/
void
vdev_config_generate_stats(vdev_t *vd, nvlist_t *nv)
{
nvlist_t *nvx;
vdev_stat_t *vs;
vdev_stat_ex_t *vsx;
vs = kmem_alloc(sizeof (*vs), KM_SLEEP);
vsx = kmem_alloc(sizeof (*vsx), KM_SLEEP);
vdev_get_stats_ex(vd, vs, vsx);
fnvlist_add_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
(uint64_t *)vs, sizeof (*vs) / sizeof (uint64_t));
/*
* Add extended stats into a special extended stats nvlist. This keeps
* all the extended stats nicely grouped together. The extended stats
* nvlist is then added to the main nvlist.
*/
nvx = fnvlist_alloc();
/* ZIOs in flight to disk */
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_SYNC_READ]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_SYNC_WRITE]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_ASYNC_READ]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_ASYNC_WRITE]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_SCRUB]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_TRIM]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE,
vsx->vsx_active_queue[ZIO_PRIORITY_REBUILD]);
/* ZIOs pending */
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_SYNC_READ]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_SYNC_WRITE]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_ASYNC_READ]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_ASYNC_WRITE]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_SCRUB]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_TRIM]);
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_REBUILD_PEND_QUEUE,
vsx->vsx_pend_queue[ZIO_PRIORITY_REBUILD]);
/* Histograms */
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO,
vsx->vsx_total_histo[ZIO_TYPE_READ],
ARRAY_SIZE(vsx->vsx_total_histo[ZIO_TYPE_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO,
vsx->vsx_total_histo[ZIO_TYPE_WRITE],
ARRAY_SIZE(vsx->vsx_total_histo[ZIO_TYPE_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO,
vsx->vsx_disk_histo[ZIO_TYPE_READ],
ARRAY_SIZE(vsx->vsx_disk_histo[ZIO_TYPE_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO,
vsx->vsx_disk_histo[ZIO_TYPE_WRITE],
ARRAY_SIZE(vsx->vsx_disk_histo[ZIO_TYPE_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_READ],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_WRITE],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_READ],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_WRITE],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_ASYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_SCRUB],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_SCRUB]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_TRIM],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_TRIM]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO,
vsx->vsx_queue_histo[ZIO_PRIORITY_REBUILD],
ARRAY_SIZE(vsx->vsx_queue_histo[ZIO_PRIORITY_REBUILD]));
/* Request sizes */
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_READ],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_WRITE],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_READ],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_WRITE],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_ASYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_SCRUB],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_SCRUB]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_TRIM],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_TRIM]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_IND_REBUILD_HISTO,
vsx->vsx_ind_histo[ZIO_PRIORITY_REBUILD],
ARRAY_SIZE(vsx->vsx_ind_histo[ZIO_PRIORITY_REBUILD]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_READ],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_WRITE],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_READ],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_READ]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_WRITE],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_ASYNC_WRITE]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_SCRUB],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_SCRUB]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_TRIM],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_TRIM]));
fnvlist_add_uint64_array(nvx, ZPOOL_CONFIG_VDEV_AGG_REBUILD_HISTO,
vsx->vsx_agg_histo[ZIO_PRIORITY_REBUILD],
ARRAY_SIZE(vsx->vsx_agg_histo[ZIO_PRIORITY_REBUILD]));
/* IO delays */
fnvlist_add_uint64(nvx, ZPOOL_CONFIG_VDEV_SLOW_IOS, vs->vs_slow_ios);
/* Add extended stats nvlist to main nvlist */
fnvlist_add_nvlist(nv, ZPOOL_CONFIG_VDEV_STATS_EX, nvx);
fnvlist_free(nvx);
kmem_free(vs, sizeof (*vs));
kmem_free(vsx, sizeof (*vsx));
}
static void
root_vdev_actions_getprogress(vdev_t *vd, nvlist_t *nvl)
{
spa_t *spa = vd->vdev_spa;
if (vd != spa->spa_root_vdev)
return;
/* provide either current or previous scan information */
pool_scan_stat_t ps;
if (spa_scan_get_stats(spa, &ps) == 0) {
fnvlist_add_uint64_array(nvl,
ZPOOL_CONFIG_SCAN_STATS, (uint64_t *)&ps,
sizeof (pool_scan_stat_t) / sizeof (uint64_t));
}
pool_removal_stat_t prs;
if (spa_removal_get_stats(spa, &prs) == 0) {
fnvlist_add_uint64_array(nvl,
ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t *)&prs,
sizeof (prs) / sizeof (uint64_t));
}
pool_checkpoint_stat_t pcs;
if (spa_checkpoint_get_stats(spa, &pcs) == 0) {
fnvlist_add_uint64_array(nvl,
ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t *)&pcs,
sizeof (pcs) / sizeof (uint64_t));
}
}
static void
top_vdev_actions_getprogress(vdev_t *vd, nvlist_t *nvl)
{
if (vd == vd->vdev_top) {
vdev_rebuild_stat_t vrs;
if (vdev_rebuild_get_stats(vd, &vrs) == 0) {
fnvlist_add_uint64_array(nvl,
ZPOOL_CONFIG_REBUILD_STATS, (uint64_t *)&vrs,
sizeof (vrs) / sizeof (uint64_t));
}
}
}
/*
* Generate the nvlist representing this vdev's config.
*/
nvlist_t *
vdev_config_generate(spa_t *spa, vdev_t *vd, boolean_t getstats,
vdev_config_flag_t flags)
{
nvlist_t *nv = NULL;
vdev_indirect_config_t *vic = &vd->vdev_indirect_config;
nv = fnvlist_alloc();
fnvlist_add_string(nv, ZPOOL_CONFIG_TYPE, vd->vdev_ops->vdev_op_type);
if (!(flags & (VDEV_CONFIG_SPARE | VDEV_CONFIG_L2CACHE)))
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ID, vd->vdev_id);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_GUID, vd->vdev_guid);
if (vd->vdev_path != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_PATH, vd->vdev_path);
if (vd->vdev_devid != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_DEVID, vd->vdev_devid);
if (vd->vdev_physpath != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_PHYS_PATH,
vd->vdev_physpath);
if (vd->vdev_enc_sysfs_path != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
vd->vdev_enc_sysfs_path);
if (vd->vdev_fru != NULL)
fnvlist_add_string(nv, ZPOOL_CONFIG_FRU, vd->vdev_fru);
if (vd->vdev_ops->vdev_op_config_generate != NULL)
vd->vdev_ops->vdev_op_config_generate(vd, nv);
if (vd->vdev_wholedisk != -1ULL) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK,
vd->vdev_wholedisk);
}
if (vd->vdev_not_present && !(flags & VDEV_CONFIG_MISSING))
fnvlist_add_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, 1);
if (vd->vdev_isspare)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_SPARE, 1);
if (!(flags & (VDEV_CONFIG_SPARE | VDEV_CONFIG_L2CACHE)) &&
vd == vd->vdev_top) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_METASLAB_ARRAY,
vd->vdev_ms_array);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_METASLAB_SHIFT,
vd->vdev_ms_shift);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ASHIFT, vd->vdev_ashift);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ASIZE,
vd->vdev_asize);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_LOG, vd->vdev_islog);
if (vd->vdev_noalloc) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_NONALLOCATING,
vd->vdev_noalloc);
}
/*
* Slog devices are removed synchronously so don't
* persist the vdev_removing flag to the label.
*/
if (vd->vdev_removing && !vd->vdev_islog) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_REMOVING,
vd->vdev_removing);
}
/* zpool command expects alloc class data */
if (getstats && vd->vdev_alloc_bias != VDEV_BIAS_NONE) {
const char *bias = NULL;
switch (vd->vdev_alloc_bias) {
case VDEV_BIAS_LOG:
bias = VDEV_ALLOC_BIAS_LOG;
break;
case VDEV_BIAS_SPECIAL:
bias = VDEV_ALLOC_BIAS_SPECIAL;
break;
case VDEV_BIAS_DEDUP:
bias = VDEV_ALLOC_BIAS_DEDUP;
break;
default:
ASSERT3U(vd->vdev_alloc_bias, ==,
VDEV_BIAS_NONE);
}
fnvlist_add_string(nv, ZPOOL_CONFIG_ALLOCATION_BIAS,
bias);
}
}
if (vd->vdev_dtl_sm != NULL) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_DTL,
space_map_object(vd->vdev_dtl_sm));
}
if (vic->vic_mapping_object != 0) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_INDIRECT_OBJECT,
vic->vic_mapping_object);
}
if (vic->vic_births_object != 0) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_INDIRECT_BIRTHS,
vic->vic_births_object);
}
if (vic->vic_prev_indirect_vdev != UINT64_MAX) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_PREV_INDIRECT_VDEV,
vic->vic_prev_indirect_vdev);
}
if (vd->vdev_crtxg)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_CREATE_TXG, vd->vdev_crtxg);
if (vd->vdev_expansion_time)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_EXPANSION_TIME,
vd->vdev_expansion_time);
if (flags & VDEV_CONFIG_MOS) {
if (vd->vdev_leaf_zap != 0) {
ASSERT(vd->vdev_ops->vdev_op_leaf);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_VDEV_LEAF_ZAP,
vd->vdev_leaf_zap);
}
if (vd->vdev_top_zap != 0) {
ASSERT(vd == vd->vdev_top);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_VDEV_TOP_ZAP,
vd->vdev_top_zap);
}
+ if (vd->vdev_ops == &vdev_root_ops && vd->vdev_root_zap != 0 &&
+ spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_AVZ_V2)) {
+ fnvlist_add_uint64(nv, ZPOOL_CONFIG_VDEV_ROOT_ZAP,
+ vd->vdev_root_zap);
+ }
+
if (vd->vdev_resilver_deferred) {
ASSERT(vd->vdev_ops->vdev_op_leaf);
ASSERT(spa->spa_resilver_deferred);
fnvlist_add_boolean(nv, ZPOOL_CONFIG_RESILVER_DEFER);
}
}
if (getstats) {
vdev_config_generate_stats(vd, nv);
root_vdev_actions_getprogress(vd, nv);
top_vdev_actions_getprogress(vd, nv);
/*
* Note: this can be called from open context
* (spa_get_stats()), so we need the rwlock to prevent
* the mapping from being changed by condensing.
*/
rw_enter(&vd->vdev_indirect_rwlock, RW_READER);
if (vd->vdev_indirect_mapping != NULL) {
ASSERT(vd->vdev_indirect_births != NULL);
vdev_indirect_mapping_t *vim =
vd->vdev_indirect_mapping;
fnvlist_add_uint64(nv, ZPOOL_CONFIG_INDIRECT_SIZE,
vdev_indirect_mapping_size(vim));
}
rw_exit(&vd->vdev_indirect_rwlock);
if (vd->vdev_mg != NULL &&
vd->vdev_mg->mg_fragmentation != ZFS_FRAG_INVALID) {
/*
* Compute approximately how much memory would be used
* for the indirect mapping if this device were to
* be removed.
*
* Note: If the frag metric is invalid, then not
* enough metaslabs have been converted to have
* histograms.
*/
uint64_t seg_count = 0;
uint64_t to_alloc = vd->vdev_stat.vs_alloc;
/*
* There are the same number of allocated segments
* as free segments, so we will have at least one
* entry per free segment. However, small free
* segments (smaller than vdev_removal_max_span)
* will be combined with adjacent allocated segments
* as a single mapping.
*/
for (int i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) {
if (i + 1 < highbit64(vdev_removal_max_span)
- 1) {
to_alloc +=
vd->vdev_mg->mg_histogram[i] <<
(i + 1);
} else {
seg_count +=
vd->vdev_mg->mg_histogram[i];
}
}
/*
* The maximum length of a mapping is
* zfs_remove_max_segment, so we need at least one entry
* per zfs_remove_max_segment of allocated data.
*/
seg_count += to_alloc / spa_remove_max_segment(spa);
fnvlist_add_uint64(nv, ZPOOL_CONFIG_INDIRECT_SIZE,
seg_count *
sizeof (vdev_indirect_mapping_entry_phys_t));
}
}
if (!vd->vdev_ops->vdev_op_leaf) {
nvlist_t **child;
uint64_t c;
ASSERT(!vd->vdev_ishole);
child = kmem_alloc(vd->vdev_children * sizeof (nvlist_t *),
KM_SLEEP);
for (c = 0; c < vd->vdev_children; c++) {
child[c] = vdev_config_generate(spa, vd->vdev_child[c],
getstats, flags);
}
fnvlist_add_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
(const nvlist_t * const *)child, vd->vdev_children);
for (c = 0; c < vd->vdev_children; c++)
nvlist_free(child[c]);
kmem_free(child, vd->vdev_children * sizeof (nvlist_t *));
} else {
const char *aux = NULL;
if (vd->vdev_offline && !vd->vdev_tmpoffline)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_OFFLINE, B_TRUE);
if (vd->vdev_resilver_txg != 0)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_RESILVER_TXG,
vd->vdev_resilver_txg);
if (vd->vdev_rebuild_txg != 0)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_REBUILD_TXG,
vd->vdev_rebuild_txg);
if (vd->vdev_faulted)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_FAULTED, B_TRUE);
if (vd->vdev_degraded)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_DEGRADED, B_TRUE);
if (vd->vdev_removed)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_REMOVED, B_TRUE);
if (vd->vdev_unspare)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_UNSPARE, B_TRUE);
if (vd->vdev_ishole)
fnvlist_add_uint64(nv, ZPOOL_CONFIG_IS_HOLE, B_TRUE);
/* Set the reason why we're FAULTED/DEGRADED. */
switch (vd->vdev_stat.vs_aux) {
case VDEV_AUX_ERR_EXCEEDED:
aux = "err_exceeded";
break;
case VDEV_AUX_EXTERNAL:
aux = "external";
break;
}
if (aux != NULL && !vd->vdev_tmpoffline) {
fnvlist_add_string(nv, ZPOOL_CONFIG_AUX_STATE, aux);
} else {
/*
* We're healthy - clear any previous AUX_STATE values.
*/
if (nvlist_exists(nv, ZPOOL_CONFIG_AUX_STATE))
nvlist_remove_all(nv, ZPOOL_CONFIG_AUX_STATE);
}
if (vd->vdev_splitting && vd->vdev_orig_guid != 0LL) {
fnvlist_add_uint64(nv, ZPOOL_CONFIG_ORIG_GUID,
vd->vdev_orig_guid);
}
}
return (nv);
}
/*
* Generate a view of the top-level vdevs. If we currently have holes
* in the namespace, then generate an array which contains a list of holey
* vdevs. Additionally, add the number of top-level children that currently
* exist.
*/
void
vdev_top_config_generate(spa_t *spa, nvlist_t *config)
{
vdev_t *rvd = spa->spa_root_vdev;
uint64_t *array;
uint_t c, idx;
array = kmem_alloc(rvd->vdev_children * sizeof (uint64_t), KM_SLEEP);
for (c = 0, idx = 0; c < rvd->vdev_children; c++) {
vdev_t *tvd = rvd->vdev_child[c];
if (tvd->vdev_ishole) {
array[idx++] = c;
}
}
if (idx) {
VERIFY(nvlist_add_uint64_array(config, ZPOOL_CONFIG_HOLE_ARRAY,
array, idx) == 0);
}
VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_VDEV_CHILDREN,
rvd->vdev_children) == 0);
kmem_free(array, rvd->vdev_children * sizeof (uint64_t));
}
/*
* Returns the configuration from the label of the given vdev. For vdevs
* which don't have a txg value stored on their label (i.e. spares/cache)
* or have not been completely initialized (txg = 0) just return
* the configuration from the first valid label we find. Otherwise,
* find the most up-to-date label that does not exceed the specified
* 'txg' value.
*/
nvlist_t *
vdev_label_read_config(vdev_t *vd, uint64_t txg)
{
spa_t *spa = vd->vdev_spa;
nvlist_t *config = NULL;
vdev_phys_t *vp[VDEV_LABELS];
abd_t *vp_abd[VDEV_LABELS];
zio_t *zio[VDEV_LABELS];
uint64_t best_txg = 0;
uint64_t label_txg = 0;
int error = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE;
ASSERT(vd->vdev_validate_thread == curthread ||
spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
if (!vdev_readable(vd))
return (NULL);
/*
* The label for a dRAID distributed spare is not stored on disk.
* Instead it is generated when needed which allows us to bypass
* the pipeline when reading the config from the label.
*/
if (vd->vdev_ops == &vdev_draid_spare_ops)
return (vdev_draid_read_config_spare(vd));
for (int l = 0; l < VDEV_LABELS; l++) {
vp_abd[l] = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
vp[l] = abd_to_buf(vp_abd[l]);
}
retry:
for (int l = 0; l < VDEV_LABELS; l++) {
zio[l] = zio_root(spa, NULL, NULL, flags);
vdev_label_read(zio[l], vd, l, vp_abd[l],
offsetof(vdev_label_t, vl_vdev_phys), sizeof (vdev_phys_t),
NULL, NULL, flags);
}
for (int l = 0; l < VDEV_LABELS; l++) {
nvlist_t *label = NULL;
if (zio_wait(zio[l]) == 0 &&
nvlist_unpack(vp[l]->vp_nvlist, sizeof (vp[l]->vp_nvlist),
&label, 0) == 0) {
/*
* Auxiliary vdevs won't have txg values in their
* labels and newly added vdevs may not have been
* completely initialized so just return the
* configuration from the first valid label we
* encounter.
*/
error = nvlist_lookup_uint64(label,
ZPOOL_CONFIG_POOL_TXG, &label_txg);
if ((error || label_txg == 0) && !config) {
config = label;
for (l++; l < VDEV_LABELS; l++)
zio_wait(zio[l]);
break;
} else if (label_txg <= txg && label_txg > best_txg) {
best_txg = label_txg;
nvlist_free(config);
config = fnvlist_dup(label);
}
}
if (label != NULL) {
nvlist_free(label);
label = NULL;
}
}
if (config == NULL && !(flags & ZIO_FLAG_TRYHARD)) {
flags |= ZIO_FLAG_TRYHARD;
goto retry;
}
/*
* We found a valid label but it didn't pass txg restrictions.
*/
if (config == NULL && label_txg != 0) {
vdev_dbgmsg(vd, "label discarded as txg is too large "
"(%llu > %llu)", (u_longlong_t)label_txg,
(u_longlong_t)txg);
}
for (int l = 0; l < VDEV_LABELS; l++) {
abd_free(vp_abd[l]);
}
return (config);
}
/*
* Determine if a device is in use. The 'spare_guid' parameter will be filled
* in with the device guid if this spare is active elsewhere on the system.
*/
static boolean_t
vdev_inuse(vdev_t *vd, uint64_t crtxg, vdev_labeltype_t reason,
uint64_t *spare_guid, uint64_t *l2cache_guid)
{
spa_t *spa = vd->vdev_spa;
uint64_t state, pool_guid, device_guid, txg, spare_pool;
uint64_t vdtxg = 0;
nvlist_t *label;
if (spare_guid)
*spare_guid = 0ULL;
if (l2cache_guid)
*l2cache_guid = 0ULL;
/*
* Read the label, if any, and perform some basic sanity checks.
*/
if ((label = vdev_label_read_config(vd, -1ULL)) == NULL)
return (B_FALSE);
(void) nvlist_lookup_uint64(label, ZPOOL_CONFIG_CREATE_TXG,
&vdtxg);
if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE,
&state) != 0 ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID,
&device_guid) != 0) {
nvlist_free(label);
return (B_FALSE);
}
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
(nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID,
&pool_guid) != 0 ||
nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_TXG,
&txg) != 0)) {
nvlist_free(label);
return (B_FALSE);
}
nvlist_free(label);
/*
* Check to see if this device indeed belongs to the pool it claims to
* be a part of. The only way this is allowed is if the device is a hot
* spare (which we check for later on).
*/
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
!spa_guid_exists(pool_guid, device_guid) &&
!spa_spare_exists(device_guid, NULL, NULL) &&
!spa_l2cache_exists(device_guid, NULL))
return (B_FALSE);
/*
* If the transaction group is zero, then this an initialized (but
* unused) label. This is only an error if the create transaction
* on-disk is the same as the one we're using now, in which case the
* user has attempted to add the same vdev multiple times in the same
* transaction.
*/
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
txg == 0 && vdtxg == crtxg)
return (B_TRUE);
/*
* Check to see if this is a spare device. We do an explicit check for
* spa_has_spare() here because it may be on our pending list of spares
* to add.
*/
if (spa_spare_exists(device_guid, &spare_pool, NULL) ||
spa_has_spare(spa, device_guid)) {
if (spare_guid)
*spare_guid = device_guid;
switch (reason) {
case VDEV_LABEL_CREATE:
return (B_TRUE);
case VDEV_LABEL_REPLACE:
return (!spa_has_spare(spa, device_guid) ||
spare_pool != 0ULL);
case VDEV_LABEL_SPARE:
return (spa_has_spare(spa, device_guid));
default:
break;
}
}
/*
* Check to see if this is an l2cache device.
*/
if (spa_l2cache_exists(device_guid, NULL) ||
spa_has_l2cache(spa, device_guid)) {
if (l2cache_guid)
*l2cache_guid = device_guid;
switch (reason) {
case VDEV_LABEL_CREATE:
return (B_TRUE);
case VDEV_LABEL_REPLACE:
return (!spa_has_l2cache(spa, device_guid));
case VDEV_LABEL_L2CACHE:
return (spa_has_l2cache(spa, device_guid));
default:
break;
}
}
/*
* We can't rely on a pool's state if it's been imported
* read-only. Instead we look to see if the pools is marked
* read-only in the namespace and set the state to active.
*/
if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
(spa = spa_by_guid(pool_guid, device_guid)) != NULL &&
spa_mode(spa) == SPA_MODE_READ)
state = POOL_STATE_ACTIVE;
/*
* If the device is marked ACTIVE, then this device is in use by another
* pool on the system.
*/
return (state == POOL_STATE_ACTIVE);
}
/*
* Initialize a vdev label. We check to make sure each leaf device is not in
* use, and writable. We put down an initial label which we will later
* overwrite with a complete label. Note that it's important to do this
* sequentially, not in parallel, so that we catch cases of multiple use of the
* same leaf vdev in the vdev we're creating -- e.g. mirroring a disk with
* itself.
*/
int
vdev_label_init(vdev_t *vd, uint64_t crtxg, vdev_labeltype_t reason)
{
spa_t *spa = vd->vdev_spa;
nvlist_t *label;
vdev_phys_t *vp;
abd_t *vp_abd;
abd_t *bootenv;
uberblock_t *ub;
abd_t *ub_abd;
zio_t *zio;
char *buf;
size_t buflen;
int error;
uint64_t spare_guid = 0, l2cache_guid = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
for (int c = 0; c < vd->vdev_children; c++)
if ((error = vdev_label_init(vd->vdev_child[c],
crtxg, reason)) != 0)
return (error);
/* Track the creation time for this vdev */
vd->vdev_crtxg = crtxg;
if (!vd->vdev_ops->vdev_op_leaf || !spa_writeable(spa))
return (0);
/*
* Dead vdevs cannot be initialized.
*/
if (vdev_is_dead(vd))
return (SET_ERROR(EIO));
/*
* Determine if the vdev is in use.
*/
if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_SPLIT &&
vdev_inuse(vd, crtxg, reason, &spare_guid, &l2cache_guid))
return (SET_ERROR(EBUSY));
/*
* If this is a request to add or replace a spare or l2cache device
* that is in use elsewhere on the system, then we must update the
* guid (which was initialized to a random value) to reflect the
* actual GUID (which is shared between multiple pools).
*/
if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_L2CACHE &&
spare_guid != 0ULL) {
uint64_t guid_delta = spare_guid - vd->vdev_guid;
vd->vdev_guid += guid_delta;
for (vdev_t *pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += guid_delta;
/*
* If this is a replacement, then we want to fallthrough to the
* rest of the code. If we're adding a spare, then it's already
* labeled appropriately and we can just return.
*/
if (reason == VDEV_LABEL_SPARE)
return (0);
ASSERT(reason == VDEV_LABEL_REPLACE ||
reason == VDEV_LABEL_SPLIT);
}
if (reason != VDEV_LABEL_REMOVE && reason != VDEV_LABEL_SPARE &&
l2cache_guid != 0ULL) {
uint64_t guid_delta = l2cache_guid - vd->vdev_guid;
vd->vdev_guid += guid_delta;
for (vdev_t *pvd = vd; pvd != NULL; pvd = pvd->vdev_parent)
pvd->vdev_guid_sum += guid_delta;
/*
* If this is a replacement, then we want to fallthrough to the
* rest of the code. If we're adding an l2cache, then it's
* already labeled appropriately and we can just return.
*/
if (reason == VDEV_LABEL_L2CACHE)
return (0);
ASSERT(reason == VDEV_LABEL_REPLACE);
}
/*
* Initialize its label.
*/
vp_abd = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
abd_zero(vp_abd, sizeof (vdev_phys_t));
vp = abd_to_buf(vp_abd);
/*
* Generate a label describing the pool and our top-level vdev.
* We mark it as being from txg 0 to indicate that it's not
* really part of an active pool just yet. The labels will
* be written again with a meaningful txg by spa_sync().
*/
if (reason == VDEV_LABEL_SPARE ||
(reason == VDEV_LABEL_REMOVE && vd->vdev_isspare)) {
/*
* For inactive hot spares, we generate a special label that
* identifies as a mutually shared hot spare. We write the
* label if we are adding a hot spare, or if we are removing an
* active hot spare (in which case we want to revert the
* labels).
*/
VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_SLEEP) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_VERSION,
spa_version(spa)) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_POOL_STATE,
POOL_STATE_SPARE) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_GUID,
vd->vdev_guid) == 0);
} else if (reason == VDEV_LABEL_L2CACHE ||
(reason == VDEV_LABEL_REMOVE && vd->vdev_isl2cache)) {
/*
* For level 2 ARC devices, add a special label.
*/
VERIFY(nvlist_alloc(&label, NV_UNIQUE_NAME, KM_SLEEP) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_VERSION,
spa_version(spa)) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_POOL_STATE,
POOL_STATE_L2CACHE) == 0);
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_GUID,
vd->vdev_guid) == 0);
} else {
uint64_t txg = 0ULL;
if (reason == VDEV_LABEL_SPLIT)
txg = spa->spa_uberblock.ub_txg;
label = spa_config_generate(spa, vd, txg, B_FALSE);
/*
* Add our creation time. This allows us to detect multiple
* vdev uses as described above, and automatically expires if we
* fail.
*/
VERIFY(nvlist_add_uint64(label, ZPOOL_CONFIG_CREATE_TXG,
crtxg) == 0);
}
buf = vp->vp_nvlist;
buflen = sizeof (vp->vp_nvlist);
error = nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_SLEEP);
if (error != 0) {
nvlist_free(label);
abd_free(vp_abd);
/* EFAULT means nvlist_pack ran out of room */
return (SET_ERROR(error == EFAULT ? ENAMETOOLONG : EINVAL));
}
/*
* Initialize uberblock template.
*/
ub_abd = abd_alloc_linear(VDEV_UBERBLOCK_RING, B_TRUE);
abd_zero(ub_abd, VDEV_UBERBLOCK_RING);
abd_copy_from_buf(ub_abd, &spa->spa_uberblock, sizeof (uberblock_t));
ub = abd_to_buf(ub_abd);
ub->ub_txg = 0;
/* Initialize the 2nd padding area. */
bootenv = abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE);
abd_zero(bootenv, VDEV_PAD_SIZE);
/*
* Write everything in parallel.
*/
retry:
zio = zio_root(spa, NULL, NULL, flags);
for (int l = 0; l < VDEV_LABELS; l++) {
vdev_label_write(zio, vd, l, vp_abd,
offsetof(vdev_label_t, vl_vdev_phys),
sizeof (vdev_phys_t), NULL, NULL, flags);
/*
* Skip the 1st padding area.
* Zero out the 2nd padding area where it might have
* left over data from previous filesystem format.
*/
vdev_label_write(zio, vd, l, bootenv,
offsetof(vdev_label_t, vl_be),
VDEV_PAD_SIZE, NULL, NULL, flags);
vdev_label_write(zio, vd, l, ub_abd,
offsetof(vdev_label_t, vl_uberblock),
VDEV_UBERBLOCK_RING, NULL, NULL, flags);
}
error = zio_wait(zio);
if (error != 0 && !(flags & ZIO_FLAG_TRYHARD)) {
flags |= ZIO_FLAG_TRYHARD;
goto retry;
}
nvlist_free(label);
abd_free(bootenv);
abd_free(ub_abd);
abd_free(vp_abd);
/*
* If this vdev hasn't been previously identified as a spare, then we
* mark it as such only if a) we are labeling it as a spare, or b) it
* exists as a spare elsewhere in the system. Do the same for
* level 2 ARC devices.
*/
if (error == 0 && !vd->vdev_isspare &&
(reason == VDEV_LABEL_SPARE ||
spa_spare_exists(vd->vdev_guid, NULL, NULL)))
spa_spare_add(vd);
if (error == 0 && !vd->vdev_isl2cache &&
(reason == VDEV_LABEL_L2CACHE ||
spa_l2cache_exists(vd->vdev_guid, NULL)))
spa_l2cache_add(vd);
return (error);
}
/*
* Done callback for vdev_label_read_bootenv_impl. If this is the first
* callback to finish, store our abd in the callback pointer. Otherwise, we
* just free our abd and return.
*/
static void
vdev_label_read_bootenv_done(zio_t *zio)
{
zio_t *rio = zio->io_private;
abd_t **cbp = rio->io_private;
ASSERT3U(zio->io_size, ==, VDEV_PAD_SIZE);
if (zio->io_error == 0) {
mutex_enter(&rio->io_lock);
if (*cbp == NULL) {
/* Will free this buffer in vdev_label_read_bootenv. */
*cbp = zio->io_abd;
} else {
abd_free(zio->io_abd);
}
mutex_exit(&rio->io_lock);
} else {
abd_free(zio->io_abd);
}
}
static void
vdev_label_read_bootenv_impl(zio_t *zio, vdev_t *vd, int flags)
{
for (int c = 0; c < vd->vdev_children; c++)
vdev_label_read_bootenv_impl(zio, vd->vdev_child[c], flags);
/*
* We just use the first label that has a correct checksum; the
* bootloader should have rewritten them all to be the same on boot,
* and any changes we made since boot have been the same across all
* labels.
*/
if (vd->vdev_ops->vdev_op_leaf && vdev_readable(vd)) {
for (int l = 0; l < VDEV_LABELS; l++) {
vdev_label_read(zio, vd, l,
abd_alloc_linear(VDEV_PAD_SIZE, B_FALSE),
offsetof(vdev_label_t, vl_be), VDEV_PAD_SIZE,
vdev_label_read_bootenv_done, zio, flags);
}
}
}
int
vdev_label_read_bootenv(vdev_t *rvd, nvlist_t *bootenv)
{
nvlist_t *config;
spa_t *spa = rvd->vdev_spa;
abd_t *abd = NULL;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE | ZIO_FLAG_TRYHARD;
ASSERT(bootenv);
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
zio_t *zio = zio_root(spa, NULL, &abd, flags);
vdev_label_read_bootenv_impl(zio, rvd, flags);
int err = zio_wait(zio);
if (abd != NULL) {
char *buf;
vdev_boot_envblock_t *vbe = abd_to_buf(abd);
vbe->vbe_version = ntohll(vbe->vbe_version);
switch (vbe->vbe_version) {
case VB_RAW:
/*
* if we have textual data in vbe_bootenv, create nvlist
* with key "envmap".
*/
fnvlist_add_uint64(bootenv, BOOTENV_VERSION, VB_RAW);
vbe->vbe_bootenv[sizeof (vbe->vbe_bootenv) - 1] = '\0';
fnvlist_add_string(bootenv, GRUB_ENVMAP,
vbe->vbe_bootenv);
break;
case VB_NVLIST:
err = nvlist_unpack(vbe->vbe_bootenv,
sizeof (vbe->vbe_bootenv), &config, 0);
if (err == 0) {
fnvlist_merge(bootenv, config);
nvlist_free(config);
break;
}
zfs_fallthrough;
default:
/* Check for FreeBSD zfs bootonce command string */
buf = abd_to_buf(abd);
if (*buf == '\0') {
fnvlist_add_uint64(bootenv, BOOTENV_VERSION,
VB_NVLIST);
break;
}
fnvlist_add_string(bootenv, FREEBSD_BOOTONCE, buf);
}
/*
* abd was allocated in vdev_label_read_bootenv_impl()
*/
abd_free(abd);
/*
* If we managed to read any successfully,
* return success.
*/
return (0);
}
return (err);
}
int
vdev_label_write_bootenv(vdev_t *vd, nvlist_t *env)
{
zio_t *zio;
spa_t *spa = vd->vdev_spa;
vdev_boot_envblock_t *bootenv;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
int error;
size_t nvsize;
char *nvbuf;
const char *tmp;
error = nvlist_size(env, &nvsize, NV_ENCODE_XDR);
if (error != 0)
return (SET_ERROR(error));
if (nvsize >= sizeof (bootenv->vbe_bootenv)) {
return (SET_ERROR(E2BIG));
}
ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL);
error = ENXIO;
for (int c = 0; c < vd->vdev_children; c++) {
int child_err;
child_err = vdev_label_write_bootenv(vd->vdev_child[c], env);
/*
* As long as any of the disks managed to write all of their
* labels successfully, return success.
*/
if (child_err == 0)
error = child_err;
}
if (!vd->vdev_ops->vdev_op_leaf || vdev_is_dead(vd) ||
!vdev_writeable(vd)) {
return (error);
}
ASSERT3U(sizeof (*bootenv), ==, VDEV_PAD_SIZE);
abd_t *abd = abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE);
abd_zero(abd, VDEV_PAD_SIZE);
bootenv = abd_borrow_buf_copy(abd, VDEV_PAD_SIZE);
nvbuf = bootenv->vbe_bootenv;
nvsize = sizeof (bootenv->vbe_bootenv);
bootenv->vbe_version = fnvlist_lookup_uint64(env, BOOTENV_VERSION);
switch (bootenv->vbe_version) {
case VB_RAW:
if (nvlist_lookup_string(env, GRUB_ENVMAP, &tmp) == 0) {
(void) strlcpy(bootenv->vbe_bootenv, tmp, nvsize);
}
error = 0;
break;
case VB_NVLIST:
error = nvlist_pack(env, &nvbuf, &nvsize, NV_ENCODE_XDR,
KM_SLEEP);
break;
default:
error = EINVAL;
break;
}
if (error == 0) {
bootenv->vbe_version = htonll(bootenv->vbe_version);
abd_return_buf_copy(abd, bootenv, VDEV_PAD_SIZE);
} else {
abd_free(abd);
return (SET_ERROR(error));
}
retry:
zio = zio_root(spa, NULL, NULL, flags);
for (int l = 0; l < VDEV_LABELS; l++) {
vdev_label_write(zio, vd, l, abd,
offsetof(vdev_label_t, vl_be),
VDEV_PAD_SIZE, NULL, NULL, flags);
}
error = zio_wait(zio);
if (error != 0 && !(flags & ZIO_FLAG_TRYHARD)) {
flags |= ZIO_FLAG_TRYHARD;
goto retry;
}
abd_free(abd);
return (error);
}
/*
* ==========================================================================
* uberblock load/sync
* ==========================================================================
*/
/*
* Consider the following situation: txg is safely synced to disk. We've
* written the first uberblock for txg + 1, and then we lose power. When we
* come back up, we fail to see the uberblock for txg + 1 because, say,
* it was on a mirrored device and the replica to which we wrote txg + 1
* is now offline. If we then make some changes and sync txg + 1, and then
* the missing replica comes back, then for a few seconds we'll have two
* conflicting uberblocks on disk with the same txg. The solution is simple:
* among uberblocks with equal txg, choose the one with the latest timestamp.
*/
static int
vdev_uberblock_compare(const uberblock_t *ub1, const uberblock_t *ub2)
{
int cmp = TREE_CMP(ub1->ub_txg, ub2->ub_txg);
if (likely(cmp))
return (cmp);
cmp = TREE_CMP(ub1->ub_timestamp, ub2->ub_timestamp);
if (likely(cmp))
return (cmp);
/*
* If MMP_VALID(ub) && MMP_SEQ_VALID(ub) then the host has an MMP-aware
* ZFS, e.g. OpenZFS >= 0.7.
*
* If one ub has MMP and the other does not, they were written by
* different hosts, which matters for MMP. So we treat no MMP/no SEQ as
* a 0 value.
*
* Since timestamp and txg are the same if we get this far, either is
* acceptable for importing the pool.
*/
unsigned int seq1 = 0;
unsigned int seq2 = 0;
if (MMP_VALID(ub1) && MMP_SEQ_VALID(ub1))
seq1 = MMP_SEQ(ub1);
if (MMP_VALID(ub2) && MMP_SEQ_VALID(ub2))
seq2 = MMP_SEQ(ub2);
return (TREE_CMP(seq1, seq2));
}
struct ubl_cbdata {
uberblock_t *ubl_ubbest; /* Best uberblock */
vdev_t *ubl_vd; /* vdev associated with the above */
};
static void
vdev_uberblock_load_done(zio_t *zio)
{
vdev_t *vd = zio->io_vd;
spa_t *spa = zio->io_spa;
zio_t *rio = zio->io_private;
uberblock_t *ub = abd_to_buf(zio->io_abd);
struct ubl_cbdata *cbp = rio->io_private;
ASSERT3U(zio->io_size, ==, VDEV_UBERBLOCK_SIZE(vd));
if (zio->io_error == 0 && uberblock_verify(ub) == 0) {
mutex_enter(&rio->io_lock);
if (ub->ub_txg <= spa->spa_load_max_txg &&
vdev_uberblock_compare(ub, cbp->ubl_ubbest) > 0) {
/*
* Keep track of the vdev in which this uberblock
* was found. We will use this information later
* to obtain the config nvlist associated with
* this uberblock.
*/
*cbp->ubl_ubbest = *ub;
cbp->ubl_vd = vd;
}
mutex_exit(&rio->io_lock);
}
abd_free(zio->io_abd);
}
static void
vdev_uberblock_load_impl(zio_t *zio, vdev_t *vd, int flags,
struct ubl_cbdata *cbp)
{
for (int c = 0; c < vd->vdev_children; c++)
vdev_uberblock_load_impl(zio, vd->vdev_child[c], flags, cbp);
if (vd->vdev_ops->vdev_op_leaf && vdev_readable(vd) &&
vd->vdev_ops != &vdev_draid_spare_ops) {
for (int l = 0; l < VDEV_LABELS; l++) {
for (int n = 0; n < VDEV_UBERBLOCK_COUNT(vd); n++) {
vdev_label_read(zio, vd, l,
abd_alloc_linear(VDEV_UBERBLOCK_SIZE(vd),
B_TRUE), VDEV_UBERBLOCK_OFFSET(vd, n),
VDEV_UBERBLOCK_SIZE(vd),
vdev_uberblock_load_done, zio, flags);
}
}
}
}
/*
* Reads the 'best' uberblock from disk along with its associated
* configuration. First, we read the uberblock array of each label of each
* vdev, keeping track of the uberblock with the highest txg in each array.
* Then, we read the configuration from the same vdev as the best uberblock.
*/
void
vdev_uberblock_load(vdev_t *rvd, uberblock_t *ub, nvlist_t **config)
{
zio_t *zio;
spa_t *spa = rvd->vdev_spa;
struct ubl_cbdata cb;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE | ZIO_FLAG_TRYHARD;
ASSERT(ub);
ASSERT(config);
memset(ub, 0, sizeof (uberblock_t));
*config = NULL;
cb.ubl_ubbest = ub;
cb.ubl_vd = NULL;
spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
zio = zio_root(spa, NULL, &cb, flags);
vdev_uberblock_load_impl(zio, rvd, flags, &cb);
(void) zio_wait(zio);
/*
* It's possible that the best uberblock was discovered on a label
* that has a configuration which was written in a future txg.
* Search all labels on this vdev to find the configuration that
* matches the txg for our uberblock.
*/
if (cb.ubl_vd != NULL) {
vdev_dbgmsg(cb.ubl_vd, "best uberblock found for spa %s. "
"txg %llu", spa->spa_name, (u_longlong_t)ub->ub_txg);
*config = vdev_label_read_config(cb.ubl_vd, ub->ub_txg);
if (*config == NULL && spa->spa_extreme_rewind) {
vdev_dbgmsg(cb.ubl_vd, "failed to read label config. "
"Trying again without txg restrictions.");
*config = vdev_label_read_config(cb.ubl_vd, UINT64_MAX);
}
if (*config == NULL) {
vdev_dbgmsg(cb.ubl_vd, "failed to read label config");
}
}
spa_config_exit(spa, SCL_ALL, FTAG);
}
/*
* For use when a leaf vdev is expanded.
* The location of labels 2 and 3 changed, and at the new location the
* uberblock rings are either empty or contain garbage. The sync will write
* new configs there because the vdev is dirty, but expansion also needs the
* uberblock rings copied. Read them from label 0 which did not move.
*
* Since the point is to populate labels {2,3} with valid uberblocks,
* we zero uberblocks we fail to read or which are not valid.
*/
static void
vdev_copy_uberblocks(vdev_t *vd)
{
abd_t *ub_abd;
zio_t *write_zio;
int locks = (SCL_L2ARC | SCL_ZIO);
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL |
ZIO_FLAG_SPECULATIVE;
ASSERT(spa_config_held(vd->vdev_spa, SCL_STATE, RW_READER) ==
SCL_STATE);
ASSERT(vd->vdev_ops->vdev_op_leaf);
/*
* No uberblocks are stored on distributed spares, they may be
* safely skipped when expanding a leaf vdev.
*/
if (vd->vdev_ops == &vdev_draid_spare_ops)
return;
spa_config_enter(vd->vdev_spa, locks, FTAG, RW_READER);
ub_abd = abd_alloc_linear(VDEV_UBERBLOCK_SIZE(vd), B_TRUE);
write_zio = zio_root(vd->vdev_spa, NULL, NULL, flags);
for (int n = 0; n < VDEV_UBERBLOCK_COUNT(vd); n++) {
const int src_label = 0;
zio_t *zio;
zio = zio_root(vd->vdev_spa, NULL, NULL, flags);
vdev_label_read(zio, vd, src_label, ub_abd,
VDEV_UBERBLOCK_OFFSET(vd, n), VDEV_UBERBLOCK_SIZE(vd),
NULL, NULL, flags);
if (zio_wait(zio) || uberblock_verify(abd_to_buf(ub_abd)))
abd_zero(ub_abd, VDEV_UBERBLOCK_SIZE(vd));
for (int l = 2; l < VDEV_LABELS; l++)
vdev_label_write(write_zio, vd, l, ub_abd,
VDEV_UBERBLOCK_OFFSET(vd, n),
VDEV_UBERBLOCK_SIZE(vd), NULL, NULL,
flags | ZIO_FLAG_DONT_PROPAGATE);
}
(void) zio_wait(write_zio);
spa_config_exit(vd->vdev_spa, locks, FTAG);
abd_free(ub_abd);
}
/*
* On success, increment root zio's count of good writes.
* We only get credit for writes to known-visible vdevs; see spa_vdev_add().
*/
static void
vdev_uberblock_sync_done(zio_t *zio)
{
uint64_t *good_writes = zio->io_private;
if (zio->io_error == 0 && zio->io_vd->vdev_top->vdev_ms_array != 0)
atomic_inc_64(good_writes);
}
/*
* Write the uberblock to all labels of all leaves of the specified vdev.
*/
static void
vdev_uberblock_sync(zio_t *zio, uint64_t *good_writes,
uberblock_t *ub, vdev_t *vd, int flags)
{
for (uint64_t c = 0; c < vd->vdev_children; c++) {
vdev_uberblock_sync(zio, good_writes,
ub, vd->vdev_child[c], flags);
}
if (!vd->vdev_ops->vdev_op_leaf)
return;
if (!vdev_writeable(vd))
return;
/*
* There's no need to write uberblocks to a distributed spare, they
* are already stored on all the leaves of the parent dRAID. For
* this same reason vdev_uberblock_load_impl() skips distributed
* spares when reading uberblocks.
*/
if (vd->vdev_ops == &vdev_draid_spare_ops)
return;
/* If the vdev was expanded, need to copy uberblock rings. */
if (vd->vdev_state == VDEV_STATE_HEALTHY &&
vd->vdev_copy_uberblocks == B_TRUE) {
vdev_copy_uberblocks(vd);
vd->vdev_copy_uberblocks = B_FALSE;
}
int m = spa_multihost(vd->vdev_spa) ? MMP_BLOCKS_PER_LABEL : 0;
int n = ub->ub_txg % (VDEV_UBERBLOCK_COUNT(vd) - m);
/* Copy the uberblock_t into the ABD */
abd_t *ub_abd = abd_alloc_for_io(VDEV_UBERBLOCK_SIZE(vd), B_TRUE);
abd_zero(ub_abd, VDEV_UBERBLOCK_SIZE(vd));
abd_copy_from_buf(ub_abd, ub, sizeof (uberblock_t));
for (int l = 0; l < VDEV_LABELS; l++)
vdev_label_write(zio, vd, l, ub_abd,
VDEV_UBERBLOCK_OFFSET(vd, n), VDEV_UBERBLOCK_SIZE(vd),
vdev_uberblock_sync_done, good_writes,
flags | ZIO_FLAG_DONT_PROPAGATE);
abd_free(ub_abd);
}
/* Sync the uberblocks to all vdevs in svd[] */
static int
vdev_uberblock_sync_list(vdev_t **svd, int svdcount, uberblock_t *ub, int flags)
{
spa_t *spa = svd[0]->vdev_spa;
zio_t *zio;
uint64_t good_writes = 0;
zio = zio_root(spa, NULL, NULL, flags);
for (int v = 0; v < svdcount; v++)
vdev_uberblock_sync(zio, &good_writes, ub, svd[v], flags);
(void) zio_wait(zio);
/*
* Flush the uberblocks to disk. This ensures that the odd labels
* are no longer needed (because the new uberblocks and the even
* labels are safely on disk), so it is safe to overwrite them.
*/
zio = zio_root(spa, NULL, NULL, flags);
for (int v = 0; v < svdcount; v++) {
if (vdev_writeable(svd[v])) {
zio_flush(zio, svd[v]);
}
}
(void) zio_wait(zio);
return (good_writes >= 1 ? 0 : EIO);
}
/*
* On success, increment the count of good writes for our top-level vdev.
*/
static void
vdev_label_sync_done(zio_t *zio)
{
uint64_t *good_writes = zio->io_private;
if (zio->io_error == 0)
atomic_inc_64(good_writes);
}
/*
* If there weren't enough good writes, indicate failure to the parent.
*/
static void
vdev_label_sync_top_done(zio_t *zio)
{
uint64_t *good_writes = zio->io_private;
if (*good_writes == 0)
zio->io_error = SET_ERROR(EIO);
kmem_free(good_writes, sizeof (uint64_t));
}
/*
* We ignore errors for log and cache devices, simply free the private data.
*/
static void
vdev_label_sync_ignore_done(zio_t *zio)
{
kmem_free(zio->io_private, sizeof (uint64_t));
}
/*
* Write all even or odd labels to all leaves of the specified vdev.
*/
static void
vdev_label_sync(zio_t *zio, uint64_t *good_writes,
vdev_t *vd, int l, uint64_t txg, int flags)
{
nvlist_t *label;
vdev_phys_t *vp;
abd_t *vp_abd;
char *buf;
size_t buflen;
for (int c = 0; c < vd->vdev_children; c++) {
vdev_label_sync(zio, good_writes,
vd->vdev_child[c], l, txg, flags);
}
if (!vd->vdev_ops->vdev_op_leaf)
return;
if (!vdev_writeable(vd))
return;
/*
* The top-level config never needs to be written to a distributed
* spare. When read vdev_dspare_label_read_config() will generate
* the config for the vdev_label_read_config().
*/
if (vd->vdev_ops == &vdev_draid_spare_ops)
return;
/*
* Generate a label describing the top-level config to which we belong.
*/
label = spa_config_generate(vd->vdev_spa, vd, txg, B_FALSE);
vp_abd = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE);
abd_zero(vp_abd, sizeof (vdev_phys_t));
vp = abd_to_buf(vp_abd);
buf = vp->vp_nvlist;
buflen = sizeof (vp->vp_nvlist);
if (!nvlist_pack(label, &buf, &buflen, NV_ENCODE_XDR, KM_SLEEP)) {
for (; l < VDEV_LABELS; l += 2) {
vdev_label_write(zio, vd, l, vp_abd,
offsetof(vdev_label_t, vl_vdev_phys),
sizeof (vdev_phys_t),
vdev_label_sync_done, good_writes,
flags | ZIO_FLAG_DONT_PROPAGATE);
}
}
abd_free(vp_abd);
nvlist_free(label);
}
static int
vdev_label_sync_list(spa_t *spa, int l, uint64_t txg, int flags)
{
list_t *dl = &spa->spa_config_dirty_list;
vdev_t *vd;
zio_t *zio;
int error;
/*
* Write the new labels to disk.
*/
zio = zio_root(spa, NULL, NULL, flags);
for (vd = list_head(dl); vd != NULL; vd = list_next(dl, vd)) {
uint64_t *good_writes;
ASSERT(!vd->vdev_ishole);
good_writes = kmem_zalloc(sizeof (uint64_t), KM_SLEEP);
zio_t *vio = zio_null(zio, spa, NULL,
(vd->vdev_islog || vd->vdev_aux != NULL) ?
vdev_label_sync_ignore_done : vdev_label_sync_top_done,
good_writes, flags);
vdev_label_sync(vio, good_writes, vd, l, txg, flags);
zio_nowait(vio);
}
error = zio_wait(zio);
/*
* Flush the new labels to disk.
*/
zio = zio_root(spa, NULL, NULL, flags);
for (vd = list_head(dl); vd != NULL; vd = list_next(dl, vd))
zio_flush(zio, vd);
(void) zio_wait(zio);
return (error);
}
/*
* Sync the uberblock and any changes to the vdev configuration.
*
* The order of operations is carefully crafted to ensure that
* if the system panics or loses power at any time, the state on disk
* is still transactionally consistent. The in-line comments below
* describe the failure semantics at each stage.
*
* Moreover, vdev_config_sync() is designed to be idempotent: if it fails
* at any time, you can just call it again, and it will resume its work.
*/
int
vdev_config_sync(vdev_t **svd, int svdcount, uint64_t txg)
{
spa_t *spa = svd[0]->vdev_spa;
uberblock_t *ub = &spa->spa_uberblock;
int error = 0;
int flags = ZIO_FLAG_CONFIG_WRITER | ZIO_FLAG_CANFAIL;
ASSERT(svdcount != 0);
retry:
/*
* Normally, we don't want to try too hard to write every label and
* uberblock. If there is a flaky disk, we don't want the rest of the
* sync process to block while we retry. But if we can't write a
* single label out, we should retry with ZIO_FLAG_TRYHARD before
* bailing out and declaring the pool faulted.
*/
if (error != 0) {
if ((flags & ZIO_FLAG_TRYHARD) != 0)
return (error);
flags |= ZIO_FLAG_TRYHARD;
}
ASSERT(ub->ub_txg <= txg);
/*
* If this isn't a resync due to I/O errors,
* and nothing changed in this transaction group,
* and the vdev configuration hasn't changed,
* then there's nothing to do.
*/
if (ub->ub_txg < txg) {
boolean_t changed = uberblock_update(ub, spa->spa_root_vdev,
txg, spa->spa_mmp.mmp_delay);
if (!changed && list_is_empty(&spa->spa_config_dirty_list))
return (0);
}
if (txg > spa_freeze_txg(spa))
return (0);
ASSERT(txg <= spa->spa_final_txg);
/*
* Flush the write cache of every disk that's been written to
* in this transaction group. This ensures that all blocks
* written in this txg will be committed to stable storage
* before any uberblock that references them.
*/
zio_t *zio = zio_root(spa, NULL, NULL, flags);
for (vdev_t *vd =
txg_list_head(&spa->spa_vdev_txg_list, TXG_CLEAN(txg)); vd != NULL;
vd = txg_list_next(&spa->spa_vdev_txg_list, vd, TXG_CLEAN(txg)))
zio_flush(zio, vd);
(void) zio_wait(zio);
/*
* Sync out the even labels (L0, L2) for every dirty vdev. If the
* system dies in the middle of this process, that's OK: all of the
* even labels that made it to disk will be newer than any uberblock,
* and will therefore be considered invalid. The odd labels (L1, L3),
* which have not yet been touched, will still be valid. We flush
* the new labels to disk to ensure that all even-label updates
* are committed to stable storage before the uberblock update.
*/
if ((error = vdev_label_sync_list(spa, 0, txg, flags)) != 0) {
if ((flags & ZIO_FLAG_TRYHARD) != 0) {
zfs_dbgmsg("vdev_label_sync_list() returned error %d "
"for pool '%s' when syncing out the even labels "
"of dirty vdevs", error, spa_name(spa));
}
goto retry;
}
/*
* Sync the uberblocks to all vdevs in svd[].
* If the system dies in the middle of this step, there are two cases
* to consider, and the on-disk state is consistent either way:
*
* (1) If none of the new uberblocks made it to disk, then the
* previous uberblock will be the newest, and the odd labels
* (which had not yet been touched) will be valid with respect
* to that uberblock.
*
* (2) If one or more new uberblocks made it to disk, then they
* will be the newest, and the even labels (which had all
* been successfully committed) will be valid with respect
* to the new uberblocks.
*/
if ((error = vdev_uberblock_sync_list(svd, svdcount, ub, flags)) != 0) {
if ((flags & ZIO_FLAG_TRYHARD) != 0) {
zfs_dbgmsg("vdev_uberblock_sync_list() returned error "
"%d for pool '%s'", error, spa_name(spa));
}
goto retry;
}
if (spa_multihost(spa))
mmp_update_uberblock(spa, ub);
/*
* Sync out odd labels for every dirty vdev. If the system dies
* in the middle of this process, the even labels and the new
* uberblocks will suffice to open the pool. The next time
* the pool is opened, the first thing we'll do -- before any
* user data is modified -- is mark every vdev dirty so that
* all labels will be brought up to date. We flush the new labels
* to disk to ensure that all odd-label updates are committed to
* stable storage before the next transaction group begins.
*/
if ((error = vdev_label_sync_list(spa, 1, txg, flags)) != 0) {
if ((flags & ZIO_FLAG_TRYHARD) != 0) {
zfs_dbgmsg("vdev_label_sync_list() returned error %d "
"for pool '%s' when syncing out the odd labels of "
"dirty vdevs", error, spa_name(spa));
}
goto retry;
}
return (0);
}
diff --git a/sys/contrib/openzfs/module/zfs/vdev_queue.c b/sys/contrib/openzfs/module/zfs/vdev_queue.c
index ec55674393ce..1a75d68abd9e 100644
--- a/sys/contrib/openzfs/module/zfs/vdev_queue.c
+++ b/sys/contrib/openzfs/module/zfs/vdev_queue.c
@@ -1,1121 +1,1124 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2009 Sun Microsystems, Inc. All rights reserved.
* Use is subject to license terms.
*/
/*
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
*/
#include <sys/zfs_context.h>
#include <sys/vdev_impl.h>
#include <sys/spa_impl.h>
#include <sys/zio.h>
#include <sys/avl.h>
#include <sys/dsl_pool.h>
#include <sys/metaslab_impl.h>
#include <sys/spa.h>
#include <sys/abd.h>
/*
* ZFS I/O Scheduler
* ---------------
*
* ZFS issues I/O operations to leaf vdevs to satisfy and complete zios. The
* I/O scheduler determines when and in what order those operations are
* issued. The I/O scheduler divides operations into five I/O classes
* prioritized in the following order: sync read, sync write, async read,
* async write, and scrub/resilver. Each queue defines the minimum and
* maximum number of concurrent operations that may be issued to the device.
* In addition, the device has an aggregate maximum. Note that the sum of the
* per-queue minimums must not exceed the aggregate maximum. If the
* sum of the per-queue maximums exceeds the aggregate maximum, then the
* number of active i/os may reach zfs_vdev_max_active, in which case no
* further i/os will be issued regardless of whether all per-queue
* minimums have been met.
*
* For many physical devices, throughput increases with the number of
* concurrent operations, but latency typically suffers. Further, physical
* devices typically have a limit at which more concurrent operations have no
* effect on throughput or can actually cause it to decrease.
*
* The scheduler selects the next operation to issue by first looking for an
* I/O class whose minimum has not been satisfied. Once all are satisfied and
* the aggregate maximum has not been hit, the scheduler looks for classes
* whose maximum has not been satisfied. Iteration through the I/O classes is
* done in the order specified above. No further operations are issued if the
* aggregate maximum number of concurrent operations has been hit or if there
* are no operations queued for an I/O class that has not hit its maximum.
* Every time an i/o is queued or an operation completes, the I/O scheduler
* looks for new operations to issue.
*
* All I/O classes have a fixed maximum number of outstanding operations
* except for the async write class. Asynchronous writes represent the data
* that is committed to stable storage during the syncing stage for
* transaction groups (see txg.c). Transaction groups enter the syncing state
* periodically so the number of queued async writes will quickly burst up and
* then bleed down to zero. Rather than servicing them as quickly as possible,
* the I/O scheduler changes the maximum number of active async write i/os
* according to the amount of dirty data in the pool (see dsl_pool.c). Since
* both throughput and latency typically increase with the number of
* concurrent operations issued to physical devices, reducing the burstiness
* in the number of concurrent operations also stabilizes the response time of
* operations from other -- and in particular synchronous -- queues. In broad
* strokes, the I/O scheduler will issue more concurrent operations from the
* async write queue as there's more dirty data in the pool.
*
* Async Writes
*
* The number of concurrent operations issued for the async write I/O class
* follows a piece-wise linear function defined by a few adjustable points.
*
* | o---------| <-- zfs_vdev_async_write_max_active
* ^ | /^ |
* | | / | |
* active | / | |
* I/O | / | |
* count | / | |
* | / | |
* |------------o | | <-- zfs_vdev_async_write_min_active
* 0|____________^______|_________|
* 0% | | 100% of zfs_dirty_data_max
* | |
* | `-- zfs_vdev_async_write_active_max_dirty_percent
* `--------- zfs_vdev_async_write_active_min_dirty_percent
*
* Until the amount of dirty data exceeds a minimum percentage of the dirty
* data allowed in the pool, the I/O scheduler will limit the number of
* concurrent operations to the minimum. As that threshold is crossed, the
* number of concurrent operations issued increases linearly to the maximum at
* the specified maximum percentage of the dirty data allowed in the pool.
*
* Ideally, the amount of dirty data on a busy pool will stay in the sloped
* part of the function between zfs_vdev_async_write_active_min_dirty_percent
* and zfs_vdev_async_write_active_max_dirty_percent. If it exceeds the
* maximum percentage, this indicates that the rate of incoming data is
* greater than the rate that the backend storage can handle. In this case, we
* must further throttle incoming writes (see dmu_tx_delay() for details).
*/
/*
* The maximum number of i/os active to each device. Ideally, this will be >=
* the sum of each queue's max_active.
*/
uint_t zfs_vdev_max_active = 1000;
/*
* Per-queue limits on the number of i/os active to each device. If the
* number of active i/os is < zfs_vdev_max_active, then the min_active comes
* into play. We will send min_active from each queue round-robin, and then
* send from queues in the order defined by zio_priority_t up to max_active.
* Some queues have additional mechanisms to limit number of active I/Os in
* addition to min_active and max_active, see below.
*
* In general, smaller max_active's will lead to lower latency of synchronous
* operations. Larger max_active's may lead to higher overall throughput,
* depending on underlying storage.
*
* The ratio of the queues' max_actives determines the balance of performance
* between reads, writes, and scrubs. E.g., increasing
* zfs_vdev_scrub_max_active will cause the scrub or resilver to complete
* more quickly, but reads and writes to have higher latency and lower
* throughput.
*/
static uint_t zfs_vdev_sync_read_min_active = 10;
static uint_t zfs_vdev_sync_read_max_active = 10;
static uint_t zfs_vdev_sync_write_min_active = 10;
static uint_t zfs_vdev_sync_write_max_active = 10;
static uint_t zfs_vdev_async_read_min_active = 1;
/* */ uint_t zfs_vdev_async_read_max_active = 3;
static uint_t zfs_vdev_async_write_min_active = 2;
/* */ uint_t zfs_vdev_async_write_max_active = 10;
static uint_t zfs_vdev_scrub_min_active = 1;
static uint_t zfs_vdev_scrub_max_active = 3;
static uint_t zfs_vdev_removal_min_active = 1;
static uint_t zfs_vdev_removal_max_active = 2;
static uint_t zfs_vdev_initializing_min_active = 1;
static uint_t zfs_vdev_initializing_max_active = 1;
static uint_t zfs_vdev_trim_min_active = 1;
static uint_t zfs_vdev_trim_max_active = 2;
static uint_t zfs_vdev_rebuild_min_active = 1;
static uint_t zfs_vdev_rebuild_max_active = 3;
/*
* When the pool has less than zfs_vdev_async_write_active_min_dirty_percent
* dirty data, use zfs_vdev_async_write_min_active. When it has more than
* zfs_vdev_async_write_active_max_dirty_percent, use
* zfs_vdev_async_write_max_active. The value is linearly interpolated
* between min and max.
*/
uint_t zfs_vdev_async_write_active_min_dirty_percent = 30;
uint_t zfs_vdev_async_write_active_max_dirty_percent = 60;
/*
* For non-interactive I/O (scrub, resilver, removal, initialize and rebuild),
* the number of concurrently-active I/O's is limited to *_min_active, unless
* the vdev is "idle". When there are no interactive I/Os active (sync or
* async), and zfs_vdev_nia_delay I/Os have completed since the last
* interactive I/O, then the vdev is considered to be "idle", and the number
* of concurrently-active non-interactive I/O's is increased to *_max_active.
*/
static uint_t zfs_vdev_nia_delay = 5;
/*
* Some HDDs tend to prioritize sequential I/O so high that concurrent
* random I/O latency reaches several seconds. On some HDDs it happens
* even if sequential I/Os are submitted one at a time, and so setting
* *_max_active to 1 does not help. To prevent non-interactive I/Os, like
* scrub, from monopolizing the device no more than zfs_vdev_nia_credit
* I/Os can be sent while there are outstanding incomplete interactive
* I/Os. This enforced wait ensures the HDD services the interactive I/O
* within a reasonable amount of time.
*/
static uint_t zfs_vdev_nia_credit = 5;
/*
* To reduce IOPs, we aggregate small adjacent I/Os into one large I/O.
* For read I/Os, we also aggregate across small adjacency gaps; for writes
* we include spans of optional I/Os to aid aggregation at the disk even when
* they aren't able to help us aggregate at this level.
*/
static uint_t zfs_vdev_aggregation_limit = 1 << 20;
static uint_t zfs_vdev_aggregation_limit_non_rotating = SPA_OLD_MAXBLOCKSIZE;
static uint_t zfs_vdev_read_gap_limit = 32 << 10;
static uint_t zfs_vdev_write_gap_limit = 4 << 10;
/*
* Define the queue depth percentage for each top-level. This percentage is
* used in conjunction with zfs_vdev_async_max_active to determine how many
* allocations a specific top-level vdev should handle. Once the queue depth
* reaches zfs_vdev_queue_depth_pct * zfs_vdev_async_write_max_active / 100
* then allocator will stop allocating blocks on that top-level device.
* The default kernel setting is 1000% which will yield 100 allocations per
* device. For userland testing, the default setting is 300% which equates
* to 30 allocations per device.
*/
#ifdef _KERNEL
uint_t zfs_vdev_queue_depth_pct = 1000;
#else
uint_t zfs_vdev_queue_depth_pct = 300;
#endif
/*
* When performing allocations for a given metaslab, we want to make sure that
* there are enough IOs to aggregate together to improve throughput. We want to
* ensure that there are at least 128k worth of IOs that can be aggregated, and
* we assume that the average allocation size is 4k, so we need the queue depth
* to be 32 per allocator to get good aggregation of sequential writes.
*/
uint_t zfs_vdev_def_queue_depth = 32;
/*
* Allow TRIM I/Os to be aggregated. This should normally not be needed since
* TRIM I/O for extents up to zfs_trim_extent_bytes_max (128M) can be submitted
* by the TRIM code in zfs_trim.c.
*/
static uint_t zfs_vdev_aggregate_trim = 0;
static int
vdev_queue_offset_compare(const void *x1, const void *x2)
{
const zio_t *z1 = (const zio_t *)x1;
const zio_t *z2 = (const zio_t *)x2;
int cmp = TREE_CMP(z1->io_offset, z2->io_offset);
if (likely(cmp))
return (cmp);
return (TREE_PCMP(z1, z2));
}
static inline avl_tree_t *
vdev_queue_class_tree(vdev_queue_t *vq, zio_priority_t p)
{
return (&vq->vq_class[p].vqc_queued_tree);
}
static inline avl_tree_t *
vdev_queue_type_tree(vdev_queue_t *vq, zio_type_t t)
{
ASSERT(t == ZIO_TYPE_READ || t == ZIO_TYPE_WRITE || t == ZIO_TYPE_TRIM);
if (t == ZIO_TYPE_READ)
return (&vq->vq_read_offset_tree);
else if (t == ZIO_TYPE_WRITE)
return (&vq->vq_write_offset_tree);
else
return (&vq->vq_trim_offset_tree);
}
static int
vdev_queue_timestamp_compare(const void *x1, const void *x2)
{
const zio_t *z1 = (const zio_t *)x1;
const zio_t *z2 = (const zio_t *)x2;
int cmp = TREE_CMP(z1->io_timestamp, z2->io_timestamp);
if (likely(cmp))
return (cmp);
return (TREE_PCMP(z1, z2));
}
static uint_t
vdev_queue_class_min_active(vdev_queue_t *vq, zio_priority_t p)
{
switch (p) {
case ZIO_PRIORITY_SYNC_READ:
return (zfs_vdev_sync_read_min_active);
case ZIO_PRIORITY_SYNC_WRITE:
return (zfs_vdev_sync_write_min_active);
case ZIO_PRIORITY_ASYNC_READ:
return (zfs_vdev_async_read_min_active);
case ZIO_PRIORITY_ASYNC_WRITE:
return (zfs_vdev_async_write_min_active);
case ZIO_PRIORITY_SCRUB:
return (vq->vq_ia_active == 0 ? zfs_vdev_scrub_min_active :
MIN(vq->vq_nia_credit, zfs_vdev_scrub_min_active));
case ZIO_PRIORITY_REMOVAL:
return (vq->vq_ia_active == 0 ? zfs_vdev_removal_min_active :
MIN(vq->vq_nia_credit, zfs_vdev_removal_min_active));
case ZIO_PRIORITY_INITIALIZING:
return (vq->vq_ia_active == 0 ?zfs_vdev_initializing_min_active:
MIN(vq->vq_nia_credit, zfs_vdev_initializing_min_active));
case ZIO_PRIORITY_TRIM:
return (zfs_vdev_trim_min_active);
case ZIO_PRIORITY_REBUILD:
return (vq->vq_ia_active == 0 ? zfs_vdev_rebuild_min_active :
MIN(vq->vq_nia_credit, zfs_vdev_rebuild_min_active));
default:
panic("invalid priority %u", p);
return (0);
}
}
static uint_t
vdev_queue_max_async_writes(spa_t *spa)
{
uint_t writes;
uint64_t dirty = 0;
dsl_pool_t *dp = spa_get_dsl(spa);
uint64_t min_bytes = zfs_dirty_data_max *
zfs_vdev_async_write_active_min_dirty_percent / 100;
uint64_t max_bytes = zfs_dirty_data_max *
zfs_vdev_async_write_active_max_dirty_percent / 100;
/*
* Async writes may occur before the assignment of the spa's
* dsl_pool_t if a self-healing zio is issued prior to the
* completion of dmu_objset_open_impl().
*/
if (dp == NULL)
return (zfs_vdev_async_write_max_active);
/*
* Sync tasks correspond to interactive user actions. To reduce the
* execution time of those actions we push data out as fast as possible.
*/
dirty = dp->dp_dirty_total;
if (dirty > max_bytes || spa_has_pending_synctask(spa))
return (zfs_vdev_async_write_max_active);
if (dirty < min_bytes)
return (zfs_vdev_async_write_min_active);
/*
* linear interpolation:
* slope = (max_writes - min_writes) / (max_bytes - min_bytes)
* move right by min_bytes
* move up by min_writes
*/
writes = (dirty - min_bytes) *
(zfs_vdev_async_write_max_active -
zfs_vdev_async_write_min_active) /
(max_bytes - min_bytes) +
zfs_vdev_async_write_min_active;
ASSERT3U(writes, >=, zfs_vdev_async_write_min_active);
ASSERT3U(writes, <=, zfs_vdev_async_write_max_active);
return (writes);
}
static uint_t
vdev_queue_class_max_active(spa_t *spa, vdev_queue_t *vq, zio_priority_t p)
{
switch (p) {
case ZIO_PRIORITY_SYNC_READ:
return (zfs_vdev_sync_read_max_active);
case ZIO_PRIORITY_SYNC_WRITE:
return (zfs_vdev_sync_write_max_active);
case ZIO_PRIORITY_ASYNC_READ:
return (zfs_vdev_async_read_max_active);
case ZIO_PRIORITY_ASYNC_WRITE:
return (vdev_queue_max_async_writes(spa));
case ZIO_PRIORITY_SCRUB:
if (vq->vq_ia_active > 0) {
return (MIN(vq->vq_nia_credit,
zfs_vdev_scrub_min_active));
} else if (vq->vq_nia_credit < zfs_vdev_nia_delay)
return (MAX(1, zfs_vdev_scrub_min_active));
return (zfs_vdev_scrub_max_active);
case ZIO_PRIORITY_REMOVAL:
if (vq->vq_ia_active > 0) {
return (MIN(vq->vq_nia_credit,
zfs_vdev_removal_min_active));
} else if (vq->vq_nia_credit < zfs_vdev_nia_delay)
return (MAX(1, zfs_vdev_removal_min_active));
return (zfs_vdev_removal_max_active);
case ZIO_PRIORITY_INITIALIZING:
if (vq->vq_ia_active > 0) {
return (MIN(vq->vq_nia_credit,
zfs_vdev_initializing_min_active));
} else if (vq->vq_nia_credit < zfs_vdev_nia_delay)
return (MAX(1, zfs_vdev_initializing_min_active));
return (zfs_vdev_initializing_max_active);
case ZIO_PRIORITY_TRIM:
return (zfs_vdev_trim_max_active);
case ZIO_PRIORITY_REBUILD:
if (vq->vq_ia_active > 0) {
return (MIN(vq->vq_nia_credit,
zfs_vdev_rebuild_min_active));
} else if (vq->vq_nia_credit < zfs_vdev_nia_delay)
return (MAX(1, zfs_vdev_rebuild_min_active));
return (zfs_vdev_rebuild_max_active);
default:
panic("invalid priority %u", p);
return (0);
}
}
/*
* Return the i/o class to issue from, or ZIO_PRIORITY_NUM_QUEUEABLE if
* there is no eligible class.
*/
static zio_priority_t
vdev_queue_class_to_issue(vdev_queue_t *vq)
{
spa_t *spa = vq->vq_vdev->vdev_spa;
zio_priority_t p, n;
if (avl_numnodes(&vq->vq_active_tree) >= zfs_vdev_max_active)
return (ZIO_PRIORITY_NUM_QUEUEABLE);
/*
* Find a queue that has not reached its minimum # outstanding i/os.
* Do round-robin to reduce starvation due to zfs_vdev_max_active
* and vq_nia_credit limits.
*/
for (n = 0; n < ZIO_PRIORITY_NUM_QUEUEABLE; n++) {
p = (vq->vq_last_prio + n + 1) % ZIO_PRIORITY_NUM_QUEUEABLE;
if (avl_numnodes(vdev_queue_class_tree(vq, p)) > 0 &&
vq->vq_class[p].vqc_active <
vdev_queue_class_min_active(vq, p)) {
vq->vq_last_prio = p;
return (p);
}
}
/*
* If we haven't found a queue, look for one that hasn't reached its
* maximum # outstanding i/os.
*/
for (p = 0; p < ZIO_PRIORITY_NUM_QUEUEABLE; p++) {
if (avl_numnodes(vdev_queue_class_tree(vq, p)) > 0 &&
vq->vq_class[p].vqc_active <
vdev_queue_class_max_active(spa, vq, p)) {
vq->vq_last_prio = p;
return (p);
}
}
/* No eligible queued i/os */
return (ZIO_PRIORITY_NUM_QUEUEABLE);
}
void
vdev_queue_init(vdev_t *vd)
{
vdev_queue_t *vq = &vd->vdev_queue;
zio_priority_t p;
mutex_init(&vq->vq_lock, NULL, MUTEX_DEFAULT, NULL);
vq->vq_vdev = vd;
taskq_init_ent(&vd->vdev_queue.vq_io_search.io_tqent);
avl_create(&vq->vq_active_tree, vdev_queue_offset_compare,
sizeof (zio_t), offsetof(struct zio, io_queue_node));
avl_create(vdev_queue_type_tree(vq, ZIO_TYPE_READ),
vdev_queue_offset_compare, sizeof (zio_t),
offsetof(struct zio, io_offset_node));
avl_create(vdev_queue_type_tree(vq, ZIO_TYPE_WRITE),
vdev_queue_offset_compare, sizeof (zio_t),
offsetof(struct zio, io_offset_node));
avl_create(vdev_queue_type_tree(vq, ZIO_TYPE_TRIM),
vdev_queue_offset_compare, sizeof (zio_t),
offsetof(struct zio, io_offset_node));
for (p = 0; p < ZIO_PRIORITY_NUM_QUEUEABLE; p++) {
int (*compfn) (const void *, const void *);
/*
* The synchronous/trim i/o queues are dispatched in FIFO rather
* than LBA order. This provides more consistent latency for
* these i/os.
*/
if (p == ZIO_PRIORITY_SYNC_READ ||
p == ZIO_PRIORITY_SYNC_WRITE ||
p == ZIO_PRIORITY_TRIM) {
compfn = vdev_queue_timestamp_compare;
} else {
compfn = vdev_queue_offset_compare;
}
avl_create(vdev_queue_class_tree(vq, p), compfn,
sizeof (zio_t), offsetof(struct zio, io_queue_node));
}
vq->vq_last_offset = 0;
}
void
vdev_queue_fini(vdev_t *vd)
{
vdev_queue_t *vq = &vd->vdev_queue;
for (zio_priority_t p = 0; p < ZIO_PRIORITY_NUM_QUEUEABLE; p++)
avl_destroy(vdev_queue_class_tree(vq, p));
avl_destroy(&vq->vq_active_tree);
avl_destroy(vdev_queue_type_tree(vq, ZIO_TYPE_READ));
avl_destroy(vdev_queue_type_tree(vq, ZIO_TYPE_WRITE));
avl_destroy(vdev_queue_type_tree(vq, ZIO_TYPE_TRIM));
mutex_destroy(&vq->vq_lock);
}
static void
vdev_queue_io_add(vdev_queue_t *vq, zio_t *zio)
{
ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE);
avl_add(vdev_queue_class_tree(vq, zio->io_priority), zio);
avl_add(vdev_queue_type_tree(vq, zio->io_type), zio);
}
static void
vdev_queue_io_remove(vdev_queue_t *vq, zio_t *zio)
{
ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE);
avl_remove(vdev_queue_class_tree(vq, zio->io_priority), zio);
avl_remove(vdev_queue_type_tree(vq, zio->io_type), zio);
}
static boolean_t
vdev_queue_is_interactive(zio_priority_t p)
{
switch (p) {
case ZIO_PRIORITY_SCRUB:
case ZIO_PRIORITY_REMOVAL:
case ZIO_PRIORITY_INITIALIZING:
case ZIO_PRIORITY_REBUILD:
return (B_FALSE);
default:
return (B_TRUE);
}
}
static void
vdev_queue_pending_add(vdev_queue_t *vq, zio_t *zio)
{
ASSERT(MUTEX_HELD(&vq->vq_lock));
ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE);
vq->vq_class[zio->io_priority].vqc_active++;
if (vdev_queue_is_interactive(zio->io_priority)) {
if (++vq->vq_ia_active == 1)
vq->vq_nia_credit = 1;
} else if (vq->vq_ia_active > 0) {
vq->vq_nia_credit--;
}
avl_add(&vq->vq_active_tree, zio);
}
static void
vdev_queue_pending_remove(vdev_queue_t *vq, zio_t *zio)
{
ASSERT(MUTEX_HELD(&vq->vq_lock));
ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE);
vq->vq_class[zio->io_priority].vqc_active--;
if (vdev_queue_is_interactive(zio->io_priority)) {
if (--vq->vq_ia_active == 0)
vq->vq_nia_credit = 0;
else
vq->vq_nia_credit = zfs_vdev_nia_credit;
} else if (vq->vq_ia_active == 0)
vq->vq_nia_credit++;
avl_remove(&vq->vq_active_tree, zio);
}
static void
vdev_queue_agg_io_done(zio_t *aio)
{
abd_free(aio->io_abd);
}
/*
* Compute the range spanned by two i/os, which is the endpoint of the last
* (lio->io_offset + lio->io_size) minus start of the first (fio->io_offset).
* Conveniently, the gap between fio and lio is given by -IO_SPAN(lio, fio);
* thus fio and lio are adjacent if and only if IO_SPAN(lio, fio) == 0.
*/
#define IO_SPAN(fio, lio) ((lio)->io_offset + (lio)->io_size - (fio)->io_offset)
#define IO_GAP(fio, lio) (-IO_SPAN(lio, fio))
/*
* Sufficiently adjacent io_offset's in ZIOs will be aggregated. We do this
* by creating a gang ABD from the adjacent ZIOs io_abd's. By using
* a gang ABD we avoid doing memory copies to and from the parent,
* child ZIOs. The gang ABD also accounts for gaps between adjacent
* io_offsets by simply getting the zero ABD for writes or allocating
* a new ABD for reads and placing them in the gang ABD as well.
*/
static zio_t *
vdev_queue_aggregate(vdev_queue_t *vq, zio_t *zio)
{
zio_t *first, *last, *aio, *dio, *mandatory, *nio;
uint64_t maxgap = 0;
uint64_t size;
uint64_t limit;
int maxblocksize;
boolean_t stretch = B_FALSE;
avl_tree_t *t = vdev_queue_type_tree(vq, zio->io_type);
zio_flag_t flags = zio->io_flags & ZIO_FLAG_AGG_INHERIT;
uint64_t next_offset;
abd_t *abd;
maxblocksize = spa_maxblocksize(vq->vq_vdev->vdev_spa);
if (vq->vq_vdev->vdev_nonrot)
limit = zfs_vdev_aggregation_limit_non_rotating;
else
limit = zfs_vdev_aggregation_limit;
limit = MIN(limit, maxblocksize);
if (zio->io_flags & ZIO_FLAG_DONT_AGGREGATE || limit == 0)
return (NULL);
/*
* While TRIM commands could be aggregated based on offset this
* behavior is disabled until it's determined to be beneficial.
*/
if (zio->io_type == ZIO_TYPE_TRIM && !zfs_vdev_aggregate_trim)
return (NULL);
/*
* I/Os to distributed spares are directly dispatched to the dRAID
* leaf vdevs for aggregation. See the comment at the end of the
* zio_vdev_io_start() function.
*/
ASSERT(vq->vq_vdev->vdev_ops != &vdev_draid_spare_ops);
first = last = zio;
if (zio->io_type == ZIO_TYPE_READ)
maxgap = zfs_vdev_read_gap_limit;
/*
* We can aggregate I/Os that are sufficiently adjacent and of
* the same flavor, as expressed by the AGG_INHERIT flags.
* The latter requirement is necessary so that certain
* attributes of the I/O, such as whether it's a normal I/O
* or a scrub/resilver, can be preserved in the aggregate.
* We can include optional I/Os, but don't allow them
* to begin a range as they add no benefit in that situation.
*/
/*
* We keep track of the last non-optional I/O.
*/
mandatory = (first->io_flags & ZIO_FLAG_OPTIONAL) ? NULL : first;
/*
* Walk backwards through sufficiently contiguous I/Os
* recording the last non-optional I/O.
*/
while ((dio = AVL_PREV(t, first)) != NULL &&
(dio->io_flags & ZIO_FLAG_AGG_INHERIT) == flags &&
IO_SPAN(dio, last) <= limit &&
IO_GAP(dio, first) <= maxgap &&
dio->io_type == zio->io_type) {
first = dio;
if (mandatory == NULL && !(first->io_flags & ZIO_FLAG_OPTIONAL))
mandatory = first;
}
/*
* Skip any initial optional I/Os.
*/
while ((first->io_flags & ZIO_FLAG_OPTIONAL) && first != last) {
first = AVL_NEXT(t, first);
ASSERT(first != NULL);
}
/*
* Walk forward through sufficiently contiguous I/Os.
* The aggregation limit does not apply to optional i/os, so that
* we can issue contiguous writes even if they are larger than the
* aggregation limit.
*/
while ((dio = AVL_NEXT(t, last)) != NULL &&
(dio->io_flags & ZIO_FLAG_AGG_INHERIT) == flags &&
(IO_SPAN(first, dio) <= limit ||
(dio->io_flags & ZIO_FLAG_OPTIONAL)) &&
IO_SPAN(first, dio) <= maxblocksize &&
IO_GAP(last, dio) <= maxgap &&
dio->io_type == zio->io_type) {
last = dio;
if (!(last->io_flags & ZIO_FLAG_OPTIONAL))
mandatory = last;
}
/*
* Now that we've established the range of the I/O aggregation
* we must decide what to do with trailing optional I/Os.
* For reads, there's nothing to do. While we are unable to
* aggregate further, it's possible that a trailing optional
* I/O would allow the underlying device to aggregate with
* subsequent I/Os. We must therefore determine if the next
* non-optional I/O is close enough to make aggregation
* worthwhile.
*/
if (zio->io_type == ZIO_TYPE_WRITE && mandatory != NULL) {
zio_t *nio = last;
while ((dio = AVL_NEXT(t, nio)) != NULL &&
IO_GAP(nio, dio) == 0 &&
IO_GAP(mandatory, dio) <= zfs_vdev_write_gap_limit) {
nio = dio;
if (!(nio->io_flags & ZIO_FLAG_OPTIONAL)) {
stretch = B_TRUE;
break;
}
}
}
if (stretch) {
/*
* We are going to include an optional io in our aggregated
* span, thus closing the write gap. Only mandatory i/os can
* start aggregated spans, so make sure that the next i/o
* after our span is mandatory.
*/
dio = AVL_NEXT(t, last);
ASSERT3P(dio, !=, NULL);
dio->io_flags &= ~ZIO_FLAG_OPTIONAL;
} else {
/* do not include the optional i/o */
while (last != mandatory && last != first) {
ASSERT(last->io_flags & ZIO_FLAG_OPTIONAL);
last = AVL_PREV(t, last);
ASSERT(last != NULL);
}
}
if (first == last)
return (NULL);
size = IO_SPAN(first, last);
ASSERT3U(size, <=, maxblocksize);
abd = abd_alloc_gang();
if (abd == NULL)
return (NULL);
aio = zio_vdev_delegated_io(first->io_vd, first->io_offset,
abd, size, first->io_type, zio->io_priority,
flags | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE,
vdev_queue_agg_io_done, NULL);
aio->io_timestamp = first->io_timestamp;
nio = first;
next_offset = first->io_offset;
do {
dio = nio;
nio = AVL_NEXT(t, dio);
ASSERT3P(dio, !=, NULL);
zio_add_child(dio, aio);
vdev_queue_io_remove(vq, dio);
if (dio->io_offset != next_offset) {
/* allocate a buffer for a read gap */
ASSERT3U(dio->io_type, ==, ZIO_TYPE_READ);
ASSERT3U(dio->io_offset, >, next_offset);
abd = abd_alloc_for_io(
dio->io_offset - next_offset, B_TRUE);
abd_gang_add(aio->io_abd, abd, B_TRUE);
}
if (dio->io_abd &&
(dio->io_size != abd_get_size(dio->io_abd))) {
/* abd size not the same as IO size */
ASSERT3U(abd_get_size(dio->io_abd), >, dio->io_size);
abd = abd_get_offset_size(dio->io_abd, 0, dio->io_size);
abd_gang_add(aio->io_abd, abd, B_TRUE);
} else {
if (dio->io_flags & ZIO_FLAG_NODATA) {
/* allocate a buffer for a write gap */
ASSERT3U(dio->io_type, ==, ZIO_TYPE_WRITE);
ASSERT3P(dio->io_abd, ==, NULL);
abd_gang_add(aio->io_abd,
abd_get_zeros(dio->io_size), B_TRUE);
} else {
/*
* We pass B_FALSE to abd_gang_add()
* because we did not allocate a new
* ABD, so it is assumed the caller
* will free this ABD.
*/
abd_gang_add(aio->io_abd, dio->io_abd,
B_FALSE);
}
}
next_offset = dio->io_offset + dio->io_size;
} while (dio != last);
ASSERT3U(abd_get_size(aio->io_abd), ==, aio->io_size);
/*
* Callers must call zio_vdev_io_bypass() and zio_execute() for
* aggregated (parent) I/Os so that we could avoid dropping the
* queue's lock here to avoid a deadlock that we could encounter
* due to lock order reversal between vq_lock and io_lock in
* zio_change_priority().
*/
return (aio);
}
static zio_t *
vdev_queue_io_to_issue(vdev_queue_t *vq)
{
zio_t *zio, *aio;
zio_priority_t p;
avl_index_t idx;
avl_tree_t *tree;
again:
ASSERT(MUTEX_HELD(&vq->vq_lock));
p = vdev_queue_class_to_issue(vq);
if (p == ZIO_PRIORITY_NUM_QUEUEABLE) {
/* No eligible queued i/os */
return (NULL);
}
/*
* For LBA-ordered queues (async / scrub / initializing), issue the
* i/o which follows the most recently issued i/o in LBA (offset) order.
*
* For FIFO queues (sync/trim), issue the i/o with the lowest timestamp.
*/
tree = vdev_queue_class_tree(vq, p);
vq->vq_io_search.io_timestamp = 0;
vq->vq_io_search.io_offset = vq->vq_last_offset - 1;
VERIFY3P(avl_find(tree, &vq->vq_io_search, &idx), ==, NULL);
zio = avl_nearest(tree, idx, AVL_AFTER);
if (zio == NULL)
zio = avl_first(tree);
ASSERT3U(zio->io_priority, ==, p);
aio = vdev_queue_aggregate(vq, zio);
if (aio != NULL) {
zio = aio;
} else {
vdev_queue_io_remove(vq, zio);
/*
* If the I/O is or was optional and therefore has no data, we
* need to simply discard it. We need to drop the vdev queue's
* lock to avoid a deadlock that we could encounter since this
* I/O will complete immediately.
*/
if (zio->io_flags & ZIO_FLAG_NODATA) {
mutex_exit(&vq->vq_lock);
zio_vdev_io_bypass(zio);
zio_execute(zio);
mutex_enter(&vq->vq_lock);
goto again;
}
}
vdev_queue_pending_add(vq, zio);
vq->vq_last_offset = zio->io_offset + zio->io_size;
return (zio);
}
zio_t *
vdev_queue_io(zio_t *zio)
{
vdev_queue_t *vq = &zio->io_vd->vdev_queue;
zio_t *dio, *nio;
zio_link_t *zl = NULL;
if (zio->io_flags & ZIO_FLAG_DONT_QUEUE)
return (zio);
/*
* Children i/os inherent their parent's priority, which might
* not match the child's i/o type. Fix it up here.
*/
if (zio->io_type == ZIO_TYPE_READ) {
ASSERT(zio->io_priority != ZIO_PRIORITY_TRIM);
if (zio->io_priority != ZIO_PRIORITY_SYNC_READ &&
zio->io_priority != ZIO_PRIORITY_ASYNC_READ &&
zio->io_priority != ZIO_PRIORITY_SCRUB &&
zio->io_priority != ZIO_PRIORITY_REMOVAL &&
zio->io_priority != ZIO_PRIORITY_INITIALIZING &&
zio->io_priority != ZIO_PRIORITY_REBUILD) {
zio->io_priority = ZIO_PRIORITY_ASYNC_READ;
}
} else if (zio->io_type == ZIO_TYPE_WRITE) {
ASSERT(zio->io_priority != ZIO_PRIORITY_TRIM);
if (zio->io_priority != ZIO_PRIORITY_SYNC_WRITE &&
zio->io_priority != ZIO_PRIORITY_ASYNC_WRITE &&
zio->io_priority != ZIO_PRIORITY_REMOVAL &&
zio->io_priority != ZIO_PRIORITY_INITIALIZING &&
zio->io_priority != ZIO_PRIORITY_REBUILD) {
zio->io_priority = ZIO_PRIORITY_ASYNC_WRITE;
}
} else {
ASSERT(zio->io_type == ZIO_TYPE_TRIM);
ASSERT(zio->io_priority == ZIO_PRIORITY_TRIM);
}
zio->io_flags |= ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE;
zio->io_timestamp = gethrtime();
mutex_enter(&vq->vq_lock);
vdev_queue_io_add(vq, zio);
nio = vdev_queue_io_to_issue(vq);
mutex_exit(&vq->vq_lock);
if (nio == NULL)
return (NULL);
if (nio->io_done == vdev_queue_agg_io_done) {
while ((dio = zio_walk_parents(nio, &zl)) != NULL) {
ASSERT3U(dio->io_type, ==, nio->io_type);
zio_vdev_io_bypass(dio);
zio_execute(dio);
}
zio_nowait(nio);
return (NULL);
}
return (nio);
}
void
vdev_queue_io_done(zio_t *zio)
{
vdev_queue_t *vq = &zio->io_vd->vdev_queue;
zio_t *dio, *nio;
zio_link_t *zl = NULL;
hrtime_t now = gethrtime();
vq->vq_io_complete_ts = now;
vq->vq_io_delta_ts = zio->io_delta = now - zio->io_timestamp;
mutex_enter(&vq->vq_lock);
vdev_queue_pending_remove(vq, zio);
while ((nio = vdev_queue_io_to_issue(vq)) != NULL) {
mutex_exit(&vq->vq_lock);
if (nio->io_done == vdev_queue_agg_io_done) {
while ((dio = zio_walk_parents(nio, &zl)) != NULL) {
ASSERT3U(dio->io_type, ==, nio->io_type);
zio_vdev_io_bypass(dio);
zio_execute(dio);
}
zio_nowait(nio);
} else {
zio_vdev_io_reissue(nio);
zio_execute(nio);
}
mutex_enter(&vq->vq_lock);
}
mutex_exit(&vq->vq_lock);
}
void
vdev_queue_change_io_priority(zio_t *zio, zio_priority_t priority)
{
vdev_queue_t *vq = &zio->io_vd->vdev_queue;
avl_tree_t *tree;
/*
* ZIO_PRIORITY_NOW is used by the vdev cache code and the aggregate zio
* code to issue IOs without adding them to the vdev queue. In this
* case, the zio is already going to be issued as quickly as possible
* and so it doesn't need any reprioritization to help.
*/
if (zio->io_priority == ZIO_PRIORITY_NOW)
return;
ASSERT3U(zio->io_priority, <, ZIO_PRIORITY_NUM_QUEUEABLE);
ASSERT3U(priority, <, ZIO_PRIORITY_NUM_QUEUEABLE);
if (zio->io_type == ZIO_TYPE_READ) {
if (priority != ZIO_PRIORITY_SYNC_READ &&
priority != ZIO_PRIORITY_ASYNC_READ &&
priority != ZIO_PRIORITY_SCRUB)
priority = ZIO_PRIORITY_ASYNC_READ;
} else {
ASSERT(zio->io_type == ZIO_TYPE_WRITE);
if (priority != ZIO_PRIORITY_SYNC_WRITE &&
priority != ZIO_PRIORITY_ASYNC_WRITE)
priority = ZIO_PRIORITY_ASYNC_WRITE;
}
mutex_enter(&vq->vq_lock);
/*
* If the zio is in none of the queues we can simply change
* the priority. If the zio is waiting to be submitted we must
* remove it from the queue and re-insert it with the new priority.
* Otherwise, the zio is currently active and we cannot change its
* priority.
*/
tree = vdev_queue_class_tree(vq, zio->io_priority);
if (avl_find(tree, zio, NULL) == zio) {
avl_remove(vdev_queue_class_tree(vq, zio->io_priority), zio);
zio->io_priority = priority;
avl_add(vdev_queue_class_tree(vq, zio->io_priority), zio);
} else if (avl_find(&vq->vq_active_tree, zio, NULL) != zio) {
zio->io_priority = priority;
}
mutex_exit(&vq->vq_lock);
}
/*
* As these two methods are only used for load calculations we're not
* concerned if we get an incorrect value on 32bit platforms due to lack of
* vq_lock mutex use here, instead we prefer to keep it lock free for
* performance.
*/
int
vdev_queue_length(vdev_t *vd)
{
return (avl_numnodes(&vd->vdev_queue.vq_active_tree));
}
uint64_t
vdev_queue_last_offset(vdev_t *vd)
{
return (vd->vdev_queue.vq_last_offset);
}
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, aggregation_limit, UINT, ZMOD_RW,
"Max vdev I/O aggregation size");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, aggregation_limit_non_rotating, UINT,
ZMOD_RW, "Max vdev I/O aggregation size for non-rotating media");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, aggregate_trim, UINT, ZMOD_RW,
"Allow TRIM I/O to be aggregated");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, read_gap_limit, UINT, ZMOD_RW,
"Aggregate read I/O over gap");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, write_gap_limit, UINT, ZMOD_RW,
"Aggregate write I/O over gap");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, max_active, UINT, ZMOD_RW,
"Maximum number of active I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, async_write_active_max_dirty_percent,
UINT, ZMOD_RW, "Async write concurrency max threshold");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, async_write_active_min_dirty_percent,
UINT, ZMOD_RW, "Async write concurrency min threshold");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, async_read_max_active, UINT, ZMOD_RW,
"Max active async read I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, async_read_min_active, UINT, ZMOD_RW,
"Min active async read I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, async_write_max_active, UINT, ZMOD_RW,
"Max active async write I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, async_write_min_active, UINT, ZMOD_RW,
"Min active async write I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, initializing_max_active, UINT, ZMOD_RW,
"Max active initializing I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, initializing_min_active, UINT, ZMOD_RW,
"Min active initializing I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, removal_max_active, UINT, ZMOD_RW,
"Max active removal I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, removal_min_active, UINT, ZMOD_RW,
"Min active removal I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, scrub_max_active, UINT, ZMOD_RW,
"Max active scrub I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, scrub_min_active, UINT, ZMOD_RW,
"Min active scrub I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, sync_read_max_active, UINT, ZMOD_RW,
"Max active sync read I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, sync_read_min_active, UINT, ZMOD_RW,
"Min active sync read I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, sync_write_max_active, UINT, ZMOD_RW,
"Max active sync write I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, sync_write_min_active, UINT, ZMOD_RW,
"Min active sync write I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, trim_max_active, UINT, ZMOD_RW,
"Max active trim/discard I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, trim_min_active, UINT, ZMOD_RW,
"Min active trim/discard I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, rebuild_max_active, UINT, ZMOD_RW,
"Max active rebuild I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, rebuild_min_active, UINT, ZMOD_RW,
"Min active rebuild I/Os per vdev");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, nia_credit, UINT, ZMOD_RW,
"Number of non-interactive I/Os to allow in sequence");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, nia_delay, UINT, ZMOD_RW,
"Number of non-interactive I/Os before _max_active");
ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, queue_depth_pct, UINT, ZMOD_RW,
"Queue depth percentage for each top-level vdev");
+
+ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, def_queue_depth, UINT, ZMOD_RW,
+ "Default queue depth for each allocator");
diff --git a/sys/contrib/openzfs/module/zfs/zfs_fm.c b/sys/contrib/openzfs/module/zfs/zfs_fm.c
index 7169e49ac46a..bdd0e96c327a 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_fm.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_fm.c
@@ -1,1598 +1,1598 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2009 Sun Microsystems, Inc. All rights reserved.
* Use is subject to license terms.
*/
/*
* Copyright (c) 2012,2021 by Delphix. All rights reserved.
*/
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/vdev.h>
#include <sys/vdev_impl.h>
#include <sys/zio.h>
#include <sys/zio_checksum.h>
#include <sys/fm/fs/zfs.h>
#include <sys/fm/protocol.h>
#include <sys/fm/util.h>
#include <sys/sysevent.h>
/*
* This general routine is responsible for generating all the different ZFS
* ereports. The payload is dependent on the class, and which arguments are
* supplied to the function:
*
* EREPORT POOL VDEV IO
* block X X X
* data X X
* device X X
* pool X
*
* If we are in a loading state, all errors are chained together by the same
* SPA-wide ENA (Error Numeric Association).
*
* For isolated I/O requests, we get the ENA from the zio_t. The propagation
* gets very complicated due to RAID-Z, gang blocks, and vdev caching. We want
* to chain together all ereports associated with a logical piece of data. For
* read I/Os, there are basically three 'types' of I/O, which form a roughly
* layered diagram:
*
* +---------------+
* | Aggregate I/O | No associated logical data or device
* +---------------+
* |
* V
* +---------------+ Reads associated with a piece of logical data.
* | Read I/O | This includes reads on behalf of RAID-Z,
* +---------------+ mirrors, gang blocks, retries, etc.
* |
* V
* +---------------+ Reads associated with a particular device, but
* | Physical I/O | no logical data. Issued as part of vdev caching
* +---------------+ and I/O aggregation.
*
* Note that 'physical I/O' here is not the same terminology as used in the rest
* of ZIO. Typically, 'physical I/O' simply means that there is no attached
* blockpointer. But I/O with no associated block pointer can still be related
* to a logical piece of data (i.e. RAID-Z requests).
*
* Purely physical I/O always have unique ENAs. They are not related to a
* particular piece of logical data, and therefore cannot be chained together.
* We still generate an ereport, but the DE doesn't correlate it with any
* logical piece of data. When such an I/O fails, the delegated I/O requests
* will issue a retry, which will trigger the 'real' ereport with the correct
* ENA.
*
* We keep track of the ENA for a ZIO chain through the 'io_logical' member.
* When a new logical I/O is issued, we set this to point to itself. Child I/Os
* then inherit this pointer, so that when it is first set subsequent failures
* will use the same ENA. For vdev cache fill and queue aggregation I/O,
* this pointer is set to NULL, and no ereport will be generated (since it
* doesn't actually correspond to any particular device or piece of data,
* and the caller will always retry without caching or queueing anyway).
*
* For checksum errors, we want to include more information about the actual
* error which occurs. Accordingly, we build an ereport when the error is
* noticed, but instead of sending it in immediately, we hang it off of the
* io_cksum_report field of the logical IO. When the logical IO completes
* (successfully or not), zfs_ereport_finish_checksum() is called with the
* good and bad versions of the buffer (if available), and we annotate the
* ereport with information about the differences.
*/
#ifdef _KERNEL
/*
* Duplicate ereport Detection
*
* Some ereports are retained momentarily for detecting duplicates. These
* are kept in a recent_events_node_t in both a time-ordered list and an AVL
* tree of recent unique ereports.
*
* The lifespan of these recent ereports is bounded (15 mins) and a cleaner
* task is used to purge stale entries.
*/
static list_t recent_events_list;
static avl_tree_t recent_events_tree;
static kmutex_t recent_events_lock;
static taskqid_t recent_events_cleaner_tqid;
/*
* Each node is about 128 bytes so 2,000 would consume 1/4 MiB.
*
* This setting can be changed dynamically and setting it to zero
* disables duplicate detection.
*/
static unsigned int zfs_zevent_retain_max = 2000;
/*
* The lifespan for a recent ereport entry. The default of 15 minutes is
* intended to outlive the zfs diagnosis engine's threshold of 10 errors
* over a period of 10 minutes.
*/
static unsigned int zfs_zevent_retain_expire_secs = 900;
typedef enum zfs_subclass {
ZSC_IO,
ZSC_DATA,
ZSC_CHECKSUM
} zfs_subclass_t;
typedef struct {
/* common criteria */
uint64_t re_pool_guid;
uint64_t re_vdev_guid;
int re_io_error;
uint64_t re_io_size;
uint64_t re_io_offset;
zfs_subclass_t re_subclass;
zio_priority_t re_io_priority;
/* logical zio criteria (optional) */
zbookmark_phys_t re_io_bookmark;
/* internal state */
avl_node_t re_tree_link;
list_node_t re_list_link;
uint64_t re_timestamp;
} recent_events_node_t;
static int
recent_events_compare(const void *a, const void *b)
{
const recent_events_node_t *node1 = a;
const recent_events_node_t *node2 = b;
int cmp;
/*
* The comparison order here is somewhat arbitrary.
* What's important is that if every criteria matches, then it
* is a duplicate (i.e. compare returns 0)
*/
if ((cmp = TREE_CMP(node1->re_subclass, node2->re_subclass)) != 0)
return (cmp);
if ((cmp = TREE_CMP(node1->re_pool_guid, node2->re_pool_guid)) != 0)
return (cmp);
if ((cmp = TREE_CMP(node1->re_vdev_guid, node2->re_vdev_guid)) != 0)
return (cmp);
if ((cmp = TREE_CMP(node1->re_io_error, node2->re_io_error)) != 0)
return (cmp);
if ((cmp = TREE_CMP(node1->re_io_priority, node2->re_io_priority)) != 0)
return (cmp);
if ((cmp = TREE_CMP(node1->re_io_size, node2->re_io_size)) != 0)
return (cmp);
if ((cmp = TREE_CMP(node1->re_io_offset, node2->re_io_offset)) != 0)
return (cmp);
const zbookmark_phys_t *zb1 = &node1->re_io_bookmark;
const zbookmark_phys_t *zb2 = &node2->re_io_bookmark;
if ((cmp = TREE_CMP(zb1->zb_objset, zb2->zb_objset)) != 0)
return (cmp);
if ((cmp = TREE_CMP(zb1->zb_object, zb2->zb_object)) != 0)
return (cmp);
if ((cmp = TREE_CMP(zb1->zb_level, zb2->zb_level)) != 0)
return (cmp);
if ((cmp = TREE_CMP(zb1->zb_blkid, zb2->zb_blkid)) != 0)
return (cmp);
return (0);
}
/*
* workaround: vdev properties don't have inheritance
*/
static uint64_t
vdev_prop_get_inherited(vdev_t *vd, vdev_prop_t prop)
{
uint64_t propdef, propval;
propdef = vdev_prop_default_numeric(prop);
switch (prop) {
case VDEV_PROP_CHECKSUM_N:
propval = vd->vdev_checksum_n;
break;
case VDEV_PROP_CHECKSUM_T:
propval = vd->vdev_checksum_t;
break;
case VDEV_PROP_IO_N:
propval = vd->vdev_io_n;
break;
case VDEV_PROP_IO_T:
propval = vd->vdev_io_t;
break;
default:
propval = propdef;
break;
}
if (propval != propdef)
return (propval);
if (vd->vdev_parent == NULL)
return (propdef);
return (vdev_prop_get_inherited(vd->vdev_parent, prop));
}
static void zfs_ereport_schedule_cleaner(void);
/*
* background task to clean stale recent event nodes.
*/
static void
zfs_ereport_cleaner(void *arg)
{
recent_events_node_t *entry;
uint64_t now = gethrtime();
/*
* purge expired entries
*/
mutex_enter(&recent_events_lock);
while ((entry = list_tail(&recent_events_list)) != NULL) {
uint64_t age = NSEC2SEC(now - entry->re_timestamp);
if (age <= zfs_zevent_retain_expire_secs)
break;
/* remove expired node */
avl_remove(&recent_events_tree, entry);
list_remove(&recent_events_list, entry);
kmem_free(entry, sizeof (*entry));
}
/* Restart the cleaner if more entries remain */
recent_events_cleaner_tqid = 0;
if (!list_is_empty(&recent_events_list))
zfs_ereport_schedule_cleaner();
mutex_exit(&recent_events_lock);
}
static void
zfs_ereport_schedule_cleaner(void)
{
ASSERT(MUTEX_HELD(&recent_events_lock));
uint64_t timeout = SEC2NSEC(zfs_zevent_retain_expire_secs + 1);
recent_events_cleaner_tqid = taskq_dispatch_delay(
system_delay_taskq, zfs_ereport_cleaner, NULL, TQ_SLEEP,
ddi_get_lbolt() + NSEC_TO_TICK(timeout));
}
/*
* Clear entries for a given vdev or all vdevs in a pool when vdev == NULL
*/
void
zfs_ereport_clear(spa_t *spa, vdev_t *vd)
{
uint64_t vdev_guid, pool_guid;
ASSERT(vd != NULL || spa != NULL);
if (vd == NULL) {
vdev_guid = 0;
pool_guid = spa_guid(spa);
} else {
vdev_guid = vd->vdev_guid;
pool_guid = 0;
}
mutex_enter(&recent_events_lock);
recent_events_node_t *next = list_head(&recent_events_list);
while (next != NULL) {
recent_events_node_t *entry = next;
next = list_next(&recent_events_list, next);
if (entry->re_vdev_guid == vdev_guid ||
entry->re_pool_guid == pool_guid) {
avl_remove(&recent_events_tree, entry);
list_remove(&recent_events_list, entry);
kmem_free(entry, sizeof (*entry));
}
}
mutex_exit(&recent_events_lock);
}
/*
* Check if an ereport would be a duplicate of one recently posted.
*
* An ereport is considered a duplicate if the set of criteria in
* recent_events_node_t all match.
*
* Only FM_EREPORT_ZFS_IO, FM_EREPORT_ZFS_DATA, and FM_EREPORT_ZFS_CHECKSUM
* are candidates for duplicate checking.
*/
static boolean_t
zfs_ereport_is_duplicate(const char *subclass, spa_t *spa, vdev_t *vd,
const zbookmark_phys_t *zb, zio_t *zio, uint64_t offset, uint64_t size)
{
recent_events_node_t search = {0}, *entry;
if (vd == NULL || zio == NULL)
return (B_FALSE);
if (zfs_zevent_retain_max == 0)
return (B_FALSE);
if (strcmp(subclass, FM_EREPORT_ZFS_IO) == 0)
search.re_subclass = ZSC_IO;
else if (strcmp(subclass, FM_EREPORT_ZFS_DATA) == 0)
search.re_subclass = ZSC_DATA;
else if (strcmp(subclass, FM_EREPORT_ZFS_CHECKSUM) == 0)
search.re_subclass = ZSC_CHECKSUM;
else
return (B_FALSE);
search.re_pool_guid = spa_guid(spa);
search.re_vdev_guid = vd->vdev_guid;
search.re_io_error = zio->io_error;
search.re_io_priority = zio->io_priority;
/* if size is supplied use it over what's in zio */
if (size) {
search.re_io_size = size;
search.re_io_offset = offset;
} else {
search.re_io_size = zio->io_size;
search.re_io_offset = zio->io_offset;
}
/* grab optional logical zio criteria */
if (zb != NULL) {
search.re_io_bookmark.zb_objset = zb->zb_objset;
search.re_io_bookmark.zb_object = zb->zb_object;
search.re_io_bookmark.zb_level = zb->zb_level;
search.re_io_bookmark.zb_blkid = zb->zb_blkid;
}
uint64_t now = gethrtime();
mutex_enter(&recent_events_lock);
/* check if we have seen this one recently */
entry = avl_find(&recent_events_tree, &search, NULL);
if (entry != NULL) {
uint64_t age = NSEC2SEC(now - entry->re_timestamp);
/*
* There is still an active cleaner (since we're here).
* Reset the last seen time for this duplicate entry
* so that its lifespand gets extended.
*/
list_remove(&recent_events_list, entry);
list_insert_head(&recent_events_list, entry);
entry->re_timestamp = now;
zfs_zevent_track_duplicate();
mutex_exit(&recent_events_lock);
return (age <= zfs_zevent_retain_expire_secs);
}
if (avl_numnodes(&recent_events_tree) >= zfs_zevent_retain_max) {
/* recycle oldest node */
entry = list_tail(&recent_events_list);
ASSERT(entry != NULL);
list_remove(&recent_events_list, entry);
avl_remove(&recent_events_tree, entry);
} else {
entry = kmem_alloc(sizeof (recent_events_node_t), KM_SLEEP);
}
/* record this as a recent ereport */
*entry = search;
avl_add(&recent_events_tree, entry);
list_insert_head(&recent_events_list, entry);
entry->re_timestamp = now;
/* Start a cleaner if not already scheduled */
if (recent_events_cleaner_tqid == 0)
zfs_ereport_schedule_cleaner();
mutex_exit(&recent_events_lock);
return (B_FALSE);
}
void
zfs_zevent_post_cb(nvlist_t *nvl, nvlist_t *detector)
{
if (nvl)
fm_nvlist_destroy(nvl, FM_NVA_FREE);
if (detector)
fm_nvlist_destroy(detector, FM_NVA_FREE);
}
/*
* We want to rate limit ZIO delay, deadman, and checksum events so as to not
* flood zevent consumers when a disk is acting up.
*
* Returns 1 if we're ratelimiting, 0 if not.
*/
static int
zfs_is_ratelimiting_event(const char *subclass, vdev_t *vd)
{
int rc = 0;
/*
* zfs_ratelimit() returns 1 if we're *not* ratelimiting and 0 if we
* are. Invert it to get our return value.
*/
if (strcmp(subclass, FM_EREPORT_ZFS_DELAY) == 0) {
rc = !zfs_ratelimit(&vd->vdev_delay_rl);
} else if (strcmp(subclass, FM_EREPORT_ZFS_DEADMAN) == 0) {
rc = !zfs_ratelimit(&vd->vdev_deadman_rl);
} else if (strcmp(subclass, FM_EREPORT_ZFS_CHECKSUM) == 0) {
rc = !zfs_ratelimit(&vd->vdev_checksum_rl);
}
if (rc) {
/* We're rate limiting */
fm_erpt_dropped_increment();
}
return (rc);
}
/*
* Return B_TRUE if the event actually posted, B_FALSE if not.
*/
static boolean_t
zfs_ereport_start(nvlist_t **ereport_out, nvlist_t **detector_out,
const char *subclass, spa_t *spa, vdev_t *vd, const zbookmark_phys_t *zb,
zio_t *zio, uint64_t stateoroffset, uint64_t size)
{
nvlist_t *ereport, *detector;
uint64_t ena;
char class[64];
if ((ereport = fm_nvlist_create(NULL)) == NULL)
return (B_FALSE);
if ((detector = fm_nvlist_create(NULL)) == NULL) {
fm_nvlist_destroy(ereport, FM_NVA_FREE);
return (B_FALSE);
}
/*
* Serialize ereport generation
*/
mutex_enter(&spa->spa_errlist_lock);
/*
* Determine the ENA to use for this event. If we are in a loading
* state, use a SPA-wide ENA. Otherwise, if we are in an I/O state, use
* a root zio-wide ENA. Otherwise, simply use a unique ENA.
*/
if (spa_load_state(spa) != SPA_LOAD_NONE) {
if (spa->spa_ena == 0)
spa->spa_ena = fm_ena_generate(0, FM_ENA_FMT1);
ena = spa->spa_ena;
} else if (zio != NULL && zio->io_logical != NULL) {
if (zio->io_logical->io_ena == 0)
zio->io_logical->io_ena =
fm_ena_generate(0, FM_ENA_FMT1);
ena = zio->io_logical->io_ena;
} else {
ena = fm_ena_generate(0, FM_ENA_FMT1);
}
/*
* Construct the full class, detector, and other standard FMA fields.
*/
(void) snprintf(class, sizeof (class), "%s.%s",
ZFS_ERROR_CLASS, subclass);
fm_fmri_zfs_set(detector, FM_ZFS_SCHEME_VERSION, spa_guid(spa),
vd != NULL ? vd->vdev_guid : 0);
fm_ereport_set(ereport, FM_EREPORT_VERSION, class, ena, detector, NULL);
/*
* Construct the per-ereport payload, depending on which parameters are
* passed in.
*/
/*
* Generic payload members common to all ereports.
*/
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_POOL, DATA_TYPE_STRING, spa_name(spa),
FM_EREPORT_PAYLOAD_ZFS_POOL_GUID, DATA_TYPE_UINT64, spa_guid(spa),
FM_EREPORT_PAYLOAD_ZFS_POOL_STATE, DATA_TYPE_UINT64,
(uint64_t)spa_state(spa),
FM_EREPORT_PAYLOAD_ZFS_POOL_CONTEXT, DATA_TYPE_INT32,
(int32_t)spa_load_state(spa), NULL);
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_POOL_FAILMODE,
DATA_TYPE_STRING,
spa_get_failmode(spa) == ZIO_FAILURE_MODE_WAIT ?
FM_EREPORT_FAILMODE_WAIT :
spa_get_failmode(spa) == ZIO_FAILURE_MODE_CONTINUE ?
FM_EREPORT_FAILMODE_CONTINUE : FM_EREPORT_FAILMODE_PANIC,
NULL);
if (vd != NULL) {
vdev_t *pvd = vd->vdev_parent;
vdev_queue_t *vq = &vd->vdev_queue;
vdev_stat_t *vs = &vd->vdev_stat;
vdev_t *spare_vd;
uint64_t *spare_guids;
char **spare_paths;
int i, spare_count;
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_VDEV_GUID,
DATA_TYPE_UINT64, vd->vdev_guid,
FM_EREPORT_PAYLOAD_ZFS_VDEV_TYPE,
DATA_TYPE_STRING, vd->vdev_ops->vdev_op_type, NULL);
if (vd->vdev_path != NULL)
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_PATH,
DATA_TYPE_STRING, vd->vdev_path, NULL);
if (vd->vdev_devid != NULL)
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_DEVID,
DATA_TYPE_STRING, vd->vdev_devid, NULL);
if (vd->vdev_fru != NULL)
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_FRU,
DATA_TYPE_STRING, vd->vdev_fru, NULL);
if (vd->vdev_enc_sysfs_path != NULL)
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_ENC_SYSFS_PATH,
DATA_TYPE_STRING, vd->vdev_enc_sysfs_path, NULL);
if (vd->vdev_ashift)
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_ASHIFT,
DATA_TYPE_UINT64, vd->vdev_ashift, NULL);
if (vq != NULL) {
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_COMP_TS,
DATA_TYPE_UINT64, vq->vq_io_complete_ts, NULL);
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_DELTA_TS,
DATA_TYPE_UINT64, vq->vq_io_delta_ts, NULL);
}
if (vs != NULL) {
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_READ_ERRORS,
DATA_TYPE_UINT64, vs->vs_read_errors,
FM_EREPORT_PAYLOAD_ZFS_VDEV_WRITE_ERRORS,
DATA_TYPE_UINT64, vs->vs_write_errors,
FM_EREPORT_PAYLOAD_ZFS_VDEV_CKSUM_ERRORS,
DATA_TYPE_UINT64, vs->vs_checksum_errors,
FM_EREPORT_PAYLOAD_ZFS_VDEV_DELAYS,
DATA_TYPE_UINT64, vs->vs_slow_ios,
NULL);
}
if (pvd != NULL) {
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_PARENT_GUID,
DATA_TYPE_UINT64, pvd->vdev_guid,
FM_EREPORT_PAYLOAD_ZFS_PARENT_TYPE,
DATA_TYPE_STRING, pvd->vdev_ops->vdev_op_type,
NULL);
if (pvd->vdev_path)
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_PARENT_PATH,
DATA_TYPE_STRING, pvd->vdev_path, NULL);
if (pvd->vdev_devid)
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_PARENT_DEVID,
DATA_TYPE_STRING, pvd->vdev_devid, NULL);
}
spare_count = spa->spa_spares.sav_count;
spare_paths = kmem_zalloc(sizeof (char *) * spare_count,
KM_SLEEP);
spare_guids = kmem_zalloc(sizeof (uint64_t) * spare_count,
KM_SLEEP);
for (i = 0; i < spare_count; i++) {
spare_vd = spa->spa_spares.sav_vdevs[i];
if (spare_vd) {
spare_paths[i] = spare_vd->vdev_path;
spare_guids[i] = spare_vd->vdev_guid;
}
}
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_VDEV_SPARE_PATHS,
DATA_TYPE_STRING_ARRAY, spare_count, spare_paths,
FM_EREPORT_PAYLOAD_ZFS_VDEV_SPARE_GUIDS,
DATA_TYPE_UINT64_ARRAY, spare_count, spare_guids, NULL);
kmem_free(spare_guids, sizeof (uint64_t) * spare_count);
kmem_free(spare_paths, sizeof (char *) * spare_count);
}
if (zio != NULL) {
/*
* Payload common to all I/Os.
*/
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_ZIO_ERR,
DATA_TYPE_INT32, zio->io_error, NULL);
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_ZIO_FLAGS,
DATA_TYPE_INT32, zio->io_flags, NULL);
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_ZIO_STAGE,
DATA_TYPE_UINT32, zio->io_stage, NULL);
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_ZIO_PIPELINE,
DATA_TYPE_UINT32, zio->io_pipeline, NULL);
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_ZIO_DELAY,
DATA_TYPE_UINT64, zio->io_delay, NULL);
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_ZIO_TIMESTAMP,
DATA_TYPE_UINT64, zio->io_timestamp, NULL);
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_ZIO_DELTA,
DATA_TYPE_UINT64, zio->io_delta, NULL);
fm_payload_set(ereport, FM_EREPORT_PAYLOAD_ZFS_ZIO_PRIORITY,
DATA_TYPE_UINT32, zio->io_priority, NULL);
/*
* If the 'size' parameter is non-zero, it indicates this is a
* RAID-Z or other I/O where the physical offset and length are
* provided for us, instead of within the zio_t.
*/
if (vd != NULL) {
if (size)
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_ZIO_OFFSET,
DATA_TYPE_UINT64, stateoroffset,
FM_EREPORT_PAYLOAD_ZFS_ZIO_SIZE,
DATA_TYPE_UINT64, size, NULL);
else
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_ZIO_OFFSET,
DATA_TYPE_UINT64, zio->io_offset,
FM_EREPORT_PAYLOAD_ZFS_ZIO_SIZE,
DATA_TYPE_UINT64, zio->io_size, NULL);
}
} else if (vd != NULL) {
/*
* If we have a vdev but no zio, this is a device fault, and the
* 'stateoroffset' parameter indicates the previous state of the
* vdev.
*/
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_PREV_STATE,
DATA_TYPE_UINT64, stateoroffset, NULL);
}
/*
* Payload for I/Os with corresponding logical information.
*/
if (zb != NULL && (zio == NULL || zio->io_logical != NULL)) {
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_ZIO_OBJSET,
DATA_TYPE_UINT64, zb->zb_objset,
FM_EREPORT_PAYLOAD_ZFS_ZIO_OBJECT,
DATA_TYPE_UINT64, zb->zb_object,
FM_EREPORT_PAYLOAD_ZFS_ZIO_LEVEL,
DATA_TYPE_INT64, zb->zb_level,
FM_EREPORT_PAYLOAD_ZFS_ZIO_BLKID,
DATA_TYPE_UINT64, zb->zb_blkid, NULL);
}
/*
* Payload for tuning the zed
*/
if (vd != NULL && strcmp(subclass, FM_EREPORT_ZFS_CHECKSUM) == 0) {
uint64_t cksum_n, cksum_t;
cksum_n = vdev_prop_get_inherited(vd, VDEV_PROP_CHECKSUM_N);
if (cksum_n != vdev_prop_default_numeric(VDEV_PROP_CHECKSUM_N))
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_CKSUM_N,
DATA_TYPE_UINT64,
cksum_n,
NULL);
cksum_t = vdev_prop_get_inherited(vd, VDEV_PROP_CHECKSUM_T);
if (cksum_t != vdev_prop_default_numeric(VDEV_PROP_CHECKSUM_T))
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_CKSUM_T,
DATA_TYPE_UINT64,
cksum_t,
NULL);
}
if (vd != NULL && strcmp(subclass, FM_EREPORT_ZFS_IO) == 0) {
uint64_t io_n, io_t;
io_n = vdev_prop_get_inherited(vd, VDEV_PROP_IO_N);
if (io_n != vdev_prop_default_numeric(VDEV_PROP_IO_N))
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_IO_N,
DATA_TYPE_UINT64,
io_n,
NULL);
io_t = vdev_prop_get_inherited(vd, VDEV_PROP_IO_T);
if (io_t != vdev_prop_default_numeric(VDEV_PROP_IO_T))
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_VDEV_IO_T,
DATA_TYPE_UINT64,
io_t,
NULL);
}
mutex_exit(&spa->spa_errlist_lock);
*ereport_out = ereport;
*detector_out = detector;
return (B_TRUE);
}
/* if it's <= 128 bytes, save the corruption directly */
#define ZFM_MAX_INLINE (128 / sizeof (uint64_t))
#define MAX_RANGES 16
typedef struct zfs_ecksum_info {
/* histograms of set and cleared bits by bit number in a 64-bit word */
- uint32_t zei_histogram_set[sizeof (uint64_t) * NBBY];
- uint32_t zei_histogram_cleared[sizeof (uint64_t) * NBBY];
+ uint8_t zei_histogram_set[sizeof (uint64_t) * NBBY];
+ uint8_t zei_histogram_cleared[sizeof (uint64_t) * NBBY];
/* inline arrays of bits set and cleared. */
uint64_t zei_bits_set[ZFM_MAX_INLINE];
uint64_t zei_bits_cleared[ZFM_MAX_INLINE];
/*
* for each range, the number of bits set and cleared. The Hamming
* distance between the good and bad buffers is the sum of them all.
*/
uint32_t zei_range_sets[MAX_RANGES];
uint32_t zei_range_clears[MAX_RANGES];
struct zei_ranges {
uint32_t zr_start;
uint32_t zr_end;
} zei_ranges[MAX_RANGES];
size_t zei_range_count;
uint32_t zei_mingap;
uint32_t zei_allowed_mingap;
} zfs_ecksum_info_t;
static void
-update_histogram(uint64_t value_arg, uint32_t *hist, uint32_t *count)
+update_histogram(uint64_t value_arg, uint8_t *hist, uint32_t *count)
{
size_t i;
size_t bits = 0;
uint64_t value = BE_64(value_arg);
/* We store the bits in big-endian (largest-first) order */
for (i = 0; i < 64; i++) {
if (value & (1ull << i)) {
hist[63 - i]++;
++bits;
}
}
/* update the count of bits changed */
*count += bits;
}
/*
* We've now filled up the range array, and need to increase "mingap" and
* shrink the range list accordingly. zei_mingap is always the smallest
* distance between array entries, so we set the new_allowed_gap to be
* one greater than that. We then go through the list, joining together
* any ranges which are closer than the new_allowed_gap.
*
* By construction, there will be at least one. We also update zei_mingap
* to the new smallest gap, to prepare for our next invocation.
*/
static void
zei_shrink_ranges(zfs_ecksum_info_t *eip)
{
uint32_t mingap = UINT32_MAX;
uint32_t new_allowed_gap = eip->zei_mingap + 1;
size_t idx, output;
size_t max = eip->zei_range_count;
struct zei_ranges *r = eip->zei_ranges;
ASSERT3U(eip->zei_range_count, >, 0);
ASSERT3U(eip->zei_range_count, <=, MAX_RANGES);
output = idx = 0;
while (idx < max - 1) {
uint32_t start = r[idx].zr_start;
uint32_t end = r[idx].zr_end;
while (idx < max - 1) {
idx++;
uint32_t nstart = r[idx].zr_start;
uint32_t nend = r[idx].zr_end;
uint32_t gap = nstart - end;
if (gap < new_allowed_gap) {
end = nend;
continue;
}
if (gap < mingap)
mingap = gap;
break;
}
r[output].zr_start = start;
r[output].zr_end = end;
output++;
}
ASSERT3U(output, <, eip->zei_range_count);
eip->zei_range_count = output;
eip->zei_mingap = mingap;
eip->zei_allowed_mingap = new_allowed_gap;
}
static void
zei_add_range(zfs_ecksum_info_t *eip, int start, int end)
{
struct zei_ranges *r = eip->zei_ranges;
size_t count = eip->zei_range_count;
if (count >= MAX_RANGES) {
zei_shrink_ranges(eip);
count = eip->zei_range_count;
}
if (count == 0) {
eip->zei_mingap = UINT32_MAX;
eip->zei_allowed_mingap = 1;
} else {
int gap = start - r[count - 1].zr_end;
if (gap < eip->zei_allowed_mingap) {
r[count - 1].zr_end = end;
return;
}
if (gap < eip->zei_mingap)
eip->zei_mingap = gap;
}
r[count].zr_start = start;
r[count].zr_end = end;
eip->zei_range_count++;
}
static size_t
zei_range_total_size(zfs_ecksum_info_t *eip)
{
struct zei_ranges *r = eip->zei_ranges;
size_t count = eip->zei_range_count;
size_t result = 0;
size_t idx;
for (idx = 0; idx < count; idx++)
result += (r[idx].zr_end - r[idx].zr_start);
return (result);
}
static zfs_ecksum_info_t *
annotate_ecksum(nvlist_t *ereport, zio_bad_cksum_t *info,
const abd_t *goodabd, const abd_t *badabd, size_t size,
boolean_t drop_if_identical)
{
const uint64_t *good;
const uint64_t *bad;
size_t nui64s = size / sizeof (uint64_t);
size_t inline_size;
int no_inline = 0;
size_t idx;
size_t range;
size_t offset = 0;
ssize_t start = -1;
zfs_ecksum_info_t *eip = kmem_zalloc(sizeof (*eip), KM_SLEEP);
/* don't do any annotation for injected checksum errors */
if (info != NULL && info->zbc_injected)
return (eip);
if (info != NULL && info->zbc_has_cksum) {
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_CKSUM_EXPECTED,
DATA_TYPE_UINT64_ARRAY,
sizeof (info->zbc_expected) / sizeof (uint64_t),
(uint64_t *)&info->zbc_expected,
FM_EREPORT_PAYLOAD_ZFS_CKSUM_ACTUAL,
DATA_TYPE_UINT64_ARRAY,
sizeof (info->zbc_actual) / sizeof (uint64_t),
(uint64_t *)&info->zbc_actual,
FM_EREPORT_PAYLOAD_ZFS_CKSUM_ALGO,
DATA_TYPE_STRING,
info->zbc_checksum_name,
NULL);
if (info->zbc_byteswapped) {
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_CKSUM_BYTESWAP,
DATA_TYPE_BOOLEAN, 1,
NULL);
}
}
if (badabd == NULL || goodabd == NULL)
return (eip);
ASSERT3U(nui64s, <=, UINT32_MAX);
ASSERT3U(size, ==, nui64s * sizeof (uint64_t));
ASSERT3U(size, <=, SPA_MAXBLOCKSIZE);
ASSERT3U(size, <=, UINT32_MAX);
good = (const uint64_t *) abd_borrow_buf_copy((abd_t *)goodabd, size);
bad = (const uint64_t *) abd_borrow_buf_copy((abd_t *)badabd, size);
/* build up the range list by comparing the two buffers. */
for (idx = 0; idx < nui64s; idx++) {
if (good[idx] == bad[idx]) {
if (start == -1)
continue;
zei_add_range(eip, start, idx);
start = -1;
} else {
if (start != -1)
continue;
start = idx;
}
}
if (start != -1)
zei_add_range(eip, start, idx);
/* See if it will fit in our inline buffers */
inline_size = zei_range_total_size(eip);
if (inline_size > ZFM_MAX_INLINE)
no_inline = 1;
/*
* If there is no change and we want to drop if the buffers are
* identical, do so.
*/
if (inline_size == 0 && drop_if_identical) {
kmem_free(eip, sizeof (*eip));
abd_return_buf((abd_t *)goodabd, (void *)good, size);
abd_return_buf((abd_t *)badabd, (void *)bad, size);
return (NULL);
}
/*
* Now walk through the ranges, filling in the details of the
* differences. Also convert our uint64_t-array offsets to byte
* offsets.
*/
for (range = 0; range < eip->zei_range_count; range++) {
size_t start = eip->zei_ranges[range].zr_start;
size_t end = eip->zei_ranges[range].zr_end;
for (idx = start; idx < end; idx++) {
uint64_t set, cleared;
// bits set in bad, but not in good
set = ((~good[idx]) & bad[idx]);
// bits set in good, but not in bad
cleared = (good[idx] & (~bad[idx]));
if (!no_inline) {
ASSERT3U(offset, <, inline_size);
eip->zei_bits_set[offset] = set;
eip->zei_bits_cleared[offset] = cleared;
offset++;
}
update_histogram(set, eip->zei_histogram_set,
&eip->zei_range_sets[range]);
update_histogram(cleared, eip->zei_histogram_cleared,
&eip->zei_range_clears[range]);
}
/* convert to byte offsets */
eip->zei_ranges[range].zr_start *= sizeof (uint64_t);
eip->zei_ranges[range].zr_end *= sizeof (uint64_t);
}
abd_return_buf((abd_t *)goodabd, (void *)good, size);
abd_return_buf((abd_t *)badabd, (void *)bad, size);
eip->zei_allowed_mingap *= sizeof (uint64_t);
inline_size *= sizeof (uint64_t);
/* fill in ereport */
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_BAD_OFFSET_RANGES,
DATA_TYPE_UINT32_ARRAY, 2 * eip->zei_range_count,
(uint32_t *)eip->zei_ranges,
FM_EREPORT_PAYLOAD_ZFS_BAD_RANGE_MIN_GAP,
DATA_TYPE_UINT32, eip->zei_allowed_mingap,
FM_EREPORT_PAYLOAD_ZFS_BAD_RANGE_SETS,
DATA_TYPE_UINT32_ARRAY, eip->zei_range_count, eip->zei_range_sets,
FM_EREPORT_PAYLOAD_ZFS_BAD_RANGE_CLEARS,
DATA_TYPE_UINT32_ARRAY, eip->zei_range_count, eip->zei_range_clears,
NULL);
if (!no_inline) {
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_BAD_SET_BITS,
DATA_TYPE_UINT8_ARRAY,
inline_size, (uint8_t *)eip->zei_bits_set,
FM_EREPORT_PAYLOAD_ZFS_BAD_CLEARED_BITS,
DATA_TYPE_UINT8_ARRAY,
inline_size, (uint8_t *)eip->zei_bits_cleared,
NULL);
} else {
fm_payload_set(ereport,
FM_EREPORT_PAYLOAD_ZFS_BAD_SET_HISTOGRAM,
- DATA_TYPE_UINT32_ARRAY,
+ DATA_TYPE_UINT8_ARRAY,
NBBY * sizeof (uint64_t), eip->zei_histogram_set,
FM_EREPORT_PAYLOAD_ZFS_BAD_CLEARED_HISTOGRAM,
- DATA_TYPE_UINT32_ARRAY,
+ DATA_TYPE_UINT8_ARRAY,
NBBY * sizeof (uint64_t), eip->zei_histogram_cleared,
NULL);
}
return (eip);
}
#else
void
zfs_ereport_clear(spa_t *spa, vdev_t *vd)
{
(void) spa, (void) vd;
}
#endif
/*
* Make sure our event is still valid for the given zio/vdev/pool. For example,
* we don't want to keep logging events for a faulted or missing vdev.
*/
boolean_t
zfs_ereport_is_valid(const char *subclass, spa_t *spa, vdev_t *vd, zio_t *zio)
{
#ifdef _KERNEL
/*
* If we are doing a spa_tryimport() or in recovery mode,
* ignore errors.
*/
if (spa_load_state(spa) == SPA_LOAD_TRYIMPORT ||
spa_load_state(spa) == SPA_LOAD_RECOVER)
return (B_FALSE);
/*
* If we are in the middle of opening a pool, and the previous attempt
* failed, don't bother logging any new ereports - we're just going to
* get the same diagnosis anyway.
*/
if (spa_load_state(spa) != SPA_LOAD_NONE &&
spa->spa_last_open_failed)
return (B_FALSE);
if (zio != NULL) {
/*
* If this is not a read or write zio, ignore the error. This
* can occur if the DKIOCFLUSHWRITECACHE ioctl fails.
*/
if (zio->io_type != ZIO_TYPE_READ &&
zio->io_type != ZIO_TYPE_WRITE)
return (B_FALSE);
if (vd != NULL) {
/*
* If the vdev has already been marked as failing due
* to a failed probe, then ignore any subsequent I/O
* errors, as the DE will automatically fault the vdev
* on the first such failure. This also catches cases
* where vdev_remove_wanted is set and the device has
* not yet been asynchronously placed into the REMOVED
* state.
*/
if (zio->io_vd == vd && !vdev_accessible(vd, zio))
return (B_FALSE);
/*
* Ignore checksum errors for reads from DTL regions of
* leaf vdevs.
*/
if (zio->io_type == ZIO_TYPE_READ &&
zio->io_error == ECKSUM &&
vd->vdev_ops->vdev_op_leaf &&
vdev_dtl_contains(vd, DTL_MISSING, zio->io_txg, 1))
return (B_FALSE);
}
}
/*
* For probe failure, we want to avoid posting ereports if we've
* already removed the device in the meantime.
*/
if (vd != NULL &&
strcmp(subclass, FM_EREPORT_ZFS_PROBE_FAILURE) == 0 &&
(vd->vdev_remove_wanted || vd->vdev_state == VDEV_STATE_REMOVED))
return (B_FALSE);
/* Ignore bogus delay events (like from ioctls or unqueued IOs) */
if ((strcmp(subclass, FM_EREPORT_ZFS_DELAY) == 0) &&
(zio != NULL) && (!zio->io_timestamp)) {
return (B_FALSE);
}
#else
(void) subclass, (void) spa, (void) vd, (void) zio;
#endif
return (B_TRUE);
}
/*
* Post an ereport for the given subclass
*
* Returns
* - 0 if an event was posted
* - EINVAL if there was a problem posting event
* - EBUSY if the event was rate limited
* - EALREADY if the event was already posted (duplicate)
*/
int
zfs_ereport_post(const char *subclass, spa_t *spa, vdev_t *vd,
const zbookmark_phys_t *zb, zio_t *zio, uint64_t state)
{
int rc = 0;
#ifdef _KERNEL
nvlist_t *ereport = NULL;
nvlist_t *detector = NULL;
if (!zfs_ereport_is_valid(subclass, spa, vd, zio))
return (EINVAL);
if (zfs_ereport_is_duplicate(subclass, spa, vd, zb, zio, 0, 0))
return (SET_ERROR(EALREADY));
if (zfs_is_ratelimiting_event(subclass, vd))
return (SET_ERROR(EBUSY));
if (!zfs_ereport_start(&ereport, &detector, subclass, spa, vd,
zb, zio, state, 0))
return (SET_ERROR(EINVAL)); /* couldn't post event */
if (ereport == NULL)
return (SET_ERROR(EINVAL));
/* Cleanup is handled by the callback function */
rc = zfs_zevent_post(ereport, detector, zfs_zevent_post_cb);
#else
(void) subclass, (void) spa, (void) vd, (void) zb, (void) zio,
(void) state;
#endif
return (rc);
}
/*
* Prepare a checksum ereport
*
* Returns
* - 0 if an event was posted
* - EINVAL if there was a problem posting event
* - EBUSY if the event was rate limited
* - EALREADY if the event was already posted (duplicate)
*/
int
zfs_ereport_start_checksum(spa_t *spa, vdev_t *vd, const zbookmark_phys_t *zb,
struct zio *zio, uint64_t offset, uint64_t length, zio_bad_cksum_t *info)
{
zio_cksum_report_t *report;
#ifdef _KERNEL
if (!zfs_ereport_is_valid(FM_EREPORT_ZFS_CHECKSUM, spa, vd, zio))
return (SET_ERROR(EINVAL));
if (zfs_ereport_is_duplicate(FM_EREPORT_ZFS_CHECKSUM, spa, vd, zb, zio,
offset, length))
return (SET_ERROR(EALREADY));
if (zfs_is_ratelimiting_event(FM_EREPORT_ZFS_CHECKSUM, vd))
return (SET_ERROR(EBUSY));
#else
(void) zb, (void) offset;
#endif
report = kmem_zalloc(sizeof (*report), KM_SLEEP);
zio_vsd_default_cksum_report(zio, report);
/* copy the checksum failure information if it was provided */
if (info != NULL) {
report->zcr_ckinfo = kmem_zalloc(sizeof (*info), KM_SLEEP);
memcpy(report->zcr_ckinfo, info, sizeof (*info));
}
report->zcr_sector = 1ULL << vd->vdev_top->vdev_ashift;
report->zcr_align =
vdev_psize_to_asize(vd->vdev_top, report->zcr_sector);
report->zcr_length = length;
#ifdef _KERNEL
(void) zfs_ereport_start(&report->zcr_ereport, &report->zcr_detector,
FM_EREPORT_ZFS_CHECKSUM, spa, vd, zb, zio, offset, length);
if (report->zcr_ereport == NULL) {
zfs_ereport_free_checksum(report);
return (0);
}
#endif
mutex_enter(&spa->spa_errlist_lock);
report->zcr_next = zio->io_logical->io_cksum_report;
zio->io_logical->io_cksum_report = report;
mutex_exit(&spa->spa_errlist_lock);
return (0);
}
void
zfs_ereport_finish_checksum(zio_cksum_report_t *report, const abd_t *good_data,
const abd_t *bad_data, boolean_t drop_if_identical)
{
#ifdef _KERNEL
zfs_ecksum_info_t *info;
info = annotate_ecksum(report->zcr_ereport, report->zcr_ckinfo,
good_data, bad_data, report->zcr_length, drop_if_identical);
if (info != NULL)
zfs_zevent_post(report->zcr_ereport,
report->zcr_detector, zfs_zevent_post_cb);
else
zfs_zevent_post_cb(report->zcr_ereport, report->zcr_detector);
report->zcr_ereport = report->zcr_detector = NULL;
if (info != NULL)
kmem_free(info, sizeof (*info));
#else
(void) report, (void) good_data, (void) bad_data,
(void) drop_if_identical;
#endif
}
void
zfs_ereport_free_checksum(zio_cksum_report_t *rpt)
{
#ifdef _KERNEL
if (rpt->zcr_ereport != NULL) {
fm_nvlist_destroy(rpt->zcr_ereport,
FM_NVA_FREE);
fm_nvlist_destroy(rpt->zcr_detector,
FM_NVA_FREE);
}
#endif
rpt->zcr_free(rpt->zcr_cbdata, rpt->zcr_cbinfo);
if (rpt->zcr_ckinfo != NULL)
kmem_free(rpt->zcr_ckinfo, sizeof (*rpt->zcr_ckinfo));
kmem_free(rpt, sizeof (*rpt));
}
/*
* Post a checksum ereport
*
* Returns
* - 0 if an event was posted
* - EINVAL if there was a problem posting event
* - EBUSY if the event was rate limited
* - EALREADY if the event was already posted (duplicate)
*/
int
zfs_ereport_post_checksum(spa_t *spa, vdev_t *vd, const zbookmark_phys_t *zb,
struct zio *zio, uint64_t offset, uint64_t length,
const abd_t *good_data, const abd_t *bad_data, zio_bad_cksum_t *zbc)
{
int rc = 0;
#ifdef _KERNEL
nvlist_t *ereport = NULL;
nvlist_t *detector = NULL;
zfs_ecksum_info_t *info;
if (!zfs_ereport_is_valid(FM_EREPORT_ZFS_CHECKSUM, spa, vd, zio))
return (SET_ERROR(EINVAL));
if (zfs_ereport_is_duplicate(FM_EREPORT_ZFS_CHECKSUM, spa, vd, zb, zio,
offset, length))
return (SET_ERROR(EALREADY));
if (zfs_is_ratelimiting_event(FM_EREPORT_ZFS_CHECKSUM, vd))
return (SET_ERROR(EBUSY));
if (!zfs_ereport_start(&ereport, &detector, FM_EREPORT_ZFS_CHECKSUM,
spa, vd, zb, zio, offset, length) || (ereport == NULL)) {
return (SET_ERROR(EINVAL));
}
info = annotate_ecksum(ereport, zbc, good_data, bad_data, length,
B_FALSE);
if (info != NULL) {
rc = zfs_zevent_post(ereport, detector, zfs_zevent_post_cb);
kmem_free(info, sizeof (*info));
}
#else
(void) spa, (void) vd, (void) zb, (void) zio, (void) offset,
(void) length, (void) good_data, (void) bad_data, (void) zbc;
#endif
return (rc);
}
/*
* The 'sysevent.fs.zfs.*' events are signals posted to notify user space of
* change in the pool. All sysevents are listed in sys/sysevent/eventdefs.h
* and are designed to be consumed by the ZFS Event Daemon (ZED). For
* additional details refer to the zed(8) man page.
*/
nvlist_t *
zfs_event_create(spa_t *spa, vdev_t *vd, const char *type, const char *name,
nvlist_t *aux)
{
nvlist_t *resource = NULL;
#ifdef _KERNEL
char class[64];
if (spa_load_state(spa) == SPA_LOAD_TRYIMPORT)
return (NULL);
if ((resource = fm_nvlist_create(NULL)) == NULL)
return (NULL);
(void) snprintf(class, sizeof (class), "%s.%s.%s", type,
ZFS_ERROR_CLASS, name);
VERIFY0(nvlist_add_uint8(resource, FM_VERSION, FM_RSRC_VERSION));
VERIFY0(nvlist_add_string(resource, FM_CLASS, class));
VERIFY0(nvlist_add_string(resource,
FM_EREPORT_PAYLOAD_ZFS_POOL, spa_name(spa)));
VERIFY0(nvlist_add_uint64(resource,
FM_EREPORT_PAYLOAD_ZFS_POOL_GUID, spa_guid(spa)));
VERIFY0(nvlist_add_uint64(resource,
FM_EREPORT_PAYLOAD_ZFS_POOL_STATE, spa_state(spa)));
VERIFY0(nvlist_add_int32(resource,
FM_EREPORT_PAYLOAD_ZFS_POOL_CONTEXT, spa_load_state(spa)));
if (vd) {
VERIFY0(nvlist_add_uint64(resource,
FM_EREPORT_PAYLOAD_ZFS_VDEV_GUID, vd->vdev_guid));
VERIFY0(nvlist_add_uint64(resource,
FM_EREPORT_PAYLOAD_ZFS_VDEV_STATE, vd->vdev_state));
if (vd->vdev_path != NULL)
VERIFY0(nvlist_add_string(resource,
FM_EREPORT_PAYLOAD_ZFS_VDEV_PATH, vd->vdev_path));
if (vd->vdev_devid != NULL)
VERIFY0(nvlist_add_string(resource,
FM_EREPORT_PAYLOAD_ZFS_VDEV_DEVID, vd->vdev_devid));
if (vd->vdev_fru != NULL)
VERIFY0(nvlist_add_string(resource,
FM_EREPORT_PAYLOAD_ZFS_VDEV_FRU, vd->vdev_fru));
if (vd->vdev_enc_sysfs_path != NULL)
VERIFY0(nvlist_add_string(resource,
FM_EREPORT_PAYLOAD_ZFS_VDEV_ENC_SYSFS_PATH,
vd->vdev_enc_sysfs_path));
}
/* also copy any optional payload data */
if (aux) {
nvpair_t *elem = NULL;
while ((elem = nvlist_next_nvpair(aux, elem)) != NULL)
(void) nvlist_add_nvpair(resource, elem);
}
#else
(void) spa, (void) vd, (void) type, (void) name, (void) aux;
#endif
return (resource);
}
static void
zfs_post_common(spa_t *spa, vdev_t *vd, const char *type, const char *name,
nvlist_t *aux)
{
#ifdef _KERNEL
nvlist_t *resource;
resource = zfs_event_create(spa, vd, type, name, aux);
if (resource)
zfs_zevent_post(resource, NULL, zfs_zevent_post_cb);
#else
(void) spa, (void) vd, (void) type, (void) name, (void) aux;
#endif
}
/*
* The 'resource.fs.zfs.removed' event is an internal signal that the given vdev
* has been removed from the system. This will cause the DE to ignore any
* recent I/O errors, inferring that they are due to the asynchronous device
* removal.
*/
void
zfs_post_remove(spa_t *spa, vdev_t *vd)
{
zfs_post_common(spa, vd, FM_RSRC_CLASS, FM_RESOURCE_REMOVED, NULL);
}
/*
* The 'resource.fs.zfs.autoreplace' event is an internal signal that the pool
* has the 'autoreplace' property set, and therefore any broken vdevs will be
* handled by higher level logic, and no vdev fault should be generated.
*/
void
zfs_post_autoreplace(spa_t *spa, vdev_t *vd)
{
zfs_post_common(spa, vd, FM_RSRC_CLASS, FM_RESOURCE_AUTOREPLACE, NULL);
}
/*
* The 'resource.fs.zfs.statechange' event is an internal signal that the
* given vdev has transitioned its state to DEGRADED or HEALTHY. This will
* cause the retire agent to repair any outstanding fault management cases
* open because the device was not found (fault.fs.zfs.device).
*/
void
zfs_post_state_change(spa_t *spa, vdev_t *vd, uint64_t laststate)
{
#ifdef _KERNEL
nvlist_t *aux;
/*
* Add optional supplemental keys to payload
*/
aux = fm_nvlist_create(NULL);
if (vd && aux) {
if (vd->vdev_physpath) {
fnvlist_add_string(aux,
FM_EREPORT_PAYLOAD_ZFS_VDEV_PHYSPATH,
vd->vdev_physpath);
}
if (vd->vdev_enc_sysfs_path) {
fnvlist_add_string(aux,
FM_EREPORT_PAYLOAD_ZFS_VDEV_ENC_SYSFS_PATH,
vd->vdev_enc_sysfs_path);
}
fnvlist_add_uint64(aux,
FM_EREPORT_PAYLOAD_ZFS_VDEV_LASTSTATE, laststate);
}
zfs_post_common(spa, vd, FM_RSRC_CLASS, FM_RESOURCE_STATECHANGE,
aux);
if (aux)
fm_nvlist_destroy(aux, FM_NVA_FREE);
#else
(void) spa, (void) vd, (void) laststate;
#endif
}
#ifdef _KERNEL
void
zfs_ereport_init(void)
{
mutex_init(&recent_events_lock, NULL, MUTEX_DEFAULT, NULL);
list_create(&recent_events_list, sizeof (recent_events_node_t),
offsetof(recent_events_node_t, re_list_link));
avl_create(&recent_events_tree, recent_events_compare,
sizeof (recent_events_node_t), offsetof(recent_events_node_t,
re_tree_link));
}
/*
* This 'early' fini needs to run before zfs_fini() which on Linux waits
* for the system_delay_taskq to drain.
*/
void
zfs_ereport_taskq_fini(void)
{
mutex_enter(&recent_events_lock);
if (recent_events_cleaner_tqid != 0) {
taskq_cancel_id(system_delay_taskq, recent_events_cleaner_tqid);
recent_events_cleaner_tqid = 0;
}
mutex_exit(&recent_events_lock);
}
void
zfs_ereport_fini(void)
{
recent_events_node_t *entry;
while ((entry = list_head(&recent_events_list)) != NULL) {
avl_remove(&recent_events_tree, entry);
list_remove(&recent_events_list, entry);
kmem_free(entry, sizeof (*entry));
}
avl_destroy(&recent_events_tree);
list_destroy(&recent_events_list);
mutex_destroy(&recent_events_lock);
}
void
zfs_ereport_snapshot_post(const char *subclass, spa_t *spa, const char *name)
{
nvlist_t *aux;
aux = fm_nvlist_create(NULL);
fnvlist_add_string(aux, FM_EREPORT_PAYLOAD_ZFS_SNAPSHOT_NAME, name);
zfs_post_common(spa, NULL, FM_RSRC_CLASS, subclass, aux);
fm_nvlist_destroy(aux, FM_NVA_FREE);
}
/*
* Post when a event when a zvol is created or removed
*
* This is currently only used by macOS, since it uses the event to create
* symlinks between the volume name (mypool/myvol) and the actual /dev
* device (/dev/disk3). For example:
*
* /var/run/zfs/dsk/mypool/myvol -> /dev/disk3
*
* name: The full name of the zvol ("mypool/myvol")
* dev_name: The full /dev name for the zvol ("/dev/disk3")
* raw_name: The raw /dev name for the zvol ("/dev/rdisk3")
*/
void
zfs_ereport_zvol_post(const char *subclass, const char *name,
const char *dev_name, const char *raw_name)
{
nvlist_t *aux;
char *r;
boolean_t locked = mutex_owned(&spa_namespace_lock);
if (!locked) mutex_enter(&spa_namespace_lock);
spa_t *spa = spa_lookup(name);
if (!locked) mutex_exit(&spa_namespace_lock);
if (spa == NULL)
return;
aux = fm_nvlist_create(NULL);
fnvlist_add_string(aux, FM_EREPORT_PAYLOAD_ZFS_DEVICE_NAME, dev_name);
fnvlist_add_string(aux, FM_EREPORT_PAYLOAD_ZFS_RAW_DEVICE_NAME,
raw_name);
r = strchr(name, '/');
if (r && r[1])
fnvlist_add_string(aux, FM_EREPORT_PAYLOAD_ZFS_VOLUME, &r[1]);
zfs_post_common(spa, NULL, FM_RSRC_CLASS, subclass, aux);
fm_nvlist_destroy(aux, FM_NVA_FREE);
}
EXPORT_SYMBOL(zfs_ereport_post);
EXPORT_SYMBOL(zfs_ereport_is_valid);
EXPORT_SYMBOL(zfs_ereport_post_checksum);
EXPORT_SYMBOL(zfs_post_remove);
EXPORT_SYMBOL(zfs_post_autoreplace);
EXPORT_SYMBOL(zfs_post_state_change);
ZFS_MODULE_PARAM(zfs_zevent, zfs_zevent_, retain_max, UINT, ZMOD_RW,
"Maximum recent zevents records to retain for duplicate checking");
ZFS_MODULE_PARAM(zfs_zevent, zfs_zevent_, retain_expire_secs, UINT, ZMOD_RW,
"Expiration time for recent zevents records");
#endif /* _KERNEL */
diff --git a/sys/contrib/openzfs/module/zfs/zfs_replay.c b/sys/contrib/openzfs/module/zfs/zfs_replay.c
index 04dfda56b3f1..09c7be853bf9 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_replay.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_replay.c
@@ -1,1223 +1,1223 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012 Cyril Plisko. All rights reserved.
* Copyright (c) 2013, 2017 by Delphix. All rights reserved.
* Copyright (c) 2021, 2022 by Pawel Jakub Dawidek
*/
#include <sys/types.h>
#include <sys/param.h>
#include <sys/sysmacros.h>
#include <sys/cmn_err.h>
#include <sys/kmem.h>
#include <sys/thread.h>
#include <sys/file.h>
#include <sys/fcntl.h>
#include <sys/vfs.h>
#include <sys/fs/zfs.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_fuid.h>
#include <sys/zfs_vnops.h>
#include <sys/spa.h>
#include <sys/zil.h>
#include <sys/byteorder.h>
#include <sys/stat.h>
#include <sys/acl.h>
#include <sys/atomic.h>
#include <sys/cred.h>
#include <sys/zpl.h>
#include <sys/dmu_objset.h>
#include <sys/zfeature.h>
/*
* NB: FreeBSD expects to be able to do vnode locking in lookup and
* hold the locks across all subsequent VOPs until vput is called.
* This means that its zfs vnops routines can't do any internal locking.
* In order to have the same contract as the Linux vnops there would
* needed to be duplicate locked vnops. If the vnops were used more widely
* in common code this would likely be preferable. However, currently
* this is the only file where this is the case.
*/
/*
* Functions to replay ZFS intent log (ZIL) records
* The functions are called through a function vector (zfs_replay_vector)
* which is indexed by the transaction type.
*/
static void
zfs_init_vattr(vattr_t *vap, uint64_t mask, uint64_t mode,
uint64_t uid, uint64_t gid, uint64_t rdev, uint64_t nodeid)
{
memset(vap, 0, sizeof (*vap));
vap->va_mask = (uint_t)mask;
vap->va_mode = mode;
#if defined(__FreeBSD__) || defined(__APPLE__)
vap->va_type = IFTOVT(mode);
#endif
vap->va_uid = (uid_t)(IS_EPHEMERAL(uid)) ? -1 : uid;
vap->va_gid = (gid_t)(IS_EPHEMERAL(gid)) ? -1 : gid;
vap->va_rdev = zfs_cmpldev(rdev);
vap->va_nodeid = nodeid;
}
static int
zfs_replay_error(void *arg1, void *arg2, boolean_t byteswap)
{
(void) arg1, (void) arg2, (void) byteswap;
return (SET_ERROR(ENOTSUP));
}
static void
zfs_replay_xvattr(lr_attr_t *lrattr, xvattr_t *xvap)
{
xoptattr_t *xoap = NULL;
uint64_t *attrs;
uint64_t *crtime;
uint32_t *bitmap;
void *scanstamp;
int i;
xvap->xva_vattr.va_mask |= ATTR_XVATTR;
if ((xoap = xva_getxoptattr(xvap)) == NULL) {
xvap->xva_vattr.va_mask &= ~ATTR_XVATTR; /* shouldn't happen */
return;
}
ASSERT(lrattr->lr_attr_masksize == xvap->xva_mapsize);
bitmap = &lrattr->lr_attr_bitmap;
for (i = 0; i != lrattr->lr_attr_masksize; i++, bitmap++)
xvap->xva_reqattrmap[i] = *bitmap;
attrs = (uint64_t *)(lrattr + lrattr->lr_attr_masksize - 1);
crtime = attrs + 1;
scanstamp = (caddr_t)(crtime + 2);
if (XVA_ISSET_REQ(xvap, XAT_HIDDEN))
xoap->xoa_hidden = ((*attrs & XAT0_HIDDEN) != 0);
if (XVA_ISSET_REQ(xvap, XAT_SYSTEM))
xoap->xoa_system = ((*attrs & XAT0_SYSTEM) != 0);
if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE))
xoap->xoa_archive = ((*attrs & XAT0_ARCHIVE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_READONLY))
xoap->xoa_readonly = ((*attrs & XAT0_READONLY) != 0);
if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE))
xoap->xoa_immutable = ((*attrs & XAT0_IMMUTABLE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK))
xoap->xoa_nounlink = ((*attrs & XAT0_NOUNLINK) != 0);
if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY))
xoap->xoa_appendonly = ((*attrs & XAT0_APPENDONLY) != 0);
if (XVA_ISSET_REQ(xvap, XAT_NODUMP))
xoap->xoa_nodump = ((*attrs & XAT0_NODUMP) != 0);
if (XVA_ISSET_REQ(xvap, XAT_OPAQUE))
xoap->xoa_opaque = ((*attrs & XAT0_OPAQUE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED))
xoap->xoa_av_modified = ((*attrs & XAT0_AV_MODIFIED) != 0);
if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED))
xoap->xoa_av_quarantined =
((*attrs & XAT0_AV_QUARANTINED) != 0);
if (XVA_ISSET_REQ(xvap, XAT_CREATETIME))
ZFS_TIME_DECODE(&xoap->xoa_createtime, crtime);
if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) {
ASSERT(!XVA_ISSET_REQ(xvap, XAT_PROJID));
memcpy(xoap->xoa_av_scanstamp, scanstamp, AV_SCANSTAMP_SZ);
} else if (XVA_ISSET_REQ(xvap, XAT_PROJID)) {
/*
* XAT_PROJID and XAT_AV_SCANSTAMP will never be valid
* at the same time, so we can share the same space.
*/
memcpy(&xoap->xoa_projid, scanstamp, sizeof (uint64_t));
}
if (XVA_ISSET_REQ(xvap, XAT_REPARSE))
xoap->xoa_reparse = ((*attrs & XAT0_REPARSE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_OFFLINE))
xoap->xoa_offline = ((*attrs & XAT0_OFFLINE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_SPARSE))
xoap->xoa_sparse = ((*attrs & XAT0_SPARSE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT))
xoap->xoa_projinherit = ((*attrs & XAT0_PROJINHERIT) != 0);
}
static int
zfs_replay_domain_cnt(uint64_t uid, uint64_t gid)
{
uint64_t uid_idx;
uint64_t gid_idx;
int domcnt = 0;
uid_idx = FUID_INDEX(uid);
gid_idx = FUID_INDEX(gid);
if (uid_idx)
domcnt++;
if (gid_idx > 0 && gid_idx != uid_idx)
domcnt++;
return (domcnt);
}
static void *
zfs_replay_fuid_domain_common(zfs_fuid_info_t *fuid_infop, void *start,
int domcnt)
{
int i;
for (i = 0; i != domcnt; i++) {
fuid_infop->z_domain_table[i] = start;
start = (caddr_t)start + strlen(start) + 1;
}
return (start);
}
/*
* Set the uid/gid in the fuid_info structure.
*/
static void
zfs_replay_fuid_ugid(zfs_fuid_info_t *fuid_infop, uint64_t uid, uint64_t gid)
{
/*
* If owner or group are log specific FUIDs then slurp up
* domain information and build zfs_fuid_info_t
*/
if (IS_EPHEMERAL(uid))
fuid_infop->z_fuid_owner = uid;
if (IS_EPHEMERAL(gid))
fuid_infop->z_fuid_group = gid;
}
/*
* Load fuid domains into fuid_info_t
*/
static zfs_fuid_info_t *
zfs_replay_fuid_domain(void *buf, void **end, uint64_t uid, uint64_t gid)
{
int domcnt;
zfs_fuid_info_t *fuid_infop;
fuid_infop = zfs_fuid_info_alloc();
domcnt = zfs_replay_domain_cnt(uid, gid);
if (domcnt == 0)
return (fuid_infop);
fuid_infop->z_domain_table =
kmem_zalloc(domcnt * sizeof (char *), KM_SLEEP);
zfs_replay_fuid_ugid(fuid_infop, uid, gid);
fuid_infop->z_domain_cnt = domcnt;
*end = zfs_replay_fuid_domain_common(fuid_infop, buf, domcnt);
return (fuid_infop);
}
/*
* load zfs_fuid_t's and fuid_domains into fuid_info_t
*/
static zfs_fuid_info_t *
zfs_replay_fuids(void *start, void **end, int idcnt, int domcnt, uint64_t uid,
uint64_t gid)
{
uint64_t *log_fuid = (uint64_t *)start;
zfs_fuid_info_t *fuid_infop;
int i;
fuid_infop = zfs_fuid_info_alloc();
fuid_infop->z_domain_cnt = domcnt;
fuid_infop->z_domain_table =
kmem_zalloc(domcnt * sizeof (char *), KM_SLEEP);
for (i = 0; i != idcnt; i++) {
zfs_fuid_t *zfuid;
zfuid = kmem_alloc(sizeof (zfs_fuid_t), KM_SLEEP);
zfuid->z_logfuid = *log_fuid;
zfuid->z_id = -1;
zfuid->z_domidx = 0;
list_insert_tail(&fuid_infop->z_fuids, zfuid);
log_fuid++;
}
zfs_replay_fuid_ugid(fuid_infop, uid, gid);
*end = zfs_replay_fuid_domain_common(fuid_infop, log_fuid, domcnt);
return (fuid_infop);
}
static void
zfs_replay_swap_attrs(lr_attr_t *lrattr)
{
/* swap the lr_attr structure */
byteswap_uint32_array(lrattr, sizeof (*lrattr));
/* swap the bitmap */
byteswap_uint32_array(lrattr + 1, (lrattr->lr_attr_masksize - 1) *
sizeof (uint32_t));
/* swap the attributes, create time + 64 bit word for attributes */
byteswap_uint64_array((caddr_t)(lrattr + 1) + (sizeof (uint32_t) *
(lrattr->lr_attr_masksize - 1)), 3 * sizeof (uint64_t));
}
/*
* Replay file create with optional ACL, xvattr information as well
* as option FUID information.
*/
static int
zfs_replay_create_acl(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_acl_create_t *lracl = arg2;
char *name = NULL; /* location determined later */
lr_create_t *lr = (lr_create_t *)lracl;
znode_t *dzp;
znode_t *zp;
xvattr_t xva;
int vflg = 0;
vsecattr_t vsec = { 0 };
lr_attr_t *lrattr;
void *aclstart;
void *fuidstart;
size_t xvatlen = 0;
uint64_t txtype;
uint64_t objid;
uint64_t dnodesize;
int error;
txtype = (lr->lr_common.lrc_txtype & ~TX_CI);
if (byteswap) {
byteswap_uint64_array(lracl, sizeof (*lracl));
if (txtype == TX_CREATE_ACL_ATTR ||
txtype == TX_MKDIR_ACL_ATTR) {
lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
zfs_replay_swap_attrs(lrattr);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
}
aclstart = (caddr_t)(lracl + 1) + xvatlen;
zfs_ace_byteswap(aclstart, lracl->lr_acl_bytes, B_FALSE);
/* swap fuids */
if (lracl->lr_fuidcnt) {
byteswap_uint64_array((caddr_t)aclstart +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes),
lracl->lr_fuidcnt * sizeof (uint64_t));
}
}
if ((error = zfs_zget(zfsvfs, lr->lr_doid, &dzp)) != 0)
return (error);
objid = LR_FOID_GET_OBJ(lr->lr_foid);
dnodesize = LR_FOID_GET_SLOTS(lr->lr_foid) << DNODE_SHIFT;
xva_init(&xva);
zfs_init_vattr(&xva.xva_vattr, ATTR_MODE | ATTR_UID | ATTR_GID,
lr->lr_mode, lr->lr_uid, lr->lr_gid, lr->lr_rdev, objid);
/*
* All forms of zfs create (create, mkdir, mkxattrdir, symlink)
* eventually end up in zfs_mknode(), which assigns the object's
* creation time, generation number, and dnode size. The generic
* zfs_create() has no concept of these attributes, so we smuggle
* the values inside the vattr's otherwise unused va_ctime,
* va_nblocks, and va_fsid fields.
*/
ZFS_TIME_DECODE(&xva.xva_vattr.va_ctime, lr->lr_crtime);
xva.xva_vattr.va_nblocks = lr->lr_gen;
xva.xva_vattr.va_fsid = dnodesize;
error = dnode_try_claim(zfsvfs->z_os, objid, dnodesize >> DNODE_SHIFT);
if (error)
goto bail;
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
switch (txtype) {
case TX_CREATE_ACL:
aclstart = (caddr_t)(lracl + 1);
fuidstart = (caddr_t)aclstart +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
zfsvfs->z_fuid_replay = zfs_replay_fuids(fuidstart,
(void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
lr->lr_uid, lr->lr_gid);
zfs_fallthrough;
case TX_CREATE_ACL_ATTR:
if (name == NULL) {
lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
xva.xva_vattr.va_mask |= ATTR_XVATTR;
zfs_replay_xvattr(lrattr, &xva);
}
vsec.vsa_mask = VSA_ACE | VSA_ACE_ACLFLAGS;
vsec.vsa_aclentp = (caddr_t)(lracl + 1) + xvatlen;
vsec.vsa_aclcnt = lracl->lr_aclcnt;
vsec.vsa_aclentsz = lracl->lr_acl_bytes;
vsec.vsa_aclflags = lracl->lr_acl_flags;
if (zfsvfs->z_fuid_replay == NULL) {
fuidstart = (caddr_t)(lracl + 1) + xvatlen +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
zfsvfs->z_fuid_replay =
zfs_replay_fuids(fuidstart,
(void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
lr->lr_uid, lr->lr_gid);
}
#if defined(__linux__)
error = zfs_create(dzp, name, &xva.xva_vattr,
- 0, 0, &zp, kcred, vflg, &vsec, kcred->user_ns);
+ 0, 0, &zp, kcred, vflg, &vsec, zfs_init_idmap);
#else
error = zfs_create(dzp, name, &xva.xva_vattr,
0, 0, &zp, kcred, vflg, &vsec, NULL);
#endif
break;
case TX_MKDIR_ACL:
aclstart = (caddr_t)(lracl + 1);
fuidstart = (caddr_t)aclstart +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
zfsvfs->z_fuid_replay = zfs_replay_fuids(fuidstart,
(void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
lr->lr_uid, lr->lr_gid);
zfs_fallthrough;
case TX_MKDIR_ACL_ATTR:
if (name == NULL) {
lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
zfs_replay_xvattr(lrattr, &xva);
}
vsec.vsa_mask = VSA_ACE | VSA_ACE_ACLFLAGS;
vsec.vsa_aclentp = (caddr_t)(lracl + 1) + xvatlen;
vsec.vsa_aclcnt = lracl->lr_aclcnt;
vsec.vsa_aclentsz = lracl->lr_acl_bytes;
vsec.vsa_aclflags = lracl->lr_acl_flags;
if (zfsvfs->z_fuid_replay == NULL) {
fuidstart = (caddr_t)(lracl + 1) + xvatlen +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
zfsvfs->z_fuid_replay =
zfs_replay_fuids(fuidstart,
(void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
lr->lr_uid, lr->lr_gid);
}
#if defined(__linux__)
error = zfs_mkdir(dzp, name, &xva.xva_vattr,
- &zp, kcred, vflg, &vsec, kcred->user_ns);
+ &zp, kcred, vflg, &vsec, zfs_init_idmap);
#else
error = zfs_mkdir(dzp, name, &xva.xva_vattr,
&zp, kcred, vflg, &vsec, NULL);
#endif
break;
default:
error = SET_ERROR(ENOTSUP);
}
bail:
if (error == 0 && zp != NULL) {
#ifdef __FreeBSD__
VOP_UNLOCK1(ZTOV(zp));
#endif
zrele(zp);
}
zrele(dzp);
if (zfsvfs->z_fuid_replay)
zfs_fuid_info_free(zfsvfs->z_fuid_replay);
zfsvfs->z_fuid_replay = NULL;
return (error);
}
static int
zfs_replay_create(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_create_t *lr = arg2;
char *name = NULL; /* location determined later */
char *link; /* symlink content follows name */
znode_t *dzp;
znode_t *zp = NULL;
xvattr_t xva;
int vflg = 0;
size_t lrsize = sizeof (lr_create_t);
lr_attr_t *lrattr;
void *start;
size_t xvatlen;
uint64_t txtype;
uint64_t objid;
uint64_t dnodesize;
int error;
txtype = (lr->lr_common.lrc_txtype & ~TX_CI);
if (byteswap) {
byteswap_uint64_array(lr, sizeof (*lr));
if (txtype == TX_CREATE_ATTR || txtype == TX_MKDIR_ATTR)
zfs_replay_swap_attrs((lr_attr_t *)(lr + 1));
}
if ((error = zfs_zget(zfsvfs, lr->lr_doid, &dzp)) != 0)
return (error);
objid = LR_FOID_GET_OBJ(lr->lr_foid);
dnodesize = LR_FOID_GET_SLOTS(lr->lr_foid) << DNODE_SHIFT;
xva_init(&xva);
zfs_init_vattr(&xva.xva_vattr, ATTR_MODE | ATTR_UID | ATTR_GID,
lr->lr_mode, lr->lr_uid, lr->lr_gid, lr->lr_rdev, objid);
/*
* All forms of zfs create (create, mkdir, mkxattrdir, symlink)
* eventually end up in zfs_mknode(), which assigns the object's
* creation time, generation number, and dnode slot count. The
* generic zfs_create() has no concept of these attributes, so
* we smuggle the values inside the vattr's otherwise unused
* va_ctime, va_nblocks, and va_fsid fields.
*/
ZFS_TIME_DECODE(&xva.xva_vattr.va_ctime, lr->lr_crtime);
xva.xva_vattr.va_nblocks = lr->lr_gen;
xva.xva_vattr.va_fsid = dnodesize;
error = dnode_try_claim(zfsvfs->z_os, objid, dnodesize >> DNODE_SHIFT);
if (error)
goto out;
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
/*
* Symlinks don't have fuid info, and CIFS never creates
* symlinks.
*
* The _ATTR versions will grab the fuid info in their subcases.
*/
if (txtype != TX_SYMLINK &&
txtype != TX_MKDIR_ATTR &&
txtype != TX_CREATE_ATTR) {
start = (lr + 1);
zfsvfs->z_fuid_replay =
zfs_replay_fuid_domain(start, &start,
lr->lr_uid, lr->lr_gid);
}
switch (txtype) {
case TX_CREATE_ATTR:
lrattr = (lr_attr_t *)(caddr_t)(lr + 1);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
zfs_replay_xvattr((lr_attr_t *)((caddr_t)lr + lrsize), &xva);
start = (caddr_t)(lr + 1) + xvatlen;
zfsvfs->z_fuid_replay =
zfs_replay_fuid_domain(start, &start,
lr->lr_uid, lr->lr_gid);
name = (char *)start;
zfs_fallthrough;
case TX_CREATE:
if (name == NULL)
name = (char *)start;
#if defined(__linux__)
error = zfs_create(dzp, name, &xva.xva_vattr,
- 0, 0, &zp, kcred, vflg, NULL, kcred->user_ns);
+ 0, 0, &zp, kcred, vflg, NULL, zfs_init_idmap);
#else
error = zfs_create(dzp, name, &xva.xva_vattr,
0, 0, &zp, kcred, vflg, NULL, NULL);
#endif
break;
case TX_MKDIR_ATTR:
lrattr = (lr_attr_t *)(caddr_t)(lr + 1);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
zfs_replay_xvattr((lr_attr_t *)((caddr_t)lr + lrsize), &xva);
start = (caddr_t)(lr + 1) + xvatlen;
zfsvfs->z_fuid_replay =
zfs_replay_fuid_domain(start, &start,
lr->lr_uid, lr->lr_gid);
name = (char *)start;
zfs_fallthrough;
case TX_MKDIR:
if (name == NULL)
name = (char *)(lr + 1);
#if defined(__linux__)
error = zfs_mkdir(dzp, name, &xva.xva_vattr,
- &zp, kcred, vflg, NULL, kcred->user_ns);
+ &zp, kcred, vflg, NULL, zfs_init_idmap);
#else
error = zfs_mkdir(dzp, name, &xva.xva_vattr,
&zp, kcred, vflg, NULL, NULL);
#endif
break;
case TX_MKXATTR:
error = zfs_make_xattrdir(dzp, &xva.xva_vattr, &zp, kcred);
break;
case TX_SYMLINK:
name = (char *)(lr + 1);
link = name + strlen(name) + 1;
#if defined(__linux__)
error = zfs_symlink(dzp, name, &xva.xva_vattr,
- link, &zp, kcred, vflg, kcred->user_ns);
+ link, &zp, kcred, vflg, zfs_init_idmap);
#else
error = zfs_symlink(dzp, name, &xva.xva_vattr,
link, &zp, kcred, vflg, NULL);
#endif
break;
default:
error = SET_ERROR(ENOTSUP);
}
out:
if (error == 0 && zp != NULL) {
#ifdef __FreeBSD__
VOP_UNLOCK1(ZTOV(zp));
#endif
zrele(zp);
}
zrele(dzp);
if (zfsvfs->z_fuid_replay)
zfs_fuid_info_free(zfsvfs->z_fuid_replay);
zfsvfs->z_fuid_replay = NULL;
return (error);
}
static int
zfs_replay_remove(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_remove_t *lr = arg2;
char *name = (char *)(lr + 1); /* name follows lr_remove_t */
znode_t *dzp;
int error;
int vflg = 0;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
if ((error = zfs_zget(zfsvfs, lr->lr_doid, &dzp)) != 0)
return (error);
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
switch ((int)lr->lr_common.lrc_txtype) {
case TX_REMOVE:
error = zfs_remove(dzp, name, kcred, vflg);
break;
case TX_RMDIR:
error = zfs_rmdir(dzp, name, NULL, kcred, vflg);
break;
default:
error = SET_ERROR(ENOTSUP);
}
zrele(dzp);
return (error);
}
static int
zfs_replay_link(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_link_t *lr = arg2;
char *name = (char *)(lr + 1); /* name follows lr_link_t */
znode_t *dzp, *zp;
int error;
int vflg = 0;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
if ((error = zfs_zget(zfsvfs, lr->lr_doid, &dzp)) != 0)
return (error);
if ((error = zfs_zget(zfsvfs, lr->lr_link_obj, &zp)) != 0) {
zrele(dzp);
return (error);
}
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
error = zfs_link(dzp, zp, name, kcred, vflg);
zrele(zp);
zrele(dzp);
return (error);
}
static int
do_zfs_replay_rename(zfsvfs_t *zfsvfs, lr_rename_t *lr, char *sname,
char *tname, uint64_t rflags, vattr_t *wo_vap)
{
znode_t *sdzp, *tdzp;
int error, vflg = 0;
/* Only Linux currently supports RENAME_* flags. */
#ifdef __linux__
VERIFY0(rflags & ~(RENAME_EXCHANGE | RENAME_WHITEOUT));
/* wo_vap must be non-NULL iff. we're doing RENAME_WHITEOUT */
VERIFY_EQUIV(rflags & RENAME_WHITEOUT, wo_vap != NULL);
#else
VERIFY0(rflags);
#endif
if ((error = zfs_zget(zfsvfs, lr->lr_sdoid, &sdzp)) != 0)
return (error);
if ((error = zfs_zget(zfsvfs, lr->lr_tdoid, &tdzp)) != 0) {
zrele(sdzp);
return (error);
}
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
#if defined(__linux__)
error = zfs_rename(sdzp, sname, tdzp, tname, kcred, vflg, rflags,
- wo_vap, kcred->user_ns);
+ wo_vap, zfs_init_idmap);
#else
error = zfs_rename(sdzp, sname, tdzp, tname, kcred, vflg, rflags,
wo_vap, NULL);
#endif
zrele(tdzp);
zrele(sdzp);
return (error);
}
static int
zfs_replay_rename(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_rename_t *lr = arg2;
char *sname = (char *)(lr + 1); /* sname and tname follow lr_rename_t */
char *tname = sname + strlen(sname) + 1;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
return (do_zfs_replay_rename(zfsvfs, lr, sname, tname, 0, NULL));
}
static int
zfs_replay_rename_exchange(void *arg1, void *arg2, boolean_t byteswap)
{
#ifdef __linux__
zfsvfs_t *zfsvfs = arg1;
lr_rename_t *lr = arg2;
char *sname = (char *)(lr + 1); /* sname and tname follow lr_rename_t */
char *tname = sname + strlen(sname) + 1;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
return (do_zfs_replay_rename(zfsvfs, lr, sname, tname, RENAME_EXCHANGE,
NULL));
#else
return (SET_ERROR(ENOTSUP));
#endif
}
static int
zfs_replay_rename_whiteout(void *arg1, void *arg2, boolean_t byteswap)
{
#ifdef __linux__
zfsvfs_t *zfsvfs = arg1;
lr_rename_whiteout_t *lr = arg2;
int error;
/* sname and tname follow lr_rename_whiteout_t */
char *sname = (char *)(lr + 1);
char *tname = sname + strlen(sname) + 1;
/* For the whiteout file. */
xvattr_t xva;
uint64_t objid;
uint64_t dnodesize;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
objid = LR_FOID_GET_OBJ(lr->lr_wfoid);
dnodesize = LR_FOID_GET_SLOTS(lr->lr_wfoid) << DNODE_SHIFT;
xva_init(&xva);
zfs_init_vattr(&xva.xva_vattr, ATTR_MODE | ATTR_UID | ATTR_GID,
lr->lr_wmode, lr->lr_wuid, lr->lr_wgid, lr->lr_wrdev, objid);
/*
* As with TX_CREATE, RENAME_WHITEOUT ends up in zfs_mknode(), which
* assigns the object's creation time, generation number, and dnode
* slot count. The generic zfs_rename() has no concept of these
* attributes, so we smuggle the values inside the vattr's otherwise
* unused va_ctime, va_nblocks, and va_fsid fields.
*/
ZFS_TIME_DECODE(&xva.xva_vattr.va_ctime, lr->lr_wcrtime);
xva.xva_vattr.va_nblocks = lr->lr_wgen;
xva.xva_vattr.va_fsid = dnodesize;
error = dnode_try_claim(zfsvfs->z_os, objid, dnodesize >> DNODE_SHIFT);
if (error)
return (error);
return (do_zfs_replay_rename(zfsvfs, &lr->lr_rename, sname, tname,
RENAME_WHITEOUT, &xva.xva_vattr));
#else
return (SET_ERROR(ENOTSUP));
#endif
}
static int
zfs_replay_write(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_write_t *lr = arg2;
char *data = (char *)(lr + 1); /* data follows lr_write_t */
znode_t *zp;
int error;
uint64_t eod, offset, length;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0) {
/*
* As we can log writes out of order, it's possible the
* file has been removed. In this case just drop the write
* and return success.
*/
if (error == ENOENT)
error = 0;
return (error);
}
offset = lr->lr_offset;
length = lr->lr_length;
eod = offset + length; /* end of data for this write */
/*
* This may be a write from a dmu_sync() for a whole block,
* and may extend beyond the current end of the file.
* We can't just replay what was written for this TX_WRITE as
* a future TX_WRITE2 may extend the eof and the data for that
* write needs to be there. So we write the whole block and
* reduce the eof. This needs to be done within the single dmu
* transaction created within vn_rdwr -> zfs_write. So a possible
* new end of file is passed through in zfsvfs->z_replay_eof
*/
zfsvfs->z_replay_eof = 0; /* 0 means don't change end of file */
/* If it's a dmu_sync() block, write the whole block */
if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) {
uint64_t blocksize = BP_GET_LSIZE(&lr->lr_blkptr);
if (length < blocksize) {
offset -= offset % blocksize;
length = blocksize;
}
if (zp->z_size < eod)
zfsvfs->z_replay_eof = eod;
}
error = zfs_write_simple(zp, data, length, offset, NULL);
zrele(zp);
zfsvfs->z_replay_eof = 0; /* safety */
return (error);
}
/*
* TX_WRITE2 are only generated when dmu_sync() returns EALREADY
* meaning the pool block is already being synced. So now that we always write
* out full blocks, all we have to do is expand the eof if
* the file is grown.
*/
static int
zfs_replay_write2(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_write_t *lr = arg2;
znode_t *zp;
int error;
uint64_t end;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
return (error);
top:
end = lr->lr_offset + lr->lr_length;
if (end > zp->z_size) {
dmu_tx_t *tx = dmu_tx_create(zfsvfs->z_os);
zp->z_size = end;
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
zrele(zp);
if (error == ERESTART) {
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
dmu_tx_abort(tx);
return (error);
}
(void) sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zfsvfs),
(void *)&zp->z_size, sizeof (uint64_t), tx);
/* Ensure the replayed seq is updated */
(void) zil_replaying(zfsvfs->z_log, tx);
dmu_tx_commit(tx);
}
zrele(zp);
return (error);
}
static int
zfs_replay_truncate(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_truncate_t *lr = arg2;
znode_t *zp;
flock64_t fl = {0};
int error;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
return (error);
fl.l_type = F_WRLCK;
fl.l_whence = SEEK_SET;
fl.l_start = lr->lr_offset;
fl.l_len = lr->lr_length;
error = zfs_space(zp, F_FREESP, &fl, O_RDWR | O_LARGEFILE,
lr->lr_offset, kcred);
zrele(zp);
return (error);
}
static int
zfs_replay_setattr(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_setattr_t *lr = arg2;
znode_t *zp;
xvattr_t xva;
vattr_t *vap = &xva.xva_vattr;
int error;
void *start;
xva_init(&xva);
if (byteswap) {
byteswap_uint64_array(lr, sizeof (*lr));
if ((lr->lr_mask & ATTR_XVATTR) &&
zfsvfs->z_version >= ZPL_VERSION_INITIAL)
zfs_replay_swap_attrs((lr_attr_t *)(lr + 1));
}
if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
return (error);
zfs_init_vattr(vap, lr->lr_mask, lr->lr_mode,
lr->lr_uid, lr->lr_gid, 0, lr->lr_foid);
vap->va_size = lr->lr_size;
ZFS_TIME_DECODE(&vap->va_atime, lr->lr_atime);
ZFS_TIME_DECODE(&vap->va_mtime, lr->lr_mtime);
gethrestime(&vap->va_ctime);
vap->va_mask |= ATTR_CTIME;
/*
* Fill in xvattr_t portions if necessary.
*/
start = (lr_setattr_t *)(lr + 1);
if (vap->va_mask & ATTR_XVATTR) {
zfs_replay_xvattr((lr_attr_t *)start, &xva);
start = (caddr_t)start +
ZIL_XVAT_SIZE(((lr_attr_t *)start)->lr_attr_masksize);
} else
xva.xva_vattr.va_mask &= ~ATTR_XVATTR;
zfsvfs->z_fuid_replay = zfs_replay_fuid_domain(start, &start,
lr->lr_uid, lr->lr_gid);
#if defined(__linux__)
- error = zfs_setattr(zp, vap, 0, kcred, kcred->user_ns);
+ error = zfs_setattr(zp, vap, 0, kcred, zfs_init_idmap);
#else
error = zfs_setattr(zp, vap, 0, kcred, NULL);
#endif
zfs_fuid_info_free(zfsvfs->z_fuid_replay);
zfsvfs->z_fuid_replay = NULL;
zrele(zp);
return (error);
}
static int
zfs_replay_setsaxattr(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_setsaxattr_t *lr = arg2;
znode_t *zp;
nvlist_t *nvl;
size_t sa_size;
char *name;
char *value;
size_t size;
int error = 0;
ASSERT(spa_feature_is_active(zfsvfs->z_os->os_spa,
SPA_FEATURE_ZILSAXATTR));
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
return (error);
rw_enter(&zp->z_xattr_lock, RW_WRITER);
mutex_enter(&zp->z_lock);
if (zp->z_xattr_cached == NULL)
error = zfs_sa_get_xattr(zp);
mutex_exit(&zp->z_lock);
if (error)
goto out;
ASSERT(zp->z_xattr_cached);
nvl = zp->z_xattr_cached;
/* Get xattr name, value and size from log record */
size = lr->lr_size;
name = (char *)(lr + 1);
if (size == 0) {
value = NULL;
error = nvlist_remove(nvl, name, DATA_TYPE_BYTE_ARRAY);
} else {
value = name + strlen(name) + 1;
/* Limited to 32k to keep nvpair memory allocations small */
if (size > DXATTR_MAX_ENTRY_SIZE) {
error = SET_ERROR(EFBIG);
goto out;
}
/* Prevent the DXATTR SA from consuming the entire SA region */
error = nvlist_size(nvl, &sa_size, NV_ENCODE_XDR);
if (error)
goto out;
if (sa_size > DXATTR_MAX_SA_SIZE) {
error = SET_ERROR(EFBIG);
goto out;
}
error = nvlist_add_byte_array(nvl, name, (uchar_t *)value,
size);
}
/*
* Update the SA for additions, modifications, and removals. On
* error drop the inconsistent cached version of the nvlist, it
* will be reconstructed from the ARC when next accessed.
*/
if (error == 0)
error = zfs_sa_set_xattr(zp, name, value, size);
if (error) {
nvlist_free(nvl);
zp->z_xattr_cached = NULL;
}
out:
rw_exit(&zp->z_xattr_lock);
zrele(zp);
return (error);
}
static int
zfs_replay_acl_v0(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_acl_v0_t *lr = arg2;
ace_t *ace = (ace_t *)(lr + 1); /* ace array follows lr_acl_t */
vsecattr_t vsa = {0};
znode_t *zp;
int error;
if (byteswap) {
byteswap_uint64_array(lr, sizeof (*lr));
zfs_oldace_byteswap(ace, lr->lr_aclcnt);
}
if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
return (error);
vsa.vsa_mask = VSA_ACE | VSA_ACECNT;
vsa.vsa_aclcnt = lr->lr_aclcnt;
vsa.vsa_aclentsz = sizeof (ace_t) * vsa.vsa_aclcnt;
vsa.vsa_aclflags = 0;
vsa.vsa_aclentp = ace;
error = zfs_setsecattr(zp, &vsa, 0, kcred);
zrele(zp);
return (error);
}
/*
* Replaying ACLs is complicated by FUID support.
* The log record may contain some optional data
* to be used for replaying FUID's. These pieces
* are the actual FUIDs that were created initially.
* The FUID table index may no longer be valid and
* during zfs_create() a new index may be assigned.
* Because of this the log will contain the original
* domain+rid in order to create a new FUID.
*
* The individual ACEs may contain an ephemeral uid/gid which is no
* longer valid and will need to be replaced with an actual FUID.
*
*/
static int
zfs_replay_acl(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_acl_t *lr = arg2;
ace_t *ace = (ace_t *)(lr + 1);
vsecattr_t vsa = {0};
znode_t *zp;
int error;
if (byteswap) {
byteswap_uint64_array(lr, sizeof (*lr));
zfs_ace_byteswap(ace, lr->lr_acl_bytes, B_FALSE);
if (lr->lr_fuidcnt) {
byteswap_uint64_array((caddr_t)ace +
ZIL_ACE_LENGTH(lr->lr_acl_bytes),
lr->lr_fuidcnt * sizeof (uint64_t));
}
}
if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
return (error);
vsa.vsa_mask = VSA_ACE | VSA_ACECNT | VSA_ACE_ACLFLAGS;
vsa.vsa_aclcnt = lr->lr_aclcnt;
vsa.vsa_aclentp = ace;
vsa.vsa_aclentsz = lr->lr_acl_bytes;
vsa.vsa_aclflags = lr->lr_acl_flags;
if (lr->lr_fuidcnt) {
void *fuidstart = (caddr_t)ace +
ZIL_ACE_LENGTH(lr->lr_acl_bytes);
zfsvfs->z_fuid_replay =
zfs_replay_fuids(fuidstart, &fuidstart,
lr->lr_fuidcnt, lr->lr_domcnt, 0, 0);
}
error = zfs_setsecattr(zp, &vsa, 0, kcred);
if (zfsvfs->z_fuid_replay)
zfs_fuid_info_free(zfsvfs->z_fuid_replay);
zfsvfs->z_fuid_replay = NULL;
zrele(zp);
return (error);
}
static int
zfs_replay_clone_range(void *arg1, void *arg2, boolean_t byteswap)
{
zfsvfs_t *zfsvfs = arg1;
lr_clone_range_t *lr = arg2;
znode_t *zp;
int error;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0) {
/*
* Clones can be logged out of order, so don't be surprised if
* the file is gone - just return success.
*/
if (error == ENOENT)
error = 0;
return (error);
}
error = zfs_clone_range_replay(zp, lr->lr_offset, lr->lr_length,
lr->lr_blksz, lr->lr_bps, lr->lr_nbps);
zrele(zp);
return (error);
}
/*
* Callback vectors for replaying records
*/
zil_replay_func_t *const zfs_replay_vector[TX_MAX_TYPE] = {
zfs_replay_error, /* no such type */
zfs_replay_create, /* TX_CREATE */
zfs_replay_create, /* TX_MKDIR */
zfs_replay_create, /* TX_MKXATTR */
zfs_replay_create, /* TX_SYMLINK */
zfs_replay_remove, /* TX_REMOVE */
zfs_replay_remove, /* TX_RMDIR */
zfs_replay_link, /* TX_LINK */
zfs_replay_rename, /* TX_RENAME */
zfs_replay_write, /* TX_WRITE */
zfs_replay_truncate, /* TX_TRUNCATE */
zfs_replay_setattr, /* TX_SETATTR */
zfs_replay_acl_v0, /* TX_ACL_V0 */
zfs_replay_acl, /* TX_ACL */
zfs_replay_create_acl, /* TX_CREATE_ACL */
zfs_replay_create, /* TX_CREATE_ATTR */
zfs_replay_create_acl, /* TX_CREATE_ACL_ATTR */
zfs_replay_create_acl, /* TX_MKDIR_ACL */
zfs_replay_create, /* TX_MKDIR_ATTR */
zfs_replay_create_acl, /* TX_MKDIR_ACL_ATTR */
zfs_replay_write2, /* TX_WRITE2 */
zfs_replay_setsaxattr, /* TX_SETSAXATTR */
zfs_replay_rename_exchange, /* TX_RENAME_EXCHANGE */
zfs_replay_rename_whiteout, /* TX_RENAME_WHITEOUT */
zfs_replay_clone_range, /* TX_CLONE_RANGE */
};
diff --git a/sys/contrib/openzfs/module/zfs/zfs_vnops.c b/sys/contrib/openzfs/module/zfs/zfs_vnops.c
index 91b594e41cda..a6a27222bf4c 100644
--- a/sys/contrib/openzfs/module/zfs/zfs_vnops.c
+++ b/sys/contrib/openzfs/module/zfs/zfs_vnops.c
@@ -1,1472 +1,1472 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2012, 2018 by Delphix. All rights reserved.
* Copyright (c) 2015 by Chunwei Chen. All rights reserved.
* Copyright 2017 Nexenta Systems, Inc.
* Copyright (c) 2021, 2022 by Pawel Jakub Dawidek
*/
/* Portions Copyright 2007 Jeremy Teo */
/* Portions Copyright 2010 Robert Milkowski */
#include <sys/types.h>
#include <sys/param.h>
#include <sys/time.h>
#include <sys/sysmacros.h>
#include <sys/vfs.h>
#include <sys/uio_impl.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/kmem.h>
#include <sys/cmn_err.h>
#include <sys/errno.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_ioctl.h>
#include <sys/fs/zfs.h>
#include <sys/dmu.h>
#include <sys/dmu_objset.h>
#include <sys/spa.h>
#include <sys/txg.h>
#include <sys/dbuf.h>
#include <sys/policy.h>
#include <sys/zfeature.h>
#include <sys/zfs_vnops.h>
#include <sys/zfs_quota.h>
#include <sys/zfs_vfsops.h>
#include <sys/zfs_znode.h>
static ulong_t zfs_fsync_sync_cnt = 4;
int
zfs_fsync(znode_t *zp, int syncflag, cred_t *cr)
{
int error = 0;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
(void) tsd_set(zfs_fsyncer_key, (void *)(uintptr_t)zfs_fsync_sync_cnt);
if (zfsvfs->z_os->os_sync != ZFS_SYNC_DISABLED) {
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
goto out;
atomic_inc_32(&zp->z_sync_writes_cnt);
zil_commit(zfsvfs->z_log, zp->z_id);
atomic_dec_32(&zp->z_sync_writes_cnt);
zfs_exit(zfsvfs, FTAG);
}
out:
tsd_set(zfs_fsyncer_key, NULL);
return (error);
}
#if defined(SEEK_HOLE) && defined(SEEK_DATA)
/*
* Lseek support for finding holes (cmd == SEEK_HOLE) and
* data (cmd == SEEK_DATA). "off" is an in/out parameter.
*/
static int
zfs_holey_common(znode_t *zp, ulong_t cmd, loff_t *off)
{
zfs_locked_range_t *lr;
uint64_t noff = (uint64_t)*off; /* new offset */
uint64_t file_sz;
int error;
boolean_t hole;
file_sz = zp->z_size;
if (noff >= file_sz) {
return (SET_ERROR(ENXIO));
}
if (cmd == F_SEEK_HOLE)
hole = B_TRUE;
else
hole = B_FALSE;
/* Flush any mmap()'d data to disk */
if (zn_has_cached_data(zp, 0, file_sz - 1))
zn_flush_cached_data(zp, B_FALSE);
lr = zfs_rangelock_enter(&zp->z_rangelock, 0, UINT64_MAX, RL_READER);
error = dmu_offset_next(ZTOZSB(zp)->z_os, zp->z_id, hole, &noff);
zfs_rangelock_exit(lr);
if (error == ESRCH)
return (SET_ERROR(ENXIO));
/* File was dirty, so fall back to using generic logic */
if (error == EBUSY) {
if (hole)
*off = file_sz;
return (0);
}
/*
* We could find a hole that begins after the logical end-of-file,
* because dmu_offset_next() only works on whole blocks. If the
* EOF falls mid-block, then indicate that the "virtual hole"
* at the end of the file begins at the logical EOF, rather than
* at the end of the last block.
*/
if (noff > file_sz) {
ASSERT(hole);
noff = file_sz;
}
if (noff < *off)
return (error);
*off = noff;
return (error);
}
int
zfs_holey(znode_t *zp, ulong_t cmd, loff_t *off)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
error = zfs_holey_common(zp, cmd, off);
zfs_exit(zfsvfs, FTAG);
return (error);
}
#endif /* SEEK_HOLE && SEEK_DATA */
int
zfs_access(znode_t *zp, int mode, int flag, cred_t *cr)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if (flag & V_ACE_MASK)
#if defined(__linux__)
error = zfs_zaccess(zp, mode, flag, B_FALSE, cr,
- kcred->user_ns);
+ zfs_init_idmap);
#else
error = zfs_zaccess(zp, mode, flag, B_FALSE, cr,
NULL);
#endif
else
#if defined(__linux__)
- error = zfs_zaccess_rwx(zp, mode, flag, cr, kcred->user_ns);
+ error = zfs_zaccess_rwx(zp, mode, flag, cr, zfs_init_idmap);
#else
error = zfs_zaccess_rwx(zp, mode, flag, cr, NULL);
#endif
zfs_exit(zfsvfs, FTAG);
return (error);
}
static uint64_t zfs_vnops_read_chunk_size = 1024 * 1024; /* Tunable */
/*
* Read bytes from specified file into supplied buffer.
*
* IN: zp - inode of file to be read from.
* uio - structure supplying read location, range info,
* and return buffer.
* ioflag - O_SYNC flags; used to provide FRSYNC semantics.
* O_DIRECT flag; used to bypass page cache.
* cr - credentials of caller.
*
* OUT: uio - updated offset and range, buffer filled.
*
* RETURN: 0 on success, error code on failure.
*
* Side Effects:
* inode - atime updated if byte count > 0
*/
int
zfs_read(struct znode *zp, zfs_uio_t *uio, int ioflag, cred_t *cr)
{
(void) cr;
int error = 0;
boolean_t frsync = B_FALSE;
zfsvfs_t *zfsvfs = ZTOZSB(zp);
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
if (zp->z_pflags & ZFS_AV_QUARANTINED) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EACCES));
}
/* We don't copy out anything useful for directories. */
if (Z_ISDIR(ZTOTYPE(zp))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EISDIR));
}
/*
* Validate file offset
*/
if (zfs_uio_offset(uio) < (offset_t)0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
/*
* Fasttrack empty reads
*/
if (zfs_uio_resid(uio) == 0) {
zfs_exit(zfsvfs, FTAG);
return (0);
}
#ifdef FRSYNC
/*
* If we're in FRSYNC mode, sync out this znode before reading it.
* Only do this for non-snapshots.
*
* Some platforms do not support FRSYNC and instead map it
* to O_SYNC, which results in unnecessary calls to zil_commit. We
* only honor FRSYNC requests on platforms which support it.
*/
frsync = !!(ioflag & FRSYNC);
#endif
if (zfsvfs->z_log &&
(frsync || zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS))
zil_commit(zfsvfs->z_log, zp->z_id);
/*
* Lock the range against changes.
*/
zfs_locked_range_t *lr = zfs_rangelock_enter(&zp->z_rangelock,
zfs_uio_offset(uio), zfs_uio_resid(uio), RL_READER);
/*
* If we are reading past end-of-file we can skip
* to the end; but we might still need to set atime.
*/
if (zfs_uio_offset(uio) >= zp->z_size) {
error = 0;
goto out;
}
ASSERT(zfs_uio_offset(uio) < zp->z_size);
#if defined(__linux__)
ssize_t start_offset = zfs_uio_offset(uio);
#endif
ssize_t n = MIN(zfs_uio_resid(uio), zp->z_size - zfs_uio_offset(uio));
ssize_t start_resid = n;
while (n > 0) {
ssize_t nbytes = MIN(n, zfs_vnops_read_chunk_size -
P2PHASE(zfs_uio_offset(uio), zfs_vnops_read_chunk_size));
#ifdef UIO_NOCOPY
if (zfs_uio_segflg(uio) == UIO_NOCOPY)
error = mappedread_sf(zp, nbytes, uio);
else
#endif
if (zn_has_cached_data(zp, zfs_uio_offset(uio),
zfs_uio_offset(uio) + nbytes - 1) && !(ioflag & O_DIRECT)) {
error = mappedread(zp, nbytes, uio);
} else {
error = dmu_read_uio_dbuf(sa_get_db(zp->z_sa_hdl),
uio, nbytes);
}
if (error) {
/* convert checksum errors into IO errors */
if (error == ECKSUM)
error = SET_ERROR(EIO);
#if defined(__linux__)
/*
* if we actually read some bytes, bubbling EFAULT
* up to become EAGAIN isn't what we want here...
*
* ...on Linux, at least. On FBSD, doing this breaks.
*/
if (error == EFAULT &&
(zfs_uio_offset(uio) - start_offset) != 0)
error = 0;
#endif
break;
}
n -= nbytes;
}
int64_t nread = start_resid - n;
dataset_kstats_update_read_kstats(&zfsvfs->z_kstat, nread);
task_io_account_read(nread);
out:
zfs_rangelock_exit(lr);
ZFS_ACCESSTIME_STAMP(zfsvfs, zp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
static void
zfs_clear_setid_bits_if_necessary(zfsvfs_t *zfsvfs, znode_t *zp, cred_t *cr,
uint64_t *clear_setid_bits_txgp, dmu_tx_t *tx)
{
zilog_t *zilog = zfsvfs->z_log;
const uint64_t uid = KUID_TO_SUID(ZTOUID(zp));
ASSERT(clear_setid_bits_txgp != NULL);
ASSERT(tx != NULL);
/*
* Clear Set-UID/Set-GID bits on successful write if not
* privileged and at least one of the execute bits is set.
*
* It would be nice to do this after all writes have
* been done, but that would still expose the ISUID/ISGID
* to another app after the partial write is committed.
*
* Note: we don't call zfs_fuid_map_id() here because
* user 0 is not an ephemeral uid.
*/
mutex_enter(&zp->z_acl_lock);
if ((zp->z_mode & (S_IXUSR | (S_IXUSR >> 3) | (S_IXUSR >> 6))) != 0 &&
(zp->z_mode & (S_ISUID | S_ISGID)) != 0 &&
secpolicy_vnode_setid_retain(zp, cr,
((zp->z_mode & S_ISUID) != 0 && uid == 0)) != 0) {
uint64_t newmode;
zp->z_mode &= ~(S_ISUID | S_ISGID);
newmode = zp->z_mode;
(void) sa_update(zp->z_sa_hdl, SA_ZPL_MODE(zfsvfs),
(void *)&newmode, sizeof (uint64_t), tx);
mutex_exit(&zp->z_acl_lock);
/*
* Make sure SUID/SGID bits will be removed when we replay the
* log. If the setid bits are keep coming back, don't log more
* than one TX_SETATTR per transaction group.
*/
if (*clear_setid_bits_txgp != dmu_tx_get_txg(tx)) {
vattr_t va = {0};
va.va_mask = ATTR_MODE;
va.va_nodeid = zp->z_id;
va.va_mode = newmode;
zfs_log_setattr(zilog, tx, TX_SETATTR, zp, &va,
ATTR_MODE, NULL);
*clear_setid_bits_txgp = dmu_tx_get_txg(tx);
}
} else {
mutex_exit(&zp->z_acl_lock);
}
}
/*
* Write the bytes to a file.
*
* IN: zp - znode of file to be written to.
* uio - structure supplying write location, range info,
* and data buffer.
* ioflag - O_APPEND flag set if in append mode.
* O_DIRECT flag; used to bypass page cache.
* cr - credentials of caller.
*
* OUT: uio - updated offset and range.
*
* RETURN: 0 if success
* error code if failure
*
* Timestamps:
* ip - ctime|mtime updated if byte count > 0
*/
int
zfs_write(znode_t *zp, zfs_uio_t *uio, int ioflag, cred_t *cr)
{
int error = 0, error1;
ssize_t start_resid = zfs_uio_resid(uio);
uint64_t clear_setid_bits_txg = 0;
/*
* Fasttrack empty write
*/
ssize_t n = start_resid;
if (n == 0)
return (0);
zfsvfs_t *zfsvfs = ZTOZSB(zp);
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
sa_bulk_attr_t bulk[4];
int count = 0;
uint64_t mtime[2], ctime[2];
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL,
&zp->z_size, 8);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&zp->z_pflags, 8);
/*
* Callers might not be able to detect properly that we are read-only,
* so check it explicitly here.
*/
if (zfs_is_readonly(zfsvfs)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EROFS));
}
/*
* If immutable or not appending then return EPERM.
* Intentionally allow ZFS_READONLY through here.
* See zfs_zaccess_common()
*/
if ((zp->z_pflags & ZFS_IMMUTABLE) ||
((zp->z_pflags & ZFS_APPENDONLY) && !(ioflag & O_APPEND) &&
(zfs_uio_offset(uio) < zp->z_size))) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
/*
* Validate file offset
*/
offset_t woff = ioflag & O_APPEND ? zp->z_size : zfs_uio_offset(uio);
if (woff < 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
const uint64_t max_blksz = zfsvfs->z_max_blksz;
/*
* Pre-fault the pages to ensure slow (eg NFS) pages
* don't hold up txg.
* Skip this if uio contains loaned arc_buf.
*/
if (zfs_uio_prefaultpages(MIN(n, max_blksz), uio)) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EFAULT));
}
/*
* If in append mode, set the io offset pointer to eof.
*/
zfs_locked_range_t *lr;
if (ioflag & O_APPEND) {
/*
* Obtain an appending range lock to guarantee file append
* semantics. We reset the write offset once we have the lock.
*/
lr = zfs_rangelock_enter(&zp->z_rangelock, 0, n, RL_APPEND);
woff = lr->lr_offset;
if (lr->lr_length == UINT64_MAX) {
/*
* We overlocked the file because this write will cause
* the file block size to increase.
* Note that zp_size cannot change with this lock held.
*/
woff = zp->z_size;
}
zfs_uio_setoffset(uio, woff);
} else {
/*
* Note that if the file block size will change as a result of
* this write, then this range lock will lock the entire file
* so that we can re-write the block safely.
*/
lr = zfs_rangelock_enter(&zp->z_rangelock, woff, n, RL_WRITER);
}
if (zn_rlimit_fsize_uio(zp, uio)) {
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EFBIG));
}
const rlim64_t limit = MAXOFFSET_T;
if (woff >= limit) {
zfs_rangelock_exit(lr);
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EFBIG));
}
if (n > limit - woff)
n = limit - woff;
uint64_t end_size = MAX(zp->z_size, woff + n);
zilog_t *zilog = zfsvfs->z_log;
const uint64_t uid = KUID_TO_SUID(ZTOUID(zp));
const uint64_t gid = KGID_TO_SGID(ZTOGID(zp));
const uint64_t projid = zp->z_projid;
/*
* Write the file in reasonable size chunks. Each chunk is written
* in a separate transaction; this keeps the intent log records small
* and allows us to do more fine-grained space accounting.
*/
while (n > 0) {
woff = zfs_uio_offset(uio);
if (zfs_id_overblockquota(zfsvfs, DMU_USERUSED_OBJECT, uid) ||
zfs_id_overblockquota(zfsvfs, DMU_GROUPUSED_OBJECT, gid) ||
(projid != ZFS_DEFAULT_PROJID &&
zfs_id_overblockquota(zfsvfs, DMU_PROJECTUSED_OBJECT,
projid))) {
error = SET_ERROR(EDQUOT);
break;
}
arc_buf_t *abuf = NULL;
if (n >= max_blksz && woff >= zp->z_size &&
P2PHASE(woff, max_blksz) == 0 &&
zp->z_blksz == max_blksz) {
/*
* This write covers a full block. "Borrow" a buffer
* from the dmu so that we can fill it before we enter
* a transaction. This avoids the possibility of
* holding up the transaction if the data copy hangs
* up on a pagefault (e.g., from an NFS server mapping).
*/
size_t cbytes;
abuf = dmu_request_arcbuf(sa_get_db(zp->z_sa_hdl),
max_blksz);
ASSERT(abuf != NULL);
ASSERT(arc_buf_size(abuf) == max_blksz);
if ((error = zfs_uiocopy(abuf->b_data, max_blksz,
UIO_WRITE, uio, &cbytes))) {
dmu_return_arcbuf(abuf);
break;
}
ASSERT3S(cbytes, ==, max_blksz);
}
/*
* Start a transaction.
*/
dmu_tx_t *tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
dmu_buf_impl_t *db = (dmu_buf_impl_t *)sa_get_db(zp->z_sa_hdl);
DB_DNODE_ENTER(db);
dmu_tx_hold_write_by_dnode(tx, DB_DNODE(db), woff,
MIN(n, max_blksz));
DB_DNODE_EXIT(db);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
if (abuf != NULL)
dmu_return_arcbuf(abuf);
break;
}
/*
* NB: We must call zfs_clear_setid_bits_if_necessary before
* committing the transaction!
*/
/*
* If rangelock_enter() over-locked we grow the blocksize
* and then reduce the lock range. This will only happen
* on the first iteration since rangelock_reduce() will
* shrink down lr_length to the appropriate size.
*/
if (lr->lr_length == UINT64_MAX) {
uint64_t new_blksz;
if (zp->z_blksz > max_blksz) {
/*
* File's blocksize is already larger than the
* "recordsize" property. Only let it grow to
* the next power of 2.
*/
ASSERT(!ISP2(zp->z_blksz));
new_blksz = MIN(end_size,
1 << highbit64(zp->z_blksz));
} else {
new_blksz = MIN(end_size, max_blksz);
}
zfs_grow_blocksize(zp, new_blksz, tx);
zfs_rangelock_reduce(lr, woff, n);
}
/*
* XXX - should we really limit each write to z_max_blksz?
* Perhaps we should use SPA_MAXBLOCKSIZE chunks?
*/
const ssize_t nbytes =
MIN(n, max_blksz - P2PHASE(woff, max_blksz));
ssize_t tx_bytes;
if (abuf == NULL) {
tx_bytes = zfs_uio_resid(uio);
zfs_uio_fault_disable(uio, B_TRUE);
error = dmu_write_uio_dbuf(sa_get_db(zp->z_sa_hdl),
uio, nbytes, tx);
zfs_uio_fault_disable(uio, B_FALSE);
#ifdef __linux__
if (error == EFAULT) {
zfs_clear_setid_bits_if_necessary(zfsvfs, zp,
cr, &clear_setid_bits_txg, tx);
dmu_tx_commit(tx);
/*
* Account for partial writes before
* continuing the loop.
* Update needs to occur before the next
* zfs_uio_prefaultpages, or prefaultpages may
* error, and we may break the loop early.
*/
if (tx_bytes != zfs_uio_resid(uio))
n -= tx_bytes - zfs_uio_resid(uio);
if (zfs_uio_prefaultpages(MIN(n, max_blksz),
uio)) {
break;
}
continue;
}
#endif
/*
* On FreeBSD, EFAULT should be propagated back to the
* VFS, which will handle faulting and will retry.
*/
if (error != 0 && error != EFAULT) {
zfs_clear_setid_bits_if_necessary(zfsvfs, zp,
cr, &clear_setid_bits_txg, tx);
dmu_tx_commit(tx);
break;
}
tx_bytes -= zfs_uio_resid(uio);
} else {
/* Implied by abuf != NULL: */
ASSERT3S(n, >=, max_blksz);
ASSERT0(P2PHASE(woff, max_blksz));
/*
* We can simplify nbytes to MIN(n, max_blksz) since
* P2PHASE(woff, max_blksz) is 0, and knowing
* n >= max_blksz lets us simplify further:
*/
ASSERT3S(nbytes, ==, max_blksz);
/*
* Thus, we're writing a full block at a block-aligned
* offset and extending the file past EOF.
*
* dmu_assign_arcbuf_by_dbuf() will directly assign the
* arc buffer to a dbuf.
*/
error = dmu_assign_arcbuf_by_dbuf(
sa_get_db(zp->z_sa_hdl), woff, abuf, tx);
if (error != 0) {
/*
* XXX This might not be necessary if
* dmu_assign_arcbuf_by_dbuf is guaranteed
* to be atomic.
*/
zfs_clear_setid_bits_if_necessary(zfsvfs, zp,
cr, &clear_setid_bits_txg, tx);
dmu_return_arcbuf(abuf);
dmu_tx_commit(tx);
break;
}
ASSERT3S(nbytes, <=, zfs_uio_resid(uio));
zfs_uioskip(uio, nbytes);
tx_bytes = nbytes;
}
if (tx_bytes &&
zn_has_cached_data(zp, woff, woff + tx_bytes - 1) &&
!(ioflag & O_DIRECT)) {
update_pages(zp, woff, tx_bytes, zfsvfs->z_os);
}
/*
* If we made no progress, we're done. If we made even
* partial progress, update the znode and ZIL accordingly.
*/
if (tx_bytes == 0) {
(void) sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zfsvfs),
(void *)&zp->z_size, sizeof (uint64_t), tx);
dmu_tx_commit(tx);
ASSERT(error != 0);
break;
}
zfs_clear_setid_bits_if_necessary(zfsvfs, zp, cr,
&clear_setid_bits_txg, tx);
zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime);
/*
* Update the file size (zp_size) if it has changed;
* account for possible concurrent updates.
*/
while ((end_size = zp->z_size) < zfs_uio_offset(uio)) {
(void) atomic_cas_64(&zp->z_size, end_size,
zfs_uio_offset(uio));
ASSERT(error == 0 || error == EFAULT);
}
/*
* If we are replaying and eof is non zero then force
* the file size to the specified eof. Note, there's no
* concurrency during replay.
*/
if (zfsvfs->z_replay && zfsvfs->z_replay_eof != 0)
zp->z_size = zfsvfs->z_replay_eof;
error1 = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
if (error1 != 0)
/* Avoid clobbering EFAULT. */
error = error1;
/*
* NB: During replay, the TX_SETATTR record logged by
* zfs_clear_setid_bits_if_necessary must precede any of
* the TX_WRITE records logged here.
*/
zfs_log_write(zilog, tx, TX_WRITE, zp, woff, tx_bytes, ioflag,
NULL, NULL);
dmu_tx_commit(tx);
if (error != 0)
break;
ASSERT3S(tx_bytes, ==, nbytes);
n -= nbytes;
if (n > 0) {
if (zfs_uio_prefaultpages(MIN(n, max_blksz), uio)) {
error = SET_ERROR(EFAULT);
break;
}
}
}
zfs_znode_update_vfs(zp);
zfs_rangelock_exit(lr);
/*
* If we're in replay mode, or we made no progress, or the
* uio data is inaccessible return an error. Otherwise, it's
* at least a partial write, so it's successful.
*/
if (zfsvfs->z_replay || zfs_uio_resid(uio) == start_resid ||
error == EFAULT) {
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (ioflag & (O_SYNC | O_DSYNC) ||
zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, zp->z_id);
const int64_t nwritten = start_resid - zfs_uio_resid(uio);
dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, nwritten);
task_io_account_write(nwritten);
zfs_exit(zfsvfs, FTAG);
return (0);
}
int
zfs_getsecattr(znode_t *zp, vsecattr_t *vsecp, int flag, cred_t *cr)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
boolean_t skipaclchk = (flag & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
error = zfs_getacl(zp, vsecp, skipaclchk, cr);
zfs_exit(zfsvfs, FTAG);
return (error);
}
int
zfs_setsecattr(znode_t *zp, vsecattr_t *vsecp, int flag, cred_t *cr)
{
zfsvfs_t *zfsvfs = ZTOZSB(zp);
int error;
boolean_t skipaclchk = (flag & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE;
zilog_t *zilog = zfsvfs->z_log;
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
error = zfs_setacl(zp, vsecp, skipaclchk, cr);
if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS)
zil_commit(zilog, 0);
zfs_exit(zfsvfs, FTAG);
return (error);
}
#ifdef ZFS_DEBUG
static int zil_fault_io = 0;
#endif
static void zfs_get_done(zgd_t *zgd, int error);
/*
* Get data to generate a TX_WRITE intent log record.
*/
int
zfs_get_data(void *arg, uint64_t gen, lr_write_t *lr, char *buf,
struct lwb *lwb, zio_t *zio)
{
zfsvfs_t *zfsvfs = arg;
objset_t *os = zfsvfs->z_os;
znode_t *zp;
uint64_t object = lr->lr_foid;
uint64_t offset = lr->lr_offset;
uint64_t size = lr->lr_length;
dmu_buf_t *db;
zgd_t *zgd;
int error = 0;
uint64_t zp_gen;
ASSERT3P(lwb, !=, NULL);
ASSERT3P(zio, !=, NULL);
ASSERT3U(size, !=, 0);
/*
* Nothing to do if the file has been removed
*/
if (zfs_zget(zfsvfs, object, &zp) != 0)
return (SET_ERROR(ENOENT));
if (zp->z_unlinked) {
/*
* Release the vnode asynchronously as we currently have the
* txg stopped from syncing.
*/
zfs_zrele_async(zp);
return (SET_ERROR(ENOENT));
}
/* check if generation number matches */
if (sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(zfsvfs), &zp_gen,
sizeof (zp_gen)) != 0) {
zfs_zrele_async(zp);
return (SET_ERROR(EIO));
}
if (zp_gen != gen) {
zfs_zrele_async(zp);
return (SET_ERROR(ENOENT));
}
zgd = kmem_zalloc(sizeof (zgd_t), KM_SLEEP);
zgd->zgd_lwb = lwb;
zgd->zgd_private = zp;
/*
* Write records come in two flavors: immediate and indirect.
* For small writes it's cheaper to store the data with the
* log record (immediate); for large writes it's cheaper to
* sync the data and get a pointer to it (indirect) so that
* we don't have to write the data twice.
*/
if (buf != NULL) { /* immediate write */
zgd->zgd_lr = zfs_rangelock_enter(&zp->z_rangelock,
offset, size, RL_READER);
/* test for truncation needs to be done while range locked */
if (offset >= zp->z_size) {
error = SET_ERROR(ENOENT);
} else {
error = dmu_read(os, object, offset, size, buf,
DMU_READ_NO_PREFETCH);
}
ASSERT(error == 0 || error == ENOENT);
} else { /* indirect write */
/*
* Have to lock the whole block to ensure when it's
* written out and its checksum is being calculated
* that no one can change the data. We need to re-check
* blocksize after we get the lock in case it's changed!
*/
for (;;) {
uint64_t blkoff;
size = zp->z_blksz;
blkoff = ISP2(size) ? P2PHASE(offset, size) : offset;
offset -= blkoff;
zgd->zgd_lr = zfs_rangelock_enter(&zp->z_rangelock,
offset, size, RL_READER);
if (zp->z_blksz == size)
break;
offset += blkoff;
zfs_rangelock_exit(zgd->zgd_lr);
}
/* test for truncation needs to be done while range locked */
if (lr->lr_offset >= zp->z_size)
error = SET_ERROR(ENOENT);
#ifdef ZFS_DEBUG
if (zil_fault_io) {
error = SET_ERROR(EIO);
zil_fault_io = 0;
}
#endif
if (error == 0)
error = dmu_buf_hold(os, object, offset, zgd, &db,
DMU_READ_NO_PREFETCH);
if (error == 0) {
blkptr_t *bp = &lr->lr_blkptr;
zgd->zgd_db = db;
zgd->zgd_bp = bp;
ASSERT(db->db_offset == offset);
ASSERT(db->db_size == size);
error = dmu_sync(zio, lr->lr_common.lrc_txg,
zfs_get_done, zgd);
ASSERT(error || lr->lr_length <= size);
/*
* On success, we need to wait for the write I/O
* initiated by dmu_sync() to complete before we can
* release this dbuf. We will finish everything up
* in the zfs_get_done() callback.
*/
if (error == 0)
return (0);
if (error == EALREADY) {
lr->lr_common.lrc_txtype = TX_WRITE2;
/*
* TX_WRITE2 relies on the data previously
* written by the TX_WRITE that caused
* EALREADY. We zero out the BP because
* it is the old, currently-on-disk BP.
*/
zgd->zgd_bp = NULL;
BP_ZERO(bp);
error = 0;
}
}
}
zfs_get_done(zgd, error);
return (error);
}
static void
zfs_get_done(zgd_t *zgd, int error)
{
(void) error;
znode_t *zp = zgd->zgd_private;
if (zgd->zgd_db)
dmu_buf_rele(zgd->zgd_db, zgd);
zfs_rangelock_exit(zgd->zgd_lr);
/*
* Release the vnode asynchronously as we currently have the
* txg stopped from syncing.
*/
zfs_zrele_async(zp);
kmem_free(zgd, sizeof (zgd_t));
}
static int
zfs_enter_two(zfsvfs_t *zfsvfs1, zfsvfs_t *zfsvfs2, const char *tag)
{
int error;
/* Swap. Not sure if the order of zfs_enter()s is important. */
if (zfsvfs1 > zfsvfs2) {
zfsvfs_t *tmpzfsvfs;
tmpzfsvfs = zfsvfs2;
zfsvfs2 = zfsvfs1;
zfsvfs1 = tmpzfsvfs;
}
error = zfs_enter(zfsvfs1, tag);
if (error != 0)
return (error);
if (zfsvfs1 != zfsvfs2) {
error = zfs_enter(zfsvfs2, tag);
if (error != 0) {
zfs_exit(zfsvfs1, tag);
return (error);
}
}
return (0);
}
static void
zfs_exit_two(zfsvfs_t *zfsvfs1, zfsvfs_t *zfsvfs2, const char *tag)
{
zfs_exit(zfsvfs1, tag);
if (zfsvfs1 != zfsvfs2)
zfs_exit(zfsvfs2, tag);
}
/*
* We split each clone request in chunks that can fit into a single ZIL
* log entry. Each ZIL log entry can fit 130816 bytes for a block cloning
* operation (see zil_max_log_data() and zfs_log_clone_range()). This gives
* us room for storing 1022 block pointers.
*
* On success, the function return the number of bytes copied in *lenp.
* Note, it doesn't return how much bytes are left to be copied.
*/
int
zfs_clone_range(znode_t *inzp, uint64_t *inoffp, znode_t *outzp,
uint64_t *outoffp, uint64_t *lenp, cred_t *cr)
{
zfsvfs_t *inzfsvfs, *outzfsvfs;
objset_t *inos, *outos;
zfs_locked_range_t *inlr, *outlr;
dmu_buf_impl_t *db;
dmu_tx_t *tx;
zilog_t *zilog;
uint64_t inoff, outoff, len, done;
uint64_t outsize, size;
int error;
int count = 0;
sa_bulk_attr_t bulk[3];
uint64_t mtime[2], ctime[2];
uint64_t uid, gid, projid;
blkptr_t *bps;
size_t maxblocks, nbps;
uint_t inblksz;
uint64_t clear_setid_bits_txg = 0;
inoff = *inoffp;
outoff = *outoffp;
len = *lenp;
done = 0;
inzfsvfs = ZTOZSB(inzp);
outzfsvfs = ZTOZSB(outzp);
inos = inzfsvfs->z_os;
outos = outzfsvfs->z_os;
/*
* Both source and destination have to belong to the same storage pool.
*/
if (dmu_objset_spa(inos) != dmu_objset_spa(outos)) {
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
/*
* We need to call zfs_enter() potentially on two different datasets,
* so we need a dedicated function for that.
*/
error = zfs_enter_two(inzfsvfs, outzfsvfs, FTAG);
if (error != 0)
return (error);
ASSERT(!outzfsvfs->z_replay);
error = zfs_verify_zp(inzp);
if (error == 0)
error = zfs_verify_zp(outzp);
if (error != 0) {
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (error);
}
if (!spa_feature_is_enabled(dmu_objset_spa(outos),
SPA_FEATURE_BLOCK_CLONING)) {
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (SET_ERROR(EXDEV));
}
/*
* We don't copy source file's flags that's why we don't allow to clone
* files that are in quarantine.
*/
if (inzp->z_pflags & ZFS_AV_QUARANTINED) {
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (SET_ERROR(EACCES));
}
if (inoff >= inzp->z_size) {
*lenp = 0;
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (0);
}
if (len > inzp->z_size - inoff) {
len = inzp->z_size - inoff;
}
if (len == 0) {
*lenp = 0;
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (0);
}
/*
* Callers might not be able to detect properly that we are read-only,
* so check it explicitly here.
*/
if (zfs_is_readonly(outzfsvfs)) {
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (SET_ERROR(EROFS));
}
/*
* If immutable or not appending then return EPERM.
* Intentionally allow ZFS_READONLY through here.
* See zfs_zaccess_common()
*/
if ((outzp->z_pflags & ZFS_IMMUTABLE) != 0) {
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (SET_ERROR(EPERM));
}
/*
* No overlapping if we are cloning within the same file.
*/
if (inzp == outzp) {
if (inoff < outoff + len && outoff < inoff + len) {
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
}
/*
* Maintain predictable lock order.
*/
if (inzp < outzp || (inzp == outzp && inoff < outoff)) {
inlr = zfs_rangelock_enter(&inzp->z_rangelock, inoff, len,
RL_READER);
outlr = zfs_rangelock_enter(&outzp->z_rangelock, outoff, len,
RL_WRITER);
} else {
outlr = zfs_rangelock_enter(&outzp->z_rangelock, outoff, len,
RL_WRITER);
inlr = zfs_rangelock_enter(&inzp->z_rangelock, inoff, len,
RL_READER);
}
inblksz = inzp->z_blksz;
/*
* We cannot clone into files with different block size.
*/
if (inblksz != outzp->z_blksz && outzp->z_size > inblksz) {
error = SET_ERROR(EXDEV);
goto unlock;
}
/*
* Offsets and len must be at block boundries.
*/
if ((inoff % inblksz) != 0 || (outoff % inblksz) != 0) {
error = SET_ERROR(EXDEV);
goto unlock;
}
/*
* Length must be multipe of blksz, except for the end of the file.
*/
if ((len % inblksz) != 0 &&
(len < inzp->z_size - inoff || len < outzp->z_size - outoff)) {
error = SET_ERROR(EXDEV);
goto unlock;
}
error = zn_rlimit_fsize(outoff + len);
if (error != 0) {
goto unlock;
}
if (inoff >= MAXOFFSET_T || outoff >= MAXOFFSET_T) {
error = SET_ERROR(EFBIG);
goto unlock;
}
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(outzfsvfs), NULL,
&mtime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(outzfsvfs), NULL,
&ctime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(outzfsvfs), NULL,
&outzp->z_size, 8);
zilog = outzfsvfs->z_log;
maxblocks = zil_max_log_data(zilog, sizeof (lr_clone_range_t)) /
sizeof (bps[0]);
uid = KUID_TO_SUID(ZTOUID(outzp));
gid = KGID_TO_SGID(ZTOGID(outzp));
projid = outzp->z_projid;
bps = kmem_alloc(sizeof (bps[0]) * maxblocks, KM_SLEEP);
/*
* Clone the file in reasonable size chunks. Each chunk is cloned
* in a separate transaction; this keeps the intent log records small
* and allows us to do more fine-grained space accounting.
*/
while (len > 0) {
size = MIN(inblksz * maxblocks, len);
if (zfs_id_overblockquota(outzfsvfs, DMU_USERUSED_OBJECT,
uid) ||
zfs_id_overblockquota(outzfsvfs, DMU_GROUPUSED_OBJECT,
gid) ||
(projid != ZFS_DEFAULT_PROJID &&
zfs_id_overblockquota(outzfsvfs, DMU_PROJECTUSED_OBJECT,
projid))) {
error = SET_ERROR(EDQUOT);
break;
}
/*
* Start a transaction.
*/
tx = dmu_tx_create(outos);
nbps = maxblocks;
error = dmu_read_l0_bps(inos, inzp->z_id, inoff, size, tx, bps,
&nbps);
if (error != 0) {
dmu_tx_abort(tx);
/*
* If we are tyring to clone a block that was created
* in the current transaction group. Return an error,
* so the caller can fallback to just copying the data.
*/
if (error == EAGAIN) {
error = SET_ERROR(EXDEV);
}
break;
}
/*
* Encrypted data is fine as long as it comes from the same
* dataset.
* TODO: We want to extend it in the future to allow cloning to
* datasets with the same keys, like clones or to be able to
* clone a file from a snapshot of an encrypted dataset into the
* dataset itself.
*/
if (BP_IS_PROTECTED(&bps[0])) {
if (inzfsvfs != outzfsvfs) {
dmu_tx_abort(tx);
error = SET_ERROR(EXDEV);
break;
}
}
dmu_tx_hold_sa(tx, outzp->z_sa_hdl, B_FALSE);
db = (dmu_buf_impl_t *)sa_get_db(outzp->z_sa_hdl);
DB_DNODE_ENTER(db);
dmu_tx_hold_clone_by_dnode(tx, DB_DNODE(db), outoff, size);
DB_DNODE_EXIT(db);
zfs_sa_upgrade_txholds(tx, outzp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error != 0) {
dmu_tx_abort(tx);
break;
}
/*
* Copy source znode's block size. This only happens on the
* first iteration since zfs_rangelock_reduce() will shrink down
* lr_len to the appropriate size.
*/
if (outlr->lr_length == UINT64_MAX) {
zfs_grow_blocksize(outzp, inblksz, tx);
/*
* Round range lock up to the block boundary, so we
* prevent appends until we are done.
*/
zfs_rangelock_reduce(outlr, outoff,
((len - 1) / inblksz + 1) * inblksz);
}
dmu_brt_clone(outos, outzp->z_id, outoff, size, tx, bps, nbps,
B_FALSE);
zfs_clear_setid_bits_if_necessary(outzfsvfs, outzp, cr,
&clear_setid_bits_txg, tx);
zfs_tstamp_update_setup(outzp, CONTENT_MODIFIED, mtime, ctime);
/*
* Update the file size (zp_size) if it has changed;
* account for possible concurrent updates.
*/
while ((outsize = outzp->z_size) < outoff + size) {
(void) atomic_cas_64(&outzp->z_size, outsize,
outoff + size);
}
error = sa_bulk_update(outzp->z_sa_hdl, bulk, count, tx);
zfs_log_clone_range(zilog, tx, TX_CLONE_RANGE, outzp, outoff,
size, inblksz, bps, nbps);
dmu_tx_commit(tx);
if (error != 0)
break;
inoff += size;
outoff += size;
len -= size;
done += size;
}
kmem_free(bps, sizeof (bps[0]) * maxblocks);
zfs_znode_update_vfs(outzp);
unlock:
zfs_rangelock_exit(outlr);
zfs_rangelock_exit(inlr);
if (done > 0) {
/*
* If we have made at least partial progress, reset the error.
*/
error = 0;
ZFS_ACCESSTIME_STAMP(inzfsvfs, inzp);
if (outos->os_sync == ZFS_SYNC_ALWAYS) {
zil_commit(zilog, outzp->z_id);
}
*inoffp += done;
*outoffp += done;
*lenp = done;
}
zfs_exit_two(inzfsvfs, outzfsvfs, FTAG);
return (error);
}
/*
* Usual pattern would be to call zfs_clone_range() from zfs_replay_clone(),
* but we cannot do that, because when replaying we don't have source znode
* available. This is why we need a dedicated replay function.
*/
int
zfs_clone_range_replay(znode_t *zp, uint64_t off, uint64_t len, uint64_t blksz,
const blkptr_t *bps, size_t nbps)
{
zfsvfs_t *zfsvfs;
dmu_buf_impl_t *db;
dmu_tx_t *tx;
int error;
int count = 0;
sa_bulk_attr_t bulk[3];
uint64_t mtime[2], ctime[2];
ASSERT3U(off, <, MAXOFFSET_T);
ASSERT3U(len, >, 0);
ASSERT3U(nbps, >, 0);
zfsvfs = ZTOZSB(zp);
ASSERT(spa_feature_is_enabled(dmu_objset_spa(zfsvfs->z_os),
SPA_FEATURE_BLOCK_CLONING));
if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0)
return (error);
ASSERT(zfsvfs->z_replay);
ASSERT(!zfs_is_readonly(zfsvfs));
if ((off % blksz) != 0) {
zfs_exit(zfsvfs, FTAG);
return (SET_ERROR(EINVAL));
}
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16);
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL,
&zp->z_size, 8);
/*
* Start a transaction.
*/
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
db = (dmu_buf_impl_t *)sa_get_db(zp->z_sa_hdl);
DB_DNODE_ENTER(db);
dmu_tx_hold_clone_by_dnode(tx, DB_DNODE(db), off, len);
DB_DNODE_EXIT(db);
zfs_sa_upgrade_txholds(tx, zp);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error != 0) {
dmu_tx_abort(tx);
zfs_exit(zfsvfs, FTAG);
return (error);
}
if (zp->z_blksz < blksz)
zfs_grow_blocksize(zp, blksz, tx);
dmu_brt_clone(zfsvfs->z_os, zp->z_id, off, len, tx, bps, nbps, B_TRUE);
zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime);
if (zp->z_size < off + len)
zp->z_size = off + len;
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
/*
* zil_replaying() not only check if we are replaying ZIL, but also
* updates the ZIL header to record replay progress.
*/
VERIFY(zil_replaying(zfsvfs->z_log, tx));
dmu_tx_commit(tx);
zfs_znode_update_vfs(zp);
zfs_exit(zfsvfs, FTAG);
return (error);
}
EXPORT_SYMBOL(zfs_access);
EXPORT_SYMBOL(zfs_fsync);
EXPORT_SYMBOL(zfs_holey);
EXPORT_SYMBOL(zfs_read);
EXPORT_SYMBOL(zfs_write);
EXPORT_SYMBOL(zfs_getsecattr);
EXPORT_SYMBOL(zfs_setsecattr);
EXPORT_SYMBOL(zfs_clone_range);
EXPORT_SYMBOL(zfs_clone_range_replay);
ZFS_MODULE_PARAM(zfs_vnops, zfs_vnops_, read_chunk_size, U64, ZMOD_RW,
"Bytes to read per chunk");
diff --git a/sys/contrib/openzfs/module/zfs/zil.c b/sys/contrib/openzfs/module/zfs/zil.c
index d1631c2ac9db..ec9da706a806 100644
--- a/sys/contrib/openzfs/module/zfs/zil.c
+++ b/sys/contrib/openzfs/module/zfs/zil.c
@@ -1,4116 +1,4123 @@
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or https://opensource.org/licenses/CDDL-1.0.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2018 by Delphix. All rights reserved.
* Copyright (c) 2014 Integros [integros.com]
* Copyright (c) 2018 Datto Inc.
*/
/* Portions Copyright 2010 Robert Milkowski */
#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/dmu.h>
#include <sys/zap.h>
#include <sys/arc.h>
#include <sys/stat.h>
#include <sys/zil.h>
#include <sys/zil_impl.h>
#include <sys/dsl_dataset.h>
#include <sys/vdev_impl.h>
#include <sys/dmu_tx.h>
#include <sys/dsl_pool.h>
#include <sys/metaslab.h>
#include <sys/trace_zfs.h>
#include <sys/abd.h>
#include <sys/brt.h>
#include <sys/wmsum.h>
/*
* The ZFS Intent Log (ZIL) saves "transaction records" (itxs) of system
* calls that change the file system. Each itx has enough information to
* be able to replay them after a system crash, power loss, or
* equivalent failure mode. These are stored in memory until either:
*
* 1. they are committed to the pool by the DMU transaction group
* (txg), at which point they can be discarded; or
* 2. they are committed to the on-disk ZIL for the dataset being
* modified (e.g. due to an fsync, O_DSYNC, or other synchronous
* requirement).
*
* In the event of a crash or power loss, the itxs contained by each
* dataset's on-disk ZIL will be replayed when that dataset is first
* instantiated (e.g. if the dataset is a normal filesystem, when it is
* first mounted).
*
* As hinted at above, there is one ZIL per dataset (both the in-memory
* representation, and the on-disk representation). The on-disk format
* consists of 3 parts:
*
* - a single, per-dataset, ZIL header; which points to a chain of
* - zero or more ZIL blocks; each of which contains
* - zero or more ZIL records
*
* A ZIL record holds the information necessary to replay a single
* system call transaction. A ZIL block can hold many ZIL records, and
* the blocks are chained together, similarly to a singly linked list.
*
* Each ZIL block contains a block pointer (blkptr_t) to the next ZIL
* block in the chain, and the ZIL header points to the first block in
* the chain.
*
* Note, there is not a fixed place in the pool to hold these ZIL
* blocks; they are dynamically allocated and freed as needed from the
* blocks available on the pool, though they can be preferentially
* allocated from a dedicated "log" vdev.
*/
/*
* This controls the amount of time that a ZIL block (lwb) will remain
* "open" when it isn't "full", and it has a thread waiting for it to be
* committed to stable storage. Please refer to the zil_commit_waiter()
* function (and the comments within it) for more details.
*/
static uint_t zfs_commit_timeout_pct = 5;
/*
* Minimal time we care to delay commit waiting for more ZIL records.
* At least FreeBSD kernel can't sleep for less than 2us at its best.
* So requests to sleep for less then 5us is a waste of CPU time with
* a risk of significant log latency increase due to oversleep.
*/
static uint64_t zil_min_commit_timeout = 5000;
/*
* See zil.h for more information about these fields.
*/
static zil_kstat_values_t zil_stats = {
{ "zil_commit_count", KSTAT_DATA_UINT64 },
{ "zil_commit_writer_count", KSTAT_DATA_UINT64 },
{ "zil_itx_count", KSTAT_DATA_UINT64 },
{ "zil_itx_indirect_count", KSTAT_DATA_UINT64 },
{ "zil_itx_indirect_bytes", KSTAT_DATA_UINT64 },
{ "zil_itx_copied_count", KSTAT_DATA_UINT64 },
{ "zil_itx_copied_bytes", KSTAT_DATA_UINT64 },
{ "zil_itx_needcopy_count", KSTAT_DATA_UINT64 },
{ "zil_itx_needcopy_bytes", KSTAT_DATA_UINT64 },
{ "zil_itx_metaslab_normal_count", KSTAT_DATA_UINT64 },
{ "zil_itx_metaslab_normal_bytes", KSTAT_DATA_UINT64 },
{ "zil_itx_metaslab_slog_count", KSTAT_DATA_UINT64 },
{ "zil_itx_metaslab_slog_bytes", KSTAT_DATA_UINT64 },
};
static zil_sums_t zil_sums_global;
static kstat_t *zil_kstats_global;
/*
* Disable intent logging replay. This global ZIL switch affects all pools.
*/
int zil_replay_disable = 0;
/*
* Disable the DKIOCFLUSHWRITECACHE commands that are normally sent to
* the disk(s) by the ZIL after an LWB write has completed. Setting this
* will cause ZIL corruption on power loss if a volatile out-of-order
* write cache is enabled.
*/
static int zil_nocacheflush = 0;
/*
* Limit SLOG write size per commit executed with synchronous priority.
* Any writes above that will be executed with lower (asynchronous) priority
* to limit potential SLOG device abuse by single active ZIL writer.
*/
static uint64_t zil_slog_bulk = 768 * 1024;
static kmem_cache_t *zil_lwb_cache;
static kmem_cache_t *zil_zcw_cache;
#define LWB_EMPTY(lwb) ((BP_GET_LSIZE(&lwb->lwb_blk) - \
sizeof (zil_chain_t)) == (lwb->lwb_sz - lwb->lwb_nused))
static int
zil_bp_compare(const void *x1, const void *x2)
{
const dva_t *dva1 = &((zil_bp_node_t *)x1)->zn_dva;
const dva_t *dva2 = &((zil_bp_node_t *)x2)->zn_dva;
int cmp = TREE_CMP(DVA_GET_VDEV(dva1), DVA_GET_VDEV(dva2));
if (likely(cmp))
return (cmp);
return (TREE_CMP(DVA_GET_OFFSET(dva1), DVA_GET_OFFSET(dva2)));
}
static void
zil_bp_tree_init(zilog_t *zilog)
{
avl_create(&zilog->zl_bp_tree, zil_bp_compare,
sizeof (zil_bp_node_t), offsetof(zil_bp_node_t, zn_node));
}
static void
zil_bp_tree_fini(zilog_t *zilog)
{
avl_tree_t *t = &zilog->zl_bp_tree;
zil_bp_node_t *zn;
void *cookie = NULL;
while ((zn = avl_destroy_nodes(t, &cookie)) != NULL)
kmem_free(zn, sizeof (zil_bp_node_t));
avl_destroy(t);
}
int
zil_bp_tree_add(zilog_t *zilog, const blkptr_t *bp)
{
avl_tree_t *t = &zilog->zl_bp_tree;
const dva_t *dva;
zil_bp_node_t *zn;
avl_index_t where;
if (BP_IS_EMBEDDED(bp))
return (0);
dva = BP_IDENTITY(bp);
if (avl_find(t, dva, &where) != NULL)
return (SET_ERROR(EEXIST));
zn = kmem_alloc(sizeof (zil_bp_node_t), KM_SLEEP);
zn->zn_dva = *dva;
avl_insert(t, zn, where);
return (0);
}
static zil_header_t *
zil_header_in_syncing_context(zilog_t *zilog)
{
return ((zil_header_t *)zilog->zl_header);
}
static void
zil_init_log_chain(zilog_t *zilog, blkptr_t *bp)
{
zio_cksum_t *zc = &bp->blk_cksum;
(void) random_get_pseudo_bytes((void *)&zc->zc_word[ZIL_ZC_GUID_0],
sizeof (zc->zc_word[ZIL_ZC_GUID_0]));
(void) random_get_pseudo_bytes((void *)&zc->zc_word[ZIL_ZC_GUID_1],
sizeof (zc->zc_word[ZIL_ZC_GUID_1]));
zc->zc_word[ZIL_ZC_OBJSET] = dmu_objset_id(zilog->zl_os);
zc->zc_word[ZIL_ZC_SEQ] = 1ULL;
}
static int
zil_kstats_global_update(kstat_t *ksp, int rw)
{
zil_kstat_values_t *zs = ksp->ks_data;
ASSERT3P(&zil_stats, ==, zs);
if (rw == KSTAT_WRITE) {
return (SET_ERROR(EACCES));
}
zil_kstat_values_update(zs, &zil_sums_global);
return (0);
}
/*
* Read a log block and make sure it's valid.
*/
static int
zil_read_log_block(zilog_t *zilog, boolean_t decrypt, const blkptr_t *bp,
blkptr_t *nbp, void *dst, char **end)
{
zio_flag_t zio_flags = ZIO_FLAG_CANFAIL;
arc_flags_t aflags = ARC_FLAG_WAIT;
arc_buf_t *abuf = NULL;
zbookmark_phys_t zb;
int error;
if (zilog->zl_header->zh_claim_txg == 0)
zio_flags |= ZIO_FLAG_SPECULATIVE | ZIO_FLAG_SCRUB;
if (!(zilog->zl_header->zh_flags & ZIL_CLAIM_LR_SEQ_VALID))
zio_flags |= ZIO_FLAG_SPECULATIVE;
if (!decrypt)
zio_flags |= ZIO_FLAG_RAW;
SET_BOOKMARK(&zb, bp->blk_cksum.zc_word[ZIL_ZC_OBJSET],
ZB_ZIL_OBJECT, ZB_ZIL_LEVEL, bp->blk_cksum.zc_word[ZIL_ZC_SEQ]);
error = arc_read(NULL, zilog->zl_spa, bp, arc_getbuf_func,
&abuf, ZIO_PRIORITY_SYNC_READ, zio_flags, &aflags, &zb);
if (error == 0) {
zio_cksum_t cksum = bp->blk_cksum;
/*
* Validate the checksummed log block.
*
* Sequence numbers should be... sequential. The checksum
* verifier for the next block should be bp's checksum plus 1.
*
* Also check the log chain linkage and size used.
*/
cksum.zc_word[ZIL_ZC_SEQ]++;
if (BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_ZILOG2) {
zil_chain_t *zilc = abuf->b_data;
char *lr = (char *)(zilc + 1);
uint64_t len = zilc->zc_nused - sizeof (zil_chain_t);
if (memcmp(&cksum, &zilc->zc_next_blk.blk_cksum,
sizeof (cksum)) || BP_IS_HOLE(&zilc->zc_next_blk)) {
error = SET_ERROR(ECKSUM);
} else {
ASSERT3U(len, <=, SPA_OLD_MAXBLOCKSIZE);
memcpy(dst, lr, len);
*end = (char *)dst + len;
*nbp = zilc->zc_next_blk;
}
} else {
char *lr = abuf->b_data;
uint64_t size = BP_GET_LSIZE(bp);
zil_chain_t *zilc = (zil_chain_t *)(lr + size) - 1;
if (memcmp(&cksum, &zilc->zc_next_blk.blk_cksum,
sizeof (cksum)) || BP_IS_HOLE(&zilc->zc_next_blk) ||
(zilc->zc_nused > (size - sizeof (*zilc)))) {
error = SET_ERROR(ECKSUM);
} else {
ASSERT3U(zilc->zc_nused, <=,
SPA_OLD_MAXBLOCKSIZE);
memcpy(dst, lr, zilc->zc_nused);
*end = (char *)dst + zilc->zc_nused;
*nbp = zilc->zc_next_blk;
}
}
arc_buf_destroy(abuf, &abuf);
}
return (error);
}
/*
* Read a TX_WRITE log data block.
*/
static int
zil_read_log_data(zilog_t *zilog, const lr_write_t *lr, void *wbuf)
{
zio_flag_t zio_flags = ZIO_FLAG_CANFAIL;
const blkptr_t *bp = &lr->lr_blkptr;
arc_flags_t aflags = ARC_FLAG_WAIT;
arc_buf_t *abuf = NULL;
zbookmark_phys_t zb;
int error;
if (BP_IS_HOLE(bp)) {
if (wbuf != NULL)
memset(wbuf, 0, MAX(BP_GET_LSIZE(bp), lr->lr_length));
return (0);
}
if (zilog->zl_header->zh_claim_txg == 0)
zio_flags |= ZIO_FLAG_SPECULATIVE | ZIO_FLAG_SCRUB;
/*
* If we are not using the resulting data, we are just checking that
* it hasn't been corrupted so we don't need to waste CPU time
* decompressing and decrypting it.
*/
if (wbuf == NULL)
zio_flags |= ZIO_FLAG_RAW;
ASSERT3U(BP_GET_LSIZE(bp), !=, 0);
SET_BOOKMARK(&zb, dmu_objset_id(zilog->zl_os), lr->lr_foid,
ZB_ZIL_LEVEL, lr->lr_offset / BP_GET_LSIZE(bp));
error = arc_read(NULL, zilog->zl_spa, bp, arc_getbuf_func, &abuf,
ZIO_PRIORITY_SYNC_READ, zio_flags, &aflags, &zb);
if (error == 0) {
if (wbuf != NULL)
memcpy(wbuf, abuf->b_data, arc_buf_size(abuf));
arc_buf_destroy(abuf, &abuf);
}
return (error);
}
void
zil_sums_init(zil_sums_t *zs)
{
wmsum_init(&zs->zil_commit_count, 0);
wmsum_init(&zs->zil_commit_writer_count, 0);
wmsum_init(&zs->zil_itx_count, 0);
wmsum_init(&zs->zil_itx_indirect_count, 0);
wmsum_init(&zs->zil_itx_indirect_bytes, 0);
wmsum_init(&zs->zil_itx_copied_count, 0);
wmsum_init(&zs->zil_itx_copied_bytes, 0);
wmsum_init(&zs->zil_itx_needcopy_count, 0);
wmsum_init(&zs->zil_itx_needcopy_bytes, 0);
wmsum_init(&zs->zil_itx_metaslab_normal_count, 0);
wmsum_init(&zs->zil_itx_metaslab_normal_bytes, 0);
wmsum_init(&zs->zil_itx_metaslab_slog_count, 0);
wmsum_init(&zs->zil_itx_metaslab_slog_bytes, 0);
}
void
zil_sums_fini(zil_sums_t *zs)
{
wmsum_fini(&zs->zil_commit_count);
wmsum_fini(&zs->zil_commit_writer_count);
wmsum_fini(&zs->zil_itx_count);
wmsum_fini(&zs->zil_itx_indirect_count);
wmsum_fini(&zs->zil_itx_indirect_bytes);
wmsum_fini(&zs->zil_itx_copied_count);
wmsum_fini(&zs->zil_itx_copied_bytes);
wmsum_fini(&zs->zil_itx_needcopy_count);
wmsum_fini(&zs->zil_itx_needcopy_bytes);
wmsum_fini(&zs->zil_itx_metaslab_normal_count);
wmsum_fini(&zs->zil_itx_metaslab_normal_bytes);
wmsum_fini(&zs->zil_itx_metaslab_slog_count);
wmsum_fini(&zs->zil_itx_metaslab_slog_bytes);
}
void
zil_kstat_values_update(zil_kstat_values_t *zs, zil_sums_t *zil_sums)
{
zs->zil_commit_count.value.ui64 =
wmsum_value(&zil_sums->zil_commit_count);
zs->zil_commit_writer_count.value.ui64 =
wmsum_value(&zil_sums->zil_commit_writer_count);
zs->zil_itx_count.value.ui64 =
wmsum_value(&zil_sums->zil_itx_count);
zs->zil_itx_indirect_count.value.ui64 =
wmsum_value(&zil_sums->zil_itx_indirect_count);
zs->zil_itx_indirect_bytes.value.ui64 =
wmsum_value(&zil_sums->zil_itx_indirect_bytes);
zs->zil_itx_copied_count.value.ui64 =
wmsum_value(&zil_sums->zil_itx_copied_count);
zs->zil_itx_copied_bytes.value.ui64 =
wmsum_value(&zil_sums->zil_itx_copied_bytes);
zs->zil_itx_needcopy_count.value.ui64 =
wmsum_value(&zil_sums->zil_itx_needcopy_count);
zs->zil_itx_needcopy_bytes.value.ui64 =
wmsum_value(&zil_sums->zil_itx_needcopy_bytes);
zs->zil_itx_metaslab_normal_count.value.ui64 =
wmsum_value(&zil_sums->zil_itx_metaslab_normal_count);
zs->zil_itx_metaslab_normal_bytes.value.ui64 =
wmsum_value(&zil_sums->zil_itx_metaslab_normal_bytes);
zs->zil_itx_metaslab_slog_count.value.ui64 =
wmsum_value(&zil_sums->zil_itx_metaslab_slog_count);
zs->zil_itx_metaslab_slog_bytes.value.ui64 =
wmsum_value(&zil_sums->zil_itx_metaslab_slog_bytes);
}
/*
* Parse the intent log, and call parse_func for each valid record within.
*/
int
zil_parse(zilog_t *zilog, zil_parse_blk_func_t *parse_blk_func,
zil_parse_lr_func_t *parse_lr_func, void *arg, uint64_t txg,
boolean_t decrypt)
{
const zil_header_t *zh = zilog->zl_header;
boolean_t claimed = !!zh->zh_claim_txg;
uint64_t claim_blk_seq = claimed ? zh->zh_claim_blk_seq : UINT64_MAX;
uint64_t claim_lr_seq = claimed ? zh->zh_claim_lr_seq : UINT64_MAX;
uint64_t max_blk_seq = 0;
uint64_t max_lr_seq = 0;
uint64_t blk_count = 0;
uint64_t lr_count = 0;
blkptr_t blk, next_blk = {{{{0}}}};
char *lrbuf, *lrp;
int error = 0;
/*
* Old logs didn't record the maximum zh_claim_lr_seq.
*/
if (!(zh->zh_flags & ZIL_CLAIM_LR_SEQ_VALID))
claim_lr_seq = UINT64_MAX;
/*
* Starting at the block pointed to by zh_log we read the log chain.
* For each block in the chain we strongly check that block to
* ensure its validity. We stop when an invalid block is found.
* For each block pointer in the chain we call parse_blk_func().
* For each record in each valid block we call parse_lr_func().
* If the log has been claimed, stop if we encounter a sequence
* number greater than the highest claimed sequence number.
*/
lrbuf = zio_buf_alloc(SPA_OLD_MAXBLOCKSIZE);
zil_bp_tree_init(zilog);
for (blk = zh->zh_log; !BP_IS_HOLE(&blk); blk = next_blk) {
uint64_t blk_seq = blk.blk_cksum.zc_word[ZIL_ZC_SEQ];
int reclen;
char *end = NULL;
if (blk_seq > claim_blk_seq)
break;
error = parse_blk_func(zilog, &blk, arg, txg);
if (error != 0)
break;
ASSERT3U(max_blk_seq, <, blk_seq);
max_blk_seq = blk_seq;
blk_count++;
if (max_lr_seq == claim_lr_seq && max_blk_seq == claim_blk_seq)
break;
error = zil_read_log_block(zilog, decrypt, &blk, &next_blk,
lrbuf, &end);
if (error != 0) {
if (claimed) {
char name[ZFS_MAX_DATASET_NAME_LEN];
dmu_objset_name(zilog->zl_os, name);
cmn_err(CE_WARN, "ZFS read log block error %d, "
"dataset %s, seq 0x%llx\n", error, name,
(u_longlong_t)blk_seq);
}
break;
}
for (lrp = lrbuf; lrp < end; lrp += reclen) {
lr_t *lr = (lr_t *)lrp;
reclen = lr->lrc_reclen;
ASSERT3U(reclen, >=, sizeof (lr_t));
if (lr->lrc_seq > claim_lr_seq)
goto done;
error = parse_lr_func(zilog, lr, arg, txg);
if (error != 0)
goto done;
ASSERT3U(max_lr_seq, <, lr->lrc_seq);
max_lr_seq = lr->lrc_seq;
lr_count++;
}
}
done:
zilog->zl_parse_error = error;
zilog->zl_parse_blk_seq = max_blk_seq;
zilog->zl_parse_lr_seq = max_lr_seq;
zilog->zl_parse_blk_count = blk_count;
zilog->zl_parse_lr_count = lr_count;
zil_bp_tree_fini(zilog);
zio_buf_free(lrbuf, SPA_OLD_MAXBLOCKSIZE);
return (error);
}
static int
zil_clear_log_block(zilog_t *zilog, const blkptr_t *bp, void *tx,
uint64_t first_txg)
{
(void) tx;
ASSERT(!BP_IS_HOLE(bp));
/*
* As we call this function from the context of a rewind to a
* checkpoint, each ZIL block whose txg is later than the txg
* that we rewind to is invalid. Thus, we return -1 so
* zil_parse() doesn't attempt to read it.
*/
if (bp->blk_birth >= first_txg)
return (-1);
if (zil_bp_tree_add(zilog, bp) != 0)
return (0);
zio_free(zilog->zl_spa, first_txg, bp);
return (0);
}
static int
zil_noop_log_record(zilog_t *zilog, const lr_t *lrc, void *tx,
uint64_t first_txg)
{
(void) zilog, (void) lrc, (void) tx, (void) first_txg;
return (0);
}
static int
zil_claim_log_block(zilog_t *zilog, const blkptr_t *bp, void *tx,
uint64_t first_txg)
{
/*
* Claim log block if not already committed and not already claimed.
* If tx == NULL, just verify that the block is claimable.
*/
if (BP_IS_HOLE(bp) || bp->blk_birth < first_txg ||
zil_bp_tree_add(zilog, bp) != 0)
return (0);
return (zio_wait(zio_claim(NULL, zilog->zl_spa,
tx == NULL ? 0 : first_txg, bp, spa_claim_notify, NULL,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE | ZIO_FLAG_SCRUB)));
}
static int
zil_claim_write(zilog_t *zilog, const lr_t *lrc, void *tx, uint64_t first_txg)
{
lr_write_t *lr = (lr_write_t *)lrc;
int error;
ASSERT(lrc->lrc_txtype == TX_WRITE);
/*
* If the block is not readable, don't claim it. This can happen
* in normal operation when a log block is written to disk before
* some of the dmu_sync() blocks it points to. In this case, the
* transaction cannot have been committed to anyone (we would have
* waited for all writes to be stable first), so it is semantically
* correct to declare this the end of the log.
*/
if (lr->lr_blkptr.blk_birth >= first_txg) {
error = zil_read_log_data(zilog, lr, NULL);
if (error != 0)
return (error);
}
return (zil_claim_log_block(zilog, &lr->lr_blkptr, tx, first_txg));
}
static int
zil_claim_clone_range(zilog_t *zilog, const lr_t *lrc, void *tx)
{
const lr_clone_range_t *lr = (const lr_clone_range_t *)lrc;
const blkptr_t *bp;
spa_t *spa;
uint_t ii;
ASSERT(lrc->lrc_txtype == TX_CLONE_RANGE);
if (tx == NULL) {
return (0);
}
/*
* XXX: Do we need to byteswap lr?
*/
spa = zilog->zl_spa;
for (ii = 0; ii < lr->lr_nbps; ii++) {
bp = &lr->lr_bps[ii];
/*
* When data in embedded into BP there is no need to create
* BRT entry as there is no data block. Just copy the BP as
* it contains the data.
*/
if (!BP_IS_HOLE(bp) && !BP_IS_EMBEDDED(bp)) {
brt_pending_add(spa, bp, tx);
}
}
return (0);
}
static int
zil_claim_log_record(zilog_t *zilog, const lr_t *lrc, void *tx,
uint64_t first_txg)
{
switch (lrc->lrc_txtype) {
case TX_WRITE:
return (zil_claim_write(zilog, lrc, tx, first_txg));
case TX_CLONE_RANGE:
return (zil_claim_clone_range(zilog, lrc, tx));
default:
return (0);
}
}
static int
zil_free_log_block(zilog_t *zilog, const blkptr_t *bp, void *tx,
uint64_t claim_txg)
{
(void) claim_txg;
zio_free(zilog->zl_spa, dmu_tx_get_txg(tx), bp);
return (0);
}
static int
zil_free_write(zilog_t *zilog, const lr_t *lrc, void *tx, uint64_t claim_txg)
{
lr_write_t *lr = (lr_write_t *)lrc;
blkptr_t *bp = &lr->lr_blkptr;
ASSERT(lrc->lrc_txtype == TX_WRITE);
/*
* If we previously claimed it, we need to free it.
*/
if (bp->blk_birth >= claim_txg && zil_bp_tree_add(zilog, bp) == 0 &&
!BP_IS_HOLE(bp)) {
zio_free(zilog->zl_spa, dmu_tx_get_txg(tx), bp);
}
return (0);
}
static int
zil_free_clone_range(zilog_t *zilog, const lr_t *lrc, void *tx)
{
const lr_clone_range_t *lr = (const lr_clone_range_t *)lrc;
const blkptr_t *bp;
spa_t *spa;
uint_t ii;
ASSERT(lrc->lrc_txtype == TX_CLONE_RANGE);
if (tx == NULL) {
return (0);
}
spa = zilog->zl_spa;
for (ii = 0; ii < lr->lr_nbps; ii++) {
bp = &lr->lr_bps[ii];
if (!BP_IS_HOLE(bp)) {
zio_free(spa, dmu_tx_get_txg(tx), bp);
}
}
return (0);
}
static int
zil_free_log_record(zilog_t *zilog, const lr_t *lrc, void *tx,
uint64_t claim_txg)
{
if (claim_txg == 0) {
return (0);
}
switch (lrc->lrc_txtype) {
case TX_WRITE:
return (zil_free_write(zilog, lrc, tx, claim_txg));
case TX_CLONE_RANGE:
return (zil_free_clone_range(zilog, lrc, tx));
default:
return (0);
}
}
static int
zil_lwb_vdev_compare(const void *x1, const void *x2)
{
const uint64_t v1 = ((zil_vdev_node_t *)x1)->zv_vdev;
const uint64_t v2 = ((zil_vdev_node_t *)x2)->zv_vdev;
return (TREE_CMP(v1, v2));
}
static lwb_t *
zil_alloc_lwb(zilog_t *zilog, blkptr_t *bp, boolean_t slog, uint64_t txg,
boolean_t fastwrite)
{
lwb_t *lwb;
lwb = kmem_cache_alloc(zil_lwb_cache, KM_SLEEP);
lwb->lwb_zilog = zilog;
lwb->lwb_blk = *bp;
lwb->lwb_fastwrite = fastwrite;
lwb->lwb_slog = slog;
lwb->lwb_state = LWB_STATE_CLOSED;
lwb->lwb_buf = zio_buf_alloc(BP_GET_LSIZE(bp));
lwb->lwb_max_txg = txg;
lwb->lwb_write_zio = NULL;
lwb->lwb_root_zio = NULL;
lwb->lwb_issued_timestamp = 0;
lwb->lwb_issued_txg = 0;
if (BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_ZILOG2) {
lwb->lwb_nused = sizeof (zil_chain_t);
lwb->lwb_sz = BP_GET_LSIZE(bp);
} else {
lwb->lwb_nused = 0;
lwb->lwb_sz = BP_GET_LSIZE(bp) - sizeof (zil_chain_t);
}
mutex_enter(&zilog->zl_lock);
list_insert_tail(&zilog->zl_lwb_list, lwb);
mutex_exit(&zilog->zl_lock);
ASSERT(!MUTEX_HELD(&lwb->lwb_vdev_lock));
ASSERT(avl_is_empty(&lwb->lwb_vdev_tree));
VERIFY(list_is_empty(&lwb->lwb_waiters));
VERIFY(list_is_empty(&lwb->lwb_itxs));
return (lwb);
}
static void
zil_free_lwb(zilog_t *zilog, lwb_t *lwb)
{
ASSERT(MUTEX_HELD(&zilog->zl_lock));
ASSERT(!MUTEX_HELD(&lwb->lwb_vdev_lock));
VERIFY(list_is_empty(&lwb->lwb_waiters));
VERIFY(list_is_empty(&lwb->lwb_itxs));
ASSERT(avl_is_empty(&lwb->lwb_vdev_tree));
ASSERT3P(lwb->lwb_write_zio, ==, NULL);
ASSERT3P(lwb->lwb_root_zio, ==, NULL);
ASSERT3U(lwb->lwb_max_txg, <=, spa_syncing_txg(zilog->zl_spa));
ASSERT(lwb->lwb_state == LWB_STATE_CLOSED ||
lwb->lwb_state == LWB_STATE_FLUSH_DONE);
/*
* Clear the zilog's field to indicate this lwb is no longer
* valid, and prevent use-after-free errors.
*/
if (zilog->zl_last_lwb_opened == lwb)
zilog->zl_last_lwb_opened = NULL;
kmem_cache_free(zil_lwb_cache, lwb);
}
/*
* Called when we create in-memory log transactions so that we know
* to cleanup the itxs at the end of spa_sync().
*/
static void
zilog_dirty(zilog_t *zilog, uint64_t txg)
{
dsl_pool_t *dp = zilog->zl_dmu_pool;
dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os);
ASSERT(spa_writeable(zilog->zl_spa));
if (ds->ds_is_snapshot)
panic("dirtying snapshot!");
if (txg_list_add(&dp->dp_dirty_zilogs, zilog, txg)) {
/* up the hold count until we can be written out */
dmu_buf_add_ref(ds->ds_dbuf, zilog);
zilog->zl_dirty_max_txg = MAX(txg, zilog->zl_dirty_max_txg);
}
}
/*
* Determine if the zil is dirty in the specified txg. Callers wanting to
* ensure that the dirty state does not change must hold the itxg_lock for
* the specified txg. Holding the lock will ensure that the zil cannot be
* dirtied (zil_itx_assign) or cleaned (zil_clean) while we check its current
* state.
*/
static boolean_t __maybe_unused
zilog_is_dirty_in_txg(zilog_t *zilog, uint64_t txg)
{
dsl_pool_t *dp = zilog->zl_dmu_pool;
if (txg_list_member(&dp->dp_dirty_zilogs, zilog, txg & TXG_MASK))
return (B_TRUE);
return (B_FALSE);
}
/*
* Determine if the zil is dirty. The zil is considered dirty if it has
* any pending itx records that have not been cleaned by zil_clean().
*/
static boolean_t
zilog_is_dirty(zilog_t *zilog)
{
dsl_pool_t *dp = zilog->zl_dmu_pool;
for (int t = 0; t < TXG_SIZE; t++) {
if (txg_list_member(&dp->dp_dirty_zilogs, zilog, t))
return (B_TRUE);
}
return (B_FALSE);
}
/*
* Its called in zil_commit context (zil_process_commit_list()/zil_create()).
* It activates SPA_FEATURE_ZILSAXATTR feature, if its enabled.
* Check dsl_dataset_feature_is_active to avoid txg_wait_synced() on every
* zil_commit.
*/
static void
zil_commit_activate_saxattr_feature(zilog_t *zilog)
{
dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os);
uint64_t txg = 0;
dmu_tx_t *tx = NULL;
if (spa_feature_is_enabled(zilog->zl_spa, SPA_FEATURE_ZILSAXATTR) &&
dmu_objset_type(zilog->zl_os) != DMU_OST_ZVOL &&
!dsl_dataset_feature_is_active(ds, SPA_FEATURE_ZILSAXATTR)) {
tx = dmu_tx_create(zilog->zl_os);
VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
dsl_dataset_dirty(ds, tx);
txg = dmu_tx_get_txg(tx);
mutex_enter(&ds->ds_lock);
ds->ds_feature_activation[SPA_FEATURE_ZILSAXATTR] =
(void *)B_TRUE;
mutex_exit(&ds->ds_lock);
dmu_tx_commit(tx);
txg_wait_synced(zilog->zl_dmu_pool, txg);
}
}
/*
* Create an on-disk intent log.
*/
static lwb_t *
zil_create(zilog_t *zilog)
{
const zil_header_t *zh = zilog->zl_header;
lwb_t *lwb = NULL;
uint64_t txg = 0;
dmu_tx_t *tx = NULL;
blkptr_t blk;
int error = 0;
boolean_t fastwrite = FALSE;
boolean_t slog = FALSE;
dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os);
/*
* Wait for any previous destroy to complete.
*/
txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg);
ASSERT(zh->zh_claim_txg == 0);
ASSERT(zh->zh_replay_seq == 0);
blk = zh->zh_log;
/*
* Allocate an initial log block if:
* - there isn't one already
* - the existing block is the wrong endianness
*/
if (BP_IS_HOLE(&blk) || BP_SHOULD_BYTESWAP(&blk)) {
tx = dmu_tx_create(zilog->zl_os);
VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
dsl_dataset_dirty(dmu_objset_ds(zilog->zl_os), tx);
txg = dmu_tx_get_txg(tx);
if (!BP_IS_HOLE(&blk)) {
zio_free(zilog->zl_spa, txg, &blk);
BP_ZERO(&blk);
}
error = zio_alloc_zil(zilog->zl_spa, zilog->zl_os, txg, &blk,
ZIL_MIN_BLKSZ, &slog);
fastwrite = TRUE;
if (error == 0)
zil_init_log_chain(zilog, &blk);
}
/*
* Allocate a log write block (lwb) for the first log block.
*/
if (error == 0)
lwb = zil_alloc_lwb(zilog, &blk, slog, txg, fastwrite);
/*
* If we just allocated the first log block, commit our transaction
* and wait for zil_sync() to stuff the block pointer into zh_log.
* (zh is part of the MOS, so we cannot modify it in open context.)
*/
if (tx != NULL) {
/*
* If "zilsaxattr" feature is enabled on zpool, then activate
* it now when we're creating the ZIL chain. We can't wait with
* this until we write the first xattr log record because we
* need to wait for the feature activation to sync out.
*/
if (spa_feature_is_enabled(zilog->zl_spa,
SPA_FEATURE_ZILSAXATTR) && dmu_objset_type(zilog->zl_os) !=
DMU_OST_ZVOL) {
mutex_enter(&ds->ds_lock);
ds->ds_feature_activation[SPA_FEATURE_ZILSAXATTR] =
(void *)B_TRUE;
mutex_exit(&ds->ds_lock);
}
dmu_tx_commit(tx);
txg_wait_synced(zilog->zl_dmu_pool, txg);
} else {
/*
* This branch covers the case where we enable the feature on a
* zpool that has existing ZIL headers.
*/
zil_commit_activate_saxattr_feature(zilog);
}
IMPLY(spa_feature_is_enabled(zilog->zl_spa, SPA_FEATURE_ZILSAXATTR) &&
dmu_objset_type(zilog->zl_os) != DMU_OST_ZVOL,
dsl_dataset_feature_is_active(ds, SPA_FEATURE_ZILSAXATTR));
ASSERT(error != 0 || memcmp(&blk, &zh->zh_log, sizeof (blk)) == 0);
IMPLY(error == 0, lwb != NULL);
return (lwb);
}
/*
* In one tx, free all log blocks and clear the log header. If keep_first
* is set, then we're replaying a log with no content. We want to keep the
* first block, however, so that the first synchronous transaction doesn't
* require a txg_wait_synced() in zil_create(). We don't need to
* txg_wait_synced() here either when keep_first is set, because both
* zil_create() and zil_destroy() will wait for any in-progress destroys
* to complete.
* Return B_TRUE if there were any entries to replay.
*/
boolean_t
zil_destroy(zilog_t *zilog, boolean_t keep_first)
{
const zil_header_t *zh = zilog->zl_header;
lwb_t *lwb;
dmu_tx_t *tx;
uint64_t txg;
/*
* Wait for any previous destroy to complete.
*/
txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg);
zilog->zl_old_header = *zh; /* debugging aid */
if (BP_IS_HOLE(&zh->zh_log))
return (B_FALSE);
tx = dmu_tx_create(zilog->zl_os);
VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
dsl_dataset_dirty(dmu_objset_ds(zilog->zl_os), tx);
txg = dmu_tx_get_txg(tx);
mutex_enter(&zilog->zl_lock);
ASSERT3U(zilog->zl_destroy_txg, <, txg);
zilog->zl_destroy_txg = txg;
zilog->zl_keep_first = keep_first;
if (!list_is_empty(&zilog->zl_lwb_list)) {
ASSERT(zh->zh_claim_txg == 0);
VERIFY(!keep_first);
while ((lwb = list_head(&zilog->zl_lwb_list)) != NULL) {
if (lwb->lwb_fastwrite)
metaslab_fastwrite_unmark(zilog->zl_spa,
&lwb->lwb_blk);
list_remove(&zilog->zl_lwb_list, lwb);
if (lwb->lwb_buf != NULL)
zio_buf_free(lwb->lwb_buf, lwb->lwb_sz);
zio_free(zilog->zl_spa, txg, &lwb->lwb_blk);
zil_free_lwb(zilog, lwb);
}
} else if (!keep_first) {
zil_destroy_sync(zilog, tx);
}
mutex_exit(&zilog->zl_lock);
dmu_tx_commit(tx);
return (B_TRUE);
}
void
zil_destroy_sync(zilog_t *zilog, dmu_tx_t *tx)
{
ASSERT(list_is_empty(&zilog->zl_lwb_list));
(void) zil_parse(zilog, zil_free_log_block,
zil_free_log_record, tx, zilog->zl_header->zh_claim_txg, B_FALSE);
}
int
zil_claim(dsl_pool_t *dp, dsl_dataset_t *ds, void *txarg)
{
dmu_tx_t *tx = txarg;
zilog_t *zilog;
uint64_t first_txg;
zil_header_t *zh;
objset_t *os;
int error;
error = dmu_objset_own_obj(dp, ds->ds_object,
DMU_OST_ANY, B_FALSE, B_FALSE, FTAG, &os);
if (error != 0) {
/*
* EBUSY indicates that the objset is inconsistent, in which
* case it can not have a ZIL.
*/
if (error != EBUSY) {
cmn_err(CE_WARN, "can't open objset for %llu, error %u",
(unsigned long long)ds->ds_object, error);
}
return (0);
}
zilog = dmu_objset_zil(os);
zh = zil_header_in_syncing_context(zilog);
ASSERT3U(tx->tx_txg, ==, spa_first_txg(zilog->zl_spa));
first_txg = spa_min_claim_txg(zilog->zl_spa);
/*
* If the spa_log_state is not set to be cleared, check whether
* the current uberblock is a checkpoint one and if the current
* header has been claimed before moving on.
*
* If the current uberblock is a checkpointed uberblock then
* one of the following scenarios took place:
*
* 1] We are currently rewinding to the checkpoint of the pool.
* 2] We crashed in the middle of a checkpoint rewind but we
* did manage to write the checkpointed uberblock to the
* vdev labels, so when we tried to import the pool again
* the checkpointed uberblock was selected from the import
* procedure.
*
* In both cases we want to zero out all the ZIL blocks, except
* the ones that have been claimed at the time of the checkpoint
* (their zh_claim_txg != 0). The reason is that these blocks
* may be corrupted since we may have reused their locations on
* disk after we took the checkpoint.
*
* We could try to set spa_log_state to SPA_LOG_CLEAR earlier
* when we first figure out whether the current uberblock is
* checkpointed or not. Unfortunately, that would discard all
* the logs, including the ones that are claimed, and we would
* leak space.
*/
if (spa_get_log_state(zilog->zl_spa) == SPA_LOG_CLEAR ||
(zilog->zl_spa->spa_uberblock.ub_checkpoint_txg != 0 &&
zh->zh_claim_txg == 0)) {
if (!BP_IS_HOLE(&zh->zh_log)) {
(void) zil_parse(zilog, zil_clear_log_block,
zil_noop_log_record, tx, first_txg, B_FALSE);
}
BP_ZERO(&zh->zh_log);
if (os->os_encrypted)
os->os_next_write_raw[tx->tx_txg & TXG_MASK] = B_TRUE;
dsl_dataset_dirty(dmu_objset_ds(os), tx);
dmu_objset_disown(os, B_FALSE, FTAG);
return (0);
}
/*
* If we are not rewinding and opening the pool normally, then
* the min_claim_txg should be equal to the first txg of the pool.
*/
ASSERT3U(first_txg, ==, spa_first_txg(zilog->zl_spa));
/*
* Claim all log blocks if we haven't already done so, and remember
* the highest claimed sequence number. This ensures that if we can
* read only part of the log now (e.g. due to a missing device),
* but we can read the entire log later, we will not try to replay
* or destroy beyond the last block we successfully claimed.
*/
ASSERT3U(zh->zh_claim_txg, <=, first_txg);
if (zh->zh_claim_txg == 0 && !BP_IS_HOLE(&zh->zh_log)) {
(void) zil_parse(zilog, zil_claim_log_block,
zil_claim_log_record, tx, first_txg, B_FALSE);
zh->zh_claim_txg = first_txg;
zh->zh_claim_blk_seq = zilog->zl_parse_blk_seq;
zh->zh_claim_lr_seq = zilog->zl_parse_lr_seq;
if (zilog->zl_parse_lr_count || zilog->zl_parse_blk_count > 1)
zh->zh_flags |= ZIL_REPLAY_NEEDED;
zh->zh_flags |= ZIL_CLAIM_LR_SEQ_VALID;
if (os->os_encrypted)
os->os_next_write_raw[tx->tx_txg & TXG_MASK] = B_TRUE;
dsl_dataset_dirty(dmu_objset_ds(os), tx);
}
ASSERT3U(first_txg, ==, (spa_last_synced_txg(zilog->zl_spa) + 1));
dmu_objset_disown(os, B_FALSE, FTAG);
return (0);
}
/*
* Check the log by walking the log chain.
* Checksum errors are ok as they indicate the end of the chain.
* Any other error (no device or read failure) returns an error.
*/
int
zil_check_log_chain(dsl_pool_t *dp, dsl_dataset_t *ds, void *tx)
{
(void) dp;
zilog_t *zilog;
objset_t *os;
blkptr_t *bp;
int error;
ASSERT(tx == NULL);
error = dmu_objset_from_ds(ds, &os);
if (error != 0) {
cmn_err(CE_WARN, "can't open objset %llu, error %d",
(unsigned long long)ds->ds_object, error);
return (0);
}
zilog = dmu_objset_zil(os);
bp = (blkptr_t *)&zilog->zl_header->zh_log;
if (!BP_IS_HOLE(bp)) {
vdev_t *vd;
boolean_t valid = B_TRUE;
/*
* Check the first block and determine if it's on a log device
* which may have been removed or faulted prior to loading this
* pool. If so, there's no point in checking the rest of the
* log as its content should have already been synced to the
* pool.
*/
spa_config_enter(os->os_spa, SCL_STATE, FTAG, RW_READER);
vd = vdev_lookup_top(os->os_spa, DVA_GET_VDEV(&bp->blk_dva[0]));
if (vd->vdev_islog && vdev_is_dead(vd))
valid = vdev_log_state_valid(vd);
spa_config_exit(os->os_spa, SCL_STATE, FTAG);
if (!valid)
return (0);
/*
* Check whether the current uberblock is checkpointed (e.g.
* we are rewinding) and whether the current header has been
* claimed or not. If it hasn't then skip verifying it. We
* do this because its ZIL blocks may be part of the pool's
* state before the rewind, which is no longer valid.
*/
zil_header_t *zh = zil_header_in_syncing_context(zilog);
if (zilog->zl_spa->spa_uberblock.ub_checkpoint_txg != 0 &&
zh->zh_claim_txg == 0)
return (0);
}
/*
* Because tx == NULL, zil_claim_log_block() will not actually claim
* any blocks, but just determine whether it is possible to do so.
* In addition to checking the log chain, zil_claim_log_block()
* will invoke zio_claim() with a done func of spa_claim_notify(),
* which will update spa_max_claim_txg. See spa_load() for details.
*/
error = zil_parse(zilog, zil_claim_log_block, zil_claim_log_record, tx,
zilog->zl_header->zh_claim_txg ? -1ULL :
spa_min_claim_txg(os->os_spa), B_FALSE);
return ((error == ECKSUM || error == ENOENT) ? 0 : error);
}
/*
* When an itx is "skipped", this function is used to properly mark the
* waiter as "done, and signal any thread(s) waiting on it. An itx can
* be skipped (and not committed to an lwb) for a variety of reasons,
* one of them being that the itx was committed via spa_sync(), prior to
* it being committed to an lwb; this can happen if a thread calling
* zil_commit() is racing with spa_sync().
*/
static void
zil_commit_waiter_skip(zil_commit_waiter_t *zcw)
{
mutex_enter(&zcw->zcw_lock);
ASSERT3B(zcw->zcw_done, ==, B_FALSE);
zcw->zcw_done = B_TRUE;
cv_broadcast(&zcw->zcw_cv);
mutex_exit(&zcw->zcw_lock);
}
/*
* This function is used when the given waiter is to be linked into an
* lwb's "lwb_waiter" list; i.e. when the itx is committed to the lwb.
* At this point, the waiter will no longer be referenced by the itx,
* and instead, will be referenced by the lwb.
*/
static void
zil_commit_waiter_link_lwb(zil_commit_waiter_t *zcw, lwb_t *lwb)
{
/*
* The lwb_waiters field of the lwb is protected by the zilog's
* zl_lock, thus it must be held when calling this function.
*/
ASSERT(MUTEX_HELD(&lwb->lwb_zilog->zl_lock));
mutex_enter(&zcw->zcw_lock);
ASSERT(!list_link_active(&zcw->zcw_node));
ASSERT3P(zcw->zcw_lwb, ==, NULL);
ASSERT3P(lwb, !=, NULL);
ASSERT(lwb->lwb_state == LWB_STATE_OPENED ||
lwb->lwb_state == LWB_STATE_ISSUED ||
lwb->lwb_state == LWB_STATE_WRITE_DONE);
list_insert_tail(&lwb->lwb_waiters, zcw);
zcw->zcw_lwb = lwb;
mutex_exit(&zcw->zcw_lock);
}
/*
* This function is used when zio_alloc_zil() fails to allocate a ZIL
* block, and the given waiter must be linked to the "nolwb waiters"
* list inside of zil_process_commit_list().
*/
static void
zil_commit_waiter_link_nolwb(zil_commit_waiter_t *zcw, list_t *nolwb)
{
mutex_enter(&zcw->zcw_lock);
ASSERT(!list_link_active(&zcw->zcw_node));
ASSERT3P(zcw->zcw_lwb, ==, NULL);
list_insert_tail(nolwb, zcw);
mutex_exit(&zcw->zcw_lock);
}
void
zil_lwb_add_block(lwb_t *lwb, const blkptr_t *bp)
{
avl_tree_t *t = &lwb->lwb_vdev_tree;
avl_index_t where;
zil_vdev_node_t *zv, zvsearch;
int ndvas = BP_GET_NDVAS(bp);
int i;
if (zil_nocacheflush)
return;
mutex_enter(&lwb->lwb_vdev_lock);
for (i = 0; i < ndvas; i++) {
zvsearch.zv_vdev = DVA_GET_VDEV(&bp->blk_dva[i]);
if (avl_find(t, &zvsearch, &where) == NULL) {
zv = kmem_alloc(sizeof (*zv), KM_SLEEP);
zv->zv_vdev = zvsearch.zv_vdev;
avl_insert(t, zv, where);
}
}
mutex_exit(&lwb->lwb_vdev_lock);
}
static void
zil_lwb_flush_defer(lwb_t *lwb, lwb_t *nlwb)
{
avl_tree_t *src = &lwb->lwb_vdev_tree;
avl_tree_t *dst = &nlwb->lwb_vdev_tree;
void *cookie = NULL;
zil_vdev_node_t *zv;
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_WRITE_DONE);
ASSERT3S(nlwb->lwb_state, !=, LWB_STATE_WRITE_DONE);
ASSERT3S(nlwb->lwb_state, !=, LWB_STATE_FLUSH_DONE);
/*
* While 'lwb' is at a point in its lifetime where lwb_vdev_tree does
* not need the protection of lwb_vdev_lock (it will only be modified
* while holding zilog->zl_lock) as its writes and those of its
* children have all completed. The younger 'nlwb' may be waiting on
* future writes to additional vdevs.
*/
mutex_enter(&nlwb->lwb_vdev_lock);
/*
* Tear down the 'lwb' vdev tree, ensuring that entries which do not
* exist in 'nlwb' are moved to it, freeing any would-be duplicates.
*/
while ((zv = avl_destroy_nodes(src, &cookie)) != NULL) {
avl_index_t where;
if (avl_find(dst, zv, &where) == NULL) {
avl_insert(dst, zv, where);
} else {
kmem_free(zv, sizeof (*zv));
}
}
mutex_exit(&nlwb->lwb_vdev_lock);
}
void
zil_lwb_add_txg(lwb_t *lwb, uint64_t txg)
{
lwb->lwb_max_txg = MAX(lwb->lwb_max_txg, txg);
}
/*
* This function is a called after all vdevs associated with a given lwb
* write have completed their DKIOCFLUSHWRITECACHE command; or as soon
* as the lwb write completes, if "zil_nocacheflush" is set. Further,
* all "previous" lwb's will have completed before this function is
* called; i.e. this function is called for all previous lwbs before
* it's called for "this" lwb (enforced via zio the dependencies
* configured in zil_lwb_set_zio_dependency()).
*
* The intention is for this function to be called as soon as the
* contents of an lwb are considered "stable" on disk, and will survive
* any sudden loss of power. At this point, any threads waiting for the
* lwb to reach this state are signalled, and the "waiter" structures
* are marked "done".
*/
static void
zil_lwb_flush_vdevs_done(zio_t *zio)
{
lwb_t *lwb = zio->io_private;
zilog_t *zilog = lwb->lwb_zilog;
zil_commit_waiter_t *zcw;
itx_t *itx;
uint64_t txg;
spa_config_exit(zilog->zl_spa, SCL_STATE, lwb);
zio_buf_free(lwb->lwb_buf, lwb->lwb_sz);
mutex_enter(&zilog->zl_lock);
/*
* If we have had an allocation failure and the txg is
* waiting to sync then we want zil_sync() to remove the lwb so
* that it's not picked up as the next new one in
* zil_process_commit_list(). zil_sync() will only remove the
* lwb if lwb_buf is null.
*/
lwb->lwb_buf = NULL;
ASSERT3U(lwb->lwb_issued_timestamp, >, 0);
zilog->zl_last_lwb_latency = (zilog->zl_last_lwb_latency * 3 +
gethrtime() - lwb->lwb_issued_timestamp) / 4;
lwb->lwb_root_zio = NULL;
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_WRITE_DONE);
lwb->lwb_state = LWB_STATE_FLUSH_DONE;
if (zilog->zl_last_lwb_opened == lwb) {
/*
* Remember the highest committed log sequence number
* for ztest. We only update this value when all the log
* writes succeeded, because ztest wants to ASSERT that
* it got the whole log chain.
*/
zilog->zl_commit_lr_seq = zilog->zl_lr_seq;
}
while ((itx = list_head(&lwb->lwb_itxs)) != NULL) {
list_remove(&lwb->lwb_itxs, itx);
zil_itx_destroy(itx);
}
while ((zcw = list_head(&lwb->lwb_waiters)) != NULL) {
mutex_enter(&zcw->zcw_lock);
ASSERT(list_link_active(&zcw->zcw_node));
list_remove(&lwb->lwb_waiters, zcw);
ASSERT3P(zcw->zcw_lwb, ==, lwb);
zcw->zcw_lwb = NULL;
/*
* We expect any ZIO errors from child ZIOs to have been
* propagated "up" to this specific LWB's root ZIO, in
* order for this error handling to work correctly. This
* includes ZIO errors from either this LWB's write or
* flush, as well as any errors from other dependent LWBs
* (e.g. a root LWB ZIO that might be a child of this LWB).
*
* With that said, it's important to note that LWB flush
* errors are not propagated up to the LWB root ZIO.
* This is incorrect behavior, and results in VDEV flush
* errors not being handled correctly here. See the
* comment above the call to "zio_flush" for details.
*/
zcw->zcw_zio_error = zio->io_error;
ASSERT3B(zcw->zcw_done, ==, B_FALSE);
zcw->zcw_done = B_TRUE;
cv_broadcast(&zcw->zcw_cv);
mutex_exit(&zcw->zcw_lock);
}
mutex_exit(&zilog->zl_lock);
mutex_enter(&zilog->zl_lwb_io_lock);
txg = lwb->lwb_issued_txg;
ASSERT3U(zilog->zl_lwb_inflight[txg & TXG_MASK], >, 0);
zilog->zl_lwb_inflight[txg & TXG_MASK]--;
if (zilog->zl_lwb_inflight[txg & TXG_MASK] == 0)
cv_broadcast(&zilog->zl_lwb_io_cv);
mutex_exit(&zilog->zl_lwb_io_lock);
}
/*
* Wait for the completion of all issued write/flush of that txg provided.
* It guarantees zil_lwb_flush_vdevs_done() is called and returned.
*/
static void
zil_lwb_flush_wait_all(zilog_t *zilog, uint64_t txg)
{
ASSERT3U(txg, ==, spa_syncing_txg(zilog->zl_spa));
mutex_enter(&zilog->zl_lwb_io_lock);
while (zilog->zl_lwb_inflight[txg & TXG_MASK] > 0)
cv_wait(&zilog->zl_lwb_io_cv, &zilog->zl_lwb_io_lock);
mutex_exit(&zilog->zl_lwb_io_lock);
#ifdef ZFS_DEBUG
mutex_enter(&zilog->zl_lock);
mutex_enter(&zilog->zl_lwb_io_lock);
lwb_t *lwb = list_head(&zilog->zl_lwb_list);
while (lwb != NULL && lwb->lwb_max_txg <= txg) {
if (lwb->lwb_issued_txg <= txg) {
ASSERT(lwb->lwb_state != LWB_STATE_ISSUED);
ASSERT(lwb->lwb_state != LWB_STATE_WRITE_DONE);
IMPLY(lwb->lwb_issued_txg > 0,
lwb->lwb_state == LWB_STATE_FLUSH_DONE);
}
IMPLY(lwb->lwb_state == LWB_STATE_FLUSH_DONE,
lwb->lwb_buf == NULL);
lwb = list_next(&zilog->zl_lwb_list, lwb);
}
mutex_exit(&zilog->zl_lwb_io_lock);
mutex_exit(&zilog->zl_lock);
#endif
}
/*
* This is called when an lwb's write zio completes. The callback's
* purpose is to issue the DKIOCFLUSHWRITECACHE commands for the vdevs
* in the lwb's lwb_vdev_tree. The tree will contain the vdevs involved
* in writing out this specific lwb's data, and in the case that cache
* flushes have been deferred, vdevs involved in writing the data for
* previous lwbs. The writes corresponding to all the vdevs in the
* lwb_vdev_tree will have completed by the time this is called, due to
* the zio dependencies configured in zil_lwb_set_zio_dependency(),
* which takes deferred flushes into account. The lwb will be "done"
* once zil_lwb_flush_vdevs_done() is called, which occurs in the zio
* completion callback for the lwb's root zio.
*/
static void
zil_lwb_write_done(zio_t *zio)
{
lwb_t *lwb = zio->io_private;
spa_t *spa = zio->io_spa;
zilog_t *zilog = lwb->lwb_zilog;
avl_tree_t *t = &lwb->lwb_vdev_tree;
void *cookie = NULL;
zil_vdev_node_t *zv;
lwb_t *nlwb;
ASSERT3S(spa_config_held(spa, SCL_STATE, RW_READER), !=, 0);
ASSERT(BP_GET_COMPRESS(zio->io_bp) == ZIO_COMPRESS_OFF);
ASSERT(BP_GET_TYPE(zio->io_bp) == DMU_OT_INTENT_LOG);
ASSERT(BP_GET_LEVEL(zio->io_bp) == 0);
ASSERT(BP_GET_BYTEORDER(zio->io_bp) == ZFS_HOST_BYTEORDER);
ASSERT(!BP_IS_GANG(zio->io_bp));
ASSERT(!BP_IS_HOLE(zio->io_bp));
ASSERT(BP_GET_FILL(zio->io_bp) == 0);
abd_free(zio->io_abd);
mutex_enter(&zilog->zl_lock);
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_ISSUED);
lwb->lwb_state = LWB_STATE_WRITE_DONE;
lwb->lwb_write_zio = NULL;
lwb->lwb_fastwrite = FALSE;
nlwb = list_next(&zilog->zl_lwb_list, lwb);
mutex_exit(&zilog->zl_lock);
if (avl_numnodes(t) == 0)
return;
/*
* If there was an IO error, we're not going to call zio_flush()
* on these vdevs, so we simply empty the tree and free the
* nodes. We avoid calling zio_flush() since there isn't any
* good reason for doing so, after the lwb block failed to be
* written out.
*
* Additionally, we don't perform any further error handling at
* this point (e.g. setting "zcw_zio_error" appropriately), as
* we expect that to occur in "zil_lwb_flush_vdevs_done" (thus,
* we expect any error seen here, to have been propagated to
* that function).
*/
if (zio->io_error != 0) {
while ((zv = avl_destroy_nodes(t, &cookie)) != NULL)
kmem_free(zv, sizeof (*zv));
return;
}
/*
* If this lwb does not have any threads waiting for it to
* complete, we want to defer issuing the DKIOCFLUSHWRITECACHE
* command to the vdevs written to by "this" lwb, and instead
* rely on the "next" lwb to handle the DKIOCFLUSHWRITECACHE
* command for those vdevs. Thus, we merge the vdev tree of
* "this" lwb with the vdev tree of the "next" lwb in the list,
* and assume the "next" lwb will handle flushing the vdevs (or
* deferring the flush(s) again).
*
* This is a useful performance optimization, especially for
* workloads with lots of async write activity and few sync
* write and/or fsync activity, as it has the potential to
* coalesce multiple flush commands to a vdev into one.
*/
if (list_head(&lwb->lwb_waiters) == NULL && nlwb != NULL) {
zil_lwb_flush_defer(lwb, nlwb);
ASSERT(avl_is_empty(&lwb->lwb_vdev_tree));
return;
}
while ((zv = avl_destroy_nodes(t, &cookie)) != NULL) {
vdev_t *vd = vdev_lookup_top(spa, zv->zv_vdev);
if (vd != NULL) {
/*
* The "ZIO_FLAG_DONT_PROPAGATE" is currently
* always used within "zio_flush". This means,
* any errors when flushing the vdev(s), will
* (unfortunately) not be handled correctly,
* since these "zio_flush" errors will not be
* propagated up to "zil_lwb_flush_vdevs_done".
*/
zio_flush(lwb->lwb_root_zio, vd);
}
kmem_free(zv, sizeof (*zv));
}
}
static void
zil_lwb_set_zio_dependency(zilog_t *zilog, lwb_t *lwb)
{
lwb_t *last_lwb_opened = zilog->zl_last_lwb_opened;
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
ASSERT(MUTEX_HELD(&zilog->zl_lock));
/*
* The zilog's "zl_last_lwb_opened" field is used to build the
* lwb/zio dependency chain, which is used to preserve the
* ordering of lwb completions that is required by the semantics
* of the ZIL. Each new lwb zio becomes a parent of the
* "previous" lwb zio, such that the new lwb's zio cannot
* complete until the "previous" lwb's zio completes.
*
* This is required by the semantics of zil_commit(); the commit
* waiters attached to the lwbs will be woken in the lwb zio's
* completion callback, so this zio dependency graph ensures the
* waiters are woken in the correct order (the same order the
* lwbs were created).
*/
if (last_lwb_opened != NULL &&
last_lwb_opened->lwb_state != LWB_STATE_FLUSH_DONE) {
ASSERT(last_lwb_opened->lwb_state == LWB_STATE_OPENED ||
last_lwb_opened->lwb_state == LWB_STATE_ISSUED ||
last_lwb_opened->lwb_state == LWB_STATE_WRITE_DONE);
ASSERT3P(last_lwb_opened->lwb_root_zio, !=, NULL);
zio_add_child(lwb->lwb_root_zio,
last_lwb_opened->lwb_root_zio);
/*
* If the previous lwb's write hasn't already completed,
* we also want to order the completion of the lwb write
* zios (above, we only order the completion of the lwb
* root zios). This is required because of how we can
* defer the DKIOCFLUSHWRITECACHE commands for each lwb.
*
* When the DKIOCFLUSHWRITECACHE commands are deferred,
* the previous lwb will rely on this lwb to flush the
* vdevs written to by that previous lwb. Thus, we need
* to ensure this lwb doesn't issue the flush until
* after the previous lwb's write completes. We ensure
* this ordering by setting the zio parent/child
* relationship here.
*
* Without this relationship on the lwb's write zio,
* it's possible for this lwb's write to complete prior
* to the previous lwb's write completing; and thus, the
* vdevs for the previous lwb would be flushed prior to
* that lwb's data being written to those vdevs (the
* vdevs are flushed in the lwb write zio's completion
* handler, zil_lwb_write_done()).
*/
if (last_lwb_opened->lwb_state != LWB_STATE_WRITE_DONE) {
ASSERT(last_lwb_opened->lwb_state == LWB_STATE_OPENED ||
last_lwb_opened->lwb_state == LWB_STATE_ISSUED);
ASSERT3P(last_lwb_opened->lwb_write_zio, !=, NULL);
zio_add_child(lwb->lwb_write_zio,
last_lwb_opened->lwb_write_zio);
}
}
}
/*
* This function's purpose is to "open" an lwb such that it is ready to
* accept new itxs being committed to it. To do this, the lwb's zio
* structures are created, and linked to the lwb. This function is
* idempotent; if the passed in lwb has already been opened, this
* function is essentially a no-op.
*/
static void
zil_lwb_write_open(zilog_t *zilog, lwb_t *lwb)
{
zbookmark_phys_t zb;
zio_priority_t prio;
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
ASSERT3P(lwb, !=, NULL);
EQUIV(lwb->lwb_root_zio == NULL, lwb->lwb_state == LWB_STATE_CLOSED);
EQUIV(lwb->lwb_root_zio != NULL, lwb->lwb_state == LWB_STATE_OPENED);
SET_BOOKMARK(&zb, lwb->lwb_blk.blk_cksum.zc_word[ZIL_ZC_OBJSET],
ZB_ZIL_OBJECT, ZB_ZIL_LEVEL,
lwb->lwb_blk.blk_cksum.zc_word[ZIL_ZC_SEQ]);
/* Lock so zil_sync() doesn't fastwrite_unmark after zio is created */
mutex_enter(&zilog->zl_lock);
if (lwb->lwb_root_zio == NULL) {
abd_t *lwb_abd = abd_get_from_buf(lwb->lwb_buf,
BP_GET_LSIZE(&lwb->lwb_blk));
if (!lwb->lwb_fastwrite) {
metaslab_fastwrite_mark(zilog->zl_spa, &lwb->lwb_blk);
lwb->lwb_fastwrite = 1;
}
if (!lwb->lwb_slog || zilog->zl_cur_used <= zil_slog_bulk)
prio = ZIO_PRIORITY_SYNC_WRITE;
else
prio = ZIO_PRIORITY_ASYNC_WRITE;
lwb->lwb_root_zio = zio_root(zilog->zl_spa,
zil_lwb_flush_vdevs_done, lwb, ZIO_FLAG_CANFAIL);
ASSERT3P(lwb->lwb_root_zio, !=, NULL);
lwb->lwb_write_zio = zio_rewrite(lwb->lwb_root_zio,
zilog->zl_spa, 0, &lwb->lwb_blk, lwb_abd,
BP_GET_LSIZE(&lwb->lwb_blk), zil_lwb_write_done, lwb,
prio, ZIO_FLAG_CANFAIL | ZIO_FLAG_FASTWRITE, &zb);
ASSERT3P(lwb->lwb_write_zio, !=, NULL);
lwb->lwb_state = LWB_STATE_OPENED;
zil_lwb_set_zio_dependency(zilog, lwb);
zilog->zl_last_lwb_opened = lwb;
}
mutex_exit(&zilog->zl_lock);
ASSERT3P(lwb->lwb_root_zio, !=, NULL);
ASSERT3P(lwb->lwb_write_zio, !=, NULL);
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_OPENED);
}
/*
* Define a limited set of intent log block sizes.
*
* These must be a multiple of 4KB. Note only the amount used (again
* aligned to 4KB) actually gets written. However, we can't always just
* allocate SPA_OLD_MAXBLOCKSIZE as the slog space could be exhausted.
*/
static const struct {
uint64_t limit;
uint64_t blksz;
} zil_block_buckets[] = {
{ 4096, 4096 }, /* non TX_WRITE */
{ 8192 + 4096, 8192 + 4096 }, /* database */
{ 32768 + 4096, 32768 + 4096 }, /* NFS writes */
{ 65536 + 4096, 65536 + 4096 }, /* 64KB writes */
{ 131072, 131072 }, /* < 128KB writes */
{ 131072 +4096, 65536 + 4096 }, /* 128KB writes */
{ UINT64_MAX, SPA_OLD_MAXBLOCKSIZE}, /* > 128KB writes */
};
/*
* Maximum block size used by the ZIL. This is picked up when the ZIL is
* initialized. Otherwise this should not be used directly; see
* zl_max_block_size instead.
*/
static uint_t zil_maxblocksize = SPA_OLD_MAXBLOCKSIZE;
/*
* Start a log block write and advance to the next log block.
* Calls are serialized.
*/
static lwb_t *
zil_lwb_write_issue(zilog_t *zilog, lwb_t *lwb)
{
lwb_t *nlwb = NULL;
zil_chain_t *zilc;
spa_t *spa = zilog->zl_spa;
blkptr_t *bp;
dmu_tx_t *tx;
uint64_t txg;
uint64_t zil_blksz, wsz;
int i, error;
boolean_t slog;
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
ASSERT3P(lwb->lwb_root_zio, !=, NULL);
ASSERT3P(lwb->lwb_write_zio, !=, NULL);
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_OPENED);
if (BP_GET_CHECKSUM(&lwb->lwb_blk) == ZIO_CHECKSUM_ZILOG2) {
zilc = (zil_chain_t *)lwb->lwb_buf;
bp = &zilc->zc_next_blk;
} else {
zilc = (zil_chain_t *)(lwb->lwb_buf + lwb->lwb_sz);
bp = &zilc->zc_next_blk;
}
ASSERT(lwb->lwb_nused <= lwb->lwb_sz);
/*
* Allocate the next block and save its address in this block
* before writing it in order to establish the log chain.
*/
tx = dmu_tx_create(zilog->zl_os);
/*
* Since we are not going to create any new dirty data, and we
* can even help with clearing the existing dirty data, we
* should not be subject to the dirty data based delays. We
* use TXG_NOTHROTTLE to bypass the delay mechanism.
*/
VERIFY0(dmu_tx_assign(tx, TXG_WAIT | TXG_NOTHROTTLE));
dsl_dataset_dirty(dmu_objset_ds(zilog->zl_os), tx);
txg = dmu_tx_get_txg(tx);
mutex_enter(&zilog->zl_lwb_io_lock);
lwb->lwb_issued_txg = txg;
zilog->zl_lwb_inflight[txg & TXG_MASK]++;
zilog->zl_lwb_max_issued_txg = MAX(txg, zilog->zl_lwb_max_issued_txg);
mutex_exit(&zilog->zl_lwb_io_lock);
/*
* Log blocks are pre-allocated. Here we select the size of the next
* block, based on size used in the last block.
* - first find the smallest bucket that will fit the block from a
* limited set of block sizes. This is because it's faster to write
* blocks allocated from the same metaslab as they are adjacent or
* close.
* - next find the maximum from the new suggested size and an array of
* previous sizes. This lessens a picket fence effect of wrongly
* guessing the size if we have a stream of say 2k, 64k, 2k, 64k
* requests.
*
* Note we only write what is used, but we can't just allocate
* the maximum block size because we can exhaust the available
* pool log space.
*/
zil_blksz = zilog->zl_cur_used + sizeof (zil_chain_t);
for (i = 0; zil_blksz > zil_block_buckets[i].limit; i++)
continue;
zil_blksz = MIN(zil_block_buckets[i].blksz, zilog->zl_max_block_size);
zilog->zl_prev_blks[zilog->zl_prev_rotor] = zil_blksz;
for (i = 0; i < ZIL_PREV_BLKS; i++)
zil_blksz = MAX(zil_blksz, zilog->zl_prev_blks[i]);
zilog->zl_prev_rotor = (zilog->zl_prev_rotor + 1) & (ZIL_PREV_BLKS - 1);
BP_ZERO(bp);
error = zio_alloc_zil(spa, zilog->zl_os, txg, bp, zil_blksz, &slog);
if (slog) {
ZIL_STAT_BUMP(zilog, zil_itx_metaslab_slog_count);
ZIL_STAT_INCR(zilog, zil_itx_metaslab_slog_bytes,
lwb->lwb_nused);
} else {
ZIL_STAT_BUMP(zilog, zil_itx_metaslab_normal_count);
ZIL_STAT_INCR(zilog, zil_itx_metaslab_normal_bytes,
lwb->lwb_nused);
}
if (error == 0) {
ASSERT3U(bp->blk_birth, ==, txg);
bp->blk_cksum = lwb->lwb_blk.blk_cksum;
bp->blk_cksum.zc_word[ZIL_ZC_SEQ]++;
/*
* Allocate a new log write block (lwb).
*/
nlwb = zil_alloc_lwb(zilog, bp, slog, txg, TRUE);
}
if (BP_GET_CHECKSUM(&lwb->lwb_blk) == ZIO_CHECKSUM_ZILOG2) {
/* For Slim ZIL only write what is used. */
wsz = P2ROUNDUP_TYPED(lwb->lwb_nused, ZIL_MIN_BLKSZ, uint64_t);
ASSERT3U(wsz, <=, lwb->lwb_sz);
zio_shrink(lwb->lwb_write_zio, wsz);
} else {
wsz = lwb->lwb_sz;
}
zilc->zc_pad = 0;
zilc->zc_nused = lwb->lwb_nused;
zilc->zc_eck.zec_cksum = lwb->lwb_blk.blk_cksum;
/*
* clear unused data for security
*/
memset(lwb->lwb_buf + lwb->lwb_nused, 0, wsz - lwb->lwb_nused);
spa_config_enter(zilog->zl_spa, SCL_STATE, lwb, RW_READER);
zil_lwb_add_block(lwb, &lwb->lwb_blk);
lwb->lwb_issued_timestamp = gethrtime();
lwb->lwb_state = LWB_STATE_ISSUED;
zio_nowait(lwb->lwb_root_zio);
zio_nowait(lwb->lwb_write_zio);
dmu_tx_commit(tx);
/*
* If there was an allocation failure then nlwb will be null which
* forces a txg_wait_synced().
*/
return (nlwb);
}
/*
* Maximum amount of data that can be put into single log block.
*/
uint64_t
zil_max_log_data(zilog_t *zilog, size_t hdrsize)
{
return (zilog->zl_max_block_size - sizeof (zil_chain_t) - hdrsize);
}
/*
* Maximum amount of log space we agree to waste to reduce number of
* WR_NEED_COPY chunks to reduce zl_get_data() overhead (~12%).
*/
static inline uint64_t
zil_max_waste_space(zilog_t *zilog)
{
return (zil_max_log_data(zilog, sizeof (lr_write_t)) / 8);
}
/*
* Maximum amount of write data for WR_COPIED. For correctness, consumers
* must fall back to WR_NEED_COPY if we can't fit the entire record into one
* maximum sized log block, because each WR_COPIED record must fit in a
* single log block. For space efficiency, we want to fit two records into a
* max-sized log block.
*/
uint64_t
zil_max_copied_data(zilog_t *zilog)
{
return ((zilog->zl_max_block_size - sizeof (zil_chain_t)) / 2 -
sizeof (lr_write_t));
}
static lwb_t *
zil_lwb_commit(zilog_t *zilog, itx_t *itx, lwb_t *lwb)
{
lr_t *lrcb, *lrc;
lr_write_t *lrwb, *lrw;
char *lr_buf;
uint64_t dlen, dnow, dpad, lwb_sp, reclen, txg, max_log_data;
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
ASSERT3P(lwb, !=, NULL);
ASSERT3P(lwb->lwb_buf, !=, NULL);
zil_lwb_write_open(zilog, lwb);
lrc = &itx->itx_lr;
lrw = (lr_write_t *)lrc;
/*
* A commit itx doesn't represent any on-disk state; instead
* it's simply used as a place holder on the commit list, and
* provides a mechanism for attaching a "commit waiter" onto the
* correct lwb (such that the waiter can be signalled upon
* completion of that lwb). Thus, we don't process this itx's
* log record if it's a commit itx (these itx's don't have log
* records), and instead link the itx's waiter onto the lwb's
* list of waiters.
*
* For more details, see the comment above zil_commit().
*/
if (lrc->lrc_txtype == TX_COMMIT) {
mutex_enter(&zilog->zl_lock);
zil_commit_waiter_link_lwb(itx->itx_private, lwb);
itx->itx_private = NULL;
mutex_exit(&zilog->zl_lock);
return (lwb);
}
if (lrc->lrc_txtype == TX_WRITE && itx->itx_wr_state == WR_NEED_COPY) {
dlen = P2ROUNDUP_TYPED(
lrw->lr_length, sizeof (uint64_t), uint64_t);
dpad = dlen - lrw->lr_length;
} else {
dlen = dpad = 0;
}
reclen = lrc->lrc_reclen;
zilog->zl_cur_used += (reclen + dlen);
txg = lrc->lrc_txg;
ASSERT3U(zilog->zl_cur_used, <, UINT64_MAX - (reclen + dlen));
cont:
/*
* If this record won't fit in the current log block, start a new one.
* For WR_NEED_COPY optimize layout for minimal number of chunks.
*/
lwb_sp = lwb->lwb_sz - lwb->lwb_nused;
max_log_data = zil_max_log_data(zilog, sizeof (lr_write_t));
if (reclen > lwb_sp || (reclen + dlen > lwb_sp &&
lwb_sp < zil_max_waste_space(zilog) &&
(dlen % max_log_data == 0 ||
lwb_sp < reclen + dlen % max_log_data))) {
lwb = zil_lwb_write_issue(zilog, lwb);
if (lwb == NULL)
return (NULL);
zil_lwb_write_open(zilog, lwb);
ASSERT(LWB_EMPTY(lwb));
lwb_sp = lwb->lwb_sz - lwb->lwb_nused;
/*
* There must be enough space in the new, empty log block to
* hold reclen. For WR_COPIED, we need to fit the whole
* record in one block, and reclen is the header size + the
* data size. For WR_NEED_COPY, we can create multiple
* records, splitting the data into multiple blocks, so we
* only need to fit one word of data per block; in this case
* reclen is just the header size (no data).
*/
ASSERT3U(reclen + MIN(dlen, sizeof (uint64_t)), <=, lwb_sp);
}
dnow = MIN(dlen, lwb_sp - reclen);
lr_buf = lwb->lwb_buf + lwb->lwb_nused;
memcpy(lr_buf, lrc, reclen);
lrcb = (lr_t *)lr_buf; /* Like lrc, but inside lwb. */
lrwb = (lr_write_t *)lrcb; /* Like lrw, but inside lwb. */
ZIL_STAT_BUMP(zilog, zil_itx_count);
/*
* If it's a write, fetch the data or get its blkptr as appropriate.
*/
if (lrc->lrc_txtype == TX_WRITE) {
if (txg > spa_freeze_txg(zilog->zl_spa))
txg_wait_synced(zilog->zl_dmu_pool, txg);
if (itx->itx_wr_state == WR_COPIED) {
ZIL_STAT_BUMP(zilog, zil_itx_copied_count);
ZIL_STAT_INCR(zilog, zil_itx_copied_bytes,
lrw->lr_length);
} else {
char *dbuf;
int error;
if (itx->itx_wr_state == WR_NEED_COPY) {
dbuf = lr_buf + reclen;
lrcb->lrc_reclen += dnow;
if (lrwb->lr_length > dnow)
lrwb->lr_length = dnow;
lrw->lr_offset += dnow;
lrw->lr_length -= dnow;
ZIL_STAT_BUMP(zilog, zil_itx_needcopy_count);
ZIL_STAT_INCR(zilog, zil_itx_needcopy_bytes,
dnow);
} else {
ASSERT3S(itx->itx_wr_state, ==, WR_INDIRECT);
dbuf = NULL;
ZIL_STAT_BUMP(zilog, zil_itx_indirect_count);
ZIL_STAT_INCR(zilog, zil_itx_indirect_bytes,
lrw->lr_length);
}
/*
* We pass in the "lwb_write_zio" rather than
* "lwb_root_zio" so that the "lwb_write_zio"
* becomes the parent of any zio's created by
* the "zl_get_data" callback. The vdevs are
* flushed after the "lwb_write_zio" completes,
* so we want to make sure that completion
* callback waits for these additional zio's,
* such that the vdevs used by those zio's will
* be included in the lwb's vdev tree, and those
* vdevs will be properly flushed. If we passed
* in "lwb_root_zio" here, then these additional
* vdevs may not be flushed; e.g. if these zio's
* completed after "lwb_write_zio" completed.
*/
error = zilog->zl_get_data(itx->itx_private,
itx->itx_gen, lrwb, dbuf, lwb,
lwb->lwb_write_zio);
if (dbuf != NULL && error == 0 && dnow == dlen)
/* Zero any padding bytes in the last block. */
memset((char *)dbuf + lrwb->lr_length, 0, dpad);
/*
* Typically, the only return values we should see from
* ->zl_get_data() are 0, EIO, ENOENT, EEXIST or
* EALREADY. However, it is also possible to see other
* error values such as ENOSPC or EINVAL from
* dmu_read() -> dnode_hold() -> dnode_hold_impl() or
* ENXIO as well as a multitude of others from the
* block layer through dmu_buf_hold() -> dbuf_read()
* -> zio_wait(), as well as through dmu_read() ->
* dnode_hold() -> dnode_hold_impl() -> dbuf_read() ->
* zio_wait(). When these errors happen, we can assume
* that neither an immediate write nor an indirect
* write occurred, so we need to fall back to
* txg_wait_synced(). This is unusual, so we print to
* dmesg whenever one of these errors occurs.
*/
switch (error) {
case 0:
break;
default:
cmn_err(CE_WARN, "zil_lwb_commit() received "
"unexpected error %d from ->zl_get_data()"
". Falling back to txg_wait_synced().",
error);
zfs_fallthrough;
case EIO:
txg_wait_synced(zilog->zl_dmu_pool, txg);
zfs_fallthrough;
case ENOENT:
zfs_fallthrough;
case EEXIST:
zfs_fallthrough;
case EALREADY:
return (lwb);
}
}
}
/*
* We're actually making an entry, so update lrc_seq to be the
* log record sequence number. Note that this is generally not
* equal to the itx sequence number because not all transactions
* are synchronous, and sometimes spa_sync() gets there first.
*/
lrcb->lrc_seq = ++zilog->zl_lr_seq;
lwb->lwb_nused += reclen + dnow;
zil_lwb_add_txg(lwb, txg);
ASSERT3U(lwb->lwb_nused, <=, lwb->lwb_sz);
ASSERT0(P2PHASE(lwb->lwb_nused, sizeof (uint64_t)));
dlen -= dnow;
if (dlen > 0) {
zilog->zl_cur_used += reclen;
goto cont;
}
return (lwb);
}
itx_t *
zil_itx_create(uint64_t txtype, size_t olrsize)
{
size_t itxsize, lrsize;
itx_t *itx;
lrsize = P2ROUNDUP_TYPED(olrsize, sizeof (uint64_t), size_t);
itxsize = offsetof(itx_t, itx_lr) + lrsize;
itx = zio_data_buf_alloc(itxsize);
itx->itx_lr.lrc_txtype = txtype;
itx->itx_lr.lrc_reclen = lrsize;
itx->itx_lr.lrc_seq = 0; /* defensive */
memset((char *)&itx->itx_lr + olrsize, 0, lrsize - olrsize);
itx->itx_sync = B_TRUE; /* default is synchronous */
itx->itx_callback = NULL;
itx->itx_callback_data = NULL;
itx->itx_size = itxsize;
return (itx);
}
void
zil_itx_destroy(itx_t *itx)
{
IMPLY(itx->itx_lr.lrc_txtype == TX_COMMIT, itx->itx_callback == NULL);
IMPLY(itx->itx_callback != NULL, itx->itx_lr.lrc_txtype != TX_COMMIT);
if (itx->itx_callback != NULL)
itx->itx_callback(itx->itx_callback_data);
zio_data_buf_free(itx, itx->itx_size);
}
/*
* Free up the sync and async itxs. The itxs_t has already been detached
* so no locks are needed.
*/
static void
zil_itxg_clean(void *arg)
{
itx_t *itx;
list_t *list;
avl_tree_t *t;
void *cookie;
itxs_t *itxs = arg;
itx_async_node_t *ian;
list = &itxs->i_sync_list;
while ((itx = list_head(list)) != NULL) {
/*
* In the general case, commit itxs will not be found
* here, as they'll be committed to an lwb via
* zil_lwb_commit(), and free'd in that function. Having
* said that, it is still possible for commit itxs to be
* found here, due to the following race:
*
* - a thread calls zil_commit() which assigns the
* commit itx to a per-txg i_sync_list
* - zil_itxg_clean() is called (e.g. via spa_sync())
* while the waiter is still on the i_sync_list
*
* There's nothing to prevent syncing the txg while the
* waiter is on the i_sync_list. This normally doesn't
* happen because spa_sync() is slower than zil_commit(),
* but if zil_commit() calls txg_wait_synced() (e.g.
* because zil_create() or zil_commit_writer_stall() is
* called) we will hit this case.
*/
if (itx->itx_lr.lrc_txtype == TX_COMMIT)
zil_commit_waiter_skip(itx->itx_private);
list_remove(list, itx);
zil_itx_destroy(itx);
}
cookie = NULL;
t = &itxs->i_async_tree;
while ((ian = avl_destroy_nodes(t, &cookie)) != NULL) {
list = &ian->ia_list;
while ((itx = list_head(list)) != NULL) {
list_remove(list, itx);
/* commit itxs should never be on the async lists. */
ASSERT3U(itx->itx_lr.lrc_txtype, !=, TX_COMMIT);
zil_itx_destroy(itx);
}
list_destroy(list);
kmem_free(ian, sizeof (itx_async_node_t));
}
avl_destroy(t);
kmem_free(itxs, sizeof (itxs_t));
}
static int
zil_aitx_compare(const void *x1, const void *x2)
{
const uint64_t o1 = ((itx_async_node_t *)x1)->ia_foid;
const uint64_t o2 = ((itx_async_node_t *)x2)->ia_foid;
return (TREE_CMP(o1, o2));
}
/*
* Remove all async itx with the given oid.
*/
void
zil_remove_async(zilog_t *zilog, uint64_t oid)
{
uint64_t otxg, txg;
itx_async_node_t *ian;
avl_tree_t *t;
avl_index_t where;
list_t clean_list;
itx_t *itx;
ASSERT(oid != 0);
list_create(&clean_list, sizeof (itx_t), offsetof(itx_t, itx_node));
if (spa_freeze_txg(zilog->zl_spa) != UINT64_MAX) /* ziltest support */
otxg = ZILTEST_TXG;
else
otxg = spa_last_synced_txg(zilog->zl_spa) + 1;
for (txg = otxg; txg < (otxg + TXG_CONCURRENT_STATES); txg++) {
itxg_t *itxg = &zilog->zl_itxg[txg & TXG_MASK];
mutex_enter(&itxg->itxg_lock);
if (itxg->itxg_txg != txg) {
mutex_exit(&itxg->itxg_lock);
continue;
}
/*
* Locate the object node and append its list.
*/
t = &itxg->itxg_itxs->i_async_tree;
ian = avl_find(t, &oid, &where);
if (ian != NULL)
list_move_tail(&clean_list, &ian->ia_list);
mutex_exit(&itxg->itxg_lock);
}
while ((itx = list_head(&clean_list)) != NULL) {
list_remove(&clean_list, itx);
/* commit itxs should never be on the async lists. */
ASSERT3U(itx->itx_lr.lrc_txtype, !=, TX_COMMIT);
zil_itx_destroy(itx);
}
list_destroy(&clean_list);
}
void
zil_itx_assign(zilog_t *zilog, itx_t *itx, dmu_tx_t *tx)
{
uint64_t txg;
itxg_t *itxg;
itxs_t *itxs, *clean = NULL;
/*
* Ensure the data of a renamed file is committed before the rename.
*/
if ((itx->itx_lr.lrc_txtype & ~TX_CI) == TX_RENAME)
zil_async_to_sync(zilog, itx->itx_oid);
if (spa_freeze_txg(zilog->zl_spa) != UINT64_MAX)
txg = ZILTEST_TXG;
else
txg = dmu_tx_get_txg(tx);
itxg = &zilog->zl_itxg[txg & TXG_MASK];
mutex_enter(&itxg->itxg_lock);
itxs = itxg->itxg_itxs;
if (itxg->itxg_txg != txg) {
if (itxs != NULL) {
/*
* The zil_clean callback hasn't got around to cleaning
* this itxg. Save the itxs for release below.
* This should be rare.
*/
zfs_dbgmsg("zil_itx_assign: missed itx cleanup for "
"txg %llu", (u_longlong_t)itxg->itxg_txg);
clean = itxg->itxg_itxs;
}
itxg->itxg_txg = txg;
itxs = itxg->itxg_itxs = kmem_zalloc(sizeof (itxs_t),
KM_SLEEP);
list_create(&itxs->i_sync_list, sizeof (itx_t),
offsetof(itx_t, itx_node));
avl_create(&itxs->i_async_tree, zil_aitx_compare,
sizeof (itx_async_node_t),
offsetof(itx_async_node_t, ia_node));
}
if (itx->itx_sync) {
list_insert_tail(&itxs->i_sync_list, itx);
} else {
avl_tree_t *t = &itxs->i_async_tree;
uint64_t foid =
LR_FOID_GET_OBJ(((lr_ooo_t *)&itx->itx_lr)->lr_foid);
itx_async_node_t *ian;
avl_index_t where;
ian = avl_find(t, &foid, &where);
if (ian == NULL) {
ian = kmem_alloc(sizeof (itx_async_node_t),
KM_SLEEP);
list_create(&ian->ia_list, sizeof (itx_t),
offsetof(itx_t, itx_node));
ian->ia_foid = foid;
avl_insert(t, ian, where);
}
list_insert_tail(&ian->ia_list, itx);
}
itx->itx_lr.lrc_txg = dmu_tx_get_txg(tx);
/*
* We don't want to dirty the ZIL using ZILTEST_TXG, because
* zil_clean() will never be called using ZILTEST_TXG. Thus, we
* need to be careful to always dirty the ZIL using the "real"
* TXG (not itxg_txg) even when the SPA is frozen.
*/
zilog_dirty(zilog, dmu_tx_get_txg(tx));
mutex_exit(&itxg->itxg_lock);
/* Release the old itxs now we've dropped the lock */
if (clean != NULL)
zil_itxg_clean(clean);
}
/*
* If there are any in-memory intent log transactions which have now been
* synced then start up a taskq to free them. We should only do this after we
* have written out the uberblocks (i.e. txg has been committed) so that
* don't inadvertently clean out in-memory log records that would be required
* by zil_commit().
*/
void
zil_clean(zilog_t *zilog, uint64_t synced_txg)
{
itxg_t *itxg = &zilog->zl_itxg[synced_txg & TXG_MASK];
itxs_t *clean_me;
ASSERT3U(synced_txg, <, ZILTEST_TXG);
mutex_enter(&itxg->itxg_lock);
if (itxg->itxg_itxs == NULL || itxg->itxg_txg == ZILTEST_TXG) {
mutex_exit(&itxg->itxg_lock);
return;
}
ASSERT3U(itxg->itxg_txg, <=, synced_txg);
ASSERT3U(itxg->itxg_txg, !=, 0);
clean_me = itxg->itxg_itxs;
itxg->itxg_itxs = NULL;
itxg->itxg_txg = 0;
mutex_exit(&itxg->itxg_lock);
/*
* Preferably start a task queue to free up the old itxs but
* if taskq_dispatch can't allocate resources to do that then
* free it in-line. This should be rare. Note, using TQ_SLEEP
* created a bad performance problem.
*/
ASSERT3P(zilog->zl_dmu_pool, !=, NULL);
ASSERT3P(zilog->zl_dmu_pool->dp_zil_clean_taskq, !=, NULL);
taskqid_t id = taskq_dispatch(zilog->zl_dmu_pool->dp_zil_clean_taskq,
zil_itxg_clean, clean_me, TQ_NOSLEEP);
if (id == TASKQID_INVALID)
zil_itxg_clean(clean_me);
}
/*
* This function will traverse the queue of itxs that need to be
* committed, and move them onto the ZIL's zl_itx_commit_list.
*/
static void
zil_get_commit_list(zilog_t *zilog)
{
uint64_t otxg, txg;
list_t *commit_list = &zilog->zl_itx_commit_list;
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
if (spa_freeze_txg(zilog->zl_spa) != UINT64_MAX) /* ziltest support */
otxg = ZILTEST_TXG;
else
otxg = spa_last_synced_txg(zilog->zl_spa) + 1;
/*
* This is inherently racy, since there is nothing to prevent
* the last synced txg from changing. That's okay since we'll
* only commit things in the future.
*/
for (txg = otxg; txg < (otxg + TXG_CONCURRENT_STATES); txg++) {
itxg_t *itxg = &zilog->zl_itxg[txg & TXG_MASK];
mutex_enter(&itxg->itxg_lock);
if (itxg->itxg_txg != txg) {
mutex_exit(&itxg->itxg_lock);
continue;
}
/*
* If we're adding itx records to the zl_itx_commit_list,
* then the zil better be dirty in this "txg". We can assert
* that here since we're holding the itxg_lock which will
* prevent spa_sync from cleaning it. Once we add the itxs
* to the zl_itx_commit_list we must commit it to disk even
* if it's unnecessary (i.e. the txg was synced).
*/
ASSERT(zilog_is_dirty_in_txg(zilog, txg) ||
spa_freeze_txg(zilog->zl_spa) != UINT64_MAX);
list_move_tail(commit_list, &itxg->itxg_itxs->i_sync_list);
mutex_exit(&itxg->itxg_lock);
}
}
/*
* Move the async itxs for a specified object to commit into sync lists.
*/
void
zil_async_to_sync(zilog_t *zilog, uint64_t foid)
{
uint64_t otxg, txg;
itx_async_node_t *ian;
avl_tree_t *t;
avl_index_t where;
if (spa_freeze_txg(zilog->zl_spa) != UINT64_MAX) /* ziltest support */
otxg = ZILTEST_TXG;
else
otxg = spa_last_synced_txg(zilog->zl_spa) + 1;
/*
* This is inherently racy, since there is nothing to prevent
* the last synced txg from changing.
*/
for (txg = otxg; txg < (otxg + TXG_CONCURRENT_STATES); txg++) {
itxg_t *itxg = &zilog->zl_itxg[txg & TXG_MASK];
mutex_enter(&itxg->itxg_lock);
if (itxg->itxg_txg != txg) {
mutex_exit(&itxg->itxg_lock);
continue;
}
/*
* If a foid is specified then find that node and append its
* list. Otherwise walk the tree appending all the lists
* to the sync list. We add to the end rather than the
* beginning to ensure the create has happened.
*/
t = &itxg->itxg_itxs->i_async_tree;
if (foid != 0) {
ian = avl_find(t, &foid, &where);
if (ian != NULL) {
list_move_tail(&itxg->itxg_itxs->i_sync_list,
&ian->ia_list);
}
} else {
void *cookie = NULL;
while ((ian = avl_destroy_nodes(t, &cookie)) != NULL) {
list_move_tail(&itxg->itxg_itxs->i_sync_list,
&ian->ia_list);
list_destroy(&ian->ia_list);
kmem_free(ian, sizeof (itx_async_node_t));
}
}
mutex_exit(&itxg->itxg_lock);
}
}
/*
* This function will prune commit itxs that are at the head of the
* commit list (it won't prune past the first non-commit itx), and
* either: a) attach them to the last lwb that's still pending
* completion, or b) skip them altogether.
*
* This is used as a performance optimization to prevent commit itxs
* from generating new lwbs when it's unnecessary to do so.
*/
static void
zil_prune_commit_list(zilog_t *zilog)
{
itx_t *itx;
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
while ((itx = list_head(&zilog->zl_itx_commit_list)) != NULL) {
lr_t *lrc = &itx->itx_lr;
if (lrc->lrc_txtype != TX_COMMIT)
break;
mutex_enter(&zilog->zl_lock);
lwb_t *last_lwb = zilog->zl_last_lwb_opened;
if (last_lwb == NULL ||
last_lwb->lwb_state == LWB_STATE_FLUSH_DONE) {
/*
* All of the itxs this waiter was waiting on
* must have already completed (or there were
* never any itx's for it to wait on), so it's
* safe to skip this waiter and mark it done.
*/
zil_commit_waiter_skip(itx->itx_private);
} else {
zil_commit_waiter_link_lwb(itx->itx_private, last_lwb);
itx->itx_private = NULL;
}
mutex_exit(&zilog->zl_lock);
list_remove(&zilog->zl_itx_commit_list, itx);
zil_itx_destroy(itx);
}
IMPLY(itx != NULL, itx->itx_lr.lrc_txtype != TX_COMMIT);
}
static void
zil_commit_writer_stall(zilog_t *zilog)
{
/*
* When zio_alloc_zil() fails to allocate the next lwb block on
* disk, we must call txg_wait_synced() to ensure all of the
* lwbs in the zilog's zl_lwb_list are synced and then freed (in
* zil_sync()), such that any subsequent ZIL writer (i.e. a call
* to zil_process_commit_list()) will have to call zil_create(),
* and start a new ZIL chain.
*
* Since zil_alloc_zil() failed, the lwb that was previously
* issued does not have a pointer to the "next" lwb on disk.
* Thus, if another ZIL writer thread was to allocate the "next"
* on-disk lwb, that block could be leaked in the event of a
* crash (because the previous lwb on-disk would not point to
* it).
*
* We must hold the zilog's zl_issuer_lock while we do this, to
* ensure no new threads enter zil_process_commit_list() until
* all lwb's in the zl_lwb_list have been synced and freed
* (which is achieved via the txg_wait_synced() call).
*/
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
txg_wait_synced(zilog->zl_dmu_pool, 0);
ASSERT3P(list_tail(&zilog->zl_lwb_list), ==, NULL);
}
/*
* This function will traverse the commit list, creating new lwbs as
* needed, and committing the itxs from the commit list to these newly
* created lwbs. Additionally, as a new lwb is created, the previous
* lwb will be issued to the zio layer to be written to disk.
*/
static void
zil_process_commit_list(zilog_t *zilog)
{
spa_t *spa = zilog->zl_spa;
list_t nolwb_itxs;
list_t nolwb_waiters;
lwb_t *lwb, *plwb;
itx_t *itx;
boolean_t first = B_TRUE;
ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock));
/*
* Return if there's nothing to commit before we dirty the fs by
* calling zil_create().
*/
if (list_head(&zilog->zl_itx_commit_list) == NULL)
return;
list_create(&nolwb_itxs, sizeof (itx_t), offsetof(itx_t, itx_node));
list_create(&nolwb_waiters, sizeof (zil_commit_waiter_t),
offsetof(zil_commit_waiter_t, zcw_node));
lwb = list_tail(&zilog->zl_lwb_list);
if (lwb == NULL) {
lwb = zil_create(zilog);
} else {
/*
* Activate SPA_FEATURE_ZILSAXATTR for the cases where ZIL will
* have already been created (zl_lwb_list not empty).
*/
zil_commit_activate_saxattr_feature(zilog);
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_ISSUED);
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_WRITE_DONE);
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_FLUSH_DONE);
first = (lwb->lwb_state != LWB_STATE_OPENED) &&
((plwb = list_prev(&zilog->zl_lwb_list, lwb)) == NULL ||
plwb->lwb_state == LWB_STATE_FLUSH_DONE);
}
while ((itx = list_head(&zilog->zl_itx_commit_list)) != NULL) {
lr_t *lrc = &itx->itx_lr;
uint64_t txg = lrc->lrc_txg;
ASSERT3U(txg, !=, 0);
if (lrc->lrc_txtype == TX_COMMIT) {
DTRACE_PROBE2(zil__process__commit__itx,
zilog_t *, zilog, itx_t *, itx);
} else {
DTRACE_PROBE2(zil__process__normal__itx,
zilog_t *, zilog, itx_t *, itx);
}
list_remove(&zilog->zl_itx_commit_list, itx);
boolean_t synced = txg <= spa_last_synced_txg(spa);
boolean_t frozen = txg > spa_freeze_txg(spa);
/*
* If the txg of this itx has already been synced out, then
* we don't need to commit this itx to an lwb. This is
* because the data of this itx will have already been
* written to the main pool. This is inherently racy, and
* it's still ok to commit an itx whose txg has already
* been synced; this will result in a write that's
* unnecessary, but will do no harm.
*
* With that said, we always want to commit TX_COMMIT itxs
* to an lwb, regardless of whether or not that itx's txg
* has been synced out. We do this to ensure any OPENED lwb
* will always have at least one zil_commit_waiter_t linked
* to the lwb.
*
* As a counter-example, if we skipped TX_COMMIT itx's
* whose txg had already been synced, the following
* situation could occur if we happened to be racing with
* spa_sync:
*
* 1. We commit a non-TX_COMMIT itx to an lwb, where the
* itx's txg is 10 and the last synced txg is 9.
* 2. spa_sync finishes syncing out txg 10.
* 3. We move to the next itx in the list, it's a TX_COMMIT
* whose txg is 10, so we skip it rather than committing
* it to the lwb used in (1).
*
* If the itx that is skipped in (3) is the last TX_COMMIT
* itx in the commit list, than it's possible for the lwb
* used in (1) to remain in the OPENED state indefinitely.
*
* To prevent the above scenario from occurring, ensuring
* that once an lwb is OPENED it will transition to ISSUED
* and eventually DONE, we always commit TX_COMMIT itx's to
* an lwb here, even if that itx's txg has already been
* synced.
*
* Finally, if the pool is frozen, we _always_ commit the
* itx. The point of freezing the pool is to prevent data
* from being written to the main pool via spa_sync, and
* instead rely solely on the ZIL to persistently store the
* data; i.e. when the pool is frozen, the last synced txg
* value can't be trusted.
*/
if (frozen || !synced || lrc->lrc_txtype == TX_COMMIT) {
if (lwb != NULL) {
lwb = zil_lwb_commit(zilog, itx, lwb);
if (lwb == NULL)
list_insert_tail(&nolwb_itxs, itx);
else
list_insert_tail(&lwb->lwb_itxs, itx);
} else {
if (lrc->lrc_txtype == TX_COMMIT) {
zil_commit_waiter_link_nolwb(
itx->itx_private, &nolwb_waiters);
}
list_insert_tail(&nolwb_itxs, itx);
}
} else {
ASSERT3S(lrc->lrc_txtype, !=, TX_COMMIT);
zil_itx_destroy(itx);
}
}
if (lwb == NULL) {
/*
* This indicates zio_alloc_zil() failed to allocate the
* "next" lwb on-disk. When this happens, we must stall
* the ZIL write pipeline; see the comment within
* zil_commit_writer_stall() for more details.
*/
zil_commit_writer_stall(zilog);
/*
* Additionally, we have to signal and mark the "nolwb"
* waiters as "done" here, since without an lwb, we
* can't do this via zil_lwb_flush_vdevs_done() like
* normal.
*/
zil_commit_waiter_t *zcw;
while ((zcw = list_head(&nolwb_waiters)) != NULL) {
zil_commit_waiter_skip(zcw);
list_remove(&nolwb_waiters, zcw);
}
/*
* And finally, we have to destroy the itx's that
* couldn't be committed to an lwb; this will also call
* the itx's callback if one exists for the itx.
*/
while ((itx = list_head(&nolwb_itxs)) != NULL) {
list_remove(&nolwb_itxs, itx);
zil_itx_destroy(itx);
}
} else {
ASSERT(list_is_empty(&nolwb_waiters));
ASSERT3P(lwb, !=, NULL);
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_ISSUED);
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_WRITE_DONE);
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_FLUSH_DONE);
/*
* At this point, the ZIL block pointed at by the "lwb"
* variable is in one of the following states: "closed"
* or "open".
*
* If it's "closed", then no itxs have been committed to
* it, so there's no point in issuing its zio (i.e. it's
* "empty").
*
* If it's "open", then it contains one or more itxs that
* eventually need to be committed to stable storage. In
* this case we intentionally do not issue the lwb's zio
* to disk yet, and instead rely on one of the following
* two mechanisms for issuing the zio:
*
* 1. Ideally, there will be more ZIL activity occurring
* on the system, such that this function will be
* immediately called again (not necessarily by the same
* thread) and this lwb's zio will be issued via
* zil_lwb_commit(). This way, the lwb is guaranteed to
* be "full" when it is issued to disk, and we'll make
* use of the lwb's size the best we can.
*
* 2. If there isn't sufficient ZIL activity occurring on
* the system, such that this lwb's zio isn't issued via
* zil_lwb_commit(), zil_commit_waiter() will issue the
* lwb's zio. If this occurs, the lwb is not guaranteed
* to be "full" by the time its zio is issued, and means
* the size of the lwb was "too large" given the amount
* of ZIL activity occurring on the system at that time.
*
* We do this for a couple of reasons:
*
* 1. To try and reduce the number of IOPs needed to
* write the same number of itxs. If an lwb has space
* available in its buffer for more itxs, and more itxs
* will be committed relatively soon (relative to the
* latency of performing a write), then it's beneficial
* to wait for these "next" itxs. This way, more itxs
* can be committed to stable storage with fewer writes.
*
* 2. To try and use the largest lwb block size that the
* incoming rate of itxs can support. Again, this is to
* try and pack as many itxs into as few lwbs as
* possible, without significantly impacting the latency
* of each individual itx.
*
* If we had no already running or open LWBs, it can be
* the workload is single-threaded. And if the ZIL write
* latency is very small or if the LWB is almost full, it
* may be cheaper to bypass the delay.
*/
if (lwb->lwb_state == LWB_STATE_OPENED && first) {
hrtime_t sleep = zilog->zl_last_lwb_latency *
zfs_commit_timeout_pct / 100;
if (sleep < zil_min_commit_timeout ||
lwb->lwb_sz - lwb->lwb_nused < lwb->lwb_sz / 8) {
lwb = zil_lwb_write_issue(zilog, lwb);
zilog->zl_cur_used = 0;
if (lwb == NULL)
zil_commit_writer_stall(zilog);
}
}
}
}
/*
* This function is responsible for ensuring the passed in commit waiter
* (and associated commit itx) is committed to an lwb. If the waiter is
* not already committed to an lwb, all itxs in the zilog's queue of
* itxs will be processed. The assumption is the passed in waiter's
* commit itx will found in the queue just like the other non-commit
* itxs, such that when the entire queue is processed, the waiter will
* have been committed to an lwb.
*
* The lwb associated with the passed in waiter is not guaranteed to
* have been issued by the time this function completes. If the lwb is
* not issued, we rely on future calls to zil_commit_writer() to issue
* the lwb, or the timeout mechanism found in zil_commit_waiter().
*/
static void
zil_commit_writer(zilog_t *zilog, zil_commit_waiter_t *zcw)
{
ASSERT(!MUTEX_HELD(&zilog->zl_lock));
ASSERT(spa_writeable(zilog->zl_spa));
mutex_enter(&zilog->zl_issuer_lock);
if (zcw->zcw_lwb != NULL || zcw->zcw_done) {
/*
* It's possible that, while we were waiting to acquire
* the "zl_issuer_lock", another thread committed this
* waiter to an lwb. If that occurs, we bail out early,
* without processing any of the zilog's queue of itxs.
*
* On certain workloads and system configurations, the
* "zl_issuer_lock" can become highly contended. In an
* attempt to reduce this contention, we immediately drop
* the lock if the waiter has already been processed.
*
* We've measured this optimization to reduce CPU spent
* contending on this lock by up to 5%, using a system
* with 32 CPUs, low latency storage (~50 usec writes),
* and 1024 threads performing sync writes.
*/
goto out;
}
ZIL_STAT_BUMP(zilog, zil_commit_writer_count);
zil_get_commit_list(zilog);
zil_prune_commit_list(zilog);
zil_process_commit_list(zilog);
out:
mutex_exit(&zilog->zl_issuer_lock);
}
static void
zil_commit_waiter_timeout(zilog_t *zilog, zil_commit_waiter_t *zcw)
{
ASSERT(!MUTEX_HELD(&zilog->zl_issuer_lock));
ASSERT(MUTEX_HELD(&zcw->zcw_lock));
ASSERT3B(zcw->zcw_done, ==, B_FALSE);
lwb_t *lwb = zcw->zcw_lwb;
ASSERT3P(lwb, !=, NULL);
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_CLOSED);
/*
* If the lwb has already been issued by another thread, we can
* immediately return since there's no work to be done (the
* point of this function is to issue the lwb). Additionally, we
* do this prior to acquiring the zl_issuer_lock, to avoid
* acquiring it when it's not necessary to do so.
*/
if (lwb->lwb_state == LWB_STATE_ISSUED ||
lwb->lwb_state == LWB_STATE_WRITE_DONE ||
lwb->lwb_state == LWB_STATE_FLUSH_DONE)
return;
/*
* In order to call zil_lwb_write_issue() we must hold the
* zilog's "zl_issuer_lock". We can't simply acquire that lock,
* since we're already holding the commit waiter's "zcw_lock",
* and those two locks are acquired in the opposite order
* elsewhere.
*/
mutex_exit(&zcw->zcw_lock);
mutex_enter(&zilog->zl_issuer_lock);
mutex_enter(&zcw->zcw_lock);
/*
* Since we just dropped and re-acquired the commit waiter's
* lock, we have to re-check to see if the waiter was marked
* "done" during that process. If the waiter was marked "done",
* the "lwb" pointer is no longer valid (it can be free'd after
* the waiter is marked "done"), so without this check we could
* wind up with a use-after-free error below.
*/
if (zcw->zcw_done)
goto out;
ASSERT3P(lwb, ==, zcw->zcw_lwb);
/*
* We've already checked this above, but since we hadn't acquired
* the zilog's zl_issuer_lock, we have to perform this check a
* second time while holding the lock.
*
* We don't need to hold the zl_lock since the lwb cannot transition
* from OPENED to ISSUED while we hold the zl_issuer_lock. The lwb
* _can_ transition from ISSUED to DONE, but it's OK to race with
* that transition since we treat the lwb the same, whether it's in
* the ISSUED or DONE states.
*
* The important thing, is we treat the lwb differently depending on
* if it's ISSUED or OPENED, and block any other threads that might
* attempt to issue this lwb. For that reason we hold the
* zl_issuer_lock when checking the lwb_state; we must not call
* zil_lwb_write_issue() if the lwb had already been issued.
*
* See the comment above the lwb_state_t structure definition for
* more details on the lwb states, and locking requirements.
*/
if (lwb->lwb_state == LWB_STATE_ISSUED ||
lwb->lwb_state == LWB_STATE_WRITE_DONE ||
lwb->lwb_state == LWB_STATE_FLUSH_DONE)
goto out;
ASSERT3S(lwb->lwb_state, ==, LWB_STATE_OPENED);
/*
* As described in the comments above zil_commit_waiter() and
* zil_process_commit_list(), we need to issue this lwb's zio
* since we've reached the commit waiter's timeout and it still
* hasn't been issued.
*/
lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb);
IMPLY(nlwb != NULL, lwb->lwb_state != LWB_STATE_OPENED);
/*
* Since the lwb's zio hadn't been issued by the time this thread
* reached its timeout, we reset the zilog's "zl_cur_used" field
* to influence the zil block size selection algorithm.
*
* By having to issue the lwb's zio here, it means the size of the
* lwb was too large, given the incoming throughput of itxs. By
* setting "zl_cur_used" to zero, we communicate this fact to the
* block size selection algorithm, so it can take this information
* into account, and potentially select a smaller size for the
* next lwb block that is allocated.
*/
zilog->zl_cur_used = 0;
if (nlwb == NULL) {
/*
* When zil_lwb_write_issue() returns NULL, this
* indicates zio_alloc_zil() failed to allocate the
* "next" lwb on-disk. When this occurs, the ZIL write
* pipeline must be stalled; see the comment within the
* zil_commit_writer_stall() function for more details.
*
* We must drop the commit waiter's lock prior to
* calling zil_commit_writer_stall() or else we can wind
* up with the following deadlock:
*
* - This thread is waiting for the txg to sync while
* holding the waiter's lock; txg_wait_synced() is
* used within txg_commit_writer_stall().
*
* - The txg can't sync because it is waiting for this
* lwb's zio callback to call dmu_tx_commit().
*
* - The lwb's zio callback can't call dmu_tx_commit()
* because it's blocked trying to acquire the waiter's
* lock, which occurs prior to calling dmu_tx_commit()
*/
mutex_exit(&zcw->zcw_lock);
zil_commit_writer_stall(zilog);
mutex_enter(&zcw->zcw_lock);
}
out:
mutex_exit(&zilog->zl_issuer_lock);
ASSERT(MUTEX_HELD(&zcw->zcw_lock));
}
/*
* This function is responsible for performing the following two tasks:
*
* 1. its primary responsibility is to block until the given "commit
* waiter" is considered "done".
*
* 2. its secondary responsibility is to issue the zio for the lwb that
* the given "commit waiter" is waiting on, if this function has
* waited "long enough" and the lwb is still in the "open" state.
*
* Given a sufficient amount of itxs being generated and written using
* the ZIL, the lwb's zio will be issued via the zil_lwb_commit()
* function. If this does not occur, this secondary responsibility will
* ensure the lwb is issued even if there is not other synchronous
* activity on the system.
*
* For more details, see zil_process_commit_list(); more specifically,
* the comment at the bottom of that function.
*/
static void
zil_commit_waiter(zilog_t *zilog, zil_commit_waiter_t *zcw)
{
ASSERT(!MUTEX_HELD(&zilog->zl_lock));
ASSERT(!MUTEX_HELD(&zilog->zl_issuer_lock));
ASSERT(spa_writeable(zilog->zl_spa));
mutex_enter(&zcw->zcw_lock);
/*
* The timeout is scaled based on the lwb latency to avoid
* significantly impacting the latency of each individual itx.
* For more details, see the comment at the bottom of the
* zil_process_commit_list() function.
*/
int pct = MAX(zfs_commit_timeout_pct, 1);
hrtime_t sleep = (zilog->zl_last_lwb_latency * pct) / 100;
hrtime_t wakeup = gethrtime() + sleep;
boolean_t timedout = B_FALSE;
while (!zcw->zcw_done) {
ASSERT(MUTEX_HELD(&zcw->zcw_lock));
lwb_t *lwb = zcw->zcw_lwb;
/*
* Usually, the waiter will have a non-NULL lwb field here,
* but it's possible for it to be NULL as a result of
* zil_commit() racing with spa_sync().
*
* When zil_clean() is called, it's possible for the itxg
* list (which may be cleaned via a taskq) to contain
* commit itxs. When this occurs, the commit waiters linked
* off of these commit itxs will not be committed to an
* lwb. Additionally, these commit waiters will not be
* marked done until zil_commit_waiter_skip() is called via
* zil_itxg_clean().
*
* Thus, it's possible for this commit waiter (i.e. the
* "zcw" variable) to be found in this "in between" state;
* where it's "zcw_lwb" field is NULL, and it hasn't yet
* been skipped, so it's "zcw_done" field is still B_FALSE.
*/
IMPLY(lwb != NULL, lwb->lwb_state != LWB_STATE_CLOSED);
if (lwb != NULL && lwb->lwb_state == LWB_STATE_OPENED) {
ASSERT3B(timedout, ==, B_FALSE);
/*
* If the lwb hasn't been issued yet, then we
* need to wait with a timeout, in case this
* function needs to issue the lwb after the
* timeout is reached; responsibility (2) from
* the comment above this function.
*/
int rc = cv_timedwait_hires(&zcw->zcw_cv,
&zcw->zcw_lock, wakeup, USEC2NSEC(1),
CALLOUT_FLAG_ABSOLUTE);
if (rc != -1 || zcw->zcw_done)
continue;
timedout = B_TRUE;
zil_commit_waiter_timeout(zilog, zcw);
if (!zcw->zcw_done) {
/*
* If the commit waiter has already been
* marked "done", it's possible for the
* waiter's lwb structure to have already
* been freed. Thus, we can only reliably
* make these assertions if the waiter
* isn't done.
*/
ASSERT3P(lwb, ==, zcw->zcw_lwb);
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED);
}
} else {
/*
* If the lwb isn't open, then it must have already
* been issued. In that case, there's no need to
* use a timeout when waiting for the lwb to
* complete.
*
* Additionally, if the lwb is NULL, the waiter
* will soon be signaled and marked done via
* zil_clean() and zil_itxg_clean(), so no timeout
* is required.
*/
IMPLY(lwb != NULL,
lwb->lwb_state == LWB_STATE_ISSUED ||
lwb->lwb_state == LWB_STATE_WRITE_DONE ||
lwb->lwb_state == LWB_STATE_FLUSH_DONE);
cv_wait(&zcw->zcw_cv, &zcw->zcw_lock);
}
}
mutex_exit(&zcw->zcw_lock);
}
static zil_commit_waiter_t *
zil_alloc_commit_waiter(void)
{
zil_commit_waiter_t *zcw = kmem_cache_alloc(zil_zcw_cache, KM_SLEEP);
cv_init(&zcw->zcw_cv, NULL, CV_DEFAULT, NULL);
mutex_init(&zcw->zcw_lock, NULL, MUTEX_DEFAULT, NULL);
list_link_init(&zcw->zcw_node);
zcw->zcw_lwb = NULL;
zcw->zcw_done = B_FALSE;
zcw->zcw_zio_error = 0;
return (zcw);
}
static void
zil_free_commit_waiter(zil_commit_waiter_t *zcw)
{
ASSERT(!list_link_active(&zcw->zcw_node));
ASSERT3P(zcw->zcw_lwb, ==, NULL);
ASSERT3B(zcw->zcw_done, ==, B_TRUE);
mutex_destroy(&zcw->zcw_lock);
cv_destroy(&zcw->zcw_cv);
kmem_cache_free(zil_zcw_cache, zcw);
}
/*
* This function is used to create a TX_COMMIT itx and assign it. This
* way, it will be linked into the ZIL's list of synchronous itxs, and
* then later committed to an lwb (or skipped) when
* zil_process_commit_list() is called.
*/
static void
zil_commit_itx_assign(zilog_t *zilog, zil_commit_waiter_t *zcw)
{
dmu_tx_t *tx = dmu_tx_create(zilog->zl_os);
- VERIFY0(dmu_tx_assign(tx, TXG_WAIT));
+
+ /*
+ * Since we are not going to create any new dirty data, and we
+ * can even help with clearing the existing dirty data, we
+ * should not be subject to the dirty data based delays. We
+ * use TXG_NOTHROTTLE to bypass the delay mechanism.
+ */
+ VERIFY0(dmu_tx_assign(tx, TXG_WAIT | TXG_NOTHROTTLE));
itx_t *itx = zil_itx_create(TX_COMMIT, sizeof (lr_t));
itx->itx_sync = B_TRUE;
itx->itx_private = zcw;
zil_itx_assign(zilog, itx, tx);
dmu_tx_commit(tx);
}
/*
* Commit ZFS Intent Log transactions (itxs) to stable storage.
*
* When writing ZIL transactions to the on-disk representation of the
* ZIL, the itxs are committed to a Log Write Block (lwb). Multiple
* itxs can be committed to a single lwb. Once a lwb is written and
* committed to stable storage (i.e. the lwb is written, and vdevs have
* been flushed), each itx that was committed to that lwb is also
* considered to be committed to stable storage.
*
* When an itx is committed to an lwb, the log record (lr_t) contained
* by the itx is copied into the lwb's zio buffer, and once this buffer
* is written to disk, it becomes an on-disk ZIL block.
*
* As itxs are generated, they're inserted into the ZIL's queue of
* uncommitted itxs. The semantics of zil_commit() are such that it will
* block until all itxs that were in the queue when it was called, are
* committed to stable storage.
*
* If "foid" is zero, this means all "synchronous" and "asynchronous"
* itxs, for all objects in the dataset, will be committed to stable
* storage prior to zil_commit() returning. If "foid" is non-zero, all
* "synchronous" itxs for all objects, but only "asynchronous" itxs
* that correspond to the foid passed in, will be committed to stable
* storage prior to zil_commit() returning.
*
* Generally speaking, when zil_commit() is called, the consumer doesn't
* actually care about _all_ of the uncommitted itxs. Instead, they're
* simply trying to waiting for a specific itx to be committed to disk,
* but the interface(s) for interacting with the ZIL don't allow such
* fine-grained communication. A better interface would allow a consumer
* to create and assign an itx, and then pass a reference to this itx to
* zil_commit(); such that zil_commit() would return as soon as that
* specific itx was committed to disk (instead of waiting for _all_
* itxs to be committed).
*
* When a thread calls zil_commit() a special "commit itx" will be
* generated, along with a corresponding "waiter" for this commit itx.
* zil_commit() will wait on this waiter's CV, such that when the waiter
* is marked done, and signaled, zil_commit() will return.
*
* This commit itx is inserted into the queue of uncommitted itxs. This
* provides an easy mechanism for determining which itxs were in the
* queue prior to zil_commit() having been called, and which itxs were
* added after zil_commit() was called.
*
* The commit itx is special; it doesn't have any on-disk representation.
* When a commit itx is "committed" to an lwb, the waiter associated
* with it is linked onto the lwb's list of waiters. Then, when that lwb
* completes, each waiter on the lwb's list is marked done and signaled
* -- allowing the thread waiting on the waiter to return from zil_commit().
*
* It's important to point out a few critical factors that allow us
* to make use of the commit itxs, commit waiters, per-lwb lists of
* commit waiters, and zio completion callbacks like we're doing:
*
* 1. The list of waiters for each lwb is traversed, and each commit
* waiter is marked "done" and signaled, in the zio completion
* callback of the lwb's zio[*].
*
* * Actually, the waiters are signaled in the zio completion
* callback of the root zio for the DKIOCFLUSHWRITECACHE commands
* that are sent to the vdevs upon completion of the lwb zio.
*
* 2. When the itxs are inserted into the ZIL's queue of uncommitted
* itxs, the order in which they are inserted is preserved[*]; as
* itxs are added to the queue, they are added to the tail of
* in-memory linked lists.
*
* When committing the itxs to lwbs (to be written to disk), they
* are committed in the same order in which the itxs were added to
* the uncommitted queue's linked list(s); i.e. the linked list of
* itxs to commit is traversed from head to tail, and each itx is
* committed to an lwb in that order.
*
* * To clarify:
*
* - the order of "sync" itxs is preserved w.r.t. other
* "sync" itxs, regardless of the corresponding objects.
* - the order of "async" itxs is preserved w.r.t. other
* "async" itxs corresponding to the same object.
* - the order of "async" itxs is *not* preserved w.r.t. other
* "async" itxs corresponding to different objects.
* - the order of "sync" itxs w.r.t. "async" itxs (or vice
* versa) is *not* preserved, even for itxs that correspond
* to the same object.
*
* For more details, see: zil_itx_assign(), zil_async_to_sync(),
* zil_get_commit_list(), and zil_process_commit_list().
*
* 3. The lwbs represent a linked list of blocks on disk. Thus, any
* lwb cannot be considered committed to stable storage, until its
* "previous" lwb is also committed to stable storage. This fact,
* coupled with the fact described above, means that itxs are
* committed in (roughly) the order in which they were generated.
* This is essential because itxs are dependent on prior itxs.
* Thus, we *must not* deem an itx as being committed to stable
* storage, until *all* prior itxs have also been committed to
* stable storage.
*
* To enforce this ordering of lwb zio's, while still leveraging as
* much of the underlying storage performance as possible, we rely
* on two fundamental concepts:
*
* 1. The creation and issuance of lwb zio's is protected by
* the zilog's "zl_issuer_lock", which ensures only a single
* thread is creating and/or issuing lwb's at a time
* 2. The "previous" lwb is a child of the "current" lwb
* (leveraging the zio parent-child dependency graph)
*
* By relying on this parent-child zio relationship, we can have
* many lwb zio's concurrently issued to the underlying storage,
* but the order in which they complete will be the same order in
* which they were created.
*/
void
zil_commit(zilog_t *zilog, uint64_t foid)
{
/*
* We should never attempt to call zil_commit on a snapshot for
* a couple of reasons:
*
* 1. A snapshot may never be modified, thus it cannot have any
* in-flight itxs that would have modified the dataset.
*
* 2. By design, when zil_commit() is called, a commit itx will
* be assigned to this zilog; as a result, the zilog will be
* dirtied. We must not dirty the zilog of a snapshot; there's
* checks in the code that enforce this invariant, and will
* cause a panic if it's not upheld.
*/
ASSERT3B(dmu_objset_is_snapshot(zilog->zl_os), ==, B_FALSE);
if (zilog->zl_sync == ZFS_SYNC_DISABLED)
return;
if (!spa_writeable(zilog->zl_spa)) {
/*
* If the SPA is not writable, there should never be any
* pending itxs waiting to be committed to disk. If that
* weren't true, we'd skip writing those itxs out, and
* would break the semantics of zil_commit(); thus, we're
* verifying that truth before we return to the caller.
*/
ASSERT(list_is_empty(&zilog->zl_lwb_list));
ASSERT3P(zilog->zl_last_lwb_opened, ==, NULL);
for (int i = 0; i < TXG_SIZE; i++)
ASSERT3P(zilog->zl_itxg[i].itxg_itxs, ==, NULL);
return;
}
/*
* If the ZIL is suspended, we don't want to dirty it by calling
* zil_commit_itx_assign() below, nor can we write out
* lwbs like would be done in zil_commit_write(). Thus, we
* simply rely on txg_wait_synced() to maintain the necessary
* semantics, and avoid calling those functions altogether.
*/
if (zilog->zl_suspend > 0) {
txg_wait_synced(zilog->zl_dmu_pool, 0);
return;
}
zil_commit_impl(zilog, foid);
}
void
zil_commit_impl(zilog_t *zilog, uint64_t foid)
{
ZIL_STAT_BUMP(zilog, zil_commit_count);
/*
* Move the "async" itxs for the specified foid to the "sync"
* queues, such that they will be later committed (or skipped)
* to an lwb when zil_process_commit_list() is called.
*
* Since these "async" itxs must be committed prior to this
* call to zil_commit returning, we must perform this operation
* before we call zil_commit_itx_assign().
*/
zil_async_to_sync(zilog, foid);
/*
* We allocate a new "waiter" structure which will initially be
* linked to the commit itx using the itx's "itx_private" field.
* Since the commit itx doesn't represent any on-disk state,
* when it's committed to an lwb, rather than copying the its
* lr_t into the lwb's buffer, the commit itx's "waiter" will be
* added to the lwb's list of waiters. Then, when the lwb is
* committed to stable storage, each waiter in the lwb's list of
* waiters will be marked "done", and signalled.
*
* We must create the waiter and assign the commit itx prior to
* calling zil_commit_writer(), or else our specific commit itx
* is not guaranteed to be committed to an lwb prior to calling
* zil_commit_waiter().
*/
zil_commit_waiter_t *zcw = zil_alloc_commit_waiter();
zil_commit_itx_assign(zilog, zcw);
zil_commit_writer(zilog, zcw);
zil_commit_waiter(zilog, zcw);
if (zcw->zcw_zio_error != 0) {
/*
* If there was an error writing out the ZIL blocks that
* this thread is waiting on, then we fallback to
* relying on spa_sync() to write out the data this
* thread is waiting on. Obviously this has performance
* implications, but the expectation is for this to be
* an exceptional case, and shouldn't occur often.
*/
DTRACE_PROBE2(zil__commit__io__error,
zilog_t *, zilog, zil_commit_waiter_t *, zcw);
txg_wait_synced(zilog->zl_dmu_pool, 0);
}
zil_free_commit_waiter(zcw);
}
/*
* Called in syncing context to free committed log blocks and update log header.
*/
void
zil_sync(zilog_t *zilog, dmu_tx_t *tx)
{
zil_header_t *zh = zil_header_in_syncing_context(zilog);
uint64_t txg = dmu_tx_get_txg(tx);
spa_t *spa = zilog->zl_spa;
uint64_t *replayed_seq = &zilog->zl_replayed_seq[txg & TXG_MASK];
lwb_t *lwb;
/*
* We don't zero out zl_destroy_txg, so make sure we don't try
* to destroy it twice.
*/
if (spa_sync_pass(spa) != 1)
return;
zil_lwb_flush_wait_all(zilog, txg);
mutex_enter(&zilog->zl_lock);
ASSERT(zilog->zl_stop_sync == 0);
if (*replayed_seq != 0) {
ASSERT(zh->zh_replay_seq < *replayed_seq);
zh->zh_replay_seq = *replayed_seq;
*replayed_seq = 0;
}
if (zilog->zl_destroy_txg == txg) {
blkptr_t blk = zh->zh_log;
dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os);
ASSERT(list_head(&zilog->zl_lwb_list) == NULL);
memset(zh, 0, sizeof (zil_header_t));
memset(zilog->zl_replayed_seq, 0,
sizeof (zilog->zl_replayed_seq));
if (zilog->zl_keep_first) {
/*
* If this block was part of log chain that couldn't
* be claimed because a device was missing during
* zil_claim(), but that device later returns,
* then this block could erroneously appear valid.
* To guard against this, assign a new GUID to the new
* log chain so it doesn't matter what blk points to.
*/
zil_init_log_chain(zilog, &blk);
zh->zh_log = blk;
} else {
/*
* A destroyed ZIL chain can't contain any TX_SETSAXATTR
* records. So, deactivate the feature for this dataset.
* We activate it again when we start a new ZIL chain.
*/
if (dsl_dataset_feature_is_active(ds,
SPA_FEATURE_ZILSAXATTR))
dsl_dataset_deactivate_feature(ds,
SPA_FEATURE_ZILSAXATTR, tx);
}
}
while ((lwb = list_head(&zilog->zl_lwb_list)) != NULL) {
zh->zh_log = lwb->lwb_blk;
if (lwb->lwb_buf != NULL || lwb->lwb_max_txg > txg)
break;
list_remove(&zilog->zl_lwb_list, lwb);
zio_free(spa, txg, &lwb->lwb_blk);
zil_free_lwb(zilog, lwb);
/*
* If we don't have anything left in the lwb list then
* we've had an allocation failure and we need to zero
* out the zil_header blkptr so that we don't end
* up freeing the same block twice.
*/
if (list_head(&zilog->zl_lwb_list) == NULL)
BP_ZERO(&zh->zh_log);
}
/*
* Remove fastwrite on any blocks that have been pre-allocated for
* the next commit. This prevents fastwrite counter pollution by
* unused, long-lived LWBs.
*/
for (; lwb != NULL; lwb = list_next(&zilog->zl_lwb_list, lwb)) {
if (lwb->lwb_fastwrite && !lwb->lwb_write_zio) {
metaslab_fastwrite_unmark(zilog->zl_spa, &lwb->lwb_blk);
lwb->lwb_fastwrite = 0;
}
}
mutex_exit(&zilog->zl_lock);
}
static int
zil_lwb_cons(void *vbuf, void *unused, int kmflag)
{
(void) unused, (void) kmflag;
lwb_t *lwb = vbuf;
list_create(&lwb->lwb_itxs, sizeof (itx_t), offsetof(itx_t, itx_node));
list_create(&lwb->lwb_waiters, sizeof (zil_commit_waiter_t),
offsetof(zil_commit_waiter_t, zcw_node));
avl_create(&lwb->lwb_vdev_tree, zil_lwb_vdev_compare,
sizeof (zil_vdev_node_t), offsetof(zil_vdev_node_t, zv_node));
mutex_init(&lwb->lwb_vdev_lock, NULL, MUTEX_DEFAULT, NULL);
return (0);
}
static void
zil_lwb_dest(void *vbuf, void *unused)
{
(void) unused;
lwb_t *lwb = vbuf;
mutex_destroy(&lwb->lwb_vdev_lock);
avl_destroy(&lwb->lwb_vdev_tree);
list_destroy(&lwb->lwb_waiters);
list_destroy(&lwb->lwb_itxs);
}
void
zil_init(void)
{
zil_lwb_cache = kmem_cache_create("zil_lwb_cache",
sizeof (lwb_t), 0, zil_lwb_cons, zil_lwb_dest, NULL, NULL, NULL, 0);
zil_zcw_cache = kmem_cache_create("zil_zcw_cache",
sizeof (zil_commit_waiter_t), 0, NULL, NULL, NULL, NULL, NULL, 0);
zil_sums_init(&zil_sums_global);
zil_kstats_global = kstat_create("zfs", 0, "zil", "misc",
KSTAT_TYPE_NAMED, sizeof (zil_stats) / sizeof (kstat_named_t),
KSTAT_FLAG_VIRTUAL);
if (zil_kstats_global != NULL) {
zil_kstats_global->ks_data = &zil_stats;
zil_kstats_global->ks_update = zil_kstats_global_update;
zil_kstats_global->ks_private = NULL;
kstat_install(zil_kstats_global);
}
}
void
zil_fini(void)
{
kmem_cache_destroy(zil_zcw_cache);
kmem_cache_destroy(zil_lwb_cache);
if (zil_kstats_global != NULL) {
kstat_delete(zil_kstats_global);
zil_kstats_global = NULL;
}
zil_sums_fini(&zil_sums_global);
}
void
zil_set_sync(zilog_t *zilog, uint64_t sync)
{
zilog->zl_sync = sync;
}
void
zil_set_logbias(zilog_t *zilog, uint64_t logbias)
{
zilog->zl_logbias = logbias;
}
zilog_t *
zil_alloc(objset_t *os, zil_header_t *zh_phys)
{
zilog_t *zilog;
zilog = kmem_zalloc(sizeof (zilog_t), KM_SLEEP);
zilog->zl_header = zh_phys;
zilog->zl_os = os;
zilog->zl_spa = dmu_objset_spa(os);
zilog->zl_dmu_pool = dmu_objset_pool(os);
zilog->zl_destroy_txg = TXG_INITIAL - 1;
zilog->zl_logbias = dmu_objset_logbias(os);
zilog->zl_sync = dmu_objset_syncprop(os);
zilog->zl_dirty_max_txg = 0;
zilog->zl_last_lwb_opened = NULL;
zilog->zl_last_lwb_latency = 0;
zilog->zl_max_block_size = zil_maxblocksize;
mutex_init(&zilog->zl_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&zilog->zl_issuer_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&zilog->zl_lwb_io_lock, NULL, MUTEX_DEFAULT, NULL);
for (int i = 0; i < TXG_SIZE; i++) {
mutex_init(&zilog->zl_itxg[i].itxg_lock, NULL,
MUTEX_DEFAULT, NULL);
}
list_create(&zilog->zl_lwb_list, sizeof (lwb_t),
offsetof(lwb_t, lwb_node));
list_create(&zilog->zl_itx_commit_list, sizeof (itx_t),
offsetof(itx_t, itx_node));
cv_init(&zilog->zl_cv_suspend, NULL, CV_DEFAULT, NULL);
cv_init(&zilog->zl_lwb_io_cv, NULL, CV_DEFAULT, NULL);
return (zilog);
}
void
zil_free(zilog_t *zilog)
{
int i;
zilog->zl_stop_sync = 1;
ASSERT0(zilog->zl_suspend);
ASSERT0(zilog->zl_suspending);
ASSERT(list_is_empty(&zilog->zl_lwb_list));
list_destroy(&zilog->zl_lwb_list);
ASSERT(list_is_empty(&zilog->zl_itx_commit_list));
list_destroy(&zilog->zl_itx_commit_list);
for (i = 0; i < TXG_SIZE; i++) {
/*
* It's possible for an itx to be generated that doesn't dirty
* a txg (e.g. ztest TX_TRUNCATE). So there's no zil_clean()
* callback to remove the entry. We remove those here.
*
* Also free up the ziltest itxs.
*/
if (zilog->zl_itxg[i].itxg_itxs)
zil_itxg_clean(zilog->zl_itxg[i].itxg_itxs);
mutex_destroy(&zilog->zl_itxg[i].itxg_lock);
}
mutex_destroy(&zilog->zl_issuer_lock);
mutex_destroy(&zilog->zl_lock);
mutex_destroy(&zilog->zl_lwb_io_lock);
cv_destroy(&zilog->zl_cv_suspend);
cv_destroy(&zilog->zl_lwb_io_cv);
kmem_free(zilog, sizeof (zilog_t));
}
/*
* Open an intent log.
*/
zilog_t *
zil_open(objset_t *os, zil_get_data_t *get_data, zil_sums_t *zil_sums)
{
zilog_t *zilog = dmu_objset_zil(os);
ASSERT3P(zilog->zl_get_data, ==, NULL);
ASSERT3P(zilog->zl_last_lwb_opened, ==, NULL);
ASSERT(list_is_empty(&zilog->zl_lwb_list));
zilog->zl_get_data = get_data;
zilog->zl_sums = zil_sums;
return (zilog);
}
/*
* Close an intent log.
*/
void
zil_close(zilog_t *zilog)
{
lwb_t *lwb;
uint64_t txg;
if (!dmu_objset_is_snapshot(zilog->zl_os)) {
zil_commit(zilog, 0);
} else {
ASSERT3P(list_tail(&zilog->zl_lwb_list), ==, NULL);
ASSERT0(zilog->zl_dirty_max_txg);
ASSERT3B(zilog_is_dirty(zilog), ==, B_FALSE);
}
mutex_enter(&zilog->zl_lock);
lwb = list_tail(&zilog->zl_lwb_list);
if (lwb == NULL)
txg = zilog->zl_dirty_max_txg;
else
txg = MAX(zilog->zl_dirty_max_txg, lwb->lwb_max_txg);
mutex_exit(&zilog->zl_lock);
/*
* zl_lwb_max_issued_txg may be larger than lwb_max_txg. It depends
* on the time when the dmu_tx transaction is assigned in
* zil_lwb_write_issue().
*/
mutex_enter(&zilog->zl_lwb_io_lock);
txg = MAX(zilog->zl_lwb_max_issued_txg, txg);
mutex_exit(&zilog->zl_lwb_io_lock);
/*
* We need to use txg_wait_synced() to wait until that txg is synced.
* zil_sync() will guarantee all lwbs up to that txg have been
* written out, flushed, and cleaned.
*/
if (txg != 0)
txg_wait_synced(zilog->zl_dmu_pool, txg);
if (zilog_is_dirty(zilog))
zfs_dbgmsg("zil (%px) is dirty, txg %llu", zilog,
(u_longlong_t)txg);
if (txg < spa_freeze_txg(zilog->zl_spa))
VERIFY(!zilog_is_dirty(zilog));
zilog->zl_get_data = NULL;
/*
* We should have only one lwb left on the list; remove it now.
*/
mutex_enter(&zilog->zl_lock);
lwb = list_head(&zilog->zl_lwb_list);
if (lwb != NULL) {
ASSERT3P(lwb, ==, list_tail(&zilog->zl_lwb_list));
ASSERT3S(lwb->lwb_state, !=, LWB_STATE_ISSUED);
if (lwb->lwb_fastwrite)
metaslab_fastwrite_unmark(zilog->zl_spa, &lwb->lwb_blk);
list_remove(&zilog->zl_lwb_list, lwb);
zio_buf_free(lwb->lwb_buf, lwb->lwb_sz);
zil_free_lwb(zilog, lwb);
}
mutex_exit(&zilog->zl_lock);
}
static const char *suspend_tag = "zil suspending";
/*
* Suspend an intent log. While in suspended mode, we still honor
* synchronous semantics, but we rely on txg_wait_synced() to do it.
* On old version pools, we suspend the log briefly when taking a
* snapshot so that it will have an empty intent log.
*
* Long holds are not really intended to be used the way we do here --
* held for such a short time. A concurrent caller of dsl_dataset_long_held()
* could fail. Therefore we take pains to only put a long hold if it is
* actually necessary. Fortunately, it will only be necessary if the
* objset is currently mounted (or the ZVOL equivalent). In that case it
* will already have a long hold, so we are not really making things any worse.
*
* Ideally, we would locate the existing long-holder (i.e. the zfsvfs_t or
* zvol_state_t), and use their mechanism to prevent their hold from being
* dropped (e.g. VFS_HOLD()). However, that would be even more pain for
* very little gain.
*
* if cookiep == NULL, this does both the suspend & resume.
* Otherwise, it returns with the dataset "long held", and the cookie
* should be passed into zil_resume().
*/
int
zil_suspend(const char *osname, void **cookiep)
{
objset_t *os;
zilog_t *zilog;
const zil_header_t *zh;
int error;
error = dmu_objset_hold(osname, suspend_tag, &os);
if (error != 0)
return (error);
zilog = dmu_objset_zil(os);
mutex_enter(&zilog->zl_lock);
zh = zilog->zl_header;
if (zh->zh_flags & ZIL_REPLAY_NEEDED) { /* unplayed log */
mutex_exit(&zilog->zl_lock);
dmu_objset_rele(os, suspend_tag);
return (SET_ERROR(EBUSY));
}
/*
* Don't put a long hold in the cases where we can avoid it. This
* is when there is no cookie so we are doing a suspend & resume
* (i.e. called from zil_vdev_offline()), and there's nothing to do
* for the suspend because it's already suspended, or there's no ZIL.
*/
if (cookiep == NULL && !zilog->zl_suspending &&
(zilog->zl_suspend > 0 || BP_IS_HOLE(&zh->zh_log))) {
mutex_exit(&zilog->zl_lock);
dmu_objset_rele(os, suspend_tag);
return (0);
}
dsl_dataset_long_hold(dmu_objset_ds(os), suspend_tag);
dsl_pool_rele(dmu_objset_pool(os), suspend_tag);
zilog->zl_suspend++;
if (zilog->zl_suspend > 1) {
/*
* Someone else is already suspending it.
* Just wait for them to finish.
*/
while (zilog->zl_suspending)
cv_wait(&zilog->zl_cv_suspend, &zilog->zl_lock);
mutex_exit(&zilog->zl_lock);
if (cookiep == NULL)
zil_resume(os);
else
*cookiep = os;
return (0);
}
/*
* If there is no pointer to an on-disk block, this ZIL must not
* be active (e.g. filesystem not mounted), so there's nothing
* to clean up.
*/
if (BP_IS_HOLE(&zh->zh_log)) {
ASSERT(cookiep != NULL); /* fast path already handled */
*cookiep = os;
mutex_exit(&zilog->zl_lock);
return (0);
}
/*
* The ZIL has work to do. Ensure that the associated encryption
* key will remain mapped while we are committing the log by
* grabbing a reference to it. If the key isn't loaded we have no
* choice but to return an error until the wrapping key is loaded.
*/
if (os->os_encrypted &&
dsl_dataset_create_key_mapping(dmu_objset_ds(os)) != 0) {
zilog->zl_suspend--;
mutex_exit(&zilog->zl_lock);
dsl_dataset_long_rele(dmu_objset_ds(os), suspend_tag);
dsl_dataset_rele(dmu_objset_ds(os), suspend_tag);
return (SET_ERROR(EACCES));
}
zilog->zl_suspending = B_TRUE;
mutex_exit(&zilog->zl_lock);
/*
* We need to use zil_commit_impl to ensure we wait for all
* LWB_STATE_OPENED and LWB_STATE_ISSUED lwbs to be committed
* to disk before proceeding. If we used zil_commit instead, it
* would just call txg_wait_synced(), because zl_suspend is set.
* txg_wait_synced() doesn't wait for these lwb's to be
* LWB_STATE_FLUSH_DONE before returning.
*/
zil_commit_impl(zilog, 0);
/*
* Now that we've ensured all lwb's are LWB_STATE_FLUSH_DONE, we
* use txg_wait_synced() to ensure the data from the zilog has
* migrated to the main pool before calling zil_destroy().
*/
txg_wait_synced(zilog->zl_dmu_pool, 0);
zil_destroy(zilog, B_FALSE);
mutex_enter(&zilog->zl_lock);
zilog->zl_suspending = B_FALSE;
cv_broadcast(&zilog->zl_cv_suspend);
mutex_exit(&zilog->zl_lock);
if (os->os_encrypted)
dsl_dataset_remove_key_mapping(dmu_objset_ds(os));
if (cookiep == NULL)
zil_resume(os);
else
*cookiep = os;
return (0);
}
void
zil_resume(void *cookie)
{
objset_t *os = cookie;
zilog_t *zilog = dmu_objset_zil(os);
mutex_enter(&zilog->zl_lock);
ASSERT(zilog->zl_suspend != 0);
zilog->zl_suspend--;
mutex_exit(&zilog->zl_lock);
dsl_dataset_long_rele(dmu_objset_ds(os), suspend_tag);
dsl_dataset_rele(dmu_objset_ds(os), suspend_tag);
}
typedef struct zil_replay_arg {
zil_replay_func_t *const *zr_replay;
void *zr_arg;
boolean_t zr_byteswap;
char *zr_lr;
} zil_replay_arg_t;
static int
zil_replay_error(zilog_t *zilog, const lr_t *lr, int error)
{
char name[ZFS_MAX_DATASET_NAME_LEN];
zilog->zl_replaying_seq--; /* didn't actually replay this one */
dmu_objset_name(zilog->zl_os, name);
cmn_err(CE_WARN, "ZFS replay transaction error %d, "
"dataset %s, seq 0x%llx, txtype %llu %s\n", error, name,
(u_longlong_t)lr->lrc_seq,
(u_longlong_t)(lr->lrc_txtype & ~TX_CI),
(lr->lrc_txtype & TX_CI) ? "CI" : "");
return (error);
}
static int
zil_replay_log_record(zilog_t *zilog, const lr_t *lr, void *zra,
uint64_t claim_txg)
{
zil_replay_arg_t *zr = zra;
const zil_header_t *zh = zilog->zl_header;
uint64_t reclen = lr->lrc_reclen;
uint64_t txtype = lr->lrc_txtype;
int error = 0;
zilog->zl_replaying_seq = lr->lrc_seq;
if (lr->lrc_seq <= zh->zh_replay_seq) /* already replayed */
return (0);
if (lr->lrc_txg < claim_txg) /* already committed */
return (0);
/* Strip case-insensitive bit, still present in log record */
txtype &= ~TX_CI;
if (txtype == 0 || txtype >= TX_MAX_TYPE)
return (zil_replay_error(zilog, lr, EINVAL));
/*
* If this record type can be logged out of order, the object
* (lr_foid) may no longer exist. That's legitimate, not an error.
*/
if (TX_OOO(txtype)) {
error = dmu_object_info(zilog->zl_os,
LR_FOID_GET_OBJ(((lr_ooo_t *)lr)->lr_foid), NULL);
if (error == ENOENT || error == EEXIST)
return (0);
}
/*
* Make a copy of the data so we can revise and extend it.
*/
memcpy(zr->zr_lr, lr, reclen);
/*
* If this is a TX_WRITE with a blkptr, suck in the data.
*/
if (txtype == TX_WRITE && reclen == sizeof (lr_write_t)) {
error = zil_read_log_data(zilog, (lr_write_t *)lr,
zr->zr_lr + reclen);
if (error != 0)
return (zil_replay_error(zilog, lr, error));
}
/*
* The log block containing this lr may have been byteswapped
* so that we can easily examine common fields like lrc_txtype.
* However, the log is a mix of different record types, and only the
* replay vectors know how to byteswap their records. Therefore, if
* the lr was byteswapped, undo it before invoking the replay vector.
*/
if (zr->zr_byteswap)
byteswap_uint64_array(zr->zr_lr, reclen);
/*
* We must now do two things atomically: replay this log record,
* and update the log header sequence number to reflect the fact that
* we did so. At the end of each replay function the sequence number
* is updated if we are in replay mode.
*/
error = zr->zr_replay[txtype](zr->zr_arg, zr->zr_lr, zr->zr_byteswap);
if (error != 0) {
/*
* The DMU's dnode layer doesn't see removes until the txg
* commits, so a subsequent claim can spuriously fail with
* EEXIST. So if we receive any error we try syncing out
* any removes then retry the transaction. Note that we
* specify B_FALSE for byteswap now, so we don't do it twice.
*/
txg_wait_synced(spa_get_dsl(zilog->zl_spa), 0);
error = zr->zr_replay[txtype](zr->zr_arg, zr->zr_lr, B_FALSE);
if (error != 0)
return (zil_replay_error(zilog, lr, error));
}
return (0);
}
static int
zil_incr_blks(zilog_t *zilog, const blkptr_t *bp, void *arg, uint64_t claim_txg)
{
(void) bp, (void) arg, (void) claim_txg;
zilog->zl_replay_blks++;
return (0);
}
/*
* If this dataset has a non-empty intent log, replay it and destroy it.
* Return B_TRUE if there were any entries to replay.
*/
boolean_t
zil_replay(objset_t *os, void *arg,
zil_replay_func_t *const replay_func[TX_MAX_TYPE])
{
zilog_t *zilog = dmu_objset_zil(os);
const zil_header_t *zh = zilog->zl_header;
zil_replay_arg_t zr;
if ((zh->zh_flags & ZIL_REPLAY_NEEDED) == 0) {
return (zil_destroy(zilog, B_TRUE));
}
zr.zr_replay = replay_func;
zr.zr_arg = arg;
zr.zr_byteswap = BP_SHOULD_BYTESWAP(&zh->zh_log);
zr.zr_lr = vmem_alloc(2 * SPA_MAXBLOCKSIZE, KM_SLEEP);
/*
* Wait for in-progress removes to sync before starting replay.
*/
txg_wait_synced(zilog->zl_dmu_pool, 0);
zilog->zl_replay = B_TRUE;
zilog->zl_replay_time = ddi_get_lbolt();
ASSERT(zilog->zl_replay_blks == 0);
(void) zil_parse(zilog, zil_incr_blks, zil_replay_log_record, &zr,
zh->zh_claim_txg, B_TRUE);
vmem_free(zr.zr_lr, 2 * SPA_MAXBLOCKSIZE);
zil_destroy(zilog, B_FALSE);
txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg);
zilog->zl_replay = B_FALSE;
return (B_TRUE);
}
boolean_t
zil_replaying(zilog_t *zilog, dmu_tx_t *tx)
{
if (zilog->zl_sync == ZFS_SYNC_DISABLED)
return (B_TRUE);
if (zilog->zl_replay) {
dsl_dataset_dirty(dmu_objset_ds(zilog->zl_os), tx);
zilog->zl_replayed_seq[dmu_tx_get_txg(tx) & TXG_MASK] =
zilog->zl_replaying_seq;
return (B_TRUE);
}
return (B_FALSE);
}
int
zil_reset(const char *osname, void *arg)
{
(void) arg;
int error = zil_suspend(osname, NULL);
/* EACCES means crypto key not loaded */
if ((error == EACCES) || (error == EBUSY))
return (SET_ERROR(error));
if (error != 0)
return (SET_ERROR(EEXIST));
return (0);
}
EXPORT_SYMBOL(zil_alloc);
EXPORT_SYMBOL(zil_free);
EXPORT_SYMBOL(zil_open);
EXPORT_SYMBOL(zil_close);
EXPORT_SYMBOL(zil_replay);
EXPORT_SYMBOL(zil_replaying);
EXPORT_SYMBOL(zil_destroy);
EXPORT_SYMBOL(zil_destroy_sync);
EXPORT_SYMBOL(zil_itx_create);
EXPORT_SYMBOL(zil_itx_destroy);
EXPORT_SYMBOL(zil_itx_assign);
EXPORT_SYMBOL(zil_commit);
EXPORT_SYMBOL(zil_claim);
EXPORT_SYMBOL(zil_check_log_chain);
EXPORT_SYMBOL(zil_sync);
EXPORT_SYMBOL(zil_clean);
EXPORT_SYMBOL(zil_suspend);
EXPORT_SYMBOL(zil_resume);
EXPORT_SYMBOL(zil_lwb_add_block);
EXPORT_SYMBOL(zil_bp_tree_add);
EXPORT_SYMBOL(zil_set_sync);
EXPORT_SYMBOL(zil_set_logbias);
EXPORT_SYMBOL(zil_sums_init);
EXPORT_SYMBOL(zil_sums_fini);
EXPORT_SYMBOL(zil_kstat_values_update);
ZFS_MODULE_PARAM(zfs, zfs_, commit_timeout_pct, UINT, ZMOD_RW,
"ZIL block open timeout percentage");
ZFS_MODULE_PARAM(zfs_zil, zil_, min_commit_timeout, U64, ZMOD_RW,
"Minimum delay we care for ZIL block commit");
ZFS_MODULE_PARAM(zfs_zil, zil_, replay_disable, INT, ZMOD_RW,
"Disable intent logging replay");
ZFS_MODULE_PARAM(zfs_zil, zil_, nocacheflush, INT, ZMOD_RW,
"Disable ZIL cache flushes");
ZFS_MODULE_PARAM(zfs_zil, zil_, slog_bulk, U64, ZMOD_RW,
"Limit in bytes slog sync writes per commit");
ZFS_MODULE_PARAM(zfs_zil, zil_, maxblocksize, UINT, ZMOD_RW,
"Limit in bytes of ZIL log block size");
diff --git a/sys/contrib/openzfs/tests/runfiles/common.run b/sys/contrib/openzfs/tests/runfiles/common.run
index bb0aacae456c..55991cfeaf78 100644
--- a/sys/contrib/openzfs/tests/runfiles/common.run
+++ b/sys/contrib/openzfs/tests/runfiles/common.run
@@ -1,977 +1,978 @@
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
# This run file contains all of the common functional tests. When
# adding a new test consider also adding it to the sanity.run file
# if the new test runs to completion in only a few seconds.
#
# Approximate run time: 4-5 hours
#
[DEFAULT]
pre = setup
quiet = False
pre_user = root
user = root
timeout = 600
post_user = root
post = cleanup
failsafe_user = root
failsafe = callbacks/zfs_failsafe
outputdir = /var/tmp/test_results
tags = ['functional']
[tests/functional/acl/off]
tests = ['dosmode', 'posixmode']
tags = ['functional', 'acl']
[tests/functional/alloc_class]
tests = ['alloc_class_001_pos', 'alloc_class_002_neg', 'alloc_class_003_pos',
'alloc_class_004_pos', 'alloc_class_005_pos', 'alloc_class_006_pos',
'alloc_class_007_pos', 'alloc_class_008_pos', 'alloc_class_009_pos',
'alloc_class_010_pos', 'alloc_class_011_neg', 'alloc_class_012_pos',
'alloc_class_013_pos']
tags = ['functional', 'alloc_class']
[tests/functional/append]
tests = ['file_append', 'threadsappend_001_pos']
tags = ['functional', 'append']
[tests/functional/arc]
tests = ['dbufstats_001_pos', 'dbufstats_002_pos', 'dbufstats_003_pos',
'arcstats_runtime_tuning']
tags = ['functional', 'arc']
[tests/functional/atime]
tests = ['atime_001_pos', 'atime_002_neg', 'root_atime_off', 'root_atime_on']
tags = ['functional', 'atime']
[tests/functional/bootfs]
tests = ['bootfs_001_pos', 'bootfs_002_neg', 'bootfs_003_pos',
'bootfs_004_neg', 'bootfs_005_neg', 'bootfs_006_pos', 'bootfs_007_pos',
'bootfs_008_pos']
tags = ['functional', 'bootfs']
[tests/functional/btree]
tests = ['btree_positive', 'btree_negative']
tags = ['functional', 'btree']
pre =
post =
[tests/functional/cache]
tests = ['cache_001_pos', 'cache_002_pos', 'cache_003_pos', 'cache_004_neg',
'cache_005_neg', 'cache_006_pos', 'cache_007_neg', 'cache_008_neg',
'cache_009_pos', 'cache_010_pos', 'cache_011_pos', 'cache_012_pos']
tags = ['functional', 'cache']
[tests/functional/cachefile]
tests = ['cachefile_001_pos', 'cachefile_002_pos', 'cachefile_003_pos',
'cachefile_004_pos']
tags = ['functional', 'cachefile']
[tests/functional/casenorm]
tests = ['case_all_values', 'norm_all_values', 'mixed_create_failure',
'sensitive_none_lookup', 'sensitive_none_delete',
'sensitive_formd_lookup', 'sensitive_formd_delete',
'insensitive_none_lookup', 'insensitive_none_delete',
'insensitive_formd_lookup', 'insensitive_formd_delete',
'mixed_none_lookup', 'mixed_none_lookup_ci', 'mixed_none_delete',
'mixed_formd_lookup', 'mixed_formd_lookup_ci', 'mixed_formd_delete']
tags = ['functional', 'casenorm']
[tests/functional/channel_program/lua_core]
tests = ['tst.args_to_lua', 'tst.divide_by_zero', 'tst.exists',
'tst.integer_illegal', 'tst.integer_overflow', 'tst.language_functions_neg',
'tst.language_functions_pos', 'tst.large_prog', 'tst.libraries',
'tst.memory_limit', 'tst.nested_neg', 'tst.nested_pos', 'tst.nvlist_to_lua',
'tst.recursive_neg', 'tst.recursive_pos', 'tst.return_large',
'tst.return_nvlist_neg', 'tst.return_nvlist_pos',
'tst.return_recursive_table', 'tst.stack_gsub', 'tst.timeout']
tags = ['functional', 'channel_program', 'lua_core']
[tests/functional/channel_program/synctask_core]
tests = ['tst.destroy_fs', 'tst.destroy_snap', 'tst.get_count_and_limit',
'tst.get_index_props', 'tst.get_mountpoint', 'tst.get_neg',
'tst.get_number_props', 'tst.get_string_props', 'tst.get_type',
'tst.get_userquota', 'tst.get_written', 'tst.inherit', 'tst.list_bookmarks',
'tst.list_children', 'tst.list_clones', 'tst.list_holds',
'tst.list_snapshots', 'tst.list_system_props',
'tst.list_user_props', 'tst.parse_args_neg','tst.promote_conflict',
'tst.promote_multiple', 'tst.promote_simple', 'tst.rollback_mult',
'tst.rollback_one', 'tst.set_props', 'tst.snapshot_destroy', 'tst.snapshot_neg',
- 'tst.snapshot_recursive', 'tst.snapshot_simple',
+ 'tst.snapshot_recursive', 'tst.snapshot_rename', 'tst.snapshot_simple',
'tst.bookmark.create', 'tst.bookmark.copy',
'tst.terminate_by_signal'
]
tags = ['functional', 'channel_program', 'synctask_core']
[tests/functional/checksum]
tests = ['run_edonr_test', 'run_sha2_test', 'run_skein_test', 'run_blake3_test',
'filetest_001_pos', 'filetest_002_pos']
tags = ['functional', 'checksum']
[tests/functional/clean_mirror]
tests = [ 'clean_mirror_001_pos', 'clean_mirror_002_pos',
'clean_mirror_003_pos', 'clean_mirror_004_pos']
tags = ['functional', 'clean_mirror']
[tests/functional/cli_root/zdb]
tests = ['zdb_002_pos', 'zdb_003_pos', 'zdb_004_pos', 'zdb_005_pos',
'zdb_006_pos', 'zdb_args_neg', 'zdb_args_pos',
'zdb_block_size_histogram', 'zdb_checksum', 'zdb_decompress',
'zdb_display_block', 'zdb_encrypted', 'zdb_label_checksum',
'zdb_object_range_neg', 'zdb_object_range_pos', 'zdb_objset_id',
'zdb_decompress_zstd', 'zdb_recover', 'zdb_recover_2']
pre =
post =
tags = ['functional', 'cli_root', 'zdb']
[tests/functional/cli_root/zfs]
tests = ['zfs_001_neg', 'zfs_002_pos']
tags = ['functional', 'cli_root', 'zfs']
[tests/functional/cli_root/zfs_bookmark]
tests = ['zfs_bookmark_cliargs']
tags = ['functional', 'cli_root', 'zfs_bookmark']
[tests/functional/cli_root/zfs_change-key]
tests = ['zfs_change-key', 'zfs_change-key_child', 'zfs_change-key_format',
'zfs_change-key_inherit', 'zfs_change-key_load', 'zfs_change-key_location',
'zfs_change-key_pbkdf2iters', 'zfs_change-key_clones']
tags = ['functional', 'cli_root', 'zfs_change-key']
[tests/functional/cli_root/zfs_clone]
tests = ['zfs_clone_001_neg', 'zfs_clone_002_pos', 'zfs_clone_003_pos',
'zfs_clone_004_pos', 'zfs_clone_005_pos', 'zfs_clone_006_pos',
'zfs_clone_007_pos', 'zfs_clone_008_neg', 'zfs_clone_009_neg',
- 'zfs_clone_010_pos', 'zfs_clone_encrypted', 'zfs_clone_deeply_nested']
+ 'zfs_clone_010_pos', 'zfs_clone_encrypted', 'zfs_clone_deeply_nested',
+ 'zfs_clone_rm_nested']
tags = ['functional', 'cli_root', 'zfs_clone']
[tests/functional/cli_root/zfs_copies]
tests = ['zfs_copies_001_pos', 'zfs_copies_002_pos', 'zfs_copies_003_pos',
'zfs_copies_004_neg', 'zfs_copies_005_neg', 'zfs_copies_006_pos']
tags = ['functional', 'cli_root', 'zfs_copies']
[tests/functional/cli_root/zfs_create]
tests = ['zfs_create_001_pos', 'zfs_create_002_pos', 'zfs_create_003_pos',
'zfs_create_004_pos', 'zfs_create_005_pos', 'zfs_create_006_pos',
'zfs_create_007_pos', 'zfs_create_008_neg', 'zfs_create_009_neg',
'zfs_create_010_neg', 'zfs_create_011_pos', 'zfs_create_012_pos',
'zfs_create_013_pos', 'zfs_create_014_pos', 'zfs_create_encrypted',
'zfs_create_crypt_combos', 'zfs_create_dryrun', 'zfs_create_nomount',
'zfs_create_verbose']
tags = ['functional', 'cli_root', 'zfs_create']
[tests/functional/cli_root/zfs_destroy]
tests = ['zfs_clone_livelist_condense_and_disable',
'zfs_clone_livelist_condense_races', 'zfs_clone_livelist_dedup',
'zfs_destroy_001_pos', 'zfs_destroy_002_pos', 'zfs_destroy_003_pos',
'zfs_destroy_004_pos', 'zfs_destroy_005_neg', 'zfs_destroy_006_neg',
'zfs_destroy_007_neg', 'zfs_destroy_008_pos', 'zfs_destroy_009_pos',
'zfs_destroy_010_pos', 'zfs_destroy_011_pos', 'zfs_destroy_012_pos',
'zfs_destroy_013_neg', 'zfs_destroy_014_pos', 'zfs_destroy_015_pos',
'zfs_destroy_016_pos', 'zfs_destroy_clone_livelist',
'zfs_destroy_dev_removal', 'zfs_destroy_dev_removal_condense']
tags = ['functional', 'cli_root', 'zfs_destroy']
[tests/functional/cli_root/zfs_diff]
tests = ['zfs_diff_changes', 'zfs_diff_cliargs', 'zfs_diff_timestamp',
'zfs_diff_types', 'zfs_diff_encrypted', 'zfs_diff_mangle']
tags = ['functional', 'cli_root', 'zfs_diff']
[tests/functional/cli_root/zfs_get]
tests = ['zfs_get_001_pos', 'zfs_get_002_pos', 'zfs_get_003_pos',
'zfs_get_004_pos', 'zfs_get_005_neg', 'zfs_get_006_neg', 'zfs_get_007_neg',
'zfs_get_008_pos', 'zfs_get_009_pos', 'zfs_get_010_neg']
tags = ['functional', 'cli_root', 'zfs_get']
[tests/functional/cli_root/zfs_ids_to_path]
tests = ['zfs_ids_to_path_001_pos']
tags = ['functional', 'cli_root', 'zfs_ids_to_path']
[tests/functional/cli_root/zfs_inherit]
tests = ['zfs_inherit_001_neg', 'zfs_inherit_002_neg', 'zfs_inherit_003_pos',
'zfs_inherit_mountpoint']
tags = ['functional', 'cli_root', 'zfs_inherit']
[tests/functional/cli_root/zfs_load-key]
tests = ['zfs_load-key', 'zfs_load-key_all', 'zfs_load-key_file',
'zfs_load-key_https', 'zfs_load-key_location', 'zfs_load-key_noop',
'zfs_load-key_recursive']
tags = ['functional', 'cli_root', 'zfs_load-key']
[tests/functional/cli_root/zfs_mount]
tests = ['zfs_mount_001_pos', 'zfs_mount_002_pos', 'zfs_mount_003_pos',
'zfs_mount_004_pos', 'zfs_mount_005_pos', 'zfs_mount_007_pos',
'zfs_mount_009_neg', 'zfs_mount_010_neg', 'zfs_mount_011_neg',
'zfs_mount_012_pos', 'zfs_mount_all_001_pos', 'zfs_mount_encrypted',
'zfs_mount_remount', 'zfs_mount_all_fail', 'zfs_mount_all_mountpoints',
'zfs_mount_test_race']
tags = ['functional', 'cli_root', 'zfs_mount']
[tests/functional/cli_root/zfs_program]
tests = ['zfs_program_json']
tags = ['functional', 'cli_root', 'zfs_program']
[tests/functional/cli_root/zfs_promote]
tests = ['zfs_promote_001_pos', 'zfs_promote_002_pos', 'zfs_promote_003_pos',
'zfs_promote_004_pos', 'zfs_promote_005_pos', 'zfs_promote_006_neg',
'zfs_promote_007_neg', 'zfs_promote_008_pos', 'zfs_promote_encryptionroot']
tags = ['functional', 'cli_root', 'zfs_promote']
[tests/functional/cli_root/zfs_property]
tests = ['zfs_written_property_001_pos']
tags = ['functional', 'cli_root', 'zfs_property']
[tests/functional/cli_root/zfs_receive]
tests = ['zfs_receive_001_pos', 'zfs_receive_002_pos', 'zfs_receive_003_pos',
'zfs_receive_004_neg', 'zfs_receive_005_neg', 'zfs_receive_006_pos',
'zfs_receive_007_neg', 'zfs_receive_008_pos', 'zfs_receive_009_neg',
'zfs_receive_010_pos', 'zfs_receive_011_pos', 'zfs_receive_012_pos',
'zfs_receive_013_pos', 'zfs_receive_014_pos', 'zfs_receive_015_pos',
'zfs_receive_016_pos', 'receive-o-x_props_override',
'receive-o-x_props_aliases',
'zfs_receive_from_encrypted', 'zfs_receive_to_encrypted',
'zfs_receive_raw', 'zfs_receive_raw_incremental', 'zfs_receive_-e',
'zfs_receive_raw_-d', 'zfs_receive_from_zstd', 'zfs_receive_new_props',
'zfs_receive_-wR-encrypted-mix', 'zfs_receive_corrective',
'zfs_receive_compressed_corrective', 'zfs_receive_large_block_corrective']
tags = ['functional', 'cli_root', 'zfs_receive']
[tests/functional/cli_root/zfs_rename]
tests = ['zfs_rename_001_pos', 'zfs_rename_002_pos', 'zfs_rename_003_pos',
'zfs_rename_004_neg', 'zfs_rename_005_neg', 'zfs_rename_006_pos',
'zfs_rename_007_pos', 'zfs_rename_008_pos', 'zfs_rename_009_neg',
'zfs_rename_010_neg', 'zfs_rename_011_pos', 'zfs_rename_012_neg',
'zfs_rename_013_pos', 'zfs_rename_014_neg', 'zfs_rename_encrypted_child',
'zfs_rename_to_encrypted', 'zfs_rename_mountpoint', 'zfs_rename_nounmount']
tags = ['functional', 'cli_root', 'zfs_rename']
[tests/functional/cli_root/zfs_reservation]
tests = ['zfs_reservation_001_pos', 'zfs_reservation_002_pos']
tags = ['functional', 'cli_root', 'zfs_reservation']
[tests/functional/cli_root/zfs_rollback]
tests = ['zfs_rollback_001_pos', 'zfs_rollback_002_pos',
'zfs_rollback_003_neg', 'zfs_rollback_004_neg']
tags = ['functional', 'cli_root', 'zfs_rollback']
[tests/functional/cli_root/zfs_send]
tests = ['zfs_send_001_pos', 'zfs_send_002_pos', 'zfs_send_003_pos',
'zfs_send_004_neg', 'zfs_send_005_pos', 'zfs_send_006_pos',
- 'zfs_send_007_pos', 'zfs_send_encrypted', 'zfs_send_raw',
- 'zfs_send_sparse', 'zfs_send-b', 'zfs_send_skip_missing']
+ 'zfs_send_007_pos', 'zfs_send_encrypted', 'zfs_send_encrypted_unloaded',
+ 'zfs_send_raw', 'zfs_send_sparse', 'zfs_send-b', 'zfs_send_skip_missing']
tags = ['functional', 'cli_root', 'zfs_send']
[tests/functional/cli_root/zfs_set]
tests = ['cache_001_pos', 'cache_002_neg', 'canmount_001_pos',
'canmount_002_pos', 'canmount_003_pos', 'canmount_004_pos',
'checksum_001_pos', 'compression_001_pos', 'mountpoint_001_pos',
'mountpoint_002_pos', 'reservation_001_neg', 'user_property_002_pos',
'share_mount_001_neg', 'snapdir_001_pos', 'onoffs_001_pos',
'user_property_001_pos', 'user_property_003_neg', 'readonly_001_pos',
'user_property_004_pos', 'version_001_neg', 'zfs_set_001_neg',
'zfs_set_002_neg', 'zfs_set_003_neg', 'property_alias_001_pos',
'mountpoint_003_pos', 'ro_props_001_pos', 'zfs_set_keylocation',
'zfs_set_feature_activation']
tags = ['functional', 'cli_root', 'zfs_set']
[tests/functional/cli_root/zfs_share]
tests = ['zfs_share_001_pos', 'zfs_share_002_pos', 'zfs_share_003_pos',
'zfs_share_004_pos', 'zfs_share_006_pos', 'zfs_share_008_neg',
'zfs_share_010_neg', 'zfs_share_011_pos', 'zfs_share_concurrent_shares']
tags = ['functional', 'cli_root', 'zfs_share']
[tests/functional/cli_root/zfs_snapshot]
tests = ['zfs_snapshot_001_neg', 'zfs_snapshot_002_neg',
'zfs_snapshot_003_neg', 'zfs_snapshot_004_neg', 'zfs_snapshot_005_neg',
'zfs_snapshot_006_pos', 'zfs_snapshot_007_neg', 'zfs_snapshot_008_neg',
'zfs_snapshot_009_pos']
tags = ['functional', 'cli_root', 'zfs_snapshot']
[tests/functional/cli_root/zfs_unload-key]
tests = ['zfs_unload-key', 'zfs_unload-key_all', 'zfs_unload-key_recursive']
tags = ['functional', 'cli_root', 'zfs_unload-key']
[tests/functional/cli_root/zfs_unmount]
tests = ['zfs_unmount_001_pos', 'zfs_unmount_002_pos', 'zfs_unmount_003_pos',
'zfs_unmount_004_pos', 'zfs_unmount_005_pos', 'zfs_unmount_006_pos',
'zfs_unmount_007_neg', 'zfs_unmount_008_neg', 'zfs_unmount_009_pos',
'zfs_unmount_all_001_pos', 'zfs_unmount_nested', 'zfs_unmount_unload_keys']
tags = ['functional', 'cli_root', 'zfs_unmount']
[tests/functional/cli_root/zfs_unshare]
tests = ['zfs_unshare_001_pos', 'zfs_unshare_002_pos', 'zfs_unshare_003_pos',
'zfs_unshare_004_neg', 'zfs_unshare_005_neg', 'zfs_unshare_006_pos',
- 'zfs_unshare_007_pos', 'zfs_unshare_008_pos']
+ 'zfs_unshare_007_pos']
tags = ['functional', 'cli_root', 'zfs_unshare']
[tests/functional/cli_root/zfs_upgrade]
tests = ['zfs_upgrade_001_pos', 'zfs_upgrade_002_pos', 'zfs_upgrade_003_pos',
'zfs_upgrade_004_pos', 'zfs_upgrade_005_pos', 'zfs_upgrade_006_neg',
'zfs_upgrade_007_neg']
tags = ['functional', 'cli_root', 'zfs_upgrade']
[tests/functional/cli_root/zfs_wait]
tests = ['zfs_wait_deleteq', 'zfs_wait_getsubopt']
tags = ['functional', 'cli_root', 'zfs_wait']
[tests/functional/cli_root/zhack]
tests = ['zhack_label_checksum']
pre =
post =
tags = ['functional', 'cli_root', 'zhack']
[tests/functional/cli_root/zpool]
tests = ['zpool_001_neg', 'zpool_002_pos', 'zpool_003_pos', 'zpool_colors']
tags = ['functional', 'cli_root', 'zpool']
[tests/functional/cli_root/zpool_add]
tests = ['zpool_add_001_pos', 'zpool_add_002_pos', 'zpool_add_003_pos',
'zpool_add_004_pos', 'zpool_add_006_pos', 'zpool_add_007_neg',
'zpool_add_008_neg', 'zpool_add_009_neg', 'zpool_add_010_pos',
'add-o_ashift', 'add_prop_ashift', 'zpool_add_dryrun_output']
tags = ['functional', 'cli_root', 'zpool_add']
[tests/functional/cli_root/zpool_attach]
tests = ['zpool_attach_001_neg', 'attach-o_ashift']
tags = ['functional', 'cli_root', 'zpool_attach']
[tests/functional/cli_root/zpool_clear]
tests = ['zpool_clear_001_pos', 'zpool_clear_002_neg', 'zpool_clear_003_neg',
'zpool_clear_readonly']
tags = ['functional', 'cli_root', 'zpool_clear']
[tests/functional/cli_root/zpool_create]
tests = ['zpool_create_001_pos', 'zpool_create_002_pos',
'zpool_create_003_pos', 'zpool_create_004_pos', 'zpool_create_005_pos',
'zpool_create_006_pos', 'zpool_create_007_neg', 'zpool_create_008_pos',
'zpool_create_009_neg', 'zpool_create_010_neg', 'zpool_create_011_neg',
'zpool_create_012_neg', 'zpool_create_014_neg', 'zpool_create_015_neg',
'zpool_create_017_neg', 'zpool_create_018_pos', 'zpool_create_019_pos',
'zpool_create_020_pos', 'zpool_create_021_pos', 'zpool_create_022_pos',
'zpool_create_023_neg', 'zpool_create_024_pos',
'zpool_create_encrypted', 'zpool_create_crypt_combos',
'zpool_create_draid_001_pos', 'zpool_create_draid_002_pos',
'zpool_create_draid_003_pos', 'zpool_create_draid_004_pos',
'zpool_create_features_001_pos', 'zpool_create_features_002_pos',
'zpool_create_features_003_pos', 'zpool_create_features_004_neg',
'zpool_create_features_005_pos', 'zpool_create_features_006_pos',
'zpool_create_features_007_pos', 'zpool_create_features_008_pos',
'zpool_create_features_009_pos', 'create-o_ashift',
'zpool_create_tempname', 'zpool_create_dryrun_output']
tags = ['functional', 'cli_root', 'zpool_create']
[tests/functional/cli_root/zpool_destroy]
tests = ['zpool_destroy_001_pos', 'zpool_destroy_002_pos',
'zpool_destroy_003_neg']
pre =
post =
tags = ['functional', 'cli_root', 'zpool_destroy']
[tests/functional/cli_root/zpool_detach]
tests = ['zpool_detach_001_neg']
tags = ['functional', 'cli_root', 'zpool_detach']
[tests/functional/cli_root/zpool_events]
tests = ['zpool_events_clear', 'zpool_events_cliargs', 'zpool_events_follow',
'zpool_events_poolname', 'zpool_events_errors', 'zpool_events_duplicates',
'zpool_events_clear_retained']
tags = ['functional', 'cli_root', 'zpool_events']
[tests/functional/cli_root/zpool_export]
tests = ['zpool_export_001_pos', 'zpool_export_002_pos',
'zpool_export_003_neg', 'zpool_export_004_pos']
tags = ['functional', 'cli_root', 'zpool_export']
[tests/functional/cli_root/zpool_get]
tests = ['zpool_get_001_pos', 'zpool_get_002_pos', 'zpool_get_003_pos',
- 'zpool_get_004_neg', 'zpool_get_005_pos']
+ 'zpool_get_004_neg', 'zpool_get_005_pos', 'vdev_get_001_pos']
tags = ['functional', 'cli_root', 'zpool_get']
[tests/functional/cli_root/zpool_history]
tests = ['zpool_history_001_neg', 'zpool_history_002_pos']
tags = ['functional', 'cli_root', 'zpool_history']
[tests/functional/cli_root/zpool_import]
tests = ['zpool_import_001_pos', 'zpool_import_002_pos',
'zpool_import_003_pos', 'zpool_import_004_pos', 'zpool_import_005_pos',
'zpool_import_006_pos', 'zpool_import_007_pos', 'zpool_import_008_pos',
'zpool_import_009_neg', 'zpool_import_010_pos', 'zpool_import_011_neg',
'zpool_import_012_pos', 'zpool_import_013_neg', 'zpool_import_014_pos',
'zpool_import_015_pos', 'zpool_import_016_pos', 'zpool_import_017_pos',
'zpool_import_features_001_pos', 'zpool_import_features_002_neg',
'zpool_import_features_003_pos', 'zpool_import_missing_001_pos',
'zpool_import_missing_002_pos', 'zpool_import_missing_003_pos',
'zpool_import_rename_001_pos', 'zpool_import_all_001_pos',
'zpool_import_encrypted', 'zpool_import_encrypted_load',
'zpool_import_errata3', 'zpool_import_errata4',
'import_cachefile_device_added',
'import_cachefile_device_removed',
'import_cachefile_device_replaced',
'import_cachefile_mirror_attached',
'import_cachefile_mirror_detached',
'import_cachefile_paths_changed',
'import_cachefile_shared_device',
'import_devices_missing',
'import_paths_changed',
'import_rewind_config_changed',
'import_rewind_device_replaced']
tags = ['functional', 'cli_root', 'zpool_import']
timeout = 1200
[tests/functional/cli_root/zpool_labelclear]
tests = ['zpool_labelclear_active', 'zpool_labelclear_exported',
'zpool_labelclear_removed', 'zpool_labelclear_valid']
pre =
post =
tags = ['functional', 'cli_root', 'zpool_labelclear']
[tests/functional/cli_root/zpool_initialize]
tests = ['zpool_initialize_attach_detach_add_remove',
'zpool_initialize_fault_export_import_online',
'zpool_initialize_import_export',
'zpool_initialize_offline_export_import_online',
'zpool_initialize_online_offline',
'zpool_initialize_split',
'zpool_initialize_start_and_cancel_neg',
'zpool_initialize_start_and_cancel_pos',
'zpool_initialize_suspend_resume',
'zpool_initialize_unsupported_vdevs',
'zpool_initialize_verify_checksums',
'zpool_initialize_verify_initialized']
pre =
tags = ['functional', 'cli_root', 'zpool_initialize']
[tests/functional/cli_root/zpool_offline]
tests = ['zpool_offline_001_pos', 'zpool_offline_002_neg',
'zpool_offline_003_pos']
tags = ['functional', 'cli_root', 'zpool_offline']
[tests/functional/cli_root/zpool_online]
tests = ['zpool_online_001_pos', 'zpool_online_002_neg']
tags = ['functional', 'cli_root', 'zpool_online']
[tests/functional/cli_root/zpool_remove]
tests = ['zpool_remove_001_neg', 'zpool_remove_002_pos',
'zpool_remove_003_pos']
tags = ['functional', 'cli_root', 'zpool_remove']
[tests/functional/cli_root/zpool_replace]
tests = ['zpool_replace_001_neg', 'replace-o_ashift', 'replace_prop_ashift']
tags = ['functional', 'cli_root', 'zpool_replace']
[tests/functional/cli_root/zpool_resilver]
tests = ['zpool_resilver_bad_args', 'zpool_resilver_restart']
tags = ['functional', 'cli_root', 'zpool_resilver']
[tests/functional/cli_root/zpool_scrub]
tests = ['zpool_scrub_001_neg', 'zpool_scrub_002_pos', 'zpool_scrub_003_pos',
'zpool_scrub_004_pos', 'zpool_scrub_005_pos',
'zpool_scrub_encrypted_unloaded', 'zpool_scrub_print_repairing',
'zpool_scrub_offline_device', 'zpool_scrub_multiple_copies']
tags = ['functional', 'cli_root', 'zpool_scrub']
[tests/functional/cli_root/zpool_set]
tests = ['zpool_set_001_pos', 'zpool_set_002_neg', 'zpool_set_003_neg',
- 'zpool_set_ashift', 'zpool_set_features',
+ 'zpool_set_ashift', 'zpool_set_features', 'vdev_set_001_pos',
'user_property_001_pos', 'user_property_002_neg']
tags = ['functional', 'cli_root', 'zpool_set']
[tests/functional/cli_root/zpool_split]
tests = ['zpool_split_cliargs', 'zpool_split_devices',
'zpool_split_encryption', 'zpool_split_props', 'zpool_split_vdevs',
'zpool_split_resilver', 'zpool_split_indirect',
'zpool_split_dryrun_output']
tags = ['functional', 'cli_root', 'zpool_split']
[tests/functional/cli_root/zpool_status]
tests = ['zpool_status_001_pos', 'zpool_status_002_pos',
'zpool_status_003_pos', 'zpool_status_004_pos',
'zpool_status_005_pos', 'zpool_status_006_pos',
'zpool_status_007_pos', 'zpool_status_features_001_pos']
tags = ['functional', 'cli_root', 'zpool_status']
[tests/functional/cli_root/zpool_sync]
tests = ['zpool_sync_001_pos', 'zpool_sync_002_neg']
tags = ['functional', 'cli_root', 'zpool_sync']
[tests/functional/cli_root/zpool_trim]
tests = ['zpool_trim_attach_detach_add_remove',
'zpool_trim_fault_export_import_online',
'zpool_trim_import_export', 'zpool_trim_multiple', 'zpool_trim_neg',
'zpool_trim_offline_export_import_online', 'zpool_trim_online_offline',
'zpool_trim_partial', 'zpool_trim_rate', 'zpool_trim_rate_neg',
'zpool_trim_secure', 'zpool_trim_split', 'zpool_trim_start_and_cancel_neg',
'zpool_trim_start_and_cancel_pos', 'zpool_trim_suspend_resume',
'zpool_trim_unsupported_vdevs', 'zpool_trim_verify_checksums',
'zpool_trim_verify_trimmed']
tags = ['functional', 'zpool_trim']
[tests/functional/cli_root/zpool_upgrade]
tests = ['zpool_upgrade_001_pos', 'zpool_upgrade_002_pos',
'zpool_upgrade_003_pos', 'zpool_upgrade_004_pos',
'zpool_upgrade_005_neg', 'zpool_upgrade_006_neg',
'zpool_upgrade_007_pos', 'zpool_upgrade_008_pos',
'zpool_upgrade_009_neg', 'zpool_upgrade_features_001_pos']
tags = ['functional', 'cli_root', 'zpool_upgrade']
[tests/functional/cli_root/zpool_wait]
tests = ['zpool_wait_discard', 'zpool_wait_freeing',
'zpool_wait_initialize_basic', 'zpool_wait_initialize_cancel',
'zpool_wait_initialize_flag', 'zpool_wait_multiple',
'zpool_wait_no_activity', 'zpool_wait_remove', 'zpool_wait_remove_cancel',
'zpool_wait_trim_basic', 'zpool_wait_trim_cancel', 'zpool_wait_trim_flag',
'zpool_wait_usage']
tags = ['functional', 'cli_root', 'zpool_wait']
[tests/functional/cli_root/zpool_wait/scan]
tests = ['zpool_wait_replace_cancel', 'zpool_wait_rebuild',
'zpool_wait_resilver', 'zpool_wait_scrub_cancel',
'zpool_wait_replace', 'zpool_wait_scrub_basic', 'zpool_wait_scrub_flag']
tags = ['functional', 'cli_root', 'zpool_wait']
[tests/functional/cli_user/misc]
tests = ['zdb_001_neg', 'zfs_001_neg', 'zfs_allow_001_neg',
'zfs_clone_001_neg', 'zfs_create_001_neg', 'zfs_destroy_001_neg',
'zfs_get_001_neg', 'zfs_inherit_001_neg', 'zfs_mount_001_neg',
'zfs_promote_001_neg', 'zfs_receive_001_neg', 'zfs_rename_001_neg',
'zfs_rollback_001_neg', 'zfs_send_001_neg', 'zfs_set_001_neg',
'zfs_share_001_neg', 'zfs_snapshot_001_neg', 'zfs_unallow_001_neg',
'zfs_unmount_001_neg', 'zfs_unshare_001_neg', 'zfs_upgrade_001_neg',
'zpool_001_neg', 'zpool_add_001_neg', 'zpool_attach_001_neg',
'zpool_clear_001_neg', 'zpool_create_001_neg', 'zpool_destroy_001_neg',
'zpool_detach_001_neg', 'zpool_export_001_neg', 'zpool_get_001_neg',
'zpool_history_001_neg', 'zpool_import_001_neg', 'zpool_import_002_neg',
'zpool_offline_001_neg', 'zpool_online_001_neg', 'zpool_remove_001_neg',
'zpool_replace_001_neg', 'zpool_scrub_001_neg', 'zpool_set_001_neg',
'zpool_status_001_neg', 'zpool_upgrade_001_neg', 'arcstat_001_pos',
'arc_summary_001_pos', 'arc_summary_002_neg', 'zpool_wait_privilege',
'zilstat_001_pos']
user =
tags = ['functional', 'cli_user', 'misc']
[tests/functional/cli_user/zfs_list]
tests = ['zfs_list_001_pos', 'zfs_list_002_pos', 'zfs_list_003_pos',
'zfs_list_004_neg', 'zfs_list_005_neg', 'zfs_list_007_pos',
'zfs_list_008_neg']
user =
tags = ['functional', 'cli_user', 'zfs_list']
[tests/functional/cli_user/zpool_iostat]
tests = ['zpool_iostat_001_neg', 'zpool_iostat_002_pos',
'zpool_iostat_003_neg', 'zpool_iostat_004_pos',
'zpool_iostat_005_pos', 'zpool_iostat_-c_disable',
'zpool_iostat_-c_homedir', 'zpool_iostat_-c_searchpath']
user =
tags = ['functional', 'cli_user', 'zpool_iostat']
[tests/functional/cli_user/zpool_list]
tests = ['zpool_list_001_pos', 'zpool_list_002_neg']
user =
tags = ['functional', 'cli_user', 'zpool_list']
[tests/functional/cli_user/zpool_status]
tests = ['zpool_status_003_pos', 'zpool_status_-c_disable',
'zpool_status_-c_homedir', 'zpool_status_-c_searchpath']
user =
tags = ['functional', 'cli_user', 'zpool_status']
[tests/functional/compression]
tests = ['compress_001_pos', 'compress_002_pos', 'compress_003_pos',
'l2arc_compressed_arc', 'l2arc_compressed_arc_disabled',
'l2arc_encrypted', 'l2arc_encrypted_no_compressed_arc']
tags = ['functional', 'compression']
[tests/functional/cp_files]
tests = ['cp_files_001_pos']
tags = ['functional', 'cp_files']
[tests/functional/crtime]
tests = ['crtime_001_pos' ]
tags = ['functional', 'crtime']
[tests/functional/ctime]
tests = ['ctime_001_pos' ]
tags = ['functional', 'ctime']
[tests/functional/deadman]
tests = ['deadman_ratelimit', 'deadman_sync', 'deadman_zio']
pre =
post =
tags = ['functional', 'deadman']
[tests/functional/delegate]
tests = ['zfs_allow_001_pos', 'zfs_allow_002_pos', 'zfs_allow_003_pos',
'zfs_allow_004_pos', 'zfs_allow_005_pos', 'zfs_allow_006_pos',
'zfs_allow_007_pos', 'zfs_allow_008_pos', 'zfs_allow_009_neg',
'zfs_allow_010_pos', 'zfs_allow_011_neg', 'zfs_allow_012_neg',
'zfs_unallow_001_pos', 'zfs_unallow_002_pos', 'zfs_unallow_003_pos',
'zfs_unallow_004_pos', 'zfs_unallow_005_pos', 'zfs_unallow_006_pos',
'zfs_unallow_007_neg', 'zfs_unallow_008_neg']
tags = ['functional', 'delegate']
[tests/functional/exec]
tests = ['exec_001_pos', 'exec_002_neg']
tags = ['functional', 'exec']
[tests/functional/fallocate]
tests = ['fallocate_punch-hole']
tags = ['functional', 'fallocate']
[tests/functional/features/async_destroy]
tests = ['async_destroy_001_pos']
tags = ['functional', 'features', 'async_destroy']
[tests/functional/features/large_dnode]
tests = ['large_dnode_001_pos', 'large_dnode_003_pos', 'large_dnode_004_neg',
'large_dnode_005_pos', 'large_dnode_007_neg', 'large_dnode_009_pos']
tags = ['functional', 'features', 'large_dnode']
[tests/functional/grow]
pre =
post =
tests = ['grow_pool_001_pos', 'grow_replicas_001_pos']
tags = ['functional', 'grow']
[tests/functional/history]
tests = ['history_001_pos', 'history_002_pos', 'history_003_pos',
'history_004_pos', 'history_005_neg', 'history_006_neg',
'history_007_pos', 'history_008_pos', 'history_009_pos',
'history_010_pos']
tags = ['functional', 'history']
[tests/functional/hkdf]
pre =
post =
tests = ['hkdf_test']
tags = ['functional', 'hkdf']
[tests/functional/inheritance]
tests = ['inherit_001_pos']
pre =
tags = ['functional', 'inheritance']
[tests/functional/io]
tests = ['sync', 'psync', 'posixaio', 'mmap']
tags = ['functional', 'io']
[tests/functional/inuse]
tests = ['inuse_004_pos', 'inuse_005_pos', 'inuse_008_pos', 'inuse_009_pos']
post =
tags = ['functional', 'inuse']
[tests/functional/large_files]
tests = ['large_files_001_pos', 'large_files_002_pos']
tags = ['functional', 'large_files']
[tests/functional/limits]
tests = ['filesystem_count', 'filesystem_limit', 'snapshot_count',
'snapshot_limit']
tags = ['functional', 'limits']
[tests/functional/link_count]
tests = ['link_count_001', 'link_count_root_inode']
tags = ['functional', 'link_count']
[tests/functional/migration]
tests = ['migration_001_pos', 'migration_002_pos', 'migration_003_pos',
'migration_004_pos', 'migration_005_pos', 'migration_006_pos',
'migration_007_pos', 'migration_008_pos', 'migration_009_pos',
'migration_010_pos', 'migration_011_pos', 'migration_012_pos']
tags = ['functional', 'migration']
[tests/functional/mmap]
tests = ['mmap_mixed', 'mmap_read_001_pos', 'mmap_seek_001_pos',
'mmap_sync_001_pos', 'mmap_write_001_pos']
tags = ['functional', 'mmap']
[tests/functional/mount]
tests = ['umount_001', 'umountall_001']
tags = ['functional', 'mount']
[tests/functional/mv_files]
tests = ['mv_files_001_pos', 'mv_files_002_pos', 'random_creation']
tags = ['functional', 'mv_files']
[tests/functional/nestedfs]
tests = ['nestedfs_001_pos']
tags = ['functional', 'nestedfs']
[tests/functional/no_space]
tests = ['enospc_001_pos', 'enospc_002_pos', 'enospc_003_pos',
'enospc_df', 'enospc_ganging', 'enospc_rm']
tags = ['functional', 'no_space']
[tests/functional/nopwrite]
tests = ['nopwrite_copies', 'nopwrite_mtime', 'nopwrite_negative',
'nopwrite_promoted_clone', 'nopwrite_recsize', 'nopwrite_sync',
'nopwrite_varying_compression', 'nopwrite_volume']
tags = ['functional', 'nopwrite']
[tests/functional/online_offline]
tests = ['online_offline_001_pos', 'online_offline_002_neg',
'online_offline_003_neg']
tags = ['functional', 'online_offline']
[tests/functional/pool_checkpoint]
tests = ['checkpoint_after_rewind', 'checkpoint_big_rewind',
'checkpoint_capacity', 'checkpoint_conf_change', 'checkpoint_discard',
'checkpoint_discard_busy', 'checkpoint_discard_many',
'checkpoint_indirect', 'checkpoint_invalid', 'checkpoint_lun_expsz',
'checkpoint_open', 'checkpoint_removal', 'checkpoint_rewind',
'checkpoint_ro_rewind', 'checkpoint_sm_scale', 'checkpoint_twice',
'checkpoint_vdev_add', 'checkpoint_zdb', 'checkpoint_zhack_feat']
tags = ['functional', 'pool_checkpoint']
timeout = 1800
[tests/functional/pool_names]
tests = ['pool_names_001_pos', 'pool_names_002_neg']
pre =
post =
tags = ['functional', 'pool_names']
[tests/functional/poolversion]
tests = ['poolversion_001_pos', 'poolversion_002_pos']
tags = ['functional', 'poolversion']
[tests/functional/pyzfs]
tests = ['pyzfs_unittest']
pre =
post =
tags = ['functional', 'pyzfs']
[tests/functional/quota]
tests = ['quota_001_pos', 'quota_002_pos', 'quota_003_pos',
'quota_004_pos', 'quota_005_pos', 'quota_006_neg']
tags = ['functional', 'quota']
[tests/functional/redacted_send]
tests = ['redacted_compressed', 'redacted_contents', 'redacted_deleted',
'redacted_disabled_feature', 'redacted_embedded', 'redacted_holes',
'redacted_incrementals', 'redacted_largeblocks', 'redacted_many_clones',
'redacted_mixed_recsize', 'redacted_mounts', 'redacted_negative',
'redacted_origin', 'redacted_panic', 'redacted_props', 'redacted_resume',
'redacted_size', 'redacted_volume']
tags = ['functional', 'redacted_send']
[tests/functional/raidz]
tests = ['raidz_001_neg', 'raidz_002_pos', 'raidz_003_pos', 'raidz_004_pos']
tags = ['functional', 'raidz']
[tests/functional/redundancy]
tests = ['redundancy_draid', 'redundancy_draid1', 'redundancy_draid2',
'redundancy_draid3', 'redundancy_draid_damaged1',
'redundancy_draid_damaged2', 'redundancy_draid_spare1',
'redundancy_draid_spare2', 'redundancy_draid_spare3', 'redundancy_mirror',
'redundancy_raidz', 'redundancy_raidz1', 'redundancy_raidz2',
'redundancy_raidz3', 'redundancy_stripe']
tags = ['functional', 'redundancy']
timeout = 1200
[tests/functional/refquota]
tests = ['refquota_001_pos', 'refquota_002_pos', 'refquota_003_pos',
'refquota_004_pos', 'refquota_005_pos', 'refquota_006_neg',
'refquota_007_neg', 'refquota_008_neg']
tags = ['functional', 'refquota']
[tests/functional/refreserv]
tests = ['refreserv_001_pos', 'refreserv_002_pos', 'refreserv_003_pos',
'refreserv_004_pos', 'refreserv_005_pos', 'refreserv_multi_raidz',
'refreserv_raidz']
tags = ['functional', 'refreserv']
[tests/functional/removal]
pre =
tests = ['removal_all_vdev', 'removal_cancel', 'removal_check_space',
'removal_condense_export', 'removal_multiple_indirection',
'removal_nopwrite', 'removal_remap_deadlists',
'removal_resume_export', 'removal_sanity', 'removal_with_add',
'removal_with_create_fs', 'removal_with_dedup',
- 'removal_with_errors', 'removal_with_export',
+ 'removal_with_errors', 'removal_with_export', 'removal_with_indirect',
'removal_with_ganging', 'removal_with_faulted',
'removal_with_remove', 'removal_with_scrub', 'removal_with_send',
'removal_with_send_recv', 'removal_with_snapshot',
'removal_with_write', 'removal_with_zdb', 'remove_expanded',
'remove_mirror', 'remove_mirror_sanity', 'remove_raidz',
- 'remove_indirect', 'remove_attach_mirror']
+ 'remove_indirect', 'remove_attach_mirror', 'removal_reservation']
tags = ['functional', 'removal']
[tests/functional/rename_dirs]
tests = ['rename_dirs_001_pos']
tags = ['functional', 'rename_dirs']
[tests/functional/replacement]
tests = ['attach_import', 'attach_multiple', 'attach_rebuild',
'attach_resilver', 'detach', 'rebuild_disabled_feature',
'rebuild_multiple', 'rebuild_raidz', 'replace_import', 'replace_rebuild',
'replace_resilver', 'resilver_restart_001', 'resilver_restart_002',
'scrub_cancel']
tags = ['functional', 'replacement']
[tests/functional/reservation]
tests = ['reservation_001_pos', 'reservation_002_pos', 'reservation_003_pos',
'reservation_004_pos', 'reservation_005_pos', 'reservation_006_pos',
'reservation_007_pos', 'reservation_008_pos', 'reservation_009_pos',
'reservation_010_pos', 'reservation_011_pos', 'reservation_012_pos',
'reservation_013_pos', 'reservation_014_pos', 'reservation_015_pos',
'reservation_016_pos', 'reservation_017_pos', 'reservation_018_pos',
'reservation_019_pos', 'reservation_020_pos', 'reservation_021_neg',
'reservation_022_pos']
tags = ['functional', 'reservation']
[tests/functional/rootpool]
tests = ['rootpool_002_neg', 'rootpool_003_neg', 'rootpool_007_pos']
tags = ['functional', 'rootpool']
[tests/functional/rsend]
tests = ['recv_dedup', 'recv_dedup_encrypted_zvol', 'rsend_001_pos',
'rsend_002_pos', 'rsend_003_pos', 'rsend_004_pos', 'rsend_005_pos',
'rsend_006_pos', 'rsend_007_pos', 'rsend_008_pos', 'rsend_009_pos',
'rsend_010_pos', 'rsend_011_pos', 'rsend_012_pos', 'rsend_013_pos',
'rsend_014_pos', 'rsend_016_neg', 'rsend_019_pos', 'rsend_020_pos',
'rsend_021_pos', 'rsend_022_pos', 'rsend_024_pos', 'rsend_025_pos',
'rsend_026_neg', 'rsend_027_pos', 'rsend_028_neg', 'rsend_029_neg',
'rsend_030_pos', 'rsend_031_pos', 'send-c_verify_ratio',
'send-c_verify_contents', 'send-c_props', 'send-c_incremental',
'send-c_volume', 'send-c_zstream_recompress', 'send-c_zstreamdump',
'send-c_lz4_disabled', 'send-c_recv_lz4_disabled',
'send-c_mixed_compression', 'send-c_stream_size_estimate',
'send-c_embedded_blocks', 'send-c_resume', 'send-cpL_varied_recsize',
'send-c_recv_dedup', 'send-L_toggle', 'send_encrypted_incremental',
'send_encrypted_freeobjects', 'send_encrypted_hierarchy',
'send_encrypted_props', 'send_encrypted_truncated_files',
'send_freeobjects', 'send_realloc_files', 'send_realloc_encrypted_files',
'send_spill_block', 'send_holds', 'send_hole_birth', 'send_mixed_raw',
'send-wR_encrypted_zvol', 'send_partial_dataset', 'send_invalid',
'send_doall', 'send_raw_spill_block', 'send_raw_ashift',
'send_raw_large_blocks']
tags = ['functional', 'rsend']
[tests/functional/scrub_mirror]
tests = ['scrub_mirror_001_pos', 'scrub_mirror_002_pos',
'scrub_mirror_003_pos', 'scrub_mirror_004_pos']
tags = ['functional', 'scrub_mirror']
[tests/functional/slog]
tests = ['slog_001_pos', 'slog_002_pos', 'slog_003_pos', 'slog_004_pos',
'slog_005_pos', 'slog_006_pos', 'slog_007_pos', 'slog_008_neg',
'slog_009_neg', 'slog_010_neg', 'slog_011_neg', 'slog_012_neg',
'slog_013_pos', 'slog_014_pos', 'slog_015_neg', 'slog_replay_fs_001',
'slog_replay_fs_002', 'slog_replay_volume', 'slog_016_pos']
tags = ['functional', 'slog']
[tests/functional/snapshot]
tests = ['clone_001_pos', 'rollback_001_pos', 'rollback_002_pos',
'rollback_003_pos', 'snapshot_001_pos', 'snapshot_002_pos',
'snapshot_003_pos', 'snapshot_004_pos', 'snapshot_005_pos',
'snapshot_006_pos', 'snapshot_007_pos', 'snapshot_008_pos',
'snapshot_009_pos', 'snapshot_010_pos', 'snapshot_011_pos',
'snapshot_012_pos', 'snapshot_013_pos', 'snapshot_014_pos',
'snapshot_017_pos', 'snapshot_018_pos']
tags = ['functional', 'snapshot']
[tests/functional/snapused]
tests = ['snapused_001_pos', 'snapused_002_pos', 'snapused_003_pos',
'snapused_004_pos', 'snapused_005_pos']
tags = ['functional', 'snapused']
[tests/functional/sparse]
tests = ['sparse_001_pos']
tags = ['functional', 'sparse']
[tests/functional/stat]
tests = ['stat_001_pos']
tags = ['functional', 'stat']
[tests/functional/suid]
tests = ['suid_write_to_suid', 'suid_write_to_sgid', 'suid_write_to_suid_sgid',
'suid_write_to_none', 'suid_write_zil_replay']
tags = ['functional', 'suid']
[tests/functional/trim]
tests = ['autotrim_integrity', 'autotrim_config', 'autotrim_trim_integrity',
'trim_integrity', 'trim_config', 'trim_l2arc']
tags = ['functional', 'trim']
[tests/functional/truncate]
tests = ['truncate_001_pos', 'truncate_002_pos', 'truncate_timestamps']
tags = ['functional', 'truncate']
[tests/functional/upgrade]
tests = ['upgrade_userobj_001_pos', 'upgrade_readonly_pool']
tags = ['functional', 'upgrade']
[tests/functional/userquota]
tests = [
'userquota_001_pos', 'userquota_002_pos', 'userquota_003_pos',
'userquota_004_pos', 'userquota_005_neg', 'userquota_006_pos',
'userquota_007_pos', 'userquota_008_pos', 'userquota_009_pos',
'userquota_010_pos', 'userquota_011_pos', 'userquota_012_neg',
'userspace_001_pos', 'userspace_002_pos', 'userspace_encrypted',
'userspace_send_encrypted', 'userspace_encrypted_13709']
tags = ['functional', 'userquota']
[tests/functional/vdev_zaps]
tests = ['vdev_zaps_001_pos', 'vdev_zaps_002_pos', 'vdev_zaps_003_pos',
'vdev_zaps_004_pos', 'vdev_zaps_005_pos', 'vdev_zaps_006_pos',
'vdev_zaps_007_pos']
tags = ['functional', 'vdev_zaps']
[tests/functional/write_dirs]
tests = ['write_dirs_001_pos', 'write_dirs_002_pos']
tags = ['functional', 'write_dirs']
[tests/functional/xattr]
tests = ['xattr_001_pos', 'xattr_002_neg', 'xattr_003_neg', 'xattr_004_pos',
'xattr_005_pos', 'xattr_006_pos', 'xattr_007_neg',
'xattr_011_pos', 'xattr_012_pos', 'xattr_013_pos', 'xattr_compat']
tags = ['functional', 'xattr']
[tests/functional/zvol/zvol_ENOSPC]
tests = ['zvol_ENOSPC_001_pos']
tags = ['functional', 'zvol', 'zvol_ENOSPC']
[tests/functional/zvol/zvol_cli]
tests = ['zvol_cli_001_pos', 'zvol_cli_002_pos', 'zvol_cli_003_neg']
tags = ['functional', 'zvol', 'zvol_cli']
[tests/functional/zvol/zvol_misc]
tests = ['zvol_misc_002_pos', 'zvol_misc_hierarchy', 'zvol_misc_rename_inuse',
'zvol_misc_snapdev', 'zvol_misc_trim', 'zvol_misc_volmode', 'zvol_misc_zil']
tags = ['functional', 'zvol', 'zvol_misc']
[tests/functional/zvol/zvol_stress]
tests = ['zvol_stress']
tags = ['functional', 'zvol', 'zvol_stress']
[tests/functional/zvol/zvol_swap]
tests = ['zvol_swap_001_pos', 'zvol_swap_002_pos', 'zvol_swap_004_pos']
tags = ['functional', 'zvol', 'zvol_swap']
[tests/functional/libzfs]
tests = ['many_fds', 'libzfs_input']
tags = ['functional', 'libzfs']
[tests/functional/log_spacemap]
tests = ['log_spacemap_import_logs']
pre =
post =
tags = ['functional', 'log_spacemap']
[tests/functional/l2arc]
tests = ['l2arc_arcstats_pos', 'l2arc_mfuonly_pos', 'l2arc_l2miss_pos',
'persist_l2arc_001_pos', 'persist_l2arc_002_pos',
'persist_l2arc_003_neg', 'persist_l2arc_004_pos', 'persist_l2arc_005_pos']
tags = ['functional', 'l2arc']
[tests/functional/zpool_influxdb]
tests = ['zpool_influxdb']
tags = ['functional', 'zpool_influxdb']
diff --git a/sys/contrib/openzfs/tests/runfiles/linux.run b/sys/contrib/openzfs/tests/runfiles/linux.run
index 15755408b5ad..4df770d61f07 100644
--- a/sys/contrib/openzfs/tests/runfiles/linux.run
+++ b/sys/contrib/openzfs/tests/runfiles/linux.run
@@ -1,205 +1,209 @@
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
[DEFAULT]
pre = setup
quiet = False
pre_user = root
user = root
timeout = 600
post_user = root
post = cleanup
failsafe_user = root
failsafe = callbacks/zfs_failsafe
outputdir = /var/tmp/test_results
tags = ['functional']
[tests/functional/acl/posix:Linux]
tests = ['posix_001_pos', 'posix_002_pos', 'posix_003_pos', 'posix_004_pos']
tags = ['functional', 'acl', 'posix']
[tests/functional/acl/posix-sa:Linux]
tests = ['posix_001_pos', 'posix_002_pos', 'posix_003_pos', 'posix_004_pos']
tags = ['functional', 'acl', 'posix-sa']
[tests/functional/atime:Linux]
tests = ['atime_003_pos', 'root_relatime_on']
tags = ['functional', 'atime']
[tests/functional/chattr:Linux]
tests = ['chattr_001_pos', 'chattr_002_neg']
tags = ['functional', 'chattr']
[tests/functional/cli_root/zfs:Linux]
tests = ['zfs_003_neg']
tags = ['functional', 'cli_root', 'zfs']
[tests/functional/cli_root/zfs_mount:Linux]
tests = ['zfs_mount_006_pos', 'zfs_mount_008_pos', 'zfs_mount_013_pos',
'zfs_mount_014_neg', 'zfs_multi_mount']
tags = ['functional', 'cli_root', 'zfs_mount']
[tests/functional/cli_root/zfs_share:Linux]
tests = ['zfs_share_005_pos', 'zfs_share_007_neg', 'zfs_share_009_neg',
'zfs_share_012_pos', 'zfs_share_013_pos']
tags = ['functional', 'cli_root', 'zfs_share']
+[tests/functional/cli_root/zfs_unshare:Linux]
+tests = ['zfs_unshare_008_pos']
+tags = ['functional', 'cli_root', 'zfs_unshare']
+
[tests/functional/cli_root/zfs_sysfs:Linux]
tests = ['zfeature_set_unsupported', 'zfs_get_unsupported',
'zfs_set_unsupported', 'zfs_sysfs_live', 'zpool_get_unsupported',
'zpool_set_unsupported']
tags = ['functional', 'cli_root', 'zfs_sysfs']
[tests/functional/cli_root/zpool_add:Linux]
tests = ['add_nested_replacing_spare']
tags = ['functional', 'cli_root', 'zpool_add']
[tests/functional/cli_root/zpool_expand:Linux]
tests = ['zpool_expand_001_pos', 'zpool_expand_002_pos',
'zpool_expand_003_neg', 'zpool_expand_004_pos', 'zpool_expand_005_pos']
tags = ['functional', 'cli_root', 'zpool_expand']
[tests/functional/cli_root/zpool_reopen:Linux]
tests = ['zpool_reopen_001_pos', 'zpool_reopen_002_pos',
'zpool_reopen_003_pos', 'zpool_reopen_004_pos', 'zpool_reopen_005_pos',
'zpool_reopen_006_neg', 'zpool_reopen_007_pos']
tags = ['functional', 'cli_root', 'zpool_reopen']
[tests/functional/cli_root/zpool_split:Linux]
tests = ['zpool_split_wholedisk']
tags = ['functional', 'cli_root', 'zpool_split']
[tests/functional/compression:Linux]
tests = ['compress_004_pos']
tags = ['functional', 'compression']
[tests/functional/devices:Linux]
tests = ['devices_001_pos', 'devices_002_neg', 'devices_003_pos']
tags = ['functional', 'devices']
[tests/functional/events:Linux]
tests = ['events_001_pos', 'events_002_pos', 'zed_rc_filter', 'zed_fd_spill',
'zed_cksum_reported', 'zed_cksum_config', 'zed_io_config']
tags = ['functional', 'events']
[tests/functional/fadvise:Linux]
tests = ['fadvise_sequential']
tags = ['functional', 'fadvise']
[tests/functional/fallocate:Linux]
tests = ['fallocate_prealloc', 'fallocate_zero-range']
tags = ['functional', 'fallocate']
[tests/functional/fault:Linux]
tests = ['auto_offline_001_pos', 'auto_online_001_pos', 'auto_online_002_pos',
'auto_replace_001_pos', 'auto_spare_001_pos', 'auto_spare_002_pos',
'auto_spare_multiple', 'auto_spare_ashift', 'auto_spare_shared',
'decrypt_fault', 'decompress_fault', 'scrub_after_resilver',
'zpool_status_-s']
tags = ['functional', 'fault']
[tests/functional/features/large_dnode:Linux]
tests = ['large_dnode_002_pos', 'large_dnode_006_pos', 'large_dnode_008_pos']
tags = ['functional', 'features', 'large_dnode']
[tests/functional/io:Linux]
tests = ['libaio', 'io_uring']
tags = ['functional', 'io']
[tests/functional/largest_pool:Linux]
tests = ['largest_pool_001_pos']
pre =
post =
tags = ['functional', 'largest_pool']
[tests/functional/mmap:Linux]
-tests = ['mmap_libaio_001_pos']
+tests = ['mmap_libaio_001_pos', 'mmap_sync_001_pos']
tags = ['functional', 'mmap']
[tests/functional/mmp:Linux]
tests = ['mmp_on_thread', 'mmp_on_uberblocks', 'mmp_on_off', 'mmp_interval',
'mmp_active_import', 'mmp_inactive_import', 'mmp_exported_import',
'mmp_write_uberblocks', 'mmp_reset_interval', 'multihost_history',
'mmp_on_zdb', 'mmp_write_distribution', 'mmp_hostid']
tags = ['functional', 'mmp']
[tests/functional/mount:Linux]
tests = ['umount_unlinked_drain']
tags = ['functional', 'mount']
[tests/functional/pam:Linux]
tests = ['pam_basic', 'pam_nounmount', 'pam_short_password']
tags = ['functional', 'pam']
[tests/functional/procfs:Linux]
tests = ['procfs_list_basic', 'procfs_list_concurrent_readers',
'procfs_list_stale_read', 'pool_state']
tags = ['functional', 'procfs']
[tests/functional/projectquota:Linux]
tests = ['projectid_001_pos', 'projectid_002_pos', 'projectid_003_pos',
'projectquota_001_pos', 'projectquota_002_pos', 'projectquota_003_pos',
'projectquota_004_neg', 'projectquota_005_pos', 'projectquota_006_pos',
'projectquota_007_pos', 'projectquota_008_pos', 'projectquota_009_pos',
'projectspace_001_pos', 'projectspace_002_pos', 'projectspace_003_pos',
'projectspace_004_pos',
'projecttree_001_pos', 'projecttree_002_pos', 'projecttree_003_neg']
tags = ['functional', 'projectquota']
[tests/functional/dos_attributes:Linux]
tests = ['read_dos_attrs_001', 'write_dos_attrs_001']
tags = ['functional', 'dos_attributes']
[tests/functional/renameat2:Linux]
tests = ['renameat2_noreplace', 'renameat2_exchange', 'renameat2_whiteout']
tags = ['functional', 'renameat2']
[tests/functional/rsend:Linux]
tests = ['send_realloc_dnode_size', 'send_encrypted_files']
tags = ['functional', 'rsend']
[tests/functional/simd:Linux]
pre =
post =
tests = ['simd_supported']
tags = ['functional', 'simd']
[tests/functional/snapshot:Linux]
tests = ['snapshot_015_pos', 'snapshot_016_pos']
tags = ['functional', 'snapshot']
[tests/functional/tmpfile:Linux]
tests = ['tmpfile_001_pos', 'tmpfile_002_pos', 'tmpfile_003_pos',
'tmpfile_stat_mode']
tags = ['functional', 'tmpfile']
[tests/functional/upgrade:Linux]
tests = ['upgrade_projectquota_001_pos']
tags = ['functional', 'upgrade']
[tests/functional/user_namespace:Linux]
tests = ['user_namespace_001', 'user_namespace_002', 'user_namespace_003',
'user_namespace_004']
tags = ['functional', 'user_namespace']
[tests/functional/userquota:Linux]
tests = ['groupspace_001_pos', 'groupspace_002_pos', 'groupspace_003_pos',
'userquota_013_pos', 'userspace_003_pos']
tags = ['functional', 'userquota']
[tests/functional/zvol/zvol_misc:Linux]
tests = ['zvol_misc_fua']
tags = ['functional', 'zvol', 'zvol_misc']
[tests/functional/idmap_mount:Linux]
tests = ['idmap_mount_001', 'idmap_mount_002', 'idmap_mount_003',
'idmap_mount_004', 'idmap_mount_005']
tags = ['functional', 'idmap_mount']
diff --git a/sys/contrib/openzfs/tests/test-runner/bin/test-runner.py.in b/sys/contrib/openzfs/tests/test-runner/bin/test-runner.py.in
index 28276ebc47e3..c454bf8d7c6a 100755
--- a/sys/contrib/openzfs/tests/test-runner/bin/test-runner.py.in
+++ b/sys/contrib/openzfs/tests/test-runner/bin/test-runner.py.in
@@ -1,1150 +1,1150 @@
#!/usr/bin/env @PYTHON_SHEBANG@
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2012, 2018 by Delphix. All rights reserved.
# Copyright (c) 2019 Datto Inc.
#
# This script must remain compatible with Python 3.6+.
#
import os
import sys
import ctypes
import re
import configparser
from datetime import datetime
from optparse import OptionParser
from pwd import getpwnam
from pwd import getpwuid
from select import select
from subprocess import PIPE
from subprocess import Popen
from subprocess import check_output
from threading import Timer
from time import time, CLOCK_MONOTONIC
from os.path import exists
BASEDIR = '/var/tmp/test_results'
TESTDIR = '/usr/share/zfs/'
KMEMLEAK_FILE = '/sys/kernel/debug/kmemleak'
KILL = 'kill'
TRUE = 'true'
SUDO = 'sudo'
LOG_FILE = 'LOG_FILE'
LOG_OUT = 'LOG_OUT'
LOG_ERR = 'LOG_ERR'
LOG_FILE_OBJ = None
+try:
+ from time import monotonic as monotonic_time
+except ImportError:
+ class timespec(ctypes.Structure):
+ _fields_ = [
+ ('tv_sec', ctypes.c_long),
+ ('tv_nsec', ctypes.c_long)
+ ]
-class timespec(ctypes.Structure):
- _fields_ = [
- ('tv_sec', ctypes.c_long),
- ('tv_nsec', ctypes.c_long)
- ]
-
-
-librt = ctypes.CDLL('librt.so.1', use_errno=True)
-clock_gettime = librt.clock_gettime
-clock_gettime.argtypes = [ctypes.c_int, ctypes.POINTER(timespec)]
-
+ librt = ctypes.CDLL('librt.so.1', use_errno=True)
+ clock_gettime = librt.clock_gettime
+ clock_gettime.argtypes = [ctypes.c_int, ctypes.POINTER(timespec)]
-def monotonic_time():
- t = timespec()
- if clock_gettime(CLOCK_MONOTONIC, ctypes.pointer(t)) != 0:
- errno_ = ctypes.get_errno()
- raise OSError(errno_, os.strerror(errno_))
- return t.tv_sec + t.tv_nsec * 1e-9
+ def monotonic_time():
+ t = timespec()
+ if clock_gettime(CLOCK_MONOTONIC, ctypes.pointer(t)) != 0:
+ errno_ = ctypes.get_errno()
+ raise OSError(errno_, os.strerror(errno_))
+ return t.tv_sec + t.tv_nsec * 1e-9
class Result(object):
total = 0
runresults = {'PASS': 0, 'FAIL': 0, 'SKIP': 0, 'KILLED': 0, 'RERAN': 0}
def __init__(self):
self.starttime = None
self.returncode = None
self.runtime = ''
self.stdout = []
self.stderr = []
self.kmemleak = ''
self.result = ''
def done(self, proc, killed, reran):
"""
Finalize the results of this Cmd.
"""
Result.total += 1
m, s = divmod(monotonic_time() - self.starttime, 60)
self.runtime = '%02d:%02d' % (m, s)
self.returncode = proc.returncode
if reran is True:
Result.runresults['RERAN'] += 1
if killed:
self.result = 'KILLED'
Result.runresults['KILLED'] += 1
elif len(self.kmemleak) > 0:
self.result = 'FAIL'
Result.runresults['FAIL'] += 1
elif self.returncode == 0:
self.result = 'PASS'
Result.runresults['PASS'] += 1
elif self.returncode == 4:
self.result = 'SKIP'
Result.runresults['SKIP'] += 1
elif self.returncode != 0:
self.result = 'FAIL'
Result.runresults['FAIL'] += 1
class Output(object):
"""
This class is a slightly modified version of the 'Stream' class found
here: http://goo.gl/aSGfv
"""
def __init__(self, stream):
self.stream = stream
self._buf = b''
self.lines = []
def fileno(self):
return self.stream.fileno()
def read(self, drain=0):
"""
Read from the file descriptor. If 'drain' set, read until EOF.
"""
while self._read() is not None:
if not drain:
break
def _read(self):
"""
Read up to 4k of data from this output stream. Collect the output
up to the last newline, and append it to any leftover data from a
previous call. The lines are stored as a (timestamp, data) tuple
for easy sorting/merging later.
"""
fd = self.fileno()
buf = os.read(fd, 4096)
if not buf:
return None
if b'\n' not in buf:
self._buf += buf
return []
buf = self._buf + buf
tmp, rest = buf.rsplit(b'\n', 1)
self._buf = rest
now = datetime.now()
rows = tmp.split(b'\n')
self.lines += [(now, r) for r in rows]
class Cmd(object):
verified_users = []
def __init__(self, pathname, identifier=None, outputdir=None,
timeout=None, user=None, tags=None):
self.pathname = pathname
self.identifier = identifier
self.outputdir = outputdir or 'BASEDIR'
"""
The timeout for tests is measured in wall-clock time
"""
self.timeout = timeout
self.user = user or ''
self.killed = False
self.reran = None
self.result = Result()
if self.timeout is None:
self.timeout = 60
def __str__(self):
return '''\
Pathname: %s
Identifier: %s
Outputdir: %s
Timeout: %d
User: %s
''' % (self.pathname, self.identifier, self.outputdir, self.timeout, self.user)
def kill_cmd(self, proc, keyboard_interrupt=False):
"""
Kill a running command due to timeout, or ^C from the keyboard. If
sudo is required, this user was verified previously.
"""
self.killed = True
do_sudo = len(self.user) != 0
signal = '-TERM'
cmd = [SUDO, KILL, signal, str(proc.pid)]
if not do_sudo:
del cmd[0]
try:
kp = Popen(cmd)
kp.wait()
except Exception:
pass
"""
If this is not a user-initiated kill and the test has not been
reran before we consider if the test needs to be reran:
If the test has spent some time hibernating and didn't run the whole
length of time before being timed out we will rerun the test.
"""
if keyboard_interrupt is False and self.reran is None:
runtime = monotonic_time() - self.result.starttime
if int(self.timeout) > runtime:
self.killed = False
self.reran = False
self.run(False)
self.reran = True
def update_cmd_privs(self, cmd, user):
"""
If a user has been specified to run this Cmd and we're not already
running as that user, prepend the appropriate sudo command to run
as that user.
"""
me = getpwuid(os.getuid())
if not user or user is me:
if os.path.isfile(cmd+'.ksh') and os.access(cmd+'.ksh', os.X_OK):
cmd += '.ksh'
if os.path.isfile(cmd+'.sh') and os.access(cmd+'.sh', os.X_OK):
cmd += '.sh'
return cmd
if not os.path.isfile(cmd):
if os.path.isfile(cmd+'.ksh') and os.access(cmd+'.ksh', os.X_OK):
cmd += '.ksh'
if os.path.isfile(cmd+'.sh') and os.access(cmd+'.sh', os.X_OK):
cmd += '.sh'
ret = '%s -E -u %s %s' % (SUDO, user, cmd)
return ret.split(' ')
def collect_output(self, proc):
"""
Read from stdout/stderr as data becomes available, until the
process is no longer running. Return the lines from the stdout and
stderr Output objects.
"""
out = Output(proc.stdout)
err = Output(proc.stderr)
res = []
while proc.returncode is None:
proc.poll()
res = select([out, err], [], [], .1)
for fd in res[0]:
fd.read()
for fd in res[0]:
fd.read(drain=1)
return out.lines, err.lines
def run(self, dryrun, kmemleak, kmsg):
"""
This is the main function that runs each individual test.
Determine whether or not the command requires sudo, and modify it
if needed. Run the command, and update the result object.
"""
if dryrun is True:
print(self)
return
privcmd = self.update_cmd_privs(self.pathname, self.user)
try:
old = os.umask(0)
if not os.path.isdir(self.outputdir):
os.makedirs(self.outputdir, mode=0o777)
os.umask(old)
except OSError as e:
fail('%s' % e)
"""
Log each test we run to /dev/kmsg (on Linux), so if there's a kernel
warning we'll be able to match it up to a particular test.
"""
if kmsg is True and exists("/dev/kmsg"):
try:
kp = Popen([SUDO, "sh", "-c",
f"echo ZTS run {self.pathname} > /dev/kmsg"])
kp.wait()
except Exception:
pass
self.result.starttime = monotonic_time()
if kmemleak:
cmd = f'{SUDO} sh -c "echo clear > {KMEMLEAK_FILE}"'
check_output(cmd, shell=True)
proc = Popen(privcmd, stdout=PIPE, stderr=PIPE)
# Allow a special timeout value of 0 to mean infinity
if int(self.timeout) == 0:
self.timeout = sys.maxsize / (10 ** 9)
t = Timer(int(self.timeout), self.kill_cmd, [proc])
try:
t.start()
self.result.stdout, self.result.stderr = self.collect_output(proc)
if kmemleak:
cmd = f'{SUDO} sh -c "echo scan > {KMEMLEAK_FILE}"'
check_output(cmd, shell=True)
cmd = f'{SUDO} cat {KMEMLEAK_FILE}'
self.result.kmemleak = check_output(cmd, shell=True)
except KeyboardInterrupt:
self.kill_cmd(proc, True)
fail('\nRun terminated at user request.')
finally:
t.cancel()
if self.reran is not False:
self.result.done(proc, self.killed, self.reran)
def skip(self):
"""
Initialize enough of the test result that we can log a skipped
command.
"""
Result.total += 1
Result.runresults['SKIP'] += 1
self.result.stdout = self.result.stderr = []
self.result.starttime = monotonic_time()
m, s = divmod(monotonic_time() - self.result.starttime, 60)
self.result.runtime = '%02d:%02d' % (m, s)
self.result.result = 'SKIP'
def log(self, options, suppress_console=False):
"""
This function is responsible for writing all output. This includes
the console output, the logfile of all results (with timestamped
merged stdout and stderr), and for each test, the unmodified
stdout/stderr/merged in its own file.
"""
logname = getpwuid(os.getuid()).pw_name
rer = ''
if self.reran is True:
rer = ' (RERAN)'
user = ' (run as %s)' % (self.user if len(self.user) else logname)
if self.identifier:
msga = 'Test (%s): %s%s ' % (self.identifier, self.pathname, user)
else:
msga = 'Test: %s%s ' % (self.pathname, user)
msgb = '[%s] [%s]%s\n' % (self.result.runtime, self.result.result, rer)
pad = ' ' * (80 - (len(msga) + len(msgb)))
result_line = msga + pad + msgb
# The result line is always written to the log file. If -q was
# specified only failures are written to the console, otherwise
# the result line is written to the console. The console output
# may be suppressed by calling log() with suppress_console=True.
write_log(bytearray(result_line, encoding='utf-8'), LOG_FILE)
if not suppress_console:
if not options.quiet:
write_log(result_line, LOG_OUT)
elif options.quiet and self.result.result != 'PASS':
write_log(result_line, LOG_OUT)
lines = sorted(self.result.stdout + self.result.stderr,
key=lambda x: x[0])
# Write timestamped output (stdout and stderr) to the logfile
for dt, line in lines:
timestamp = bytearray(dt.strftime("%H:%M:%S.%f ")[:11],
encoding='utf-8')
write_log(b'%s %s\n' % (timestamp, line), LOG_FILE)
# Write the separate stdout/stderr/merged files, if the data exists
if len(self.result.stdout):
with open(os.path.join(self.outputdir, 'stdout'), 'wb') as out:
for _, line in self.result.stdout:
os.write(out.fileno(), b'%s\n' % line)
if len(self.result.stderr):
with open(os.path.join(self.outputdir, 'stderr'), 'wb') as err:
for _, line in self.result.stderr:
os.write(err.fileno(), b'%s\n' % line)
if len(self.result.stdout) and len(self.result.stderr):
with open(os.path.join(self.outputdir, 'merged'), 'wb') as merged:
for _, line in lines:
os.write(merged.fileno(), b'%s\n' % line)
if len(self.result.kmemleak):
with open(os.path.join(self.outputdir, 'kmemleak'), 'wb') as kmem:
kmem.write(self.result.kmemleak)
class Test(Cmd):
props = ['outputdir', 'timeout', 'user', 'pre', 'pre_user', 'post',
'post_user', 'failsafe', 'failsafe_user', 'tags']
def __init__(self, pathname,
pre=None, pre_user=None, post=None, post_user=None,
failsafe=None, failsafe_user=None, tags=None, **kwargs):
super(Test, self).__init__(pathname, **kwargs)
self.pre = pre or ''
self.pre_user = pre_user or ''
self.post = post or ''
self.post_user = post_user or ''
self.failsafe = failsafe or ''
self.failsafe_user = failsafe_user or ''
self.tags = tags or []
def __str__(self):
post_user = pre_user = failsafe_user = ''
if len(self.pre_user):
pre_user = ' (as %s)' % (self.pre_user)
if len(self.post_user):
post_user = ' (as %s)' % (self.post_user)
if len(self.failsafe_user):
failsafe_user = ' (as %s)' % (self.failsafe_user)
return '''\
Pathname: %s
Identifier: %s
Outputdir: %s
Timeout: %d
User: %s
Pre: %s%s
Post: %s%s
Failsafe: %s%s
Tags: %s
''' % (self.pathname, self.identifier, self.outputdir, self.timeout, self.user,
self.pre, pre_user, self.post, post_user, self.failsafe,
failsafe_user, self.tags)
def verify(self):
"""
Check the pre/post/failsafe scripts, user and Test. Omit the Test from
this run if there are any problems.
"""
files = [self.pre, self.pathname, self.post, self.failsafe]
users = [self.pre_user, self.user, self.post_user, self.failsafe_user]
for f in [f for f in files if len(f)]:
if not verify_file(f):
write_log("Warning: Test '%s' not added to this run because"
" it failed verification.\n" % f, LOG_ERR)
return False
for user in [user for user in users if len(user)]:
if not verify_user(user):
write_log("Not adding Test '%s' to this run.\n" %
self.pathname, LOG_ERR)
return False
return True
def run(self, options):
"""
Create Cmd instances for the pre/post/failsafe scripts. If the pre
script doesn't pass, skip this Test. Run the post script regardless.
If the Test is killed, also run the failsafe script.
"""
odir = os.path.join(self.outputdir, os.path.basename(self.pre))
pretest = Cmd(self.pre, identifier=self.identifier, outputdir=odir,
timeout=self.timeout, user=self.pre_user)
test = Cmd(self.pathname, identifier=self.identifier,
outputdir=self.outputdir, timeout=self.timeout,
user=self.user)
odir = os.path.join(self.outputdir, os.path.basename(self.failsafe))
failsafe = Cmd(self.failsafe, identifier=self.identifier,
outputdir=odir, timeout=self.timeout,
user=self.failsafe_user)
odir = os.path.join(self.outputdir, os.path.basename(self.post))
posttest = Cmd(self.post, identifier=self.identifier, outputdir=odir,
timeout=self.timeout, user=self.post_user)
cont = True
if len(pretest.pathname):
pretest.run(options.dryrun, False, options.kmsg)
cont = pretest.result.result == 'PASS'
pretest.log(options)
if cont:
test.run(options.dryrun, options.kmemleak, options.kmsg)
if test.result.result == 'KILLED' and len(failsafe.pathname):
failsafe.run(options.dryrun, False, options.kmsg)
failsafe.log(options, suppress_console=True)
else:
test.skip()
test.log(options)
if len(posttest.pathname):
posttest.run(options.dryrun, False, options.kmsg)
posttest.log(options)
class TestGroup(Test):
props = Test.props + ['tests']
def __init__(self, pathname, tests=None, **kwargs):
super(TestGroup, self).__init__(pathname, **kwargs)
self.tests = tests or []
def __str__(self):
post_user = pre_user = failsafe_user = ''
if len(self.pre_user):
pre_user = ' (as %s)' % (self.pre_user)
if len(self.post_user):
post_user = ' (as %s)' % (self.post_user)
if len(self.failsafe_user):
failsafe_user = ' (as %s)' % (self.failsafe_user)
return '''\
Pathname: %s
Identifier: %s
Outputdir: %s
Tests: %s
Timeout: %s
User: %s
Pre: %s%s
Post: %s%s
Failsafe: %s%s
Tags: %s
''' % (self.pathname, self.identifier, self.outputdir, self.tests,
self.timeout, self.user, self.pre, pre_user, self.post, post_user,
self.failsafe, failsafe_user, self.tags)
def filter(self, keeplist):
self.tests = [x for x in self.tests if x in keeplist]
def verify(self):
"""
Check the pre/post/failsafe scripts, user and tests in this TestGroup.
Omit the TestGroup entirely, or simply delete the relevant tests in the
group, if that's all that's required.
"""
# If the pre/post/failsafe scripts are relative pathnames, convert to
# absolute, so they stand a chance of passing verification.
if len(self.pre) and not os.path.isabs(self.pre):
self.pre = os.path.join(self.pathname, self.pre)
if len(self.post) and not os.path.isabs(self.post):
self.post = os.path.join(self.pathname, self.post)
if len(self.failsafe) and not os.path.isabs(self.failsafe):
self.post = os.path.join(self.pathname, self.post)
auxfiles = [self.pre, self.post, self.failsafe]
users = [self.pre_user, self.user, self.post_user, self.failsafe_user]
for f in [f for f in auxfiles if len(f)]:
if f != self.failsafe and self.pathname != os.path.dirname(f):
write_log("Warning: TestGroup '%s' not added to this run. "
"Auxiliary script '%s' exists in a different "
"directory.\n" % (self.pathname, f), LOG_ERR)
return False
if not verify_file(f):
write_log("Warning: TestGroup '%s' not added to this run. "
"Auxiliary script '%s' failed verification.\n" %
(self.pathname, f), LOG_ERR)
return False
for user in [user for user in users if len(user)]:
if not verify_user(user):
write_log("Not adding TestGroup '%s' to this run.\n" %
self.pathname, LOG_ERR)
return False
# If one of the tests is invalid, delete it, log it, and drive on.
for test in self.tests:
if not verify_file(os.path.join(self.pathname, test)):
del self.tests[self.tests.index(test)]
write_log("Warning: Test '%s' removed from TestGroup '%s' "
"because it failed verification.\n" %
(test, self.pathname), LOG_ERR)
return len(self.tests) != 0
def run(self, options):
"""
Create Cmd instances for the pre/post/failsafe scripts. If the pre
script doesn't pass, skip all the tests in this TestGroup. Run the
post script regardless. Run the failsafe script when a test is killed.
"""
# tags assigned to this test group also include the test names
if options.tags and not set(self.tags).intersection(set(options.tags)):
return
odir = os.path.join(self.outputdir, os.path.basename(self.pre))
pretest = Cmd(self.pre, outputdir=odir, timeout=self.timeout,
user=self.pre_user, identifier=self.identifier)
odir = os.path.join(self.outputdir, os.path.basename(self.post))
posttest = Cmd(self.post, outputdir=odir, timeout=self.timeout,
user=self.post_user, identifier=self.identifier)
cont = True
if len(pretest.pathname):
pretest.run(options.dryrun, False, options.kmsg)
cont = pretest.result.result == 'PASS'
pretest.log(options)
for fname in self.tests:
odir = os.path.join(self.outputdir, fname)
test = Cmd(os.path.join(self.pathname, fname), outputdir=odir,
timeout=self.timeout, user=self.user,
identifier=self.identifier)
odir = os.path.join(odir, os.path.basename(self.failsafe))
failsafe = Cmd(self.failsafe, outputdir=odir, timeout=self.timeout,
user=self.failsafe_user, identifier=self.identifier)
if cont:
test.run(options.dryrun, options.kmemleak, options.kmsg)
if test.result.result == 'KILLED' and len(failsafe.pathname):
failsafe.run(options.dryrun, False, options.kmsg)
failsafe.log(options, suppress_console=True)
else:
test.skip()
test.log(options)
if len(posttest.pathname):
posttest.run(options.dryrun, False, options.kmsg)
posttest.log(options)
class TestRun(object):
props = ['quiet', 'outputdir']
def __init__(self, options):
self.tests = {}
self.testgroups = {}
self.starttime = time()
self.timestamp = datetime.now().strftime('%Y%m%dT%H%M%S')
self.outputdir = os.path.join(options.outputdir, self.timestamp)
self.setup_logging(options)
self.defaults = [
('outputdir', BASEDIR),
('quiet', False),
('timeout', 60),
('user', ''),
('pre', ''),
('pre_user', ''),
('post', ''),
('post_user', ''),
('failsafe', ''),
('failsafe_user', ''),
('tags', [])
]
def __str__(self):
s = 'TestRun:\n outputdir: %s\n' % self.outputdir
s += 'TESTS:\n'
for key in sorted(self.tests.keys()):
s += '%s%s' % (self.tests[key].__str__(), '\n')
s += 'TESTGROUPS:\n'
for key in sorted(self.testgroups.keys()):
s += '%s%s' % (self.testgroups[key].__str__(), '\n')
return s
def addtest(self, pathname, options):
"""
Create a new Test, and apply any properties that were passed in
from the command line. If it passes verification, add it to the
TestRun.
"""
test = Test(pathname)
for prop in Test.props:
setattr(test, prop, getattr(options, prop))
if test.verify():
self.tests[pathname] = test
def addtestgroup(self, dirname, filenames, options):
"""
Create a new TestGroup, and apply any properties that were passed
in from the command line. If it passes verification, add it to the
TestRun.
"""
if dirname not in self.testgroups:
testgroup = TestGroup(dirname)
for prop in Test.props:
setattr(testgroup, prop, getattr(options, prop))
# Prevent pre/post/failsafe scripts from running as regular tests
for f in [testgroup.pre, testgroup.post, testgroup.failsafe]:
if f in filenames:
del filenames[filenames.index(f)]
self.testgroups[dirname] = testgroup
self.testgroups[dirname].tests = sorted(filenames)
testgroup.verify()
def filter(self, keeplist):
for group in list(self.testgroups.keys()):
if group not in keeplist:
del self.testgroups[group]
continue
g = self.testgroups[group]
if g.pre and os.path.basename(g.pre) in keeplist[group]:
continue
g.filter(keeplist[group])
for test in list(self.tests.keys()):
directory, base = os.path.split(test)
if directory not in keeplist or base not in keeplist[directory]:
del self.tests[test]
def read(self, options):
"""
Read in the specified runfiles, and apply the TestRun properties
listed in the 'DEFAULT' section to our TestRun. Then read each
section, and apply the appropriate properties to the Test or
TestGroup. Properties from individual sections override those set
in the 'DEFAULT' section. If the Test or TestGroup passes
verification, add it to the TestRun.
"""
config = configparser.RawConfigParser()
parsed = config.read(options.runfiles)
failed = options.runfiles - set(parsed)
if len(failed):
files = ' '.join(sorted(failed))
fail("Couldn't read config files: %s" % files)
for opt in TestRun.props:
if config.has_option('DEFAULT', opt):
setattr(self, opt, config.get('DEFAULT', opt))
self.outputdir = os.path.join(self.outputdir, self.timestamp)
testdir = options.testdir
for section in config.sections():
if 'tests' in config.options(section):
parts = section.split(':', 1)
sectiondir = parts[0]
identifier = parts[1] if len(parts) == 2 else None
if os.path.isdir(sectiondir):
pathname = sectiondir
elif os.path.isdir(os.path.join(testdir, sectiondir)):
pathname = os.path.join(testdir, sectiondir)
else:
pathname = sectiondir
testgroup = TestGroup(os.path.abspath(pathname),
identifier=identifier)
for prop in TestGroup.props:
for sect in ['DEFAULT', section]:
if config.has_option(sect, prop):
if prop == 'tags':
setattr(testgroup, prop,
eval(config.get(sect, prop)))
elif prop == 'failsafe':
failsafe = config.get(sect, prop)
setattr(testgroup, prop,
os.path.join(testdir, failsafe))
else:
setattr(testgroup, prop,
config.get(sect, prop))
# Repopulate tests using eval to convert the string to a list
testgroup.tests = eval(config.get(section, 'tests'))
if testgroup.verify():
self.testgroups[section] = testgroup
else:
test = Test(section)
for prop in Test.props:
for sect in ['DEFAULT', section]:
if config.has_option(sect, prop):
if prop == 'failsafe':
failsafe = config.get(sect, prop)
setattr(test, prop,
os.path.join(testdir, failsafe))
else:
setattr(test, prop, config.get(sect, prop))
if test.verify():
self.tests[section] = test
def write(self, options):
"""
Create a configuration file for editing and later use. The
'DEFAULT' section of the config file is created from the
properties that were specified on the command line. Tests are
simply added as sections that inherit everything from the
'DEFAULT' section. TestGroups are the same, except they get an
option including all the tests to run in that directory.
"""
defaults = dict([(prop, getattr(options, prop)) for prop, _ in
self.defaults])
config = configparser.RawConfigParser(defaults)
for test in sorted(self.tests.keys()):
config.add_section(test)
for prop in Test.props:
if prop not in self.props:
config.set(test, prop,
getattr(self.tests[test], prop))
for testgroup in sorted(self.testgroups.keys()):
config.add_section(testgroup)
config.set(testgroup, 'tests', self.testgroups[testgroup].tests)
for prop in TestGroup.props:
if prop not in self.props:
config.set(testgroup, prop,
getattr(self.testgroups[testgroup], prop))
try:
with open(options.template, 'w') as f:
return config.write(f)
except IOError:
fail('Could not open \'%s\' for writing.' % options.template)
def complete_outputdirs(self):
"""
Collect all the pathnames for Tests, and TestGroups. Work
backwards one pathname component at a time, to create a unique
directory name in which to deposit test output. Tests will be able
to write output files directly in the newly modified outputdir.
TestGroups will be able to create one subdirectory per test in the
outputdir, and are guaranteed uniqueness because a group can only
contain files in one directory. Pre and post tests will create a
directory rooted at the outputdir of the Test or TestGroup in
question for their output. Failsafe scripts will create a directory
rooted at the outputdir of each Test for their output.
"""
done = False
components = 0
tmp_dict = dict(list(self.tests.items()) +
list(self.testgroups.items()))
total = len(tmp_dict)
base = self.outputdir
while not done:
paths = []
components -= 1
for testfile in list(tmp_dict.keys()):
uniq = '/'.join(testfile.split('/')[components:]).lstrip('/')
if uniq not in paths:
paths.append(uniq)
tmp_dict[testfile].outputdir = os.path.join(base, uniq)
else:
break
done = total == len(paths)
def setup_logging(self, options):
"""
This function creates the output directory and gets a file object
for the logfile. This function must be called before write_log()
can be used.
"""
if options.dryrun is True:
return
global LOG_FILE_OBJ
if not options.template:
try:
old = os.umask(0)
os.makedirs(self.outputdir, mode=0o777)
os.umask(old)
filename = os.path.join(self.outputdir, 'log')
LOG_FILE_OBJ = open(filename, buffering=0, mode='wb')
except OSError as e:
fail('%s' % e)
def run(self, options):
"""
Walk through all the Tests and TestGroups, calling run().
"""
try:
os.chdir(self.outputdir)
except OSError:
fail('Could not change to directory %s' % self.outputdir)
# make a symlink to the output for the currently running test
logsymlink = os.path.join(self.outputdir, '../current')
if os.path.islink(logsymlink):
os.unlink(logsymlink)
if not os.path.exists(logsymlink):
os.symlink(self.outputdir, logsymlink)
else:
write_log('Could not make a symlink to directory %s\n' %
self.outputdir, LOG_ERR)
if options.kmemleak:
cmd = f'{SUDO} -c "echo scan=0 > {KMEMLEAK_FILE}"'
check_output(cmd, shell=True)
iteration = 0
while iteration < options.iterations:
for test in sorted(self.tests.keys()):
self.tests[test].run(options)
for testgroup in sorted(self.testgroups.keys()):
self.testgroups[testgroup].run(options)
iteration += 1
def summary(self):
if Result.total == 0:
return 2
print('\nResults Summary')
for key in list(Result.runresults.keys()):
if Result.runresults[key] != 0:
print('%s\t% 4d' % (key, Result.runresults[key]))
m, s = divmod(time() - self.starttime, 60)
h, m = divmod(m, 60)
print('\nRunning Time:\t%02d:%02d:%02d' % (h, m, s))
print('Percent passed:\t%.1f%%' % ((float(Result.runresults['PASS']) /
float(Result.total)) * 100))
print('Log directory:\t%s' % self.outputdir)
if Result.runresults['FAIL'] > 0:
return 1
if Result.runresults['KILLED'] > 0:
return 1
if Result.runresults['RERAN'] > 0:
return 3
return 0
def write_log(msg, target):
"""
Write the provided message to standard out, standard error or
the logfile. If specifying LOG_FILE, then `msg` must be a bytes
like object. This way we can still handle output from tests that
may be in unexpected encodings.
"""
if target == LOG_OUT:
os.write(sys.stdout.fileno(), bytearray(msg, encoding='utf-8'))
elif target == LOG_ERR:
os.write(sys.stderr.fileno(), bytearray(msg, encoding='utf-8'))
elif target == LOG_FILE:
os.write(LOG_FILE_OBJ.fileno(), msg)
else:
fail('log_msg called with unknown target "%s"' % target)
def verify_file(pathname):
"""
Verify that the supplied pathname is an executable regular file.
"""
if os.path.isdir(pathname) or os.path.islink(pathname):
return False
for ext in '', '.ksh', '.sh':
script_path = pathname + ext
if os.path.isfile(script_path) and os.access(script_path, os.X_OK):
return True
return False
def verify_user(user):
"""
Verify that the specified user exists on this system, and can execute
sudo without being prompted for a password.
"""
testcmd = [SUDO, '-n', '-u', user, TRUE]
if user in Cmd.verified_users:
return True
try:
getpwnam(user)
except KeyError:
write_log("Warning: user '%s' does not exist.\n" % user,
LOG_ERR)
return False
p = Popen(testcmd)
p.wait()
if p.returncode != 0:
write_log("Warning: user '%s' cannot use passwordless sudo.\n" % user,
LOG_ERR)
return False
else:
Cmd.verified_users.append(user)
return True
def find_tests(testrun, options):
"""
For the given list of pathnames, add files as Tests. For directories,
if do_groups is True, add the directory as a TestGroup. If False,
recursively search for executable files.
"""
for p in sorted(options.pathnames):
if os.path.isdir(p):
for dirname, _, filenames in os.walk(p):
if options.do_groups:
testrun.addtestgroup(dirname, filenames, options)
else:
for f in sorted(filenames):
testrun.addtest(os.path.join(dirname, f), options)
else:
testrun.addtest(p, options)
def filter_tests(testrun, options):
try:
fh = open(options.logfile, "r")
except Exception as e:
fail('%s' % e)
failed = {}
while True:
line = fh.readline()
if not line:
break
m = re.match(r'Test: .*(tests/.*)/(\S+).*\[FAIL\]', line)
if not m:
continue
group, test = m.group(1, 2)
try:
failed[group].append(test)
except KeyError:
failed[group] = [test]
fh.close()
testrun.filter(failed)
def fail(retstr, ret=1):
print('%s: %s' % (sys.argv[0], retstr))
exit(ret)
def kmemleak_cb(option, opt_str, value, parser):
if not os.path.exists(KMEMLEAK_FILE):
fail(f"File '{KMEMLEAK_FILE}' doesn't exist. " +
"Enable CONFIG_DEBUG_KMEMLEAK in kernel configuration.")
setattr(parser.values, option.dest, True)
def options_cb(option, opt_str, value, parser):
path_options = ['outputdir', 'template', 'testdir', 'logfile']
if opt_str in parser.rargs:
fail('%s may only be specified once.' % opt_str)
if option.dest == 'runfiles':
parser.values.cmd = 'rdconfig'
value = set(os.path.abspath(p) for p in value.split(','))
if option.dest == 'tags':
value = [x.strip() for x in value.split(',')]
if option.dest in path_options:
setattr(parser.values, option.dest, os.path.abspath(value))
else:
setattr(parser.values, option.dest, value)
def parse_args():
parser = OptionParser()
parser.add_option('-c', action='callback', callback=options_cb,
type='string', dest='runfiles', metavar='runfiles',
help='Specify tests to run via config files.')
parser.add_option('-d', action='store_true', default=False, dest='dryrun',
help='Dry run. Print tests, but take no other action.')
parser.add_option('-l', action='callback', callback=options_cb,
default=None, dest='logfile', metavar='logfile',
type='string',
help='Read logfile and re-run tests which failed.')
parser.add_option('-g', action='store_true', default=False,
dest='do_groups', help='Make directories TestGroups.')
parser.add_option('-o', action='callback', callback=options_cb,
default=BASEDIR, dest='outputdir', type='string',
metavar='outputdir', help='Specify an output directory.')
parser.add_option('-i', action='callback', callback=options_cb,
default=TESTDIR, dest='testdir', type='string',
metavar='testdir', help='Specify a test directory.')
parser.add_option('-K', action='store_true', default=False, dest='kmsg',
help='Log tests names to /dev/kmsg')
parser.add_option('-m', action='callback', callback=kmemleak_cb,
default=False, dest='kmemleak',
help='Enable kmemleak reporting (Linux only)')
parser.add_option('-p', action='callback', callback=options_cb,
default='', dest='pre', metavar='script',
type='string', help='Specify a pre script.')
parser.add_option('-P', action='callback', callback=options_cb,
default='', dest='post', metavar='script',
type='string', help='Specify a post script.')
parser.add_option('-q', action='store_true', default=False, dest='quiet',
help='Silence on the console during a test run.')
parser.add_option('-s', action='callback', callback=options_cb,
default='', dest='failsafe', metavar='script',
type='string', help='Specify a failsafe script.')
parser.add_option('-S', action='callback', callback=options_cb,
default='', dest='failsafe_user',
metavar='failsafe_user', type='string',
help='Specify a user to execute the failsafe script.')
parser.add_option('-t', action='callback', callback=options_cb, default=60,
dest='timeout', metavar='seconds', type='int',
help='Timeout (in seconds) for an individual test.')
parser.add_option('-u', action='callback', callback=options_cb,
default='', dest='user', metavar='user', type='string',
help='Specify a different user name to run as.')
parser.add_option('-w', action='callback', callback=options_cb,
default=None, dest='template', metavar='template',
type='string', help='Create a new config file.')
parser.add_option('-x', action='callback', callback=options_cb, default='',
dest='pre_user', metavar='pre_user', type='string',
help='Specify a user to execute the pre script.')
parser.add_option('-X', action='callback', callback=options_cb, default='',
dest='post_user', metavar='post_user', type='string',
help='Specify a user to execute the post script.')
parser.add_option('-T', action='callback', callback=options_cb, default='',
dest='tags', metavar='tags', type='string',
help='Specify tags to execute specific test groups.')
parser.add_option('-I', action='callback', callback=options_cb, default=1,
dest='iterations', metavar='iterations', type='int',
help='Number of times to run the test run.')
(options, pathnames) = parser.parse_args()
if options.runfiles and len(pathnames):
fail('Extraneous arguments.')
options.pathnames = [os.path.abspath(path) for path in pathnames]
return options
def main():
options = parse_args()
testrun = TestRun(options)
if options.runfiles:
testrun.read(options)
else:
find_tests(testrun, options)
if options.logfile:
filter_tests(testrun, options)
if options.template:
testrun.write(options)
exit(0)
testrun.complete_outputdirs()
testrun.run(options)
exit(testrun.summary())
if __name__ == '__main__':
main()
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am b/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
index 9b46beaf460f..74295b86ddc2 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am
@@ -1,2034 +1,2038 @@
CLEANFILES =
dist_noinst_DATA =
include $(top_srcdir)/config/Substfiles.am
datadir_zfs_tests_testsdir = $(datadir)/$(PACKAGE)/zfs-tests/tests
nobase_dist_datadir_zfs_tests_tests_DATA = \
perf/nfs-sample.cfg \
perf/perf.shlib \
\
perf/fio/mkfiles.fio \
perf/fio/random_reads.fio \
perf/fio/random_readwrite.fio \
perf/fio/random_readwrite_fixed.fio \
perf/fio/random_writes.fio \
perf/fio/sequential_reads.fio \
perf/fio/sequential_readwrite.fio \
perf/fio/sequential_writes.fio
nobase_dist_datadir_zfs_tests_tests_SCRIPTS = \
perf/regression/random_reads.ksh \
perf/regression/random_readwrite.ksh \
perf/regression/random_readwrite_fixed.ksh \
perf/regression/random_writes.ksh \
perf/regression/random_writes_zil.ksh \
perf/regression/sequential_reads_arc_cached_clone.ksh \
perf/regression/sequential_reads_arc_cached.ksh \
perf/regression/sequential_reads_dbuf_cached.ksh \
perf/regression/sequential_reads.ksh \
perf/regression/sequential_writes.ksh \
perf/regression/setup.ksh \
\
perf/scripts/prefetch_io.sh
# These lists can be regenerated by running make regen-tests at the root, or, on a *clean* source:
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -executable -name '*.in' | sort | sed 's/\.in$//;s/^/\t/;$!s/$/ \\/'
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' -executable -name '*.in' | sort | sed 's/\.in$//;s/^/\t/;$!s/$/ \\/'
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -name '*.in' ! -name '*.c' | grep -Fe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/'
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/'
# find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/'
#
# simd and tmpfile are Linux-only and not installed elsewhere
#
# C programs are specced in ../Makefile.am above as part of the main Makefile
find_common := find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po'
regen:
@$(MAKE) -C $(top_builddir) clean
@$(MAKE) clean
$(SED) $(ac_inplace) '/^# -- >8 --/q' Makefile.am
echo >> Makefile.am
echo 'nobase_nodist_datadir_zfs_tests_tests_DATA = \' >> Makefile.am
$(find_common) ! -executable -name '*.in' | sort | sed 's/\.in$$//;s/^/\t/;$$!s/$$/ \\/' >> Makefile.am
echo 'nobase_nodist_datadir_zfs_tests_tests_SCRIPTS = \' >> Makefile.am
$(find_common) -executable -name '*.in' | sort | sed 's/\.in$$//;s/^/\t/;$$!s/$$/ \\/' >> Makefile.am
echo >> Makefile.am
echo 'SUBSTFILES += $$(nobase_nodist_datadir_zfs_tests_tests_DATA) $$(nobase_nodist_datadir_zfs_tests_tests_SCRIPTS)' >> Makefile.am
echo >> Makefile.am
echo 'if BUILD_LINUX' >> Makefile.am
echo 'nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \' >> Makefile.am
$(find_common) ! -name '*.in' ! -name '*.c' | grep -Fe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am
echo 'endif' >> Makefile.am
echo >> Makefile.am
echo 'nobase_dist_datadir_zfs_tests_tests_DATA += \' >> Makefile.am
$(find_common) ! -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am
echo >> Makefile.am
echo 'nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \' >> Makefile.am
$(find_common) -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am
# -- >8 --
nobase_nodist_datadir_zfs_tests_tests_DATA = \
functional/pam/utilities.kshlib
nobase_nodist_datadir_zfs_tests_tests_SCRIPTS = \
functional/pyzfs/pyzfs_unittest.ksh
SUBSTFILES += $(nobase_nodist_datadir_zfs_tests_tests_DATA) $(nobase_nodist_datadir_zfs_tests_tests_SCRIPTS)
if BUILD_LINUX
nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \
functional/simd/simd_supported.ksh \
functional/tmpfile/cleanup.ksh \
functional/tmpfile/setup.ksh
endif
nobase_dist_datadir_zfs_tests_tests_DATA += \
functional/acl/acl.cfg \
functional/acl/acl_common.kshlib \
functional/alloc_class/alloc_class.cfg \
functional/alloc_class/alloc_class.kshlib \
functional/atime/atime.cfg \
functional/atime/atime_common.kshlib \
functional/cache/cache.cfg \
functional/cache/cache.kshlib \
functional/cachefile/cachefile.cfg \
functional/cachefile/cachefile.kshlib \
functional/casenorm/casenorm.cfg \
functional/casenorm/casenorm.kshlib \
functional/channel_program/channel_common.kshlib \
functional/channel_program/lua_core/tst.args_to_lua.out \
functional/channel_program/lua_core/tst.args_to_lua.zcp \
functional/channel_program/lua_core/tst.divide_by_zero.err \
functional/channel_program/lua_core/tst.divide_by_zero.zcp \
functional/channel_program/lua_core/tst.exists.zcp \
functional/channel_program/lua_core/tst.large_prog.out \
functional/channel_program/lua_core/tst.large_prog.zcp \
functional/channel_program/lua_core/tst.lib_base.lua \
functional/channel_program/lua_core/tst.lib_coroutine.lua \
functional/channel_program/lua_core/tst.lib_strings.lua \
functional/channel_program/lua_core/tst.lib_table.lua \
functional/channel_program/lua_core/tst.nested_neg.zcp \
functional/channel_program/lua_core/tst.nested_pos.zcp \
functional/channel_program/lua_core/tst.recursive.zcp \
functional/channel_program/lua_core/tst.return_large.zcp \
functional/channel_program/lua_core/tst.return_recursive_table.zcp \
functional/channel_program/lua_core/tst.stack_gsub.err \
functional/channel_program/lua_core/tst.stack_gsub.zcp \
functional/channel_program/lua_core/tst.timeout.zcp \
functional/channel_program/synctask_core/tst.bookmark.copy.zcp \
functional/channel_program/synctask_core/tst.bookmark.create.zcp \
functional/channel_program/synctask_core/tst.get_index_props.out \
functional/channel_program/synctask_core/tst.get_index_props.zcp \
functional/channel_program/synctask_core/tst.get_number_props.out \
functional/channel_program/synctask_core/tst.get_number_props.zcp \
functional/channel_program/synctask_core/tst.get_string_props.out \
functional/channel_program/synctask_core/tst.get_string_props.zcp \
functional/channel_program/synctask_core/tst.promote_conflict.zcp \
functional/channel_program/synctask_core/tst.set_props.zcp \
functional/channel_program/synctask_core/tst.snapshot_destroy.zcp \
functional/channel_program/synctask_core/tst.snapshot_neg.zcp \
functional/channel_program/synctask_core/tst.snapshot_recursive.zcp \
functional/channel_program/synctask_core/tst.snapshot_rename.zcp \
functional/channel_program/synctask_core/tst.snapshot_simple.zcp \
functional/checksum/default.cfg \
functional/clean_mirror/clean_mirror_common.kshlib \
functional/clean_mirror/default.cfg \
functional/cli_root/cli_common.kshlib \
functional/cli_root/zfs_copies/zfs_copies.cfg \
functional/cli_root/zfs_copies/zfs_copies.kshlib \
functional/cli_root/zfs_create/properties.kshlib \
functional/cli_root/zfs_create/zfs_create.cfg \
functional/cli_root/zfs_create/zfs_create_common.kshlib \
functional/cli_root/zfs_destroy/zfs_destroy.cfg \
functional/cli_root/zfs_destroy/zfs_destroy_common.kshlib \
functional/cli_root/zfs_get/zfs_get_common.kshlib \
functional/cli_root/zfs_get/zfs_get_list_d.kshlib \
functional/cli_root/zfs_jail/jail.conf \
functional/cli_root/zfs_load-key/HEXKEY \
functional/cli_root/zfs_load-key/PASSPHRASE \
functional/cli_root/zfs_load-key/RAWKEY \
functional/cli_root/zfs_load-key/zfs_load-key.cfg \
functional/cli_root/zfs_load-key/zfs_load-key_common.kshlib \
functional/cli_root/zfs_mount/zfs_mount.cfg \
functional/cli_root/zfs_mount/zfs_mount.kshlib \
functional/cli_root/zfs_promote/zfs_promote.cfg \
functional/cli_root/zfs_receive/zstd_test_data.txt \
functional/cli_root/zfs_rename/zfs_rename.cfg \
functional/cli_root/zfs_rename/zfs_rename.kshlib \
functional/cli_root/zfs_rollback/zfs_rollback.cfg \
functional/cli_root/zfs_rollback/zfs_rollback_common.kshlib \
functional/cli_root/zfs_send/zfs_send.cfg \
functional/cli_root/zfs_set/zfs_set_common.kshlib \
functional/cli_root/zfs_share/zfs_share.cfg \
functional/cli_root/zfs_snapshot/zfs_snapshot.cfg \
functional/cli_root/zfs_unmount/zfs_unmount.cfg \
functional/cli_root/zfs_unmount/zfs_unmount.kshlib \
functional/cli_root/zfs_upgrade/zfs_upgrade.kshlib \
functional/cli_root/zfs_wait/zfs_wait.kshlib \
functional/cli_root/zpool_add/zpool_add.cfg \
functional/cli_root/zpool_add/zpool_add.kshlib \
functional/cli_root/zpool_clear/zpool_clear.cfg \
functional/cli_root/zpool_create/draidcfg.gz \
functional/cli_root/zpool_create/zpool_create.cfg \
functional/cli_root/zpool_create/zpool_create.shlib \
functional/cli_root/zpool_destroy/zpool_destroy.cfg \
functional/cli_root/zpool_events/zpool_events.cfg \
functional/cli_root/zpool_events/zpool_events.kshlib \
functional/cli_root/zpool_expand/zpool_expand.cfg \
functional/cli_root/zpool_export/zpool_export.cfg \
functional/cli_root/zpool_export/zpool_export.kshlib \
+ functional/cli_root/zpool_get/vdev_get.cfg \
functional/cli_root/zpool_get/zpool_get.cfg \
functional/cli_root/zpool_get/zpool_get_parsable.cfg \
functional/cli_root/zpool_import/blockfiles/cryptv0.dat.bz2 \
functional/cli_root/zpool_import/blockfiles/missing_ivset.dat.bz2 \
functional/cli_root/zpool_import/blockfiles/unclean_export.dat.bz2 \
functional/cli_root/zpool_import/zpool_import.cfg \
functional/cli_root/zpool_import/zpool_import.kshlib \
functional/cli_root/zpool_initialize/zpool_initialize.kshlib \
functional/cli_root/zpool_labelclear/labelclear.cfg \
functional/cli_root/zpool_remove/zpool_remove.cfg \
functional/cli_root/zpool_reopen/zpool_reopen.cfg \
functional/cli_root/zpool_reopen/zpool_reopen.shlib \
functional/cli_root/zpool_resilver/zpool_resilver.cfg \
functional/cli_root/zpool_scrub/zpool_scrub.cfg \
functional/cli_root/zpool_split/zpool_split.cfg \
functional/cli_root/zpool_trim/zpool_trim.kshlib \
functional/cli_root/zpool_upgrade/blockfiles/zfs-broken-mirror1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-broken-mirror2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v10.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v11.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v12.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v13.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v14.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v15.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz21.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz22.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz23.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe1.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe2.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe3.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v4.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v5.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v6.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v7.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v8.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v999.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v9.dat.bz2 \
functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-vBROKEN.dat.bz2 \
functional/cli_root/zpool_upgrade/zpool_upgrade.cfg \
functional/cli_root/zpool_upgrade/zpool_upgrade.kshlib \
functional/cli_root/zpool_wait/zpool_wait.kshlib \
functional/cli_user/misc/misc.cfg \
functional/cli_user/zfs_list/zfs_list.cfg \
functional/cli_user/zfs_list/zfs_list.kshlib \
functional/compression/compress.cfg \
functional/compression/testpool_zstd.tar.gz \
functional/deadman/deadman.cfg \
functional/delegate/delegate.cfg \
functional/delegate/delegate_common.kshlib \
functional/devices/devices.cfg \
functional/devices/devices_common.kshlib \
functional/events/events.cfg \
functional/events/events_common.kshlib \
functional/fault/fault.cfg \
functional/grow/grow.cfg \
functional/history/history.cfg \
functional/history/history_common.kshlib \
functional/history/i386.migratedpool.DAT.Z \
functional/history/i386.orig_history.txt \
functional/history/sparc.migratedpool.DAT.Z \
functional/history/sparc.orig_history.txt \
functional/history/zfs-pool-v4.dat.Z \
functional/inheritance/config001.cfg \
functional/inheritance/config002.cfg \
functional/inheritance/config003.cfg \
functional/inheritance/config004.cfg \
functional/inheritance/config005.cfg \
functional/inheritance/config006.cfg \
functional/inheritance/config007.cfg \
functional/inheritance/config008.cfg \
functional/inheritance/config009.cfg \
functional/inheritance/config010.cfg \
functional/inheritance/config011.cfg \
functional/inheritance/config012.cfg \
functional/inheritance/config013.cfg \
functional/inheritance/config014.cfg \
functional/inheritance/config015.cfg \
functional/inheritance/config016.cfg \
functional/inheritance/config017.cfg \
functional/inheritance/config018.cfg \
functional/inheritance/config019.cfg \
functional/inheritance/config020.cfg \
functional/inheritance/config021.cfg \
functional/inheritance/config022.cfg \
functional/inheritance/config023.cfg \
functional/inheritance/config024.cfg \
functional/inheritance/inherit.kshlib \
functional/inheritance/README.config \
functional/inheritance/README.state \
functional/inheritance/state001.cfg \
functional/inheritance/state002.cfg \
functional/inheritance/state003.cfg \
functional/inheritance/state004.cfg \
functional/inheritance/state005.cfg \
functional/inheritance/state006.cfg \
functional/inheritance/state007.cfg \
functional/inheritance/state008.cfg \
functional/inheritance/state009.cfg \
functional/inheritance/state010.cfg \
functional/inheritance/state011.cfg \
functional/inheritance/state012.cfg \
functional/inheritance/state013.cfg \
functional/inheritance/state014.cfg \
functional/inheritance/state015.cfg \
functional/inheritance/state016.cfg \
functional/inheritance/state017.cfg \
functional/inheritance/state018.cfg \
functional/inheritance/state019.cfg \
functional/inheritance/state020.cfg \
functional/inheritance/state021.cfg \
functional/inheritance/state022.cfg \
functional/inheritance/state023.cfg \
functional/inheritance/state024.cfg \
functional/inuse/inuse.cfg \
functional/io/io.cfg \
functional/l2arc/l2arc.cfg \
functional/largest_pool/largest_pool.cfg \
functional/migration/migration.cfg \
functional/migration/migration.kshlib \
functional/mmap/mmap.cfg \
functional/mmp/mmp.cfg \
functional/mmp/mmp.kshlib \
functional/mv_files/mv_files.cfg \
functional/mv_files/mv_files_common.kshlib \
functional/nopwrite/nopwrite.shlib \
functional/no_space/enospc.cfg \
functional/online_offline/online_offline.cfg \
functional/pool_checkpoint/pool_checkpoint.kshlib \
functional/projectquota/projectquota.cfg \
functional/projectquota/projectquota_common.kshlib \
functional/quota/quota.cfg \
functional/quota/quota.kshlib \
functional/redacted_send/redacted.cfg \
functional/redacted_send/redacted.kshlib \
functional/redundancy/redundancy.cfg \
functional/redundancy/redundancy.kshlib \
functional/refreserv/refreserv.cfg \
functional/removal/removal.kshlib \
functional/replacement/replacement.cfg \
functional/reservation/reservation.cfg \
functional/reservation/reservation.shlib \
functional/rsend/dedup_encrypted_zvol.bz2 \
functional/rsend/dedup_encrypted_zvol.zsend.bz2 \
functional/rsend/dedup.zsend.bz2 \
functional/rsend/fs.tar.gz \
functional/rsend/rsend.cfg \
functional/rsend/rsend.kshlib \
functional/scrub_mirror/default.cfg \
functional/scrub_mirror/scrub_mirror_common.kshlib \
functional/slog/slog.cfg \
functional/slog/slog.kshlib \
functional/snapshot/snapshot.cfg \
functional/snapused/snapused.kshlib \
functional/sparse/sparse.cfg \
functional/trim/trim.cfg \
functional/trim/trim.kshlib \
functional/truncate/truncate.cfg \
functional/upgrade/upgrade_common.kshlib \
functional/user_namespace/user_namespace.cfg \
functional/user_namespace/user_namespace_common.kshlib \
functional/userquota/13709_reproducer.bz2 \
functional/userquota/userquota.cfg \
functional/userquota/userquota_common.kshlib \
functional/vdev_zaps/vdev_zaps.kshlib \
functional/xattr/xattr.cfg \
functional/xattr/xattr_common.kshlib \
functional/zvol/zvol.cfg \
functional/zvol/zvol_cli/zvol_cli.cfg \
functional/zvol/zvol_common.shlib \
functional/zvol/zvol_ENOSPC/zvol_ENOSPC.cfg \
functional/zvol/zvol_misc/zvol_misc_common.kshlib \
functional/zvol/zvol_swap/zvol_swap.cfg \
functional/idmap_mount/idmap_mount.cfg \
functional/idmap_mount/idmap_mount_common.kshlib
nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \
functional/acl/off/cleanup.ksh \
functional/acl/off/dosmode.ksh \
functional/acl/off/posixmode.ksh \
functional/acl/off/setup.ksh \
functional/acl/posix/cleanup.ksh \
functional/acl/posix/posix_001_pos.ksh \
functional/acl/posix/posix_002_pos.ksh \
functional/acl/posix/posix_003_pos.ksh \
functional/acl/posix/posix_004_pos.ksh \
functional/acl/posix-sa/cleanup.ksh \
functional/acl/posix-sa/posix_001_pos.ksh \
functional/acl/posix-sa/posix_002_pos.ksh \
functional/acl/posix-sa/posix_003_pos.ksh \
functional/acl/posix-sa/posix_004_pos.ksh \
functional/acl/posix-sa/setup.ksh \
functional/acl/posix/setup.ksh \
functional/alloc_class/alloc_class_001_pos.ksh \
functional/alloc_class/alloc_class_002_neg.ksh \
functional/alloc_class/alloc_class_003_pos.ksh \
functional/alloc_class/alloc_class_004_pos.ksh \
functional/alloc_class/alloc_class_005_pos.ksh \
functional/alloc_class/alloc_class_006_pos.ksh \
functional/alloc_class/alloc_class_007_pos.ksh \
functional/alloc_class/alloc_class_008_pos.ksh \
functional/alloc_class/alloc_class_009_pos.ksh \
functional/alloc_class/alloc_class_010_pos.ksh \
functional/alloc_class/alloc_class_011_neg.ksh \
functional/alloc_class/alloc_class_012_pos.ksh \
functional/alloc_class/alloc_class_013_pos.ksh \
functional/alloc_class/cleanup.ksh \
functional/alloc_class/setup.ksh \
functional/append/file_append.ksh \
functional/append/threadsappend_001_pos.ksh \
functional/append/cleanup.ksh \
functional/append/setup.ksh \
functional/arc/arcstats_runtime_tuning.ksh \
functional/arc/cleanup.ksh \
functional/arc/dbufstats_001_pos.ksh \
functional/arc/dbufstats_002_pos.ksh \
functional/arc/dbufstats_003_pos.ksh \
functional/arc/setup.ksh \
functional/atime/atime_001_pos.ksh \
functional/atime/atime_002_neg.ksh \
functional/atime/atime_003_pos.ksh \
functional/atime/cleanup.ksh \
functional/atime/root_atime_off.ksh \
functional/atime/root_atime_on.ksh \
functional/atime/root_relatime_on.ksh \
functional/atime/setup.ksh \
functional/bootfs/bootfs_001_pos.ksh \
functional/bootfs/bootfs_002_neg.ksh \
functional/bootfs/bootfs_003_pos.ksh \
functional/bootfs/bootfs_004_neg.ksh \
functional/bootfs/bootfs_005_neg.ksh \
functional/bootfs/bootfs_006_pos.ksh \
functional/bootfs/bootfs_007_pos.ksh \
functional/bootfs/bootfs_008_pos.ksh \
functional/bootfs/cleanup.ksh \
functional/bootfs/setup.ksh \
functional/btree/btree_negative.ksh \
functional/btree/btree_positive.ksh \
functional/cache/cache_001_pos.ksh \
functional/cache/cache_002_pos.ksh \
functional/cache/cache_003_pos.ksh \
functional/cache/cache_004_neg.ksh \
functional/cache/cache_005_neg.ksh \
functional/cache/cache_006_pos.ksh \
functional/cache/cache_007_neg.ksh \
functional/cache/cache_008_neg.ksh \
functional/cache/cache_009_pos.ksh \
functional/cache/cache_010_pos.ksh \
functional/cache/cache_011_pos.ksh \
functional/cache/cache_012_pos.ksh \
functional/cache/cleanup.ksh \
functional/cachefile/cachefile_001_pos.ksh \
functional/cachefile/cachefile_002_pos.ksh \
functional/cachefile/cachefile_003_pos.ksh \
functional/cachefile/cachefile_004_pos.ksh \
functional/cachefile/cleanup.ksh \
functional/cachefile/setup.ksh \
functional/cache/setup.ksh \
functional/casenorm/case_all_values.ksh \
functional/casenorm/cleanup.ksh \
functional/casenorm/insensitive_formd_delete.ksh \
functional/casenorm/insensitive_formd_lookup.ksh \
functional/casenorm/insensitive_none_delete.ksh \
functional/casenorm/insensitive_none_lookup.ksh \
functional/casenorm/mixed_create_failure.ksh \
functional/casenorm/mixed_formd_delete.ksh \
functional/casenorm/mixed_formd_lookup_ci.ksh \
functional/casenorm/mixed_formd_lookup.ksh \
functional/casenorm/mixed_none_delete.ksh \
functional/casenorm/mixed_none_lookup_ci.ksh \
functional/casenorm/mixed_none_lookup.ksh \
functional/casenorm/norm_all_values.ksh \
functional/casenorm/sensitive_formd_delete.ksh \
functional/casenorm/sensitive_formd_lookup.ksh \
functional/casenorm/sensitive_none_delete.ksh \
functional/casenorm/sensitive_none_lookup.ksh \
functional/casenorm/setup.ksh \
functional/channel_program/lua_core/cleanup.ksh \
functional/channel_program/lua_core/setup.ksh \
functional/channel_program/lua_core/tst.args_to_lua.ksh \
functional/channel_program/lua_core/tst.divide_by_zero.ksh \
functional/channel_program/lua_core/tst.exists.ksh \
functional/channel_program/lua_core/tst.integer_illegal.ksh \
functional/channel_program/lua_core/tst.integer_overflow.ksh \
functional/channel_program/lua_core/tst.language_functions_neg.ksh \
functional/channel_program/lua_core/tst.language_functions_pos.ksh \
functional/channel_program/lua_core/tst.large_prog.ksh \
functional/channel_program/lua_core/tst.libraries.ksh \
functional/channel_program/lua_core/tst.memory_limit.ksh \
functional/channel_program/lua_core/tst.nested_neg.ksh \
functional/channel_program/lua_core/tst.nested_pos.ksh \
functional/channel_program/lua_core/tst.nvlist_to_lua.ksh \
functional/channel_program/lua_core/tst.recursive_neg.ksh \
functional/channel_program/lua_core/tst.recursive_pos.ksh \
functional/channel_program/lua_core/tst.return_large.ksh \
functional/channel_program/lua_core/tst.return_nvlist_neg.ksh \
functional/channel_program/lua_core/tst.return_nvlist_pos.ksh \
functional/channel_program/lua_core/tst.return_recursive_table.ksh \
functional/channel_program/lua_core/tst.stack_gsub.ksh \
functional/channel_program/lua_core/tst.timeout.ksh \
functional/channel_program/synctask_core/cleanup.ksh \
functional/channel_program/synctask_core/setup.ksh \
functional/channel_program/synctask_core/tst.bookmark.copy.ksh \
functional/channel_program/synctask_core/tst.bookmark.create.ksh \
functional/channel_program/synctask_core/tst.destroy_fs.ksh \
functional/channel_program/synctask_core/tst.destroy_snap.ksh \
functional/channel_program/synctask_core/tst.get_count_and_limit.ksh \
functional/channel_program/synctask_core/tst.get_index_props.ksh \
functional/channel_program/synctask_core/tst.get_mountpoint.ksh \
functional/channel_program/synctask_core/tst.get_neg.ksh \
functional/channel_program/synctask_core/tst.get_number_props.ksh \
functional/channel_program/synctask_core/tst.get_string_props.ksh \
functional/channel_program/synctask_core/tst.get_type.ksh \
functional/channel_program/synctask_core/tst.get_userquota.ksh \
functional/channel_program/synctask_core/tst.get_written.ksh \
functional/channel_program/synctask_core/tst.inherit.ksh \
functional/channel_program/synctask_core/tst.list_bookmarks.ksh \
functional/channel_program/synctask_core/tst.list_children.ksh \
functional/channel_program/synctask_core/tst.list_clones.ksh \
functional/channel_program/synctask_core/tst.list_holds.ksh \
functional/channel_program/synctask_core/tst.list_snapshots.ksh \
functional/channel_program/synctask_core/tst.list_system_props.ksh \
functional/channel_program/synctask_core/tst.list_user_props.ksh \
functional/channel_program/synctask_core/tst.parse_args_neg.ksh \
functional/channel_program/synctask_core/tst.promote_conflict.ksh \
functional/channel_program/synctask_core/tst.promote_multiple.ksh \
functional/channel_program/synctask_core/tst.promote_simple.ksh \
functional/channel_program/synctask_core/tst.rollback_mult.ksh \
functional/channel_program/synctask_core/tst.rollback_one.ksh \
functional/channel_program/synctask_core/tst.set_props.ksh \
functional/channel_program/synctask_core/tst.snapshot_destroy.ksh \
functional/channel_program/synctask_core/tst.snapshot_neg.ksh \
functional/channel_program/synctask_core/tst.snapshot_recursive.ksh \
functional/channel_program/synctask_core/tst.snapshot_rename.ksh \
functional/channel_program/synctask_core/tst.snapshot_simple.ksh \
functional/channel_program/synctask_core/tst.terminate_by_signal.ksh \
functional/chattr/chattr_001_pos.ksh \
functional/chattr/chattr_002_neg.ksh \
functional/chattr/cleanup.ksh \
functional/chattr/setup.ksh \
functional/checksum/cleanup.ksh \
functional/checksum/filetest_001_pos.ksh \
functional/checksum/filetest_002_pos.ksh \
functional/checksum/run_blake3_test.ksh \
functional/checksum/run_edonr_test.ksh \
functional/checksum/run_sha2_test.ksh \
functional/checksum/run_skein_test.ksh \
functional/checksum/setup.ksh \
functional/clean_mirror/clean_mirror_001_pos.ksh \
functional/clean_mirror/clean_mirror_002_pos.ksh \
functional/clean_mirror/clean_mirror_003_pos.ksh \
functional/clean_mirror/clean_mirror_004_pos.ksh \
functional/clean_mirror/cleanup.ksh \
functional/clean_mirror/setup.ksh \
functional/cli_root/zdb/zdb_002_pos.ksh \
functional/cli_root/zdb/zdb_003_pos.ksh \
functional/cli_root/zdb/zdb_004_pos.ksh \
functional/cli_root/zdb/zdb_005_pos.ksh \
functional/cli_root/zdb/zdb_006_pos.ksh \
functional/cli_root/zdb/zdb_args_neg.ksh \
functional/cli_root/zdb/zdb_args_pos.ksh \
functional/cli_root/zdb/zdb_block_size_histogram.ksh \
functional/cli_root/zdb/zdb_checksum.ksh \
functional/cli_root/zdb/zdb_decompress.ksh \
functional/cli_root/zdb/zdb_decompress_zstd.ksh \
functional/cli_root/zdb/zdb_display_block.ksh \
functional/cli_root/zdb/zdb_encrypted.ksh \
functional/cli_root/zdb/zdb_label_checksum.ksh \
functional/cli_root/zdb/zdb_object_range_neg.ksh \
functional/cli_root/zdb/zdb_object_range_pos.ksh \
functional/cli_root/zdb/zdb_objset_id.ksh \
functional/cli_root/zdb/zdb_recover_2.ksh \
functional/cli_root/zdb/zdb_recover.ksh \
functional/cli_root/zfs_bookmark/cleanup.ksh \
functional/cli_root/zfs_bookmark/setup.ksh \
functional/cli_root/zfs_bookmark/zfs_bookmark_cliargs.ksh \
functional/cli_root/zfs_change-key/cleanup.ksh \
functional/cli_root/zfs_change-key/setup.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_child.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_clones.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_format.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_inherit.ksh \
functional/cli_root/zfs_change-key/zfs_change-key.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_load.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_location.ksh \
functional/cli_root/zfs_change-key/zfs_change-key_pbkdf2iters.ksh \
functional/cli_root/zfs/cleanup.ksh \
functional/cli_root/zfs_clone/cleanup.ksh \
functional/cli_root/zfs_clone/setup.ksh \
functional/cli_root/zfs_clone/zfs_clone_001_neg.ksh \
functional/cli_root/zfs_clone/zfs_clone_002_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_003_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_004_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_005_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_006_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_007_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_008_neg.ksh \
functional/cli_root/zfs_clone/zfs_clone_009_neg.ksh \
functional/cli_root/zfs_clone/zfs_clone_010_pos.ksh \
functional/cli_root/zfs_clone/zfs_clone_deeply_nested.ksh \
functional/cli_root/zfs_clone/zfs_clone_encrypted.ksh \
functional/cli_root/zfs_clone/zfs_clone_rm_nested.ksh \
functional/cli_root/zfs_copies/cleanup.ksh \
functional/cli_root/zfs_copies/setup.ksh \
functional/cli_root/zfs_copies/zfs_copies_001_pos.ksh \
functional/cli_root/zfs_copies/zfs_copies_002_pos.ksh \
functional/cli_root/zfs_copies/zfs_copies_003_pos.ksh \
functional/cli_root/zfs_copies/zfs_copies_004_neg.ksh \
functional/cli_root/zfs_copies/zfs_copies_005_neg.ksh \
functional/cli_root/zfs_copies/zfs_copies_006_pos.ksh \
functional/cli_root/zfs_create/cleanup.ksh \
functional/cli_root/zfs_create/setup.ksh \
functional/cli_root/zfs_create/zfs_create_001_pos.ksh \
functional/cli_root/zfs_create/zfs_create_002_pos.ksh \
functional/cli_root/zfs_create/zfs_create_003_pos.ksh \
functional/cli_root/zfs_create/zfs_create_004_pos.ksh \
functional/cli_root/zfs_create/zfs_create_005_pos.ksh \
functional/cli_root/zfs_create/zfs_create_006_pos.ksh \
functional/cli_root/zfs_create/zfs_create_007_pos.ksh \
functional/cli_root/zfs_create/zfs_create_008_neg.ksh \
functional/cli_root/zfs_create/zfs_create_009_neg.ksh \
functional/cli_root/zfs_create/zfs_create_010_neg.ksh \
functional/cli_root/zfs_create/zfs_create_011_pos.ksh \
functional/cli_root/zfs_create/zfs_create_012_pos.ksh \
functional/cli_root/zfs_create/zfs_create_013_pos.ksh \
functional/cli_root/zfs_create/zfs_create_014_pos.ksh \
functional/cli_root/zfs_create/zfs_create_crypt_combos.ksh \
functional/cli_root/zfs_create/zfs_create_dryrun.ksh \
functional/cli_root/zfs_create/zfs_create_encrypted.ksh \
functional/cli_root/zfs_create/zfs_create_nomount.ksh \
functional/cli_root/zfs_create/zfs_create_verbose.ksh \
functional/cli_root/zfs_destroy/cleanup.ksh \
functional/cli_root/zfs_destroy/setup.ksh \
functional/cli_root/zfs_destroy/zfs_clone_livelist_condense_and_disable.ksh \
functional/cli_root/zfs_destroy/zfs_clone_livelist_condense_races.ksh \
functional/cli_root/zfs_destroy/zfs_clone_livelist_dedup.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_001_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_002_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_003_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_004_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_005_neg.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_006_neg.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_007_neg.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_008_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_009_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_010_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_011_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_012_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_013_neg.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_014_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_015_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_016_pos.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_clone_livelist.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_dev_removal_condense.ksh \
functional/cli_root/zfs_destroy/zfs_destroy_dev_removal.ksh \
functional/cli_root/zfs_diff/cleanup.ksh \
functional/cli_root/zfs_diff/setup.ksh \
functional/cli_root/zfs_diff/zfs_diff_changes.ksh \
functional/cli_root/zfs_diff/zfs_diff_cliargs.ksh \
functional/cli_root/zfs_diff/zfs_diff_encrypted.ksh \
functional/cli_root/zfs_diff/zfs_diff_mangle.ksh \
functional/cli_root/zfs_diff/zfs_diff_timestamp.ksh \
functional/cli_root/zfs_diff/zfs_diff_types.ksh \
functional/cli_root/zfs_get/cleanup.ksh \
functional/cli_root/zfs_get/setup.ksh \
functional/cli_root/zfs_get/zfs_get_001_pos.ksh \
functional/cli_root/zfs_get/zfs_get_002_pos.ksh \
functional/cli_root/zfs_get/zfs_get_003_pos.ksh \
functional/cli_root/zfs_get/zfs_get_004_pos.ksh \
functional/cli_root/zfs_get/zfs_get_005_neg.ksh \
functional/cli_root/zfs_get/zfs_get_006_neg.ksh \
functional/cli_root/zfs_get/zfs_get_007_neg.ksh \
functional/cli_root/zfs_get/zfs_get_008_pos.ksh \
functional/cli_root/zfs_get/zfs_get_009_pos.ksh \
functional/cli_root/zfs_get/zfs_get_010_neg.ksh \
functional/cli_root/zfs_ids_to_path/cleanup.ksh \
functional/cli_root/zfs_ids_to_path/setup.ksh \
functional/cli_root/zfs_ids_to_path/zfs_ids_to_path_001_pos.ksh \
functional/cli_root/zfs_inherit/cleanup.ksh \
functional/cli_root/zfs_inherit/setup.ksh \
functional/cli_root/zfs_inherit/zfs_inherit_001_neg.ksh \
functional/cli_root/zfs_inherit/zfs_inherit_002_neg.ksh \
functional/cli_root/zfs_inherit/zfs_inherit_003_pos.ksh \
functional/cli_root/zfs_inherit/zfs_inherit_mountpoint.ksh \
functional/cli_root/zfs_jail/cleanup.ksh \
functional/cli_root/zfs_jail/setup.ksh \
functional/cli_root/zfs_jail/zfs_jail_001_pos.ksh \
functional/cli_root/zfs_load-key/cleanup.ksh \
functional/cli_root/zfs_load-key/setup.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_all.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_file.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_https.ksh \
functional/cli_root/zfs_load-key/zfs_load-key.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_location.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_noop.ksh \
functional/cli_root/zfs_load-key/zfs_load-key_recursive.ksh \
functional/cli_root/zfs_mount/cleanup.ksh \
functional/cli_root/zfs_mount/setup.ksh \
functional/cli_root/zfs_mount/zfs_mount_001_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_002_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_003_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_004_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_005_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_006_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_007_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_008_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_009_neg.ksh \
functional/cli_root/zfs_mount/zfs_mount_010_neg.ksh \
functional/cli_root/zfs_mount/zfs_mount_011_neg.ksh \
functional/cli_root/zfs_mount/zfs_mount_012_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_013_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_014_neg.ksh \
functional/cli_root/zfs_mount/zfs_mount_all_001_pos.ksh \
functional/cli_root/zfs_mount/zfs_mount_all_fail.ksh \
functional/cli_root/zfs_mount/zfs_mount_all_mountpoints.ksh \
functional/cli_root/zfs_mount/zfs_mount_encrypted.ksh \
functional/cli_root/zfs_mount/zfs_mount_remount.ksh \
functional/cli_root/zfs_mount/zfs_mount_test_race.ksh \
functional/cli_root/zfs_mount/zfs_multi_mount.ksh \
functional/cli_root/zfs_program/cleanup.ksh \
functional/cli_root/zfs_program/setup.ksh \
functional/cli_root/zfs_program/zfs_program_json.ksh \
functional/cli_root/zfs_promote/cleanup.ksh \
functional/cli_root/zfs_promote/setup.ksh \
functional/cli_root/zfs_promote/zfs_promote_001_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_002_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_003_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_004_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_005_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_006_neg.ksh \
functional/cli_root/zfs_promote/zfs_promote_007_neg.ksh \
functional/cli_root/zfs_promote/zfs_promote_008_pos.ksh \
functional/cli_root/zfs_promote/zfs_promote_encryptionroot.ksh \
functional/cli_root/zfs_property/cleanup.ksh \
functional/cli_root/zfs_property/setup.ksh \
functional/cli_root/zfs_property/zfs_written_property_001_pos.ksh \
functional/cli_root/zfs_receive/cleanup.ksh \
functional/cli_root/zfs_receive/receive-o-x_props_aliases.ksh \
functional/cli_root/zfs_receive/receive-o-x_props_override.ksh \
functional/cli_root/zfs_receive/setup.ksh \
functional/cli_root/zfs_receive/zfs_receive_001_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_002_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_003_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_004_neg.ksh \
functional/cli_root/zfs_receive/zfs_receive_005_neg.ksh \
functional/cli_root/zfs_receive/zfs_receive_006_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_007_neg.ksh \
functional/cli_root/zfs_receive/zfs_receive_008_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_009_neg.ksh \
functional/cli_root/zfs_receive/zfs_receive_010_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_011_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_012_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_013_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_014_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_015_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_016_pos.ksh \
functional/cli_root/zfs_receive/zfs_receive_-e.ksh \
functional/cli_root/zfs_receive/zfs_receive_from_encrypted.ksh \
functional/cli_root/zfs_receive/zfs_receive_from_zstd.ksh \
functional/cli_root/zfs_receive/zfs_receive_new_props.ksh \
functional/cli_root/zfs_receive/zfs_receive_raw_-d.ksh \
functional/cli_root/zfs_receive/zfs_receive_raw_incremental.ksh \
functional/cli_root/zfs_receive/zfs_receive_raw.ksh \
functional/cli_root/zfs_receive/zfs_receive_to_encrypted.ksh \
functional/cli_root/zfs_receive/zfs_receive_-wR-encrypted-mix.ksh \
functional/cli_root/zfs_receive/zfs_receive_corrective.ksh \
functional/cli_root/zfs_receive/zfs_receive_compressed_corrective.ksh \
functional/cli_root/zfs_receive/zfs_receive_large_block_corrective.ksh \
functional/cli_root/zfs_rename/cleanup.ksh \
functional/cli_root/zfs_rename/setup.ksh \
functional/cli_root/zfs_rename/zfs_rename_001_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_002_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_003_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_004_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_005_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_006_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_007_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_008_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_009_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_010_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_011_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_012_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_013_pos.ksh \
functional/cli_root/zfs_rename/zfs_rename_014_neg.ksh \
functional/cli_root/zfs_rename/zfs_rename_encrypted_child.ksh \
functional/cli_root/zfs_rename/zfs_rename_mountpoint.ksh \
functional/cli_root/zfs_rename/zfs_rename_nounmount.ksh \
functional/cli_root/zfs_rename/zfs_rename_to_encrypted.ksh \
functional/cli_root/zfs_reservation/cleanup.ksh \
functional/cli_root/zfs_reservation/setup.ksh \
functional/cli_root/zfs_reservation/zfs_reservation_001_pos.ksh \
functional/cli_root/zfs_reservation/zfs_reservation_002_pos.ksh \
functional/cli_root/zfs_rollback/cleanup.ksh \
functional/cli_root/zfs_rollback/setup.ksh \
functional/cli_root/zfs_rollback/zfs_rollback_001_pos.ksh \
functional/cli_root/zfs_rollback/zfs_rollback_002_pos.ksh \
functional/cli_root/zfs_rollback/zfs_rollback_003_neg.ksh \
functional/cli_root/zfs_rollback/zfs_rollback_004_neg.ksh \
functional/cli_root/zfs_send/cleanup.ksh \
functional/cli_root/zfs_send/setup.ksh \
functional/cli_root/zfs_send/zfs_send_001_pos.ksh \
functional/cli_root/zfs_send/zfs_send_002_pos.ksh \
functional/cli_root/zfs_send/zfs_send_003_pos.ksh \
functional/cli_root/zfs_send/zfs_send_004_neg.ksh \
functional/cli_root/zfs_send/zfs_send_005_pos.ksh \
functional/cli_root/zfs_send/zfs_send_006_pos.ksh \
functional/cli_root/zfs_send/zfs_send_007_pos.ksh \
functional/cli_root/zfs_send/zfs_send-b.ksh \
functional/cli_root/zfs_send/zfs_send_encrypted.ksh \
functional/cli_root/zfs_send/zfs_send_encrypted_unloaded.ksh \
functional/cli_root/zfs_send/zfs_send_raw.ksh \
functional/cli_root/zfs_send/zfs_send_skip_missing.ksh \
functional/cli_root/zfs_send/zfs_send_sparse.ksh \
functional/cli_root/zfs_set/cache_001_pos.ksh \
functional/cli_root/zfs_set/cache_002_neg.ksh \
functional/cli_root/zfs_set/canmount_001_pos.ksh \
functional/cli_root/zfs_set/canmount_002_pos.ksh \
functional/cli_root/zfs_set/canmount_003_pos.ksh \
functional/cli_root/zfs_set/canmount_004_pos.ksh \
functional/cli_root/zfs_set/checksum_001_pos.ksh \
functional/cli_root/zfs_set/cleanup.ksh \
functional/cli_root/zfs_set/compression_001_pos.ksh \
functional/cli_root/zfs_set/mountpoint_001_pos.ksh \
functional/cli_root/zfs_set/mountpoint_002_pos.ksh \
functional/cli_root/zfs_set/mountpoint_003_pos.ksh \
functional/cli_root/zfs_set/onoffs_001_pos.ksh \
functional/cli_root/zfs_set/property_alias_001_pos.ksh \
functional/cli_root/zfs_set/readonly_001_pos.ksh \
functional/cli_root/zfs_set/reservation_001_neg.ksh \
functional/cli_root/zfs_set/ro_props_001_pos.ksh \
functional/cli_root/zfs_set/setup.ksh \
functional/cli_root/zfs_set/share_mount_001_neg.ksh \
functional/cli_root/zfs_set/snapdir_001_pos.ksh \
functional/cli_root/zfs/setup.ksh \
functional/cli_root/zfs_set/user_property_001_pos.ksh \
functional/cli_root/zfs_set/user_property_002_pos.ksh \
functional/cli_root/zfs_set/user_property_003_neg.ksh \
functional/cli_root/zfs_set/user_property_004_pos.ksh \
functional/cli_root/zfs_set/version_001_neg.ksh \
functional/cli_root/zfs_set/zfs_set_001_neg.ksh \
functional/cli_root/zfs_set/zfs_set_002_neg.ksh \
functional/cli_root/zfs_set/zfs_set_003_neg.ksh \
functional/cli_root/zfs_set/zfs_set_feature_activation.ksh \
functional/cli_root/zfs_set/zfs_set_keylocation.ksh \
functional/cli_root/zfs_share/cleanup.ksh \
functional/cli_root/zfs_share/setup.ksh \
functional/cli_root/zfs_share/zfs_share_001_pos.ksh \
functional/cli_root/zfs_share/zfs_share_002_pos.ksh \
functional/cli_root/zfs_share/zfs_share_003_pos.ksh \
functional/cli_root/zfs_share/zfs_share_004_pos.ksh \
functional/cli_root/zfs_share/zfs_share_005_pos.ksh \
functional/cli_root/zfs_share/zfs_share_006_pos.ksh \
functional/cli_root/zfs_share/zfs_share_007_neg.ksh \
functional/cli_root/zfs_share/zfs_share_008_neg.ksh \
functional/cli_root/zfs_share/zfs_share_009_neg.ksh \
functional/cli_root/zfs_share/zfs_share_010_neg.ksh \
functional/cli_root/zfs_share/zfs_share_011_pos.ksh \
functional/cli_root/zfs_share/zfs_share_012_pos.ksh \
functional/cli_root/zfs_share/zfs_share_013_pos.ksh \
functional/cli_root/zfs_share/zfs_share_concurrent_shares.ksh \
functional/cli_root/zfs_snapshot/cleanup.ksh \
functional/cli_root/zfs_snapshot/setup.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_001_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_002_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_003_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_004_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_005_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_006_pos.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_007_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_008_neg.ksh \
functional/cli_root/zfs_snapshot/zfs_snapshot_009_pos.ksh \
functional/cli_root/zfs_sysfs/cleanup.ksh \
functional/cli_root/zfs_sysfs/setup.ksh \
functional/cli_root/zfs_sysfs/zfeature_set_unsupported.ksh \
functional/cli_root/zfs_sysfs/zfs_get_unsupported.ksh \
functional/cli_root/zfs_sysfs/zfs_set_unsupported.ksh \
functional/cli_root/zfs_sysfs/zfs_sysfs_live.ksh \
functional/cli_root/zfs_sysfs/zpool_get_unsupported.ksh \
functional/cli_root/zfs_sysfs/zpool_set_unsupported.ksh \
functional/cli_root/zfs_unload-key/cleanup.ksh \
functional/cli_root/zfs_unload-key/setup.ksh \
functional/cli_root/zfs_unload-key/zfs_unload-key_all.ksh \
functional/cli_root/zfs_unload-key/zfs_unload-key.ksh \
functional/cli_root/zfs_unload-key/zfs_unload-key_recursive.ksh \
functional/cli_root/zfs_unmount/cleanup.ksh \
functional/cli_root/zfs_unmount/setup.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_001_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_002_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_003_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_004_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_005_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_006_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_007_neg.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_008_neg.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_009_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_all_001_pos.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_nested.ksh \
functional/cli_root/zfs_unmount/zfs_unmount_unload_keys.ksh \
functional/cli_root/zfs_unshare/cleanup.ksh \
functional/cli_root/zfs_unshare/setup.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_001_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_002_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_003_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_004_neg.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_005_neg.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_006_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_007_pos.ksh \
functional/cli_root/zfs_unshare/zfs_unshare_008_pos.ksh \
functional/cli_root/zfs_upgrade/cleanup.ksh \
functional/cli_root/zfs_upgrade/setup.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_001_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_002_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_003_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_004_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_005_pos.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_006_neg.ksh \
functional/cli_root/zfs_upgrade/zfs_upgrade_007_neg.ksh \
functional/cli_root/zfs_wait/cleanup.ksh \
functional/cli_root/zfs_wait/setup.ksh \
functional/cli_root/zfs_wait/zfs_wait_deleteq.ksh \
functional/cli_root/zfs_wait/zfs_wait_getsubopt.ksh \
functional/cli_root/zfs/zfs_001_neg.ksh \
functional/cli_root/zfs/zfs_002_pos.ksh \
functional/cli_root/zfs/zfs_003_neg.ksh \
functional/cli_root/zhack/zhack_label_checksum.ksh \
functional/cli_root/zpool_add/add_nested_replacing_spare.ksh \
functional/cli_root/zpool_add/add-o_ashift.ksh \
functional/cli_root/zpool_add/add_prop_ashift.ksh \
functional/cli_root/zpool_add/cleanup.ksh \
functional/cli_root/zpool_add/setup.ksh \
functional/cli_root/zpool_add/zpool_add_001_pos.ksh \
functional/cli_root/zpool_add/zpool_add_002_pos.ksh \
functional/cli_root/zpool_add/zpool_add_003_pos.ksh \
functional/cli_root/zpool_add/zpool_add_004_pos.ksh \
functional/cli_root/zpool_add/zpool_add_005_pos.ksh \
functional/cli_root/zpool_add/zpool_add_006_pos.ksh \
functional/cli_root/zpool_add/zpool_add_007_neg.ksh \
functional/cli_root/zpool_add/zpool_add_008_neg.ksh \
functional/cli_root/zpool_add/zpool_add_009_neg.ksh \
functional/cli_root/zpool_add/zpool_add_010_pos.ksh \
functional/cli_root/zpool_add/zpool_add_dryrun_output.ksh \
functional/cli_root/zpool_attach/attach-o_ashift.ksh \
functional/cli_root/zpool_attach/cleanup.ksh \
functional/cli_root/zpool_attach/setup.ksh \
functional/cli_root/zpool_attach/zpool_attach_001_neg.ksh \
functional/cli_root/zpool/cleanup.ksh \
functional/cli_root/zpool_clear/cleanup.ksh \
functional/cli_root/zpool_clear/setup.ksh \
functional/cli_root/zpool_clear/zpool_clear_001_pos.ksh \
functional/cli_root/zpool_clear/zpool_clear_002_neg.ksh \
functional/cli_root/zpool_clear/zpool_clear_003_neg.ksh \
functional/cli_root/zpool_clear/zpool_clear_readonly.ksh \
functional/cli_root/zpool_create/cleanup.ksh \
functional/cli_root/zpool_create/create-o_ashift.ksh \
functional/cli_root/zpool_create/setup.ksh \
functional/cli_root/zpool_create/zpool_create_001_pos.ksh \
functional/cli_root/zpool_create/zpool_create_002_pos.ksh \
functional/cli_root/zpool_create/zpool_create_003_pos.ksh \
functional/cli_root/zpool_create/zpool_create_004_pos.ksh \
functional/cli_root/zpool_create/zpool_create_005_pos.ksh \
functional/cli_root/zpool_create/zpool_create_006_pos.ksh \
functional/cli_root/zpool_create/zpool_create_007_neg.ksh \
functional/cli_root/zpool_create/zpool_create_008_pos.ksh \
functional/cli_root/zpool_create/zpool_create_009_neg.ksh \
functional/cli_root/zpool_create/zpool_create_010_neg.ksh \
functional/cli_root/zpool_create/zpool_create_011_neg.ksh \
functional/cli_root/zpool_create/zpool_create_012_neg.ksh \
functional/cli_root/zpool_create/zpool_create_014_neg.ksh \
functional/cli_root/zpool_create/zpool_create_015_neg.ksh \
functional/cli_root/zpool_create/zpool_create_016_pos.ksh \
functional/cli_root/zpool_create/zpool_create_017_neg.ksh \
functional/cli_root/zpool_create/zpool_create_018_pos.ksh \
functional/cli_root/zpool_create/zpool_create_019_pos.ksh \
functional/cli_root/zpool_create/zpool_create_020_pos.ksh \
functional/cli_root/zpool_create/zpool_create_021_pos.ksh \
functional/cli_root/zpool_create/zpool_create_022_pos.ksh \
functional/cli_root/zpool_create/zpool_create_023_neg.ksh \
functional/cli_root/zpool_create/zpool_create_024_pos.ksh \
functional/cli_root/zpool_create/zpool_create_crypt_combos.ksh \
functional/cli_root/zpool_create/zpool_create_draid_001_pos.ksh \
functional/cli_root/zpool_create/zpool_create_draid_002_pos.ksh \
functional/cli_root/zpool_create/zpool_create_draid_003_pos.ksh \
functional/cli_root/zpool_create/zpool_create_draid_004_pos.ksh \
functional/cli_root/zpool_create/zpool_create_dryrun_output.ksh \
functional/cli_root/zpool_create/zpool_create_encrypted.ksh \
functional/cli_root/zpool_create/zpool_create_features_001_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_002_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_003_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_004_neg.ksh \
functional/cli_root/zpool_create/zpool_create_features_005_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_006_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_007_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_008_pos.ksh \
functional/cli_root/zpool_create/zpool_create_features_009_pos.ksh \
functional/cli_root/zpool_create/zpool_create_tempname.ksh \
functional/cli_root/zpool_destroy/zpool_destroy_001_pos.ksh \
functional/cli_root/zpool_destroy/zpool_destroy_002_pos.ksh \
functional/cli_root/zpool_destroy/zpool_destroy_003_neg.ksh \
functional/cli_root/zpool_detach/cleanup.ksh \
functional/cli_root/zpool_detach/setup.ksh \
functional/cli_root/zpool_detach/zpool_detach_001_neg.ksh \
functional/cli_root/zpool_events/cleanup.ksh \
functional/cli_root/zpool_events/setup.ksh \
functional/cli_root/zpool_events/zpool_events_clear.ksh \
functional/cli_root/zpool_events/zpool_events_clear_retained.ksh \
functional/cli_root/zpool_events/zpool_events_cliargs.ksh \
functional/cli_root/zpool_events/zpool_events_duplicates.ksh \
functional/cli_root/zpool_events/zpool_events_errors.ksh \
functional/cli_root/zpool_events/zpool_events_follow.ksh \
functional/cli_root/zpool_events/zpool_events_poolname.ksh \
functional/cli_root/zpool_expand/cleanup.ksh \
functional/cli_root/zpool_expand/setup.ksh \
functional/cli_root/zpool_expand/zpool_expand_001_pos.ksh \
functional/cli_root/zpool_expand/zpool_expand_002_pos.ksh \
functional/cli_root/zpool_expand/zpool_expand_003_neg.ksh \
functional/cli_root/zpool_expand/zpool_expand_004_pos.ksh \
functional/cli_root/zpool_expand/zpool_expand_005_pos.ksh \
functional/cli_root/zpool_export/cleanup.ksh \
functional/cli_root/zpool_export/setup.ksh \
functional/cli_root/zpool_export/zpool_export_001_pos.ksh \
functional/cli_root/zpool_export/zpool_export_002_pos.ksh \
functional/cli_root/zpool_export/zpool_export_003_neg.ksh \
functional/cli_root/zpool_export/zpool_export_004_pos.ksh \
functional/cli_root/zpool_get/cleanup.ksh \
functional/cli_root/zpool_get/setup.ksh \
+ functional/cli_root/zpool_get/vdev_get_001_pos.ksh \
functional/cli_root/zpool_get/zpool_get_001_pos.ksh \
functional/cli_root/zpool_get/zpool_get_002_pos.ksh \
functional/cli_root/zpool_get/zpool_get_003_pos.ksh \
functional/cli_root/zpool_get/zpool_get_004_neg.ksh \
functional/cli_root/zpool_get/zpool_get_005_pos.ksh \
functional/cli_root/zpool_history/cleanup.ksh \
functional/cli_root/zpool_history/setup.ksh \
functional/cli_root/zpool_history/zpool_history_001_neg.ksh \
functional/cli_root/zpool_history/zpool_history_002_pos.ksh \
functional/cli_root/zpool_import/cleanup.ksh \
functional/cli_root/zpool_import/import_cachefile_device_added.ksh \
functional/cli_root/zpool_import/import_cachefile_device_removed.ksh \
functional/cli_root/zpool_import/import_cachefile_device_replaced.ksh \
functional/cli_root/zpool_import/import_cachefile_mirror_attached.ksh \
functional/cli_root/zpool_import/import_cachefile_mirror_detached.ksh \
functional/cli_root/zpool_import/import_cachefile_paths_changed.ksh \
functional/cli_root/zpool_import/import_cachefile_shared_device.ksh \
functional/cli_root/zpool_import/import_devices_missing.ksh \
functional/cli_root/zpool_import/import_paths_changed.ksh \
functional/cli_root/zpool_import/import_rewind_config_changed.ksh \
functional/cli_root/zpool_import/import_rewind_device_replaced.ksh \
functional/cli_root/zpool_import/setup.ksh \
functional/cli_root/zpool_import/zpool_import_001_pos.ksh \
functional/cli_root/zpool_import/zpool_import_002_pos.ksh \
functional/cli_root/zpool_import/zpool_import_003_pos.ksh \
functional/cli_root/zpool_import/zpool_import_004_pos.ksh \
functional/cli_root/zpool_import/zpool_import_005_pos.ksh \
functional/cli_root/zpool_import/zpool_import_006_pos.ksh \
functional/cli_root/zpool_import/zpool_import_007_pos.ksh \
functional/cli_root/zpool_import/zpool_import_008_pos.ksh \
functional/cli_root/zpool_import/zpool_import_009_neg.ksh \
functional/cli_root/zpool_import/zpool_import_010_pos.ksh \
functional/cli_root/zpool_import/zpool_import_011_neg.ksh \
functional/cli_root/zpool_import/zpool_import_012_pos.ksh \
functional/cli_root/zpool_import/zpool_import_013_neg.ksh \
functional/cli_root/zpool_import/zpool_import_014_pos.ksh \
functional/cli_root/zpool_import/zpool_import_015_pos.ksh \
functional/cli_root/zpool_import/zpool_import_016_pos.ksh \
functional/cli_root/zpool_import/zpool_import_017_pos.ksh \
functional/cli_root/zpool_import/zpool_import_all_001_pos.ksh \
functional/cli_root/zpool_import/zpool_import_encrypted.ksh \
functional/cli_root/zpool_import/zpool_import_encrypted_load.ksh \
functional/cli_root/zpool_import/zpool_import_errata3.ksh \
functional/cli_root/zpool_import/zpool_import_errata4.ksh \
functional/cli_root/zpool_import/zpool_import_features_001_pos.ksh \
functional/cli_root/zpool_import/zpool_import_features_002_neg.ksh \
functional/cli_root/zpool_import/zpool_import_features_003_pos.ksh \
functional/cli_root/zpool_import/zpool_import_missing_001_pos.ksh \
functional/cli_root/zpool_import/zpool_import_missing_002_pos.ksh \
functional/cli_root/zpool_import/zpool_import_missing_003_pos.ksh \
functional/cli_root/zpool_import/zpool_import_rename_001_pos.ksh \
functional/cli_root/zpool_initialize/cleanup.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_attach_detach_add_remove.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_fault_export_import_online.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_import_export.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_offline_export_import_online.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_online_offline.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_split.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_start_and_cancel_neg.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_start_and_cancel_pos.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_suspend_resume.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_unsupported_vdevs.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_verify_checksums.ksh \
functional/cli_root/zpool_initialize/zpool_initialize_verify_initialized.ksh \
functional/cli_root/zpool_labelclear/zpool_labelclear_active.ksh \
functional/cli_root/zpool_labelclear/zpool_labelclear_exported.ksh \
functional/cli_root/zpool_labelclear/zpool_labelclear_removed.ksh \
functional/cli_root/zpool_labelclear/zpool_labelclear_valid.ksh \
functional/cli_root/zpool_offline/cleanup.ksh \
functional/cli_root/zpool_offline/setup.ksh \
functional/cli_root/zpool_offline/zpool_offline_001_pos.ksh \
functional/cli_root/zpool_offline/zpool_offline_002_neg.ksh \
functional/cli_root/zpool_offline/zpool_offline_003_pos.ksh \
functional/cli_root/zpool_online/cleanup.ksh \
functional/cli_root/zpool_online/setup.ksh \
functional/cli_root/zpool_online/zpool_online_001_pos.ksh \
functional/cli_root/zpool_online/zpool_online_002_neg.ksh \
functional/cli_root/zpool_remove/cleanup.ksh \
functional/cli_root/zpool_remove/setup.ksh \
functional/cli_root/zpool_remove/zpool_remove_001_neg.ksh \
functional/cli_root/zpool_remove/zpool_remove_002_pos.ksh \
functional/cli_root/zpool_remove/zpool_remove_003_pos.ksh \
functional/cli_root/zpool_reopen/cleanup.ksh \
functional/cli_root/zpool_reopen/setup.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_001_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_002_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_003_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_004_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_005_pos.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_006_neg.ksh \
functional/cli_root/zpool_reopen/zpool_reopen_007_pos.ksh \
functional/cli_root/zpool_replace/cleanup.ksh \
functional/cli_root/zpool_replace/replace-o_ashift.ksh \
functional/cli_root/zpool_replace/replace_prop_ashift.ksh \
functional/cli_root/zpool_replace/setup.ksh \
functional/cli_root/zpool_replace/zpool_replace_001_neg.ksh \
functional/cli_root/zpool_resilver/cleanup.ksh \
functional/cli_root/zpool_resilver/setup.ksh \
functional/cli_root/zpool_resilver/zpool_resilver_bad_args.ksh \
functional/cli_root/zpool_resilver/zpool_resilver_restart.ksh \
functional/cli_root/zpool_scrub/cleanup.ksh \
functional/cli_root/zpool_scrub/setup.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_001_neg.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_002_pos.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_003_pos.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_004_pos.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_005_pos.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_encrypted_unloaded.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_multiple_copies.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_offline_device.ksh \
functional/cli_root/zpool_scrub/zpool_scrub_print_repairing.ksh \
functional/cli_root/zpool_set/cleanup.ksh \
functional/cli_root/zpool_set/setup.ksh \
functional/cli_root/zpool/setup.ksh \
+ functional/cli_root/zpool_set/vdev_set_001_pos.ksh \
functional/cli_root/zpool_set/zpool_set_common.kshlib \
functional/cli_root/zpool_set/zpool_set_001_pos.ksh \
functional/cli_root/zpool_set/zpool_set_002_neg.ksh \
functional/cli_root/zpool_set/zpool_set_003_neg.ksh \
functional/cli_root/zpool_set/zpool_set_ashift.ksh \
functional/cli_root/zpool_set/user_property_001_pos.ksh \
functional/cli_root/zpool_set/user_property_002_neg.ksh \
functional/cli_root/zpool_set/zpool_set_features.ksh \
functional/cli_root/zpool_split/cleanup.ksh \
functional/cli_root/zpool_split/setup.ksh \
functional/cli_root/zpool_split/zpool_split_cliargs.ksh \
functional/cli_root/zpool_split/zpool_split_devices.ksh \
functional/cli_root/zpool_split/zpool_split_dryrun_output.ksh \
functional/cli_root/zpool_split/zpool_split_encryption.ksh \
functional/cli_root/zpool_split/zpool_split_indirect.ksh \
functional/cli_root/zpool_split/zpool_split_props.ksh \
functional/cli_root/zpool_split/zpool_split_resilver.ksh \
functional/cli_root/zpool_split/zpool_split_vdevs.ksh \
functional/cli_root/zpool_split/zpool_split_wholedisk.ksh \
functional/cli_root/zpool_status/cleanup.ksh \
functional/cli_root/zpool_status/setup.ksh \
functional/cli_root/zpool_status/zpool_status_001_pos.ksh \
functional/cli_root/zpool_status/zpool_status_002_pos.ksh \
functional/cli_root/zpool_status/zpool_status_003_pos.ksh \
functional/cli_root/zpool_status/zpool_status_004_pos.ksh \
functional/cli_root/zpool_status/zpool_status_005_pos.ksh \
functional/cli_root/zpool_status/zpool_status_006_pos.ksh \
functional/cli_root/zpool_status/zpool_status_007_pos.ksh \
functional/cli_root/zpool_status/zpool_status_features_001_pos.ksh \
functional/cli_root/zpool_sync/cleanup.ksh \
functional/cli_root/zpool_sync/setup.ksh \
functional/cli_root/zpool_sync/zpool_sync_001_pos.ksh \
functional/cli_root/zpool_sync/zpool_sync_002_neg.ksh \
functional/cli_root/zpool_trim/cleanup.ksh \
functional/cli_root/zpool_trim/setup.ksh \
functional/cli_root/zpool_trim/zpool_trim_attach_detach_add_remove.ksh \
functional/cli_root/zpool_trim/zpool_trim_fault_export_import_online.ksh \
functional/cli_root/zpool_trim/zpool_trim_import_export.ksh \
functional/cli_root/zpool_trim/zpool_trim_multiple.ksh \
functional/cli_root/zpool_trim/zpool_trim_neg.ksh \
functional/cli_root/zpool_trim/zpool_trim_offline_export_import_online.ksh \
functional/cli_root/zpool_trim/zpool_trim_online_offline.ksh \
functional/cli_root/zpool_trim/zpool_trim_partial.ksh \
functional/cli_root/zpool_trim/zpool_trim_rate.ksh \
functional/cli_root/zpool_trim/zpool_trim_rate_neg.ksh \
functional/cli_root/zpool_trim/zpool_trim_secure.ksh \
functional/cli_root/zpool_trim/zpool_trim_split.ksh \
functional/cli_root/zpool_trim/zpool_trim_start_and_cancel_neg.ksh \
functional/cli_root/zpool_trim/zpool_trim_start_and_cancel_pos.ksh \
functional/cli_root/zpool_trim/zpool_trim_suspend_resume.ksh \
functional/cli_root/zpool_trim/zpool_trim_unsupported_vdevs.ksh \
functional/cli_root/zpool_trim/zpool_trim_verify_checksums.ksh \
functional/cli_root/zpool_trim/zpool_trim_verify_trimmed.ksh \
functional/cli_root/zpool_upgrade/cleanup.ksh \
functional/cli_root/zpool_upgrade/setup.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_001_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_002_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_003_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_004_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_005_neg.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_006_neg.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_007_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_008_pos.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_009_neg.ksh \
functional/cli_root/zpool_upgrade/zpool_upgrade_features_001_pos.ksh \
functional/cli_root/zpool_wait/cleanup.ksh \
functional/cli_root/zpool_wait/scan/cleanup.ksh \
functional/cli_root/zpool_wait/scan/setup.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_rebuild.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_replace_cancel.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_replace.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_resilver.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_scrub_basic.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_scrub_cancel.ksh \
functional/cli_root/zpool_wait/scan/zpool_wait_scrub_flag.ksh \
functional/cli_root/zpool_wait/setup.ksh \
functional/cli_root/zpool_wait/zpool_wait_discard.ksh \
functional/cli_root/zpool_wait/zpool_wait_freeing.ksh \
functional/cli_root/zpool_wait/zpool_wait_initialize_basic.ksh \
functional/cli_root/zpool_wait/zpool_wait_initialize_cancel.ksh \
functional/cli_root/zpool_wait/zpool_wait_initialize_flag.ksh \
functional/cli_root/zpool_wait/zpool_wait_multiple.ksh \
functional/cli_root/zpool_wait/zpool_wait_no_activity.ksh \
functional/cli_root/zpool_wait/zpool_wait_remove_cancel.ksh \
functional/cli_root/zpool_wait/zpool_wait_remove.ksh \
functional/cli_root/zpool_wait/zpool_wait_trim_basic.ksh \
functional/cli_root/zpool_wait/zpool_wait_trim_cancel.ksh \
functional/cli_root/zpool_wait/zpool_wait_trim_flag.ksh \
functional/cli_root/zpool_wait/zpool_wait_usage.ksh \
functional/cli_root/zpool/zpool_001_neg.ksh \
functional/cli_root/zpool/zpool_002_pos.ksh \
functional/cli_root/zpool/zpool_003_pos.ksh \
functional/cli_root/zpool/zpool_colors.ksh \
functional/cli_user/misc/arcstat_001_pos.ksh \
functional/cli_user/misc/arc_summary_001_pos.ksh \
functional/cli_user/misc/arc_summary_002_neg.ksh \
functional/cli_user/misc/zilstat_001_pos.ksh \
functional/cli_user/misc/cleanup.ksh \
functional/cli_user/misc/setup.ksh \
functional/cli_user/misc/zdb_001_neg.ksh \
functional/cli_user/misc/zfs_001_neg.ksh \
functional/cli_user/misc/zfs_allow_001_neg.ksh \
functional/cli_user/misc/zfs_clone_001_neg.ksh \
functional/cli_user/misc/zfs_create_001_neg.ksh \
functional/cli_user/misc/zfs_destroy_001_neg.ksh \
functional/cli_user/misc/zfs_get_001_neg.ksh \
functional/cli_user/misc/zfs_inherit_001_neg.ksh \
functional/cli_user/misc/zfs_mount_001_neg.ksh \
functional/cli_user/misc/zfs_promote_001_neg.ksh \
functional/cli_user/misc/zfs_receive_001_neg.ksh \
functional/cli_user/misc/zfs_rename_001_neg.ksh \
functional/cli_user/misc/zfs_rollback_001_neg.ksh \
functional/cli_user/misc/zfs_send_001_neg.ksh \
functional/cli_user/misc/zfs_set_001_neg.ksh \
functional/cli_user/misc/zfs_share_001_neg.ksh \
functional/cli_user/misc/zfs_snapshot_001_neg.ksh \
functional/cli_user/misc/zfs_unallow_001_neg.ksh \
functional/cli_user/misc/zfs_unmount_001_neg.ksh \
functional/cli_user/misc/zfs_unshare_001_neg.ksh \
functional/cli_user/misc/zfs_upgrade_001_neg.ksh \
functional/cli_user/misc/zpool_001_neg.ksh \
functional/cli_user/misc/zpool_add_001_neg.ksh \
functional/cli_user/misc/zpool_attach_001_neg.ksh \
functional/cli_user/misc/zpool_clear_001_neg.ksh \
functional/cli_user/misc/zpool_create_001_neg.ksh \
functional/cli_user/misc/zpool_destroy_001_neg.ksh \
functional/cli_user/misc/zpool_detach_001_neg.ksh \
functional/cli_user/misc/zpool_export_001_neg.ksh \
functional/cli_user/misc/zpool_get_001_neg.ksh \
functional/cli_user/misc/zpool_history_001_neg.ksh \
functional/cli_user/misc/zpool_import_001_neg.ksh \
functional/cli_user/misc/zpool_import_002_neg.ksh \
functional/cli_user/misc/zpool_offline_001_neg.ksh \
functional/cli_user/misc/zpool_online_001_neg.ksh \
functional/cli_user/misc/zpool_remove_001_neg.ksh \
functional/cli_user/misc/zpool_replace_001_neg.ksh \
functional/cli_user/misc/zpool_scrub_001_neg.ksh \
functional/cli_user/misc/zpool_set_001_neg.ksh \
functional/cli_user/misc/zpool_status_001_neg.ksh \
functional/cli_user/misc/zpool_upgrade_001_neg.ksh \
functional/cli_user/misc/zpool_wait_privilege.ksh \
functional/cli_user/zfs_list/cleanup.ksh \
functional/cli_user/zfs_list/setup.ksh \
functional/cli_user/zfs_list/zfs_list_001_pos.ksh \
functional/cli_user/zfs_list/zfs_list_002_pos.ksh \
functional/cli_user/zfs_list/zfs_list_003_pos.ksh \
functional/cli_user/zfs_list/zfs_list_004_neg.ksh \
functional/cli_user/zfs_list/zfs_list_005_neg.ksh \
functional/cli_user/zfs_list/zfs_list_007_pos.ksh \
functional/cli_user/zfs_list/zfs_list_008_neg.ksh \
functional/cli_user/zpool_iostat/cleanup.ksh \
functional/cli_user/zpool_iostat/setup.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_001_neg.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_002_pos.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_003_neg.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_004_pos.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_005_pos.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_-c_disable.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_-c_homedir.ksh \
functional/cli_user/zpool_iostat/zpool_iostat_-c_searchpath.ksh \
functional/cli_user/zpool_list/cleanup.ksh \
functional/cli_user/zpool_list/setup.ksh \
functional/cli_user/zpool_list/zpool_list_001_pos.ksh \
functional/cli_user/zpool_list/zpool_list_002_neg.ksh \
functional/cli_user/zpool_status/cleanup.ksh \
functional/cli_user/zpool_status/setup.ksh \
functional/cli_user/zpool_status/zpool_status_003_pos.ksh \
functional/cli_user/zpool_status/zpool_status_-c_disable.ksh \
functional/cli_user/zpool_status/zpool_status_-c_homedir.ksh \
functional/cli_user/zpool_status/zpool_status_-c_searchpath.ksh \
functional/compression/cleanup.ksh \
functional/compression/compress_001_pos.ksh \
functional/compression/compress_002_pos.ksh \
functional/compression/compress_003_pos.ksh \
functional/compression/compress_004_pos.ksh \
functional/compression/compress_zstd_bswap.ksh \
functional/compression/l2arc_compressed_arc_disabled.ksh \
functional/compression/l2arc_compressed_arc.ksh \
functional/compression/l2arc_encrypted.ksh \
functional/compression/l2arc_encrypted_no_compressed_arc.ksh \
functional/compression/setup.ksh \
functional/cp_files/cleanup.ksh \
functional/cp_files/cp_files_001_pos.ksh \
functional/cp_files/setup.ksh \
functional/crtime/cleanup.ksh \
functional/crtime/crtime_001_pos.ksh \
functional/crtime/setup.ksh \
functional/ctime/cleanup.ksh \
functional/ctime/ctime_001_pos.ksh \
functional/ctime/setup.ksh \
functional/deadman/deadman_ratelimit.ksh \
functional/deadman/deadman_sync.ksh \
functional/deadman/deadman_zio.ksh \
functional/delegate/cleanup.ksh \
functional/delegate/setup.ksh \
functional/delegate/zfs_allow_001_pos.ksh \
functional/delegate/zfs_allow_002_pos.ksh \
functional/delegate/zfs_allow_003_pos.ksh \
functional/delegate/zfs_allow_004_pos.ksh \
functional/delegate/zfs_allow_005_pos.ksh \
functional/delegate/zfs_allow_006_pos.ksh \
functional/delegate/zfs_allow_007_pos.ksh \
functional/delegate/zfs_allow_008_pos.ksh \
functional/delegate/zfs_allow_009_neg.ksh \
functional/delegate/zfs_allow_010_pos.ksh \
functional/delegate/zfs_allow_011_neg.ksh \
functional/delegate/zfs_allow_012_neg.ksh \
functional/delegate/zfs_unallow_001_pos.ksh \
functional/delegate/zfs_unallow_002_pos.ksh \
functional/delegate/zfs_unallow_003_pos.ksh \
functional/delegate/zfs_unallow_004_pos.ksh \
functional/delegate/zfs_unallow_005_pos.ksh \
functional/delegate/zfs_unallow_006_pos.ksh \
functional/delegate/zfs_unallow_007_neg.ksh \
functional/delegate/zfs_unallow_008_neg.ksh \
functional/devices/cleanup.ksh \
functional/devices/devices_001_pos.ksh \
functional/devices/devices_002_neg.ksh \
functional/devices/devices_003_pos.ksh \
functional/devices/setup.ksh \
functional/dos_attributes/cleanup.ksh \
functional/dos_attributes/read_dos_attrs_001.ksh \
functional/dos_attributes/setup.ksh \
functional/dos_attributes/write_dos_attrs_001.ksh \
functional/events/cleanup.ksh \
functional/events/events_001_pos.ksh \
functional/events/events_002_pos.ksh \
functional/events/setup.ksh \
functional/events/zed_cksum_config.ksh \
functional/events/zed_cksum_reported.ksh \
functional/events/zed_fd_spill.ksh \
functional/events/zed_io_config.ksh \
functional/events/zed_rc_filter.ksh \
functional/exec/cleanup.ksh \
functional/exec/exec_001_pos.ksh \
functional/exec/exec_002_neg.ksh \
functional/exec/setup.ksh \
functional/fadvise/cleanup.ksh \
functional/fadvise/fadvise_sequential.ksh \
functional/fadvise/setup.ksh \
functional/fallocate/cleanup.ksh \
functional/fallocate/fallocate_prealloc.ksh \
functional/fallocate/fallocate_punch-hole.ksh \
functional/fallocate/fallocate_zero-range.ksh \
functional/fallocate/setup.ksh \
functional/fault/auto_offline_001_pos.ksh \
functional/fault/auto_online_001_pos.ksh \
functional/fault/auto_online_002_pos.ksh \
functional/fault/auto_replace_001_pos.ksh \
functional/fault/auto_spare_001_pos.ksh \
functional/fault/auto_spare_002_pos.ksh \
functional/fault/auto_spare_ashift.ksh \
functional/fault/auto_spare_multiple.ksh \
functional/fault/auto_spare_shared.ksh \
functional/fault/cleanup.ksh \
functional/fault/decompress_fault.ksh \
functional/fault/decrypt_fault.ksh \
functional/fault/scrub_after_resilver.ksh \
functional/fault/setup.ksh \
functional/fault/zpool_status_-s.ksh \
functional/features/async_destroy/async_destroy_001_pos.ksh \
functional/features/async_destroy/cleanup.ksh \
functional/features/async_destroy/setup.ksh \
functional/features/large_dnode/cleanup.ksh \
functional/features/large_dnode/large_dnode_001_pos.ksh \
functional/features/large_dnode/large_dnode_002_pos.ksh \
functional/features/large_dnode/large_dnode_003_pos.ksh \
functional/features/large_dnode/large_dnode_004_neg.ksh \
functional/features/large_dnode/large_dnode_005_pos.ksh \
functional/features/large_dnode/large_dnode_006_pos.ksh \
functional/features/large_dnode/large_dnode_007_neg.ksh \
functional/features/large_dnode/large_dnode_008_pos.ksh \
functional/features/large_dnode/large_dnode_009_pos.ksh \
functional/features/large_dnode/setup.ksh \
functional/grow/grow_pool_001_pos.ksh \
functional/grow/grow_replicas_001_pos.ksh \
functional/history/cleanup.ksh \
functional/history/history_001_pos.ksh \
functional/history/history_002_pos.ksh \
functional/history/history_003_pos.ksh \
functional/history/history_004_pos.ksh \
functional/history/history_005_neg.ksh \
functional/history/history_006_neg.ksh \
functional/history/history_007_pos.ksh \
functional/history/history_008_pos.ksh \
functional/history/history_009_pos.ksh \
functional/history/history_010_pos.ksh \
functional/history/setup.ksh \
functional/inheritance/cleanup.ksh \
functional/inheritance/inherit_001_pos.ksh \
functional/inuse/inuse_001_pos.ksh \
functional/inuse/inuse_003_pos.ksh \
functional/inuse/inuse_004_pos.ksh \
functional/inuse/inuse_005_pos.ksh \
functional/inuse/inuse_006_pos.ksh \
functional/inuse/inuse_007_pos.ksh \
functional/inuse/inuse_008_pos.ksh \
functional/inuse/inuse_009_pos.ksh \
functional/inuse/setup.ksh \
functional/io/cleanup.ksh \
functional/io/io_uring.ksh \
functional/io/libaio.ksh \
functional/io/mmap.ksh \
functional/io/posixaio.ksh \
functional/io/psync.ksh \
functional/io/setup.ksh \
functional/io/sync.ksh \
functional/l2arc/cleanup.ksh \
functional/l2arc/l2arc_arcstats_pos.ksh \
functional/l2arc/l2arc_l2miss_pos.ksh \
functional/l2arc/l2arc_mfuonly_pos.ksh \
functional/l2arc/persist_l2arc_001_pos.ksh \
functional/l2arc/persist_l2arc_002_pos.ksh \
functional/l2arc/persist_l2arc_003_neg.ksh \
functional/l2arc/persist_l2arc_004_pos.ksh \
functional/l2arc/persist_l2arc_005_pos.ksh \
functional/l2arc/setup.ksh \
functional/large_files/cleanup.ksh \
functional/large_files/large_files_001_pos.ksh \
functional/large_files/large_files_002_pos.ksh \
functional/large_files/setup.ksh \
functional/largest_pool/largest_pool_001_pos.ksh \
functional/libzfs/cleanup.ksh \
functional/libzfs/libzfs_input.ksh \
functional/libzfs/setup.ksh \
functional/limits/cleanup.ksh \
functional/limits/filesystem_count.ksh \
functional/limits/filesystem_limit.ksh \
functional/limits/setup.ksh \
functional/limits/snapshot_count.ksh \
functional/limits/snapshot_limit.ksh \
functional/link_count/cleanup.ksh \
functional/link_count/link_count_001.ksh \
functional/link_count/link_count_root_inode.ksh \
functional/link_count/setup.ksh \
functional/log_spacemap/log_spacemap_import_logs.ksh \
functional/migration/cleanup.ksh \
functional/migration/migration_001_pos.ksh \
functional/migration/migration_002_pos.ksh \
functional/migration/migration_003_pos.ksh \
functional/migration/migration_004_pos.ksh \
functional/migration/migration_005_pos.ksh \
functional/migration/migration_006_pos.ksh \
functional/migration/migration_007_pos.ksh \
functional/migration/migration_008_pos.ksh \
functional/migration/migration_009_pos.ksh \
functional/migration/migration_010_pos.ksh \
functional/migration/migration_011_pos.ksh \
functional/migration/migration_012_pos.ksh \
functional/migration/setup.ksh \
functional/mmap/cleanup.ksh \
functional/mmap/mmap_libaio_001_pos.ksh \
functional/mmap/mmap_mixed.ksh \
functional/mmap/mmap_read_001_pos.ksh \
functional/mmap/mmap_seek_001_pos.ksh \
functional/mmap/mmap_sync_001_pos.ksh \
functional/mmap/mmap_write_001_pos.ksh \
functional/mmap/setup.ksh \
functional/mmp/cleanup.ksh \
functional/mmp/mmp_active_import.ksh \
functional/mmp/mmp_exported_import.ksh \
functional/mmp/mmp_hostid.ksh \
functional/mmp/mmp_inactive_import.ksh \
functional/mmp/mmp_interval.ksh \
functional/mmp/mmp_on_off.ksh \
functional/mmp/mmp_on_thread.ksh \
functional/mmp/mmp_on_uberblocks.ksh \
functional/mmp/mmp_on_zdb.ksh \
functional/mmp/mmp_reset_interval.ksh \
functional/mmp/mmp_write_distribution.ksh \
functional/mmp/mmp_write_uberblocks.ksh \
functional/mmp/multihost_history.ksh \
functional/mmp/setup.ksh \
functional/mount/cleanup.ksh \
functional/mount/setup.ksh \
functional/mount/umount_001.ksh \
functional/mount/umountall_001.ksh \
functional/mount/umount_unlinked_drain.ksh \
functional/mv_files/cleanup.ksh \
functional/mv_files/mv_files_001_pos.ksh \
functional/mv_files/mv_files_002_pos.ksh \
functional/mv_files/random_creation.ksh \
functional/mv_files/setup.ksh \
functional/nestedfs/cleanup.ksh \
functional/nestedfs/nestedfs_001_pos.ksh \
functional/nestedfs/setup.ksh \
functional/nopwrite/cleanup.ksh \
functional/nopwrite/nopwrite_copies.ksh \
functional/nopwrite/nopwrite_mtime.ksh \
functional/nopwrite/nopwrite_negative.ksh \
functional/nopwrite/nopwrite_promoted_clone.ksh \
functional/nopwrite/nopwrite_recsize.ksh \
functional/nopwrite/nopwrite_sync.ksh \
functional/nopwrite/nopwrite_varying_compression.ksh \
functional/nopwrite/nopwrite_volume.ksh \
functional/nopwrite/setup.ksh \
functional/no_space/cleanup.ksh \
functional/no_space/enospc_001_pos.ksh \
functional/no_space/enospc_002_pos.ksh \
functional/no_space/enospc_003_pos.ksh \
functional/no_space/enospc_df.ksh \
functional/no_space/enospc_ganging.ksh \
functional/no_space/enospc_rm.ksh \
functional/no_space/setup.ksh \
functional/online_offline/cleanup.ksh \
functional/online_offline/online_offline_001_pos.ksh \
functional/online_offline/online_offline_002_neg.ksh \
functional/online_offline/online_offline_003_neg.ksh \
functional/online_offline/setup.ksh \
functional/pam/cleanup.ksh \
functional/pam/pam_basic.ksh \
functional/pam/pam_nounmount.ksh \
functional/pam/pam_short_password.ksh \
functional/pam/setup.ksh \
functional/pool_checkpoint/checkpoint_after_rewind.ksh \
functional/pool_checkpoint/checkpoint_big_rewind.ksh \
functional/pool_checkpoint/checkpoint_capacity.ksh \
functional/pool_checkpoint/checkpoint_conf_change.ksh \
functional/pool_checkpoint/checkpoint_discard_busy.ksh \
functional/pool_checkpoint/checkpoint_discard.ksh \
functional/pool_checkpoint/checkpoint_discard_many.ksh \
functional/pool_checkpoint/checkpoint_indirect.ksh \
functional/pool_checkpoint/checkpoint_invalid.ksh \
functional/pool_checkpoint/checkpoint_lun_expsz.ksh \
functional/pool_checkpoint/checkpoint_open.ksh \
functional/pool_checkpoint/checkpoint_removal.ksh \
functional/pool_checkpoint/checkpoint_rewind.ksh \
functional/pool_checkpoint/checkpoint_ro_rewind.ksh \
functional/pool_checkpoint/checkpoint_sm_scale.ksh \
functional/pool_checkpoint/checkpoint_twice.ksh \
functional/pool_checkpoint/checkpoint_vdev_add.ksh \
functional/pool_checkpoint/checkpoint_zdb.ksh \
functional/pool_checkpoint/checkpoint_zhack_feat.ksh \
functional/pool_checkpoint/cleanup.ksh \
functional/pool_checkpoint/setup.ksh \
functional/pool_names/pool_names_001_pos.ksh \
functional/pool_names/pool_names_002_neg.ksh \
functional/poolversion/cleanup.ksh \
functional/poolversion/poolversion_001_pos.ksh \
functional/poolversion/poolversion_002_pos.ksh \
functional/poolversion/setup.ksh \
functional/privilege/cleanup.ksh \
functional/privilege/privilege_001_pos.ksh \
functional/privilege/privilege_002_pos.ksh \
functional/privilege/setup.ksh \
functional/procfs/cleanup.ksh \
functional/procfs/pool_state.ksh \
functional/procfs/procfs_list_basic.ksh \
functional/procfs/procfs_list_concurrent_readers.ksh \
functional/procfs/procfs_list_stale_read.ksh \
functional/procfs/setup.ksh \
functional/projectquota/cleanup.ksh \
functional/projectquota/projectid_001_pos.ksh \
functional/projectquota/projectid_002_pos.ksh \
functional/projectquota/projectid_003_pos.ksh \
functional/projectquota/projectquota_001_pos.ksh \
functional/projectquota/projectquota_002_pos.ksh \
functional/projectquota/projectquota_003_pos.ksh \
functional/projectquota/projectquota_004_neg.ksh \
functional/projectquota/projectquota_005_pos.ksh \
functional/projectquota/projectquota_006_pos.ksh \
functional/projectquota/projectquota_007_pos.ksh \
functional/projectquota/projectquota_008_pos.ksh \
functional/projectquota/projectquota_009_pos.ksh \
functional/projectquota/projectspace_001_pos.ksh \
functional/projectquota/projectspace_002_pos.ksh \
functional/projectquota/projectspace_003_pos.ksh \
functional/projectquota/projectspace_004_pos.ksh \
functional/projectquota/projecttree_001_pos.ksh \
functional/projectquota/projecttree_002_pos.ksh \
functional/projectquota/projecttree_003_neg.ksh \
functional/projectquota/setup.ksh \
functional/quota/cleanup.ksh \
functional/quota/quota_001_pos.ksh \
functional/quota/quota_002_pos.ksh \
functional/quota/quota_003_pos.ksh \
functional/quota/quota_004_pos.ksh \
functional/quota/quota_005_pos.ksh \
functional/quota/quota_006_neg.ksh \
functional/quota/setup.ksh \
functional/raidz/cleanup.ksh \
functional/raidz/raidz_001_neg.ksh \
functional/raidz/raidz_002_pos.ksh \
functional/raidz/raidz_003_pos.ksh \
functional/raidz/raidz_004_pos.ksh \
functional/raidz/setup.ksh \
functional/redacted_send/cleanup.ksh \
functional/redacted_send/redacted_compressed.ksh \
functional/redacted_send/redacted_contents.ksh \
functional/redacted_send/redacted_deleted.ksh \
functional/redacted_send/redacted_disabled_feature.ksh \
functional/redacted_send/redacted_embedded.ksh \
functional/redacted_send/redacted_holes.ksh \
functional/redacted_send/redacted_incrementals.ksh \
functional/redacted_send/redacted_largeblocks.ksh \
functional/redacted_send/redacted_many_clones.ksh \
functional/redacted_send/redacted_mixed_recsize.ksh \
functional/redacted_send/redacted_mounts.ksh \
functional/redacted_send/redacted_negative.ksh \
functional/redacted_send/redacted_origin.ksh \
functional/redacted_send/redacted_panic.ksh \
functional/redacted_send/redacted_props.ksh \
functional/redacted_send/redacted_resume.ksh \
functional/redacted_send/redacted_size.ksh \
functional/redacted_send/redacted_volume.ksh \
functional/redacted_send/setup.ksh \
functional/redundancy/cleanup.ksh \
functional/redundancy/redundancy_draid1.ksh \
functional/redundancy/redundancy_draid2.ksh \
functional/redundancy/redundancy_draid3.ksh \
functional/redundancy/redundancy_draid_damaged1.ksh \
functional/redundancy/redundancy_draid_damaged2.ksh \
functional/redundancy/redundancy_draid.ksh \
functional/redundancy/redundancy_draid_spare1.ksh \
functional/redundancy/redundancy_draid_spare2.ksh \
functional/redundancy/redundancy_draid_spare3.ksh \
functional/redundancy/redundancy_mirror.ksh \
functional/redundancy/redundancy_raidz1.ksh \
functional/redundancy/redundancy_raidz2.ksh \
functional/redundancy/redundancy_raidz3.ksh \
functional/redundancy/redundancy_raidz.ksh \
functional/redundancy/redundancy_stripe.ksh \
functional/redundancy/setup.ksh \
functional/refquota/cleanup.ksh \
functional/refquota/refquota_001_pos.ksh \
functional/refquota/refquota_002_pos.ksh \
functional/refquota/refquota_003_pos.ksh \
functional/refquota/refquota_004_pos.ksh \
functional/refquota/refquota_005_pos.ksh \
functional/refquota/refquota_006_neg.ksh \
functional/refquota/refquota_007_neg.ksh \
functional/refquota/refquota_008_neg.ksh \
functional/refquota/setup.ksh \
functional/refreserv/cleanup.ksh \
functional/refreserv/refreserv_001_pos.ksh \
functional/refreserv/refreserv_002_pos.ksh \
functional/refreserv/refreserv_003_pos.ksh \
functional/refreserv/refreserv_004_pos.ksh \
functional/refreserv/refreserv_005_pos.ksh \
functional/refreserv/refreserv_multi_raidz.ksh \
functional/refreserv/refreserv_raidz.ksh \
functional/refreserv/setup.ksh \
functional/removal/cleanup.ksh \
functional/removal/removal_all_vdev.ksh \
functional/removal/removal_cancel.ksh \
functional/removal/removal_check_space.ksh \
functional/removal/removal_condense_export.ksh \
functional/removal/removal_multiple_indirection.ksh \
functional/removal/removal_nopwrite.ksh \
functional/removal/removal_remap_deadlists.ksh \
functional/removal/removal_reservation.ksh \
functional/removal/removal_resume_export.ksh \
functional/removal/removal_sanity.ksh \
functional/removal/removal_with_add.ksh \
functional/removal/removal_with_create_fs.ksh \
functional/removal/removal_with_dedup.ksh \
functional/removal/removal_with_errors.ksh \
functional/removal/removal_with_export.ksh \
functional/removal/removal_with_faulted.ksh \
functional/removal/removal_with_ganging.ksh \
+ functional/removal/removal_with_indirect.ksh \
functional/removal/removal_with_remove.ksh \
functional/removal/removal_with_scrub.ksh \
functional/removal/removal_with_send.ksh \
functional/removal/removal_with_send_recv.ksh \
functional/removal/removal_with_snapshot.ksh \
functional/removal/removal_with_write.ksh \
functional/removal/removal_with_zdb.ksh \
functional/removal/remove_attach_mirror.ksh \
functional/removal/remove_expanded.ksh \
functional/removal/remove_indirect.ksh \
functional/removal/remove_mirror.ksh \
functional/removal/remove_mirror_sanity.ksh \
functional/removal/remove_raidz.ksh \
functional/rename_dirs/cleanup.ksh \
functional/rename_dirs/rename_dirs_001_pos.ksh \
functional/rename_dirs/setup.ksh \
functional/renameat2/cleanup.ksh \
functional/renameat2/setup.ksh \
functional/renameat2/renameat2_exchange.ksh \
functional/renameat2/renameat2_noreplace.ksh \
functional/renameat2/renameat2_whiteout.ksh \
functional/replacement/attach_import.ksh \
functional/replacement/attach_multiple.ksh \
functional/replacement/attach_rebuild.ksh \
functional/replacement/attach_resilver.ksh \
functional/replacement/cleanup.ksh \
functional/replacement/detach.ksh \
functional/replacement/rebuild_disabled_feature.ksh \
functional/replacement/rebuild_multiple.ksh \
functional/replacement/rebuild_raidz.ksh \
functional/replacement/replace_import.ksh \
functional/replacement/replace_rebuild.ksh \
functional/replacement/replace_resilver.ksh \
functional/replacement/resilver_restart_001.ksh \
functional/replacement/resilver_restart_002.ksh \
functional/replacement/scrub_cancel.ksh \
functional/replacement/setup.ksh \
functional/reservation/cleanup.ksh \
functional/reservation/reservation_001_pos.ksh \
functional/reservation/reservation_002_pos.ksh \
functional/reservation/reservation_003_pos.ksh \
functional/reservation/reservation_004_pos.ksh \
functional/reservation/reservation_005_pos.ksh \
functional/reservation/reservation_006_pos.ksh \
functional/reservation/reservation_007_pos.ksh \
functional/reservation/reservation_008_pos.ksh \
functional/reservation/reservation_009_pos.ksh \
functional/reservation/reservation_010_pos.ksh \
functional/reservation/reservation_011_pos.ksh \
functional/reservation/reservation_012_pos.ksh \
functional/reservation/reservation_013_pos.ksh \
functional/reservation/reservation_014_pos.ksh \
functional/reservation/reservation_015_pos.ksh \
functional/reservation/reservation_016_pos.ksh \
functional/reservation/reservation_017_pos.ksh \
functional/reservation/reservation_018_pos.ksh \
functional/reservation/reservation_019_pos.ksh \
functional/reservation/reservation_020_pos.ksh \
functional/reservation/reservation_021_neg.ksh \
functional/reservation/reservation_022_pos.ksh \
functional/reservation/setup.ksh \
functional/rootpool/cleanup.ksh \
functional/rootpool/rootpool_002_neg.ksh \
functional/rootpool/rootpool_003_neg.ksh \
functional/rootpool/rootpool_007_pos.ksh \
functional/rootpool/setup.ksh \
functional/rsend/cleanup.ksh \
functional/rsend/recv_dedup_encrypted_zvol.ksh \
functional/rsend/recv_dedup.ksh \
functional/rsend/rsend_001_pos.ksh \
functional/rsend/rsend_002_pos.ksh \
functional/rsend/rsend_003_pos.ksh \
functional/rsend/rsend_004_pos.ksh \
functional/rsend/rsend_005_pos.ksh \
functional/rsend/rsend_006_pos.ksh \
functional/rsend/rsend_007_pos.ksh \
functional/rsend/rsend_008_pos.ksh \
functional/rsend/rsend_009_pos.ksh \
functional/rsend/rsend_010_pos.ksh \
functional/rsend/rsend_011_pos.ksh \
functional/rsend/rsend_012_pos.ksh \
functional/rsend/rsend_013_pos.ksh \
functional/rsend/rsend_014_pos.ksh \
functional/rsend/rsend_016_neg.ksh \
functional/rsend/rsend_019_pos.ksh \
functional/rsend/rsend_020_pos.ksh \
functional/rsend/rsend_021_pos.ksh \
functional/rsend/rsend_022_pos.ksh \
functional/rsend/rsend_024_pos.ksh \
functional/rsend/rsend_025_pos.ksh \
functional/rsend/rsend_026_neg.ksh \
functional/rsend/rsend_027_pos.ksh \
functional/rsend/rsend_028_neg.ksh \
functional/rsend/rsend_029_neg.ksh \
functional/rsend/rsend_030_pos.ksh \
functional/rsend/rsend_031_pos.ksh \
functional/rsend/send-c_embedded_blocks.ksh \
functional/rsend/send-c_incremental.ksh \
functional/rsend/send-c_lz4_disabled.ksh \
functional/rsend/send-c_mixed_compression.ksh \
functional/rsend/send-c_props.ksh \
functional/rsend/send-c_recv_dedup.ksh \
functional/rsend/send-c_recv_lz4_disabled.ksh \
functional/rsend/send-c_resume.ksh \
functional/rsend/send-c_stream_size_estimate.ksh \
functional/rsend/send-c_verify_contents.ksh \
functional/rsend/send-c_verify_ratio.ksh \
functional/rsend/send-c_volume.ksh \
functional/rsend/send-c_zstream_recompress.ksh \
functional/rsend/send-c_zstreamdump.ksh \
functional/rsend/send-cpL_varied_recsize.ksh \
functional/rsend/send_doall.ksh \
functional/rsend/send_encrypted_incremental.ksh \
functional/rsend/send_encrypted_files.ksh \
functional/rsend/send_encrypted_freeobjects.ksh \
functional/rsend/send_encrypted_hierarchy.ksh \
functional/rsend/send_encrypted_props.ksh \
functional/rsend/send_encrypted_truncated_files.ksh \
functional/rsend/send_freeobjects.ksh \
functional/rsend/send_holds.ksh \
functional/rsend/send_hole_birth.ksh \
functional/rsend/send_invalid.ksh \
functional/rsend/send-L_toggle.ksh \
functional/rsend/send_mixed_raw.ksh \
functional/rsend/send_partial_dataset.ksh \
functional/rsend/send_raw_ashift.ksh \
functional/rsend/send_raw_spill_block.ksh \
functional/rsend/send_raw_large_blocks.ksh \
functional/rsend/send_realloc_dnode_size.ksh \
functional/rsend/send_realloc_encrypted_files.ksh \
functional/rsend/send_realloc_files.ksh \
functional/rsend/send_spill_block.ksh \
functional/rsend/send-wR_encrypted_zvol.ksh \
functional/rsend/setup.ksh \
functional/scrub_mirror/cleanup.ksh \
functional/scrub_mirror/scrub_mirror_001_pos.ksh \
functional/scrub_mirror/scrub_mirror_002_pos.ksh \
functional/scrub_mirror/scrub_mirror_003_pos.ksh \
functional/scrub_mirror/scrub_mirror_004_pos.ksh \
functional/scrub_mirror/setup.ksh \
functional/slog/cleanup.ksh \
functional/slog/setup.ksh \
functional/slog/slog_001_pos.ksh \
functional/slog/slog_002_pos.ksh \
functional/slog/slog_003_pos.ksh \
functional/slog/slog_004_pos.ksh \
functional/slog/slog_005_pos.ksh \
functional/slog/slog_006_pos.ksh \
functional/slog/slog_007_pos.ksh \
functional/slog/slog_008_neg.ksh \
functional/slog/slog_009_neg.ksh \
functional/slog/slog_010_neg.ksh \
functional/slog/slog_011_neg.ksh \
functional/slog/slog_012_neg.ksh \
functional/slog/slog_013_pos.ksh \
functional/slog/slog_014_pos.ksh \
functional/slog/slog_015_neg.ksh \
functional/slog/slog_016_pos.ksh \
functional/slog/slog_replay_fs_001.ksh \
functional/slog/slog_replay_fs_002.ksh \
functional/slog/slog_replay_volume.ksh \
functional/snapshot/cleanup.ksh \
functional/snapshot/clone_001_pos.ksh \
functional/snapshot/rollback_001_pos.ksh \
functional/snapshot/rollback_002_pos.ksh \
functional/snapshot/rollback_003_pos.ksh \
functional/snapshot/setup.ksh \
functional/snapshot/snapshot_001_pos.ksh \
functional/snapshot/snapshot_002_pos.ksh \
functional/snapshot/snapshot_003_pos.ksh \
functional/snapshot/snapshot_004_pos.ksh \
functional/snapshot/snapshot_005_pos.ksh \
functional/snapshot/snapshot_006_pos.ksh \
functional/snapshot/snapshot_007_pos.ksh \
functional/snapshot/snapshot_008_pos.ksh \
functional/snapshot/snapshot_009_pos.ksh \
functional/snapshot/snapshot_010_pos.ksh \
functional/snapshot/snapshot_011_pos.ksh \
functional/snapshot/snapshot_012_pos.ksh \
functional/snapshot/snapshot_013_pos.ksh \
functional/snapshot/snapshot_014_pos.ksh \
functional/snapshot/snapshot_015_pos.ksh \
functional/snapshot/snapshot_016_pos.ksh \
functional/snapshot/snapshot_017_pos.ksh \
functional/snapshot/snapshot_018_pos.ksh \
functional/snapused/cleanup.ksh \
functional/snapused/setup.ksh \
functional/snapused/snapused_001_pos.ksh \
functional/snapused/snapused_002_pos.ksh \
functional/snapused/snapused_003_pos.ksh \
functional/snapused/snapused_004_pos.ksh \
functional/snapused/snapused_005_pos.ksh \
functional/sparse/cleanup.ksh \
functional/sparse/setup.ksh \
functional/sparse/sparse_001_pos.ksh \
functional/stat/cleanup.ksh \
functional/stat/setup.ksh \
functional/stat/stat_001_pos.ksh \
functional/suid/cleanup.ksh \
functional/suid/setup.ksh \
functional/suid/suid_write_to_none.ksh \
functional/suid/suid_write_to_sgid.ksh \
functional/suid/suid_write_to_suid.ksh \
functional/suid/suid_write_to_suid_sgid.ksh \
functional/suid/suid_write_zil_replay.ksh \
functional/trim/autotrim_config.ksh \
functional/trim/autotrim_integrity.ksh \
functional/trim/autotrim_trim_integrity.ksh \
functional/trim/cleanup.ksh \
functional/trim/setup.ksh \
functional/trim/trim_config.ksh \
functional/trim/trim_integrity.ksh \
functional/trim/trim_l2arc.ksh \
functional/truncate/cleanup.ksh \
functional/truncate/setup.ksh \
functional/truncate/truncate_001_pos.ksh \
functional/truncate/truncate_002_pos.ksh \
functional/truncate/truncate_timestamps.ksh \
functional/upgrade/cleanup.ksh \
functional/upgrade/setup.ksh \
functional/upgrade/upgrade_projectquota_001_pos.ksh \
functional/upgrade/upgrade_readonly_pool.ksh \
functional/upgrade/upgrade_userobj_001_pos.ksh \
functional/user_namespace/cleanup.ksh \
functional/user_namespace/setup.ksh \
functional/user_namespace/user_namespace_001.ksh \
functional/user_namespace/user_namespace_002.ksh \
functional/user_namespace/user_namespace_003.ksh \
functional/user_namespace/user_namespace_004.ksh \
functional/userquota/cleanup.ksh \
functional/userquota/groupspace_001_pos.ksh \
functional/userquota/groupspace_002_pos.ksh \
functional/userquota/groupspace_003_pos.ksh \
functional/userquota/setup.ksh \
functional/userquota/userquota_001_pos.ksh \
functional/userquota/userquota_002_pos.ksh \
functional/userquota/userquota_003_pos.ksh \
functional/userquota/userquota_004_pos.ksh \
functional/userquota/userquota_005_neg.ksh \
functional/userquota/userquota_006_pos.ksh \
functional/userquota/userquota_007_pos.ksh \
functional/userquota/userquota_008_pos.ksh \
functional/userquota/userquota_009_pos.ksh \
functional/userquota/userquota_010_pos.ksh \
functional/userquota/userquota_011_pos.ksh \
functional/userquota/userquota_012_neg.ksh \
functional/userquota/userquota_013_pos.ksh \
functional/userquota/userspace_001_pos.ksh \
functional/userquota/userspace_002_pos.ksh \
functional/userquota/userspace_003_pos.ksh \
functional/userquota/userspace_encrypted.ksh \
functional/userquota/userspace_send_encrypted.ksh \
functional/userquota/userspace_encrypted_13709.ksh \
functional/vdev_zaps/cleanup.ksh \
functional/vdev_zaps/setup.ksh \
functional/vdev_zaps/vdev_zaps_001_pos.ksh \
functional/vdev_zaps/vdev_zaps_002_pos.ksh \
functional/vdev_zaps/vdev_zaps_003_pos.ksh \
functional/vdev_zaps/vdev_zaps_004_pos.ksh \
functional/vdev_zaps/vdev_zaps_005_pos.ksh \
functional/vdev_zaps/vdev_zaps_006_pos.ksh \
functional/vdev_zaps/vdev_zaps_007_pos.ksh \
functional/write_dirs/cleanup.ksh \
functional/write_dirs/setup.ksh \
functional/write_dirs/write_dirs_001_pos.ksh \
functional/write_dirs/write_dirs_002_pos.ksh \
functional/xattr/cleanup.ksh \
functional/xattr/setup.ksh \
functional/xattr/xattr_001_pos.ksh \
functional/xattr/xattr_002_neg.ksh \
functional/xattr/xattr_003_neg.ksh \
functional/xattr/xattr_004_pos.ksh \
functional/xattr/xattr_005_pos.ksh \
functional/xattr/xattr_006_pos.ksh \
functional/xattr/xattr_007_neg.ksh \
functional/xattr/xattr_008_pos.ksh \
functional/xattr/xattr_009_neg.ksh \
functional/xattr/xattr_010_neg.ksh \
functional/xattr/xattr_011_pos.ksh \
functional/xattr/xattr_012_pos.ksh \
functional/xattr/xattr_013_pos.ksh \
functional/xattr/xattr_compat.ksh \
functional/zpool_influxdb/cleanup.ksh \
functional/zpool_influxdb/setup.ksh \
functional/zpool_influxdb/zpool_influxdb.ksh \
functional/zvol/zvol_cli/cleanup.ksh \
functional/zvol/zvol_cli/setup.ksh \
functional/zvol/zvol_cli/zvol_cli_001_pos.ksh \
functional/zvol/zvol_cli/zvol_cli_002_pos.ksh \
functional/zvol/zvol_cli/zvol_cli_003_neg.ksh \
functional/zvol/zvol_ENOSPC/cleanup.ksh \
functional/zvol/zvol_ENOSPC/setup.ksh \
functional/zvol/zvol_ENOSPC/zvol_ENOSPC_001_pos.ksh \
functional/zvol/zvol_misc/cleanup.ksh \
functional/zvol/zvol_misc/setup.ksh \
functional/zvol/zvol_misc/zvol_misc_001_neg.ksh \
functional/zvol/zvol_misc/zvol_misc_002_pos.ksh \
functional/zvol/zvol_misc/zvol_misc_003_neg.ksh \
functional/zvol/zvol_misc/zvol_misc_004_pos.ksh \
functional/zvol/zvol_misc/zvol_misc_005_neg.ksh \
functional/zvol/zvol_misc/zvol_misc_006_pos.ksh \
functional/zvol/zvol_misc/zvol_misc_fua.ksh \
functional/zvol/zvol_misc/zvol_misc_hierarchy.ksh \
functional/zvol/zvol_misc/zvol_misc_rename_inuse.ksh \
functional/zvol/zvol_misc/zvol_misc_snapdev.ksh \
functional/zvol/zvol_misc/zvol_misc_trim.ksh \
functional/zvol/zvol_misc/zvol_misc_volmode.ksh \
functional/zvol/zvol_misc/zvol_misc_zil.ksh \
functional/zvol/zvol_stress/cleanup.ksh \
functional/zvol/zvol_stress/setup.ksh \
functional/zvol/zvol_stress/zvol_stress.ksh \
functional/zvol/zvol_swap/cleanup.ksh \
functional/zvol/zvol_swap/setup.ksh \
functional/zvol/zvol_swap/zvol_swap_001_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_002_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_003_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_004_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_005_pos.ksh \
functional/zvol/zvol_swap/zvol_swap_006_pos.ksh \
functional/idmap_mount/cleanup.ksh \
functional/idmap_mount/setup.ksh \
functional/idmap_mount/idmap_mount_001.ksh \
functional/idmap_mount/idmap_mount_002.ksh \
functional/idmap_mount/idmap_mount_003.ksh \
functional/idmap_mount/idmap_mount_004.ksh \
functional/idmap_mount/idmap_mount_005.ksh
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zdb/zdb_encrypted.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zdb/zdb_encrypted.ksh
index 4572f64947a1..0218c2ea1033 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zdb/zdb_encrypted.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zdb/zdb_encrypted.ksh
@@ -1,69 +1,69 @@
#!/bin/ksh -p
#
# CDDL HEADER START
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
# CDDL HEADER END
#
#
# Copyright (c) 2017, Datto, Inc. All rights reserved.
# Copyright (c) 2023, Rob Norris <robn@despairlabs.com>
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/tests/functional/cli_root/zfs_load-key/zfs_load-key_common.kshlib
#
# DESCRIPTION:
# 'zdb -K ...' should enable reading from an encrypt dataset
#
# STRATEGY:
# 1. Create an encrypted dataset
# 2. Write some data to a file
# 3. Run zdb -dddd on the file, confirm it can't be read
# 4. Run zdb -K ... -ddddd on the file, confirm it can be read
#
verify_runnable "both"
dataset="$TESTPOOL/$TESTFS2"
file="$TESTDIR2/somefile"
function cleanup
{
datasetexists $dataset && destroy_dataset $dataset -f
default_cleanup_noexit
}
log_onexit cleanup
log_must default_setup_noexit $DISKS
log_assert "'zdb -K' should enable reading from an encrypted dataset"
log_must eval "echo $PASSPHRASE | zfs create -o mountpoint=$TESTDIR2" \
"-o encryption=on -o keyformat=passphrase $dataset"
echo 'my great encrypted text' > $file
-obj="$(ls -i $file | cut -d' ' -f1)"
-size="$(wc -c < $file)"
+typeset -i obj=$(ls -i $file | cut -d' ' -f1)
+typeset -i size=$(wc -c < $file)
log_note "test file $file is objid $obj, size $size"
sync_pool $TESTPOOL true
log_must eval "zdb -dddd $dataset $obj | grep -q 'object encrypted'"
log_must eval "zdb -K $PASSPHRASE -dddd $dataset $obj | grep -q 'size\s$size$'"
log_pass "'zdb -K' enables reading from an encrypted dataset"
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/vdev_get.cfg b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/vdev_get.cfg
new file mode 100644
index 000000000000..71a64d4fae7a
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/vdev_get.cfg
@@ -0,0 +1,73 @@
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2022, Klara Inc.
+#
+
+# Set the expected properties of a vdev
+typeset -a properties=(
+ capacity
+ state
+ guid
+ asize
+ psize
+ ashift
+ size
+ free
+ allocated
+ comment
+ expandsize
+ fragmentation
+ bootsize
+ parity
+ path
+ devid
+ physpath
+ encpath
+ fru
+ parent
+ children
+ numchildren
+ read_errors
+ write_errors
+ checksum_errors
+ initialize_errors
+ null_ops
+ read_ops
+ write_ops
+ free_ops
+ claim_ops
+ trim_ops
+ null_bytes
+ read_bytes
+ write_bytes
+ free_bytes
+ claim_bytes
+ trim_bytes
+ removing
+ allocating
+ failfast
+ checksum_n
+ checksum_t
+ io_n
+ io_t
+)
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/vdev_get_001_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/vdev_get_001_pos.ksh
new file mode 100755
index 000000000000..bca2337861d4
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/vdev_get_001_pos.ksh
@@ -0,0 +1,62 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2022, Klara Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+. $STF_SUITE/tests/functional/cli_root/zpool_get/vdev_get.cfg
+
+#
+# DESCRIPTION:
+#
+# zpool get <pool> root works as expected
+#
+# STRATEGY:
+#
+# 1. use zpool get to retrieve properties from root vdev
+# 2. verify expected properties match detected properties
+#
+
+log_assert "zpool get all on root vdev"
+
+EXPECT="$(zpool get -H all ${TESTPOOL} root | wc -l)"
+if [ $? -ne 0 ]; then
+ log_fail "cannot retrieve properties from root vdev"
+fi
+
+i=0;
+while [ $i -lt "${#properties[@]}" ]
+do
+ log_must zpool get -H "${properties[$i]}" "$TESTPOOL" root
+ i=$(($i+1))
+done
+
+EXPECT=$((EXPECT))
+if [ $i -gt $EXPECT ]; then
+ log_fail "found vdev properties not in vdev_get.cfg: $i/$EXPECT."
+elif [ $i -lt $EXPECT ]; then
+ log_fail "expected properties not found in vdev_get.cfg: $i/$EXPECT."
+fi
+
+log_pass "zpool get all on root vdev"
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/zpool_get.cfg b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/zpool_get.cfg
index 097cd52e4777..160a0ca2e6db 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/zpool_get.cfg
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_get/zpool_get.cfg
@@ -1,108 +1,109 @@
#
# CDDL HEADER START
#
# The contents of this file are subject to the terms of the
# Common Development and Distribution License (the "License").
# You may not use this file except in compliance with the License.
#
# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
# or https://opensource.org/licenses/CDDL-1.0.
# See the License for the specific language governing permissions
# and limitations under the License.
#
# When distributing Covered Code, include this CDDL HEADER in each
# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
# If applicable, add the following below this CDDL HEADER, with the
# fields enclosed by brackets "[]" replaced with your own identifying
# information: Portions Copyright [yyyy] [name of copyright owner]
#
# CDDL HEADER END
#
#
# Copyright 2009 Sun Microsystems, Inc. All rights reserved.
# Use is subject to license terms.
#
#
# Copyright (c) 2013, 2014 by Delphix. All rights reserved.
# Copyright 2016 Nexenta Systems, Inc. All rights reserved.
#
# Set the expected properties of zpool
typeset -a properties=(
"size"
"capacity"
"altroot"
"health"
"guid"
"load_guid"
"version"
"bootfs"
"delegation"
"autoreplace"
"cachefile"
"checkpoint"
"failmode"
"listsnapshots"
"autoexpand"
"dedupratio"
"free"
"allocated"
"readonly"
"comment"
"expandsize"
"freeing"
"fragmentation"
"leaked"
"multihost"
"autotrim"
"compatibility"
"bcloneused"
"bclonesaved"
"bcloneratio"
"feature@async_destroy"
"feature@empty_bpobj"
"feature@lz4_compress"
"feature@multi_vdev_crash_dump"
"feature@spacemap_histogram"
"feature@enabled_txg"
"feature@hole_birth"
"feature@extensible_dataset"
"feature@embedded_data"
"feature@bookmarks"
"feature@filesystem_limits"
"feature@large_blocks"
"feature@sha512"
"feature@skein"
"feature@edonr"
"feature@device_removal"
"feature@obsolete_counts"
"feature@zpool_checkpoint"
"feature@spacemap_v2"
"feature@redaction_bookmarks"
"feature@redacted_datasets"
"feature@bookmark_written"
"feature@log_spacemap"
"feature@device_rebuild"
"feature@draid"
)
if is_linux || is_freebsd; then
properties+=(
"ashift"
"feature@large_dnode"
"feature@userobj_accounting"
"feature@encryption"
"feature@project_quota"
"feature@allocation_classes"
"feature@resilver_defer"
"feature@bookmark_v2"
"feature@livelist"
"feature@zstd_compress"
"feature@zilsaxattr"
"feature@head_errlog"
"feature@blake3"
"feature@block_cloning"
+ "feature@vdev_zaps_v2"
)
fi
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_set/vdev_set_001_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_set/vdev_set_001_pos.ksh
new file mode 100755
index 000000000000..a1f3efb90577
--- /dev/null
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_set/vdev_set_001_pos.ksh
@@ -0,0 +1,52 @@
+#!/bin/ksh -p
+#
+# CDDL HEADER START
+#
+# The contents of this file are subject to the terms of the
+# Common Development and Distribution License (the "License").
+# You may not use this file except in compliance with the License.
+#
+# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+# or https://opensource.org/licenses/CDDL-1.0.
+# See the License for the specific language governing permissions
+# and limitations under the License.
+#
+# When distributing Covered Code, include this CDDL HEADER in each
+# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+# If applicable, add the following below this CDDL HEADER, with the
+# fields enclosed by brackets "[]" replaced with your own identifying
+# information: Portions Copyright [yyyy] [name of copyright owner]
+#
+# CDDL HEADER END
+#
+
+#
+# Copyright (c) 2022, Klara Inc.
+#
+
+. $STF_SUITE/include/libtest.shlib
+
+#
+# DESCRIPTION:
+#
+# zpool set comment property on root vdev
+#
+# STRATEGY:
+# 1. set a property on root vdev
+# 2. verify the property is set
+#
+
+log_assert "zpool set comment property on root vdev"
+
+log_must zpool set comment="openzfs" ${TESTPOOL} root
+
+COMMENT="$(zpool get -H -o value comment ${TESTPOOL} root)"
+if [ $? -ne 0 ]; then
+ log_fail "cant retrieve comment property from root vdev"
+fi
+
+if [ "$COMMENT" != "openzfs" ]; then
+ log_fail "unexpected value for comment property: $COMMENT != \"openzfs\""
+fi
+
+log_pass "zpool set comment property on root vdev"
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/rsend/send-c_volume.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/rsend/send-c_volume.ksh
index 988ed91b9918..1bf234823459 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/rsend/send-c_volume.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/rsend/send-c_volume.ksh
@@ -1,78 +1,83 @@
#!/bin/ksh -p
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2015 by Delphix. All rights reserved.
#
. $STF_SUITE/tests/functional/rsend/rsend.kshlib
#
# Description:
# Verify that compressed send correctly handles volumes
#
# Strategy:
# 1. Write compressible data into a volume, take a snap
# 2. Verify the compressed stream is the correct size, and has the correct data
# 3. Repeat step 2 for an incremental compressed stream
#
function cleanup
{
+ rm $BACKDIR/copy
log_must_busy zfs destroy -r $vol
cleanup_pool $POOL2
}
verify_runnable "both"
log_assert "Verify compressed send works with volumes"
log_onexit cleanup
typeset vol="$POOL/newvol"
typeset vol2="$POOL2/newvol"
typeset voldev="$ZVOL_DEVDIR/$POOL/newvol"
typeset voldev2="$ZVOL_DEVDIR/$POOL2/newvol"
typeset data1=$BACKDIR/file.0
typeset data2=$BACKDIR/file.1
typeset megs=8
log_must zfs create -V 256m -o compress=lz4 $vol
write_compressible $BACKDIR ${megs}m 2
md5_1=$(md5digest $data1)
md5_2=$(md5digest $data2)
log_must dd if=$data1 of=$voldev bs=1024k
log_must zfs snapshot $vol@snap
log_must eval "zfs send -c $vol@snap >$BACKDIR/full"
log_must eval "zfs recv -d $POOL2 <$BACKDIR/full"
verify_stream_size $BACKDIR/full $vol
verify_stream_size $BACKDIR/full $vol2
-md5=$(dd if=$voldev2 bs=1024k count=$megs 2>/dev/null | md5digest)
+block_device_wait $voldev2
+log_must dd if=$voldev2 of=$BACKDIR/copy bs=1024k count=$megs
+md5=$(md5digest $BACKDIR/copy)
[[ $md5 = $md5_1 ]] || log_fail "md5 mismatch: $md5 != $md5_1"
# Repeat, for an incremental send
log_must dd seek=$megs if=$data2 of=$voldev bs=1024k
log_must zfs snapshot $vol@snap2
log_must eval "zfs send -c -i snap $vol@snap2 >$BACKDIR/inc"
log_must eval "zfs recv -d $POOL2 <$BACKDIR/inc"
verify_stream_size $BACKDIR/inc $vol 90 $vol@snap
verify_stream_size $BACKDIR/inc $vol2 90 $vol2@snap
-md5=$(dd skip=$megs if=$voldev2 bs=1024k count=$megs 2>/dev/null | md5digest)
+block_device_wait $voldev2
+log_must dd skip=$megs if=$voldev2 of=$BACKDIR/copy bs=1024k count=$megs
+md5=$(md5digest $BACKDIR/copy)
[[ $md5 = $md5_2 ]] || log_fail "md5 mismatch: $md5 != $md5_2"
log_pass "Verify compressed send works with volumes"
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps.kshlib b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps.kshlib
index ad5bd9e7f81b..c68a5b2c4c83 100644
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps.kshlib
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps.kshlib
@@ -1,88 +1,101 @@
#
# CDDL HEADER START
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
# CDDL HEADER END
#
#
# Copyright (c) 2015, 2016 by Delphix. All rights reserved.
#
function get_conf_section # regex conf
{
typeset regex="$1"
typeset conf="$2"
awk -v r="$1" '$0 ~ r, 0 {if($0 ~ r) next; if(/children\[/) exit; print}' "$conf"
}
function get_leaf_vd_zap # dsk conf
{
get_conf_section "$1" "$2" | awk '/com.delphix:vdev_zap_leaf: [0-9]+/ {print $2}'
}
function get_top_vd_zap # dsk conf
{
get_conf_section "$1" "$2" | awk '/com.delphix:vdev_zap_top: [0-9]+/ {print $2}'
}
+function get_root_vd_zap # conf
+{
+ awk '/com.klarasystems:vdev_zap_root: [0-9]+/ {print $2}' "$1"
+}
function assert_has_sentinel # conf
{
log_must grep -q "com.delphix:has_per_vdev_zaps" "$1"
}
function assert_zap_common # pool vd lvl zapobj
{
typeset pool=$1
typeset vd="$2"
typeset lvl=$3
typeset zapobj=$4
if [ -z "$zapobj" ]; then
log_fail "$vd on $pool has no $lvl ZAP in config"
elif ! zdb -d $pool $zapobj | grep -q 'zap'; then
log_fail "$vd on $pool has no $lvl ZAP in MOS"
fi
}
+function assert_root_zap # pool conf
+{
+ typeset pool=$1
+ typeset conf=$2
+
+ root_zap=$(get_root_vd_zap $conf)
+ assert_zap_common $pool "root vdev" "root" $root_zap
+}
+
function assert_top_zap # pool vd conf
{
typeset pool=$1
typeset vd="$2"
typeset conf=$3
top_zap=$(get_top_vd_zap "$vd" $conf)
assert_zap_common $pool "$vd" "top" $top_zap
}
function assert_leaf_zap # pool vd conf
{
typeset pool=$1
typeset vd="$2"
typeset conf=$3
leaf_zap=$(get_leaf_vd_zap "$vd" $conf)
assert_zap_common $pool "$vd" "leaf" $leaf_zap
}
#
# Code common to setup/teardown for each test.
#
function cleanup
{
datasetexists $TESTPOOL && log_must zpool destroy -f $TESTPOOL
[ -e $conf ] && log_must rm -f "$conf"
poolexists $POOL2 && log_must zpool destroy -f $POOL2
}
log_onexit cleanup
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_001_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_001_pos.ksh
index b67cc6d973e6..bdc8dcd468ae 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_001_pos.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_001_pos.ksh
@@ -1,42 +1,43 @@
#!/bin/ksh
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2015 by Delphix. All rights reserved.
#
#
# Description:
# Verify that per-vdev ZAPs are created with one vdev.
#
# Strategy:
# 1. Create a pool with one disk.
-# 2. Verify that the disk has a top and leaf ZAP in its config and the MOS.
+# 2. Verify that the disk has a root, top and leaf ZAP in its config and the MOS.
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/tests/functional/vdev_zaps/vdev_zaps.kshlib
log_assert "Per-vdev ZAPs are created on pool creation with one disk."
DISK=${DISKS%% *}
log_must zpool create -f $TESTPOOL $DISK
conf="$TESTDIR/vz001"
log_must eval "zdb -PC $TESTPOOL > $conf"
+assert_root_zap $TESTPOOL "$conf"
assert_top_zap $TESTPOOL $DISK "$conf"
assert_leaf_zap $TESTPOOL $DISK "$conf"
assert_has_sentinel "$conf"
log_pass "Per-vdev ZAPs are created in a one-disk pool."
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_002_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_002_pos.ksh
index c571973b080b..35c4f64fa463 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_002_pos.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_002_pos.ksh
@@ -1,44 +1,45 @@
#!/bin/ksh
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2015 by Delphix. All rights reserved.
#
#
# Description:
# Verify that per-vdev ZAPs are created with multiple vdevs.
#
# Strategy:
# 1. Create a pool with multiple disks.
# 2. Verify that each has both a top and leaf zap.
# 3. Verify that each of those ZAPs exists in the MOS.
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/tests/functional/vdev_zaps/vdev_zaps.kshlib
log_assert "Per-vdev ZAPs are created on pool creation with many disks."
log_must zpool create -f $TESTPOOL $DISKS
conf="$TESTDIR/vz002"
log_must eval "zdb -PC $TESTPOOL > $conf"
assert_has_sentinel "$conf"
+assert_root_zap $TESTPOOL "$conf"
for DISK in $DISKS; do
assert_top_zap $TESTPOOL $DISK "$conf"
assert_leaf_zap $TESTPOOL $DISK "$conf"
done
log_pass
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_003_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_003_pos.ksh
index 015729576a7d..bb6875c339c3 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_003_pos.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_003_pos.ksh
@@ -1,47 +1,48 @@
#!/bin/ksh
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2015 by Delphix. All rights reserved.
#
#
# Description:
# Verify that per-vdev ZAPs are created with multi-level vdev tree.
#
# Strategy:
# 1. Create a pool with a multi-disk mirror.
# 2. Verify that mirror has top ZAP but no leaf ZAP.
# 3. Verify that each disk has a leaf ZAP but no top ZAP.
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/tests/functional/vdev_zaps/vdev_zaps.kshlib
log_assert "Per-vdev ZAPs are created on pool creation with multi-level vdev "\
"trees."
log_must zpool create -f $TESTPOOL mirror $DISKS
conf="$TESTDIR/vz003"
log_must eval "zdb -PC $TESTPOOL > $conf"
assert_has_sentinel "$conf"
+assert_root_zap $TESTPOOL "$conf"
assert_top_zap $TESTPOOL "type: 'mirror'" "$conf"
for DISK in $DISKS; do
assert_leaf_zap $TESTPOOL $DISK "$conf"
top_zap=$(get_top_vd_zap $DISK "$conf")
[[ -n "$top_zap" ]] && log_fail "Leaf vdev $DISK has top-level ZAP."
done
log_pass
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_004_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_004_pos.ksh
index 3d0f55d5a9a7..e82e398c6d27 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_004_pos.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_004_pos.ksh
@@ -1,95 +1,96 @@
#!/bin/ksh
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2015 by Delphix. All rights reserved.
#
#
# Description:
# Verify that per-vdev ZAPs are properly transferred on attach/detach.
#
# Strategy:
# 1. Create a pool with one disk. Verify that it has a top and leaf ZAP.
# 2. Attach a disk.
# 3. Verify that top-level and leaf-level ZAPs were transferred properly.
# 4. Verify that the newly-attached disk has a leaf ZAP.
# 5. Detach the original disk.
# 6. Verify that top-level and leaf-level ZAPs were transferred properly.
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/tests/functional/vdev_zaps/vdev_zaps.kshlib
log_assert "Per-vdev ZAPs are transferred properly on attach/detach"
DISK=${DISKS%% *}
log_must zpool create -f $TESTPOOL $DISK
# Make the pool.
conf="$TESTDIR/vz004"
log_must eval "zdb -PC $TESTPOOL > $conf"
assert_has_sentinel "$conf"
+assert_root_zap $TESTPOOL "$conf"
orig_top=$(get_top_vd_zap $DISK $conf)
orig_leaf=$(get_leaf_vd_zap $DISK $conf)
assert_zap_common $TESTPOOL $DISK "top" $orig_top
#
# Attach a disk.
#
read -r _ disk2 _ <<<"$DISKS"
log_must zpool attach $TESTPOOL $DISK $disk2
log_must zpool wait -t resilver $TESTPOOL
log_must eval "zdb -PC $TESTPOOL > $conf"
# Ensure top-level ZAP was transferred successfully.
new_top=$(get_top_vd_zap "type: 'mirror'" $conf)
if [[ "$new_top" -ne "$orig_top" ]]; then
log_fail "Top-level ZAP wasn't transferred successfully on attach."
fi
# Ensure leaf ZAP of original disk was transferred successfully.
new_leaf=$(get_leaf_vd_zap $DISK $conf)
if [[ "$new_leaf" -ne "$orig_leaf" ]]; then
log_fail "$DISK used to have leaf-level ZAP $orig_leaf, now has "\
"$new_leaf"
fi
# Ensure original disk no longer has top-level ZAP.
dsk1_top=$(get_top_vd_zap $DISK $conf)
[[ -n "$dsk1_top" ]] && log_fail "$DISK has top-level ZAP, but is only leaf."
# Ensure attached disk got a leaf-level ZAP but not a top-level ZAP.
dsk2_top=$(get_top_vd_zap $disk2 $conf)
dsk2_leaf=$(get_leaf_vd_zap $disk2 $conf)
[[ -n "$dsk2_top" ]] && log_fail "Attached disk $disk2 has top ZAP."
[[ -z "$dsk2_leaf" ]] && log_fail "Attached disk $disk2 has no leaf ZAP."
#
# Detach original disk.
#
log_must zpool detach $TESTPOOL $DISK
log_must eval "zdb -PC $TESTPOOL > $conf"
final_top=$(get_top_vd_zap $disk2 $conf)
final_leaf=$(get_leaf_vd_zap $disk2 $conf)
# Make sure top ZAP was successfully transferred.
[[ "$final_top" -ne "$orig_top" ]] && log_fail "Lost top-level ZAP when "\
"promoting $disk2 (expected $orig_top, found $final_top)"
# Make sure leaf ZAP was successfully transferred.
[[ "$final_leaf" -ne "$dsk2_leaf" ]] && log_fail "$disk2 lost its leaf ZAP "\
"on promotion (expected $dsk2_leaf, got $final_leaf)"
log_pass
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_005_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_005_pos.ksh
index 1d82218bf283..4b9b45e149d7 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_005_pos.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_005_pos.ksh
@@ -1,63 +1,64 @@
#!/bin/ksh
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2015 by Delphix. All rights reserved.
#
#
# Description:
# Verify that per-vdev ZAPs persist when the pool is exported and imported.
#
# Strategy:
# 1. Create a pool with a disk.
# 2. Export the pool and re-import it.
# 3. Verify that the ZAPs aren't different.
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/tests/functional/vdev_zaps/vdev_zaps.kshlib
log_assert "Per-vdev ZAPs persist across export/import."
DISK=${DISKS%% *}
log_must zpool create -f $TESTPOOL $DISK
# Make the pool.
conf="$TESTDIR/vz005"
log_must eval "zdb -PC $TESTPOOL > $conf"
assert_has_sentinel "$conf"
+assert_root_zap $TESTPOOL "$conf"
orig_top=$(get_top_vd_zap $DISK $conf)
orig_leaf=$(get_leaf_vd_zap $DISK $conf)
assert_zap_common $TESTPOOL $DISK "top" $orig_top
assert_zap_common $TESTPOOL $DISK "leaf" $orig_leaf
sync_all_pools
# Export the pool.
log_must zpool export $TESTPOOL
# Import the pool.
log_must zpool import $TESTPOOL
# Verify that ZAPs persisted.
log_must eval "zdb -PC $TESTPOOL > $conf"
new_top=$(get_top_vd_zap $DISK $conf)
new_leaf=$(get_leaf_vd_zap $DISK $conf)
[[ "$new_top" -ne "$orig_top" ]] && log_fail "Top ZAP ($new_top) after "\
"import does not match top ZAP before export ($orig_top)"
[[ "$new_leaf" -ne "$orig_leaf" ]] && log_fail "Leaf ZAP ($new_leaf) after "\
"import does not match leaf ZAP before export ($orig_leaf)"
log_pass
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_006_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_006_pos.ksh
index ce94336c7c5d..2ac493b8b0d2 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_006_pos.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_006_pos.ksh
@@ -1,46 +1,47 @@
#!/bin/ksh
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2015, 2016 by Delphix. All rights reserved.
#
#
# Description:
# Verify that top-level per-vdev ZAPs are created for added devices
#
# Strategy:
# 1. Create a pool with one disk.
# 2. Add a disk.
# 3. Verify its ZAPs were created.
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/tests/functional/vdev_zaps/vdev_zaps.kshlib
DISK_ARR=($DISKS)
DISK=${DISK_ARR[0]}
log_must zpool create -f $TESTPOOL $DISK
log_assert "Per-vdev ZAPs are created for added vdevs."
log_must zpool add -f $TESTPOOL ${DISK_ARR[1]}
conf="$TESTDIR/vz006"
log_must eval "zdb -PC $TESTPOOL > $conf"
assert_has_sentinel "$conf"
+assert_root_zap $TESTPOOL "$conf"
orig_top=$(get_top_vd_zap ${DISK_ARR[1]} $conf)
assert_zap_common $TESTPOOL ${DISK_ARR[1]} "top" $orig_top
assert_leaf_zap $TESTPOOL ${DISK_ARR[1]} "$conf"
log_pass
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_007_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_007_pos.ksh
index c7f12c633706..c7a4a62de436 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_007_pos.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/vdev_zaps/vdev_zaps_007_pos.ksh
@@ -1,74 +1,75 @@
#!/bin/ksh
#
# This file and its contents are supplied under the terms of the
# Common Development and Distribution License ("CDDL"), version 1.0.
# You may only use this file in accordance with the terms of version
# 1.0 of the CDDL.
#
# A full copy of the text of the CDDL should have accompanied this
# source. A copy of the CDDL is also available via the Internet at
# http://www.illumos.org/license/CDDL.
#
#
# Copyright (c) 2015 by Delphix. All rights reserved.
#
#
# Description:
# Verify that ZAPs are handled properly during mirror pool splitting.
#
# Strategy:
# 1. Create a pool with a two-way mirror.
# 2. Split the pool.
# 3. Verify that the ZAPs in the old pool persisted.
# 4. Import the new pool.
# 5. Verify that the ZAPs in the new pool persisted.
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/tests/functional/vdev_zaps/vdev_zaps.kshlib
DISK_ARR=($DISKS)
POOL2=${TESTPOOL}2
log_must zpool create -f $TESTPOOL mirror ${DISK_ARR[0]} ${DISK_ARR[1]}
log_assert "Per-vdev ZAPs persist correctly on the original pool after split."
conf="$TESTDIR/vz007"
log_must eval "zdb -PC $TESTPOOL > $conf"
assert_has_sentinel "$conf"
+assert_root_zap $TESTPOOL "$conf"
orig_top=$(get_top_vd_zap "type: 'mirror'" $conf)
orig_leaf0=$(get_leaf_vd_zap ${DISK_ARR[0]} $conf)
orig_leaf1=$(get_leaf_vd_zap ${DISK_ARR[1]} $conf)
assert_zap_common $TESTPOOL "type: 'mirror'" "top" $orig_top
assert_zap_common $TESTPOOL ${DISK_ARR[0]} "leaf" $orig_leaf0
assert_zap_common $TESTPOOL ${DISK_ARR[1]} "leaf" $orig_leaf1
log_must zpool split $TESTPOOL $POOL2 ${DISK_ARR[1]}
# Make sure old pool's ZAPs are consistent.
log_must eval "zdb -PC $TESTPOOL > $conf"
new_leaf0=$(get_leaf_vd_zap ${DISK_ARR[0]} $conf)
new_top_s0=$(get_top_vd_zap ${DISK_ARR[0]} $conf)
[[ "$new_leaf0" -ne "$orig_leaf0" ]] && log_fail "Leaf ZAP in original pool "\
"didn't persist (expected $orig_leaf0, got $new_leaf0)"
[[ "$new_top_s0" -ne "$orig_top" ]] && log_fail "Top ZAP in original pool "\
"didn't persist (expected $orig_top, got $new_top_s0)"
log_assert "Per-vdev ZAPs persist on the new pool after import."
# Import the split pool.
log_must zpool import $POOL2
log_must eval "zdb -PC $POOL2 > $conf"
new_leaf1=$(get_leaf_vd_zap ${DISK_ARR[1]} $conf)
new_top_s1=$(get_top_vd_zap ${DISK_ARR[1]} $conf)
[[ "$new_leaf1" -ne "$orig_leaf1" ]] && log_fail "Leaf ZAP in new pool "\
"didn't persist (expected $orig_leaf1, got $new_leaf1)"
[[ "$new_top_s1" -ne "$orig_top" ]] && log_fail "Top ZAP in new pool "\
"didn't persist (expected $orig_top, got $new_top_s1)"
log_pass
diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/zvol/zvol_misc/zvol_misc_trim.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/zvol/zvol_misc/zvol_misc_trim.ksh
index 2c4ef28ab826..46cac3ecb6c2 100755
--- a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/zvol/zvol_misc/zvol_misc_trim.ksh
+++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/zvol/zvol_misc/zvol_misc_trim.ksh
@@ -1,137 +1,137 @@
#!/bin/ksh -p
#
# CDDL HEADER START
#
# The contents of this file are subject to the terms of the
# Common Development and Distribution License (the "License").
# You may not use this file except in compliance with the License.
#
# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
# or https://opensource.org/licenses/CDDL-1.0.
# See the License for the specific language governing permissions
# and limitations under the License.
#
# When distributing Covered Code, include this CDDL HEADER in each
# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
# If applicable, add the following below this CDDL HEADER, with the
# fields enclosed by brackets "[]" replaced with your own identifying
# information: Portions Copyright [yyyy] [name of copyright owner]
#
# CDDL HEADER END
#
#
# Copyright (c) 2022 by Lawrence Livermore National Security, LLC.
#
. $STF_SUITE/include/libtest.shlib
. $STF_SUITE/include/math.shlib
. $STF_SUITE/tests/functional/zvol/zvol_common.shlib
#
# DESCRIPTION:
# Verify we can TRIM a zvol
#
# STRATEGY:
# 1. TRIM the entire zvol to remove data from older tests
# 2. Create a 5MB data file
# 3. Write the file to the zvol
# 4. Observe 5MB of used space on the zvol
# 5. TRIM the first 1MB and last 2MB of the 5MB block of data.
# 6. Observe 2MB of used space on the zvol
# 7. Verify the trimmed regions are zero'd on the zvol
verify_runnable "global"
if is_linux ; then
# We need '--force' here since the prior tests may leave a filesystem
# on the zvol, and blkdiscard will see that filesystem and print a
# warning unless you force it.
#
# Only blkdiscard >= v2.36 supports --force, so we need to
# check for it.
if blkdiscard --help | grep -q '\-\-force' ; then
trimcmd='blkdiscard --force'
else
trimcmd='blkdiscard'
fi
else
# By default, FreeBSD 'trim' always does a dry-run. '-f' makes
# it perform the actual operation.
trimcmd='trim -f'
fi
if ! is_physical_device $DISKS; then
log_unsupported "This directory cannot be run on raw files."
fi
typeset datafile1="$(mktemp zvol_misc_flags1.XXXXXX)"
typeset datafile2="$(mktemp zvol_misc_flags2.XXXXXX)"
typeset zvolpath=${ZVOL_DEVDIR}/$TESTPOOL/$TESTVOL
function cleanup
{
rm "$datafile1" "$datafile2"
}
function do_test {
# Wait for udev to create symlinks to our zvol
block_device_wait $zvolpath
# Create a data file
log_must dd if=/dev/urandom of="$datafile1" bs=1M count=5
# Write to zvol
log_must dd if=$datafile1 of=$zvolpath conv=fsync
sync_pool
# Record how much space we've used (should be 5MB, with 128k
# of tolerance).
before="$(get_prop refer $TESTPOOL/$TESTVOL)"
log_must within_tolerance $before 5242880 131072
# We currently have 5MB of random data on the zvol.
# Trim the first 1MB and also trim 2MB at offset 3MB.
log_must $trimcmd -l $((1 * 1048576)) $zvolpath
log_must $trimcmd -o $((3 * 1048576)) -l $((2 * 1048576)) $zvolpath
sync_pool
# After trimming 3MB, the zvol should have 2MB of data (with 128k of
# tolerance).
after="$(get_prop refer $TESTPOOL/$TESTVOL)"
log_must within_tolerance $after 2097152 131072
# Make the same holes in our test data
log_must dd if=/dev/zero of="$datafile1" bs=1M count=1 conv=notrunc
log_must dd if=/dev/zero of="$datafile1" bs=1M count=2 seek=3 conv=notrunc
# Extract data from our zvol
log_must dd if=$zvolpath of="$datafile2" bs=1M count=5
# Compare the data we expect with what's on our zvol. diff will return
# non-zero if they differ.
log_must diff $datafile1 $datafile2
log_must rm $datafile1 $datafile2
}
log_assert "Verify that a ZFS volume can be TRIMed"
log_onexit cleanup
log_must zfs set compression=off $TESTPOOL/$TESTVOL
# Remove old data from previous tests
log_must $trimcmd $zvolpath
set_blk_mq 1
-log_must zpool export $TESTPOOL
+log_must_busy zpool export $TESTPOOL
log_must zpool import $TESTPOOL
do_test
set_blk_mq 0
-log_must zpool export $TESTPOOL
+log_must_busy zpool export $TESTPOOL
log_must zpool import $TESTPOOL
do_test
log_pass "ZFS volumes can be trimmed"
diff --git a/sys/modules/zfs/zfs_config.h b/sys/modules/zfs/zfs_config.h
index ea7471e1efaa..f489f0b6c69a 100644
--- a/sys/modules/zfs/zfs_config.h
+++ b/sys/modules/zfs/zfs_config.h
@@ -1,1033 +1,1087 @@
/*
* $FreeBSD$
*/
/* zfs_config.h. Generated from zfs_config.h.in by configure. */
/* zfs_config.h.in. Generated from configure.ac by autoheader. */
/* Define to 1 if translation of program messages to the user's native
language is requested. */
/* #undef ENABLE_NLS */
/* bio_end_io_t wants 1 arg */
/* #undef HAVE_1ARG_BIO_END_IO_T */
/* lookup_bdev() wants 1 arg */
/* #undef HAVE_1ARG_LOOKUP_BDEV */
/* submit_bio() wants 1 arg */
/* #undef HAVE_1ARG_SUBMIT_BIO */
/* bdi_setup_and_register() wants 2 args */
/* #undef HAVE_2ARGS_BDI_SETUP_AND_REGISTER */
/* vfs_getattr wants 2 args */
/* #undef HAVE_2ARGS_VFS_GETATTR */
/* zlib_deflate_workspacesize() wants 2 args */
/* #undef HAVE_2ARGS_ZLIB_DEFLATE_WORKSPACESIZE */
/* bdi_setup_and_register() wants 3 args */
/* #undef HAVE_3ARGS_BDI_SETUP_AND_REGISTER */
/* vfs_getattr wants 3 args */
/* #undef HAVE_3ARGS_VFS_GETATTR */
/* vfs_getattr wants 4 args */
/* #undef HAVE_4ARGS_VFS_GETATTR */
/* kernel has access_ok with 'type' parameter */
/* #undef HAVE_ACCESS_OK_TYPE */
/* posix_acl has refcount_t */
/* #undef HAVE_ACL_REFCOUNT */
/* add_disk() returns int */
/* #undef HAVE_ADD_DISK_RET */
/* Define if host toolchain supports AES */
#define HAVE_AES 1
/* Define if you have [rt] */
#define HAVE_AIO_H 1
#ifdef __amd64__
#ifndef RESCUE
/* Define if host toolchain supports AVX */
#define HAVE_AVX 1
#endif
/* Define if host toolchain supports AVX2 */
#define HAVE_AVX2 1
/* Define if host toolchain supports AVX512BW */
#define HAVE_AVX512BW 1
/* Define if host toolchain supports AVX512CD */
#define HAVE_AVX512CD 1
/* Define if host toolchain supports AVX512DQ */
#define HAVE_AVX512DQ 1
/* Define if host toolchain supports AVX512ER */
#define HAVE_AVX512ER 1
/* Define if host toolchain supports AVX512F */
#define HAVE_AVX512F 1
/* Define if host toolchain supports AVX512IFMA */
#define HAVE_AVX512IFMA 1
/* Define if host toolchain supports AVX512PF */
#define HAVE_AVX512PF 1
/* Define if host toolchain supports AVX512VBMI */
#define HAVE_AVX512VBMI 1
/* Define if host toolchain supports AVX512VL */
#define HAVE_AVX512VL 1
#endif
/* bdevname() is available */
/* #undef HAVE_BDEVNAME */
/* bdev_check_media_change() exists */
/* #undef HAVE_BDEV_CHECK_MEDIA_CHANGE */
/* bdev_*_io_acct() available */
-/* #undef HAVE_BDEV_IO_ACCT */
+/* #undef HAVE_BDEV_IO_ACCT_63 */
+
+/* bdev_*_io_acct() available */
+/* #undef HAVE_BDEV_IO_ACCT_OLD */
/* bdev_kobj() exists */
/* #undef HAVE_BDEV_KOBJ */
/* bdev_max_discard_sectors() is available */
/* #undef HAVE_BDEV_MAX_DISCARD_SECTORS */
/* bdev_max_secure_erase_sectors() is available */
/* #undef HAVE_BDEV_MAX_SECURE_ERASE_SECTORS */
/* block_device_operations->submit_bio() returns void */
/* #undef HAVE_BDEV_SUBMIT_BIO_RETURNS_VOID */
/* bdev_whole() is available */
/* #undef HAVE_BDEV_WHOLE */
/* bio_alloc() takes 4 arguments */
/* #undef HAVE_BIO_ALLOC_4ARG */
/* bio->bi_bdev->bd_disk exists */
/* #undef HAVE_BIO_BDEV_DISK */
/* bio->bi_opf is defined */
/* #undef HAVE_BIO_BI_OPF */
/* bio->bi_status exists */
/* #undef HAVE_BIO_BI_STATUS */
/* bio has bi_iter */
/* #undef HAVE_BIO_BVEC_ITER */
/* bio_*_io_acct() available */
/* #undef HAVE_BIO_IO_ACCT */
/* bio_max_segs() is implemented */
/* #undef HAVE_BIO_MAX_SEGS */
/* bio_set_dev() is available */
/* #undef HAVE_BIO_SET_DEV */
/* bio_set_dev() GPL-only */
/* #undef HAVE_BIO_SET_DEV_GPL_ONLY */
/* bio_set_dev() is a macro */
/* #undef HAVE_BIO_SET_DEV_MACRO */
/* bio_set_op_attrs is available */
/* #undef HAVE_BIO_SET_OP_ATTRS */
/* blkdev_get_by_path() handles ERESTARTSYS */
/* #undef HAVE_BLKDEV_GET_ERESTARTSYS */
/* blkdev_issue_discard() is available */
/* #undef HAVE_BLKDEV_ISSUE_DISCARD */
/* blkdev_issue_secure_erase() is available */
/* #undef HAVE_BLKDEV_ISSUE_SECURE_ERASE */
/* blkdev_reread_part() exists */
/* #undef HAVE_BLKDEV_REREAD_PART */
/* blkg_tryget() is available */
/* #undef HAVE_BLKG_TRYGET */
/* blkg_tryget() GPL-only */
/* #undef HAVE_BLKG_TRYGET_GPL_ONLY */
/* blk_alloc_disk() exists */
/* #undef HAVE_BLK_ALLOC_DISK */
/* blk_alloc_queue() expects request function */
/* #undef HAVE_BLK_ALLOC_QUEUE_REQUEST_FN */
/* blk_alloc_queue_rh() expects request function */
/* #undef HAVE_BLK_ALLOC_QUEUE_REQUEST_FN_RH */
/* blk_cleanup_disk() exists */
/* #undef HAVE_BLK_CLEANUP_DISK */
/* block multiqueue is available */
/* #undef HAVE_BLK_MQ */
/* blk queue backing_dev_info is dynamic */
/* #undef HAVE_BLK_QUEUE_BDI_DYNAMIC */
/* blk_queue_discard() is available */
/* #undef HAVE_BLK_QUEUE_DISCARD */
/* blk_queue_flag_clear() exists */
/* #undef HAVE_BLK_QUEUE_FLAG_CLEAR */
/* blk_queue_flag_set() exists */
/* #undef HAVE_BLK_QUEUE_FLAG_SET */
/* blk_queue_flush() is available */
/* #undef HAVE_BLK_QUEUE_FLUSH */
/* blk_queue_flush() is GPL-only */
/* #undef HAVE_BLK_QUEUE_FLUSH_GPL_ONLY */
/* blk_queue_secdiscard() is available */
/* #undef HAVE_BLK_QUEUE_SECDISCARD */
/* blk_queue_secure_erase() is available */
/* #undef HAVE_BLK_QUEUE_SECURE_ERASE */
/* blk_queue_update_readahead() exists */
/* #undef HAVE_BLK_QUEUE_UPDATE_READAHEAD */
/* blk_queue_write_cache() exists */
/* #undef HAVE_BLK_QUEUE_WRITE_CACHE */
/* blk_queue_write_cache() is GPL-only */
/* #undef HAVE_BLK_QUEUE_WRITE_CACHE_GPL_ONLY */
/* Define if revalidate_disk() in block_device_operations */
/* #undef HAVE_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK */
/* Define to 1 if you have the Mac OS X function CFLocaleCopyCurrent in the
CoreFoundation framework. */
/* #undef HAVE_CFLOCALECOPYCURRENT */
/* Define to 1 if you have the Mac OS X function
CFLocaleCopyPreferredLanguages in the CoreFoundation framework. */
/* #undef HAVE_CFLOCALECOPYPREFERREDLANGUAGES */
/* Define to 1 if you have the Mac OS X function CFPreferencesCopyAppValue in
the CoreFoundation framework. */
/* #undef HAVE_CFPREFERENCESCOPYAPPVALUE */
/* check_disk_change() exists */
/* #undef HAVE_CHECK_DISK_CHANGE */
/* clear_inode() is available */
/* #undef HAVE_CLEAR_INODE */
/* dentry uses const struct dentry_operations */
/* #undef HAVE_CONST_DENTRY_OPERATIONS */
/* copy_from_iter() is available */
/* #undef HAVE_COPY_FROM_ITER */
/* copy_to_iter() is available */
/* #undef HAVE_COPY_TO_ITER */
/* cpu_has_feature() is GPL-only */
/* #undef HAVE_CPU_HAS_FEATURE_GPL_ONLY */
/* yes */
/* #undef HAVE_CPU_HOTPLUG */
/* current_time() exists */
/* #undef HAVE_CURRENT_TIME */
/* Define if the GNU dcgettext() function is already present or preinstalled.
*/
/* #undef HAVE_DCGETTEXT */
/* DECLARE_EVENT_CLASS() is available */
/* #undef HAVE_DECLARE_EVENT_CLASS */
/* dentry aliases are in d_u member */
/* #undef HAVE_DENTRY_D_U_ALIASES */
/* dequeue_signal() takes 4 arguments */
/* #undef HAVE_DEQUEUE_SIGNAL_4ARG */
/* lookup_bdev() wants dev_t arg */
/* #undef HAVE_DEVT_LOOKUP_BDEV */
/* sops->dirty_inode() wants flags */
/* #undef HAVE_DIRTY_INODE_WITH_FLAGS */
/* disk_*_io_acct() available */
/* #undef HAVE_DISK_IO_ACCT */
/* disk_update_readahead() exists */
/* #undef HAVE_DISK_UPDATE_READAHEAD */
/* Define to 1 if you have the <dlfcn.h> header file. */
#define HAVE_DLFCN_H 1
/* d_make_root() is available */
/* #undef HAVE_D_MAKE_ROOT */
/* d_prune_aliases() is available */
/* #undef HAVE_D_PRUNE_ALIASES */
/* dops->d_revalidate() operation takes nameidata */
/* #undef HAVE_D_REVALIDATE_NAMEIDATA */
/* eops->encode_fh() wants child and parent inodes */
/* #undef HAVE_ENCODE_FH_WITH_INODE */
/* sops->evict_inode() exists */
/* #undef HAVE_EVICT_INODE */
/* FALLOC_FL_ZERO_RANGE is defined */
/* #undef HAVE_FALLOC_FL_ZERO_RANGE */
/* fault_in_iov_iter_readable() is available */
/* #undef HAVE_FAULT_IN_IOV_ITER_READABLE */
/* filemap_range_has_page() is available */
/* #undef HAVE_FILEMAP_RANGE_HAS_PAGE */
/* fops->aio_fsync() exists */
/* #undef HAVE_FILE_AIO_FSYNC */
/* file_dentry() is available */
/* #undef HAVE_FILE_DENTRY */
/* fops->fadvise() exists */
/* #undef HAVE_FILE_FADVISE */
/* file_inode() is available */
/* #undef HAVE_FILE_INODE */
/* flush_dcache_page() is GPL-only */
/* #undef HAVE_FLUSH_DCACHE_PAGE_GPL_ONLY */
/* iops->follow_link() cookie */
/* #undef HAVE_FOLLOW_LINK_COOKIE */
/* iops->follow_link() nameidata */
/* #undef HAVE_FOLLOW_LINK_NAMEIDATA */
/* Define if compiler supports -Wformat-overflow */
/* #undef HAVE_FORMAT_OVERFLOW */
/* fops->fsync() with range */
/* #undef HAVE_FSYNC_RANGE */
/* fops->fsync() without dentry */
/* #undef HAVE_FSYNC_WITHOUT_DENTRY */
/* yes */
/* #undef HAVE_GENERIC_FADVISE */
+/* generic_fillattr requires struct mnt_idmap* */
+/* #undef HAVE_GENERIC_FILLATTR_IDMAP */
+
/* generic_fillattr requires struct user_namespace* */
/* #undef HAVE_GENERIC_FILLATTR_USERNS */
/* generic_*_io_acct() 3 arg available */
/* #undef HAVE_GENERIC_IO_ACCT_3ARG */
/* generic_*_io_acct() 4 arg available */
/* #undef HAVE_GENERIC_IO_ACCT_4ARG */
/* generic_readlink is global */
/* #undef HAVE_GENERIC_READLINK */
/* generic_setxattr() exists */
/* #undef HAVE_GENERIC_SETXATTR */
/* generic_write_checks() takes kiocb */
/* #undef HAVE_GENERIC_WRITE_CHECKS_KIOCB */
/* Define if the GNU gettext() function is already present or preinstalled. */
/* #undef HAVE_GETTEXT */
/* iops->get_acl() exists */
/* #undef HAVE_GET_ACL */
/* iops->get_acl() takes rcu */
/* #undef HAVE_GET_ACL_RCU */
/* has iops->get_inode_acl() */
/* #undef HAVE_GET_INODE_ACL */
/* iops->get_link() cookie */
/* #undef HAVE_GET_LINK_COOKIE */
/* iops->get_link() delayed */
/* #undef HAVE_GET_LINK_DELAYED */
/* group_info->gid exists */
/* #undef HAVE_GROUP_INFO_GID */
/* has_capability() is available */
/* #undef HAVE_HAS_CAPABILITY */
/* iattr->ia_vfsuid and iattr->ia_vfsgid exist */
/* #undef HAVE_IATTR_VFSID */
/* Define if you have the iconv() function and it works. */
#define HAVE_ICONV 1
+/* iops->getattr() takes struct mnt_idmap* */
+/* #undef HAVE_IDMAP_IOPS_GETATTR */
+
+/* iops->setattr() takes struct mnt_idmap* */
+/* #undef HAVE_IDMAP_IOPS_SETATTR */
+
/* APIs for idmapped mount are present */
/* #undef HAVE_IDMAP_MNT_API */
/* Define if compiler supports -Wimplicit-fallthrough */
/* #undef HAVE_IMPLICIT_FALLTHROUGH */
/* Define if compiler supports -Winfinite-recursion */
/* #undef HAVE_INFINITE_RECURSION */
/* yes */
/* #undef HAVE_INODE_LOCK_SHARED */
/* inode_owner_or_capable() exists */
/* #undef HAVE_INODE_OWNER_OR_CAPABLE */
+/* inode_owner_or_capable() takes mnt_idmap */
+/* #undef HAVE_INODE_OWNER_OR_CAPABLE_IDMAP */
+
/* inode_owner_or_capable() takes user_ns */
-/* #undef HAVE_INODE_OWNER_OR_CAPABLE_IDMAPPED */
+/* #undef HAVE_INODE_OWNER_OR_CAPABLE_USERNS */
/* inode_set_flags() exists */
/* #undef HAVE_INODE_SET_FLAGS */
/* inode_set_iversion() exists */
/* #undef HAVE_INODE_SET_IVERSION */
/* inode->i_*time's are timespec64 */
/* #undef HAVE_INODE_TIMESPEC64_TIMES */
/* timestamp_truncate() exists */
/* #undef HAVE_INODE_TIMESTAMP_TRUNCATE */
/* Define to 1 if you have the <inttypes.h> header file. */
#define HAVE_INTTYPES_H 1
/* in_compat_syscall() is available */
/* #undef HAVE_IN_COMPAT_SYSCALL */
+/* iops->create() takes struct mnt_idmap* */
+/* #undef HAVE_IOPS_CREATE_IDMAP */
+
/* iops->create() takes struct user_namespace* */
/* #undef HAVE_IOPS_CREATE_USERNS */
+/* iops->mkdir() takes struct mnt_idmap* */
+/* #undef HAVE_IOPS_MKDIR_IDMAP */
+
/* iops->mkdir() takes struct user_namespace* */
/* #undef HAVE_IOPS_MKDIR_USERNS */
+/* iops->mknod() takes struct mnt_idmap* */
+/* #undef HAVE_IOPS_MKNOD_IDMAP */
+
/* iops->mknod() takes struct user_namespace* */
/* #undef HAVE_IOPS_MKNOD_USERNS */
+/* iops->permission() takes struct mnt_idmap* */
+/* #undef HAVE_IOPS_PERMISSION_IDMAP */
+
/* iops->permission() takes struct user_namespace* */
/* #undef HAVE_IOPS_PERMISSION_USERNS */
+/* iops->rename() takes struct mnt_idmap* */
+/* #undef HAVE_IOPS_RENAME_IDMAP */
+
/* iops->rename() takes struct user_namespace* */
/* #undef HAVE_IOPS_RENAME_USERNS */
+/* iops->setattr() exists */
+/* #undef HAVE_IOPS_SETATTR */
+
+/* iops->symlink() takes struct mnt_idmap* */
+/* #undef HAVE_IOPS_SYMLINK_IDMAP */
+
/* iops->symlink() takes struct user_namespace* */
/* #undef HAVE_IOPS_SYMLINK_USERNS */
/* iov_iter_advance() is available */
/* #undef HAVE_IOV_ITER_ADVANCE */
/* iov_iter_count() is available */
/* #undef HAVE_IOV_ITER_COUNT */
/* iov_iter_fault_in_readable() is available */
/* #undef HAVE_IOV_ITER_FAULT_IN_READABLE */
/* iov_iter_revert() is available */
/* #undef HAVE_IOV_ITER_REVERT */
/* iov_iter_type() is available */
/* #undef HAVE_IOV_ITER_TYPE */
/* iov_iter types are available */
/* #undef HAVE_IOV_ITER_TYPES */
/* yes */
/* #undef HAVE_IO_SCHEDULE_TIMEOUT */
/* Define to 1 if you have the `issetugid' function. */
#define HAVE_ISSETUGID 1
/* kernel has kernel_fpu_* functions */
/* #undef HAVE_KERNEL_FPU */
/* kernel has asm/fpu/api.h */
/* #undef HAVE_KERNEL_FPU_API_HEADER */
/* kernel fpu internal */
/* #undef HAVE_KERNEL_FPU_INTERNAL */
/* kernel has asm/fpu/internal.h */
/* #undef HAVE_KERNEL_FPU_INTERNAL_HEADER */
/* uncached_acl_sentinel() exists */
/* #undef HAVE_KERNEL_GET_ACL_HANDLE_CACHE */
/* Define if compiler supports -Winfinite-recursion */
/* #undef HAVE_KERNEL_INFINITE_RECURSION */
/* kernel does stack verification */
/* #undef HAVE_KERNEL_OBJTOOL */
/* kernel has linux/objtool.h */
/* #undef HAVE_KERNEL_OBJTOOL_HEADER */
/* kernel_read() take loff_t pointer */
/* #undef HAVE_KERNEL_READ_PPOS */
/* timer_list.function gets a timer_list */
/* #undef HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST */
/* struct timer_list has a flags member */
/* #undef HAVE_KERNEL_TIMER_LIST_FLAGS */
/* timer_setup() is available */
/* #undef HAVE_KERNEL_TIMER_SETUP */
/* kernel_write() take loff_t pointer */
/* #undef HAVE_KERNEL_WRITE_PPOS */
/* kmem_cache_create_usercopy() exists */
/* #undef HAVE_KMEM_CACHE_CREATE_USERCOPY */
/* kstrtoul() exists */
/* #undef HAVE_KSTRTOUL */
/* ktime_get_coarse_real_ts64() exists */
/* #undef HAVE_KTIME_GET_COARSE_REAL_TS64 */
/* ktime_get_raw_ts64() exists */
/* #undef HAVE_KTIME_GET_RAW_TS64 */
/* kvmalloc exists */
/* #undef HAVE_KVMALLOC */
/* Define if you have [aio] */
/* #undef HAVE_LIBAIO */
/* Define if you have [blkid] */
/* #undef HAVE_LIBBLKID */
/* Define if you have [crypto] */
#define HAVE_LIBCRYPTO 1
/* Define if you have [tirpc] */
/* #undef HAVE_LIBTIRPC */
/* Define if you have [udev] */
/* #undef HAVE_LIBUDEV */
/* Define if you have [uuid] */
/* #undef HAVE_LIBUUID */
/* linux/blk-cgroup.h exists */
/* #undef HAVE_LINUX_BLK_CGROUP_HEADER */
/* lseek_execute() is available */
/* #undef HAVE_LSEEK_EXECUTE */
/* makedev() is declared in sys/mkdev.h */
/* #undef HAVE_MAKEDEV_IN_MKDEV */
/* makedev() is declared in sys/sysmacros.h */
/* #undef HAVE_MAKEDEV_IN_SYSMACROS */
/* Noting that make_request_fn() returns blk_qc_t */
/* #undef HAVE_MAKE_REQUEST_FN_RET_QC */
/* Noting that make_request_fn() returns void */
/* #undef HAVE_MAKE_REQUEST_FN_RET_VOID */
/* iops->mkdir() takes umode_t */
/* #undef HAVE_MKDIR_UMODE_T */
/* Define to 1 if you have the `mlockall' function. */
#define HAVE_MLOCKALL 1
/* lookup_bdev() wants mode arg */
/* #undef HAVE_MODE_LOOKUP_BDEV */
/* Define if host toolchain supports MOVBE */
#define HAVE_MOVBE 1
/* new_sync_read()/new_sync_write() are available */
/* #undef HAVE_NEW_SYNC_READ */
/* folio_wait_bit() exists */
/* #undef HAVE_PAGEMAP_FOLIO_WAIT_BIT */
/* part_to_dev() exists */
/* #undef HAVE_PART_TO_DEV */
/* iops->getattr() takes a path */
/* #undef HAVE_PATH_IOPS_GETATTR */
/* Define if host toolchain supports PCLMULQDQ */
#define HAVE_PCLMULQDQ 1
/* percpu_counter_add_batch() is defined */
/* #undef HAVE_PERCPU_COUNTER_ADD_BATCH */
/* percpu_counter_init() wants gfp_t */
/* #undef HAVE_PERCPU_COUNTER_INIT_WITH_GFP */
/* posix_acl_chmod() exists */
/* #undef HAVE_POSIX_ACL_CHMOD */
/* posix_acl_from_xattr() needs user_ns */
/* #undef HAVE_POSIX_ACL_FROM_XATTR_USERNS */
/* posix_acl_release() is available */
/* #undef HAVE_POSIX_ACL_RELEASE */
/* posix_acl_release() is GPL-only */
/* #undef HAVE_POSIX_ACL_RELEASE_GPL_ONLY */
/* posix_acl_valid() wants user namespace */
/* #undef HAVE_POSIX_ACL_VALID_WITH_NS */
/* proc_ops structure exists */
/* #undef HAVE_PROC_OPS_STRUCT */
/* iops->put_link() cookie */
/* #undef HAVE_PUT_LINK_COOKIE */
/* iops->put_link() delayed */
/* #undef HAVE_PUT_LINK_DELAYED */
/* iops->put_link() nameidata */
/* #undef HAVE_PUT_LINK_NAMEIDATA */
/* If available, contains the Python version number currently in use. */
#define HAVE_PYTHON "3.7"
/* qat is enabled and existed */
/* #undef HAVE_QAT */
/* register_shrinker is vararg */
/* #undef HAVE_REGISTER_SHRINKER_VARARG */
/* iops->rename2() exists */
/* #undef HAVE_RENAME2 */
/* struct inode_operations_wrapper takes .rename2() */
/* #undef HAVE_RENAME2_OPERATIONS_WRAPPER */
/* iops->rename() wants flags */
/* #undef HAVE_RENAME_WANTS_FLAGS */
/* REQ_DISCARD is defined */
/* #undef HAVE_REQ_DISCARD */
/* REQ_FLUSH is defined */
/* #undef HAVE_REQ_FLUSH */
/* REQ_OP_DISCARD is defined */
/* #undef HAVE_REQ_OP_DISCARD */
/* REQ_OP_FLUSH is defined */
/* #undef HAVE_REQ_OP_FLUSH */
/* REQ_OP_SECURE_ERASE is defined */
/* #undef HAVE_REQ_OP_SECURE_ERASE */
/* REQ_PREFLUSH is defined */
/* #undef HAVE_REQ_PREFLUSH */
/* revalidate_disk() is available */
/* #undef HAVE_REVALIDATE_DISK */
/* revalidate_disk_size() is available */
/* #undef HAVE_REVALIDATE_DISK_SIZE */
/* struct rw_semaphore has member activity */
/* #undef HAVE_RWSEM_ACTIVITY */
/* struct rw_semaphore has atomic_long_t member count */
/* #undef HAVE_RWSEM_ATOMIC_LONG_COUNT */
/* linux/sched/signal.h exists */
/* #undef HAVE_SCHED_SIGNAL_HEADER */
/* Define to 1 if you have the <security/pam_modules.h> header file. */
#define HAVE_SECURITY_PAM_MODULES_H 1
+/* setattr_prepare() accepts mnt_idmap */
+/* #undef HAVE_SETATTR_PREPARE_IDMAP */
+
/* setattr_prepare() is available, doesn't accept user_namespace */
/* #undef HAVE_SETATTR_PREPARE_NO_USERNS */
/* setattr_prepare() accepts user_namespace */
/* #undef HAVE_SETATTR_PREPARE_USERNS */
/* iops->set_acl() exists, takes 3 args */
/* #undef HAVE_SET_ACL */
+/* iops->set_acl() takes 4 args, arg1 is struct mnt_idmap * */
+/* #undef HAVE_SET_ACL_IDMAP_DENTRY */
+
/* iops->set_acl() takes 4 args */
/* #undef HAVE_SET_ACL_USERNS */
/* iops->set_acl() takes 4 args, arg2 is struct dentry * */
/* #undef HAVE_SET_ACL_USERNS_DENTRY_ARG2 */
/* set_cached_acl() is usable */
/* #undef HAVE_SET_CACHED_ACL_USABLE */
/* set_special_state() exists */
/* #undef HAVE_SET_SPECIAL_STATE */
/* struct shrink_control exists */
/* #undef HAVE_SHRINK_CONTROL_STRUCT */
/* kernel_siginfo_t exists */
/* #undef HAVE_SIGINFO */
/* signal_stop() exists */
/* #undef HAVE_SIGNAL_STOP */
/* new shrinker callback wants 2 args */
/* #undef HAVE_SINGLE_SHRINKER_CALLBACK */
/* cs->count_objects exists */
/* #undef HAVE_SPLIT_SHRINKER_CALLBACK */
#if defined(__amd64__) || defined(__i386__)
/* Define if host toolchain supports SSE */
#define HAVE_SSE 1
/* Define if host toolchain supports SSE2 */
#define HAVE_SSE2 1
/* Define if host toolchain supports SSE3 */
#define HAVE_SSE3 1
/* Define if host toolchain supports SSE4.1 */
#define HAVE_SSE4_1 1
/* Define if host toolchain supports SSE4.2 */
#define HAVE_SSE4_2 1
/* Define if host toolchain supports SSSE3 */
#define HAVE_SSSE3 1
#endif
/* STACK_FRAME_NON_STANDARD is defined */
/* #undef HAVE_STACK_FRAME_NON_STANDARD */
/* standalone <linux/stdarg.h> exists */
/* #undef HAVE_STANDALONE_LINUX_STDARG */
/* Define to 1 if you have the <stdint.h> header file. */
#define HAVE_STDINT_H 1
/* Define to 1 if you have the <stdio.h> header file. */
#define HAVE_STDIO_H 1
/* Define to 1 if you have the <stdlib.h> header file. */
#define HAVE_STDLIB_H 1
/* Define to 1 if you have the <strings.h> header file. */
#define HAVE_STRINGS_H 1
/* Define to 1 if you have the <string.h> header file. */
#define HAVE_STRING_H 1
/* Define to 1 if you have the `strlcat' function. */
#define HAVE_STRLCAT 1
/* Define to 1 if you have the `strlcpy' function. */
#define HAVE_STRLCPY 1
/* submit_bio is member of struct block_device_operations */
/* #undef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS */
/* super_setup_bdi_name() exits */
/* #undef HAVE_SUPER_SETUP_BDI_NAME */
/* super_block->s_user_ns exists */
/* #undef HAVE_SUPER_USER_NS */
/* struct kobj_type has default_groups */
/* #undef HAVE_SYSFS_DEFAULT_GROUPS */
/* Define to 1 if you have the <sys/stat.h> header file. */
#define HAVE_SYS_STAT_H 1
/* Define to 1 if you have the <sys/types.h> header file. */
#define HAVE_SYS_TYPES_H 1
/* i_op->tmpfile() exists */
/* #undef HAVE_TMPFILE */
/* i_op->tmpfile() uses old dentry signature */
/* #undef HAVE_TMPFILE_DENTRY */
+/* i_op->tmpfile() has mnt_idmap */
+/* #undef HAVE_TMPFILE_IDMAP */
+
/* i_op->tmpfile() has userns */
/* #undef HAVE_TMPFILE_USERNS */
/* totalhigh_pages() exists */
/* #undef HAVE_TOTALHIGH_PAGES */
/* kernel has totalram_pages() */
/* #undef HAVE_TOTALRAM_PAGES_FUNC */
/* Define to 1 if you have the `udev_device_get_is_initialized' function. */
/* #undef HAVE_UDEV_DEVICE_GET_IS_INITIALIZED */
/* kernel has __kernel_fpu_* functions */
/* #undef HAVE_UNDERSCORE_KERNEL_FPU */
/* Define to 1 if you have the <unistd.h> header file. */
#define HAVE_UNISTD_H 1
/* iops->getattr() takes struct user_namespace* */
/* #undef HAVE_USERNS_IOPS_GETATTR */
+/* iops->setattr() takes struct user_namespace* */
+/* #undef HAVE_USERNS_IOPS_SETATTR */
+
/* user_namespace->ns.inum exists */
/* #undef HAVE_USER_NS_COMMON_INUM */
/* iops->getattr() takes a vfsmount */
/* #undef HAVE_VFSMOUNT_IOPS_GETATTR */
/* aops->direct_IO() uses iovec */
/* #undef HAVE_VFS_DIRECT_IO_IOVEC */
/* aops->direct_IO() uses iov_iter without rw */
/* #undef HAVE_VFS_DIRECT_IO_ITER */
/* aops->direct_IO() uses iov_iter with offset */
/* #undef HAVE_VFS_DIRECT_IO_ITER_OFFSET */
/* aops->direct_IO() uses iov_iter with rw and offset */
/* #undef HAVE_VFS_DIRECT_IO_ITER_RW_OFFSET */
/* filemap_dirty_folio exists */
/* #undef HAVE_VFS_FILEMAP_DIRTY_FOLIO */
/* All required iov_iter interfaces are available */
/* #undef HAVE_VFS_IOV_ITER */
/* fops->iterate() is available */
/* #undef HAVE_VFS_ITERATE */
/* fops->iterate_shared() is available */
/* #undef HAVE_VFS_ITERATE_SHARED */
/* fops->readdir() is available */
/* #undef HAVE_VFS_READDIR */
/* address_space_operations->readpages exists */
/* #undef HAVE_VFS_READPAGES */
/* read_folio exists */
/* #undef HAVE_VFS_READ_FOLIO */
/* fops->read/write_iter() are available */
/* #undef HAVE_VFS_RW_ITERATE */
/* __set_page_dirty_nobuffers exists */
/* #undef HAVE_VFS_SET_PAGE_DIRTY_NOBUFFERS */
/* __vmalloc page flags exists */
/* #undef HAVE_VMALLOC_PAGE_KERNEL */
/* yes */
/* #undef HAVE_WAIT_ON_BIT_ACTION */
/* wait_queue_entry_t exists */
/* #undef HAVE_WAIT_QUEUE_ENTRY_T */
/* wq_head->head and wq_entry->entry exist */
/* #undef HAVE_WAIT_QUEUE_HEAD_ENTRY */
+/* int (*writepage_t)() takes struct folio* */
+/* #undef HAVE_WRITEPAGE_T_FOLIO */
+
/* xattr_handler->get() wants dentry */
/* #undef HAVE_XATTR_GET_DENTRY */
/* xattr_handler->get() wants both dentry and inode */
/* #undef HAVE_XATTR_GET_DENTRY_INODE */
/* xattr_handler->get() wants dentry and inode and flags */
/* #undef HAVE_XATTR_GET_DENTRY_INODE_FLAGS */
/* xattr_handler->get() wants xattr_handler */
/* #undef HAVE_XATTR_GET_HANDLER */
/* xattr_handler has name */
/* #undef HAVE_XATTR_HANDLER_NAME */
/* xattr_handler->list() wants dentry */
/* #undef HAVE_XATTR_LIST_DENTRY */
/* xattr_handler->list() wants xattr_handler */
/* #undef HAVE_XATTR_LIST_HANDLER */
/* xattr_handler->list() wants simple */
/* #undef HAVE_XATTR_LIST_SIMPLE */
/* xattr_handler->set() wants dentry */
/* #undef HAVE_XATTR_SET_DENTRY */
/* xattr_handler->set() wants both dentry and inode */
/* #undef HAVE_XATTR_SET_DENTRY_INODE */
/* xattr_handler->set() wants xattr_handler */
/* #undef HAVE_XATTR_SET_HANDLER */
+/* xattr_handler->set() takes mnt_idmap */
+/* #undef HAVE_XATTR_SET_IDMAP */
+
/* xattr_handler->set() takes user_namespace */
/* #undef HAVE_XATTR_SET_USERNS */
/* Define if host toolchain supports XSAVE */
#define HAVE_XSAVE 1
/* Define if host toolchain supports XSAVEOPT */
#define HAVE_XSAVEOPT 1
/* Define if host toolchain supports XSAVES */
#define HAVE_XSAVES 1
/* ZERO_PAGE() is GPL-only */
/* #undef HAVE_ZERO_PAGE_GPL_ONLY */
/* Define if you have [z] */
#define HAVE_ZLIB 1
/* __posix_acl_chmod() exists */
/* #undef HAVE___POSIX_ACL_CHMOD */
/* kernel exports FPU functions */
/* #undef KERNEL_EXPORTS_X86_FPU */
/* TBD: fetch(3) support */
#if 0
/* whether the chosen libfetch is to be loaded at run-time */
#define LIBFETCH_DYNAMIC 1
/* libfetch is fetch(3) */
#define LIBFETCH_IS_FETCH 1
/* libfetch is libcurl */
#define LIBFETCH_IS_LIBCURL 0
/* soname of chosen libfetch */
#define LIBFETCH_SONAME "libfetch.so.6"
#endif
/* Define to the sub-directory where libtool stores uninstalled libraries. */
#define LT_OBJDIR ".libs/"
/* make_request_fn() return type */
/* #undef MAKE_REQUEST_FN_RET */
/* struct shrink_control has nid */
/* #undef SHRINK_CONTROL_HAS_NID */
/* using complete_and_exit() instead */
/* #undef SPL_KTHREAD_COMPLETE_AND_EXIT */
/* Defined for legacy compatibility. */
#define SPL_META_ALIAS ZFS_META_ALIAS
/* Defined for legacy compatibility. */
#define SPL_META_RELEASE ZFS_META_RELEASE
/* Defined for legacy compatibility. */
#define SPL_META_VERSION ZFS_META_VERSION
/* pde_data() is PDE_DATA() */
/* #undef SPL_PDE_DATA */
/* Define to 1 if all of the C90 standard headers exist (not just the ones
required in a freestanding environment). This macro is provided for
backward compatibility; new code need not use it. */
#define SYSTEM_FREEBSD 1
/* True if ZFS is to be compiled for a Linux system */
/* #undef SYSTEM_LINUX */
/* Version number of package */
/* #undef ZFS_DEBUG */
/* /dev/zfs minor */
/* #undef ZFS_DEVICE_MINOR */
/* enum node_stat_item contains NR_FILE_PAGES */
/* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_FILE_PAGES */
/* enum node_stat_item contains NR_INACTIVE_ANON */
/* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_INACTIVE_ANON */
/* enum node_stat_item contains NR_INACTIVE_FILE */
/* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_INACTIVE_FILE */
/* enum zone_stat_item contains NR_FILE_PAGES */
/* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_FILE_PAGES */
/* enum zone_stat_item contains NR_INACTIVE_ANON */
/* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_INACTIVE_ANON */
/* enum zone_stat_item contains NR_INACTIVE_FILE */
/* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_INACTIVE_FILE */
/* GENHD_FL_EXT_DEVT flag is not available */
/* #undef ZFS_GENHD_FL_EXT_DEVT */
/* GENHD_FL_NO_PART_SCAN flag is available */
/* #undef ZFS_GENHD_FL_NO_PART */
/* global_node_page_state() exists */
/* #undef ZFS_GLOBAL_NODE_PAGE_STATE */
/* global_zone_page_state() exists */
/* #undef ZFS_GLOBAL_ZONE_PAGE_STATE */
/* Define to 1 if GPL-only symbols can be used */
/* #undef ZFS_IS_GPL_COMPATIBLE */
/* Define the project alias string. */
-#define ZFS_META_ALIAS "zfs-2.1.99-FreeBSD_g431083f75"
+#define ZFS_META_ALIAS "zfs-2.1.99-FreeBSD_gd96e29576"
/* Define the project author. */
#define ZFS_META_AUTHOR "OpenZFS"
/* Define the project release date. */
/* #undef ZFS_META_DATA */
/* Define the maximum compatible kernel version. */
-#define ZFS_META_KVER_MAX "6.1"
+#define ZFS_META_KVER_MAX "6.2"
/* Define the minimum compatible kernel version. */
#define ZFS_META_KVER_MIN "3.10"
/* Define the project license. */
#define ZFS_META_LICENSE "CDDL"
/* Define the libtool library 'age' version information. */
/* #undef ZFS_META_LT_AGE */
/* Define the libtool library 'current' version information. */
/* #undef ZFS_META_LT_CURRENT */
/* Define the libtool library 'revision' version information. */
/* #undef ZFS_META_LT_REVISION */
/* Define the project name. */
#define ZFS_META_NAME "zfs"
/* Define the project release. */
-#define ZFS_META_RELEASE "FreeBSD_g431083f75"
+#define ZFS_META_RELEASE "FreeBSD_gd96e29576"
/* Define the project version. */
#define ZFS_META_VERSION "2.1.99"
/* count is located in percpu_ref.data */
/* #undef ZFS_PERCPU_REF_COUNT_IN_DATA */
diff --git a/sys/modules/zfs/zfs_gitrev.h b/sys/modules/zfs/zfs_gitrev.h
index c30d690c0ad8..e0a986d41c02 100644
--- a/sys/modules/zfs/zfs_gitrev.h
+++ b/sys/modules/zfs/zfs_gitrev.h
@@ -1 +1 @@
-#define ZFS_META_GITREV "zfs-2.1.99-1860-g431083f75"
+#define ZFS_META_GITREV "zfs-2.1.99-1919-gd96e29576"

File Metadata

Mime Type
application/octet-stream
Expires
Sun, May 19, 9:46 AM (2 d)
Storage Engine
chunks
Storage Format
Chunks
Storage Handle
pPYRYczSQYv9
Default Alt Text
(5 MB)

Event Timeline