Index: vendor/illumos/dist/cmd/zpool/zpool_main.c =================================================================== --- vendor/illumos/dist/cmd/zpool/zpool_main.c (revision 354960) +++ vendor/illumos/dist/cmd/zpool/zpool_main.c (revision 354961) @@ -1,6550 +1,6595 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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) 2012 by Frederik Wessels. All rights reserved. * Copyright (c) 2013 by Prasad Joshi (sTec). All rights reserved. * Copyright 2016 Igor Kozhukhov . * Copyright 2016 Nexenta Systems, Inc. * Copyright (c) 2017 Datto Inc. * Copyright (c) 2017, Intel Corporation. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zpool_util.h" #include "zfs_comutil.h" #include "zfeature_common.h" #include "statcommon.h" 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_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_get(int, char **); static int zpool_do_set(int, char **); static int zpool_do_sync(int, char **); /* * 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_STATUS, HELP_UPGRADE, HELP_GET, HELP_SET, HELP_SPLIT, HELP_SYNC, HELP_REGUID, HELP_REOPEN } zpool_help_t; 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[] = { { "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 }, { 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 }, { "get", zpool_do_get, HELP_GET }, { "set", zpool_do_set, HELP_SET }, { "sync", zpool_do_sync, HELP_SYNC }, }; #define NCOMMAND (sizeof (command_table) / sizeof (command_table[0])) #define VDEV_ALLOC_CLASS_LOGS "logs" static zpool_command_t *current_command; 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] ...\n")); case HELP_ATTACH: return (gettext("\tattach [-f] " "\n")); case HELP_CLEAR: return (gettext("\tclear [-nF] [device]\n")); case HELP_CREATE: return (gettext("\tcreate [-fnd] [-B] " "[-o property=value] ... \n" "\t [-O file-system-property=value] ...\n" "\t [-m mountpoint] [-R root] [-t tempname] " " ...\n")); case HELP_CHECKPOINT: return (gettext("\tcheckpoint [--discard] ...\n")); case HELP_DESTROY: return (gettext("\tdestroy [-f] \n")); case HELP_DETACH: return (gettext("\tdetach \n")); case HELP_EXPORT: return (gettext("\texport [-f] ...\n")); case HELP_HISTORY: return (gettext("\thistory [-il] [] ...\n")); case HELP_IMPORT: return (gettext("\timport [-d dir] [-D]\n" "\timport [-o mntopts] [-o property=value] ... \n" "\t [-d dir | -c cachefile] [-D] [-f] [-m] [-N] " "[-R root] [-F [-n]] -a\n" "\timport [-o mntopts] [-o property=value] ... \n" "\t [-d dir | -c cachefile] [-D] [-f] [-m] [-N] " "[-R root] [-F [-n]] [-t]\n" "\t [--rewind-to-checkpoint] [newpool]\n")); case HELP_IOSTAT: return (gettext("\tiostat [-gLPv] [-T d|u] [pool] ... " "[interval [count]]\n")); case HELP_LABELCLEAR: return (gettext("\tlabelclear [-f] \n")); case HELP_LIST: return (gettext("\tlist [-gHLpPv] [-o property[,...]] " "[-T d|u] [pool] ... [interval [count]]\n")); case HELP_OFFLINE: return (gettext("\toffline [-t] ...\n")); case HELP_ONLINE: return (gettext("\tonline ...\n")); case HELP_REPLACE: return (gettext("\treplace [-f] " "[new-device]\n")); case HELP_REMOVE: return (gettext("\tremove [-nps] ...\n")); case HELP_REOPEN: return (gettext("\treopen \n")); case HELP_INITIALIZE: return (gettext("\tinitialize [-cs] [ ...]\n")); case HELP_SCRUB: return (gettext("\tscrub [-s | -p] ...\n")); + case HELP_RESILVER: + return (gettext("\tresilver ...\n")); case HELP_STATUS: return (gettext("\tstatus [-DgLPvx] [-T d|u] [pool] ... " "[interval [count]]\n")); case HELP_UPGRADE: return (gettext("\tupgrade\n" "\tupgrade -v\n" "\tupgrade [-V version] <-a | pool ...>\n")); case HELP_GET: return (gettext("\tget [-Hp] [-o \"all\" | field[,...]] " "<\"all\" | property[,...]> ...\n")); case HELP_SET: return (gettext("\tset \n")); case HELP_SPLIT: return (gettext("\tsplit [-gLnP] [-R altroot] [-o mntopts]\n" "\t [-o property=value] " "[ ...]\n")); case HELP_REGUID: return (gettext("\treguid \n")); case HELP_SYNC: return (gettext("\tsync [pool] ...\n")); } abort(); /* NOTREACHED */ } /* * Callback routine that will print out a pool property value. */ static int print_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); } /* * 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. */ 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 && ((strcmp(current_command->name, "set") == 0) || (strcmp(current_command->name, "get") == 0) || (strcmp(current_command->name, "list") == 0))) { (void) fprintf(fp, gettext("\nthe following properties are supported:\n")); (void) fprintf(fp, "\n\t%-19s %s %s\n\n", "PROPERTY", "EDIT", "VALUES"); /* Iterate over all properties */ (void) zprop_iter(print_prop_cb, fp, B_FALSE, B_TRUE, ZFS_TYPE_POOL); (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.\nSee zpool-features(5).\n")); } /* * See comments at end of main(). */ if (getenv("ZFS_ABORT") != NULL) { (void) printf("dumping core by request\n"); abort(); } exit(requested ? 0 : 2); } /* * 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; char *class = ""; (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); } } 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, char *propval, nvlist_t **props, boolean_t poolprop) { zpool_prop_t prop = ZPROP_INVAL; zfs_prop_t fprop; nvlist_t *proplist; const char *normnm; 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); if ((prop = zpool_name_to_prop(propname)) == ZPROP_INVAL && !zpool_prop_feature(propname)) { (void) fprintf(stderr, gettext("property '%s' is " "not a valid pool 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 (zpool_prop_feature(propname)) normnm = propname; else normnm = zpool_prop_to_name(prop); } else { if ((fprop = zfs_name_to_prop(propname)) != ZPROP_INVAL) { normnm = zfs_prop_to_name(fprop); } else { normnm = propname; } } 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, char *propval, nvlist_t **props, boolean_t poolprop) { char *pval; if (nvlist_lookup_string(*props, propname, &pval) == 0) return (0); return (add_prop_list(propname, propval, props, poolprop)); } /* * zpool add [-fgLnP] [-o property=value] ... * * -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. * -P Display full path for vdev name. * * Adds the given vdevs to 'pool'. As with create, the bulk of this work is * handled by get_vdev_spec(), 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; zpool_boot_label_t boot_type; uint64_t boot_size; int ret; zpool_handle_t *zhp; nvlist_t *config; /* check options */ while ((c = getopt(argc, argv, "fgLnP")) != -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 '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); } if (zpool_is_bootable(zhp)) boot_type = ZPOOL_COPY_BOOT_LABEL; else boot_type = ZPOOL_NO_BOOT_LABEL; /* pass off to get_vdev_spec for processing */ boot_size = zpool_get_prop_int(zhp, ZPOOL_PROP_BOOTSIZE, NULL); nvroot = make_root_vdev(zhp, force, !force, B_FALSE, dryrun, boot_type, boot_size, argc, argv); if (nvroot == NULL) { zpool_close(zhp); return (1); } if (dryrun) { nvlist_t *poolnvroot; verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &poolnvroot) == 0); (void) printf(gettext("would update '%s' to the following " "configuration:\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' */ 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); 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); 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); ret = 0; } else { ret = (zpool_add(zhp, nvroot) != 0); } nvlist_free(nvroot); zpool_close(zhp); return (ret); } /* * zpool remove ... * * 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; boolean_t stop = B_FALSE; boolean_t noop = B_FALSE; boolean_t parsable = B_FALSE; char c; /* check options */ while ((c = getopt(argc, argv, "nps")) != -1) { switch (c) { case 'n': noop = B_TRUE; break; case 'p': parsable = B_TRUE; break; case 's': stop = 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) { (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; } 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], 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; } } } return (ret); } /* * zpool labelclear [-f] * * -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; struct stat st; int c, fd, 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); } /* * 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 dsk path and appending s0. */ (void) strlcpy(vdev, argv[0], sizeof (vdev)); if (vdev[0] != '/' && stat(vdev, &st) != 0) { char *s; (void) snprintf(vdev, sizeof (vdev), "%s/%s", ZFS_DISK_ROOT, argv[0]); if ((s = strrchr(argv[0], 's')) == NULL || !isdigit(*(s + 1))) (void) strlcat(vdev, "s0", sizeof (vdev)); if (stat(vdev, &st) != 0) { (void) fprintf(stderr, gettext( "failed to find device %s, try specifying absolute " "path instead\n"), argv[0]); return (1); } } if ((fd = open(vdev, O_RDWR)) < 0) { (void) fprintf(stderr, gettext("failed to open %s: %s\n"), vdev, strerror(errno)); return (1); } if (zpool_read_label(fd, &config) != 0) { (void) fprintf(stderr, gettext("failed to read label from %s\n"), vdev); return (1); } 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); return (1); } if (!inuse) goto wipe_label; switch (state) { default: case POOL_STATE_ACTIVE: case POOL_STATE_SPARE: case POOL_STATE_L2CACHE: (void) fprintf(stderr, gettext( "%s is a member (%s) of pool \"%s\"\n"), vdev, zpool_pool_state_to_name(state), name); 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] [-B] [-o property=value] ... * [-O file-system-property=value] ... * [-R root] [-m mountpoint] [-t tempname] ... * * -B Create boot partition. * -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 * '/' * -t Use the temporary name until the pool is exported. * -o Set property=value. * -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 get_vdev_spec() in zpool_vdev.c. Once * we get the nvlist back from get_vdev_spec(), we either print out the contents * (if '-n' was specified), or pass it to libzfs to do the creation. */ #define SYSTEM256 (256 * 1024 * 1024) int zpool_do_create(int argc, char **argv) { boolean_t force = B_FALSE; boolean_t dryrun = B_FALSE; boolean_t enable_all_pool_feat = B_TRUE; zpool_boot_label_t boot_type = ZPOOL_NO_BOOT_LABEL; uint64_t boot_size = 0; int c; nvlist_t *nvroot = NULL; char *poolname; char *tname = NULL; int ret = 1; char *altroot = NULL; char *mountpoint = NULL; nvlist_t *fsprops = NULL; nvlist_t *props = NULL; char *propval; /* check options */ while ((c = getopt(argc, argv, ":fndBR:m:o:O:t:")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case 'n': dryrun = B_TRUE; break; case 'd': enable_all_pool_feat = B_FALSE; break; case 'B': /* * We should create the system partition. * Also make sure the size is set. */ boot_type = ZPOOL_CREATE_BOOT_LABEL; if (boot_size == 0) boot_size = SYSTEM256; 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, B_TRUE)) 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; /* * Get bootsize value for make_root_vdev(). */ if (zpool_name_to_prop(optarg) == ZPOOL_PROP_BOOTSIZE) { if (zfs_nicestrtonum(g_zfs, propval, &boot_size) < 0 || boot_size == 0) { (void) fprintf(stderr, gettext("bad boot partition size " "'%s': %s\n"), propval, libzfs_error_description(g_zfs)); 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_all_pool_feat = B_FALSE; } } if (zpool_name_to_prop(optarg) == ZPOOL_PROP_ALTROOT) altroot = 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, B_TRUE)) 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; } /* * Make sure the bootsize is set when ZPOOL_CREATE_BOOT_LABEL is used, * and not set otherwise. */ if (boot_type == ZPOOL_CREATE_BOOT_LABEL) { const char *propname; char *strptr, *buf = NULL; int rv; propname = zpool_prop_to_name(ZPOOL_PROP_BOOTSIZE); if (nvlist_lookup_string(props, propname, &strptr) != 0) { (void) asprintf(&buf, "%" PRIu64, boot_size); if (buf == NULL) { (void) fprintf(stderr, gettext("internal error: out of memory\n")); goto errout; } rv = add_prop_list(propname, buf, &props, B_TRUE); free(buf); if (rv != 0) goto errout; } } else { const char *propname; char *strptr; propname = zpool_prop_to_name(ZPOOL_PROP_BOOTSIZE); if (nvlist_lookup_string(props, propname, &strptr) == 0) { (void) fprintf(stderr, gettext("error: setting boot " "partition size requires option '-B'\n")); goto errout; } } /* pass off to get_vdev_spec for bulk processing */ nvroot = make_root_vdev(NULL, force, !force, B_FALSE, dryrun, boot_type, boot_size, 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); ret = 0; } else { /* * Hand off to libzfs. */ if (enable_all_pool_feat) { spa_feature_t i; for (i = 0; i < SPA_FEATURES; i++) { char propname[MAXPATHLEN]; zfeature_info_t *feat = &spa_feature_table[i]; (void) snprintf(propname, sizeof (propname), "feature@%s", feat->fi_uname); /* * Skip feature if user specified it manually * on the command line. */ if (nvlist_exists(props, propname)) continue; 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_shareall(pool); 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 * * -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)); 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); } /* * zpool export [-f] ... * * -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) { boolean_t force = B_FALSE; boolean_t hardforce = B_FALSE; int c; zpool_handle_t *zhp; int ret; int i; /* check options */ while ((c = getopt(argc, argv, "fF")) != -1) { switch (c) { 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); } } argc -= optind; argv += optind; /* check arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool argument\n")); usage(B_FALSE); } ret = 0; for (i = 0; i < argc; i++) { if ((zhp = zpool_open_canfail(g_zfs, argv[i])) == NULL) { ret = 1; continue; } if (zpool_disable_datasets(zhp, force) != 0) { ret = 1; zpool_close(zhp); continue; } /* The history must be logged as part of the export */ log_history = B_FALSE; if (hardforce) { if (zpool_export_force(zhp, history_str) != 0) ret = 1; } else if (zpool_export(zhp, force, history_str) != 0) { ret = 1; } zpool_close(zhp); } 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) { char *name; nvlist_t **child; uint_t c, children; int ret; name = zpool_vdev_name(g_zfs, zhp, nv, name_flags | VDEV_NAME_TYPE_ID); if (strlen(name) + depth > max) max = strlen(name) + depth; free(name); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0) { for (c = 0; c < children; c++) if ((ret = max_width(zhp, child[c], depth + 2, max, name_flags)) > max) max = ret; } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0) { for (c = 0; c < children; c++) if ((ret = max_width(zhp, child[c], depth + 2, max, name_flags)) > max) max = ret; } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { for (c = 0; c < children; c++) if ((ret = max_width(zhp, child[c], depth + 2, max, name_flags)) > max) max = ret; } 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_explain; boolean_t cb_first; boolean_t cb_dedup_stats; boolean_t cb_print_status; } status_cbdata_t; /* * 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) { nvlist_t **child; uint_t c, 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; char *type; 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, &c) == 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 = "INUSE"; else if (vs->vs_state == VDEV_STATE_HEALTHY) state = "AVAIL"; } (void) printf("\t%*s%-*s %-8s", depth, "", cb->cb_namewidth - depth, name, state); if (!isspare) { 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)); (void) printf(" %5s %5s %5s", rbuf, wbuf, cbuf); } if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, ¬present) == 0) { char *path; verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0); (void) printf(" was %s", path); } else 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_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; default: (void) printf(gettext("corrupted data")); break; } } (void) nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &c); - if (ps != NULL && ps->pss_state == DSS_SCANNING && - vs->vs_scan_processed != 0 && children == 0) { - (void) printf(gettext(" (%s)"), - (ps->pss_func == POOL_SCAN_RESILVER) ? - "resilvering" : "repairing"); + if (ps != NULL && ps->pss_state == DSS_SCANNING && children == 0) { + 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)")); + } } 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), ", suspended, started at %s", tbuf); break; case VDEV_INITIALIZE_ACTIVE: (void) snprintf(zbuf, sizeof (zbuf), ", started at %s", tbuf); break; case VDEV_INITIALIZE_COMPLETE: (void) snprintf(zbuf, sizeof (zbuf), ", completed at %s", tbuf); break; } (void) printf(gettext(" (%d%% initialized%s)"), initialize_pct, zbuf); } (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; 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); 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; char *type, *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; 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; char *bias = NULL; char *type = NULL; (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); if (is_log) { 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) != 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); else print_import_config(cb, name, child[c], 2); free(name); } } /* * Display the status for the given pool. */ static void show_import(nvlist_t *config) { uint64_t pool_state; vdev_stat_t *vs; char *name; uint64_t guid; uint64_t hostid = 0; char *msgid; char *hostname = "unknown"; nvlist_t *nvroot, *nvinfo; int reason; const char *health; uint_t vsc; 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); (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: (void) printf(gettext(" status: One or more devices are " "missing from the system.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_R: case ZPOOL_STATUS_CORRUPT_LABEL_NR: (void) printf(gettext(" status: 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: (void) printf(gettext(" status: One or more devices " "are offlined.\n")); break; case ZPOOL_STATUS_CORRUPT_POOL: (void) printf(gettext(" status: The pool metadata is " "corrupted.\n")); break; case ZPOOL_STATUS_VERSION_OLDER: (void) printf(gettext(" status: The pool is formatted using a " "legacy on-disk version.\n")); break; case ZPOOL_STATUS_VERSION_NEWER: (void) printf(gettext(" status: The pool is formatted using an " "incompatible version.\n")); break; case ZPOOL_STATUS_FEAT_DISABLED: (void) printf(gettext(" status: Some supported features are " "not enabled on the pool.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_READ: (void) printf(gettext("status: The pool uses the following " "feature(s) not supported on this system:\n")); zpool_print_unsup_feat(config); break; case ZPOOL_STATUS_UNSUP_FEAT_WRITE: (void) printf(gettext("status: 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); break; case ZPOOL_STATUS_HOSTID_ACTIVE: (void) printf(gettext(" status: The pool is currently " "imported by another system.\n")); break; case ZPOOL_STATUS_HOSTID_REQUIRED: (void) printf(gettext(" status: 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: (void) printf(gettext(" status: The pool was last accessed by " "another system.\n")); break; case ZPOOL_STATUS_FAULTED_DEV_R: case ZPOOL_STATUS_FAULTED_DEV_NR: (void) printf(gettext(" status: One or more devices are " "faulted.\n")); break; case ZPOOL_STATUS_BAD_LOG: (void) printf(gettext(" status: An intent log record cannot be " "read.\n")); break; case ZPOOL_STATUS_RESILVERING: (void) printf(gettext(" status: One or more devices were being " "resilvered.\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_HOSTID_MISMATCH) { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric " "identifier and\n\tthe '-f' flag.\n")); } 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: (void) printf(gettext("action: 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: (void) printf(gettext("action: 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=%lx)\n\tbefore it " "can be safely imported.\n"), hostname, (unsigned long) hostid); break; case ZPOOL_STATUS_HOSTID_REQUIRED: (void) printf(gettext(" action: Check the SMF " "svc:/system/hostid service.\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: http://illumos.org/msg/%s\n"), msgid); (void) printf(gettext(" config:\n\n")); cb.cb_namewidth = max_width(NULL, nvroot, 0, 0, 0); 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")); } } 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) { zpool_handle_t *zhp; 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) { char *hostname = ""; 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%lx)\nExport the pool on the other system, " "then run 'zpool import'.\n"), name, hostname, (unsigned long) 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.\n"), name); } else { char *hostname = ""; uint64_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=%lx) at %s" "The pool can be imported, use 'zpool import -f' " "to import the pool.\n"), name, hostname, (unsigned long)hostid, ctime((time_t *)×tamp)); } return (1); } if (zpool_import_props(g_zfs, config, newname, props, flags) != 0) return (1); if (newname != NULL) name = (char *)newname; if ((zhp = zpool_open_canfail(g_zfs, name)) == NULL) return (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 (0); } /* * zpool checkpoint * checkpoint --discard * * -d Discard the checkpoint from a checkpointed * --discard pool. * * 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; char *pool; zpool_handle_t *zhp; int c, err; struct option long_options[] = { {"discard", no_argument, NULL, 'd'}, {0, 0, 0, 0} }; discard = B_FALSE; while ((c = getopt_long(argc, argv, ":d", long_options, NULL)) != -1) { switch (c) { case 'd': discard = 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 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); 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] * [-d dir | -c cachefile] [-f] -a * import [-o mntopts] [-o prop=value] ... [-R root] [-D] * [-d dir | -c cachefile] [-f] [-n] [-F] [-t] * [newpool] * * -c Read pool information from a cachefile instead of searching * devices. * * -d Scan in a specific directory, other than /dev/dsk. 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 Use newpool as a temporary pool name instead of renaming * the pool. * * -T Specify a starting txg to use for import. This option is * intentionally undocumented option for testing purposes. * * -a Import all pools found. * * -o Set property=value and/or temporary mount options (without '='). * * --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; 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; nvpair_t *elem; nvlist_t *config; uint64_t searchguid = 0; char *searchname = NULL; char *propval; nvlist_t *found_config; nvlist_t *policy = NULL; nvlist_t *props = NULL; boolean_t first; 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; uint64_t pool_state, 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:DEfFmnNo:rR:tT:VX", long_options, NULL)) != -1) { switch (c) { case 'a': do_all = B_TRUE; break; case 'c': cachefile = optarg; break; case 'd': if (searchdirs == NULL) { searchdirs = safe_malloc(sizeof (char *)); } else { char **tmp = safe_malloc((nsearch + 1) * sizeof (char *)); bcopy(searchdirs, tmp, nsearch * sizeof (char *)); free(searchdirs); searchdirs = tmp; } 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 '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, B_TRUE)) goto error; break; case 't': flags |= ZFS_IMPORT_TEMP_NAME; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props, B_TRUE)) 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 ((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; if (searchdirs == NULL) { searchdirs = safe_malloc(sizeof (char *)); searchdirs[0] = ZFS_DISK_ROOT; nsearch = 1; } /* 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 SYS_CONFIG privilege. We do this explicitly * here because otherwise any attempt to discover pools will * silently fail. */ if (argc == 0 && !priv_ineffect(PRIV_SYS_CONFIG)) { (void) fprintf(stderr, gettext("cannot " "discover pools: permission denied\n")); free(searchdirs); nvlist_free(policy); return (1); } } /* * Depending on the arguments given, we do one of the following: * * Iterate through all pools and display information about * each one. * * -a Iterate through all pools and try to import each one. * * 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; } found_config = NULL; /* * User specified a name or guid. Ensure it's unique. */ idata.unique = B_TRUE; } idata.path = searchdirs; idata.paths = nsearch; idata.poolname = searchname; idata.guid = searchguid; idata.cachefile = cachefile; idata.policy = policy; pools = zpool_search_import(g_zfs, &idata); if (pools != NULL && idata.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 'zpool import " "[-t] ' to give it a new temporary " "or permanent name\n")); err = 1; } else if (pools == NULL && idata.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); nvlist_free(policy); return (1); } /* * 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. */ err = 0; elem = NULL; 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, policy) == 0); if (argc == 0) { if (first) first = B_FALSE; else if (!do_all) (void) printf("\n"); if (do_all) { err |= do_import(config, NULL, mntopts, props, flags); } else { show_import(config); } } else if (searchname != NULL) { 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, searchname) == 0) { if (found_config != NULL) { (void) fprintf(stderr, gettext( "cannot import '%s': more than " "one matching pool\n"), searchname); (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 == searchguid) found_config = config; } } /* * If we were searching for a specific pool, verify that we found a * pool, and then do the import. */ if (argc != 0 && err == 0) { if (found_config == NULL) { (void) fprintf(stderr, gettext("cannot import '%s': " "no such pool available\n"), argv[0]); err = B_TRUE; } else { err |= do_import(found_config, argc == 1 ? NULL : argv[1], mntopts, props, flags); } } /* * If we were just looking for pools, report an error if none were * found. */ if (argc == 0 && first) (void) fprintf(stderr, gettext("no pools available to import\n")); error: nvlist_free(props); nvlist_free(pools); nvlist_free(policy); free(searchdirs); 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, zpool_sync_one, &force); return (ret); } typedef struct iostat_cbdata { boolean_t cb_verbose; int cb_name_flags; int cb_namewidth; int cb_iteration; boolean_t cb_scripted; zpool_list_t *cb_list; } iostat_cbdata_t; static void print_iostat_separator(iostat_cbdata_t *cb) { int i = 0; for (i = 0; i < cb->cb_namewidth; i++) (void) printf("-"); (void) printf(" ----- ----- ----- ----- ----- -----\n"); } static void print_iostat_header(iostat_cbdata_t *cb) { (void) printf("%*s capacity operations bandwidth\n", cb->cb_namewidth, ""); (void) printf("%-*s alloc free read write read write\n", cb->cb_namewidth, "pool"); print_iostat_separator(cb); } /* * Display a single statistic. */ static void print_one_stat(uint64_t value) { char buf[64]; zfs_nicenum(value, buf, sizeof (buf)); (void) printf(" %5s", buf); } static const char *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; vdev_stat_t *oldvs, *newvs; vdev_stat_t zerovs = { 0 }; char *vname; int ret = 0; uint64_t tdelta; double scale; if (strcmp(name, VDEV_TYPE_INDIRECT) == 0) return (ret); if (oldnv != NULL) { verify(nvlist_lookup_uint64_array(oldnv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&oldvs, &c) == 0); } else { oldvs = &zerovs; } verify(nvlist_lookup_uint64_array(newnv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&newvs, &c) == 0); 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), ""); tdelta = newvs->vs_timestamp - oldvs->vs_timestamp; if (tdelta == 0) scale = 1.0; else scale = (double)NANOSEC / tdelta; /* only toplevel vdevs have capacity stats */ if (newvs->vs_space == 0) { (void) printf(" - -"); } else { print_one_stat(newvs->vs_alloc); print_one_stat(newvs->vs_space - newvs->vs_alloc); } print_one_stat((uint64_t)(scale * (newvs->vs_ops[ZIO_TYPE_READ] - oldvs->vs_ops[ZIO_TYPE_READ]))); print_one_stat((uint64_t)(scale * (newvs->vs_ops[ZIO_TYPE_WRITE] - oldvs->vs_ops[ZIO_TYPE_WRITE]))); print_one_stat((uint64_t)(scale * (newvs->vs_bytes[ZIO_TYPE_READ] - oldvs->vs_bytes[ZIO_TYPE_READ]))); print_one_stat((uint64_t)(scale * (newvs->vs_bytes[ZIO_TYPE_WRITE] - oldvs->vs_bytes[ZIO_TYPE_WRITE]))); (void) printf("\n"); if (!cb->cb_verbose) return (ret); if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_CHILDREN, &newchild, &children) != 0) return (ret); if (oldnv && nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_CHILDREN, &oldchild, &c) != 0) return (ret); /* * 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_name_flags); 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 < 3; n++) { for (c = 0; c < children; c++) { uint64_t islog = B_FALSE; char *bias = NULL; 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; vname = zpool_vdev_name(g_zfs, zhp, newchild[c], cb->cb_name_flags); 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 && nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_L2CACHE, &oldchild, &c) != 0) return (ret); if (children > 0) { (void) printf("%-*s - - - - - " "-\n", cb->cb_namewidth, "cache"); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, zhp, newchild[c], cb->cb_name_flags); 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. */ int print_iostat(zpool_handle_t *zhp, void *data) { iostat_cbdata_t *cb = data; nvlist_t *oldconfig, *newconfig; nvlist_t *oldnvroot, *newnvroot; 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); /* * Print out the statistics for the pool. */ print_vdev_stats(zhp, zpool_get_name(zhp), oldnvroot, newnvroot, cb, 0); if (cb->cb_verbose) print_iostat_separator(cb); return (0); } int get_namewidth(zpool_handle_t *zhp, void *data) { iostat_cbdata_t *cb = data; nvlist_t *config, *nvroot; if ((config = zpool_get_config(zhp, NULL)) != NULL) { verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); if (!cb->cb_verbose) cb->cb_namewidth = strlen(zpool_get_name(zhp)); else cb->cb_namewidth = max_width(zhp, nvroot, 0, cb->cb_namewidth, cb->cb_name_flags); } /* * The width must fall into the range [10,38]. The upper limit is the * maximum we can have and still fit in 80 columns. */ if (cb->cb_namewidth < 10) cb->cb_namewidth = 10; if (cb->cb_namewidth > 38) cb->cb_namewidth = 38; return (0); } /* * Parse the input string, get the 'interval' and 'count' value if there is one. */ static void get_interval_count(int *argcp, char **argv, unsigned long *iv, unsigned long *cnt) { unsigned long interval = 0, count = 0; int argc = *argcp, errno; /* * Determine if the last argument is an integer or a pool name */ if (argc > 0 && isdigit(argv[argc - 1][0])) { char *end; errno = 0; interval = strtoul(argv[argc - 1], &end, 10); 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 && isdigit(argv[argc - 1][0])) { char *end; errno = 0; count = interval; interval = strtoul(argv[argc - 1], &end, 10); 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); } /* * zpool iostat [-gLPv] [-T d|u] [pool] ... [interval [count]] * * -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 * -T Display a timestamp in date(1) or Unix format * * 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; unsigned long interval = 0, count = 0; zpool_list_t *list; boolean_t verbose = B_FALSE; boolean_t guid = B_FALSE; boolean_t follow_links = B_FALSE; boolean_t full_name = B_FALSE; iostat_cbdata_t cb = { 0 }; /* check options */ while ((c = getopt(argc, argv, "gLPT:v")) != -1) { switch (c) { 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 '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; get_interval_count(&argc, argv, &interval, &count); /* * Construct the list of all interesting pools. */ ret = 0; if ((list = pool_list_get(argc, argv, NULL, &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); } /* * Enter the main iostat loop. */ cb.cb_list = list; cb.cb_verbose = verbose; if (guid) cb.cb_name_flags |= VDEV_NAME_GUID; if (follow_links) cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS; if (full_name) cb.cb_name_flags |= VDEV_NAME_PATH; cb.cb_iteration = 0; cb.cb_namewidth = 0; for (;;) { pool_list_update(list); if ((npools = pool_list_count(list)) == 0) break; /* * 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, &cb); if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); /* * If it's the first time, or verbose mode, print the header. */ if (++cb.cb_iteration == 1 || verbose) print_iostat_header(&cb); (void) 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. */ if (npools > 1 && !verbose) print_iostat_separator(&cb); if (verbose) (void) printf("\n"); /* * 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) sleep(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) printf(" "); else first = B_FALSE; right_justify = B_FALSE; if (pl->pl_prop != ZPROP_INVAL) { 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) printf("%s", header); else if (right_justify) (void) printf("%*s", width, header); else (void) printf("%-*s", width, header); } (void) printf("\n"); } /* * 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]; 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) printf("\t"); else (void) printf(" "); } else { first = B_FALSE; } right_justify = B_FALSE; if (pl->pl_prop != ZPROP_INVAL) { 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 { 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) printf("%s", propstr); else if (right_justify) (void) printf("%*s", width, propstr); else (void) printf("%-*s", width, propstr); } (void) printf("\n"); } static void print_one_column(zpool_prop_t prop, uint64_t value, boolean_t scripted, boolean_t valid) { char propval[64]; boolean_t fixed; size_t width = zprop_width(prop, &fixed, ZFS_TYPE_POOL); switch (prop) { case ZPOOL_PROP_EXPANDSZ: case ZPOOL_PROP_CHECKPOINT: if (value == 0) (void) strlcpy(propval, "-", sizeof (propval)); else zfs_nicenum(value, propval, sizeof (propval)); break; case ZPOOL_PROP_FRAGMENTATION: if (value == ZFS_FRAG_INVALID) { (void) strlcpy(propval, "-", sizeof (propval)); } else { (void) snprintf(propval, sizeof (propval), "%llu%%", value); } break; case ZPOOL_PROP_CAPACITY: (void) snprintf(propval, sizeof (propval), value < 1000 ? "%1.2f%%" : value < 10000 ? "%2.1f%%" : "%3.0f%%", value / 100.0); break; default: zfs_nicenum(value, propval, sizeof (propval)); } if (!valid) (void) strlcpy(propval, "-", sizeof (propval)); if (scripted) (void) printf("\t%s", propval); else (void) printf(" %*s", width, propval); } /* * print static default line per vdev */ void print_list_stats(zpool_handle_t *zhp, const char *name, nvlist_t *nv, list_cbdata_t *cb, int depth) { nvlist_t **child; vdev_stat_t *vs; uint_t c, children; char *vname; boolean_t scripted = cb->cb_scripted; uint64_t islog = B_FALSE; 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; 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. */ print_one_column(ZPOOL_PROP_SIZE, vs->vs_space, scripted, toplevel); print_one_column(ZPOOL_PROP_ALLOCATED, vs->vs_alloc, scripted, toplevel); print_one_column(ZPOOL_PROP_FREE, vs->vs_space - vs->vs_alloc, scripted, toplevel); print_one_column(ZPOOL_PROP_CHECKPOINT, vs->vs_checkpoint_space, scripted, toplevel); print_one_column(ZPOOL_PROP_EXPANDSZ, vs->vs_esize, scripted, B_TRUE); print_one_column(ZPOOL_PROP_FRAGMENTATION, vs->vs_fragmentation, scripted, (vs->vs_fragmentation != ZFS_FRAG_INVALID && toplevel)); cap = (vs->vs_space == 0) ? 0 : (vs->vs_alloc * 10000 / vs->vs_space); print_one_column(ZPOOL_PROP_CAPACITY, cap, scripted, toplevel); (void) printf("\n"); } 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); print_list_stats(zhp, vname, child[c], cb, depth + 2); free(vname); } /* list the classes: 'logs', 'dedup', and 'special' */ for (uint_t n = 0; n < 3; n++) { boolean_t printed = B_FALSE; for (c = 0; c < children; c++) { char *bias = NULL; 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); print_list_stats(zhp, vname, child[c], cb, depth + 2); 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); 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); free(vname); } } } /* * Generic callback function to list a pool. */ int list_callback(zpool_handle_t *zhp, void *data) { list_cbdata_t *cbp = data; nvlist_t *config; nvlist_t *nvroot; config = zpool_get_config(zhp, NULL); if (cbp->cb_verbose) { config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); } if (cbp->cb_verbose) cbp->cb_namewidth = max_width(zhp, nvroot, 0, 0, cbp->cb_name_flags); print_pool(zhp, cbp); if (cbp->cb_verbose) print_list_stats(zhp, NULL, nvroot, cbp, 0); return (0); } /* * zpool list [-gHLP] [-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 Diplay 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; list_cbdata_t cb = { 0 }; static char default_props[] = "name,size,allocated,free,checkpoint,expandsize,fragmentation," "capacity,dedupratio,health,altroot"; char *props = default_props; unsigned long interval = 0, count = 0; zpool_list_t *list; boolean_t first = B_TRUE; /* 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, &ret)) == NULL) return (1); if (pool_list_count(list) == 0) break; cb.cb_namewidth = 0; (void) pool_list_iter(list, B_FALSE, get_namewidth, &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) sleep(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; int c; nvlist_t *nvroot; char *poolname, *old_disk, *new_disk; zpool_handle_t *zhp; zpool_boot_label_t boot_type; uint64_t boot_size; 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; /* 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 specification\n")); usage(B_FALSE); } old_disk = argv[1]; if (argc < 3) { if (!replacing) { (void) fprintf(stderr, gettext("missing 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) return (1); if (zpool_get_config(zhp, NULL) == NULL) { (void) fprintf(stderr, gettext("pool '%s' is unavailable\n"), poolname); zpool_close(zhp); return (1); } if (zpool_is_bootable(zhp)) boot_type = ZPOOL_COPY_BOOT_LABEL; else boot_type = ZPOOL_NO_BOOT_LABEL; boot_size = zpool_get_prop_int(zhp, ZPOOL_PROP_BOOTSIZE, NULL); nvroot = make_root_vdev(zhp, force, B_FALSE, replacing, B_FALSE, boot_type, boot_size, argc, argv); if (nvroot == NULL) { zpool_close(zhp); return (1); } ret = zpool_vdev_attach(zhp, old_disk, new_disk, nvroot, replacing); nvlist_free(nvroot); zpool_close(zhp); return (ret); } /* * zpool replace [-f] * * -f Force attach, even if appears to be in use. * * Replace with . */ /* ARGSUSED */ int zpool_do_replace(int argc, char **argv) { return (zpool_do_attach_or_replace(argc, argv, B_TRUE)); } /* * zpool attach [-f] * * -f Force attach, even if appears to be in use. * * Attach to the mirror containing . If is not * part of a mirror, then will be transformed into a mirror of * and . In either case, 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] * * -f Force detach of , even if DTLs argue against it * (not supported yet) * * Detach a device from a mirror. The operation will be refused if * is the last device in the mirror, or if the DTLs indicate that this device * has the only valid copy of some data. */ /* ARGSUSED */ 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, "f")) != -1) { switch (c) { case 'f': 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 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] [ ...] * * -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. * * 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; 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:no: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 '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); } 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) 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); } nvlist_free(config); } zpool_close(zhp); if (ret != 0 || flags.dryrun || !flags.import) 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) return (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); return (ret); } /* * zpool online ... */ 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, "et")) != -1) { switch (c) { case 'e': flags |= ZFS_ONLINE_EXPAND; break; case 't': 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 (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")); } } else { ret = 1; } } zpool_close(zhp); return (ret); } /* * zpool offline [-ft] ... * * -f Force the device into the offline state, even if doing * so would appear to compromise pool availability. * (not supported yet) * * -t Only take the device off-line temporarily. The offline * state will not be persistent across reboots. */ /* ARGSUSED */ 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; /* check options */ while ((c = getopt(argc, argv, "ft")) != -1) { switch (c) { case 't': istmp = B_TRUE; break; case 'f': 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 (zpool_vdev_offline(zhp, argv[i], istmp) != 0) ret = 1; } zpool_close(zhp); return (ret); } /* * zpool clear [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 */ 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 * * 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; zpool_handle_t *zhp; char *pool; /* check options */ while ((c = getopt(argc, argv, "")) != -1) { switch (c) { case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc--; argv++; 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); } pool = argv[0]; if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) return (1); ret = zpool_reopen(zhp); zpool_close(zhp); return (ret); } typedef struct scrub_cbdata { int cb_type; int cb_argc; char **cb_argv; 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); } 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 scrub '%s': pool is " + (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); } /* * zpool scrub [-s | -p] ... * * -s Stop. Stops any in-progress scrub. * -p Pause. Pause in-progress scrub. */ int zpool_do_scrub(int argc, char **argv) { int c; scrub_cbdata_t cb; cb.cb_type = POOL_SCAN_SCRUB; cb.cb_scrub_cmd = POOL_SCRUB_NORMAL; /* check options */ while ((c = getopt(argc, argv, "sp")) != -1) { switch (c) { case 's': cb.cb_type = POOL_SCAN_NONE; break; case 'p': cb.cb_scrub_cmd = POOL_SCRUB_PAUSE; 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); } cb.cb_argc = argc; cb.cb_argv = argv; + 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, scrub_callback, &cb)); +} + +/* + * zpool resilver ... + * + * 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; + cb.cb_argc = argc; + cb.cb_argv = argv; + + /* 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, scrub_callback, &cb)); } 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, B_FALSE); fnvlist_add_boolean(res, path); free(path); return; } for (i = 0; i < children; i++) { zpool_collect_leaves(zhp, child[i], res); } } /* * zpool initialize [-cs] [ ...] * 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. */ int zpool_do_initialize(int argc, char **argv) { int c; char *poolname; zpool_handle_t *zhp; nvlist_t *vdevs; int err = 0; struct option long_options[] = { {"cancel", no_argument, NULL, 'c'}, {"suspend", no_argument, NULL, 's'}, {0, 0, 0, 0} }; pool_initialize_func_t cmd_type = POOL_INITIALIZE_DO; while ((c = getopt_long(argc, argv, "cs", long_options, NULL)) != -1) { switch (c) { case 'c': if (cmd_type != POOL_INITIALIZE_DO) { (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_DO) { (void) fprintf(stderr, gettext("-s cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_INITIALIZE_SUSPEND; 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); } 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 { int i; for (i = 1; i < argc; i++) { fnvlist_add_boolean(vdevs, argv[i]); } } err = zpool_initialize(zhp, cmd_type, vdevs); fnvlist_free(vdevs); zpool_close(zhp); return (err); } /* * Print out detailed scrub status. */ static void print_scan_status(pool_scan_stat_t *ps) { time_t start, end, pause; uint64_t total_secs_left; uint64_t elapsed, secs_left, mins_left, hours_left, days_left; uint64_t pass_scanned, scanned, pass_issued, issued, total; uint_t 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]; (void) printf(gettext(" scan: ")); /* 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_nicenum(ps->pss_processed, processed_buf, sizeof (processed_buf)); assert(ps->pss_func == POOL_SCAN_SCRUB || ps->pss_func == POOL_SCAN_RESILVER); /* * Scan is finished or canceled. */ if (ps->pss_state == DSS_FINISHED) { total_secs_left = end - start; days_left = total_secs_left / 60 / 60 / 24; hours_left = (total_secs_left / 60 / 60) % 24; mins_left = (total_secs_left / 60) % 60; secs_left = (total_secs_left % 60); if (ps->pss_func == POOL_SCAN_SCRUB) { (void) printf(gettext("scrub repaired %s " "in %llu days %02llu:%02llu:%02llu " "with %llu errors on %s"), processed_buf, (u_longlong_t)days_left, (u_longlong_t)hours_left, (u_longlong_t)mins_left, (u_longlong_t)secs_left, (u_longlong_t)ps->pss_errors, ctime(&end)); } else if (ps->pss_func == POOL_SCAN_RESILVER) { (void) printf(gettext("resilvered %s " "in %llu days %02llu:%02llu:%02llu " "with %llu errors on %s"), processed_buf, (u_longlong_t)days_left, (u_longlong_t)hours_left, (u_longlong_t)mins_left, (u_longlong_t)secs_left, (u_longlong_t)ps->pss_errors, ctime(&end)); } return; } else if (ps->pss_state == DSS_CANCELED) { if (ps->pss_func == POOL_SCAN_SCRUB) { (void) printf(gettext("scrub canceled on %s"), ctime(&end)); } else if (ps->pss_func == POOL_SCAN_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 (ps->pss_func == POOL_SCAN_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 (ps->pss_func == POOL_SCAN_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; total_secs_left = (issue_rate != 0 && total >= issued) ? ((total - issued) / issue_rate) : UINT64_MAX; days_left = total_secs_left / 60 / 60 / 24; hours_left = (total_secs_left / 60 / 60) % 24; mins_left = (total_secs_left / 60) % 60; secs_left = (total_secs_left % 60); /* format all of the numbers we will be reporting */ zfs_nicenum(scanned, scanned_buf, sizeof (scanned_buf)); zfs_nicenum(issued, issued_buf, sizeof (issued_buf)); zfs_nicenum(total, total_buf, sizeof (total_buf)); zfs_nicenum(scan_rate, srate_buf, sizeof (srate_buf)); zfs_nicenum(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 (ps->pss_func == POOL_SCAN_RESILVER) { (void) printf(gettext("\t%s resilvered, %.2f%% done"), processed_buf, 100 * fraction_done); } else if (ps->pss_func == POOL_SCAN_SCRUB) { (void) printf(gettext("\t%s repaired, %.2f%% done"), processed_buf, 100 * fraction_done); } if (pause == 0) { if (total_secs_left != UINT64_MAX && issue_rate >= 10 * 1024 * 1024) { (void) printf(gettext(", %llu days " "%02llu:%02llu:%02llu to go\n"), (u_longlong_t)days_left, (u_longlong_t)hours_left, (u_longlong_t)mins_left, (u_longlong_t)secs_left); } else { (void) printf(gettext(", no estimated " "completion time\n")); } } else { (void) printf(gettext("\n")); } } /* * 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. */ 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")); } } /* * 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); (void) printf(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(" %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")); } } if (prs->prs_mapping_memory > 0) { char mem_buf[7]; zfs_nicenum(prs->prs_mapping_memory, mem_buf, sizeof (mem_buf)); (void) printf(gettext(" %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"), 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) { (void) printf("errors: List of errors unavailable " "(insufficient privileges)\n"); 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); 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); 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; /* * 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; } (void) printf("DDT entries %llu, size %llu on disk, %llu in core\n", (u_longlong_t)ddo->ddo_count, (u_longlong_t)ddo->ddo_dspace, (u_longlong_t)ddo->ddo_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: http://illumos.org/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. */ int status_callback(zpool_handle_t *zhp, void *data) { status_cbdata_t *cbp = data; nvlist_t *config, *nvroot; char *msgid; int reason; const char *health; uint_t c; vdev_stat_t *vs; config = zpool_get_config(zhp, NULL); reason = zpool_get_status(zhp, &msgid); 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)) { 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_state_to_name(vs->vs_state, vs->vs_aux); (void) printf(gettext(" pool: %s\n"), zpool_get_name(zhp)); (void) printf(gettext(" state: %s\n"), health); switch (reason) { case ZPOOL_STATUS_MISSING_DEV_R: (void) printf(gettext("status: One or more devices could not " "be opened. Sufficient replicas exist for\n\tthe pool to " "continue functioning in a degraded state.\n")); (void) printf(gettext("action: Attach the missing device and " "online it using 'zpool online'.\n")); break; case ZPOOL_STATUS_MISSING_DEV_NR: (void) printf(gettext("status: One or more devices could not " "be opened. There are insufficient\n\treplicas for the " "pool to continue functioning.\n")); (void) printf(gettext("action: Attach the missing device and " "online it using 'zpool online'.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_R: (void) printf(gettext("status: 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")); (void) printf(gettext("action: Replace the device using " "'zpool replace'.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_NR: (void) printf(gettext("status: 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: (void) printf(gettext("status: One or more devices has " "experienced an unrecoverable error. An\n\tattempt was " "made to correct the error. Applications are " "unaffected.\n")); (void) printf(gettext("action: 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: (void) printf(gettext("status: 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")); (void) printf(gettext("action: Online the device using " "'zpool online' or replace the device with\n\t'zpool " "replace'.\n")); break; case ZPOOL_STATUS_REMOVED_DEV: (void) printf(gettext("status: 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")); (void) printf(gettext("action: Online the device using " "'zpool online' or replace the device with\n\t'zpool " "replace'.\n")); break; case ZPOOL_STATUS_RESILVERING: (void) printf(gettext("status: One or more devices is " "currently being resilvered. The pool will\n\tcontinue " "to function, possibly in a degraded state.\n")); (void) printf(gettext("action: Wait for the resilver to " "complete.\n")); break; case ZPOOL_STATUS_CORRUPT_DATA: (void) printf(gettext("status: One or more devices has " "experienced an error resulting in data\n\tcorruption. " "Applications may be affected.\n")); (void) printf(gettext("action: Restore the file in question " "if possible. Otherwise restore the\n\tentire pool from " "backup.\n")); break; case ZPOOL_STATUS_CORRUPT_POOL: (void) printf(gettext("status: 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: (void) printf(gettext("status: The pool is formatted using a " "legacy on-disk format. The pool can\n\tstill be used, " "but some features are unavailable.\n")); (void) printf(gettext("action: Upgrade the pool using 'zpool " "upgrade'. Once this is done, the\n\tpool will no longer " "be accessible on software that does not support feature\n" "\tflags.\n")); break; case ZPOOL_STATUS_VERSION_NEWER: (void) printf(gettext("status: The pool has been upgraded to a " "newer, incompatible on-disk version.\n\tThe pool cannot " "be accessed on this system.\n")); (void) printf(gettext("action: Access the pool from a system " "running more recent software, or\n\trestore the pool from " "backup.\n")); break; case ZPOOL_STATUS_FEAT_DISABLED: (void) printf(gettext("status: Some supported features are not " "enabled on the pool. The pool can\n\tstill be used, but " "some features are unavailable.\n")); (void) printf(gettext("action: 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(5) for details.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_READ: (void) printf(gettext("status: 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"); (void) printf(gettext("action: 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: (void) printf(gettext("status: 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"); (void) printf(gettext("action: 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: (void) printf(gettext("status: 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")); (void) printf(gettext("action: Replace the faulted device, " "or use 'zpool clear' to mark the device\n\trepaired.\n")); break; case ZPOOL_STATUS_FAULTED_DEV_NR: (void) printf(gettext("status: One or more devices are " "faulted in response to persistent errors. There are " "insufficient replicas for the pool to\n\tcontinue " "functioning.\n")); (void) printf(gettext("action: 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: (void) printf(gettext("status: The pool is suspended because " "multihost writes failed or were delayed;\n\tanother " "system could import the pool undetected.\n")); (void) printf(gettext("action: 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: (void) printf(gettext("status: One or more devices are " "faulted in response to IO failures.\n")); (void) printf(gettext("action: Make sure the affected devices " "are connected, then run 'zpool clear'.\n")); break; case ZPOOL_STATUS_BAD_LOG: (void) printf(gettext("status: An intent log record " "could not be read.\n" "\tWaiting for adminstrator intervention to fix the " "faulted pool.\n")); (void) printf(gettext("action: Either restore the affected " "device(s) and run 'zpool online',\n" "\tor ignore the intent log records by running " "'zpool clear'.\n")); break; default: /* * The remaining errors can't actually be generated, yet. */ assert(reason == ZPOOL_STATUS_OK); } if (msgid != NULL) (void) printf(gettext(" see: http://illumos.org/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_scan_stat_t *ps = NULL; pool_removal_stat_t *prs = NULL; (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &c); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c); print_scan_status(ps); print_checkpoint_scan_warning(ps, pcs); print_removal_status(zhp, prs); print_checkpoint_status(pcs); cbp->cb_namewidth = max_width(zhp, nvroot, 0, 0, cbp->cb_name_flags); if (cbp->cb_namewidth < 10) cbp->cb_namewidth = 10; (void) printf(gettext("config:\n\n")); (void) printf(gettext("\t%-*s %-8s %5s %5s %5s\n"), cbp->cb_namewidth, "NAME", "STATE", "READ", "WRITE", "CKSUM"); print_status_config(zhp, cbp, zpool_get_name(zhp), nvroot, 0, B_FALSE); 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) { nvlist_t *nverrlist = NULL; /* * If the approximate error count is small, get a * precise count by fetching the entire log and * uniquifying the results. */ if (nerr > 0 && nerr < 100 && !cbp->cb_verbose && zpool_get_errlog(zhp, &nverrlist) == 0) { nvpair_t *elem; elem = NULL; nerr = 0; while ((elem = nvlist_next_nvpair(nverrlist, elem)) != NULL) { nerr++; } } nvlist_free(nverrlist); (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 [-gLPvx] [-T d|u] [pool] ... [interval [count]] * * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -P Display full path for vdev name. * -v Display complete error logs * -x Display only pools with potential problems * -D Display dedup status (undocumented) * -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; unsigned long interval = 0, count = 0; status_cbdata_t cb = { 0 }; /* check options */ while ((c = getopt(argc, argv, "gLPvxDT:")) != -1) { switch (c) { 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_name_flags |= VDEV_NAME_PATH; 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': get_timestamp_arg(*optarg); 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 (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); ret = for_each_pool(argc, argv, B_TRUE, NULL, status_callback, &cb); if (argc == 0 && cb.cb_count == 0) (void) printf(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) sleep(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 upgrade_version(zpool_handle_t *zhp, uint64_t version) { int ret; nvlist_t *config; uint64_t oldversion; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &oldversion) == 0); assert(SPA_VERSION_IS_SUPPORTED(oldversion)); assert(oldversion < version); 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), oldversion); } else { (void) printf(gettext("Successfully upgraded " "'%s' from version %llu to version %llu.\n"), zpool_get_name(zhp), oldversion, 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); 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 (!nvlist_exists(enabled, fguid)) { 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 printnl = 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); printnl = 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; printnl = B_TRUE; } } if (printnl) { (void) printf(gettext("\n")); } 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")); (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 (!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(5) for " "details.\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); } } } return (0); } /* ARGSUSED */ static int upgrade_one(zpool_handle_t *zhp, void *data) { boolean_t printnl = B_FALSE; upgrade_cbdata_t *cbp = data; uint64_t cur_version; int ret; if (strcmp("log", zpool_get_name(zhp)) == 0) { (void) printf(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), 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), cbp->cb_version); return (0); } if (cur_version != cbp->cb_version) { printnl = 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) { printnl = B_TRUE; } else if (cur_version == SPA_VERSION) { (void) printf(gettext("Pool '%s' already has all " "supported features enabled.\n"), zpool_get_name(zhp)); } } if (printnl) { (void) printf(gettext("\n")); } 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(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]; 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 features " "enabled.\n")); } else { (void) printf(gettext("All pools are already " "formatted with version %llu or higher.\n"), 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 features enabled.\n")); } else { (void) printf(gettext("\n")); } } else { ret = for_each_pool(argc, argv, B_FALSE, NULL, upgrade_one, &cb); } return (ret); } typedef struct hist_cbdata { boolean_t first; boolean_t longfmt; boolean_t internal; } hist_cbdata_t; /* * Print out the command history for a specific pool. */ static int get_history_one(zpool_handle_t *zhp, void *data) { nvlist_t *nvhis; nvlist_t **records; uint_t numrecords; int ret, i; hist_cbdata_t *cb = (hist_cbdata_t *)data; cb->first = B_FALSE; (void) printf(gettext("History for '%s':\n"), zpool_get_name(zhp)); if ((ret = zpool_get_history(zhp, &nvhis)) != 0) return (ret); 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[30] = ""; 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_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], 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, 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), 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_ERRNO)) { (void) printf(" errno: %lld\n", 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"); } (void) printf("\n"); nvlist_free(nvhis); return (ret); } /* * zpool history * * 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, get_history_one, &cbdata); if (argc == 0 && cbdata.first == B_TRUE) { (void) printf(gettext("no pools available\n")); return (0); } return (ret); } static int get_callback(zpool_handle_t *zhp, void *data) { zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data; char value[MAXNAMELEN]; zprop_source_t srctype; zprop_list_t *pl; 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 && (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[,...]> ... * * -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 Diplay 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 *value; 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; /* 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': bzero(&cb.cb_columns, sizeof (cb.cb_columns)); i = 0; while (*optarg != '\0') { static char *col_subopts[] = { "name", "property", "value", "source", "all", NULL }; if (i == ZFS_GET_NCOLS) { (void) fprintf(stderr, gettext("too " "many fields given to -o " "option\n")); usage(B_FALSE); } switch (getsubopt(&optarg, col_subopts, &value)) { case 0: cb.cb_columns[i++] = GET_COL_NAME; break; case 1: cb.cb_columns[i++] = GET_COL_PROPERTY; break; case 2: cb.cb_columns[i++] = GET_COL_VALUE; break; case 3: cb.cb_columns[i++] = GET_COL_SOURCE; break; case 4: if (i > 0) { (void) fprintf(stderr, gettext("\"all\" conflicts " "with specific fields " "given to -o option\n")); usage(B_FALSE); } 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; i = ZFS_GET_NCOLS; break; default: (void) fprintf(stderr, gettext("invalid column name " "'%s'\n"), value); 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 property " "argument\n")); usage(B_FALSE); } if (zprop_get_list(g_zfs, argv[0], &cb.cb_proplist, ZFS_TYPE_POOL) != 0) usage(B_FALSE); argc--; argv++; 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, 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); } typedef struct set_cbdata { char *cb_propname; char *cb_value; boolean_t cb_any_successful; } set_cbdata_t; int set_callback(zpool_handle_t *zhp, void *data) { int error; set_cbdata_t *cb = (set_cbdata_t *)data; error = zpool_set_prop(zhp, cb->cb_propname, cb->cb_value); if (!error) cb->cb_any_successful = B_TRUE; return (error); } int zpool_do_set(int argc, char **argv) { set_cbdata_t cb = { 0 }; int error; 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 > 3) { (void) fprintf(stderr, gettext("too many pool names\n")); usage(B_FALSE); } cb.cb_propname = argv[1]; 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++; error = for_each_pool(argc - 2, argv + 2, B_TRUE, NULL, set_callback, &cb); return (error); } static int find_command_idx(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); } int main(int argc, char **argv) { int ret = 0; int i; char *cmdname; (void) setlocale(LC_ALL, ""); (void) textdomain(TEXT_DOMAIN); if ((g_zfs = libzfs_init()) == NULL) { (void) fprintf(stderr, gettext("internal error: failed to " "initialize ZFS library\n")); return (1); } libzfs_print_on_error(g_zfs, B_TRUE); 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) usage(B_TRUE); zfs_save_arguments(argc, argv, history_str, sizeof (history_str)); /* * Run the appropriate command. */ if (find_command_idx(cmdname, &i) == 0) { current_command = &command_table[i]; ret = command_table[i].func(argc - 1, argv + 1); } else if (strchr(cmdname, '=')) { verify(find_command_idx("set", &i) == 0); current_command = &command_table[i]; ret = command_table[i].func(argc, argv); } else if (strcmp(cmdname, "freeze") == 0 && argc == 3) { /* * 'freeze' is a vile debugging abomination, so we treat * it as such. */ char buf[16384]; int fd = open(ZFS_DEV, O_RDWR); (void) strcpy((void *)buf, argv[2]); return (!!ioctl(fd, ZFS_IOC_POOL_FREEZE, buf)); } else { (void) fprintf(stderr, gettext("unrecognized " "command '%s'\n"), cmdname); usage(B_FALSE); } 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); } Index: vendor/illumos/dist/lib/libzfs/common/libzfs.h =================================================================== --- vendor/illumos/dist/lib/libzfs/common/libzfs.h (revision 354960) +++ vendor/illumos/dist/lib/libzfs/common/libzfs.h (revision 354961) @@ -1,852 +1,853 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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 Pawel Jakub Dawidek. All rights reserved. * Copyright (c) 2011, 2017 by Delphix. All rights reserved. * Copyright 2019 Joyent, Inc. * Copyright (c) 2013 Steven Hartland. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright 2016 Nexenta Systems, Inc. * Copyright (c) 2017 Datto Inc. */ #ifndef _LIBZFS_H #define _LIBZFS_H #include #include #include #include #include #include #include #include #include #include #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, /* currently resilvering */ 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, /* unshare(1M) failed */ EZFS_SHARENFSFAILED, /* share(1M) failed */ 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_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_TOOMANY, /* argument list too long */ EZFS_INITIALIZING, /* currently initializing */ EZFS_NO_INITIALIZE, /* no active initialize */ + EZFS_NO_RESILVER_DEFER, /* pool doesn't support resilver_defer */ EZFS_UNKNOWN } zfs_error_t; /* * UEFI boot support parameters. When creating whole disk boot pool, * zpool create should allow to create EFI System partition for UEFI boot * program. In case of BIOS, the EFI System partition is not used * even if it does exist. */ typedef enum zpool_boot_label { ZPOOL_NO_BOOT_LABEL = 0, ZPOOL_CREATE_BOOT_LABEL, ZPOOL_COPY_BOOT_LABEL } zpool_boot_label_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; /* * Library initialization */ extern libzfs_handle_t *libzfs_init(void); extern void libzfs_fini(libzfs_handle_t *); extern libzfs_handle_t *zpool_get_handle(zpool_handle_t *); extern libzfs_handle_t *zfs_get_handle(zfs_handle_t *); extern void libzfs_print_on_error(libzfs_handle_t *, boolean_t); extern void zfs_save_arguments(int argc, char **, char *, int); extern int zpool_log_history(libzfs_handle_t *, const char *); extern int libzfs_errno(libzfs_handle_t *); extern const char *libzfs_error_action(libzfs_handle_t *); extern const char *libzfs_error_description(libzfs_handle_t *); extern int zfs_standard_error(libzfs_handle_t *, int, const char *); extern void libzfs_mnttab_init(libzfs_handle_t *); extern void libzfs_mnttab_fini(libzfs_handle_t *); extern void libzfs_mnttab_cache(libzfs_handle_t *, boolean_t); extern int libzfs_mnttab_find(libzfs_handle_t *, const char *, struct mnttab *); extern void libzfs_mnttab_add(libzfs_handle_t *, const char *, const char *, const char *); extern void libzfs_mnttab_remove(libzfs_handle_t *, const char *); /* * Basic handle functions */ extern zpool_handle_t *zpool_open(libzfs_handle_t *, const char *); extern zpool_handle_t *zpool_open_canfail(libzfs_handle_t *, const char *); extern void zpool_close(zpool_handle_t *); extern const char *zpool_get_name(zpool_handle_t *); extern int zpool_get_state(zpool_handle_t *); extern const char *zpool_state_to_name(vdev_state_t, vdev_aux_t); extern const char *zpool_pool_state_to_name(pool_state_t); extern void zpool_free_handles(libzfs_handle_t *); /* * Iterate over all active pools in the system. */ typedef int (*zpool_iter_f)(zpool_handle_t *, void *); extern int zpool_iter(libzfs_handle_t *, zpool_iter_f, void *); extern boolean_t zpool_skip_pool(const char *); /* * Functions to create and destroy pools */ extern int zpool_create(libzfs_handle_t *, const char *, nvlist_t *, nvlist_t *, nvlist_t *); extern int zpool_destroy(zpool_handle_t *, const char *); extern int zpool_add(zpool_handle_t *, nvlist_t *); typedef struct splitflags { /* do not split, but return the config that would be split off */ int dryrun : 1; /* after splitting, import the pool */ int import : 1; int name_flags; } splitflags_t; /* * Functions to manipulate pool and vdev state */ extern int zpool_scan(zpool_handle_t *, pool_scan_func_t, pool_scrub_cmd_t); extern int zpool_initialize(zpool_handle_t *, pool_initialize_func_t, nvlist_t *); extern int zpool_clear(zpool_handle_t *, const char *, nvlist_t *); extern int zpool_reguid(zpool_handle_t *); extern int zpool_reopen(zpool_handle_t *); extern int zpool_sync_one(zpool_handle_t *, void *); extern int zpool_vdev_online(zpool_handle_t *, const char *, int, vdev_state_t *); extern int zpool_vdev_offline(zpool_handle_t *, const char *, boolean_t); extern int zpool_vdev_attach(zpool_handle_t *, const char *, const char *, nvlist_t *, int); extern int zpool_vdev_detach(zpool_handle_t *, const char *); extern int zpool_vdev_remove(zpool_handle_t *, const char *); extern int zpool_vdev_remove_cancel(zpool_handle_t *); extern int zpool_vdev_indirect_size(zpool_handle_t *, const char *, uint64_t *); extern int zpool_vdev_split(zpool_handle_t *, char *, nvlist_t **, nvlist_t *, splitflags_t); extern int zpool_vdev_fault(zpool_handle_t *, uint64_t, vdev_aux_t); extern int zpool_vdev_degrade(zpool_handle_t *, uint64_t, vdev_aux_t); extern int zpool_vdev_clear(zpool_handle_t *, uint64_t); extern nvlist_t *zpool_find_vdev(zpool_handle_t *, const char *, boolean_t *, boolean_t *, boolean_t *); extern nvlist_t *zpool_find_vdev_by_physpath(zpool_handle_t *, const char *, boolean_t *, boolean_t *, boolean_t *); extern int zpool_label_disk(libzfs_handle_t *, zpool_handle_t *, const char *, zpool_boot_label_t, uint64_t, int *); /* * Functions to manage pool properties */ extern int zpool_set_prop(zpool_handle_t *, const char *, const char *); extern int zpool_get_prop(zpool_handle_t *, zpool_prop_t, char *, size_t proplen, zprop_source_t *, boolean_t); extern uint64_t zpool_get_prop_int(zpool_handle_t *, zpool_prop_t, zprop_source_t *); extern const char *zpool_prop_to_name(zpool_prop_t); extern const char *zpool_prop_values(zpool_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) */ /* * 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 */ /* * Finally, the following indicates a healthy pool. */ ZPOOL_STATUS_OK } zpool_status_t; extern zpool_status_t zpool_get_status(zpool_handle_t *, char **); extern zpool_status_t zpool_import_status(nvlist_t *, char **); extern void zpool_dump_ddt(const ddt_stat_t *dds, const ddt_histogram_t *ddh); /* * Statistics and configuration functions. */ extern nvlist_t *zpool_get_config(zpool_handle_t *, nvlist_t **); extern nvlist_t *zpool_get_features(zpool_handle_t *); extern int zpool_refresh_stats(zpool_handle_t *, boolean_t *); extern int zpool_get_errlog(zpool_handle_t *, nvlist_t **); extern boolean_t zpool_is_bootable(zpool_handle_t *); /* * Import and export functions */ extern int zpool_export(zpool_handle_t *, boolean_t, const char *); extern int zpool_export_force(zpool_handle_t *, const char *); extern int zpool_import(libzfs_handle_t *, nvlist_t *, const char *, char *altroot); extern int zpool_import_props(libzfs_handle_t *, nvlist_t *, const char *, nvlist_t *, int); extern void zpool_print_unsup_feat(nvlist_t *config); /* * Search for pools to import */ typedef struct importargs { char **path; /* a list of paths to search */ int paths; /* number of paths to search */ char *poolname; /* name of a pool to find */ uint64_t guid; /* guid of a pool to find */ char *cachefile; /* cachefile to use for import */ int can_be_active : 1; /* can the pool be active? */ int unique : 1; /* does 'poolname' already exist? */ int exists : 1; /* set on return if pool already exists */ nvlist_t *policy; /* load policy (max txg, rewind, etc.) */ } importargs_t; extern nvlist_t *zpool_search_import(libzfs_handle_t *, importargs_t *); extern int zpool_tryimport(libzfs_handle_t *hdl, char *target, nvlist_t **configp, importargs_t *args); /* legacy pool search routines */ extern nvlist_t *zpool_find_import(libzfs_handle_t *, int, char **); extern nvlist_t *zpool_find_import_cached(libzfs_handle_t *, const char *, char *, uint64_t); /* * Miscellaneous pool functions */ struct zfs_cmd; extern const char *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; extern char *zpool_vdev_name(libzfs_handle_t *, zpool_handle_t *, nvlist_t *, int name_flags); extern int zpool_upgrade(zpool_handle_t *, uint64_t); extern int zpool_get_history(zpool_handle_t *, nvlist_t **); extern int zpool_history_unpack(char *, uint64_t, uint64_t *, nvlist_t ***, uint_t *); extern void zpool_obj_to_path(zpool_handle_t *, uint64_t, uint64_t, char *, size_t len); extern int zfs_ioctl(libzfs_handle_t *, int, struct zfs_cmd *); extern int zpool_get_physpath(zpool_handle_t *, char *, size_t); extern void zpool_explain_recover(libzfs_handle_t *, const char *, int, nvlist_t *); extern int zpool_checkpoint(zpool_handle_t *); extern int zpool_discard_checkpoint(zpool_handle_t *); /* * Basic handle manipulations. These functions do not create or destroy the * underlying datasets, only the references to them. */ extern zfs_handle_t *zfs_open(libzfs_handle_t *, const char *, int); extern zfs_handle_t *zfs_handle_dup(zfs_handle_t *); extern void zfs_close(zfs_handle_t *); extern zfs_type_t zfs_get_type(const zfs_handle_t *); extern const char *zfs_get_name(const zfs_handle_t *); extern zpool_handle_t *zfs_get_pool_handle(const zfs_handle_t *); extern 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 */ extern const char *zfs_prop_default_string(zfs_prop_t); extern uint64_t zfs_prop_default_numeric(zfs_prop_t); extern const char *zfs_prop_column_name(zfs_prop_t); extern boolean_t zfs_prop_align_right(zfs_prop_t); extern nvlist_t *zfs_valid_proplist(libzfs_handle_t *, zfs_type_t, nvlist_t *, uint64_t, zfs_handle_t *, zpool_handle_t *, const char *); extern const char *zfs_prop_to_name(zfs_prop_t); extern int zfs_prop_set(zfs_handle_t *, const char *, const char *); extern int zfs_prop_set_list(zfs_handle_t *, nvlist_t *); extern int zfs_prop_get(zfs_handle_t *, zfs_prop_t, char *, size_t, zprop_source_t *, char *, size_t, boolean_t); extern int zfs_prop_get_recvd(zfs_handle_t *, const char *, char *, size_t, boolean_t); extern int zfs_prop_get_numeric(zfs_handle_t *, zfs_prop_t, uint64_t *, zprop_source_t *, char *, size_t); extern int zfs_prop_get_userquota_int(zfs_handle_t *zhp, const char *propname, uint64_t *propvalue); extern int zfs_prop_get_userquota(zfs_handle_t *zhp, const char *propname, char *propbuf, int proplen, boolean_t literal); extern int zfs_prop_get_written_int(zfs_handle_t *zhp, const char *propname, uint64_t *propvalue); extern int zfs_prop_get_written(zfs_handle_t *zhp, const char *propname, char *propbuf, int proplen, boolean_t literal); extern int zfs_prop_get_feature(zfs_handle_t *zhp, const char *propname, char *buf, size_t len); extern uint64_t zfs_prop_get_int(zfs_handle_t *, zfs_prop_t); extern int zfs_prop_inherit(zfs_handle_t *, const char *, boolean_t); extern const char *zfs_prop_values(zfs_prop_t); extern int zfs_prop_is_string(zfs_prop_t prop); extern nvlist_t *zfs_get_user_props(zfs_handle_t *); extern nvlist_t *zfs_get_recvd_props(zfs_handle_t *); extern nvlist_t *zfs_get_clones_nvl(zfs_handle_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; extern int zfs_expand_proplist(zfs_handle_t *, zprop_list_t **, boolean_t, boolean_t); extern void zfs_prune_proplist(zfs_handle_t *, uint8_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 */ extern int zpool_expand_proplist(zpool_handle_t *, zprop_list_t **); extern int zpool_prop_get_feature(zpool_handle_t *, const char *, char *, size_t); extern const char *zpool_prop_default_string(zpool_prop_t); extern uint64_t zpool_prop_default_numeric(zpool_prop_t); extern const char *zpool_prop_column_name(zpool_prop_t); extern boolean_t zpool_prop_align_right(zpool_prop_t); /* * Functions shared by zfs and zpool property management. */ extern int zprop_iter(zprop_func func, void *cb, boolean_t show_all, boolean_t ordered, zfs_type_t type); extern int zprop_get_list(libzfs_handle_t *, char *, zprop_list_t **, zfs_type_t); extern 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 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; } zprop_get_cbdata_t; void zprop_print_one_property(const char *, zprop_get_cbdata_t *, const char *, const char *, zprop_source_t, const char *, const char *); /* * Iterator functions. */ typedef int (*zfs_iter_f)(zfs_handle_t *, void *); extern int zfs_iter_root(libzfs_handle_t *, zfs_iter_f, void *); extern int zfs_iter_children(zfs_handle_t *, zfs_iter_f, void *); extern int zfs_iter_dependents(zfs_handle_t *, boolean_t, zfs_iter_f, void *); extern int zfs_iter_filesystems(zfs_handle_t *, zfs_iter_f, void *); extern int zfs_iter_snapshots(zfs_handle_t *, boolean_t, zfs_iter_f, void *); extern int zfs_iter_snapshots_sorted(zfs_handle_t *, zfs_iter_f, void *); extern int zfs_iter_snapspec(zfs_handle_t *, const char *, zfs_iter_f, void *); extern int zfs_iter_bookmarks(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; void zfs_foreach_mountpoint(libzfs_handle_t *, zfs_handle_t **, size_t, zfs_iter_f, void *, boolean_t); void libzfs_add_handle(get_all_cb_t *, zfs_handle_t *); /* * Functions to create and destroy datasets. */ extern int zfs_create(libzfs_handle_t *, const char *, zfs_type_t, nvlist_t *); extern int zfs_create_ancestors(libzfs_handle_t *, const char *); extern int zfs_destroy(zfs_handle_t *, boolean_t); extern int zfs_destroy_snaps(zfs_handle_t *, char *, boolean_t); extern int zfs_destroy_snaps_nvl(libzfs_handle_t *, nvlist_t *, boolean_t); extern int zfs_clone(zfs_handle_t *, const char *, nvlist_t *); extern int zfs_snapshot(libzfs_handle_t *, const char *, boolean_t, nvlist_t *); extern int zfs_snapshot_nvl(libzfs_handle_t *hdl, nvlist_t *snaps, nvlist_t *props); extern int zfs_rollback(zfs_handle_t *, zfs_handle_t *, boolean_t); extern int zfs_rename(zfs_handle_t *, const char *, boolean_t, boolean_t); typedef struct sendflags { /* print informational messages (ie, -v was specified) */ boolean_t verbose; /* recursive send (ie, -R) */ boolean_t replicate; /* for incrementals, do all intermediate snapshots */ boolean_t doall; /* if dataset is a clone, do incremental from its origin */ boolean_t fromorigin; /* do deduplication */ boolean_t dedup; /* 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; /* 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; } sendflags_t; typedef boolean_t (snapfilter_cb_t)(zfs_handle_t *, void *); extern int zfs_send(zfs_handle_t *, const char *, const char *, sendflags_t *, int, snapfilter_cb_t, void *, nvlist_t **); extern int zfs_send_one(zfs_handle_t *, const char *, int, enum lzc_send_flags); extern int zfs_send_resume(libzfs_handle_t *, sendflags_t *, int outfd, const char *); extern nvlist_t *zfs_send_resume_token_to_nvlist(libzfs_handle_t *hdl, const char *token); extern int zfs_promote(zfs_handle_t *); extern int zfs_hold(zfs_handle_t *, const char *, const char *, boolean_t, int); extern int zfs_hold_nvl(zfs_handle_t *, int, nvlist_t *); extern int zfs_release(zfs_handle_t *, const char *, const char *, boolean_t); extern int zfs_get_holds(zfs_handle_t *, nvlist_t **); extern uint64_t zvol_volsize_to_reservation(uint64_t, nvlist_t *); typedef int (*zfs_userspace_cb_t)(void *arg, const char *domain, uid_t rid, uint64_t space); extern int zfs_userspace(zfs_handle_t *, zfs_userquota_prop_t, zfs_userspace_cb_t, void *); extern int zfs_get_fsacl(zfs_handle_t *, nvlist_t **); extern 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; } recvflags_t; extern int zfs_receive(libzfs_handle_t *, const char *, nvlist_t *, recvflags_t *, int, avl_tree_t *); typedef enum diff_flags { ZFS_DIFF_PARSEABLE = 0x1, ZFS_DIFF_TIMESTAMP = 0x2, ZFS_DIFF_CLASSIFY = 0x4 } diff_flags_t; extern int zfs_show_diffs(zfs_handle_t *, int, const char *, const char *, int); /* * Miscellaneous functions. */ extern const char *zfs_type_to_name(zfs_type_t); extern void zfs_refresh_properties(zfs_handle_t *); extern int zfs_name_valid(const char *, zfs_type_t); extern zfs_handle_t *zfs_path_to_zhandle(libzfs_handle_t *, char *, zfs_type_t); extern boolean_t zfs_dataset_exists(libzfs_handle_t *, const char *, zfs_type_t); extern int zfs_spa_version(zfs_handle_t *, int *); extern boolean_t zfs_bookmark_exists(const char *path); extern ulong_t get_system_hostid(void); /* * Mount support functions. */ extern boolean_t is_mounted(libzfs_handle_t *, const char *special, char **); extern boolean_t zfs_is_mounted(zfs_handle_t *, char **); extern int zfs_mount(zfs_handle_t *, const char *, int); extern int zfs_unmount(zfs_handle_t *, const char *, int); extern int zfs_unmountall(zfs_handle_t *, int); /* * Share support functions. */ extern boolean_t zfs_is_shared(zfs_handle_t *); extern int zfs_share(zfs_handle_t *); extern int zfs_unshare(zfs_handle_t *); /* * Protocol-specific share support functions. */ extern boolean_t zfs_is_shared_nfs(zfs_handle_t *, char **); extern boolean_t zfs_is_shared_smb(zfs_handle_t *, char **); extern int zfs_share_nfs(zfs_handle_t *); extern int zfs_share_smb(zfs_handle_t *); extern int zfs_shareall(zfs_handle_t *); extern int zfs_unshare_nfs(zfs_handle_t *, const char *); extern int zfs_unshare_smb(zfs_handle_t *, const char *); extern int zfs_unshareall_nfs(zfs_handle_t *); extern int zfs_unshareall_smb(zfs_handle_t *); extern int zfs_unshareall_bypath(zfs_handle_t *, const char *); extern int zfs_unshareall(zfs_handle_t *); extern int zfs_deleg_share_nfs(libzfs_handle_t *, char *, char *, char *, void *, void *, int, zfs_share_op_t); /* * When dealing with nvlists, verify() is extremely useful */ #ifdef NDEBUG #define verify(EX) ((void)(EX)) #else #define verify(EX) assert(EX) #endif /* * Utility function to convert a number to a human-readable form. */ extern void zfs_nicenum(uint64_t, char *, size_t); extern int zfs_nicestrtonum(libzfs_handle_t *, const char *, uint64_t *); /* * Given a device or file, determine if it is part of a pool. */ extern int zpool_in_use(libzfs_handle_t *, int, pool_state_t *, char **, boolean_t *); /* * Label manipulation. */ extern int zpool_read_label(int, nvlist_t **); extern int zpool_clear_label(int); /* is this zvol valid for use as a dump device? */ extern int zvol_check_dump_config(char *); /* * Management interfaces for SMB ACL files */ int zfs_smb_acl_add(libzfs_handle_t *, char *, char *, char *); int zfs_smb_acl_remove(libzfs_handle_t *, char *, char *, char *); int zfs_smb_acl_purge(libzfs_handle_t *, char *, char *); 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. */ extern int zpool_enable_datasets(zpool_handle_t *, const char *, int); extern int zpool_disable_datasets(zpool_handle_t *, boolean_t); /* * Mappings between vdev and FRU. */ extern void libzfs_fru_refresh(libzfs_handle_t *); extern const char *libzfs_fru_lookup(libzfs_handle_t *, const char *); extern const char *libzfs_fru_devpath(libzfs_handle_t *, const char *); extern boolean_t libzfs_fru_compare(libzfs_handle_t *, const char *, const char *); extern boolean_t libzfs_fru_notself(libzfs_handle_t *, const char *); extern int zpool_fru_set(zpool_handle_t *, uint64_t, const char *); extern int zfs_get_hole_count(const char *, uint64_t *, uint64_t *); extern int zfs_remap_indirects(libzfs_handle_t *hdl, const char *); /* Allow consumers to initialize libshare externally for optimal performance */ extern int zfs_init_libshare_arg(libzfs_handle_t *, int, void *); /* * For most consumers, zfs_init_libshare_arg is sufficient on its own, and * zfs_uninit_libshare is unnecessary. zfs_uninit_libshare should only be called * if the caller has already initialized libshare for one set of zfs handles, * and wishes to share or unshare filesystems outside of that set. In that case, * the caller should uninitialize libshare, and then re-initialize it with the * new handles being shared or unshared. */ extern void zfs_uninit_libshare(libzfs_handle_t *); #ifdef __cplusplus } #endif #endif /* _LIBZFS_H */ Index: vendor/illumos/dist/lib/libzfs/common/libzfs_pool.c =================================================================== --- vendor/illumos/dist/lib/libzfs/common/libzfs_pool.c (revision 354960) +++ vendor/illumos/dist/lib/libzfs/common/libzfs_pool.c (revision 354961) @@ -1,4568 +1,4574 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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, 2017 by Delphix. All rights reserved. * Copyright (c) 2013, Joyent, Inc. All rights reserved. * Copyright 2016 Nexenta Systems, Inc. * Copyright 2016 Igor Kozhukhov * Copyright (c) 2017 Datto Inc. * Copyright (c) 2017, Intel Corporation. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zfs_namecheck.h" #include "zfs_prop.h" #include "libzfs_impl.h" #include "zfs_comutil.h" #include "zfeature_common.h" static int read_efi_label(nvlist_t *, diskaddr_t *, boolean_t *); static boolean_t zpool_vdev_is_interior(const char *name); #define BACKUP_SLICE "s2" typedef struct prop_flags { int create:1; /* Validate property on creation */ int import:1; /* Validate property on import */ } 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)); if (zcmd_alloc_dst_nvlist(hdl, &zc, 0) != 0) return (-1); while (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_GET_PROPS, &zc) != 0) { if (errno == ENOMEM) { if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) { zcmd_free_nvlists(&zc); return (-1); } } 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); } static 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 char * zpool_get_prop_string(zpool_handle_t *zhp, zpool_prop_t prop, zprop_source_t *src) { nvlist_t *nv, *nvl; uint64_t ival; char *value; zprop_source_t source; nvl = zhp->zpool_props; if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) { verify(nvlist_lookup_uint64(nv, ZPROP_SOURCE, &ival) == 0); source = ival; verify(nvlist_lookup_string(nv, ZPROP_VALUE, &value) == 0); } else { source = ZPROP_SRC_DEFAULT; if ((value = (char *)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) { verify(nvlist_lookup_uint64(nv, ZPROP_SOURCE, &value) == 0); source = value; verify(nvlist_lookup_uint64(nv, ZPROP_VALUE, &value) == 0); } 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) { 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")); } /* * 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; nvlist_t *nvroot; vdev_stat_t *vs; uint_t vsc; 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, "FAULTED", len); break; case ZPOOL_PROP_GUID: intval = zpool_get_prop_int(zhp, prop, &src); (void) snprintf(buf, len, "%llu", intval); break; case ZPOOL_PROP_ALTROOT: case ZPOOL_PROP_CACHEFILE: case ZPOOL_PROP_COMMENT: if (zhp->zpool_props != NULL || zpool_get_all_props(zhp) == 0) { (void) strlcpy(buf, zpool_get_prop_string(zhp, prop, &src), len); break; } /* 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: if (literal) { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } else { (void) zfs_nicenum(intval, buf, len); } break; case ZPOOL_PROP_BOOTSIZE: 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_nicenum(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 { (void) snprintf(buf, len, "%llu%%", (u_longlong_t)intval); } break; case ZPOOL_PROP_DEDUPRATIO: (void) snprintf(buf, len, "%llu.%02llux", (u_longlong_t)(intval / 100), (u_longlong_t)(intval % 100)); break; case ZPOOL_PROP_HEALTH: verify(nvlist_lookup_nvlist(zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc) == 0); (void) strlcpy(buf, zpool_state_to_name(intval, vs->vs_aux), len); break; case ZPOOL_PROP_VERSION: if (intval >= SPA_VERSION_FEATURES) { (void) snprintf(buf, len, "-"); break; } /* FALLTHROUGH */ default: (void) snprintf(buf, len, "%llu", 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); } /* * 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, char *bootfs) { int len = strlen(pool); 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); } boolean_t zpool_is_bootable(zpool_handle_t *zhp) { char bootfs[ZFS_MAX_DATASET_NAME_LEN]; return (zpool_get_prop(zhp, ZPOOL_PROP_BOOTFS, bootfs, sizeof (bootfs), NULL, B_FALSE) == 0 && strncmp(bootfs, "-", sizeof (bootfs)) != 0); } /* * 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; char *strval; uint64_t intval; char *slash, *check; struct stat64 statbuf; zpool_handle_t *zhp; 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); 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, "invalid feature '%s'"), 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) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' can only be set to " "'enabled'"), 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; } /* * 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 (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 %d is invalid."), propname, intval); (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); goto error; } break; case ZPOOL_PROP_BOOTSIZE: if (!flags.create) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' can only be set during pool " "creation"), propname); (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 (strval[0] != '\0' && !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; } *slash = '\0'; if (strval[0] != '\0' && (stat64(strval, &statbuf) != 0 || !S_ISDIR(statbuf.st_mode))) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is not a valid directory"), strval); (void) zfs_error(hdl, EZFS_BADPATH, errbuf); goto error; } *slash = '/'; 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_TNAME: if (!flags.create) { 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; } 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; default: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s'(%d) not defined"), propname, prop); 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[1024]; 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)); if (zcmd_write_src_nvlist(zhp->zpool_hdl, &zc, nvl) != 0) { nvlist_free(nvl); return (-1); } 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) { libzfs_handle_t *hdl = zhp->zpool_hdl; zprop_list_t *entry; char buf[ZFS_MAXPROPLEN]; nvlist_t *features = NULL; zprop_list_t **last; boolean_t firstexpand = (NULL == *plp); if (zprop_expand_list(hdl, plp, ZFS_TYPE_POOL) != 0) return (-1); last = plp; while (*last != NULL) last = &(*last)->pl_next; if ((*plp)->pl_all) features = zpool_get_features(zhp); if ((*plp)->pl_all && firstexpand) { for (int i = 0; i < SPA_FEATURES; i++) { zprop_list_t *entry = zfs_alloc(hdl, sizeof (zprop_list_t)); entry->pl_prop = ZPROP_INVAL; 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 (nvpair_t *nvp = nvlist_next_nvpair(features, NULL); nvp != NULL; nvp = nvlist_next_nvpair(features, nvp)) { char *propname; boolean_t found; zprop_list_t *entry; 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_INVAL; 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) continue; if (entry->pl_prop != ZPROP_INVAL && zpool_get_prop(zhp, entry->pl_prop, buf, sizeof (buf), NULL, B_FALSE) == 0) { if (strlen(buf) > entry->pl_width) entry->pl_width = strlen(buf); } } 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 * unecessary 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); } /* * Don't start the slice at the default block of 34; many storage * devices will use a stripe width of 128k, so start there instead. */ #define NEW_START_BLOCK 256 /* * 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, "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; 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); } if ((zhp = zfs_alloc(hdl, sizeof (zpool_handle_t))) == NULL) return (NULL); 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; if ((zhp = zfs_alloc(hdl, sizeof (zpool_handle_t))) == NULL) return (-1); 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++) { 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); } /* * 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; char msg[1024]; int ret = -1; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot create '%s'"), pool); if (!zpool_name_valid(hdl, B_FALSE, pool)) return (zfs_error(hdl, EZFS_INVALIDNAME, msg)); if (zcmd_write_conf_nvlist(hdl, &zc, nvroot) != 0) return (-1); 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, msg)) == NULL) { goto create_failed; } } if (fsprops) { uint64_t zoned; 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, msg)) == 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, msg); goto create_failed; } if (!zc_props && (nvlist_alloc(&zc_props, NV_UNIQUE_NAME, 0) != 0)) { goto create_failed; } if (nvlist_add_nvlist(zc_props, ZPOOL_ROOTFS_PROPS, zc_fsprops) != 0) { goto create_failed; } } if (zc_props && zcmd_write_src_nvlist(hdl, &zc, zc_props) != 0) goto create_failed; (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); 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")); return (zfs_error(hdl, EZFS_BADDEV, msg)); 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, msg)); 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_nicenum(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, msg)); case ENOSPC: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more devices is out of space")); return (zfs_error(hdl, EZFS_BADDEV, msg)); case ENOTBLK: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cache device must be a disk or disk slice")); return (zfs_error(hdl, EZFS_BADDEV, msg)); default: return (zpool_standard_error(hdl, errno, msg)); } } create_failed: zcmd_free_nvlists(&zc); nvlist_free(zc_props); nvlist_free(zc_fsprops); 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 msg[1024]; 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(msg, sizeof (msg), 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, msg); } else { (void) zpool_standard_error(hdl, errno, msg); } 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 msg[1024]; int error; error = lzc_pool_checkpoint(zhp->zpool_name); if (error != 0) { (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot checkpoint '%s'"), zhp->zpool_name); (void) zpool_standard_error(hdl, error, msg); 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 msg[1024]; int error; error = lzc_pool_checkpoint_discard(zhp->zpool_name); if (error != 0) { (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot discard checkpoint in '%s'"), zhp->zpool_name); (void) zpool_standard_error(hdl, error, msg); 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 msg[1024]; nvlist_t **spares, **l2cache; uint_t nspares, nl2cache; (void) snprintf(msg, sizeof (msg), 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, msg)); } 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, msg)); } if (zcmd_write_conf_nvlist(hdl, &zc, nvroot) != 0) return (-1); (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, msg); break; case EINVAL: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid config; a pool with removing/removed " "vdevs does not support adding raidz vdevs")); (void) zfs_error(hdl, EZFS_BADDEV, msg); break; case EOVERFLOW: /* * This occurrs 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_nicenum(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, msg); break; case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded to add these vdevs")); (void) zfs_error(hdl, EZFS_BADVERSION, msg); break; case EDOM: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "root pool can not have multiple vdevs" " or separate logs")); (void) zfs_error(hdl, EZFS_POOL_NOTSUP, msg); break; case ENOTBLK: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cache device must be a disk or disk slice")); (void) zfs_error(hdl, EZFS_BADDEV, msg); break; default: (void) zpool_standard_error(hdl, errno, msg); } 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 }; char msg[1024]; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot export '%s'"), zhp->zpool_name); (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(zhp->zpool_hdl, EZFS_ACTIVE_SPARE, msg)); default: return (zpool_standard_error_fmt(zhp->zpool_hdl, errno, msg)); } } 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, 0, &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", (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", 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, 0, &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. "), (loss + 30) / 60); } else if (loss > 0) { (void) printf(dgettext(TEXT_DOMAIN, "Approximately %lld seconds of data\n" "\tmust be discarded, irreversibly. "), 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; verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0); verify(nvlist_lookup_nvlist(nvinfo, ZPOOL_CONFIG_UNSUP_FEAT, &unsup_feat) == 0); for (nvpair_t *nvp = nvlist_next_nvpair(unsup_feat, NULL); nvp != NULL; nvp = nvlist_next_nvpair(unsup_feat, nvp)) { char *desc; verify(nvpair_type(nvp) == DATA_TYPE_STRING); verify(nvpair_value_string(nvp, &desc) == 0); 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; char *thename; char *origname; int ret; int error = 0; char errbuf[1024]; verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &origname) == 0); (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 = (char *)newname; } else { thename = origname; } if (props != NULL) { uint64_t version; prop_flags_t flags = { .create = B_FALSE, .import = B_TRUE }; verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); if ((props = zpool_valid_proplist(hdl, origname, props, version, flags, errbuf)) == NULL) return (-1); if (zcmd_write_src_nvlist(hdl, &zc, props) != 0) { nvlist_free(props); return (-1); } nvlist_free(props); } (void) strlcpy(zc.zc_name, thename, sizeof (zc.zc_name)); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &zc.zc_guid) == 0); if (zcmd_write_conf_nvlist(hdl, &zc, config) != 0) { zcmd_free_nvlists(&zc); return (-1); } if (zcmd_alloc_dst_nvlist(hdl, &zc, zc.zc_nvlist_conf_size * 2) != 0) { zcmd_free_nvlists(&zc); return (-1); } zc.zc_cookie = flags; while ((ret = zfs_ioctl(hdl, ZFS_IOC_POOL_IMPORT, &zc)) != 0 && errno == ENOMEM) { if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) { zcmd_free_nvlists(&zc); return (-1); } } 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) { char *hostname = ""; 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.\n")); } (void) zfs_error_aux(hdl, 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 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 (0); } return (ret); } /* * 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 msg[1024]; 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(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot pause scrubbing %s"), zc.zc_name); } else { assert(cmd == POOL_SCRUB_NORMAL); (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot scrub %s"), zc.zc_name); } + } else if (func == POOL_SCAN_RESILVER) { + assert(cmd == POOL_SCRUB_NORMAL); + (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, + "cannot restart resilver on %s"), zc.zc_name); } else if (func == POOL_SCAN_NONE) { (void) snprintf(msg, sizeof (msg), 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; verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &psc); if (ps && ps->pss_func == POOL_SCAN_SCRUB) { if (cmd == POOL_SCRUB_PAUSE) return (zfs_error(hdl, EZFS_SCRUB_PAUSED, msg)); else return (zfs_error(hdl, EZFS_SCRUBBING, msg)); } else { return (zfs_error(hdl, EZFS_RESILVERING, msg)); } } else if (err == ENOENT) { return (zfs_error(hdl, EZFS_NO_SCRUB, msg)); + } else if (err == ENOTSUP && func == POOL_SCAN_RESILVER) { + return (zfs_error(hdl, EZFS_NO_RESILVER_DEFER, msg)); } else { return (zpool_standard_error(hdl, err, msg)); } } 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. */ int zpool_initialize(zpool_handle_t *zhp, pool_initialize_func_t cmd_type, nvlist_t *vds) { char msg[1024]; libzfs_handle_t *hdl = zhp->zpool_hdl; nvlist_t *errlist; /* translate vdev names to guids */ nvlist_t *vdev_guids = fnvlist_alloc(); nvlist_t *guids_to_paths = fnvlist_alloc(); boolean_t spare, cache; nvlist_t *tgt; nvpair_t *elem; for (elem = nvlist_next_nvpair(vds, NULL); elem != NULL; elem = nvlist_next_nvpair(vds, elem)) { char *vd_path = nvpair_name(elem); tgt = zpool_find_vdev(zhp, vd_path, &spare, &cache, NULL); if ((tgt == NULL) || cache || spare) { (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot initialize '%s'"), vd_path); int err = (tgt == NULL) ? EZFS_NODEVICE : (spare ? EZFS_ISSPARE : EZFS_ISL2CACHE); fnvlist_free(vdev_guids); fnvlist_free(guids_to_paths); return (zfs_error(hdl, err, msg)); } uint64_t guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); fnvlist_add_uint64(vdev_guids, vd_path, guid); (void) snprintf(msg, sizeof (msg), "%llu", guid); fnvlist_add_string(guids_to_paths, msg, vd_path); } int err = lzc_initialize(zhp->zpool_name, cmd_type, vdev_guids, &errlist); fnvlist_free(vdev_guids); if (err == 0) { fnvlist_free(guids_to_paths); return (0); } nvlist_t *vd_errlist = NULL; if (errlist != NULL) { vd_errlist = fnvlist_lookup_nvlist(errlist, ZPOOL_INITIALIZE_VDEVS); } (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "operation failed")); 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)); char *path = fnvlist_lookup_string(guids_to_paths, nvpair_name(elem)); (void) zfs_error_fmt(hdl, vd_error, "cannot initialize '%s'", path); } fnvlist_free(guids_to_paths); if (vd_errlist != NULL) return (-1); return (zpool_standard_error(hdl, err, msg)); } /* * This provides a very minimal check whether a given string is likely a * c#t#d# style string. Users of this are expected to do their own * verification of the s# part. */ #define CTD_CHECK(str) (str && str[0] == 'c' && isdigit(str[1])) /* * More elaborate version for ones which may start with "/dev/dsk/" * and the like. */ static int ctd_check_path(char *str) { /* * If it starts with a slash, check the last component. */ if (str && str[0] == '/') { char *tmp = strrchr(str, '/'); /* * If it ends in "/old", check the second-to-last * component of the string instead. */ if (tmp != str && strcmp(tmp, "/old") == 0) { for (tmp--; *tmp != '/'; tmp--) ; } str = tmp + 1; } return (CTD_CHECK(str)); } /* * 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; 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, theguid; verify(nvpair_value_uint64(pair, &srchval) == 0); verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &theguid) == 0); if (theguid == srchval) return (nv); } break; case DATA_TYPE_STRING: { char *srchval, *val; verify(nvpair_value_string(pair, &srchval) == 0); if (nvlist_lookup_string(nv, srchkey, &val) != 0) break; /* * Search for the requested value. Special cases: * * - ZPOOL_CONFIG_PATH for whole disk entries. To support * UEFI boot, these end in "s0" or "s0/old" or "s1" or * "s1/old". The "s0" or "s1" part is hidden from the user, * but included in the string, so this matches around it. * - 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 && ctd_check_path(val)) { uint64_t wholedisk = 0; (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &wholedisk); if (wholedisk) { int slen = strlen(srchval); int vlen = strlen(val); if (slen != vlen - 2) break; /* * make_leaf_vdev() should only set * wholedisk for ZPOOL_CONFIG_PATHs which * will include "/dev/dsk/", giving plenty of * room for the indices used next. */ ASSERT(vlen >= 6); /* * strings identical except trailing "s0" */ if ((strcmp(&val[vlen - 2], "s0") == 0 || strcmp(&val[vlen - 2], "s1") == 0) && strncmp(srchval, val, slen) == 0) return (nv); /* * strings identical except trailing "s0/old" */ if ((strcmp(&val[vlen - 6], "s0/old") == 0 || strcmp(&val[vlen - 6], "s1/old") == 0) && strcmp(&srchval[slen - 4], "/old") == 0 && strncmp(srchval, val, slen - 4) == 0) return (nv); break; } } else if (strcmp(srchkey, ZPOOL_CONFIG_TYPE) == 0 && val) { 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)); verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID, &id) == 0); errno = 0; vdev_id = strtoull(idx, &end, 10); 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 (minus the "/devices" prefix), 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; verify(nvlist_alloc(&search, NV_UNIQUE_NAME, KM_SLEEP) == 0); verify(nvlist_add_string(search, ZPOOL_CONFIG_PHYS_PATH, ppath) == 0); verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); *avail_spare = B_FALSE; *l2cache = B_FALSE; if (log != NULL) *log = B_FALSE; ret = vdev_to_nvlist_iter(nvroot, search, avail_spare, l2cache, log); nvlist_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_MIRROR, strlen(VDEV_TYPE_MIRROR)) == 0) 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 buf[MAXPATHLEN]; char *end; nvlist_t *nvroot, *search, *ret; uint64_t guid; verify(nvlist_alloc(&search, NV_UNIQUE_NAME, KM_SLEEP) == 0); guid = strtoull(path, &end, 10); if (guid != 0 && *end == '\0') { verify(nvlist_add_uint64(search, ZPOOL_CONFIG_GUID, guid) == 0); } else if (zpool_vdev_is_interior(path)) { verify(nvlist_add_string(search, ZPOOL_CONFIG_TYPE, path) == 0); } else if (path[0] != '/') { (void) snprintf(buf, sizeof (buf), "%s/%s", ZFS_DISK_ROOT, path); verify(nvlist_add_string(search, ZPOOL_CONFIG_PATH, buf) == 0); } else { verify(nvlist_add_string(search, ZPOOL_CONFIG_PATH, path) == 0); } verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); *avail_spare = B_FALSE; *l2cache = B_FALSE; if (log != NULL) *log = B_FALSE; ret = vdev_to_nvlist_iter(nvroot, search, avail_spare, l2cache, log); nvlist_free(search); return (ret); } static int vdev_is_online(nvlist_t *nv) { uint64_t ival; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_OFFLINE, &ival) == 0 || nvlist_lookup_uint64(nv, ZPOOL_CONFIG_FAULTED, &ival) == 0 || nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVED, &ival) == 0) return (0); return (1); } /* * Helper function for zpool_get_physpaths(). */ static int vdev_get_one_physpath(nvlist_t *config, char *physpath, size_t physpath_size, size_t *bytes_written) { size_t bytes_left, pos, rsz; char *tmppath; const char *format; if (nvlist_lookup_string(config, ZPOOL_CONFIG_PHYS_PATH, &tmppath) != 0) return (EZFS_NODEVICE); pos = *bytes_written; bytes_left = physpath_size - pos; format = (pos == 0) ? "%s" : " %s"; rsz = snprintf(physpath + pos, bytes_left, format, tmppath); *bytes_written += rsz; if (rsz >= bytes_left) { /* if physpath was not copied properly, clear it */ if (bytes_left != 0) { physpath[pos] = 0; } return (EZFS_NOSPC); } return (0); } static int vdev_get_physpaths(nvlist_t *nv, char *physpath, size_t phypath_size, size_t *rsz, boolean_t is_spare) { char *type; int ret; if (nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) != 0) return (EZFS_INVALCONFIG); if (strcmp(type, VDEV_TYPE_DISK) == 0) { /* * An active spare device has ZPOOL_CONFIG_IS_SPARE set. * For a spare vdev, we only want to boot from the active * spare device. */ if (is_spare) { uint64_t spare = 0; (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_SPARE, &spare); if (!spare) return (EZFS_INVALCONFIG); } if (vdev_is_online(nv)) { if ((ret = vdev_get_one_physpath(nv, physpath, phypath_size, rsz)) != 0) return (ret); } } else if (strcmp(type, VDEV_TYPE_MIRROR) == 0 || strcmp(type, VDEV_TYPE_RAIDZ) == 0 || strcmp(type, VDEV_TYPE_REPLACING) == 0 || (is_spare = (strcmp(type, VDEV_TYPE_SPARE) == 0))) { nvlist_t **child; uint_t count; int i, ret; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &count) != 0) return (EZFS_INVALCONFIG); for (i = 0; i < count; i++) { ret = vdev_get_physpaths(child[i], physpath, phypath_size, rsz, is_spare); if (ret == EZFS_NOSPC) return (ret); } } return (EZFS_POOL_INVALARG); } /* * Get phys_path for a root pool config. * Return 0 on success; non-zero on failure. */ static int zpool_get_config_physpath(nvlist_t *config, char *physpath, size_t phypath_size) { size_t rsz; nvlist_t *vdev_root; nvlist_t **child; uint_t count; char *type; rsz = 0; if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &vdev_root) != 0) return (EZFS_INVALCONFIG); if (nvlist_lookup_string(vdev_root, ZPOOL_CONFIG_TYPE, &type) != 0 || nvlist_lookup_nvlist_array(vdev_root, ZPOOL_CONFIG_CHILDREN, &child, &count) != 0) return (EZFS_INVALCONFIG); /* * root pool can only have a single top-level vdev. */ if (strcmp(type, VDEV_TYPE_ROOT) != 0 || count != 1) return (EZFS_POOL_INVALARG); (void) vdev_get_physpaths(child[0], physpath, phypath_size, &rsz, B_FALSE); /* No online devices */ if (rsz == 0) return (EZFS_NODEVICE); return (0); } /* * Get phys_path for a root pool * Return 0 on success; non-zero on failure. */ int zpool_get_physpath(zpool_handle_t *zhp, char *physpath, size_t phypath_size) { return (zpool_get_config_physpath(zhp->zpool_config, physpath, phypath_size)); } /* * If the device has being dynamically expanded then we need to relabel * the disk to use the new unallocated space. */ static int zpool_relabel_disk(libzfs_handle_t *hdl, const char *name) { char path[MAXPATHLEN]; char errbuf[1024]; int fd, error; int (*_efi_use_whole_disk)(int); if ((_efi_use_whole_disk = (int (*)(int))dlsym(RTLD_DEFAULT, "efi_use_whole_disk")) == NULL) return (-1); (void) snprintf(path, sizeof (path), "%s/%s", ZFS_RDISK_ROOT, name); if ((fd = open(path, O_RDWR | O_NDELAY)) < 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cannot " "relabel '%s': unable to open device"), name); return (zfs_error(hdl, EZFS_OPENFAILED, errbuf)); } /* * It's possible that we might encounter an error if the device * does not have any unallocated space left. If so, we simply * ignore that error and continue on. */ error = _efi_use_whole_disk(fd); (void) close(fd); if (error && error != VT_ENOSPC) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cannot " "relabel '%s': unable to read disk capacity"), name); return (zfs_error(hdl, EZFS_NOCAP, errbuf)); } return (0); } /* * 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 msg[1024]; char *pathname; nvlist_t *tgt; boolean_t avail_spare, l2cache, islog; libzfs_handle_t *hdl = zhp->zpool_hdl; if (flags & ZFS_ONLINE_EXPAND) { (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot expand %s"), path); } else { (void) snprintf(msg, sizeof (msg), 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, msg)); verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, msg)); 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, msg)); } if (wholedisk) { pathname += strlen(ZFS_DISK_ROOT) + 1; (void) zpool_relabel_disk(hdl, pathname); } } 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, msg)); } return (zpool_standard_error(hdl, errno, msg)); } *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 msg[1024]; nvlist_t *tgt; boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), 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, msg)); verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, msg)); 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, msg)); case EEXIST: /* * The log device has unplayed logs */ return (zfs_error(hdl, EZFS_UNPLAYED_LOGS, msg)); default: return (zpool_standard_error(hdl, errno, msg)); } } /* * 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 msg[1024]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot fault %llu"), 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 (ioctl(hdl->libzfs_fd, 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, msg)); default: return (zpool_standard_error(hdl, errno, msg)); } } /* * 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 msg[1024]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot degrade %llu"), 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 (ioctl(hdl->libzfs_fd, ZFS_IOC_VDEV_SET_STATE, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, msg)); } /* * 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; char *type; if (nvlist_lookup_nvlist_array(search, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { verify(nvlist_lookup_string(search, ZPOOL_CONFIG_TYPE, &type) == 0); if (strcmp(type, VDEV_TYPE_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) { zfs_cmd_t zc = { 0 }; char msg[1024]; 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; boolean_t rootpool = zpool_is_bootable(zhp); if (replacing) (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot replace %s with %s"), old_disk, new_disk); else (void) snprintf(msg, sizeof (msg), 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, msg)); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, msg)); if (l2cache) return (zfs_error(hdl, EZFS_ISL2CACHE, msg)); verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); zc.zc_cookie = replacing; 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, msg)); } verify(nvlist_lookup_nvlist(zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE, &config_root) == 0); 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, msg)); } free(newname); if (zcmd_write_conf_nvlist(hdl, &zc, nvroot) != 0) return (-1); ret = zfs_ioctl(hdl, ZFS_IOC_VDEV_ATTACH, &zc); zcmd_free_nvlists(&zc); if (ret == 0) { if (rootpool) { /* * XXX need a better way to prevent user from * booting up a half-baked vdev. */ (void) fprintf(stderr, dgettext(TEXT_DOMAIN, "Make " "sure to wait until resilver is done " "before rebooting.\n")); } 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 (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 { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "can only attach to mirrors and top-level " "disks")); } (void) zfs_error(hdl, EZFS_BADTARGET, msg); 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, msg); 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, msg); 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, msg); break; case EDOM: /* * The new device has a different alignment requirement. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "devices have different sector alignment")); (void) zfs_error(hdl, EZFS_BADDEV, msg); break; case ENAMETOOLONG: /* * The resulting top-level vdev spec won't fit in the label. */ (void) zfs_error(hdl, EZFS_DEVOVERFLOW, msg); break; default: (void) zpool_standard_error(hdl, errno, msg); } return (-1); } /* * Detach the specified device. */ int zpool_vdev_detach(zpool_handle_t *zhp, const char *path) { zfs_cmd_t zc = { 0 }; char msg[1024]; nvlist_t *tgt; boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), 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, msg)); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, msg)); if (l2cache) return (zfs_error(hdl, EZFS_ISL2CACHE, msg)); verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); 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, msg); break; case EBUSY: /* * There are no other replicas of this device. */ (void) zfs_error(hdl, EZFS_NOREPLICAS, msg); break; default: (void) zpool_standard_error(hdl, errno, msg); } 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 msg[1024]; 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; boolean_t freelist = B_FALSE, memory_err = B_TRUE; int retval = 0; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "Unable to split %s"), zhp->zpool_name); if (!zpool_name_valid(hdl, B_FALSE, newname)) return (zfs_error(hdl, EZFS_INVALIDNAME, msg)); if ((config = zpool_get_config(zhp, NULL)) == NULL) { (void) fprintf(stderr, gettext("Internal error: unable to " "retrieve pool configuration\n")); return (-1); } verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &tree) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &vers) == 0); 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, msg)) == NULL) 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; 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; verify(nvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE, &type) == 0); 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, msg); goto out; } 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; } /* 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, msg); 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, 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)); if (zcmd_write_conf_nvlist(hdl, &zc, newconfig) != 0) goto out; if (zc_props != NULL && zcmd_write_src_nvlist(hdl, &zc, zc_props) != 0) goto out; if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SPLIT, &zc) != 0) { retval = zpool_standard_error(hdl, errno, msg); 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 msg[1024]; nvlist_t *tgt; boolean_t avail_spare, l2cache, islog; libzfs_handle_t *hdl = zhp->zpool_hdl; uint64_t version; (void) snprintf(msg, sizeof (msg), 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, &islog)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, msg)); 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, msg)); } if (!islog && !avail_spare && !l2cache && zpool_is_bootable(zhp)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "root pool can not have removed devices, " "because GRUB does not understand them")); return (zfs_error(hdl, EINVAL, msg)); } 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, msg); break; case EBUSY: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Pool busy; removal may already be in progress")); (void) zfs_error(hdl, EZFS_BUSY, msg); break; default: (void) zpool_standard_error(hdl, errno, msg); } return (-1); } int zpool_vdev_remove_cancel(zpool_handle_t *zhp) { zfs_cmd_t zc = { 0 }; char msg[1024]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), 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, msg)); } int zpool_vdev_indirect_size(zpool_handle_t *zhp, const char *path, uint64_t *sizep) { char msg[1024]; nvlist_t *tgt; boolean_t avail_spare, l2cache, islog; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), 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, msg)); 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, msg)); } 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 msg[1024]; 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(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot clear errors for %s"), path); else (void) snprintf(msg, sizeof (msg), 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, msg)); /* * Don't allow error clearing for hot spares. Do allow * error clearing for l2cache devices. */ if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, msg)); verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); } zpool_get_load_policy(rewindnvl, &policy); zc.zc_cookie = policy.zlp_rewind; if (zcmd_alloc_dst_nvlist(hdl, &zc, zhp->zpool_config_size * 2) != 0) return (-1); if (zcmd_write_src_nvlist(hdl, &zc, rewindnvl) != 0) return (-1); while ((error = zfs_ioctl(hdl, ZFS_IOC_CLEAR, &zc)) != 0 && errno == ENOMEM) { if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) { zcmd_free_nvlists(&zc); return (-1); } } 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, msg)); } /* * 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 msg[1024]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot clear errors for %llx"), guid); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_guid = guid; zc.zc_cookie = ZPOOL_NO_REWIND; if (ioctl(hdl->libzfs_fd, ZFS_IOC_CLEAR, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, msg)); } /* * Change the GUID for a pool. */ int zpool_reguid(zpool_handle_t *zhp) { char msg[1024]; libzfs_handle_t *hdl = zhp->zpool_hdl; zfs_cmd_t zc = { 0 }; (void) snprintf(msg, sizeof (msg), 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, msg)); } /* * Reopen the pool. */ int zpool_reopen(zpool_handle_t *zhp) { zfs_cmd_t zc = { 0 }; char msg[1024]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot reopen '%s'"), zhp->zpool_name); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if (zfs_ioctl(hdl, ZFS_IOC_POOL_REOPEN, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, msg)); } /* 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); } /* * Convert from a devid string to a path. */ static char * devid_to_path(char *devid_str) { ddi_devid_t devid; char *minor; char *path; devid_nmlist_t *list = NULL; int ret; if (devid_str_decode(devid_str, &devid, &minor) != 0) return (NULL); ret = devid_deviceid_to_nmlist("/dev", devid, minor, &list); devid_str_free(minor); devid_free(devid); if (ret != 0) return (NULL); /* * In a case the strdup() fails, we will just return NULL below. */ path = strdup(list[0].devname); devid_free_nmlist(list); return (path); } /* * Convert from a path to a devid string. */ static char * path_to_devid(const char *path) { int fd; ddi_devid_t devid; char *minor, *ret; if ((fd = open(path, O_RDONLY)) < 0) return (NULL); minor = NULL; ret = NULL; if (devid_get(fd, &devid) == 0) { if (devid_get_minor_name(fd, &minor) == 0) ret = devid_str_encode(devid, minor); if (minor != NULL) devid_str_free(minor); devid_free(devid); } (void) close(fd); return (ret); } /* * Issue the necessary ioctl() to update the stored path value for the vdev. We * ignore any failure here, since a common case is for an unprivileged user to * type 'zpool status', and we'll display the correct information anyway. */ static void set_path(zpool_handle_t *zhp, nvlist_t *nv, const char *path) { zfs_cmd_t zc = { 0 }; (void) strncpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); (void) strncpy(zc.zc_value, path, sizeof (zc.zc_value)); verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); (void) ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_SETPATH, &zc); } /* * 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) { char *path, *devid, *env; uint64_t value; char buf[64]; vdev_stat_t *vs; uint_t vsc; env = getenv("ZPOOL_VDEV_NAME_PATH"); if (env && (strtoul(env, NULL, 0) > 0 || !strncasecmp(env, "YES", 3) || !strncasecmp(env, "ON", 2))) name_flags |= VDEV_NAME_PATH; env = getenv("ZPOOL_VDEV_NAME_GUID"); if (env && (strtoul(env, NULL, 0) > 0 || !strncasecmp(env, "YES", 3) || !strncasecmp(env, "ON", 2))) name_flags |= VDEV_NAME_GUID; env = getenv("ZPOOL_VDEV_NAME_FOLLOW_LINKS"); if (env && (strtoul(env, NULL, 0) > 0 || !strncasecmp(env, "YES", 3) || !strncasecmp(env, "ON", 2))) name_flags |= VDEV_NAME_FOLLOW_LINKS; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &value) == 0 || name_flags & VDEV_NAME_GUID) { 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, &path) == 0) { /* * If the device is dead (faulted, offline, etc) then don't * bother opening it. Otherwise we may be forcing the user to * open a misbehaving device, which can have undesirable * effects. */ if ((nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc) != 0 || vs->vs_state >= VDEV_STATE_DEGRADED) && zhp != NULL && nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &devid) == 0) { /* * Determine if the current path is correct. */ char *newdevid = path_to_devid(path); if (newdevid == NULL || strcmp(devid, newdevid) != 0) { char *newpath; if ((newpath = devid_to_path(devid)) != NULL) { /* * Update the path appropriately. */ set_path(zhp, nv, newpath); if (nvlist_add_string(nv, ZPOOL_CONFIG_PATH, newpath) == 0) verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0); free(newpath); } } if (newdevid) devid_str_free(newdevid); } if (name_flags & VDEV_NAME_FOLLOW_LINKS) { char *rp = realpath(path, NULL); if (rp) { strlcpy(buf, rp, sizeof (buf)); path = buf; free(rp); } } if (strncmp(path, ZFS_DISK_ROOTD, strlen(ZFS_DISK_ROOTD)) == 0) path += strlen(ZFS_DISK_ROOTD); /* * Remove the partition from the path it this is a whole disk. */ if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &value) == 0 && value && !(name_flags & VDEV_NAME_PATH)) { int pathlen = strlen(path); char *tmp = zfs_strdup(hdl, path); /* * If it starts with c#, and ends with "s0" or "s1", * chop the slice off, or if it ends with "s0/old" or * "s1/old", remove the slice from the middle. */ if (CTD_CHECK(tmp)) { if (strcmp(&tmp[pathlen - 2], "s0") == 0 || strcmp(&tmp[pathlen - 2], "s1") == 0) { tmp[pathlen - 2] = '\0'; } else if (pathlen > 6 && (strcmp(&tmp[pathlen - 6], "s0/old") == 0 || strcmp(&tmp[pathlen - 6], "s1/old") == 0)) { (void) strcpy(&tmp[pathlen - 6], "/old"); } } return (tmp); } } else { verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &path) == 0); /* * If it's a raidz device, we need to stick in the parity level. */ if (strcmp(path, VDEV_TYPE_RAIDZ) == 0) { verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY, &value) == 0); (void) snprintf(buf, sizeof (buf), "%s%llu", path, (u_longlong_t)value); path = buf; } /* * We identify each top-level vdev by using a * naming convention. */ if (name_flags & VDEV_NAME_TYPE_ID) { uint64_t id; verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID, &id) == 0); (void) snprintf(buf, sizeof (buf), "%s-%llu", path, (u_longlong_t)id); path = buf; } } 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 }; uint64_t count; zbookmark_phys_t *zb = NULL; int i; /* * Retrieve the raw error list from the kernel. If the number of errors * has increased, allocate more space and continue until we get the * entire list. */ verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_ERRCOUNT, &count) == 0); if (count == 0) return (0); if ((zc.zc_nvlist_dst = (uintptr_t)zfs_alloc(zhp->zpool_hdl, count * sizeof (zbookmark_phys_t))) == (uintptr_t)NULL) return (-1); zc.zc_nvlist_dst_size = count; (void) strcpy(zc.zc_name, zhp->zpool_name); for (;;) { if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_ERROR_LOG, &zc) != 0) { free((void *)(uintptr_t)zc.zc_nvlist_dst); if (errno == ENOMEM) { void *dst; count = zc.zc_nvlist_dst_size; dst = zfs_alloc(zhp->zpool_hdl, count * sizeof (zbookmark_phys_t)); if (dst == NULL) return (-1); zc.zc_nvlist_dst = (uintptr_t)dst; } else { return (-1); } } 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 boomarks * _not_ copied as part of the process. So we point the start of our * array appropriate and decrement the total number of elements. */ zb = ((zbookmark_phys_t *)(uintptr_t)zc.zc_nvlist_dst) + zc.zc_nvlist_dst_size; count -= zc.zc_nvlist_dst_size; qsort(zb, count, 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 (i = 0; i < count; 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((void *)(uintptr_t)zc.zc_nvlist_dst); return (0); nomem: free((void *)(uintptr_t)zc.zc_nvlist_dst); 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) { (void) strlcpy(string, basename(argv[0]), len); for (int 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; int err; args = fnvlist_alloc(); fnvlist_add_string(args, "message", message); err = zcmd_write_src_nvlist(hdl, &zc, args); if (err == 0) err = ioctl(hdl->libzfs_fd, 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 (ioctl(hdl->libzfs_fd, 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); } /* * Process the buffer of nvlists, unpacking and storing each nvlist record * into 'records'. 'leftover' is set to the number of bytes that weren't * processed as there wasn't a complete record. */ int zpool_history_unpack(char *buf, uint64_t bytes_read, uint64_t *leftover, nvlist_t ***records, uint_t *numrecords) { uint64_t reclen; nvlist_t *nv; int i; while (bytes_read > sizeof (reclen)) { /* get length of packed record (stored as little endian) */ for (i = 0, reclen = 0; i < sizeof (reclen); i++) reclen += (uint64_t)(((uchar_t *)buf)[i]) << (8*i); if (bytes_read < sizeof (reclen) + reclen) break; /* unpack record */ if (nvlist_unpack(buf + sizeof (reclen), reclen, &nv, 0) != 0) return (ENOMEM); bytes_read -= sizeof (reclen) + reclen; buf += sizeof (reclen) + reclen; /* add record to nvlist array */ (*numrecords)++; if (ISP2(*numrecords + 1)) { *records = realloc(*records, *numrecords * 2 * sizeof (nvlist_t *)); } (*records)[*numrecords - 1] = nv; } *leftover = bytes_read; return (0); } /* * Retrieve the command history of a pool. */ int zpool_get_history(zpool_handle_t *zhp, nvlist_t **nvhisp) { char *buf; int buflen = 128 * 1024; uint64_t off = 0; nvlist_t **records = NULL; uint_t numrecords = 0; int err, i; buf = malloc(buflen); if (buf == NULL) return (ENOMEM); do { 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) break; if ((err = zpool_history_unpack(buf, bytes_read, &leftover, &records, &numrecords)) != 0) 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 = malloc(buflen); if (buf == NULL) return (ENOMEM); } /* CONSTCOND */ } while (1); free(buf); if (!err) { verify(nvlist_alloc(nvhisp, NV_UNIQUE_NAME, 0) == 0); verify(nvlist_add_nvlist_array(*nvhisp, ZPOOL_HIST_RECORD, records, numrecords) == 0); } for (i = 0; i < numrecords; i++) nvlist_free(records[i]); free(records); return (err); } void zpool_obj_to_path(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj, char *pathname, size_t len) { 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, ":<0x%llx>", obj); return; } /* get the dataset's name */ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_obj = dsobj; if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_DSOBJ_TO_DSNAME, &zc) != 0) { /* just write out a path of two object numbers */ (void) snprintf(pathname, len, "<0x%llx>:<0x%llx>", dsobj, obj); return; } (void) strlcpy(dsname, zc.zc_value, sizeof (dsname)); /* find out if the dataset is mounted */ mounted = 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 (ioctl(zhp->zpool_hdl->libzfs_fd, 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, obj); } free(mntpnt); } /* * Read the EFI label from the config, if a label does not exist then * pass back the error to the caller. If the caller has passed a non-NULL * diskaddr argument then we set it to the starting address of the EFI * partition. If the caller has passed a non-NULL boolean argument, then * we set it to indicate if the disk does have efi system partition. */ static int read_efi_label(nvlist_t *config, diskaddr_t *sb, boolean_t *system) { char *path; int fd; char diskname[MAXPATHLEN]; boolean_t boot = B_FALSE; int err = -1; int slice; if (nvlist_lookup_string(config, ZPOOL_CONFIG_PATH, &path) != 0) return (err); (void) snprintf(diskname, sizeof (diskname), "%s%s", ZFS_RDISK_ROOT, strrchr(path, '/')); if ((fd = open(diskname, O_RDONLY|O_NDELAY)) >= 0) { struct dk_gpt *vtoc; if ((err = efi_alloc_and_read(fd, &vtoc)) >= 0) { for (slice = 0; slice < vtoc->efi_nparts; slice++) { if (vtoc->efi_parts[slice].p_tag == V_SYSTEM) boot = B_TRUE; if (vtoc->efi_parts[slice].p_tag == V_USR) break; } if (sb != NULL && vtoc->efi_parts[slice].p_tag == V_USR) *sb = vtoc->efi_parts[slice].p_start; if (system != NULL) *system = boot; efi_free(vtoc); } (void) close(fd); } return (err); } /* * determine where a partition starts on a disk in the current * configuration */ static diskaddr_t find_start_block(nvlist_t *config) { nvlist_t **child; uint_t c, children; diskaddr_t sb = MAXOFFSET_T; uint64_t wholedisk; if (nvlist_lookup_nvlist_array(config, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) { if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_WHOLE_DISK, &wholedisk) != 0 || !wholedisk) { return (MAXOFFSET_T); } if (read_efi_label(config, &sb, NULL) < 0) sb = MAXOFFSET_T; return (sb); } for (c = 0; c < children; c++) { sb = find_start_block(child[c]); if (sb != MAXOFFSET_T) { return (sb); } } return (MAXOFFSET_T); } /* * Label an individual disk. The name provided is the short name, * stripped of any leading /dev path. */ int zpool_label_disk(libzfs_handle_t *hdl, zpool_handle_t *zhp, const char *name, zpool_boot_label_t boot_type, uint64_t boot_size, int *slice) { char path[MAXPATHLEN]; struct dk_gpt *vtoc; int fd; size_t resv = EFI_MIN_RESV_SIZE; uint64_t slice_size; diskaddr_t start_block; char errbuf[1024]; /* prepare an error message just in case */ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot label '%s'"), name); if (zhp) { nvlist_t *nvroot; verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); if (zhp->zpool_start_block == 0) start_block = find_start_block(nvroot); else start_block = zhp->zpool_start_block; zhp->zpool_start_block = start_block; } else { /* new pool */ start_block = NEW_START_BLOCK; } (void) snprintf(path, sizeof (path), "%s/%s%s", ZFS_RDISK_ROOT, name, BACKUP_SLICE); if ((fd = open(path, O_RDWR | O_NDELAY)) < 0) { /* * This shouldn't happen. We've long since verified that this * is a valid device. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "unable to open device")); return (zfs_error(hdl, EZFS_OPENFAILED, errbuf)); } if (efi_alloc_and_init(fd, EFI_NUMPAR, &vtoc) != 0) { /* * The only way this can fail is if we run out of memory, or we * were unable to read the disk's capacity */ if (errno == ENOMEM) (void) no_memory(hdl); (void) close(fd); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "unable to read disk capacity"), name); return (zfs_error(hdl, EZFS_NOCAP, errbuf)); } /* * Why we use V_USR: V_BACKUP confuses users, and is considered * disposable by some EFI utilities (since EFI doesn't have a backup * slice). V_UNASSIGNED is supposed to be used only for zero size * partitions, and efi_write() will fail if we use it. V_ROOT, V_BOOT, * etc. were all pretty specific. V_USR is as close to reality as we * can get, in the absence of V_OTHER. */ /* first fix the partition start block */ if (start_block == MAXOFFSET_T) start_block = NEW_START_BLOCK; /* * EFI System partition is using slice 0. * ZFS is on slice 1 and slice 8 is reserved. * We assume the GPT partition table without system * partition has zfs p_start == NEW_START_BLOCK. * If start_block != NEW_START_BLOCK, it means we have * system partition. Correct solution would be to query/cache vtoc * from existing vdev member. */ if (boot_type == ZPOOL_CREATE_BOOT_LABEL) { if (boot_size % vtoc->efi_lbasize != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "boot partition size must be a multiple of %d"), vtoc->efi_lbasize); (void) close(fd); efi_free(vtoc); return (zfs_error(hdl, EZFS_LABELFAILED, errbuf)); } /* * System partition size checks. * Note the 1MB is quite arbitrary value, since we * are creating dedicated pool, it should be enough * to hold fat + efi bootloader. May need to be * adjusted if the bootloader size will grow. */ if (boot_size < 1024 * 1024) { char buf[64]; zfs_nicenum(boot_size, buf, sizeof (buf)); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Specified size %s for EFI System partition is too " "small, the minimum size is 1MB."), buf); (void) close(fd); efi_free(vtoc); return (zfs_error(hdl, EZFS_LABELFAILED, errbuf)); } /* 33MB is tested with mkfs -F pcfs */ if (hdl->libzfs_printerr && ((vtoc->efi_lbasize == 512 && boot_size < 33 * 1024 * 1024) || (vtoc->efi_lbasize == 4096 && boot_size < 256 * 1024 * 1024))) { char buf[64]; zfs_nicenum(boot_size, buf, sizeof (buf)); (void) fprintf(stderr, dgettext(TEXT_DOMAIN, "Warning: EFI System partition size %s is " "not allowing to create FAT32 file\nsystem, which " "may result in unbootable system.\n"), buf); } /* Adjust zfs partition start by size of system partition. */ start_block += boot_size / vtoc->efi_lbasize; } if (start_block == NEW_START_BLOCK) { /* * Use default layout. * ZFS is on slice 0 and slice 8 is reserved. */ slice_size = vtoc->efi_last_u_lba + 1; slice_size -= EFI_MIN_RESV_SIZE; slice_size -= start_block; if (slice != NULL) *slice = 0; vtoc->efi_parts[0].p_start = start_block; vtoc->efi_parts[0].p_size = slice_size; vtoc->efi_parts[0].p_tag = V_USR; (void) strcpy(vtoc->efi_parts[0].p_name, "zfs"); vtoc->efi_parts[8].p_start = slice_size + start_block; vtoc->efi_parts[8].p_size = resv; vtoc->efi_parts[8].p_tag = V_RESERVED; } else { slice_size = start_block - NEW_START_BLOCK; vtoc->efi_parts[0].p_start = NEW_START_BLOCK; vtoc->efi_parts[0].p_size = slice_size; vtoc->efi_parts[0].p_tag = V_SYSTEM; (void) strcpy(vtoc->efi_parts[0].p_name, "loader"); if (slice != NULL) *slice = 1; /* prepare slice 1 */ slice_size = vtoc->efi_last_u_lba + 1 - slice_size; slice_size -= resv; slice_size -= NEW_START_BLOCK; vtoc->efi_parts[1].p_start = start_block; vtoc->efi_parts[1].p_size = slice_size; vtoc->efi_parts[1].p_tag = V_USR; (void) strcpy(vtoc->efi_parts[1].p_name, "zfs"); vtoc->efi_parts[8].p_start = slice_size + start_block; vtoc->efi_parts[8].p_size = resv; vtoc->efi_parts[8].p_tag = V_RESERVED; } if (efi_write(fd, vtoc) != 0) { /* * Some block drivers (like pcata) may not support EFI * GPT labels. Print out a helpful error message dir- * ecting the user to manually label the disk and give * a specific slice. */ (void) close(fd); efi_free(vtoc); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "try using fdisk(1M) and then provide a specific slice")); return (zfs_error(hdl, EZFS_LABELFAILED, errbuf)); } (void) close(fd); efi_free(vtoc); return (0); } static boolean_t supported_dump_vdev_type(libzfs_handle_t *hdl, nvlist_t *config, char *errbuf) { char *type; nvlist_t **child; uint_t children, c; verify(nvlist_lookup_string(config, ZPOOL_CONFIG_TYPE, &type) == 0); if (strcmp(type, VDEV_TYPE_FILE) == 0 || strcmp(type, VDEV_TYPE_HOLE) == 0 || strcmp(type, VDEV_TYPE_MISSING) == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "vdev type '%s' is not supported"), type); (void) zfs_error(hdl, EZFS_VDEVNOTSUP, errbuf); return (B_FALSE); } if (nvlist_lookup_nvlist_array(config, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { for (c = 0; c < children; c++) { if (!supported_dump_vdev_type(hdl, child[c], errbuf)) return (B_FALSE); } } return (B_TRUE); } /* * Check if this zvol is allowable for use as a dump device; zero if * it is, > 0 if it isn't, < 0 if it isn't a zvol. * * Allowable storage configurations include mirrors, all raidz variants, and * pools with log, cache, and spare devices. Pools which are backed by files or * have missing/hole vdevs are not suitable. */ int zvol_check_dump_config(char *arg) { zpool_handle_t *zhp = NULL; nvlist_t *config, *nvroot; char *p, *volname; nvlist_t **top; uint_t toplevels; libzfs_handle_t *hdl; char errbuf[1024]; char poolname[ZFS_MAX_DATASET_NAME_LEN]; int pathlen = strlen(ZVOL_FULL_DEV_DIR); int ret = 1; if (strncmp(arg, ZVOL_FULL_DEV_DIR, pathlen)) { return (-1); } (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "dump is not supported on device '%s'"), arg); if ((hdl = libzfs_init()) == NULL) return (1); libzfs_print_on_error(hdl, B_TRUE); volname = arg + pathlen; /* check the configuration of the pool */ if ((p = strchr(volname, '/')) == NULL) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "malformed dataset name")); (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf); return (1); } else if (p - volname >= ZFS_MAX_DATASET_NAME_LEN) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dataset name is too long")); (void) zfs_error(hdl, EZFS_NAMETOOLONG, errbuf); return (1); } else { (void) strncpy(poolname, volname, p - volname); poolname[p - volname] = '\0'; } if ((zhp = zpool_open(hdl, poolname)) == NULL) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "could not open pool '%s'"), poolname); (void) zfs_error(hdl, EZFS_OPENFAILED, errbuf); goto out; } config = zpool_get_config(zhp, NULL); if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "could not obtain vdev configuration for '%s'"), poolname); (void) zfs_error(hdl, EZFS_INVALCONFIG, errbuf); goto out; } verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &top, &toplevels) == 0); if (!supported_dump_vdev_type(hdl, top[0], errbuf)) { goto out; } ret = 0; out: if (zhp) zpool_close(zhp); libzfs_fini(hdl); return (ret); } Index: vendor/illumos/dist/lib/libzfs/common/libzfs_util.c =================================================================== --- vendor/illumos/dist/lib/libzfs/common/libzfs_util.c (revision 354960) +++ vendor/illumos/dist/lib/libzfs/common/libzfs_util.c (revision 354961) @@ -1,1625 +1,1628 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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, 2017 by Delphix. All rights reserved. * Copyright 2016 Igor Kozhukhov * Copyright (c) 2017 Datto Inc. */ /* * Internal utility routines for the ZFS library. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "libzfs_impl.h" #include "zfs_prop.h" #include "zfs_comutil.h" #include "zfeature_common.h" 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, "umount failed")); case EZFS_UNSHARENFSFAILED: return (dgettext(TEXT_DOMAIN, "unshare(1M) failed")); case EZFS_SHARENFSFAILED: return (dgettext(TEXT_DOMAIN, "share(1M) failed")); case EZFS_UNSHARESMBFAILED: return (dgettext(TEXT_DOMAIN, "smb remove share failed")); case EZFS_SHARESMBFAILED: return (dgettext(TEXT_DOMAIN, "smb add share 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_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_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_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_NO_RESILVER_DEFER: + return (dgettext(TEXT_DOMAIN, "this action requires the " + "resilver_defer feature")); case EZFS_UNKNOWN: return (dgettext(TEXT_DOMAIN, "unknown error")); default: assert(hdl->libzfs_error == 0); return (dgettext(TEXT_DOMAIN, "no error")); } } /*PRINTFLIKE2*/ 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\n"), 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)); } /*PRINTFLIKE3*/ 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); } return (0); } int zfs_standard_error(libzfs_handle_t *hdl, int error, const char *msg) { return (zfs_standard_error_fmt(hdl, error, "%s", msg)); } /*PRINTFLIKE3*/ 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); return (-1); 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; default: zfs_error_aux(hdl, strerror(error)); zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap); break; } va_end(ap); return (-1); } int zpool_standard_error(libzfs_handle_t *hdl, int error, const char *msg) { return (zpool_standard_error_fmt(hdl, error, "%s", msg)); } /*PRINTFLIKE3*/ 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; 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); return (-1); 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; /* There is no pending operation to cancel */ case ENOTACTIVE: zfs_verror(hdl, EZFS_NO_PENDING, 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; default: zfs_error_aux(hdl, 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. */ /*PRINTFLIKE2*/ 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); 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); } bzero((char *)ret + oldsize, (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); } /* * Convert a number to an appropriately human-readable output. */ void zfs_nicenum(uint64_t num, char *buf, size_t buflen) { nicenum(num, buf, buflen); } void libzfs_print_on_error(libzfs_handle_t *hdl, boolean_t printerr) { hdl->libzfs_printerr = printerr; } libzfs_handle_t * libzfs_init(void) { libzfs_handle_t *hdl; if ((hdl = calloc(1, sizeof (libzfs_handle_t))) == NULL) { return (NULL); } if ((hdl->libzfs_fd = open(ZFS_DEV, O_RDWR)) < 0) { free(hdl); return (NULL); } if ((hdl->libzfs_mnttab = fopen(MNTTAB, "rF")) == NULL) { (void) close(hdl->libzfs_fd); free(hdl); return (NULL); } hdl->libzfs_sharetab = fopen("/etc/dfs/sharetab", "rF"); if (libzfs_core_init() != 0) { (void) close(hdl->libzfs_fd); (void) fclose(hdl->libzfs_mnttab); (void) fclose(hdl->libzfs_sharetab); free(hdl); return (NULL); } zfs_prop_init(); zpool_prop_init(); zpool_feature_init(); libzfs_mnttab_init(hdl); if (getenv("ZFS_PROP_DEBUG") != NULL) { hdl->libzfs_prop_debug = B_TRUE; } return (hdl); } void libzfs_fini(libzfs_handle_t *hdl) { (void) close(hdl->libzfs_fd); if (hdl->libzfs_mnttab) (void) fclose(hdl->libzfs_mnttab); if (hdl->libzfs_sharetab) (void) fclose(hdl->libzfs_sharetab); zfs_uninit_libshare(hdl); zpool_free_handles(hdl); libzfs_fru_clear(hdl, B_TRUE); namespace_clear(hdl); libzfs_mnttab_fini(hdl); libzfs_core_fini(); 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, char *path, zfs_type_t argtype) { struct stat64 statbuf; struct extmnttab entry; int ret; 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 (stat64(path, &statbuf) != 0) { (void) fprintf(stderr, "%s: %s\n", path, strerror(errno)); return (NULL); } rewind(hdl->libzfs_mnttab); while ((ret = getextmntent(hdl->libzfs_mnttab, &entry, 0)) == 0) { if (makedevice(entry.mnt_major, entry.mnt_minor) == statbuf.st_dev) { break; } } if (ret != 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(). */ int zcmd_alloc_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, size_t len) { if (len == 0) len = 16 * 1024; zc->zc_nvlist_dst_size = len; zc->zc_nvlist_dst = (uint64_t)(uintptr_t)zfs_alloc(hdl, zc->zc_nvlist_dst_size); if (zc->zc_nvlist_dst == 0) return (-1); return (0); } /* * 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. */ int 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); if (zc->zc_nvlist_dst == 0) return (-1); return (0); } /* * 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 = NULL; zc->zc_nvlist_src = NULL; zc->zc_nvlist_dst = NULL; } static int zcmd_write_nvlist_com(libzfs_handle_t *hdl, uint64_t *outnv, uint64_t *outlen, nvlist_t *nvl) { char *packed; size_t len; verify(nvlist_size(nvl, &len, NV_ENCODE_NATIVE) == 0); if ((packed = zfs_alloc(hdl, len)) == NULL) return (-1); verify(nvlist_pack(nvl, &packed, &len, NV_ENCODE_NATIVE, 0) == 0); *outnv = (uint64_t)(uintptr_t)packed; *outlen = len; return (0); } int zcmd_write_conf_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl) { return (zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_conf, &zc->zc_nvlist_conf_size, nvl)); } int zcmd_write_src_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl) { return (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); } int zfs_ioctl(libzfs_handle_t *hdl, int request, zfs_cmd_t *zc) { return (ioctl(hdl->libzfs_fd, request, zc)); } /* * ================================================================ * 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 = cbp->cb_proplist; 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 : 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_INVAL) { const char *propname = (type == ZFS_TYPE_POOL) ? zpool_prop_to_name(pl->pl_prop) : zfs_prop_to_name(pl->pl_prop); len = strlen(propname); if (len > cbp->cb_colwidths[GET_COL_PROPERTY]) cbp->cb_colwidths[GET_COL_PROPERTY] = len; } else { 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 : 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 (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)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid numeric suffix '%s'"), buf); return (-1); } /* * We want to allow trailing 'b' characters for 'GB' or 'Mb'. But don't * allow 'BB' - that's just weird. */ if (buf[1] == '\0' || (toupper(buf[1]) == 'B' && buf[2] == '\0' && toupper(buf[0]) != 'B')) return (10*i); 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); if (fval > 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, char **svalp, uint64_t *ivalp, const char *errbuf) { data_type_t datatype = nvpair_type(elem); zprop_type_t proptype; const char *propname; char *value; boolean_t isnone = B_FALSE; boolean_t isauto = B_FALSE; if (type == ZFS_TYPE_POOL) { proptype = zpool_prop_get_type(prop); propname = zpool_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; } (void) nvpair_value_string(elem, svalp); 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 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, char *propname, zprop_list_t **listp, zfs_type_t type) { int prop; zprop_list_t *entry; prop = zprop_name_to_prop(propname, type); if (prop != ZPROP_INVAL && !zprop_valid_for_type(prop, type)) prop = ZPROP_INVAL; /* * When no property table entry can be found, return failure if * this is a pool property or if this isn't a user-defined * dataset property, */ if (prop == ZPROP_INVAL && ((type == ZFS_TYPE_POOL && !zpool_prop_feature(propname) && !zpool_prop_unsupported(propname)) || (type == ZFS_TYPE_DATASET && !zfs_prop_user(propname) && !zfs_prop_userquota(propname) && !zfs_prop_written(propname)))) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid property '%s'"), propname); return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN, "bad property list"))); } if ((entry = zfs_alloc(hdl, sizeof (zprop_list_t))) == NULL) return (-1); entry->pl_prop = prop; if (prop == ZPROP_INVAL) { if ((entry->pl_user_prop = zfs_strdup(hdl, propname)) == NULL) { free(entry); return (-1); } 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"))); } /* * It would be nice to use getsubopt() here, but the inclusion of column * aliases makes this more effort than it's worth. */ while (*props != '\0') { size_t len; char *p; char c; if ((p = strchr(props, ',')) == NULL) { len = strlen(props); p = props + len; } else { len = p - props; } /* * Check for empty options. */ if (len == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "empty property name")); return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN, "bad property list"))); } /* * Check all regular property names. */ c = props[len]; props[len] = '\0'; if (strcmp(props, "space") == 0) { static char *spaceprops[] = { "name", "avail", "used", "usedbysnapshots", "usedbydataset", "usedbyrefreservation", "usedbychildren", NULL }; int i; for (i = 0; spaceprops[i]; i++) { if (addlist(hdl, spaceprops[i], listp, type)) return (-1); listp = &(*listp)->pl_next; } } else { if (addlist(hdl, props, listp, type)) return (-1); listp = &(*listp)->pl_next; } props = p; if (c == ',') props++; } 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; int zprop_expand_list_cb(int prop, void *cb) { zprop_list_t *entry; expand_data_t *edp = cb; if ((entry = zfs_alloc(edp->hdl, sizeof (zprop_list_t))) == NULL) return (ZPROP_INVAL); 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. */ if ((entry = zfs_alloc(hdl, sizeof (zprop_list_t))) == NULL) return (-1); entry->pl_prop = (type == ZFS_TYPE_POOL) ? ZPOOL_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)); } /* * zfs_get_hole_count retrieves the number of holes (blocks which are * zero-filled) in the specified file using the _FIO_COUNT_FILLED ioctl. It * also optionally fetches the block size when bs is non-NULL. With hole count * and block size the full space consumed by the holes of a file can be * calculated. * * On success, zero is returned, the count argument is set to the * number of holes, and the bs argument is set to the block size (if it is * not NULL). On error, a non-zero errno is returned and the values in count * and bs are undefined. */ int zfs_get_hole_count(const char *path, uint64_t *count, uint64_t *bs) { int fd, err; struct stat64 ss; uint64_t fill; fd = open(path, O_RDONLY | O_LARGEFILE); if (fd == -1) return (errno); if (ioctl(fd, _FIO_COUNT_FILLED, &fill) == -1) { err = errno; (void) close(fd); return (err); } if (fstat64(fd, &ss) == -1) { err = errno; (void) close(fd); return (err); } *count = (ss.st_size + ss.st_blksize - 1) / ss.st_blksize - fill; VERIFY3S(*count, >=, 0); if (bs != NULL) { *bs = ss.st_blksize; } if (close(fd) == -1) { return (errno); } return (0); } ulong_t get_system_hostid(void) { char *env; /* * Allow the hostid to be subverted for testing. */ env = getenv("ZFS_HOSTID"); if (env) { ulong_t hostid = strtoull(env, NULL, 16); return (hostid & 0xFFFFFFFF); } return (gethostid()); } Index: vendor/illumos/dist/man/man1m/zpool.1m =================================================================== --- vendor/illumos/dist/man/man1m/zpool.1m (revision 354960) +++ vendor/illumos/dist/man/man1m/zpool.1m (revision 354961) @@ -1,2245 +1,2261 @@ .\" .\" 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 http://www.opensolaris.org/os/licensing. .\" 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, 2017 by Delphix. All rights reserved. .\" Copyright 2017 Nexenta Systems, Inc. .\" Copyright (c) 2017 Datto Inc. .\" Copyright (c) 2017 George Melikov. All Rights Reserved. .\" Copyright 2019 Joyent, Inc. .\" -.Dd April 27, 2018 +.Dd May 15, 2019 .Dt ZPOOL 1M .Os .Sh NAME .Nm zpool .Nd configure ZFS storage pools .Sh SYNOPSIS .Nm .Fl \? .Nm .Cm add .Op Fl fgLnP .Ar pool vdev Ns ... .Nm .Cm attach .Op Fl f .Ar pool device new_device .Nm .Cm checkpoint .Op Fl d, -discard .Ar pool .Nm .Cm clear .Ar pool .Op Ar device .Nm .Cm create .Op Fl dfn .Op Fl B .Op Fl m Ar mountpoint .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Oo Fl O Ar file-system-property Ns = Ns Ar value Oc Ns ... .Op Fl R Ar root .Op Fl t Ar tempname .Ar pool vdev Ns ... .Nm .Cm destroy .Op Fl f .Ar pool .Nm .Cm detach .Ar pool device .Nm .Cm export .Op Fl f .Ar pool Ns ... .Nm .Cm get .Op Fl Hp .Op Fl o Ar field Ns Oo , Ns Ar field Oc Ns ... .Sy all Ns | Ns Ar property Ns Oo , Ns Ar property Oc Ns ... .Ar pool Ns ... .Nm .Cm history .Op Fl il .Oo Ar pool Oc Ns ... .Nm .Cm import .Op Fl D .Op Fl d Ar dir .Nm .Cm import .Fl a .Op Fl DfmN .Op Fl F Op Fl n .Op Fl c Ar cachefile Ns | Ns Fl d Ar dir .Op Fl o Ar mntopts .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Op Fl R Ar root .Nm .Cm import .Op Fl Dfmt .Op Fl F Op Fl n .Op Fl -rewind-to-checkpoint .Op Fl c Ar cachefile Ns | Ns Fl d Ar dir .Op Fl o Ar mntopts .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Op Fl R Ar root .Ar pool Ns | Ns Ar id .Op Ar newpool .Nm .Cm initialize .Op Fl cs .Ar pool .Op Ar device Ns ... .Nm .Cm iostat .Op Fl v .Op Fl T Sy u Ns | Ns Sy d .Op Fl gLP .Oo Ar pool Oc Ns ... .Op Ar interval Op Ar count .Nm .Cm labelclear .Op Fl f .Ar device .Nm .Cm list .Op Fl HgLpPv .Op Fl o Ar property Ns Oo , Ns Ar property Oc Ns ... .Op Fl T Sy u Ns | Ns Sy d .Oo Ar pool Oc Ns ... .Op Ar interval Op Ar count .Nm .Cm offline .Op Fl t .Ar pool Ar device Ns ... .Nm .Cm online .Op Fl e .Ar pool Ar device Ns ... .Nm .Cm reguid .Ar pool .Nm .Cm reopen .Ar pool .Nm .Cm remove .Op Fl np .Ar pool Ar device Ns ... .Nm .Cm remove .Fl s .Ar pool .Nm .Cm replace .Op Fl f .Ar pool Ar device Op Ar new_device .Nm +.Cm resilver +.Ar pool Ns ... +.Nm .Cm scrub .Op Fl s | Fl p .Ar pool Ns ... .Nm .Cm set .Ar property Ns = Ns Ar value .Ar pool .Nm .Cm split .Op Fl gLnP .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Op Fl R Ar root .Ar pool newpool .Nm .Cm status .Op Fl DgLPvx .Op Fl T Sy u Ns | Ns Sy d .Oo Ar pool Oc Ns ... .Op Ar interval Op Ar count .Nm .Cm sync .Oo Ar pool Oc Ns ... .Nm .Cm upgrade .Nm .Cm upgrade .Fl v .Nm .Cm upgrade .Op Fl V Ar version .Fl a Ns | Ns Ar pool Ns ... .Sh DESCRIPTION The .Nm command configures ZFS storage pools. A storage pool is a collection of devices that provides physical storage and data replication for ZFS datasets. All datasets within a storage pool share the same space. See .Xr zfs 1M for information on managing datasets. .Ss Virtual Devices (vdevs) 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 Ds .It Sy disk A block device, typically located under .Pa /dev/dsk . 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/dsk .Pc . A whole disk can be specified by omitting the slice or partition designation. For example, .Pa c0t0d0 is equivalent to .Pa /dev/dsk/c0t0d0s2 . 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 of which it is a part. 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 N disks of size X can hold X bytes and can withstand (N-1) devices failing before data integrity is compromised. .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 .Pq in which data and parity become inconsistent after a power loss . Data and parity is striped across all disks within a raidz group. .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 N disks of size X with P parity disks can hold approximately (N-P)*X bytes and can withstand P device(s) failing before data integrity is compromised. 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 spare A special 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 allocating dedup data. 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 devices. .It Sy special A device dedicated solely for allocating various kinds of internal metadata, and optionally small file data. 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 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. The keywords .Sy mirror and .Sy raidz are used to distinguish where a group ends and another begins. For example, the following creates two root vdevs, each a mirror of two disks: .Bd -literal # zpool create mypool mirror c0t0d0 c0t1d0 mirror c1t0d0 c1t1d0 .Ed .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. .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: online, degraded, or 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 mirror or raidz device, 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 .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 .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 If a device is removed and later re-attached to the system, ZFS attempts to put the device online automatically. Device attach 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 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, .Bd -literal # zpool create pool mirror c0d0 c1d0 spare c2d0 c3d0 .Ed .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. .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 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 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 3C 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: .Bd -literal # zpool create pool c0d0 c1d0 log c2d0 .Ed .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 imported and exported as part of the larger pool. 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 allow 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: .Bd -literal # zpool create pool c0d0 c1d0 cache c2d0 c3d0 .Ed .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 considered volatile, as is the case with other system caches. .Ss Pool checkpoint Before starting critical procedures that include destructive actions (e.g .Nm zfs Cm destroy ), 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, while a pool has a checkpoint certain operations are not allowed. 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: .Bd -literal # zpool checkpoint pool .Ed .Pp To later rewind to its checkpointed state, you need to first export it and then rewind it during import: .Bd -literal # zpool export pool # zpool import --rewind-to-checkpoint pool .Ed .Pp To discard the checkpoint from a pool: .Bd -literal # zpool checkpoint -d pool .Ed .Pp Dataset reservations (controlled by the .Nm reservation or .Nm refreservation zfs 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 The allocations in the special class are dedicated to specific block types. By default this includes all metadata, the indirect blocks of user data, and any dedup data. The class can also be provisioned to accept a limited percentage of small file data blocks. .Pp A pool must always have at least one general (non-specified) vdev before other devices can be assigned to the special class. If the special class becomes full, then allocations intended for it will spill back into the normal class. .Pp Dedup data can be excluded from the special class by setting the .Sy zfs_ddt_data_is_special zfs kernel variable to false (0). .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 dataset property. It defaults to zero so you must opt-in by setting it to a non-zero value. See .Xr zfs 1M for more info on setting this property. .Ss Properties Each pool has several properties associated with it. Some properties are read-only statistics while others are configurable and change the behavior of the pool. .Pp The following are read-only properties: .Bl -tag -width Ds .It Cm allocated Amount of storage space used within the pool. .It Sy bootsize The size of the system boot partition. This property can only be set at pool creation time and is read-only once pool is created. Setting this property implies using the .Fl B option. .It Sy capacity Percentage of pool space used. This property can also be referred to by its shortened column name, .Sy cap . .It Sy expandsize Amount of uninitialized space within the pool or device that can be used to increase the total capacity of the pool. Uninitialized space consists of any space on an EFI labeled vdev which has not been brought online .Po e.g, using .Nm zpool Cm online Fl e .Pc . This space occurs when a LUN is dynamically expanded. .It Sy fragmentation The amount of fragmentation in the pool. .It Sy free The amount of free space available in the pool. .It Sy freeing After a file system or snapshot is destroyed, the space it was using is returned to the pool asynchronously. .Sy freeing is the amount of space remaining to be reclaimed. Over time .Sy freeing will decrease while .Sy free increases. .It Sy health The current health of the pool. Health can be one of .Sy ONLINE , DEGRADED , FAULTED , OFFLINE, REMOVED , UNAVAIL . .It Sy guid A unique identifier for the pool. .It Sy size Total size of the storage pool. .It Sy unsupported@ Ns Em feature_guid Information about unsupported features that are enabled on the pool. See .Xr zpool-features 5 for details. .El .Pp The space usage properties report actual physical space available to the storage pool. The physical space can be different from the total amount of space that any contained datasets can actually use. The amount of space used in a raidz configuration depends on the characteristics of the data being written. In addition, ZFS reserves some space for internal accounting that the .Xr zfs 1M command takes into account, but the .Nm command does not. For non-full pools of a reasonable size, these effects should be invisible. For small pools, or pools that are close to being completely full, these discrepancies may become more noticeable. .Pp The following property can be set at creation time and import time: .Bl -tag -width Ds .It Sy altroot Alternate root directory. If set, this directory is prepended to any mount points within the pool. This can be used when examining an unknown pool where the mount points cannot be trusted, or in an alternate boot environment, where the typical paths are not valid. .Sy altroot is not a persistent property. It is valid only while the system is up. Setting .Sy altroot defaults to using .Sy cachefile Ns = Ns Sy none , though this may be overridden using an explicit setting. .El .Pp The following property can be set only at import time: .Bl -tag -width Ds .It Sy readonly Ns = Ns Sy on Ns | Ns Sy off If set to .Sy on , the pool will be imported in read-only mode. This property can also be referred to by its shortened column name, .Sy rdonly . .El .Pp The following properties can be set at creation time and import time, and later changed with the .Nm zpool Cm set command: .Bl -tag -width Ds .It Sy autoexpand Ns = Ns Sy on Ns | Ns Sy off Controls automatic pool expansion when the underlying LUN is grown. If set to .Sy on , the pool will be resized according to the size of the expanded device. If the device is part of a mirror or raidz then all devices within that mirror/raidz group must be expanded before the new space is made available to the pool. The default behavior is .Sy off . This property can also be referred to by its shortened column name, .Sy expand . .It Sy autoreplace Ns = Ns Sy on Ns | Ns Sy off Controls automatic device replacement. If set to .Sy off , device replacement must be initiated by the administrator by using the .Nm zpool Cm replace command. If set to .Sy on , any new device, found in the same physical location as a device that previously belonged to the pool, is automatically formatted and replaced. The default behavior is .Sy off . This property can also be referred to by its shortened column name, .Sy replace . .It Sy bootfs Ns = Ns Ar pool Ns / Ns Ar dataset Identifies the default bootable dataset for the root pool. This property is expected to be set mainly by the installation and upgrade programs. .It Sy cachefile Ns = Ns Ar path Ns | Ns Sy none Controls the location of where the pool configuration is cached. Discovering all pools on system startup requires a cached copy of the configuration data that is stored on the root file system. All pools in this cache are automatically imported when the system boots. Some environments, such as install and clustering, need to cache this information in a different location so that pools are not automatically imported. Setting this property caches the pool configuration in a different location that can later be imported with .Nm zpool Cm import Fl c . Setting it to the special value .Sy none creates a temporary pool that is never cached, and the special value .Qq .Pq empty string uses the default location. .Pp Multiple pools can share the same cache file. Because the kernel destroys and recreates this file when pools are added and removed, care should be taken when attempting to access this file. When the last pool using a .Sy cachefile is exported or destroyed, the file is removed. .It Sy comment Ns = Ns Ar text A text string consisting of printable ASCII characters that will be stored such that it is available even if the pool becomes faulted. An administrator can provide additional information about a pool using this property. .It Sy dedupditto Ns = Ns Ar number Threshold for the number of block ditto copies. If the reference count for a deduplicated block increases above this number, a new ditto copy of this block is automatically stored. The default setting is .Sy 0 which causes no ditto copies to be created for deduplicated blocks. The minimum legal nonzero setting is .Sy 100 . .It Sy delegation Ns = Ns Sy on Ns | Ns Sy off Controls whether a non-privileged user is granted access based on the dataset permissions defined on the dataset. See .Xr zfs 1M for more information on ZFS delegated administration. .It Sy failmode Ns = Ns Sy wait Ns | Ns Sy continue Ns | Ns Sy panic Controls the system behavior in the event of catastrophic pool failure. This condition is typically a result of a loss of connectivity to the underlying storage device(s) or a failure of all devices within the pool. The behavior of such an event is determined as follows: .Bl -tag -width "continue" .It Sy wait Blocks all I/O access until the device connectivity is recovered and the errors are cleared. This is the default behavior. .It Sy continue Returns .Er EIO to any new write I/O requests but allows reads to any of the remaining healthy devices. Any write requests that have yet to be committed to disk would be blocked. .It Sy panic Prints out a message to the console and generates a system crash dump. .El .It Sy feature@ Ns Ar feature_name Ns = Ns Sy enabled The value of this property is the current state of .Ar feature_name . The only valid value when setting this property is .Sy enabled which moves .Ar feature_name to the enabled state. See .Xr zpool-features 5 for details on feature states. .It Sy listsnapshots Ns = Ns Sy on Ns | Ns Sy off Controls whether information about snapshots associated with this pool is output when .Nm zfs Cm list is run without the .Fl t option. The default value is .Sy off . This property can also be referred to by its shortened name, .Sy listsnaps . .It Sy multihost Ns = Ns Sy on Ns | Ns Sy off Controls whether a pool activity check should be performed during .Nm zpool Cm import . When a pool is determined to be active it cannot be imported, even with the .Fl f option. This property is intended to be used in failover configurations where multiple hosts have access to a pool on shared storage. .sp Multihost provides protection on import only. It does not protect against an individual device being used in multiple pools, regardless of the type of vdev. See the discussion under .Sy zpool create. .sp When this property is on, periodic writes to storage occur to show the pool is in use. See .Sy zfs_multihost_interval in the .Xr zfs-module-parameters 5 man page. In order to enable this property each host must set a unique hostid. The default value is .Sy off . .It Sy version Ns = Ns Ar version The current on-disk version of the pool. This can be increased, but never decreased. The preferred method of updating pools is with the .Nm zpool Cm upgrade command, though this property can be used when a specific version is needed for backwards compatibility. Once feature flags are enabled on a pool this property will no longer have a value. .El .Ss Subcommands All subcommands that modify state are logged persistently to the pool in their original form. .Pp The .Nm command provides subcommands to create and destroy storage pools, add capacity to storage pools, and provide information about the storage pools. The following subcommands are supported: .Bl -tag -width Ds .It Xo .Nm .Fl \? .Xc Displays a help message. .It Xo .Nm .Cm add .Op Fl fgLnP .Ar pool vdev Ns ... .Xc Adds the specified virtual devices to the given pool. The .Ar vdev specification is described in the .Sx Virtual Devices section. The behavior of the .Fl f option, and the device checks performed are described in the .Nm zpool Cm create subcommand. .Bl -tag -width Ds .It Fl f Forces use of .Ar vdev Ns s , even if they appear in use or specify a conflicting replication level. Not all devices can be overridden in this manner. .It Fl g Display .Ar vdev , GUIDs instead of the normal device names. These GUIDs can be used in place of device names for the zpool detach/offline/remove/replace commands. .It Fl L Display real paths for .Ar vdev Ns s resolving all symbolic links. This can be used to look up the current block device name regardless of the /dev/disk/ path used to open it. .It Fl n Displays the configuration that would be used without actually adding the .Ar vdev Ns s . The actual pool creation can still fail due to insufficient privileges or device sharing. .It Fl P Display real paths for .Ar vdev Ns s instead of only the last component of the path. This can be used in conjunction with the .Fl L flag. .El .It Xo .Nm .Cm attach .Op Fl f .Ar pool device new_device .Xc Attaches .Ar new_device to the existing .Ar device . The existing device cannot be part of a raidz configuration. If .Ar device is not currently part of a mirrored configuration, .Ar device automatically transforms into a two-way mirror of .Ar device and .Ar new_device . If .Ar device is part of a two-way mirror, attaching .Ar new_device creates a three-way mirror, and so on. In either case, .Ar new_device begins to resilver immediately. .Bl -tag -width Ds .It Fl f Forces use of .Ar new_device , even if its appears to be in use. Not all devices can be overridden in this manner. .El .It Xo .Nm .Cm checkpoint .Op Fl d, -discard .Ar pool .Xc Checkpoints the current state of .Ar pool , which can be later restored by .Nm zpool Cm import --rewind-to-checkpoint . The existence of a checkpoint in a pool prohibits the following .Nm zpool commands: .Cm remove , .Cm attach , .Cm detach , .Cm split , and .Cm reguid . In addition, it may break reservation boundaries if the pool lacks free space. The .Nm zpool Cm status command indicates the existence of a checkpoint or the progress of discarding a checkpoint from a pool. The .Nm zpool Cm list command reports how much space the checkpoint takes from the pool. .Bl -tag -width Ds .It Fl d, -discard Discards an existing checkpoint from .Ar pool . .El .It Xo .Nm .Cm clear .Ar pool .Op Ar device .Xc Clears device errors in a pool. If no arguments are specified, all device errors within the pool are cleared. If one or more devices is specified, only those errors associated with the specified device or devices are cleared. If multihost is enabled, and the pool has been suspended, this will not resume I/O. While the pool was suspended, it may have been imported on another host, and resuming I/O could result in pool damage. .It Xo .Nm .Cm create .Op Fl dfn .Op Fl B .Op Fl m Ar mountpoint .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Oo Fl O Ar file-system-property Ns = Ns Ar value Oc Ns ... .Op Fl R Ar root .Op Fl t Ar tempname .Ar pool vdev Ns ... .Xc Creates a new storage pool containing the virtual devices specified on the command line. The pool name must begin with a letter, and can only contain alphanumeric characters as well as underscore .Pq Qq Sy _ , dash .Pq Qq Sy - , and period .Pq Qq Sy \&. . The pool names .Sy mirror , .Sy raidz , .Sy spare and .Sy log are reserved, as are names beginning with the pattern .Sy c[0-9] . The .Ar vdev specification is described in the .Sx Virtual Devices section. .Pp The command attempts to verify that each device specified is accessible and not currently in use by another subsystem. However this check is not robust enough to detect simultaneous attempts to use a new device in different pools, even if .Sy multihost is .Sy enabled. The administrator must ensure that simultaneous invocations of any combination of .Sy zpool replace , .Sy zpool create , .Sy zpool add , or .Sy zpool labelclear , do not refer to the same device. Using the same device in two pools will result in pool corruption. .sp There are some uses, such as being currently mounted, or specified as the dedicated dump device, that prevents a device from ever being used by ZFS. Other uses, such as having a preexisting UFS file system, can be overridden with the .Fl f option. .Pp The command also checks that the replication strategy for the pool is consistent. An attempt to combine redundant and non-redundant storage in a single pool, or to mix disks and files, results in an error unless .Fl f is specified. The use of differently sized devices within a single raidz or mirror group is also flagged as an error unless .Fl f is specified. .Pp Unless the .Fl R option is specified, the default mount point is .Pa / Ns Ar pool . The mount point must not exist or must be empty, or else the root dataset cannot be mounted. This can be overridden with the .Fl m option. .Pp By default all supported features are enabled on the new pool unless the .Fl d option is specified. .Bl -tag -width Ds .It Fl B Create whole disk pool with EFI System partition to support booting system with UEFI firmware. Default size is 256MB. To create boot partition with custom size, set the .Sy bootsize property with the .Fl o option. See the .Sx Properties section for details. .It Fl d Do not enable any features on the new pool. Individual features can be enabled by setting their corresponding properties to .Sy enabled with the .Fl o option. See .Xr zpool-features 5 for details about feature properties. .It Fl f Forces use of .Ar vdev Ns s , even if they appear in use or specify a conflicting replication level. Not all devices can be overridden in this manner. .It Fl m Ar mountpoint Sets the mount point for the root dataset. The default mount point is .Pa /pool or .Pa altroot/pool if .Ar altroot is specified. The mount point must be an absolute path, .Sy legacy , or .Sy none . For more information on dataset mount points, see .Xr zfs 1M . .It Fl n Displays the configuration that would be used without actually creating the pool. The actual pool creation can still fail due to insufficient privileges or device sharing. .It Fl o Ar property Ns = Ns Ar value Sets the given pool properties. See the .Sx Properties section for a list of valid properties that can be set. .It Fl O Ar file-system-property Ns = Ns Ar value Sets the given file system properties in the root file system of the pool. See the .Sx Properties section of .Xr zfs 1M for a list of valid properties that can be set. .It Fl R Ar root Equivalent to .Fl o Sy cachefile Ns = Ns Sy none Fl o Sy altroot Ns = Ns Ar root .It Fl t Ar tempname Sets the in-core pool name to .Pa tempname while the on-disk name will be the name specified as the pool name .Pa pool . This will set the default cachefile property to .Sy none. This is intended to handle name space collisions when creating pools for other systems, such as virtual machines or physical machines whose pools live on network block devices. .El .It Xo .Nm .Cm destroy .Op Fl f .Ar pool .Xc Destroys the given pool, freeing up any devices for other use. This command tries to unmount any active datasets before destroying the pool. .Bl -tag -width Ds .It Fl f Forces any active datasets contained within the pool to be unmounted. .El .It Xo .Nm .Cm detach .Ar pool device .Xc Detaches .Ar device from a mirror. The operation is refused if there are no other valid replicas of the data. .It Xo .Nm .Cm export .Op Fl f .Ar pool Ns ... .Xc Exports the given pools from the system. All devices are marked as exported, but are still considered in use by other subsystems. The devices can be moved between systems .Pq even those of different endianness and imported as long as a sufficient number of devices are present. .Pp Before exporting the pool, all datasets within the pool are unmounted. A pool can not be exported if it has a shared spare that is currently being used. .Pp For pools to be portable, you must give the .Nm command whole disks, not just slices, so that ZFS can label the disks with portable EFI labels. Otherwise, disk drivers on platforms of different endianness will not recognize the disks. .Bl -tag -width Ds .It Fl f Forcefully unmount all datasets, using the .Nm unmount Fl f command. .Pp This command will forcefully export the pool even if it has a shared spare that is currently being used. This may lead to potential data corruption. .El .It Xo .Nm .Cm get .Op Fl Hp .Op Fl o Ar field Ns Oo , Ns Ar field Oc Ns ... .Sy all Ns | Ns Ar property Ns Oo , Ns Ar property Oc Ns ... .Ar pool Ns ... .Xc Retrieves the given list of properties .Po or all properties if .Sy all is used .Pc for the specified storage pool(s). These properties are displayed with the following fields: .Bd -literal name Name of storage pool property Property name value Property value source Property source, either 'default' or 'local'. .Ed .Pp See the .Sx Properties section for more information on the available pool properties. .Bl -tag -width Ds .It Fl H Scripted mode. Do not display headers, and separate fields by a single tab instead of arbitrary space. .It Fl o Ar field A comma-separated list of columns to display. .Sy name Ns \&, Ns Sy property Ns \&, Ns Sy value Ns \&, Ns Sy source is the default value. .It Fl p Display numbers in parsable (exact) values. .El .It Xo .Nm .Cm history .Op Fl il .Oo Ar pool Oc Ns ... .Xc Displays the command history of the specified pool(s) or all pools if no pool is specified. .Bl -tag -width Ds .It Fl i Displays internally logged ZFS events in addition to user initiated events. .It Fl l Displays log records in long format, which in addition to standard format includes, the user name, the hostname, and the zone in which the operation was performed. .El .It Xo .Nm .Cm import .Op Fl D .Op Fl d Ar dir .Xc Lists pools available to import. If the .Fl d option is not specified, this command searches for devices in .Pa /dev/dsk . The .Fl d option can be specified multiple times, and all directories are searched. If the device appears to be part of an exported pool, this command displays a summary of the pool with the name of the pool, a numeric identifier, as well as the vdev layout and current health of the device for each device or file. Destroyed pools, pools that were previously destroyed with the .Nm zpool Cm destroy command, are not listed unless the .Fl D option is specified. .Pp The numeric identifier is unique, and can be used instead of the pool name when multiple exported pools of the same name are available. .Bl -tag -width Ds .It Fl c Ar cachefile Reads configuration from the given .Ar cachefile that was created with the .Sy cachefile pool property. This .Ar cachefile is used instead of searching for devices. .It Fl d Ar dir Searches for devices or files in .Ar dir . The .Fl d option can be specified multiple times. .It Fl D Lists destroyed pools only. .El .It Xo .Nm .Cm import .Fl a .Op Fl DfmN .Op Fl F Op Fl n .Op Fl c Ar cachefile Ns | Ns Fl d Ar dir .Op Fl o Ar mntopts .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Op Fl R Ar root .Xc Imports all pools found in the search directories. Identical to the previous command, except that all pools with a sufficient number of devices available are imported. Destroyed pools, pools that were previously destroyed with the .Nm zpool Cm destroy command, will not be imported unless the .Fl D option is specified. .Bl -tag -width Ds .It Fl a Searches for and imports all pools found. .It Fl c Ar cachefile Reads configuration from the given .Ar cachefile that was created with the .Sy cachefile pool property. This .Ar cachefile is used instead of searching for devices. .It Fl d Ar dir Searches for devices or files in .Ar dir . The .Fl d option can be specified multiple times. This option is incompatible with the .Fl c option. .It Fl D Imports destroyed pools only. The .Fl f option is also required. .It Fl f Forces import, even if the pool appears to be potentially active. .It Fl F Recovery mode for a non-importable pool. Attempt to return the pool to an importable state by discarding the last few transactions. Not all damaged pools can be recovered by using this option. If successful, the data from the discarded transactions is irretrievably lost. This option is ignored if the pool is importable or already imported. .It Fl m Allows a pool to import when there is a missing log device. Recent transactions can be lost because the log device will be discarded. .It Fl n Used with the .Fl F recovery option. Determines whether a non-importable pool can be made importable again, but does not actually perform the pool recovery. For more details about pool recovery mode, see the .Fl F option, above. .It Fl N Import the pool without mounting any file systems. .It Fl o Ar mntopts Comma-separated list of mount options to use when mounting datasets within the pool. See .Xr zfs 1M for a description of dataset properties and mount options. .It Fl o Ar property Ns = Ns Ar value Sets the specified property on the imported pool. See the .Sx Properties section for more information on the available pool properties. .It Fl R Ar root Sets the .Sy cachefile property to .Sy none and the .Sy altroot property to .Ar root . .El .It Xo .Nm .Cm import .Op Fl Dfmt .Op Fl F Op Fl n .Op Fl -rewind-to-checkpoint .Op Fl c Ar cachefile Ns | Ns Fl d Ar dir .Op Fl o Ar mntopts .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Op Fl R Ar root .Ar pool Ns | Ns Ar id .Op Ar newpool .Xc Imports a specific pool. A pool can be identified by its name or the numeric identifier. If .Ar newpool is specified, the pool is imported using the name .Ar newpool . Otherwise, it is imported with the same name as its exported name. .Pp If a device is removed from a system without running .Nm zpool Cm export first, the device appears as potentially active. It cannot be determined if this was a failed export, or whether the device is really in use from another host. To import a pool in this state, the .Fl f option is required. .Bl -tag -width Ds .It Fl c Ar cachefile Reads configuration from the given .Ar cachefile that was created with the .Sy cachefile pool property. This .Ar cachefile is used instead of searching for devices. .It Fl d Ar dir Searches for devices or files in .Ar dir . The .Fl d option can be specified multiple times. This option is incompatible with the .Fl c option. .It Fl D Imports destroyed pool. The .Fl f option is also required. .It Fl f Forces import, even if the pool appears to be potentially active. .It Fl F Recovery mode for a non-importable pool. Attempt to return the pool to an importable state by discarding the last few transactions. Not all damaged pools can be recovered by using this option. If successful, the data from the discarded transactions is irretrievably lost. This option is ignored if the pool is importable or already imported. .It Fl m Allows a pool to import when there is a missing log device. Recent transactions can be lost because the log device will be discarded. .It Fl n Used with the .Fl F recovery option. Determines whether a non-importable pool can be made importable again, but does not actually perform the pool recovery. For more details about pool recovery mode, see the .Fl F option, above. .It Fl o Ar mntopts Comma-separated list of mount options to use when mounting datasets within the pool. See .Xr zfs 1M for a description of dataset properties and mount options. .It Fl o Ar property Ns = Ns Ar value Sets the specified property on the imported pool. See the .Sx Properties section for more information on the available pool properties. .It Fl R Ar root Sets the .Sy cachefile property to .Sy none and the .Sy altroot property to .Ar root . .It Fl t Used with .Ar newpool . Specifies that .Ar newpool is temporary. Temporary pool names last until export. Ensures that the original pool name will be used in all label updates and therefore is retained upon export. Will also set .Sy cachefile property to .Sy none when not explicitly specified. .It Fl -rewind-to-checkpoint Rewinds pool to the checkpointed state. Once the pool is imported with this flag there is no way to undo the rewind. All changes and data that were written after the checkpoint are lost! The only exception is when the .Sy readonly mounting option is enabled. In this case, the checkpointed state of the pool is opened and an administrator can see how the pool would look like if they were to fully rewind. .El .It Xo .Nm .Cm initialize .Op Fl cs .Ar pool .Op Ar device Ns ... .Xc Begins initializing by writing to all unallocated regions on the specified devices, or all eligible devices in the pool if no individual devices are specified. Only leaf data or log devices may be initialized. .Bl -tag -width Ds .It Fl c, -cancel Cancel initializing on the specified devices, or all eligible devices if none are specified. If one or more target devices are invalid or are not currently being initialized, the command will fail and no cancellation will occur on any device. .It Fl s -suspend Suspend initializing on the specified devices, or all eligible devices if none are specified. If one or more target devices are invalid or are not currently being initialized, the command will fail and no suspension will occur on any device. Initializing can then be resumed by running .Nm zpool Cm initialize with no flags on the relevant target devices. .El .It Xo .Nm .Cm iostat .Op Fl T Sy u Ns | Ns Sy d .Op Fl gLPv .Oo Ar pool Oc Ns ... .Op Ar interval Op Ar count .Xc Displays I/O statistics for the given pools. When given an .Ar interval , the statistics are printed every .Ar interval seconds until ^C is pressed. If no .Ar pool Ns s are specified, statistics for every pool in the system is shown. If .Ar count is specified, the command exits after .Ar count reports are printed. .Bl -tag -width Ds .It Fl T Sy u Ns | Ns Sy d Display a time stamp. Specify .Sy u for a printed representation of the internal representation of time. See .Xr time 2 . Specify .Sy d for standard date format. See .Xr date 1 . .It Fl g Display vdev GUIDs instead of the normal device names. These GUIDs can be used in place of device names for the zpool detach/offline/remove/replace commands. .It Fl L Display real paths for vdevs resolving all symbolic links. This can be used to look up the current block device name regardless of the .Pa /dev/disk/ path used to open it. .It Fl P Display full paths for vdevs instead of only the last component of the path. This can be used in conjunction with the .Fl L flag. .It Fl v Verbose statistics Reports usage statistics for individual vdevs within the pool, in addition to the pool-wide statistics. .El .It Xo .Nm .Cm labelclear .Op Fl f .Ar device .Xc Removes ZFS label information from the specified .Ar device . The .Ar device must not be part of an active pool configuration. .Bl -tag -width Ds .It Fl f Treat exported or foreign devices as inactive. .El .It Xo .Nm .Cm list .Op Fl HgLpPv .Op Fl o Ar property Ns Oo , Ns Ar property Oc Ns ... .Op Fl T Sy u Ns | Ns Sy d .Oo Ar pool Oc Ns ... .Op Ar interval Op Ar count .Xc Lists the given pools along with a health status and space usage. If no .Ar pool Ns s are specified, all pools in the system are listed. When given an .Ar interval , the information is printed every .Ar interval seconds until ^C is pressed. If .Ar count is specified, the command exits after .Ar count reports are printed. .Bl -tag -width Ds .It Fl g Display vdev GUIDs instead of the normal device names. These GUIDs can be used in place of device names for the zpool detach/offline/remove/replace commands. .It Fl H Scripted mode. Do not display headers, and separate fields by a single tab instead of arbitrary space. .It Fl o Ar property Comma-separated list of properties to display. See the .Sx Properties section for a list of valid properties. The default list is .Cm name , size , allocated , free , checkpoint, expandsize , fragmentation , capacity , .Cm dedupratio , health , altroot . .It Fl L Display real paths for vdevs resolving all symbolic links. This can be used to look up the current block device name regardless of the /dev/disk/ path used to open it. .It Fl p Display numbers in parsable .Pq exact values. .It Fl P Display full paths for vdevs instead of only the last component of the path. This can be used in conjunction with the .Fl L flag. .It Fl T Sy u Ns | Ns Sy d Display a time stamp. Specify .Fl u for a printed representation of the internal representation of time. See .Xr time 2 . Specify .Fl d for standard date format. See .Xr date 1 . .It Fl v Verbose statistics. Reports usage statistics for individual vdevs within the pool, in addition to the pool-wise statistics. .El .It Xo .Nm .Cm offline .Op Fl t .Ar pool Ar device Ns ... .Xc Takes the specified physical device offline. While the .Ar device is offline, no attempt is made to read or write to the device. This command is not applicable to spares. .Bl -tag -width Ds .It Fl t Temporary. Upon reboot, the specified physical device reverts to its previous state. .El .It Xo .Nm .Cm online .Op Fl e .Ar pool Ar device Ns ... .Xc Brings the specified physical device online. This command is not applicable to spares. .Bl -tag -width Ds .It Fl e Expand the device to use all available space. If the device is part of a mirror or raidz then all devices must be expanded before the new space will become available to the pool. .El .It Xo .Nm .Cm reguid .Ar pool .Xc Generates a new unique identifier for the pool. You must ensure that all devices in this pool are online and healthy before performing this action. .It Xo .Nm .Cm reopen .Ar pool .Xc Reopen all the vdevs associated with the pool. .It Xo .Nm .Cm remove .Op Fl np .Ar pool Ar device Ns ... .Xc Removes the specified device from the pool. This command currently only supports removing hot spares, cache, log devices and mirrored top-level vdevs (mirror of leaf devices); but not raidz. .sp Removing a top-level vdev reduces the total amount of space in the storage pool. The specified device will be evacuated by copying all allocated space from it to the other devices in the pool. In this case, the .Nm zpool Cm remove command initiates the removal and returns, while the evacuation continues in the background. The removal progress can be monitored with .Nm zpool Cm status. This feature must be enabled to be used, see .Xr zpool-features 5 .Pp A mirrored top-level device (log or data) can be removed by specifying the top-level mirror for the same. Non-log devices or data devices that are part of a mirrored configuration can be removed using the .Nm zpool Cm detach command. .Bl -tag -width Ds .It Fl n Do not actually perform the removal ("no-op"). Instead, print the estimated amount of memory that will be used by the mapping table after the removal completes. This is nonzero only for top-level vdevs. .El .Bl -tag -width Ds .It Fl p Used in conjunction with the .Fl n flag, displays numbers as parsable (exact) values. .El .It Xo .Nm .Cm remove .Fl s .Ar pool .Xc Stops and cancels an in-progress removal of a top-level vdev. .It Xo .Nm .Cm replace .Op Fl f .Ar pool Ar device Op Ar new_device .Xc Replaces .Ar old_device with .Ar new_device . This is equivalent to attaching .Ar new_device , waiting for it to resilver, and then detaching .Ar old_device . .Pp The size of .Ar new_device must be greater than or equal to the minimum size of all the devices in a mirror or raidz configuration. .Pp .Ar new_device is required if the pool is not redundant. If .Ar new_device is not specified, it defaults to .Ar old_device . This form of replacement is useful after an existing disk has failed and has been physically replaced. In this case, the new disk may have the same .Pa /dev/dsk path as the old device, even though it is actually a different disk. ZFS recognizes this. .Bl -tag -width Ds .It Fl f Forces use of .Ar new_device , even if its appears to be in use. Not all devices can be overridden in this manner. .El +.It Xo +.Nm +.Cm resilver +.Ar pool Ns ... +.Xc +Starts a resilver. +If an existing resilver is already running it will be restarted from the +beginning. +Any drives that were scheduled for a deferred resilver will be added to the +new one. +This requires the +.Sy resilver_defer +feature. .It Xo .Nm .Cm scrub .Op Fl s | Fl p .Ar pool Ns ... .Xc Begins a scrub or resumes a paused scrub. The scrub examines all data in the specified pools to verify that it checksums correctly. For replicated .Pq mirror or raidz devices, ZFS automatically repairs any damage discovered during the scrub. The .Nm zpool Cm status command reports the progress of the scrub and summarizes the results of the scrub upon completion. .Pp Scrubbing and resilvering are very similar operations. The difference is that resilvering only examines data that ZFS knows to be out of date .Po for example, when attaching a new device to a mirror or replacing an existing device .Pc , whereas scrubbing examines all data to discover silent errors due to hardware faults or disk failure. .Pp Because scrubbing and resilvering are I/O-intensive operations, ZFS only allows one at a time. If a scrub is paused, the .Nm zpool Cm scrub resumes it. If a resilver is in progress, ZFS does not allow a scrub to be started until the resilver completes. .Pp Note that, due to changes in pool data on a live system, it is possible for scrubs to progress slightly beyond 100% completion. During this period, no completion time estimate will be provided. .Bl -tag -width Ds .It Fl s Stop scrubbing. .El .Bl -tag -width Ds .It Fl p Pause scrubbing. Scrub pause state and progress are periodically synced to disk. If the system is restarted or pool is exported during a paused scrub, even after import, scrub will remain paused until it is resumed. Once resumed the scrub will pick up from the place where it was last checkpointed to disk. To resume a paused scrub issue .Nm zpool Cm scrub again. .El .It Xo .Nm .Cm set .Ar property Ns = Ns Ar value .Ar pool .Xc Sets the given property on the specified pool. See the .Sx Properties section for more information on what properties can be set and acceptable values. .It Xo .Nm .Cm split .Op Fl gLnP .Oo Fl o Ar property Ns = Ns Ar value Oc Ns ... .Op Fl R Ar root .Ar pool newpool .Xc Splits devices off .Ar pool creating .Ar newpool . All vdevs in .Ar pool must be mirrors. At the time of the split, .Ar newpool will be a replica of .Ar pool . .Bl -tag -width Ds .It Fl g Display vdev GUIDs instead of the normal device names. These GUIDs can be used in place of device names for the zpool detach/offline/remove/replace commands. .It Fl L Display real paths for vdevs resolving all symbolic links. This can be used to look up the current block device name regardless of the .Pa /dev/disk/ path used to open it. .It Fl n Do dry run, do not actually perform the split. Print out the expected configuration of .Ar newpool . .It Fl P Display full paths for vdevs instead of only the last component of the path. This can be used in conjunction with the .Fl L flag. .It Fl o Ar property Ns = Ns Ar value Sets the specified property for .Ar newpool . See the .Sx Properties section for more information on the available pool properties. .It Fl R Ar root Set .Sy altroot for .Ar newpool to .Ar root and automatically import it. .El .It Xo .Nm .Cm status .Op Fl DgLPvx .Op Fl T Sy u Ns | Ns Sy d .Oo Ar pool Oc Ns ... .Op Ar interval Op Ar count .Xc Displays the detailed health status for the given pools. If no .Ar pool is specified, then the status of each pool in the system is displayed. For more information on pool and device health, see the .Sx Device Failure and Recovery section. .Pp If a scrub or resilver is in progress, this command reports the percentage done and the estimated time to completion. Both of these are only approximate, because the amount of data in the pool and the other workloads on the system can change. .Bl -tag -width Ds .It Fl D Display a histogram of deduplication statistics, showing the allocated .Pq physically present on disk and referenced .Pq logically referenced in the pool block counts and sizes by reference count. .It Fl g Display vdev GUIDs instead of the normal device names. These GUIDs can be used in place of device names for the zpool detach/offline/remove/replace commands. .It Fl L Display real paths for vdevs resolving all symbolic links. This can be used to look up the current block device name regardless of the .Pa /dev/disk/ path used to open it. .It Fl P Display full paths for vdevs instead of only the last component of the path. This can be used in conjunction with the .Fl L flag. .It Fl T Sy u Ns | Ns Sy d Display a time stamp. Specify .Fl u for a printed representation of the internal representation of time. See .Xr time 2 . Specify .Fl d for standard date format. See .Xr date 1 . .It Fl v Displays verbose data error information, printing out a complete list of all data errors since the last complete pool scrub. .It Fl x Only display status for pools that are exhibiting errors or are otherwise unavailable. Warnings about pools not using the latest on-disk format will not be included. .El .It Xo .Nm .Cm sync .Oo Ar pool Oc Ns ... .Xc Forces all in-core dirty data to be written to the primary pool storage and not the ZIL. It will also update administrative information including quota reporting. Without arguments, .Nm zpool Cm sync will sync all pools on the system. Otherwise, it will only sync the specified .Ar pool . .It Xo .Nm .Cm upgrade .Xc Displays pools which do not have all supported features enabled and pools formatted using a legacy ZFS version number. These pools can continue to be used, but some features may not be available. Use .Nm zpool Cm upgrade Fl a to enable all features on all pools. .It Xo .Nm .Cm upgrade .Fl v .Xc Displays legacy ZFS versions supported by the current software. See .Xr zpool-features 5 for a description of feature flags features supported by the current software. .It Xo .Nm .Cm upgrade .Op Fl V Ar version .Fl a Ns | Ns Ar pool Ns ... .Xc Enables all supported features on the given pool. Once this is done, the pool will no longer be accessible on systems that do not support feature flags. See .Xr zpool-features 5 for details on compatibility with systems that support feature flags, but do not support all features enabled on the pool. .Bl -tag -width Ds .It Fl a Enables all supported features on all pools. .It Fl V Ar version Upgrade to the specified legacy version. If the .Fl V flag is specified, no features will be enabled on the pool. This option can only be used to increase the version number up to the last supported legacy version number. .El .El .Sh EXIT STATUS The following exit values are returned: .Bl -tag -width Ds .It Sy 0 Successful completion. .It Sy 1 An error occurred. .It Sy 2 Invalid command line options were specified. .El .Sh EXAMPLES .Bl -tag -width Ds .It Sy Example 1 No Creating a RAID-Z Storage Pool The following command creates a pool with a single raidz root vdev that consists of six disks. .Bd -literal # zpool create tank raidz c0t0d0 c0t1d0 c0t2d0 c0t3d0 c0t4d0 c0t5d0 .Ed .It Sy Example 2 No Creating a Mirrored Storage Pool The following command creates a pool with two mirrors, where each mirror contains two disks. .Bd -literal # zpool create tank mirror c0t0d0 c0t1d0 mirror c0t2d0 c0t3d0 .Ed .It Sy Example 3 No Creating a ZFS Storage Pool by Using Slices The following command creates an unmirrored pool using two disk slices. .Bd -literal # zpool create tank /dev/dsk/c0t0d0s1 c0t1d0s4 .Ed .It Sy Example 4 No Creating a ZFS Storage Pool by Using Files The following command creates an unmirrored pool using files. While not recommended, a pool based on files can be useful for experimental purposes. .Bd -literal # zpool create tank /path/to/file/a /path/to/file/b .Ed .It Sy Example 5 No Adding a Mirror to a ZFS Storage Pool The following command adds two mirrored disks to the pool .Em tank , assuming the pool is already made up of two-way mirrors. The additional space is immediately available to any datasets within the pool. .Bd -literal # zpool add tank mirror c1t0d0 c1t1d0 .Ed .It Sy Example 6 No Listing Available ZFS Storage Pools The following command lists all available pools on the system. In this case, the pool .Em zion is faulted due to a missing device. The results from this command are similar to the following: .Bd -literal # zpool list NAME SIZE ALLOC FREE FRAG EXPANDSZ CAP DEDUP HEALTH ALTROOT rpool 19.9G 8.43G 11.4G 33% - 42% 1.00x ONLINE - tank 61.5G 20.0G 41.5G 48% - 32% 1.00x ONLINE - zion - - - - - - - FAULTED - .Ed .It Sy Example 7 No Destroying a ZFS Storage Pool The following command destroys the pool .Em tank and any datasets contained within. .Bd -literal # zpool destroy -f tank .Ed .It Sy Example 8 No Exporting a ZFS Storage Pool The following command exports the devices in pool .Em tank so that they can be relocated or later imported. .Bd -literal # zpool export tank .Ed .It Sy Example 9 No Importing a ZFS Storage Pool The following command displays available pools, and then imports the pool .Em tank for use on the system. The results from this command are similar to the following: .Bd -literal # zpool import pool: tank id: 15451357997522795478 state: ONLINE action: The pool can be imported using its name or numeric identifier. config: tank ONLINE mirror ONLINE c1t2d0 ONLINE c1t3d0 ONLINE # zpool import tank .Ed .It Sy Example 10 No Upgrading All ZFS Storage Pools to the Current Version The following command upgrades all ZFS Storage pools to the current version of the software. .Bd -literal # zpool upgrade -a This system is currently running ZFS version 2. .Ed .It Sy Example 11 No Managing Hot Spares The following command creates a new pool with an available hot spare: .Bd -literal # zpool create tank mirror c0t0d0 c0t1d0 spare c0t2d0 .Ed .Pp If one of the disks were to fail, the pool would be reduced to the degraded state. The failed device can be replaced using the following command: .Bd -literal # zpool replace tank c0t0d0 c0t3d0 .Ed .Pp Once the data has been resilvered, the spare is automatically removed and is made available for use should another device fail. The hot spare can be permanently removed from the pool using the following command: .Bd -literal # zpool remove tank c0t2d0 .Ed .It Sy Example 12 No Creating a ZFS Pool with Mirrored Separate Intent Logs The following command creates a ZFS storage pool consisting of two, two-way mirrors and mirrored log devices: .Bd -literal # zpool create pool mirror c0d0 c1d0 mirror c2d0 c3d0 log mirror \e c4d0 c5d0 .Ed .It Sy Example 13 No Adding Cache Devices to a ZFS Pool The following command adds two disks for use as cache devices to a ZFS storage pool: .Bd -literal # zpool add pool cache c2d0 c3d0 .Ed .Pp Once added, the cache devices gradually fill with content from main memory. Depending on the size of your cache devices, it could take over an hour for them to fill. Capacity and reads can be monitored using the .Cm iostat option as follows: .Bd -literal # zpool iostat -v pool 5 .Ed .It Sy Example 14 No Removing a Mirrored top-level (Log or Data) Device The following commands remove the mirrored log device .Sy mirror-2 and mirrored top-level data device .Sy mirror-1 . .Pp Given this configuration: .Bd -literal pool: tank state: ONLINE scrub: none requested config: NAME STATE READ WRITE CKSUM tank ONLINE 0 0 0 mirror-0 ONLINE 0 0 0 c6t0d0 ONLINE 0 0 0 c6t1d0 ONLINE 0 0 0 mirror-1 ONLINE 0 0 0 c6t2d0 ONLINE 0 0 0 c6t3d0 ONLINE 0 0 0 logs mirror-2 ONLINE 0 0 0 c4t0d0 ONLINE 0 0 0 c4t1d0 ONLINE 0 0 0 .Ed .Pp The command to remove the mirrored log .Sy mirror-2 is: .Bd -literal # zpool remove tank mirror-2 .Ed .Pp The command to remove the mirrored data .Sy mirror-1 is: .Bd -literal # zpool remove tank mirror-1 .Ed .It Sy Example 15 No Displaying expanded space on a device The following command displays the detailed information for the pool .Em data . This pool is comprised of a single raidz vdev where one of its devices increased its capacity by 10GB. In this example, the pool will not be able to utilize this extra capacity until all the devices under the raidz vdev have been expanded. .Bd -literal # zpool list -v data NAME SIZE ALLOC FREE FRAG EXPANDSZ CAP DEDUP HEALTH ALTROOT data 23.9G 14.6G 9.30G 48% - 61% 1.00x ONLINE - raidz1 23.9G 14.6G 9.30G 48% - c1t1d0 - - - - - c1t2d0 - - - - 10G c1t3d0 - - - - - .Ed .El .Sh ENVIRONMENT VARIABLES .Bl -tag -width "ZPOOL_VDEV_NAME_GUID" .It Ev ZPOOL_VDEV_NAME_GUID Cause .Nm zpool subcommands to output vdev guids by default. This behavior is identical to the .Nm zpool status -g command line option. .El .Bl -tag -width "ZPOOL_VDEV_NAME_FOLLOW_LINKS" .It Ev ZPOOL_VDEV_NAME_FOLLOW_LINKS Cause .Nm zpool subcommands to follow links for vdev names by default. This behavior is identical to the .Nm zpool status -L command line option. .El .Bl -tag -width "ZPOOL_VDEV_NAME_PATH" .It Ev ZPOOL_VDEV_NAME_PATH Cause .Nm zpool subcommands to output full vdev path names by default. This behavior is identical to the .Nm zpool status -P command line option. .El .Sh INTERFACE STABILITY .Sy Evolving .Sh SEE ALSO .Xr zfs 1M , .Xr attributes 5 , .Xr zpool-features 5 Index: vendor/illumos/dist/man/man5/zpool-features.5 =================================================================== --- vendor/illumos/dist/man/man5/zpool-features.5 (revision 354960) +++ vendor/illumos/dist/man/man5/zpool-features.5 (revision 354961) @@ -1,687 +1,708 @@ '\" te .\" Copyright (c) 2013, 2017 by Delphix. All rights reserved. .\" Copyright (c) 2013 by Saso Kiselkov. All rights reserved. .\" Copyright (c) 2014, Joyent, Inc. All rights reserved. .\" Copyright (c) 2014 Integros [integros.com] .\" 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 http://www.opensolaris.org/os/licensing. .\" .\" 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] -.TH ZPOOL-FEATURES 5 "Apr 19, 2019" +.TH ZPOOL-FEATURES 5 "May 15, 2019" .SH NAME zpool\-features \- ZFS pool feature descriptions .SH DESCRIPTION .LP ZFS pool on\-disk format versions are specified via "features" which replace the old on\-disk format numbers (the last supported on\-disk format number is 28). To enable a feature on a pool use the \fBupgrade\fR subcommand of the \fBzpool\fR(1M) command, or set the \fBfeature@\fR\fIfeature_name\fR property to \fBenabled\fR. .sp .LP The pool format does not affect file system version compatibility or the ability to send file systems between pools. .sp .LP 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 \fBactive\fR on the pool. If the pool was created by another software version this set may include unsupported features. .SS "Identifying features" .LP Every feature has a guid of the form \fIcom.example:feature_name\fR. The reverse 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. .sp .LP Each supported feature also has a short name. By convention a feature's short name is the portion of its guid which follows the ':' (e.g. \fIcom.example:feature_name\fR would have the short name \fIfeature_name\fR), however a feature's short name may differ across ZFS implementations if following the convention would result in name conflicts. .SS "Feature states" .LP Features can be in one of three states: .sp .ne 2 .na \fB\fBactive\fR\fR .ad .RS 12n 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 (see "Read\-only compatibility"). .RE .sp .ne 2 .na \fB\fBenabled\fR\fR .ad .RS 12n 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 \fBactive\fR state. Some features may support returning to the \fBenabled\fR state after becoming \fBactive\fR. See feature\-specific documentation for details. .RE .sp .ne 2 .na \fBdisabled\fR .ad .RS 12n This feature's on\-disk format changes have not been made and will not be made unless an administrator moves the feature to the \fBenabled\fR state. Features cannot be disabled once they have been enabled. .RE .sp .LP The state of supported features is exposed through pool properties of the form \fIfeature@short_name\fR. .SS "Read\-only compatibility" .LP 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 "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 \fBreadonly\fR property during import (see \fBzpool\fR(1M) for details on importing pools). .SS "Unsupported features" .LP For each unsupported feature enabled on an imported pool a pool property named \fIunsupported@feature_guid\fR will indicate why the import was allowed despite the unsupported feature. Possible values for this property are: .sp .ne 2 .na \fB\fBinactive\fR\fR .ad .RS 12n The feature is in the \fBenabled\fR state and therefore the pool's on\-disk format is still compatible with software that does not support this feature. .RE .sp .ne 2 .na \fB\fBreadonly\fR\fR .ad .RS 12n The feature is read\-only compatible and the pool has been imported in read\-only mode. .RE .SS "Feature dependencies" .LP Some features depend on other features being enabled in order to function properly. Enabling a feature will automatically enable any features it depends on. .SH FEATURES .LP The following features are supported on this system: .sp .ne 2 .na \fB\fBasync_destroy\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:async_destroy READ\-ONLY COMPATIBLE yes DEPENDENCIES none .TE Destroying a file system requires traversing all of its data in order to return its used space to the pool. Without \fBasync_destroy\fR 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. When \fBasync_destroy\fR 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 \fBfreeing\fR property. This feature is only \fBactive\fR while \fBfreeing\fR is non\-zero. .RE .sp .ne 2 .na \fB\fBempty_bpobj\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:empty_bpobj READ\-ONLY COMPATIBLE yes DEPENDENCIES none .TE 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. When there are many snapshots, each snapshot uses many Block Pointer Objects (bpobj's) to track blocks associated with that snapshot. However, in common use cases, most of these bpobj's are empty. This feature allows us to create each bpobj on-demand, thus eliminating the empty bpobjs. This feature is \fBactive\fR while there are any filesystems, volumes, or snapshots which were created after enabling this feature. .RE .sp .ne 2 .na \fB\fBfilesystem_limits\fR\fR .ad .RS 4n .TS l l . GUID com.joyent:filesystem_limits READ\-ONLY COMPATIBLE yes DEPENDENCIES extensible_dataset .TE 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. This feature is \fBactive\fR once either of the limit properties has been set on a dataset. Once activated the feature is never deactivated. .RE .sp .ne 2 .na \fB\fBlz4_compress\fR\fR .ad .RS 4n .TS l l . GUID org.illumos:lz4_compress READ\-ONLY COMPATIBLE no DEPENDENCIES none .TE \fBlz4\fR 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 \fBlzjb\fR compression. Typically, \fBlz4\fR compression is approximately 50% faster on compressible data and 200% faster on incompressible data than \fBlzjb\fR. It is also approximately 80% faster on decompression, while giving approximately 10% better compression ratio. When the \fBlz4_compress\fR feature is set to \fBenabled\fR, the administrator can turn on \fBlz4\fR compression on any dataset on the pool using the \fBzfs\fR(1M) command. Also, all newly written metadata will be compressed with \fBlz4\fR algorithm. Since this feature is not read-only compatible, this operation will render the pool unimportable on systems without support for the \fBlz4_compress\fR feature. Booting off of \fBlz4\fR-compressed root pools is supported. This feature becomes \fBactive\fR as soon as it is enabled and will never return to being \fBenabled\fB. .RE .sp .ne 2 .na \fB\fBspacemap_histogram\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:spacemap_histogram READ\-ONLY COMPATIBLE yes DEPENDENCIES none .TE This features allows ZFS to maintain more information about how free space is organized within the pool. If this feature is \fBenabled\fR, ZFS will set this feature to \fBactive\fR when a new space map object is created or an existing space map is upgraded to the new format. Once the feature is \fBactive\fR, it will remain in that state until the pool is destroyed. .RE .sp .ne 2 .na \fB\fBmulti_vdev_crash_dump\fR\fR .ad .RS 4n .TS l l . GUID com.joyent:multi_vdev_crash_dump READ\-ONLY COMPATIBLE no DEPENDENCIES none .TE 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. When the \fBmulti_vdev_crash_dump\fR feature is set to \fBenabled\fR, the administrator can use the \fBdumpadm\fR(1M) command to configure a dump device on a pool comprised of multiple vdevs. .RE .sp .ne 2 .na \fB\fBextensible_dataset\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:extensible_dataset READ\-ONLY COMPATIBLE no DEPENDENCIES none .TE This feature allows more flexible use of internal ZFS data structures, and exists for other features to depend on. This feature will be \fBactive\fR when the first dependent feature uses it, and will be returned to the \fBenabled\fR state when all datasets that use this feature are destroyed. .RE .sp .ne 2 .na \fB\fBbookmarks\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:bookmarks READ\-ONLY COMPATIBLE yes DEPENDENCIES extensible_dataset .TE This feature enables use of the \fBzfs bookmark\fR subcommand. This feature is \fBactive\fR while any bookmarks exist in the pool. All bookmarks in the pool can be listed by running \fBzfs list -t bookmark -r \fIpoolname\fR\fR. .RE .sp .ne 2 .na \fB\fBenabled_txg\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:enabled_txg READ\-ONLY COMPATIBLE yes DEPENDENCIES none .TE 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. This feature becomes \fBactive\fR as soon as it is enabled and will never return to being \fBenabled\fB. .RE .sp .ne 2 .na \fB\fBhole_birth\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:hole_birth READ\-ONLY COMPATIBLE no DEPENDENCIES enabled_txg .TE This feature improves performance of incremental sends ("zfs send -i") and receives for objects with many holes. The most common case of hole-filled objects is zvols. An incremental send stream from snapshot \fBA\fR to snapshot \fBB\fR contains information about every block that changed between \fBA\fR and \fBB\fR. Blocks which did not change between those snapshots can be identified and omitted from the stream using a piece of metadata called the 'block birth time', but birth times are not recorded for holes (blocks filled only with zeroes). Since holes created after \fBA\fR cannot be distinguished from holes created before \fBA\fR, information about every hole in the entire filesystem or zvol is included in the send stream. For workloads where holes are rare this is not a problem. However, when incrementally replicating filesystems or zvols with many holes (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. Once the \fBhole_birth\fR 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. This feature becomes \fBactive\fR as soon as it is enabled and will never return to being \fBenabled\fB. .RE .sp .ne 2 .na \fB\fBembedded_data\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:embedded_data READ\-ONLY COMPATIBLE no DEPENDENCIES none .TE 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. When this feature is enabled, the contents of highly-compressible blocks are stored in the block "pointer" itself (a misnomer in this case, as it contains the compresseed data, rather than a pointer to its location on disk). Thus the space of the block (one sector, typically 512 bytes or 4KB) is saved, and no additional i/o is needed to read and write the data block. This feature becomes \fBactive\fR as soon as it is enabled and will never return to being \fBenabled\fR. .RE .sp .ne 2 .na \fB\fBdevice_removal\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:device_removal READ\-ONLY COMPATIBLE no DEPENDENCIES none .TE This feature enables the "zpool remove" subcommand to remove top-level vdevs, evacuating them to reduce the total size of the pool. This feature becomes \fBactive\fR when the "zpool remove" command is used on a top-level vdev, and will never return to being \fBenabled\fR. .RE .sp .ne 2 .na \fB\fBobsolete_counts\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:obsolete_counts READ\-ONLY COMPATIBLE yes DEPENDENCIES device_removal .TE This feature is an enhancement of 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 "obsolete", i.e. no longer needed. See also the \fBzfs remap\fR subcommand in \fBzfs\fR(1M). This feature becomes \fBactive\fR when the "zpool remove" command is used on a top-level vdev, and will never return to being \fBenabled\fR. .RE .sp .ne 2 .na \fB\fBzpool_checkpoint\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:zpool_checkpoint READ\-ONLY COMPATIBLE yes DEPENDENCIES none .TE This feature enables the "zpool checkpoint" subcommand that can checkpoint the state of the pool at the time it was issued and later rewind back to it or discard it. This feature becomes \fBactive\fR when the "zpool checkpoint" command is used to checkpoint the pool. The feature will only return back to being \fBenabled\fR when the pool is rewound or the checkpoint has been discarded. .RE .sp .ne 2 .na \fB\fBspacemap_v2\fR\fR .ad .RS 4n .TS l l . GUID com.delphix:spacemap_v2 READ\-ONLY COMPATIBLE yes DEPENDENCIES none .TE This feature enables the use of the new space map encoding which consists of two words (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. This feature becomes \fBactive\fR once it is \fBenabled\fR, and never returns back to being \fBenabled\fR. .RE .sp .ne 2 .na \fB\fBlarge_blocks\fR\fR .ad .RS 4n .TS l l . GUID org.open-zfs:large_block READ\-ONLY COMPATIBLE no DEPENDENCIES extensible_dataset .TE The \fBlarge_block\fR feature allows the record size on a dataset to be set larger than 128KB. This feature becomes \fBactive\fR once a \fBrecordsize\fR property has been set larger than 128KB, and will return to being \fBenabled\fR once all filesystems that have ever had their recordsize larger than 128KB are destroyed. .RE .ne 2 .na \fB\fBlarge_dnode\fR\fR .ad .RS 4n .TS l l . GUID org.zfsonlinux:large_dnode READ\-ONLY COMPATIBLE no DEPENDENCIES extensible_dataset .TE The \fBlarge_dnode\fR feature allows the size of dnodes in a dataset to be set larger than 512B. This feature becomes \fBactive\fR once a dataset contains an object with a dnode larger than 512B, which occurs as a result of setting the \fBdnodesize\fR dataset property to a value other than \fBlegacy\fR. The feature will return to being \fBenabled\fR once all filesystems that have ever contained a dnode larger than 512B 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. .RE .sp .ne 2 .na \fB\fBsha512\fR\fR .ad .RS 4n .TS l l . GUID org.illumos:sha512 READ\-ONLY COMPATIBLE no DEPENDENCIES extensible_dataset .TE This feature enables the use of the SHA-512/256 truncated hash algorithm (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 \fBskein\fR and \fBedonr\fR algorithms. When the \fBsha512\fR feature is set to \fBenabled\fR, the administrator can turn on the \fBsha512\fR checksum on any dataset using the \fBzfs set checksum=sha512\fR(1M) command. This feature becomes \fBactive\fR once a \fBchecksum\fR property has been set to \fBsha512\fR, and will return to being \fBenabled\fR once all filesystems that have ever had their checksum set to \fBsha512\fR are destroyed. Booting off of pools utilizing SHA-512/256 is supported. .RE .sp .ne 2 .na \fB\fBskein\fR\fR .ad .RS 4n .TS l l . GUID org.illumos:skein READ\-ONLY COMPATIBLE no DEPENDENCIES extensible_dataset .TE 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 (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 (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. When the \fBskein\fR feature is set to \fBenabled\fR, the administrator can turn on the \fBskein\fR checksum on any dataset using the \fBzfs set checksum=skein\fR(1M) command. This feature becomes \fBactive\fR once a \fBchecksum\fR property has been set to \fBskein\fR, and will return to being \fBenabled\fR once all filesystems that have ever had their checksum set to \fBskein\fR are destroyed. Booting off of pools using \fBskein\fR is supported. .RE .sp .ne 2 .na \fB\fBedonr\fR\fR .ad .RS 4n .TS l l . GUID org.illumos:edonr READ\-ONLY COMPATIBLE no DEPENDENCIES extensible_dataset .TE This feature enables the use of the Edon-R hash algorithm for checksum, including for nopwrite (if compression is also enabled, an overwrite of a block whose checksum matches the data being written will be ignored). In an abundance of caution, Edon-R can not be used with dedup (without verification). Edon-R is a very high-performance hash algorithm that was part of the NIST SHA-3 competition. It provides extremely high hash performance (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 (stored on the pool) before being fed the data block to be checksummed. Thus the produced checksums are unique to a given pool. When the \fBedonr\fR feature is set to \fBenabled\fR, the administrator can turn on the \fBedonr\fR checksum on any dataset using the \fBzfs set checksum=edonr\fR(1M) command. This feature becomes \fBactive\fR once a \fBchecksum\fR property has been set to \fBedonr\fR, and will return to being \fBenabled\fR once all filesystems that have ever had their checksum set to \fBedonr\fR are destroyed. Booting off of pools using \fBedonr\fR is supported. .RE .sp .ne 2 .na \fB\fBallocation_classes\fR\fR .ad .RS 4n .TS l l . GUID com.intel:allocation_classes READ\-ONLY COMPATIBLE yes DEPENDENCIES none .TE This feature enables support for separate allocation classes. This feature becomes \fBactive\fR when a dedicated allocation class vdev (dedup or special) is created with zpool create or zpool add. With device removal, it can be returned to the \fBenabled\fR state if all the top-level vdevs from an allocation class are removed. + +.RE +.sp +.ne 2 +.na +\fB\fBresilver_defer\fR\fR +.ad +.RS 4n +.TS +l l . +GUID com.datto:resilver_defer +READ\-ONLY COMPATIBLE yes +DEPENDENCIES none +.TE + +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. + +This feature becomes \fBactive\fR once a resilver has been deferred, and +returns to being \fBenabled\fR when the deferred resilver begins. .SH "SEE ALSO" \fBzpool\fR(1M) Index: vendor-sys/illumos/dist/common/zfs/zfeature_common.c =================================================================== --- vendor-sys/illumos/dist/common/zfs/zfeature_common.c (revision 354960) +++ vendor-sys/illumos/dist/common/zfs/zfeature_common.c (revision 354961) @@ -1,309 +1,312 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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) 2014 Integros [integros.com] * Copyright (c) 2017, Intel Corporation. */ #ifdef _KERNEL #include #else #include #include #endif #include #include #include #include #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 (strcmp(guid, feature->fi_guid) == 0) return (B_TRUE); } return (B_FALSE); } 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 (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 void zfeature_register(spa_feature_t fid, const char *guid, const char *name, const char *desc, zfeature_flags_t flags, const spa_feature_t *deps) { zfeature_info_t *feature = &spa_feature_table[fid]; static 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; feature->fi_feature = fid; feature->fi_guid = guid; feature->fi_uname = name; feature->fi_desc = desc; feature->fi_flags = flags; feature->fi_depends = deps; } void zpool_feature_init(void) { zfeature_register(SPA_FEATURE_ASYNC_DESTROY, "com.delphix:async_destroy", "async_destroy", "Destroy filesystems asynchronously.", ZFEATURE_FLAG_READONLY_COMPAT, NULL); zfeature_register(SPA_FEATURE_EMPTY_BPOBJ, "com.delphix:empty_bpobj", "empty_bpobj", "Snapshots use less space.", ZFEATURE_FLAG_READONLY_COMPAT, NULL); zfeature_register(SPA_FEATURE_LZ4_COMPRESS, "org.illumos:lz4_compress", "lz4_compress", "LZ4 compression algorithm support.", ZFEATURE_FLAG_ACTIVATE_ON_ENABLE, NULL); 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, NULL); zfeature_register(SPA_FEATURE_SPACEMAP_HISTOGRAM, "com.delphix:spacemap_histogram", "spacemap_histogram", "Spacemaps maintain space histograms.", ZFEATURE_FLAG_READONLY_COMPAT, NULL); zfeature_register(SPA_FEATURE_ENABLED_TXG, "com.delphix:enabled_txg", "enabled_txg", "Record txg at which a feature is enabled", ZFEATURE_FLAG_READONLY_COMPAT, NULL); static 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, hole_birth_deps); zfeature_register(SPA_FEATURE_EXTENSIBLE_DATASET, "com.delphix:extensible_dataset", "extensible_dataset", "Enhanced dataset functionality, used by other features.", 0, NULL); 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, bookmarks_deps); 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, filesystem_limits_deps); 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, NULL); zfeature_register(SPA_FEATURE_POOL_CHECKPOINT, "com.delphix:zpool_checkpoint", "zpool_checkpoint", "Pool state can be checkpointed, allowing rewind later.", ZFEATURE_FLAG_READONLY_COMPAT, NULL); 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, NULL); 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, large_blocks_deps); { 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, large_dnode_deps); } 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, sha512_deps); 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, skein_deps); 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, edonr_deps); 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, NULL); 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, obsolete_counts_deps); - { zfeature_register(SPA_FEATURE_ALLOCATION_CLASSES, "org.zfsonlinux:allocation_classes", "allocation_classes", "Support for separate allocation classes.", ZFEATURE_FLAG_READONLY_COMPAT, NULL); - } + + zfeature_register(SPA_FEATURE_RESILVER_DEFER, + "com.datto:resilver_defer", "resilver_defer", + "Support for defering new resilvers when one is already running.", + ZFEATURE_FLAG_READONLY_COMPAT, NULL); } Index: vendor-sys/illumos/dist/common/zfs/zfeature_common.h =================================================================== --- vendor-sys/illumos/dist/common/zfs/zfeature_common.h (revision 354960) +++ vendor-sys/illumos/dist/common/zfs/zfeature_common.h (revision 354961) @@ -1,110 +1,111 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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, 2017 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 Integros [integros.com] * Copyright (c) 2017, Intel Corporation. */ #ifndef _ZFEATURE_COMMON_H #define _ZFEATURE_COMMON_H #include #include #include #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_DEVICE_REMOVAL, SPA_FEATURE_OBSOLETE_COUNTS, SPA_FEATURE_POOL_CHECKPOINT, SPA_FEATURE_SPACEMAP_V2, SPA_FEATURE_ALLOCATION_CLASSES, + SPA_FEATURE_RESILVER_DEFER, 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 read 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 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; /* 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 extern zfeature_info_t spa_feature_table[SPA_FEATURES]; extern boolean_t zfeature_is_valid_guid(const char *); extern boolean_t zfeature_is_supported(const char *); extern int zfeature_lookup_name(const char *, spa_feature_t *); extern boolean_t zfeature_depends_on(spa_feature_t, spa_feature_t); extern void zpool_feature_init(void); #ifdef __cplusplus } #endif #endif /* _ZFEATURE_COMMON_H */ Index: vendor-sys/illumos/dist/uts/common/fs/zfs/dsl_scan.c =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/dsl_scan.c (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/dsl_scan.c (revision 354961) @@ -1,3979 +1,4115 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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) 2008, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2018 by Delphix. All rights reserved. * Copyright 2016 Gary Mills * Copyright (c) 2011, 2017 by Delphix. All rights reserved. * Copyright 2017 Joyent, Inc. * Copyright (c) 2017 Datto Inc. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef _KERNEL #include #endif /* * Grand theory statement on scan queue sorting * * Scanning is implemented by recursively traversing all indirection levels * in an object and reading all blocks referenced from said objects. This * results in us approximately traversing the object from lowest logical * offset to the highest. For best performance, we would want the logical * blocks to be physically contiguous. However, this is frequently not the * case with pools given the allocation patterns of copy-on-write filesystems. * So instead, we put the I/Os into a reordering queue and issue them in a * way that will most benefit physical disks (LBA-order). * * Queue management: * * Ideally, we would want to scan all metadata and queue up all block I/O * prior to starting to issue it, because that allows us to do an optimal * sorting job. This can however consume large amounts of memory. Therefore * we continuously monitor the size of the queues and constrain them to 5% * (zfs_scan_mem_lim_fact) of physmem. If the queues grow larger than this * limit, we clear out a few of the largest extents at the head of the queues * to make room for more scanning. Hopefully, these extents will be fairly * large and contiguous, allowing us to approach sequential I/O throughput * even without a fully sorted tree. * * Metadata scanning takes place in dsl_scan_visit(), which is called from * dsl_scan_sync() every spa_sync(). If we have either fully scanned all * metadata on the pool, or we need to make room in memory because our * queues are too large, dsl_scan_visit() is postponed and * scan_io_queues_run() is called from dsl_scan_sync() instead. This implies * that metadata scanning and queued I/O issuing are mutually exclusive. This * allows us to provide maximum sequential I/O throughput for the majority of * I/O's issued since sequential I/O performance is significantly negatively * impacted if it is interleaved with random I/O. * * Implementation Notes * * One side effect of the queued scanning algorithm is that the scanning code * needs to be notified whenever a block is freed. This is needed to allow * the scanning code to remove these I/Os from the issuing queue. Additionally, * we do not attempt to queue gang blocks to be issued sequentially since this * is very hard to do and would have an extremely limited performance benefit. * Instead, we simply issue gang I/Os as soon as we find them using the legacy * algorithm. * * Backwards compatibility * * This new algorithm is backwards compatible with the legacy on-disk data * structures (and therefore does not require a new feature flag). * Periodically during scanning (see zfs_scan_checkpoint_intval), the scan * will stop scanning metadata (in logical order) and wait for all outstanding * sorted I/O to complete. Once this is done, we write out a checkpoint * bookmark, indicating that we have scanned everything logically before it. * If the pool is imported on a machine without the new sorting algorithm, * the scan simply resumes from the last checkpoint using the legacy algorithm. */ typedef int (scan_cb_t)(dsl_pool_t *, const blkptr_t *, const zbookmark_phys_t *); static scan_cb_t dsl_scan_scrub_cb; static int scan_ds_queue_compare(const void *a, const void *b); static int scan_prefetch_queue_compare(const void *a, const void *b); static void scan_ds_queue_clear(dsl_scan_t *scn); static boolean_t scan_ds_queue_contains(dsl_scan_t *scn, uint64_t dsobj, uint64_t *txg); static void scan_ds_queue_insert(dsl_scan_t *scn, uint64_t dsobj, uint64_t txg); static void scan_ds_queue_remove(dsl_scan_t *scn, uint64_t dsobj); static void scan_ds_queue_sync(dsl_scan_t *scn, dmu_tx_t *tx); extern int zfs_vdev_async_write_active_min_dirty_percent; /* * By default zfs will check to ensure it is not over the hard memory * limit before each txg. If finer-grained control of this is needed * this value can be set to 1 to enable checking before scanning each * block. */ int zfs_scan_strict_mem_lim = B_FALSE; /* * Maximum number of parallelly executing I/Os per top-level vdev. * Tune with care. Very high settings (hundreds) are known to trigger * some firmware bugs and resets on certain SSDs. */ int zfs_top_maxinflight = 32; /* maximum I/Os per top-level */ unsigned int zfs_resilver_delay = 2; /* number of ticks to delay resilver */ unsigned int zfs_scrub_delay = 4; /* number of ticks to delay scrub */ unsigned int zfs_scan_idle = 50; /* idle window in clock ticks */ /* * Maximum number of parallelly executed bytes per leaf vdev. We attempt * to strike a balance here between keeping the vdev queues full of I/Os * at all times and not overflowing the queues to cause long latency, * which would cause long txg sync times. No matter what, we will not * overload the drives with I/O, since that is protected by * zfs_vdev_scrub_max_active. */ unsigned long zfs_scan_vdev_limit = 4 << 20; int zfs_scan_issue_strategy = 0; int zfs_scan_legacy = B_FALSE; /* don't queue & sort zios, go direct */ uint64_t zfs_scan_max_ext_gap = 2 << 20; /* in bytes */ unsigned int zfs_scan_checkpoint_intval = 7200; /* seconds */ #define ZFS_SCAN_CHECKPOINT_INTVAL SEC_TO_TICK(zfs_scan_checkpoint_intval) /* * fill_weight is non-tunable at runtime, so we copy it at module init from * zfs_scan_fill_weight. Runtime adjustments to zfs_scan_fill_weight would * break queue sorting. */ uint64_t zfs_scan_fill_weight = 3; static uint64_t fill_weight; /* See dsl_scan_should_clear() for details on the memory limit tunables */ uint64_t zfs_scan_mem_lim_min = 16 << 20; /* bytes */ uint64_t zfs_scan_mem_lim_soft_max = 128 << 20; /* bytes */ int zfs_scan_mem_lim_fact = 20; /* fraction of physmem */ int zfs_scan_mem_lim_soft_fact = 20; /* fraction of mem lim above */ unsigned int zfs_scrub_min_time_ms = 1000; /* min millisecs to scrub per txg */ unsigned int zfs_free_min_time_ms = 1000; /* min millisecs to free per txg */ /* min millisecs to obsolete per txg */ unsigned int zfs_obsolete_min_time_ms = 500; /* min millisecs to resilver per txg */ unsigned int zfs_resilver_min_time_ms = 3000; +int zfs_scan_suspend_progress = 0; /* set to prevent scans from progressing */ boolean_t zfs_no_scrub_io = B_FALSE; /* set to disable scrub i/o */ boolean_t zfs_no_scrub_prefetch = B_FALSE; /* set to disable scrub prefetch */ enum ddt_class zfs_scrub_ddt_class_max = DDT_CLASS_DUPLICATE; /* max number of blocks to free in a single TXG */ uint64_t zfs_async_block_max_blocks = UINT64_MAX; +int zfs_resilver_disable_defer = 0; /* set to disable resilver deferring */ + /* * We wait a few txgs after importing a pool to begin scanning so that * the import / mounting code isn't held up by scrub / resilver IO. * Unfortunately, it is a bit difficult to determine exactly how long * this will take since userspace will trigger fs mounts asynchronously * and the kernel will create zvol minors asynchronously. As a result, * the value provided here is a bit arbitrary, but represents a * reasonable estimate of how many txgs it will take to finish fully * importing a pool */ #define SCAN_IMPORT_WAIT_TXGS 5 #define DSL_SCAN_IS_SCRUB_RESILVER(scn) \ ((scn)->scn_phys.scn_func == POOL_SCAN_SCRUB || \ (scn)->scn_phys.scn_func == POOL_SCAN_RESILVER) extern int zfs_txg_timeout; /* * Enable/disable the processing of the free_bpobj object. */ boolean_t zfs_free_bpobj_enabled = B_TRUE; /* the order has to match pool_scan_type */ static scan_cb_t *scan_funcs[POOL_SCAN_FUNCS] = { NULL, dsl_scan_scrub_cb, /* POOL_SCAN_SCRUB */ dsl_scan_scrub_cb, /* POOL_SCAN_RESILVER */ }; /* In core node for the scn->scn_queue. Represents a dataset to be scanned */ typedef struct { uint64_t sds_dsobj; uint64_t sds_txg; avl_node_t sds_node; } scan_ds_t; /* * This controls what conditions are placed on dsl_scan_sync_state(): * SYNC_OPTIONAL) write out scn_phys iff scn_bytes_pending == 0 * SYNC_MANDATORY) write out scn_phys always. scn_bytes_pending must be 0. * SYNC_CACHED) if scn_bytes_pending == 0, write out scn_phys. Otherwise * write out the scn_phys_cached version. * See dsl_scan_sync_state for details. */ typedef enum { SYNC_OPTIONAL, SYNC_MANDATORY, SYNC_CACHED } state_sync_type_t; /* * This struct represents the minimum information needed to reconstruct a * zio for sequential scanning. This is useful because many of these will * accumulate in the sequential IO queues before being issued, so saving * memory matters here. */ typedef struct scan_io { /* fields from blkptr_t */ uint64_t sio_blk_prop; uint64_t sio_phys_birth; uint64_t sio_birth; zio_cksum_t sio_cksum; uint32_t sio_nr_dvas; /* fields from zio_t */ uint32_t sio_flags; zbookmark_phys_t sio_zb; /* members for queue sorting */ union { avl_node_t sio_addr_node; /* link into issuing queue */ list_node_t sio_list_node; /* link for issuing to disk */ } sio_nodes; /* * There may be up to SPA_DVAS_PER_BP DVAs here from the bp, * depending on how many were in the original bp. Only the * first DVA is really used for sorting and issuing purposes. * The other DVAs (if provided) simply exist so that the zio * layer can find additional copies to repair from in the * event of an error. This array must go at the end of the * struct to allow this for the variable number of elements. */ dva_t sio_dva[0]; } scan_io_t; #define SIO_SET_OFFSET(sio, x) DVA_SET_OFFSET(&(sio)->sio_dva[0], x) #define SIO_SET_ASIZE(sio, x) DVA_SET_ASIZE(&(sio)->sio_dva[0], x) #define SIO_GET_OFFSET(sio) DVA_GET_OFFSET(&(sio)->sio_dva[0]) #define SIO_GET_ASIZE(sio) DVA_GET_ASIZE(&(sio)->sio_dva[0]) #define SIO_GET_END_OFFSET(sio) \ (SIO_GET_OFFSET(sio) + SIO_GET_ASIZE(sio)) #define SIO_GET_MUSED(sio) \ (sizeof (scan_io_t) + ((sio)->sio_nr_dvas * sizeof (dva_t))) struct dsl_scan_io_queue { dsl_scan_t *q_scn; /* associated dsl_scan_t */ vdev_t *q_vd; /* top-level vdev that this queue represents */ /* trees used for sorting I/Os and extents of I/Os */ range_tree_t *q_exts_by_addr; avl_tree_t q_exts_by_size; avl_tree_t q_sios_by_addr; uint64_t q_sio_memused; /* members for zio rate limiting */ uint64_t q_maxinflight_bytes; uint64_t q_inflight_bytes; kcondvar_t q_zio_cv; /* used under vd->vdev_scan_io_queue_lock */ /* per txg statistics */ uint64_t q_total_seg_size_this_txg; uint64_t q_segs_this_txg; uint64_t q_total_zio_size_this_txg; uint64_t q_zios_this_txg; }; /* private data for dsl_scan_prefetch_cb() */ typedef struct scan_prefetch_ctx { zfs_refcount_t spc_refcnt; /* refcount for memory management */ dsl_scan_t *spc_scn; /* dsl_scan_t for the pool */ boolean_t spc_root; /* is this prefetch for an objset? */ uint8_t spc_indblkshift; /* dn_indblkshift of current dnode */ uint16_t spc_datablkszsec; /* dn_idatablkszsec of current dnode */ } scan_prefetch_ctx_t; /* private data for dsl_scan_prefetch() */ typedef struct scan_prefetch_issue_ctx { avl_node_t spic_avl_node; /* link into scn->scn_prefetch_queue */ scan_prefetch_ctx_t *spic_spc; /* spc for the callback */ blkptr_t spic_bp; /* bp to prefetch */ zbookmark_phys_t spic_zb; /* bookmark to prefetch */ } scan_prefetch_issue_ctx_t; static void scan_exec_io(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb, dsl_scan_io_queue_t *queue); static void scan_io_queue_insert_impl(dsl_scan_io_queue_t *queue, scan_io_t *sio); static dsl_scan_io_queue_t *scan_io_queue_create(vdev_t *vd); static void scan_io_queues_destroy(dsl_scan_t *scn); static kmem_cache_t *sio_cache[SPA_DVAS_PER_BP]; /* sio->sio_nr_dvas must be set so we know which cache to free from */ static void sio_free(scan_io_t *sio) { ASSERT3U(sio->sio_nr_dvas, >, 0); ASSERT3U(sio->sio_nr_dvas, <=, SPA_DVAS_PER_BP); kmem_cache_free(sio_cache[sio->sio_nr_dvas - 1], sio); } /* It is up to the caller to set sio->sio_nr_dvas for freeing */ static scan_io_t * sio_alloc(unsigned short nr_dvas) { ASSERT3U(nr_dvas, >, 0); ASSERT3U(nr_dvas, <=, SPA_DVAS_PER_BP); return (kmem_cache_alloc(sio_cache[nr_dvas - 1], KM_SLEEP)); } void scan_init(void) { /* * This is used in ext_size_compare() to weight segments * based on how sparse they are. This cannot be changed * mid-scan and the tree comparison functions don't currently * have a mechansim for passing additional context to the * compare functions. Thus we store this value globally and * we only allow it to be set at module intiailization time */ fill_weight = zfs_scan_fill_weight; for (int i = 0; i < SPA_DVAS_PER_BP; i++) { char name[36]; (void) sprintf(name, "sio_cache_%d", i); sio_cache[i] = kmem_cache_create(name, (sizeof (scan_io_t) + ((i + 1) * sizeof (dva_t))), 0, NULL, NULL, NULL, NULL, NULL, 0); } } void scan_fini(void) { for (int i = 0; i < SPA_DVAS_PER_BP; i++) { kmem_cache_destroy(sio_cache[i]); } } static inline boolean_t dsl_scan_is_running(const dsl_scan_t *scn) { return (scn->scn_phys.scn_state == DSS_SCANNING); } boolean_t dsl_scan_resilvering(dsl_pool_t *dp) { return (dsl_scan_is_running(dp->dp_scan) && dp->dp_scan->scn_phys.scn_func == POOL_SCAN_RESILVER); } static inline void sio2bp(const scan_io_t *sio, blkptr_t *bp) { bzero(bp, sizeof (*bp)); bp->blk_prop = sio->sio_blk_prop; bp->blk_phys_birth = sio->sio_phys_birth; bp->blk_birth = sio->sio_birth; bp->blk_fill = 1; /* we always only work with data pointers */ bp->blk_cksum = sio->sio_cksum; ASSERT3U(sio->sio_nr_dvas, >, 0); ASSERT3U(sio->sio_nr_dvas, <=, SPA_DVAS_PER_BP); bcopy(sio->sio_dva, bp->blk_dva, sio->sio_nr_dvas * sizeof (dva_t)); } static inline void bp2sio(const blkptr_t *bp, scan_io_t *sio, int dva_i) { sio->sio_blk_prop = bp->blk_prop; sio->sio_phys_birth = bp->blk_phys_birth; sio->sio_birth = bp->blk_birth; sio->sio_cksum = bp->blk_cksum; sio->sio_nr_dvas = BP_GET_NDVAS(bp); /* * Copy the DVAs to the sio. We need all copies of the block so * that the self healing code can use the alternate copies if the * first is corrupted. We want the DVA at index dva_i to be first * in the sio since this is the primary one that we want to issue. */ for (int i = 0, j = dva_i; i < sio->sio_nr_dvas; i++, j++) { sio->sio_dva[i] = bp->blk_dva[j % sio->sio_nr_dvas]; } } int dsl_scan_init(dsl_pool_t *dp, uint64_t txg) { int err; dsl_scan_t *scn; spa_t *spa = dp->dp_spa; uint64_t f; scn = dp->dp_scan = kmem_zalloc(sizeof (dsl_scan_t), KM_SLEEP); scn->scn_dp = dp; /* * It's possible that we're resuming a scan after a reboot so * make sure that the scan_async_destroying flag is initialized * appropriately. */ ASSERT(!scn->scn_async_destroying); scn->scn_async_destroying = spa_feature_is_active(dp->dp_spa, SPA_FEATURE_ASYNC_DESTROY); - bcopy(&scn->scn_phys, &scn->scn_phys_cached, sizeof (scn->scn_phys)); avl_create(&scn->scn_queue, scan_ds_queue_compare, sizeof (scan_ds_t), offsetof(scan_ds_t, sds_node)); avl_create(&scn->scn_prefetch_queue, scan_prefetch_queue_compare, sizeof (scan_prefetch_issue_ctx_t), offsetof(scan_prefetch_issue_ctx_t, spic_avl_node)); err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, "scrub_func", sizeof (uint64_t), 1, &f); if (err == 0) { /* * There was an old-style scrub in progress. Restart a * new-style scrub from the beginning. */ scn->scn_restart_txg = txg; zfs_dbgmsg("old-style scrub was in progress; " "restarting new-style scrub in txg %llu", (longlong_t)scn->scn_restart_txg); /* * Load the queue obj from the old location so that it * can be freed by dsl_scan_done(). */ (void) zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, "scrub_queue", sizeof (uint64_t), 1, &scn->scn_phys.scn_queue_obj); } else { err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys); if (err == ENOENT) return (0); else if (err) return (err); /* * We might be restarting after a reboot, so jump the issued * counter to how far we've scanned. We know we're consistent * up to here. */ scn->scn_issued_before_pass = scn->scn_phys.scn_examined; if (dsl_scan_is_running(scn) && spa_prev_software_version(dp->dp_spa) < SPA_VERSION_SCAN) { /* * A new-type scrub was in progress on an old * pool, and the pool was accessed by old * software. Restart from the beginning, since * the old software may have changed the pool in * the meantime. */ scn->scn_restart_txg = txg; zfs_dbgmsg("new-style scrub was modified " "by old software; restarting in txg %llu", (longlong_t)scn->scn_restart_txg); } } + bcopy(&scn->scn_phys, &scn->scn_phys_cached, sizeof (scn->scn_phys)); + /* reload the queue into the in-core state */ if (scn->scn_phys.scn_queue_obj != 0) { zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, dp->dp_meta_objset, scn->scn_phys.scn_queue_obj); zap_cursor_retrieve(&zc, &za) == 0; (void) zap_cursor_advance(&zc)) { scan_ds_queue_insert(scn, zfs_strtonum(za.za_name, NULL), za.za_first_integer); } zap_cursor_fini(&zc); } spa_scan_stat_init(spa); return (0); } void dsl_scan_fini(dsl_pool_t *dp) { if (dp->dp_scan != NULL) { dsl_scan_t *scn = dp->dp_scan; if (scn->scn_taskq != NULL) taskq_destroy(scn->scn_taskq); scan_ds_queue_clear(scn); avl_destroy(&scn->scn_queue); avl_destroy(&scn->scn_prefetch_queue); kmem_free(dp->dp_scan, sizeof (dsl_scan_t)); dp->dp_scan = NULL; } } static boolean_t dsl_scan_restarting(dsl_scan_t *scn, dmu_tx_t *tx) { return (scn->scn_restart_txg != 0 && scn->scn_restart_txg <= tx->tx_txg); } boolean_t dsl_scan_scrubbing(const dsl_pool_t *dp) { dsl_scan_phys_t *scn_phys = &dp->dp_scan->scn_phys; return (scn_phys->scn_state == DSS_SCANNING && scn_phys->scn_func == POOL_SCAN_SCRUB); } boolean_t dsl_scan_is_paused_scrub(const dsl_scan_t *scn) { return (dsl_scan_scrubbing(scn->scn_dp) && scn->scn_phys.scn_flags & DSF_SCRUB_PAUSED); } /* * Writes out a persistent dsl_scan_phys_t record to the pool directory. * Because we can be running in the block sorting algorithm, we do not always * want to write out the record, only when it is "safe" to do so. This safety * condition is achieved by making sure that the sorting queues are empty * (scn_bytes_pending == 0). When this condition is not true, the sync'd state * is inconsistent with how much actual scanning progress has been made. The * kind of sync to be performed is specified by the sync_type argument. If the * sync is optional, we only sync if the queues are empty. If the sync is * mandatory, we do a hard ASSERT to make sure that the queues are empty. The * third possible state is a "cached" sync. This is done in response to: * 1) The dataset that was in the last sync'd dsl_scan_phys_t having been * destroyed, so we wouldn't be able to restart scanning from it. * 2) The snapshot that was in the last sync'd dsl_scan_phys_t having been * superseded by a newer snapshot. * 3) The dataset that was in the last sync'd dsl_scan_phys_t having been * swapped with its clone. * In all cases, a cached sync simply rewrites the last record we've written, * just slightly modified. For the modifications that are performed to the * last written dsl_scan_phys_t, see dsl_scan_ds_destroyed, * dsl_scan_ds_snapshotted and dsl_scan_ds_clone_swapped. */ static void dsl_scan_sync_state(dsl_scan_t *scn, dmu_tx_t *tx, state_sync_type_t sync_type) { int i; spa_t *spa = scn->scn_dp->dp_spa; ASSERT(sync_type != SYNC_MANDATORY || scn->scn_bytes_pending == 0); if (scn->scn_bytes_pending == 0) { for (i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; dsl_scan_io_queue_t *q = vd->vdev_scan_io_queue; if (q == NULL) continue; mutex_enter(&vd->vdev_scan_io_queue_lock); ASSERT3P(avl_first(&q->q_sios_by_addr), ==, NULL); ASSERT3P(avl_first(&q->q_exts_by_size), ==, NULL); ASSERT3P(range_tree_first(q->q_exts_by_addr), ==, NULL); mutex_exit(&vd->vdev_scan_io_queue_lock); } if (scn->scn_phys.scn_queue_obj != 0) scan_ds_queue_sync(scn, tx); VERIFY0(zap_update(scn->scn_dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys, tx)); bcopy(&scn->scn_phys, &scn->scn_phys_cached, sizeof (scn->scn_phys)); if (scn->scn_checkpointing) zfs_dbgmsg("finish scan checkpoint"); scn->scn_checkpointing = B_FALSE; scn->scn_last_checkpoint = ddi_get_lbolt(); } else if (sync_type == SYNC_CACHED) { VERIFY0(zap_update(scn->scn_dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys_cached, tx)); } } /* ARGSUSED */ static int dsl_scan_setup_check(void *arg, dmu_tx_t *tx) { dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; if (dsl_scan_is_running(scn)) return (SET_ERROR(EBUSY)); return (0); } static void dsl_scan_setup_sync(void *arg, dmu_tx_t *tx) { dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; pool_scan_func_t *funcp = arg; dmu_object_type_t ot = 0; dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; ASSERT(!dsl_scan_is_running(scn)); ASSERT(*funcp > POOL_SCAN_NONE && *funcp < POOL_SCAN_FUNCS); bzero(&scn->scn_phys, sizeof (scn->scn_phys)); scn->scn_phys.scn_func = *funcp; scn->scn_phys.scn_state = DSS_SCANNING; scn->scn_phys.scn_min_txg = 0; scn->scn_phys.scn_max_txg = tx->tx_txg; scn->scn_phys.scn_ddt_class_max = DDT_CLASSES - 1; /* the entire DDT */ scn->scn_phys.scn_start_time = gethrestime_sec(); scn->scn_phys.scn_errors = 0; scn->scn_phys.scn_to_examine = spa->spa_root_vdev->vdev_stat.vs_alloc; scn->scn_issued_before_pass = 0; scn->scn_restart_txg = 0; scn->scn_done_txg = 0; scn->scn_last_checkpoint = 0; scn->scn_checkpointing = B_FALSE; spa_scan_stat_init(spa); if (DSL_SCAN_IS_SCRUB_RESILVER(scn)) { scn->scn_phys.scn_ddt_class_max = zfs_scrub_ddt_class_max; /* rewrite all disk labels */ vdev_config_dirty(spa->spa_root_vdev); if (vdev_resilver_needed(spa->spa_root_vdev, &scn->scn_phys.scn_min_txg, &scn->scn_phys.scn_max_txg)) { spa_event_notify(spa, NULL, NULL, ESC_ZFS_RESILVER_START); } else { spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_START); } spa->spa_scrub_started = B_TRUE; /* * If this is an incremental scrub, limit the DDT scrub phase * to just the auto-ditto class (for correctness); the rest * of the scrub should go faster using top-down pruning. */ if (scn->scn_phys.scn_min_txg > TXG_INITIAL) scn->scn_phys.scn_ddt_class_max = DDT_CLASS_DITTO; } /* back to the generic stuff */ if (dp->dp_blkstats == NULL) { dp->dp_blkstats = kmem_alloc(sizeof (zfs_all_blkstats_t), KM_SLEEP); mutex_init(&dp->dp_blkstats->zab_lock, NULL, MUTEX_DEFAULT, NULL); } bzero(&dp->dp_blkstats->zab_type, sizeof (dp->dp_blkstats->zab_type)); if (spa_version(spa) < SPA_VERSION_DSL_SCRUB) ot = DMU_OT_ZAP_OTHER; scn->scn_phys.scn_queue_obj = zap_create(dp->dp_meta_objset, ot ? ot : DMU_OT_SCAN_QUEUE, DMU_OT_NONE, 0, tx); bcopy(&scn->scn_phys, &scn->scn_phys_cached, sizeof (scn->scn_phys)); dsl_scan_sync_state(scn, tx, SYNC_MANDATORY); spa_history_log_internal(spa, "scan setup", tx, "func=%u mintxg=%llu maxtxg=%llu", *funcp, scn->scn_phys.scn_min_txg, scn->scn_phys.scn_max_txg); } /* * Called by the ZFS_IOC_POOL_SCAN ioctl to start a scrub or resilver. * Can also be called to resume a paused scrub. */ int dsl_scan(dsl_pool_t *dp, pool_scan_func_t func) { spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; /* * Purge all vdev caches and probe all devices. We do this here * rather than in sync context because this requires a writer lock * on the spa_config lock, which we can't do from sync context. The * spa_scrub_reopen flag indicates that vdev_open() should not * attempt to start another scrub. */ spa_vdev_state_enter(spa, SCL_NONE); spa->spa_scrub_reopen = B_TRUE; vdev_reopen(spa->spa_root_vdev); spa->spa_scrub_reopen = B_FALSE; (void) spa_vdev_state_exit(spa, NULL, 0); + if (func == POOL_SCAN_RESILVER) { + dsl_resilver_restart(spa->spa_dsl_pool, 0); + return (0); + } + if (func == POOL_SCAN_SCRUB && dsl_scan_is_paused_scrub(scn)) { /* got scrub start cmd, resume paused scrub */ int err = dsl_scrub_set_pause_resume(scn->scn_dp, POOL_SCRUB_NORMAL); if (err == 0) { spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_RESUME); return (ECANCELED); } return (SET_ERROR(err)); } return (dsl_sync_task(spa_name(spa), dsl_scan_setup_check, dsl_scan_setup_sync, &func, 0, ZFS_SPACE_CHECK_EXTRA_RESERVED)); } +/* + * Sets the resilver defer flag to B_FALSE 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. + */ +static boolean_t +dsl_scan_clear_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++) { + resilver_needed |= + dsl_scan_clear_deferred(vd->vdev_child[c], 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); + + if (vd->vdev_resilver_deferred) + vd->vdev_resilver_deferred = B_FALSE; + + return (!vdev_is_dead(vd) && !vd->vdev_offline && + vdev_resilver_needed(vd, NULL, NULL)); +} + /* ARGSUSED */ static void dsl_scan_done(dsl_scan_t *scn, boolean_t complete, dmu_tx_t *tx) { static const char *old_names[] = { "scrub_bookmark", "scrub_ddt_bookmark", "scrub_ddt_class_max", "scrub_queue", "scrub_min_txg", "scrub_max_txg", "scrub_func", "scrub_errors", NULL }; dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; int i; /* Remove any remnants of an old-style scrub. */ for (i = 0; old_names[i]; i++) { (void) zap_remove(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, old_names[i], tx); } if (scn->scn_phys.scn_queue_obj != 0) { VERIFY0(dmu_object_free(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, tx)); scn->scn_phys.scn_queue_obj = 0; } scan_ds_queue_clear(scn); scn->scn_phys.scn_flags &= ~DSF_SCRUB_PAUSED; /* * If we were "restarted" from a stopped state, don't bother * with anything else. */ if (!dsl_scan_is_running(scn)) { ASSERT(!scn->scn_is_sorted); return; } if (scn->scn_is_sorted) { scan_io_queues_destroy(scn); scn->scn_is_sorted = B_FALSE; if (scn->scn_taskq != NULL) { taskq_destroy(scn->scn_taskq); scn->scn_taskq = NULL; } } scn->scn_phys.scn_state = complete ? DSS_FINISHED : DSS_CANCELED; if (dsl_scan_restarting(scn, tx)) spa_history_log_internal(spa, "scan aborted, restarting", tx, "errors=%llu", spa_get_errlog_size(spa)); else if (!complete) spa_history_log_internal(spa, "scan cancelled", tx, "errors=%llu", spa_get_errlog_size(spa)); else spa_history_log_internal(spa, "scan done", tx, "errors=%llu", spa_get_errlog_size(spa)); if (DSL_SCAN_IS_SCRUB_RESILVER(scn)) { spa->spa_scrub_started = B_FALSE; spa->spa_scrub_active = B_FALSE; /* * If the scrub/resilver completed, update all DTLs to * reflect this. Whether it succeeded or not, vacate * all temporary scrub DTLs. * * As the scrub does not currently support traversing * data that have been freed but are part of a checkpoint, * we don't mark the scrub as done in the DTLs as faults * may still exist in those vdevs. */ if (complete && !spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) { vdev_dtl_reassess(spa->spa_root_vdev, tx->tx_txg, scn->scn_phys.scn_max_txg, B_TRUE); spa_event_notify(spa, NULL, NULL, scn->scn_phys.scn_min_txg ? ESC_ZFS_RESILVER_FINISH : ESC_ZFS_SCRUB_FINISH); } else { vdev_dtl_reassess(spa->spa_root_vdev, tx->tx_txg, 0, B_TRUE); } spa_errlog_rotate(spa); /* * We may have finished replacing a device. * Let the async thread assess this and handle the detach. */ spa_async_request(spa, SPA_ASYNC_RESILVER_DONE); + + /* + * Clear any deferred_resilver flags in the config. + * If there are drives that need resilvering, kick + * off an asynchronous request to start resilver. + * dsl_scan_clear_deferred() may update the config + * before the resilver can restart. In the event of + * a crash during this period, the spa loading code + * will find the drives that need to be resilvered + * when the machine reboots and start the resilver then. + */ + boolean_t resilver_needed = + dsl_scan_clear_deferred(spa->spa_root_vdev, tx); + if (resilver_needed) { + spa_history_log_internal(spa, + "starting deferred resilver", tx, + "errors=%llu", spa_get_errlog_size(spa)); + spa_async_request(spa, SPA_ASYNC_RESILVER); + } } scn->scn_phys.scn_end_time = gethrestime_sec(); ASSERT(!dsl_scan_is_running(scn)); } /* ARGSUSED */ static int dsl_scan_cancel_check(void *arg, dmu_tx_t *tx) { dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; if (!dsl_scan_is_running(scn)) return (SET_ERROR(ENOENT)); return (0); } /* ARGSUSED */ static void dsl_scan_cancel_sync(void *arg, dmu_tx_t *tx) { dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; dsl_scan_done(scn, B_FALSE, tx); dsl_scan_sync_state(scn, tx, SYNC_MANDATORY); spa_event_notify(scn->scn_dp->dp_spa, NULL, NULL, ESC_ZFS_SCRUB_ABORT); } int dsl_scan_cancel(dsl_pool_t *dp) { return (dsl_sync_task(spa_name(dp->dp_spa), dsl_scan_cancel_check, dsl_scan_cancel_sync, NULL, 3, ZFS_SPACE_CHECK_RESERVED)); } static int dsl_scrub_pause_resume_check(void *arg, dmu_tx_t *tx) { pool_scrub_cmd_t *cmd = arg; dsl_pool_t *dp = dmu_tx_pool(tx); dsl_scan_t *scn = dp->dp_scan; if (*cmd == POOL_SCRUB_PAUSE) { /* can't pause a scrub when there is no in-progress scrub */ if (!dsl_scan_scrubbing(dp)) return (SET_ERROR(ENOENT)); /* can't pause a paused scrub */ if (dsl_scan_is_paused_scrub(scn)) return (SET_ERROR(EBUSY)); } else if (*cmd != POOL_SCRUB_NORMAL) { return (SET_ERROR(ENOTSUP)); } return (0); } static void dsl_scrub_pause_resume_sync(void *arg, dmu_tx_t *tx) { pool_scrub_cmd_t *cmd = arg; dsl_pool_t *dp = dmu_tx_pool(tx); spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; if (*cmd == POOL_SCRUB_PAUSE) { /* can't pause a scrub when there is no in-progress scrub */ spa->spa_scan_pass_scrub_pause = gethrestime_sec(); scn->scn_phys.scn_flags |= DSF_SCRUB_PAUSED; + scn->scn_phys_cached.scn_flags |= DSF_SCRUB_PAUSED; dsl_scan_sync_state(scn, tx, SYNC_CACHED); spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_PAUSED); } else { ASSERT3U(*cmd, ==, POOL_SCRUB_NORMAL); if (dsl_scan_is_paused_scrub(scn)) { /* * We need to keep track of how much time we spend * paused per pass so that we can adjust the scrub rate * shown in the output of 'zpool status' */ spa->spa_scan_pass_scrub_spent_paused += gethrestime_sec() - spa->spa_scan_pass_scrub_pause; spa->spa_scan_pass_scrub_pause = 0; scn->scn_phys.scn_flags &= ~DSF_SCRUB_PAUSED; + scn->scn_phys_cached.scn_flags &= ~DSF_SCRUB_PAUSED; dsl_scan_sync_state(scn, tx, SYNC_CACHED); } } } /* * Set scrub pause/resume state if it makes sense to do so */ int dsl_scrub_set_pause_resume(const dsl_pool_t *dp, pool_scrub_cmd_t cmd) { return (dsl_sync_task(spa_name(dp->dp_spa), dsl_scrub_pause_resume_check, dsl_scrub_pause_resume_sync, &cmd, 3, ZFS_SPACE_CHECK_RESERVED)); } /* start a new scan, or restart an existing one. */ void dsl_resilver_restart(dsl_pool_t *dp, uint64_t txg) { if (txg == 0) { dmu_tx_t *tx; tx = dmu_tx_create_dd(dp->dp_mos_dir); VERIFY(0 == dmu_tx_assign(tx, TXG_WAIT)); txg = dmu_tx_get_txg(tx); dp->dp_scan->scn_restart_txg = txg; dmu_tx_commit(tx); } else { dp->dp_scan->scn_restart_txg = txg; } zfs_dbgmsg("restarting resilver txg=%llu", txg); } void dsl_free(dsl_pool_t *dp, uint64_t txg, const blkptr_t *bp) { zio_free(dp->dp_spa, txg, bp); } void dsl_free_sync(zio_t *pio, dsl_pool_t *dp, uint64_t txg, const blkptr_t *bpp) { ASSERT(dsl_pool_sync_context(dp)); zio_nowait(zio_free_sync(pio, dp->dp_spa, txg, bpp, pio->io_flags)); } static int scan_ds_queue_compare(const void *a, const void *b) { const scan_ds_t *sds_a = a, *sds_b = b; if (sds_a->sds_dsobj < sds_b->sds_dsobj) return (-1); if (sds_a->sds_dsobj == sds_b->sds_dsobj) return (0); return (1); } static void scan_ds_queue_clear(dsl_scan_t *scn) { void *cookie = NULL; scan_ds_t *sds; while ((sds = avl_destroy_nodes(&scn->scn_queue, &cookie)) != NULL) { kmem_free(sds, sizeof (*sds)); } } static boolean_t scan_ds_queue_contains(dsl_scan_t *scn, uint64_t dsobj, uint64_t *txg) { scan_ds_t srch, *sds; srch.sds_dsobj = dsobj; sds = avl_find(&scn->scn_queue, &srch, NULL); if (sds != NULL && txg != NULL) *txg = sds->sds_txg; return (sds != NULL); } static void scan_ds_queue_insert(dsl_scan_t *scn, uint64_t dsobj, uint64_t txg) { scan_ds_t *sds; avl_index_t where; sds = kmem_zalloc(sizeof (*sds), KM_SLEEP); sds->sds_dsobj = dsobj; sds->sds_txg = txg; VERIFY3P(avl_find(&scn->scn_queue, sds, &where), ==, NULL); avl_insert(&scn->scn_queue, sds, where); } static void scan_ds_queue_remove(dsl_scan_t *scn, uint64_t dsobj) { scan_ds_t srch, *sds; srch.sds_dsobj = dsobj; sds = avl_find(&scn->scn_queue, &srch, NULL); VERIFY(sds != NULL); avl_remove(&scn->scn_queue, sds); kmem_free(sds, sizeof (*sds)); } static void scan_ds_queue_sync(dsl_scan_t *scn, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; dmu_object_type_t ot = (spa_version(spa) >= SPA_VERSION_DSL_SCRUB) ? DMU_OT_SCAN_QUEUE : DMU_OT_ZAP_OTHER; ASSERT0(scn->scn_bytes_pending); ASSERT(scn->scn_phys.scn_queue_obj != 0); VERIFY0(dmu_object_free(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, tx)); scn->scn_phys.scn_queue_obj = zap_create(dp->dp_meta_objset, ot, DMU_OT_NONE, 0, tx); for (scan_ds_t *sds = avl_first(&scn->scn_queue); sds != NULL; sds = AVL_NEXT(&scn->scn_queue, sds)) { VERIFY0(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, sds->sds_dsobj, sds->sds_txg, tx)); } } /* * Computes the memory limit state that we're currently in. A sorted scan * needs quite a bit of memory to hold the sorting queue, so we need to * reasonably constrain the size so it doesn't impact overall system * performance. We compute two limits: * 1) Hard memory limit: if the amount of memory used by the sorting * queues on a pool gets above this value, we stop the metadata * scanning portion and start issuing the queued up and sorted * I/Os to reduce memory usage. * This limit is calculated as a fraction of physmem (by default 5%). * We constrain the lower bound of the hard limit to an absolute * minimum of zfs_scan_mem_lim_min (default: 16 MiB). We also constrain * the upper bound to 5% of the total pool size - no chance we'll * ever need that much memory, but just to keep the value in check. * 2) Soft memory limit: once we hit the hard memory limit, we start * issuing I/O to reduce queue memory usage, but we don't want to * completely empty out the queues, since we might be able to find I/Os * that will fill in the gaps of our non-sequential IOs at some point * in the future. So we stop the issuing of I/Os once the amount of * memory used drops below the soft limit (at which point we stop issuing * I/O and start scanning metadata again). * * This limit is calculated by subtracting a fraction of the hard * limit from the hard limit. By default this fraction is 5%, so * the soft limit is 95% of the hard limit. We cap the size of the * difference between the hard and soft limits at an absolute * maximum of zfs_scan_mem_lim_soft_max (default: 128 MiB) - this is * sufficient to not cause too frequent switching between the * metadata scan and I/O issue (even at 2k recordsize, 128 MiB's * worth of queues is about 1.2 GiB of on-pool data, so scanning * that should take at least a decent fraction of a second). */ static boolean_t dsl_scan_should_clear(dsl_scan_t *scn) { vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev; uint64_t mlim_hard, mlim_soft, mused; uint64_t alloc = metaslab_class_get_alloc(spa_normal_class( scn->scn_dp->dp_spa)); mlim_hard = MAX((physmem / zfs_scan_mem_lim_fact) * PAGESIZE, zfs_scan_mem_lim_min); mlim_hard = MIN(mlim_hard, alloc / 20); mlim_soft = mlim_hard - MIN(mlim_hard / zfs_scan_mem_lim_soft_fact, zfs_scan_mem_lim_soft_max); mused = 0; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *tvd = rvd->vdev_child[i]; dsl_scan_io_queue_t *queue; mutex_enter(&tvd->vdev_scan_io_queue_lock); queue = tvd->vdev_scan_io_queue; if (queue != NULL) { /* # extents in exts_by_size = # in exts_by_addr */ mused += avl_numnodes(&queue->q_exts_by_size) * sizeof (range_seg_t) + queue->q_sio_memused; } mutex_exit(&tvd->vdev_scan_io_queue_lock); } dprintf("current scan memory usage: %llu bytes\n", (longlong_t)mused); if (mused == 0) ASSERT0(scn->scn_bytes_pending); /* * If we are above our hard limit, we need to clear out memory. * If we are below our soft limit, we need to accumulate sequential IOs. * Otherwise, we should keep doing whatever we are currently doing. */ if (mused >= mlim_hard) return (B_TRUE); else if (mused < mlim_soft) return (B_FALSE); else return (scn->scn_clearing); } static boolean_t dsl_scan_check_suspend(dsl_scan_t *scn, const zbookmark_phys_t *zb) { /* we never skip user/group accounting objects */ if (zb && (int64_t)zb->zb_object < 0) return (B_FALSE); if (scn->scn_suspending) return (B_TRUE); /* we're already suspending */ if (!ZB_IS_ZERO(&scn->scn_phys.scn_bookmark)) return (B_FALSE); /* we're resuming */ /* We only know how to resume from level-0 blocks. */ if (zb && zb->zb_level != 0) return (B_FALSE); /* * We suspend if: * - we have scanned for at least the minimum time (default 1 sec * for scrub, 3 sec for resilver), and either we have sufficient * dirty data that we are starting to write more quickly * (default 30%), or someone is explicitly waiting for this txg * to complete. * or * - the spa is shutting down because this pool is being exported * or the machine is rebooting. * or * - the scan queue has reached its memory use limit */ hrtime_t curr_time_ns = gethrtime(); uint64_t scan_time_ns = curr_time_ns - scn->scn_sync_start_time; uint64_t sync_time_ns = curr_time_ns - scn->scn_dp->dp_spa->spa_sync_starttime; int dirty_pct = scn->scn_dp->dp_dirty_total * 100 / zfs_dirty_data_max; int mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; if ((NSEC2MSEC(scan_time_ns) > mintime && (dirty_pct >= zfs_vdev_async_write_active_min_dirty_percent || txg_sync_waiting(scn->scn_dp) || NSEC2SEC(sync_time_ns) >= zfs_txg_timeout)) || spa_shutting_down(scn->scn_dp->dp_spa) || (zfs_scan_strict_mem_lim && dsl_scan_should_clear(scn))) { if (zb) { dprintf("suspending at bookmark %llx/%llx/%llx/%llx\n", (longlong_t)zb->zb_objset, (longlong_t)zb->zb_object, (longlong_t)zb->zb_level, (longlong_t)zb->zb_blkid); scn->scn_phys.scn_bookmark = *zb; } else { dsl_scan_phys_t *scnp = &scn->scn_phys; dprintf("suspending at DDT bookmark " "%llx/%llx/%llx/%llx\n", (longlong_t)scnp->scn_ddt_bookmark.ddb_class, (longlong_t)scnp->scn_ddt_bookmark.ddb_type, (longlong_t)scnp->scn_ddt_bookmark.ddb_checksum, (longlong_t)scnp->scn_ddt_bookmark.ddb_cursor); } scn->scn_suspending = B_TRUE; return (B_TRUE); } return (B_FALSE); } typedef struct zil_scan_arg { dsl_pool_t *zsa_dp; zil_header_t *zsa_zh; } zil_scan_arg_t; /* ARGSUSED */ static int dsl_scan_zil_block(zilog_t *zilog, blkptr_t *bp, void *arg, uint64_t claim_txg) { zil_scan_arg_t *zsa = arg; dsl_pool_t *dp = zsa->zsa_dp; dsl_scan_t *scn = dp->dp_scan; zil_header_t *zh = zsa->zsa_zh; zbookmark_phys_t zb; if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) return (0); /* * One block ("stubby") can be allocated a long time ago; we * want to visit that one because it has been allocated * (on-disk) even if it hasn't been claimed (even though for * scrub there's nothing to do to it). */ if (claim_txg == 0 && bp->blk_birth >= spa_min_claim_txg(dp->dp_spa)) return (0); SET_BOOKMARK(&zb, zh->zh_log.blk_cksum.zc_word[ZIL_ZC_OBJSET], ZB_ZIL_OBJECT, ZB_ZIL_LEVEL, bp->blk_cksum.zc_word[ZIL_ZC_SEQ]); VERIFY(0 == scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb)); return (0); } /* ARGSUSED */ static int dsl_scan_zil_record(zilog_t *zilog, lr_t *lrc, void *arg, uint64_t claim_txg) { if (lrc->lrc_txtype == TX_WRITE) { zil_scan_arg_t *zsa = arg; dsl_pool_t *dp = zsa->zsa_dp; dsl_scan_t *scn = dp->dp_scan; zil_header_t *zh = zsa->zsa_zh; lr_write_t *lr = (lr_write_t *)lrc; blkptr_t *bp = &lr->lr_blkptr; zbookmark_phys_t zb; if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) return (0); /* * birth can be < claim_txg if this record's txg is * already txg sync'ed (but this log block contains * other records that are not synced) */ if (claim_txg == 0 || bp->blk_birth < claim_txg) return (0); SET_BOOKMARK(&zb, zh->zh_log.blk_cksum.zc_word[ZIL_ZC_OBJSET], lr->lr_foid, ZB_ZIL_LEVEL, lr->lr_offset / BP_GET_LSIZE(bp)); VERIFY(0 == scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb)); } return (0); } static void dsl_scan_zil(dsl_pool_t *dp, zil_header_t *zh) { uint64_t claim_txg = zh->zh_claim_txg; zil_scan_arg_t zsa = { dp, zh }; zilog_t *zilog; ASSERT(spa_writeable(dp->dp_spa)); /* * We only want to visit blocks that have been claimed * but not yet replayed. */ if (claim_txg == 0) return; zilog = zil_alloc(dp->dp_meta_objset, zh); (void) zil_parse(zilog, dsl_scan_zil_block, dsl_scan_zil_record, &zsa, claim_txg); zil_free(zilog); } /* * We compare scan_prefetch_issue_ctx_t's based on their bookmarks. The idea * here is to sort the AVL tree by the order each block will be needed. */ static int scan_prefetch_queue_compare(const void *a, const void *b) { const scan_prefetch_issue_ctx_t *spic_a = a, *spic_b = b; const scan_prefetch_ctx_t *spc_a = spic_a->spic_spc; const scan_prefetch_ctx_t *spc_b = spic_b->spic_spc; return (zbookmark_compare(spc_a->spc_datablkszsec, spc_a->spc_indblkshift, spc_b->spc_datablkszsec, spc_b->spc_indblkshift, &spic_a->spic_zb, &spic_b->spic_zb)); } static void scan_prefetch_ctx_rele(scan_prefetch_ctx_t *spc, void *tag) { if (zfs_refcount_remove(&spc->spc_refcnt, tag) == 0) { zfs_refcount_destroy(&spc->spc_refcnt); kmem_free(spc, sizeof (scan_prefetch_ctx_t)); } } static scan_prefetch_ctx_t * scan_prefetch_ctx_create(dsl_scan_t *scn, dnode_phys_t *dnp, void *tag) { scan_prefetch_ctx_t *spc; spc = kmem_alloc(sizeof (scan_prefetch_ctx_t), KM_SLEEP); zfs_refcount_create(&spc->spc_refcnt); zfs_refcount_add(&spc->spc_refcnt, tag); spc->spc_scn = scn; if (dnp != NULL) { spc->spc_datablkszsec = dnp->dn_datablkszsec; spc->spc_indblkshift = dnp->dn_indblkshift; spc->spc_root = B_FALSE; } else { spc->spc_datablkszsec = 0; spc->spc_indblkshift = 0; spc->spc_root = B_TRUE; } return (spc); } static void scan_prefetch_ctx_add_ref(scan_prefetch_ctx_t *spc, void *tag) { zfs_refcount_add(&spc->spc_refcnt, tag); } static boolean_t dsl_scan_check_prefetch_resume(scan_prefetch_ctx_t *spc, const zbookmark_phys_t *zb) { zbookmark_phys_t *last_zb = &spc->spc_scn->scn_prefetch_bookmark; dnode_phys_t tmp_dnp; dnode_phys_t *dnp = (spc->spc_root) ? NULL : &tmp_dnp; if (zb->zb_objset != last_zb->zb_objset) return (B_TRUE); if ((int64_t)zb->zb_object < 0) return (B_FALSE); tmp_dnp.dn_datablkszsec = spc->spc_datablkszsec; tmp_dnp.dn_indblkshift = spc->spc_indblkshift; if (zbookmark_subtree_completed(dnp, zb, last_zb)) return (B_TRUE); return (B_FALSE); } static void dsl_scan_prefetch(scan_prefetch_ctx_t *spc, blkptr_t *bp, zbookmark_phys_t *zb) { avl_index_t idx; dsl_scan_t *scn = spc->spc_scn; spa_t *spa = scn->scn_dp->dp_spa; scan_prefetch_issue_ctx_t *spic; if (zfs_no_scrub_prefetch) return; if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg || (BP_GET_LEVEL(bp) == 0 && BP_GET_TYPE(bp) != DMU_OT_DNODE && BP_GET_TYPE(bp) != DMU_OT_OBJSET)) return; if (dsl_scan_check_prefetch_resume(spc, zb)) return; scan_prefetch_ctx_add_ref(spc, scn); spic = kmem_alloc(sizeof (scan_prefetch_issue_ctx_t), KM_SLEEP); spic->spic_spc = spc; spic->spic_bp = *bp; spic->spic_zb = *zb; /* * Add the IO to the queue of blocks to prefetch. This allows us to * prioritize blocks that we will need first for the main traversal * thread. */ mutex_enter(&spa->spa_scrub_lock); if (avl_find(&scn->scn_prefetch_queue, spic, &idx) != NULL) { /* this block is already queued for prefetch */ kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); scan_prefetch_ctx_rele(spc, scn); mutex_exit(&spa->spa_scrub_lock); return; } avl_insert(&scn->scn_prefetch_queue, spic, idx); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); } static void dsl_scan_prefetch_dnode(dsl_scan_t *scn, dnode_phys_t *dnp, uint64_t objset, uint64_t object) { int i; zbookmark_phys_t zb; scan_prefetch_ctx_t *spc; if (dnp->dn_nblkptr == 0 && !(dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR)) return; SET_BOOKMARK(&zb, objset, object, 0, 0); spc = scan_prefetch_ctx_create(scn, dnp, FTAG); for (i = 0; i < dnp->dn_nblkptr; i++) { zb.zb_level = BP_GET_LEVEL(&dnp->dn_blkptr[i]); zb.zb_blkid = i; dsl_scan_prefetch(spc, &dnp->dn_blkptr[i], &zb); } if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { zb.zb_level = 0; zb.zb_blkid = DMU_SPILL_BLKID; dsl_scan_prefetch(spc, &dnp->dn_spill, &zb); } scan_prefetch_ctx_rele(spc, FTAG); } void dsl_scan_prefetch_cb(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp, arc_buf_t *buf, void *private) { scan_prefetch_ctx_t *spc = private; dsl_scan_t *scn = spc->spc_scn; spa_t *spa = scn->scn_dp->dp_spa; /* broadcast that the IO has completed for rate limitting purposes */ mutex_enter(&spa->spa_scrub_lock); ASSERT3U(spa->spa_scrub_inflight, >=, BP_GET_PSIZE(bp)); spa->spa_scrub_inflight -= BP_GET_PSIZE(bp); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); /* if there was an error or we are done prefetching, just cleanup */ if (buf == NULL || scn->scn_suspending) goto out; if (BP_GET_LEVEL(bp) > 0) { int i; blkptr_t *cbp; int epb = BP_GET_LSIZE(bp) >> SPA_BLKPTRSHIFT; zbookmark_phys_t czb; for (i = 0, cbp = buf->b_data; i < epb; i++, cbp++) { SET_BOOKMARK(&czb, zb->zb_objset, zb->zb_object, zb->zb_level - 1, zb->zb_blkid * epb + i); dsl_scan_prefetch(spc, cbp, &czb); } } else if (BP_GET_TYPE(bp) == DMU_OT_DNODE) { dnode_phys_t *cdnp = buf->b_data; int i; int epb = BP_GET_LSIZE(bp) >> DNODE_SHIFT; for (i = 0, cdnp = buf->b_data; i < epb; i += cdnp->dn_extra_slots + 1, cdnp += cdnp->dn_extra_slots + 1) { dsl_scan_prefetch_dnode(scn, cdnp, zb->zb_objset, zb->zb_blkid * epb + i); } } else if (BP_GET_TYPE(bp) == DMU_OT_OBJSET) { objset_phys_t *osp = buf->b_data; dsl_scan_prefetch_dnode(scn, &osp->os_meta_dnode, zb->zb_objset, DMU_META_DNODE_OBJECT); if (OBJSET_BUF_HAS_USERUSED(buf)) { dsl_scan_prefetch_dnode(scn, &osp->os_groupused_dnode, zb->zb_objset, DMU_GROUPUSED_OBJECT); dsl_scan_prefetch_dnode(scn, &osp->os_userused_dnode, zb->zb_objset, DMU_USERUSED_OBJECT); } } out: if (buf != NULL) arc_buf_destroy(buf, private); scan_prefetch_ctx_rele(spc, scn); } /* ARGSUSED */ static void dsl_scan_prefetch_thread(void *arg) { dsl_scan_t *scn = arg; spa_t *spa = scn->scn_dp->dp_spa; vdev_t *rvd = spa->spa_root_vdev; uint64_t maxinflight = rvd->vdev_children * zfs_top_maxinflight; scan_prefetch_issue_ctx_t *spic; /* loop until we are told to stop */ while (!scn->scn_prefetch_stop) { arc_flags_t flags = ARC_FLAG_NOWAIT | ARC_FLAG_PRESCIENT_PREFETCH | ARC_FLAG_PREFETCH; int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCAN_THREAD; mutex_enter(&spa->spa_scrub_lock); /* * Wait until we have an IO to issue and are not above our * maximum in flight limit. */ while (!scn->scn_prefetch_stop && (avl_numnodes(&scn->scn_prefetch_queue) == 0 || spa->spa_scrub_inflight >= scn->scn_maxinflight_bytes)) { cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock); } /* recheck if we should stop since we waited for the cv */ if (scn->scn_prefetch_stop) { mutex_exit(&spa->spa_scrub_lock); break; } /* remove the prefetch IO from the tree */ spic = avl_first(&scn->scn_prefetch_queue); spa->spa_scrub_inflight += BP_GET_PSIZE(&spic->spic_bp); avl_remove(&scn->scn_prefetch_queue, spic); mutex_exit(&spa->spa_scrub_lock); /* issue the prefetch asynchronously */ (void) arc_read(scn->scn_zio_root, scn->scn_dp->dp_spa, &spic->spic_bp, dsl_scan_prefetch_cb, spic->spic_spc, ZIO_PRIORITY_SCRUB, zio_flags, &flags, &spic->spic_zb); kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); } ASSERT(scn->scn_prefetch_stop); /* free any prefetches we didn't get to complete */ mutex_enter(&spa->spa_scrub_lock); while ((spic = avl_first(&scn->scn_prefetch_queue)) != NULL) { avl_remove(&scn->scn_prefetch_queue, spic); scan_prefetch_ctx_rele(spic->spic_spc, scn); kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); } ASSERT0(avl_numnodes(&scn->scn_prefetch_queue)); mutex_exit(&spa->spa_scrub_lock); } static boolean_t dsl_scan_check_resume(dsl_scan_t *scn, const dnode_phys_t *dnp, const zbookmark_phys_t *zb) { /* * We never skip over user/group accounting objects (obj<0) */ if (!ZB_IS_ZERO(&scn->scn_phys.scn_bookmark) && (int64_t)zb->zb_object >= 0) { /* * If we already visited this bp & everything below (in * a prior txg sync), don't bother doing it again. */ if (zbookmark_subtree_completed(dnp, zb, &scn->scn_phys.scn_bookmark)) return (B_TRUE); /* * If we found the block we're trying to resume from, or * we went past it to a different object, zero it out to * indicate that it's OK to start checking for suspending * again. */ if (bcmp(zb, &scn->scn_phys.scn_bookmark, sizeof (*zb)) == 0 || zb->zb_object > scn->scn_phys.scn_bookmark.zb_object) { dprintf("resuming at %llx/%llx/%llx/%llx\n", (longlong_t)zb->zb_objset, (longlong_t)zb->zb_object, (longlong_t)zb->zb_level, (longlong_t)zb->zb_blkid); bzero(&scn->scn_phys.scn_bookmark, sizeof (*zb)); } } return (B_FALSE); } static void dsl_scan_visitbp(blkptr_t *bp, const zbookmark_phys_t *zb, dnode_phys_t *dnp, dsl_dataset_t *ds, dsl_scan_t *scn, dmu_objset_type_t ostype, dmu_tx_t *tx); static void dsl_scan_visitdnode( dsl_scan_t *, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, uint64_t object, dmu_tx_t *tx); /* * Return nonzero on i/o error. * Return new buf to write out in *bufp. */ static int dsl_scan_recurse(dsl_scan_t *scn, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, const blkptr_t *bp, const zbookmark_phys_t *zb, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCAN_THREAD; int err; if (BP_GET_LEVEL(bp) > 0) { arc_flags_t flags = ARC_FLAG_WAIT; int i; blkptr_t *cbp; int epb = BP_GET_LSIZE(bp) >> SPA_BLKPTRSHIFT; arc_buf_t *buf; err = arc_read(NULL, dp->dp_spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } for (i = 0, cbp = buf->b_data; 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); dsl_scan_visitbp(cbp, &czb, dnp, ds, scn, ostype, tx); } arc_buf_destroy(buf, &buf); } else if (BP_GET_TYPE(bp) == DMU_OT_DNODE) { arc_flags_t flags = ARC_FLAG_WAIT; dnode_phys_t *cdnp; int i; int epb = BP_GET_LSIZE(bp) >> DNODE_SHIFT; arc_buf_t *buf; err = arc_read(NULL, dp->dp_spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } for (i = 0, cdnp = buf->b_data; i < epb; i += cdnp->dn_extra_slots + 1, cdnp += cdnp->dn_extra_slots + 1) { dsl_scan_visitdnode(scn, ds, ostype, cdnp, zb->zb_blkid * epb + i, tx); } arc_buf_destroy(buf, &buf); } else if (BP_GET_TYPE(bp) == DMU_OT_OBJSET) { arc_flags_t flags = ARC_FLAG_WAIT; objset_phys_t *osp; arc_buf_t *buf; err = arc_read(NULL, dp->dp_spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } osp = buf->b_data; dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_meta_dnode, DMU_META_DNODE_OBJECT, tx); if (OBJSET_BUF_HAS_USERUSED(buf)) { /* * We also always visit user/group accounting * objects, and never skip them, even if we are * suspending. This is necessary so that the space * deltas from this txg get integrated. */ dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_groupused_dnode, DMU_GROUPUSED_OBJECT, tx); dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_userused_dnode, DMU_USERUSED_OBJECT, tx); } arc_buf_destroy(buf, &buf); } return (0); } static void dsl_scan_visitdnode(dsl_scan_t *scn, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, uint64_t object, dmu_tx_t *tx) { int j; for (j = 0; j < dnp->dn_nblkptr; j++) { zbookmark_phys_t czb; SET_BOOKMARK(&czb, ds ? ds->ds_object : 0, object, dnp->dn_nlevels - 1, j); dsl_scan_visitbp(&dnp->dn_blkptr[j], &czb, dnp, ds, scn, ostype, tx); } if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { zbookmark_phys_t czb; SET_BOOKMARK(&czb, ds ? ds->ds_object : 0, object, 0, DMU_SPILL_BLKID); dsl_scan_visitbp(DN_SPILL_BLKPTR(dnp), &czb, dnp, ds, scn, ostype, tx); } } /* * The arguments are in this order because mdb can only print the * first 5; we want them to be useful. */ static void dsl_scan_visitbp(blkptr_t *bp, const zbookmark_phys_t *zb, dnode_phys_t *dnp, dsl_dataset_t *ds, dsl_scan_t *scn, dmu_objset_type_t ostype, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; blkptr_t *bp_toread = NULL; if (dsl_scan_check_suspend(scn, zb)) return; if (dsl_scan_check_resume(scn, dnp, zb)) return; scn->scn_visited_this_txg++; /* * This debugging is commented out to conserve stack space. This * function is called recursively and the debugging addes several * bytes to the stack for each call. It can be commented back in * if required to debug an issue in dsl_scan_visitbp(). * * dprintf_bp(bp, * "visiting ds=%p/%llu zb=%llx/%llx/%llx/%llx bp=%p", * ds, ds ? ds->ds_object : 0, * zb->zb_objset, zb->zb_object, zb->zb_level, zb->zb_blkid, * bp); */ if (BP_IS_HOLE(bp)) { scn->scn_holes_this_txg++; return; } if (bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) { scn->scn_lt_min_this_txg++; return; } bp_toread = kmem_alloc(sizeof (blkptr_t), KM_SLEEP); *bp_toread = *bp; if (dsl_scan_recurse(scn, ds, ostype, dnp, bp_toread, zb, tx) != 0) goto out; /* * If dsl_scan_ddt() has already visited this block, it will have * already done any translations or scrubbing, so don't call the * callback again. */ if (ddt_class_contains(dp->dp_spa, scn->scn_phys.scn_ddt_class_max, bp)) { scn->scn_ddt_contained_this_txg++; goto out; } /* * If this block is from the future (after cur_max_txg), then we * are doing this on behalf of a deleted snapshot, and we will * revisit the future block on the next pass of this dataset. * Don't scan it now unless we need to because something * under it was modified. */ if (BP_PHYSICAL_BIRTH(bp) > scn->scn_phys.scn_cur_max_txg) { scn->scn_gt_max_this_txg++; goto out; } scan_funcs[scn->scn_phys.scn_func](dp, bp, zb); out: kmem_free(bp_toread, sizeof (blkptr_t)); } static void dsl_scan_visit_rootbp(dsl_scan_t *scn, dsl_dataset_t *ds, blkptr_t *bp, dmu_tx_t *tx) { zbookmark_phys_t zb; scan_prefetch_ctx_t *spc; SET_BOOKMARK(&zb, ds ? ds->ds_object : DMU_META_OBJSET, ZB_ROOT_OBJECT, ZB_ROOT_LEVEL, ZB_ROOT_BLKID); if (ZB_IS_ZERO(&scn->scn_phys.scn_bookmark)) { SET_BOOKMARK(&scn->scn_prefetch_bookmark, zb.zb_objset, 0, 0, 0); } else { scn->scn_prefetch_bookmark = scn->scn_phys.scn_bookmark; } scn->scn_objsets_visited_this_txg++; spc = scan_prefetch_ctx_create(scn, NULL, FTAG); dsl_scan_prefetch(spc, bp, &zb); scan_prefetch_ctx_rele(spc, FTAG); dsl_scan_visitbp(bp, &zb, NULL, ds, scn, DMU_OST_NONE, tx); dprintf_ds(ds, "finished scan%s", ""); } static void ds_destroyed_scn_phys(dsl_dataset_t *ds, dsl_scan_phys_t *scn_phys) { if (scn_phys->scn_bookmark.zb_objset == ds->ds_object) { if (ds->ds_is_snapshot) { /* * Note: * - scn_cur_{min,max}_txg stays the same. * - Setting the flag is not really necessary if * scn_cur_max_txg == scn_max_txg, because there * is nothing after this snapshot that we care * about. However, we set it anyway and then * ignore it when we retraverse it in * dsl_scan_visitds(). */ scn_phys->scn_bookmark.zb_objset = dsl_dataset_phys(ds)->ds_next_snap_obj; zfs_dbgmsg("destroying ds %llu; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds->ds_object, (u_longlong_t)dsl_dataset_phys(ds)-> ds_next_snap_obj); scn_phys->scn_flags |= DSF_VISIT_DS_AGAIN; } else { SET_BOOKMARK(&scn_phys->scn_bookmark, ZB_DESTROYED_OBJSET, 0, 0, 0); zfs_dbgmsg("destroying ds %llu; currently traversing; " "reset bookmark to -1,0,0,0", (u_longlong_t)ds->ds_object); } } } /* * Invoked when a dataset is destroyed. We need to make sure that: * * 1) If it is the dataset that was currently being scanned, we write * a new dsl_scan_phys_t and marking the objset reference in it * as destroyed. * 2) Remove it from the work queue, if it was present. * * If the dataset was actually a snapshot, instead of marking the dataset * as destroyed, we instead substitute the next snapshot in line. */ void dsl_scan_ds_destroyed(dsl_dataset_t *ds, dmu_tx_t *tx) { dsl_pool_t *dp = ds->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg; if (!dsl_scan_is_running(scn)) return; ds_destroyed_scn_phys(ds, &scn->scn_phys); ds_destroyed_scn_phys(ds, &scn->scn_phys_cached); if (scan_ds_queue_contains(scn, ds->ds_object, &mintxg)) { scan_ds_queue_remove(scn, ds->ds_object); if (ds->ds_is_snapshot) scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_next_snap_obj, mintxg); } if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, &mintxg) == 0) { ASSERT3U(dsl_dataset_phys(ds)->ds_num_children, <=, 1); VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, tx)); if (ds->ds_is_snapshot) { /* * We keep the same mintxg; it could be > * ds_creation_txg if the previous snapshot was * deleted too. */ VERIFY(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, dsl_dataset_phys(ds)->ds_next_snap_obj, mintxg, tx) == 0); zfs_dbgmsg("destroying ds %llu; in queue; " "replacing with %llu", (u_longlong_t)ds->ds_object, (u_longlong_t)dsl_dataset_phys(ds)-> ds_next_snap_obj); } else { zfs_dbgmsg("destroying ds %llu; in queue; removing", (u_longlong_t)ds->ds_object); } } /* * dsl_scan_sync() should be called after this, and should sync * out our changed state, but just to be safe, do it here. */ dsl_scan_sync_state(scn, tx, SYNC_CACHED); } static void ds_snapshotted_bookmark(dsl_dataset_t *ds, zbookmark_phys_t *scn_bookmark) { if (scn_bookmark->zb_objset == ds->ds_object) { scn_bookmark->zb_objset = dsl_dataset_phys(ds)->ds_prev_snap_obj; zfs_dbgmsg("snapshotting ds %llu; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds->ds_object, (u_longlong_t)dsl_dataset_phys(ds)->ds_prev_snap_obj); } } /* * Called when a dataset is snapshotted. If we were currently traversing * this snapshot, we reset our bookmark to point at the newly created * snapshot. We also modify our work queue to remove the old snapshot and * replace with the new one. */ void dsl_scan_ds_snapshotted(dsl_dataset_t *ds, dmu_tx_t *tx) { dsl_pool_t *dp = ds->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg; if (!dsl_scan_is_running(scn)) return; ASSERT(dsl_dataset_phys(ds)->ds_prev_snap_obj != 0); ds_snapshotted_bookmark(ds, &scn->scn_phys.scn_bookmark); ds_snapshotted_bookmark(ds, &scn->scn_phys_cached.scn_bookmark); if (scan_ds_queue_contains(scn, ds->ds_object, &mintxg)) { scan_ds_queue_remove(scn, ds->ds_object); scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_prev_snap_obj, mintxg); } if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, &mintxg) == 0) { VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, tx)); VERIFY(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, dsl_dataset_phys(ds)->ds_prev_snap_obj, mintxg, tx) == 0); zfs_dbgmsg("snapshotting ds %llu; in queue; " "replacing with %llu", (u_longlong_t)ds->ds_object, (u_longlong_t)dsl_dataset_phys(ds)->ds_prev_snap_obj); } dsl_scan_sync_state(scn, tx, SYNC_CACHED); } static void ds_clone_swapped_bookmark(dsl_dataset_t *ds1, dsl_dataset_t *ds2, zbookmark_phys_t *scn_bookmark) { if (scn_bookmark->zb_objset == ds1->ds_object) { scn_bookmark->zb_objset = ds2->ds_object; zfs_dbgmsg("clone_swap ds %llu; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds1->ds_object, (u_longlong_t)ds2->ds_object); } else if (scn_bookmark->zb_objset == ds2->ds_object) { scn_bookmark->zb_objset = ds1->ds_object; zfs_dbgmsg("clone_swap ds %llu; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds2->ds_object, (u_longlong_t)ds1->ds_object); } } /* * Called when a parent dataset and its clone are swapped. If we were * currently traversing the dataset, we need to switch to traversing the * newly promoted parent. */ void dsl_scan_ds_clone_swapped(dsl_dataset_t *ds1, dsl_dataset_t *ds2, dmu_tx_t *tx) { dsl_pool_t *dp = ds1->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg; if (!dsl_scan_is_running(scn)) return; ds_clone_swapped_bookmark(ds1, ds2, &scn->scn_phys.scn_bookmark); ds_clone_swapped_bookmark(ds1, ds2, &scn->scn_phys_cached.scn_bookmark); if (scan_ds_queue_contains(scn, ds1->ds_object, &mintxg)) { scan_ds_queue_remove(scn, ds1->ds_object); scan_ds_queue_insert(scn, ds2->ds_object, mintxg); } if (scan_ds_queue_contains(scn, ds2->ds_object, &mintxg)) { scan_ds_queue_remove(scn, ds2->ds_object); scan_ds_queue_insert(scn, ds1->ds_object, mintxg); } if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, &mintxg) == 0) { int err; ASSERT3U(mintxg, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, tx)); err = zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, mintxg, tx); VERIFY(err == 0 || err == EEXIST); if (err == EEXIST) { /* Both were there to begin with */ VERIFY(0 == zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, mintxg, tx)); } zfs_dbgmsg("clone_swap ds %llu; in queue; " "replacing with %llu", (u_longlong_t)ds1->ds_object, (u_longlong_t)ds2->ds_object); } if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, &mintxg) == 0) { ASSERT3U(mintxg, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, tx)); VERIFY(0 == zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, mintxg, tx)); zfs_dbgmsg("clone_swap ds %llu; in queue; " "replacing with %llu", (u_longlong_t)ds2->ds_object, (u_longlong_t)ds1->ds_object); } dsl_scan_sync_state(scn, tx, SYNC_CACHED); } /* ARGSUSED */ static int enqueue_clones_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg) { uint64_t originobj = *(uint64_t *)arg; dsl_dataset_t *ds; int err; dsl_scan_t *scn = dp->dp_scan; if (dsl_dir_phys(hds->ds_dir)->dd_origin_obj != originobj) return (0); err = dsl_dataset_hold_obj(dp, hds->ds_object, FTAG, &ds); if (err) return (err); while (dsl_dataset_phys(ds)->ds_prev_snap_obj != originobj) { dsl_dataset_t *prev; err = dsl_dataset_hold_obj(dp, dsl_dataset_phys(ds)->ds_prev_snap_obj, FTAG, &prev); dsl_dataset_rele(ds, FTAG); if (err) return (err); ds = prev; } scan_ds_queue_insert(scn, ds->ds_object, dsl_dataset_phys(ds)->ds_prev_snap_txg); dsl_dataset_rele(ds, FTAG); return (0); } static void dsl_scan_visitds(dsl_scan_t *scn, uint64_t dsobj, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; dsl_dataset_t *ds; VERIFY3U(0, ==, dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds)); if (scn->scn_phys.scn_cur_min_txg >= scn->scn_phys.scn_max_txg) { /* * This can happen if this snapshot was created after the * scan started, and we already completed a previous snapshot * that was created after the scan started. This snapshot * only references blocks with: * * birth < our ds_creation_txg * cur_min_txg is no less than ds_creation_txg. * We have already visited these blocks. * or * birth > scn_max_txg * The scan requested not to visit these blocks. * * Subsequent snapshots (and clones) can reference our * blocks, or blocks with even higher birth times. * Therefore we do not need to visit them either, * so we do not add them to the work queue. * * Note that checking for cur_min_txg >= cur_max_txg * is not sufficient, because in that case we may need to * visit subsequent snapshots. This happens when min_txg > 0, * which raises cur_min_txg. In this case we will visit * this dataset but skip all of its blocks, because the * rootbp's birth time is < cur_min_txg. Then we will * add the next snapshots/clones to the work queue. */ char *dsname = kmem_alloc(MAXNAMELEN, KM_SLEEP); dsl_dataset_name(ds, dsname); zfs_dbgmsg("scanning dataset %llu (%s) is unnecessary because " "cur_min_txg (%llu) >= max_txg (%llu)", (longlong_t)dsobj, dsname, (longlong_t)scn->scn_phys.scn_cur_min_txg, (longlong_t)scn->scn_phys.scn_max_txg); kmem_free(dsname, MAXNAMELEN); goto out; } /* * Only the ZIL in the head (non-snapshot) is valid. Even though * snapshots can have ZIL block pointers (which may be the same * BP as in the head), they must be ignored. In addition, $ORIGIN * doesn't have a objset (i.e. its ds_bp is a hole) so we don't * need to look for a ZIL in it either. So we traverse the ZIL here, * rather than in scan_recurse(), because the regular snapshot * block-sharing rules don't apply to it. */ if (DSL_SCAN_IS_SCRUB_RESILVER(scn) && !dsl_dataset_is_snapshot(ds) && (dp->dp_origin_snap == NULL || ds->ds_dir != dp->dp_origin_snap->ds_dir)) { objset_t *os; if (dmu_objset_from_ds(ds, &os) != 0) { goto out; } dsl_scan_zil(dp, &os->os_zil_header); } /* * Iterate over the bps in this ds. */ dmu_buf_will_dirty(ds->ds_dbuf, tx); rrw_enter(&ds->ds_bp_rwlock, RW_READER, FTAG); dsl_scan_visit_rootbp(scn, ds, &dsl_dataset_phys(ds)->ds_bp, tx); rrw_exit(&ds->ds_bp_rwlock, FTAG); char *dsname = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP); dsl_dataset_name(ds, dsname); zfs_dbgmsg("scanned dataset %llu (%s) with min=%llu max=%llu; " "suspending=%u", (longlong_t)dsobj, dsname, (longlong_t)scn->scn_phys.scn_cur_min_txg, (longlong_t)scn->scn_phys.scn_cur_max_txg, (int)scn->scn_suspending); kmem_free(dsname, ZFS_MAX_DATASET_NAME_LEN); if (scn->scn_suspending) goto out; /* * We've finished this pass over this dataset. */ /* * If we did not completely visit this dataset, do another pass. */ if (scn->scn_phys.scn_flags & DSF_VISIT_DS_AGAIN) { zfs_dbgmsg("incomplete pass; visiting again"); scn->scn_phys.scn_flags &= ~DSF_VISIT_DS_AGAIN; scan_ds_queue_insert(scn, ds->ds_object, scn->scn_phys.scn_cur_max_txg); goto out; } /* * Add descendent datasets to work queue. */ if (dsl_dataset_phys(ds)->ds_next_snap_obj != 0) { scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_next_snap_obj, dsl_dataset_phys(ds)->ds_creation_txg); } if (dsl_dataset_phys(ds)->ds_num_children > 1) { boolean_t usenext = B_FALSE; if (dsl_dataset_phys(ds)->ds_next_clones_obj != 0) { uint64_t count; /* * A bug in a previous version of the code could * cause upgrade_clones_cb() to not set * ds_next_snap_obj when it should, leading to a * missing entry. Therefore we can only use the * next_clones_obj when its count is correct. */ int err = zap_count(dp->dp_meta_objset, dsl_dataset_phys(ds)->ds_next_clones_obj, &count); if (err == 0 && count == dsl_dataset_phys(ds)->ds_num_children - 1) usenext = B_TRUE; } if (usenext) { zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, dp->dp_meta_objset, dsl_dataset_phys(ds)->ds_next_clones_obj); zap_cursor_retrieve(&zc, &za) == 0; (void) zap_cursor_advance(&zc)) { scan_ds_queue_insert(scn, zfs_strtonum(za.za_name, NULL), dsl_dataset_phys(ds)->ds_creation_txg); } zap_cursor_fini(&zc); } else { VERIFY0(dmu_objset_find_dp(dp, dp->dp_root_dir_obj, enqueue_clones_cb, &ds->ds_object, DS_FIND_CHILDREN)); } } out: dsl_dataset_rele(ds, FTAG); } /* ARGSUSED */ static int enqueue_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg) { dsl_dataset_t *ds; int err; dsl_scan_t *scn = dp->dp_scan; err = dsl_dataset_hold_obj(dp, hds->ds_object, FTAG, &ds); if (err) return (err); while (dsl_dataset_phys(ds)->ds_prev_snap_obj != 0) { dsl_dataset_t *prev; err = dsl_dataset_hold_obj(dp, dsl_dataset_phys(ds)->ds_prev_snap_obj, FTAG, &prev); if (err) { dsl_dataset_rele(ds, FTAG); return (err); } /* * If this is a clone, we don't need to worry about it for now. */ if (dsl_dataset_phys(prev)->ds_next_snap_obj != ds->ds_object) { dsl_dataset_rele(ds, FTAG); dsl_dataset_rele(prev, FTAG); return (0); } dsl_dataset_rele(ds, FTAG); ds = prev; } scan_ds_queue_insert(scn, ds->ds_object, dsl_dataset_phys(ds)->ds_prev_snap_txg); dsl_dataset_rele(ds, FTAG); return (0); } /* ARGSUSED */ void dsl_scan_ddt_entry(dsl_scan_t *scn, enum zio_checksum checksum, ddt_entry_t *dde, dmu_tx_t *tx) { const ddt_key_t *ddk = &dde->dde_key; ddt_phys_t *ddp = dde->dde_phys; blkptr_t bp; zbookmark_phys_t zb = { 0 }; int p; if (scn->scn_phys.scn_state != DSS_SCANNING) return; + /* + * This function is special because it is the only thing + * that can add scan_io_t's to the vdev scan queues from + * outside dsl_scan_sync(). For the most part this is ok + * as long as it is called from within syncing context. + * However, dsl_scan_sync() expects that no new sio's will + * be added between when all the work for a scan is done + * and the next txg when the scan is actually marked as + * completed. This check ensures we do not issue new sio's + * during this period. + */ + if (scn->scn_done_txg != 0) + return; + for (p = 0; p < DDT_PHYS_TYPES; p++, ddp++) { if (ddp->ddp_phys_birth == 0 || ddp->ddp_phys_birth > scn->scn_phys.scn_max_txg) continue; ddt_bp_create(checksum, ddk, ddp, &bp); scn->scn_visited_this_txg++; scan_funcs[scn->scn_phys.scn_func](scn->scn_dp, &bp, &zb); } } /* * Scrub/dedup interaction. * * If there are N references to a deduped block, we don't want to scrub it * N times -- ideally, we should scrub it exactly once. * * We leverage the fact that the dde's replication class (enum ddt_class) * is ordered from highest replication class (DDT_CLASS_DITTO) to lowest * (DDT_CLASS_UNIQUE) so that we may walk the DDT in that order. * * To prevent excess scrubbing, the scrub begins by walking the DDT * to find all blocks with refcnt > 1, and scrubs each of these once. * Since there are two replication classes which contain blocks with * refcnt > 1, we scrub the highest replication class (DDT_CLASS_DITTO) first. * Finally the top-down scrub begins, only visiting blocks with refcnt == 1. * * There would be nothing more to say if a block's refcnt couldn't change * during a scrub, but of course it can so we must account for changes * in a block's replication class. * * Here's an example of what can occur: * * If a block has refcnt > 1 during the DDT scrub phase, but has refcnt == 1 * when visited during the top-down scrub phase, it will be scrubbed twice. * This negates our scrub optimization, but is otherwise harmless. * * If a block has refcnt == 1 during the DDT scrub phase, but has refcnt > 1 * on each visit during the top-down scrub phase, it will never be scrubbed. * To catch this, ddt_sync_entry() notifies the scrub code whenever a block's * reference class transitions to a higher level (i.e DDT_CLASS_UNIQUE to * DDT_CLASS_DUPLICATE); if it transitions from refcnt == 1 to refcnt > 1 * while a scrub is in progress, it scrubs the block right then. */ static void dsl_scan_ddt(dsl_scan_t *scn, dmu_tx_t *tx) { ddt_bookmark_t *ddb = &scn->scn_phys.scn_ddt_bookmark; ddt_entry_t dde = { 0 }; int error; uint64_t n = 0; while ((error = ddt_walk(scn->scn_dp->dp_spa, ddb, &dde)) == 0) { ddt_t *ddt; if (ddb->ddb_class > scn->scn_phys.scn_ddt_class_max) break; dprintf("visiting ddb=%llu/%llu/%llu/%llx\n", (longlong_t)ddb->ddb_class, (longlong_t)ddb->ddb_type, (longlong_t)ddb->ddb_checksum, (longlong_t)ddb->ddb_cursor); /* There should be no pending changes to the dedup table */ ddt = scn->scn_dp->dp_spa->spa_ddt[ddb->ddb_checksum]; ASSERT(avl_first(&ddt->ddt_tree) == NULL); dsl_scan_ddt_entry(scn, ddb->ddb_checksum, &dde, tx); n++; if (dsl_scan_check_suspend(scn, NULL)) break; } zfs_dbgmsg("scanned %llu ddt entries with class_max = %u; " "suspending=%u", (longlong_t)n, (int)scn->scn_phys.scn_ddt_class_max, (int)scn->scn_suspending); ASSERT(error == 0 || error == ENOENT); ASSERT(error != ENOENT || ddb->ddb_class > scn->scn_phys.scn_ddt_class_max); } static uint64_t dsl_scan_ds_maxtxg(dsl_dataset_t *ds) { uint64_t smt = ds->ds_dir->dd_pool->dp_scan->scn_phys.scn_max_txg; if (ds->ds_is_snapshot) return (MIN(smt, dsl_dataset_phys(ds)->ds_creation_txg)); return (smt); } static void dsl_scan_visit(dsl_scan_t *scn, dmu_tx_t *tx) { scan_ds_t *sds; dsl_pool_t *dp = scn->scn_dp; if (scn->scn_phys.scn_ddt_bookmark.ddb_class <= scn->scn_phys.scn_ddt_class_max) { scn->scn_phys.scn_cur_min_txg = scn->scn_phys.scn_min_txg; scn->scn_phys.scn_cur_max_txg = scn->scn_phys.scn_max_txg; dsl_scan_ddt(scn, tx); if (scn->scn_suspending) return; } if (scn->scn_phys.scn_bookmark.zb_objset == DMU_META_OBJSET) { /* First do the MOS & ORIGIN */ scn->scn_phys.scn_cur_min_txg = scn->scn_phys.scn_min_txg; scn->scn_phys.scn_cur_max_txg = scn->scn_phys.scn_max_txg; dsl_scan_visit_rootbp(scn, NULL, &dp->dp_meta_rootbp, tx); spa_set_rootblkptr(dp->dp_spa, &dp->dp_meta_rootbp); if (scn->scn_suspending) return; if (spa_version(dp->dp_spa) < SPA_VERSION_DSL_SCRUB) { VERIFY0(dmu_objset_find_dp(dp, dp->dp_root_dir_obj, enqueue_cb, NULL, DS_FIND_CHILDREN)); } else { dsl_scan_visitds(scn, dp->dp_origin_snap->ds_object, tx); } ASSERT(!scn->scn_suspending); } else if (scn->scn_phys.scn_bookmark.zb_objset != ZB_DESTROYED_OBJSET) { uint64_t dsobj = scn->scn_phys.scn_bookmark.zb_objset; /* * If we were suspended, continue from here. Note if the * ds we were suspended on was deleted, the zb_objset may * be -1, so we will skip this and find a new objset * below. */ dsl_scan_visitds(scn, dsobj, tx); if (scn->scn_suspending) return; } /* * In case we suspended right at the end of the ds, zero the * bookmark so we don't think that we're still trying to resume. */ bzero(&scn->scn_phys.scn_bookmark, sizeof (zbookmark_phys_t)); /* * Keep pulling things out of the dataset avl queue. Updates to the * persistent zap-object-as-queue happen only at checkpoints. */ while ((sds = avl_first(&scn->scn_queue)) != NULL) { dsl_dataset_t *ds; uint64_t dsobj = sds->sds_dsobj; uint64_t txg = sds->sds_txg; /* dequeue and free the ds from the queue */ scan_ds_queue_remove(scn, dsobj); sds = NULL; /* must not be touched after removal */ /* Set up min / max txg */ VERIFY3U(0, ==, dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds)); if (txg != 0) { scn->scn_phys.scn_cur_min_txg = MAX(scn->scn_phys.scn_min_txg, txg); } else { scn->scn_phys.scn_cur_min_txg = MAX(scn->scn_phys.scn_min_txg, dsl_dataset_phys(ds)->ds_prev_snap_txg); } scn->scn_phys.scn_cur_max_txg = dsl_scan_ds_maxtxg(ds); dsl_dataset_rele(ds, FTAG); dsl_scan_visitds(scn, dsobj, tx); if (scn->scn_suspending) return; } /* No more objsets to fetch, we're done */ scn->scn_phys.scn_bookmark.zb_objset = ZB_DESTROYED_OBJSET; ASSERT0(scn->scn_suspending); } static uint64_t dsl_scan_count_leaves(vdev_t *vd) { uint64_t i, leaves = 0; /* we only count leaves that belong to the main pool and are readable */ if (vd->vdev_islog || vd->vdev_isspare || vd->vdev_isl2cache || !vdev_readable(vd)) return (0); if (vd->vdev_ops->vdev_op_leaf) return (1); for (i = 0; i < vd->vdev_children; i++) { leaves += dsl_scan_count_leaves(vd->vdev_child[i]); } return (leaves); } static void scan_io_queues_update_zio_stats(dsl_scan_io_queue_t *q, const blkptr_t *bp) { int i; uint64_t cur_size = 0; for (i = 0; i < BP_GET_NDVAS(bp); i++) { cur_size += DVA_GET_ASIZE(&bp->blk_dva[i]); } q->q_total_zio_size_this_txg += cur_size; q->q_zios_this_txg++; } static void scan_io_queues_update_seg_stats(dsl_scan_io_queue_t *q, uint64_t start, uint64_t end) { q->q_total_seg_size_this_txg += end - start; q->q_segs_this_txg++; } static boolean_t scan_io_queue_check_suspend(dsl_scan_t *scn) { /* See comment in dsl_scan_check_suspend() */ uint64_t curr_time_ns = gethrtime(); uint64_t scan_time_ns = curr_time_ns - scn->scn_sync_start_time; uint64_t sync_time_ns = curr_time_ns - scn->scn_dp->dp_spa->spa_sync_starttime; int dirty_pct = scn->scn_dp->dp_dirty_total * 100 / zfs_dirty_data_max; int mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; return ((NSEC2MSEC(scan_time_ns) > mintime && (dirty_pct >= zfs_vdev_async_write_active_min_dirty_percent || txg_sync_waiting(scn->scn_dp) || NSEC2SEC(sync_time_ns) >= zfs_txg_timeout)) || spa_shutting_down(scn->scn_dp->dp_spa)); } /* * Given a list of scan_io_t's in io_list, this issues the io's out to * disk. This consumes the io_list and frees the scan_io_t's. This is * called when emptying queues, either when we're up against the memory * limit or when we have finished scanning. Returns B_TRUE if we stopped * processing the list before we finished. Any zios that were not issued * will remain in the io_list. */ static boolean_t scan_io_queue_issue(dsl_scan_io_queue_t *queue, list_t *io_list) { dsl_scan_t *scn = queue->q_scn; scan_io_t *sio; int64_t bytes_issued = 0; boolean_t suspended = B_FALSE; while ((sio = list_head(io_list)) != NULL) { blkptr_t bp; if (scan_io_queue_check_suspend(scn)) { suspended = B_TRUE; break; } sio2bp(sio, &bp); bytes_issued += SIO_GET_ASIZE(sio); scan_exec_io(scn->scn_dp, &bp, sio->sio_flags, &sio->sio_zb, queue); (void) list_remove_head(io_list); scan_io_queues_update_zio_stats(queue, &bp); sio_free(sio); } atomic_add_64(&scn->scn_bytes_pending, -bytes_issued); return (suspended); } /* * Given a range_seg_t (extent) and a list, this function passes over a * scan queue and gathers up the appropriate ios which fit into that * scan seg (starting from lowest LBA). At the end, we remove the segment * from the q_exts_by_addr range tree. */ static boolean_t scan_io_queue_gather(dsl_scan_io_queue_t *queue, range_seg_t *rs, list_t *list) { scan_io_t *srch_sio, *sio, *next_sio; avl_index_t idx; uint_t num_sios = 0; int64_t bytes_issued = 0; ASSERT(rs != NULL); ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); srch_sio = sio_alloc(1); srch_sio->sio_nr_dvas = 1; SIO_SET_OFFSET(srch_sio, rs->rs_start); /* * The exact start of the extent might not contain any matching zios, * so if that's the case, examine the next one in the tree. */ sio = avl_find(&queue->q_sios_by_addr, srch_sio, &idx); sio_free(srch_sio); if (sio == NULL) sio = avl_nearest(&queue->q_sios_by_addr, idx, AVL_AFTER); while (sio != NULL && SIO_GET_OFFSET(sio) < rs->rs_end && num_sios <= 32) { ASSERT3U(SIO_GET_OFFSET(sio), >=, rs->rs_start); ASSERT3U(SIO_GET_END_OFFSET(sio), <=, rs->rs_end); next_sio = AVL_NEXT(&queue->q_sios_by_addr, sio); avl_remove(&queue->q_sios_by_addr, sio); queue->q_sio_memused -= SIO_GET_MUSED(sio); bytes_issued += SIO_GET_ASIZE(sio); num_sios++; list_insert_tail(list, sio); sio = next_sio; } /* * We limit the number of sios we process at once to 32 to avoid * biting off more than we can chew. If we didn't take everything * in the segment we update it to reflect the work we were able to * complete. Otherwise, we remove it from the range tree entirely. */ if (sio != NULL && SIO_GET_OFFSET(sio) < rs->rs_end) { range_tree_adjust_fill(queue->q_exts_by_addr, rs, -bytes_issued); range_tree_resize_segment(queue->q_exts_by_addr, rs, SIO_GET_OFFSET(sio), rs->rs_end - SIO_GET_OFFSET(sio)); return (B_TRUE); } else { range_tree_remove(queue->q_exts_by_addr, rs->rs_start, rs->rs_end - rs->rs_start); return (B_FALSE); } } /* * This is called from the queue emptying thread and selects the next * extent from which we are to issue io's. The behavior of this function * depends on the state of the scan, the current memory consumption and * whether or not we are performing a scan shutdown. * 1) We select extents in an elevator algorithm (LBA-order) if the scan * needs to perform a checkpoint * 2) We select the largest available extent if we are up against the * memory limit. * 3) Otherwise we don't select any extents. */ static const range_seg_t * scan_io_queue_fetch_ext(dsl_scan_io_queue_t *queue) { dsl_scan_t *scn = queue->q_scn; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); ASSERT(scn->scn_is_sorted); /* handle tunable overrides */ if (scn->scn_checkpointing || scn->scn_clearing) { if (zfs_scan_issue_strategy == 1) { return (range_tree_first(queue->q_exts_by_addr)); } else if (zfs_scan_issue_strategy == 2) { return (avl_first(&queue->q_exts_by_size)); } } /* * During normal clearing, we want to issue our largest segments * first, keeping IO as sequential as possible, and leaving the * smaller extents for later with the hope that they might eventually * grow to larger sequential segments. However, when the scan is * checkpointing, no new extents will be added to the sorting queue, * so the way we are sorted now is as good as it will ever get. * In this case, we instead switch to issuing extents in LBA order. */ if (scn->scn_checkpointing) { return (range_tree_first(queue->q_exts_by_addr)); } else if (scn->scn_clearing) { return (avl_first(&queue->q_exts_by_size)); } else { return (NULL); } } static void scan_io_queues_run_one(void *arg) { dsl_scan_io_queue_t *queue = arg; kmutex_t *q_lock = &queue->q_vd->vdev_scan_io_queue_lock; boolean_t suspended = B_FALSE; range_seg_t *rs = NULL; scan_io_t *sio = NULL; list_t sio_list; uint64_t bytes_per_leaf = zfs_scan_vdev_limit; uint64_t nr_leaves = dsl_scan_count_leaves(queue->q_vd); ASSERT(queue->q_scn->scn_is_sorted); list_create(&sio_list, sizeof (scan_io_t), offsetof(scan_io_t, sio_nodes.sio_list_node)); mutex_enter(q_lock); /* calculate maximum in-flight bytes for this txg (min 1MB) */ queue->q_maxinflight_bytes = MAX(nr_leaves * bytes_per_leaf, 1ULL << 20); /* reset per-queue scan statistics for this txg */ queue->q_total_seg_size_this_txg = 0; queue->q_segs_this_txg = 0; queue->q_total_zio_size_this_txg = 0; queue->q_zios_this_txg = 0; /* loop until we have run out of time or sios */ while ((rs = (range_seg_t *)scan_io_queue_fetch_ext(queue)) != NULL) { uint64_t seg_start = 0, seg_end = 0; boolean_t more_left = B_TRUE; ASSERT(list_is_empty(&sio_list)); /* loop while we still have sios left to process in this rs */ while (more_left) { scan_io_t *first_sio, *last_sio; /* * We have selected which extent needs to be * processed next. Gather up the corresponding sios. */ more_left = scan_io_queue_gather(queue, rs, &sio_list); ASSERT(!list_is_empty(&sio_list)); first_sio = list_head(&sio_list); last_sio = list_tail(&sio_list); seg_end = SIO_GET_END_OFFSET(last_sio); if (seg_start == 0) seg_start = SIO_GET_OFFSET(first_sio); /* * Issuing sios can take a long time so drop the * queue lock. The sio queue won't be updated by * other threads since we're in syncing context so * we can be sure that our trees will remain exactly * as we left them. */ mutex_exit(q_lock); suspended = scan_io_queue_issue(queue, &sio_list); mutex_enter(q_lock); if (suspended) break; } /* update statistics for debugging purposes */ scan_io_queues_update_seg_stats(queue, seg_start, seg_end); if (suspended) break; } /* * If we were suspended in the middle of processing, * requeue any unfinished sios and exit. */ while ((sio = list_head(&sio_list)) != NULL) { list_remove(&sio_list, sio); scan_io_queue_insert_impl(queue, sio); } mutex_exit(q_lock); list_destroy(&sio_list); } /* * Performs an emptying run on all scan queues in the pool. This just * punches out one thread per top-level vdev, each of which processes * only that vdev's scan queue. We can parallelize the I/O here because * we know that each queue's io's only affect its own top-level vdev. * * This function waits for the queue runs to complete, and must be * called from dsl_scan_sync (or in general, syncing context). */ static void scan_io_queues_run(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; ASSERT(scn->scn_is_sorted); ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER)); if (scn->scn_bytes_pending == 0) return; if (scn->scn_taskq == NULL) { char *tq_name = kmem_zalloc(ZFS_MAX_DATASET_NAME_LEN + 16, KM_SLEEP); int nthreads = spa->spa_root_vdev->vdev_children; /* * We need to make this taskq *always* execute as many * threads in parallel as we have top-level vdevs and no * less, otherwise strange serialization of the calls to * scan_io_queues_run_one can occur during spa_sync runs * and that significantly impacts performance. */ (void) snprintf(tq_name, ZFS_MAX_DATASET_NAME_LEN + 16, "dsl_scan_tq_%s", spa->spa_name); scn->scn_taskq = taskq_create(tq_name, nthreads, minclsyspri, nthreads, nthreads, TASKQ_PREPOPULATE); kmem_free(tq_name, ZFS_MAX_DATASET_NAME_LEN + 16); } for (uint64_t i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; mutex_enter(&vd->vdev_scan_io_queue_lock); if (vd->vdev_scan_io_queue != NULL) { VERIFY(taskq_dispatch(scn->scn_taskq, scan_io_queues_run_one, vd->vdev_scan_io_queue, TQ_SLEEP) != TASKQID_INVALID); } mutex_exit(&vd->vdev_scan_io_queue_lock); } /* * Wait for the queues to finish issuing thir IOs for this run * before we return. There may still be IOs in flight at this * point. */ taskq_wait(scn->scn_taskq); } static boolean_t dsl_scan_async_block_should_pause(dsl_scan_t *scn) { uint64_t elapsed_nanosecs; if (zfs_recover) return (B_FALSE); if (scn->scn_visited_this_txg >= zfs_async_block_max_blocks) return (B_TRUE); elapsed_nanosecs = gethrtime() - scn->scn_sync_start_time; return (elapsed_nanosecs / NANOSEC > zfs_txg_timeout || (NSEC2MSEC(elapsed_nanosecs) > scn->scn_async_block_min_time_ms && txg_sync_waiting(scn->scn_dp)) || spa_shutting_down(scn->scn_dp->dp_spa)); } static int dsl_scan_free_block_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { dsl_scan_t *scn = arg; if (!scn->scn_is_bptree || (BP_GET_LEVEL(bp) == 0 && BP_GET_TYPE(bp) != DMU_OT_OBJSET)) { if (dsl_scan_async_block_should_pause(scn)) return (SET_ERROR(ERESTART)); } zio_nowait(zio_free_sync(scn->scn_zio_root, scn->scn_dp->dp_spa, dmu_tx_get_txg(tx), bp, 0)); dsl_dir_diduse_space(tx->tx_pool->dp_free_dir, DD_USED_HEAD, -bp_get_dsize_sync(scn->scn_dp->dp_spa, bp), -BP_GET_PSIZE(bp), -BP_GET_UCSIZE(bp), tx); scn->scn_visited_this_txg++; return (0); } static void dsl_scan_update_stats(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; uint64_t i; uint64_t seg_size_total = 0, zio_size_total = 0; uint64_t seg_count_total = 0, zio_count_total = 0; for (i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; dsl_scan_io_queue_t *queue = vd->vdev_scan_io_queue; if (queue == NULL) continue; seg_size_total += queue->q_total_seg_size_this_txg; zio_size_total += queue->q_total_zio_size_this_txg; seg_count_total += queue->q_segs_this_txg; zio_count_total += queue->q_zios_this_txg; } if (seg_count_total == 0 || zio_count_total == 0) { scn->scn_avg_seg_size_this_txg = 0; scn->scn_avg_zio_size_this_txg = 0; scn->scn_segs_this_txg = 0; scn->scn_zios_this_txg = 0; return; } scn->scn_avg_seg_size_this_txg = seg_size_total / seg_count_total; scn->scn_avg_zio_size_this_txg = zio_size_total / zio_count_total; scn->scn_segs_this_txg = seg_count_total; scn->scn_zios_this_txg = zio_count_total; } static int dsl_scan_obsolete_block_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { dsl_scan_t *scn = arg; const dva_t *dva = &bp->blk_dva[0]; if (dsl_scan_async_block_should_pause(scn)) return (SET_ERROR(ERESTART)); spa_vdev_indirect_mark_obsolete(scn->scn_dp->dp_spa, DVA_GET_VDEV(dva), DVA_GET_OFFSET(dva), DVA_GET_ASIZE(dva), tx); scn->scn_visited_this_txg++; return (0); } boolean_t dsl_scan_active(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; uint64_t used = 0, comp, uncomp; if (spa->spa_load_state != SPA_LOAD_NONE) return (B_FALSE); if (spa_shutting_down(spa)) return (B_FALSE); if ((dsl_scan_is_running(scn) && !dsl_scan_is_paused_scrub(scn)) || (scn->scn_async_destroying && !scn->scn_async_stalled)) return (B_TRUE); if (spa_version(scn->scn_dp->dp_spa) >= SPA_VERSION_DEADLISTS) { (void) bpobj_space(&scn->scn_dp->dp_free_bpobj, &used, &comp, &uncomp); } return (used != 0); } static boolean_t +dsl_scan_check_deferred(vdev_t *vd) +{ + boolean_t need_resilver = B_FALSE; + + for (int c = 0; c < vd->vdev_children; c++) { + need_resilver |= + dsl_scan_check_deferred(vd->vdev_child[c]); + } + + if (!vdev_is_concrete(vd) || vd->vdev_aux || + !vd->vdev_ops->vdev_op_leaf) + return (need_resilver); + + if (!vd->vdev_resilver_deferred) + need_resilver = B_TRUE; + + return (need_resilver); +} + +static boolean_t dsl_scan_need_resilver(spa_t *spa, const dva_t *dva, size_t psize, uint64_t phys_birth) { vdev_t *vd; vd = vdev_lookup_top(spa, DVA_GET_VDEV(dva)); if (vd->vdev_ops == &vdev_indirect_ops) { /* * The indirect vdev can point to multiple * vdevs. For simplicity, always create * the resilver zio_t. zio_vdev_io_start() * will bypass the child resilver i/o's if * they are on vdevs that don't have DTL's. */ return (B_TRUE); } if (DVA_GET_GANG(dva)) { /* * Gang members may be spread across multiple * vdevs, so the best estimate we have is the * scrub range, which has already been checked. * XXX -- it would be better to change our * allocation policy to ensure that all * gang members reside on the same vdev. */ return (B_TRUE); } /* * Check if the txg falls within the range which must be * resilvered. DVAs outside this range can always be skipped. */ if (!vdev_dtl_contains(vd, DTL_PARTIAL, phys_birth, 1)) return (B_FALSE); /* * Check if the top-level vdev must resilver this offset. * When the offset does not intersect with a dirty leaf DTL * then it may be possible to skip the resilver IO. The psize * is provided instead of asize to simplify the check for RAIDZ. */ if (!vdev_dtl_need_resilver(vd, DVA_GET_OFFSET(dva), psize)) return (B_FALSE); + /* + * Check that this top-level vdev has a device under it which + * is resilvering and is not deferred. + */ + if (!dsl_scan_check_deferred(vd)) + return (B_FALSE); + return (B_TRUE); } static int dsl_process_async_destroys(dsl_pool_t *dp, dmu_tx_t *tx) { int err = 0; dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; if (spa_suspend_async_destroy(spa)) return (0); if (zfs_free_bpobj_enabled && spa_version(spa) >= SPA_VERSION_DEADLISTS) { scn->scn_is_bptree = B_FALSE; scn->scn_async_block_min_time_ms = zfs_free_min_time_ms; scn->scn_zio_root = zio_root(spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED); err = bpobj_iterate(&dp->dp_free_bpobj, dsl_scan_free_block_cb, scn, tx); VERIFY0(zio_wait(scn->scn_zio_root)); scn->scn_zio_root = NULL; if (err != 0 && err != ERESTART) zfs_panic_recover("error %u from bpobj_iterate()", err); } if (err == 0 && spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY)) { ASSERT(scn->scn_async_destroying); scn->scn_is_bptree = B_TRUE; scn->scn_zio_root = zio_root(spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED); err = bptree_iterate(dp->dp_meta_objset, dp->dp_bptree_obj, B_TRUE, dsl_scan_free_block_cb, scn, tx); VERIFY0(zio_wait(scn->scn_zio_root)); scn->scn_zio_root = NULL; if (err == EIO || err == ECKSUM) { err = 0; } else if (err != 0 && err != ERESTART) { zfs_panic_recover("error %u from " "traverse_dataset_destroyed()", err); } if (bptree_is_empty(dp->dp_meta_objset, dp->dp_bptree_obj)) { /* finished; deactivate async destroy feature */ spa_feature_decr(spa, SPA_FEATURE_ASYNC_DESTROY, tx); ASSERT(!spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY)); VERIFY0(zap_remove(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_BPTREE_OBJ, tx)); VERIFY0(bptree_free(dp->dp_meta_objset, dp->dp_bptree_obj, tx)); dp->dp_bptree_obj = 0; scn->scn_async_destroying = B_FALSE; scn->scn_async_stalled = B_FALSE; } else { /* * If we didn't make progress, mark the async * destroy as stalled, so that we will not initiate * a spa_sync() on its behalf. Note that we only * check this if we are not finished, because if the * bptree had no blocks for us to visit, we can * finish without "making progress". */ scn->scn_async_stalled = (scn->scn_visited_this_txg == 0); } } if (scn->scn_visited_this_txg) { zfs_dbgmsg("freed %llu blocks in %llums from " "free_bpobj/bptree txg %llu; err=%d", (longlong_t)scn->scn_visited_this_txg, (longlong_t) NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), (longlong_t)tx->tx_txg, err); scn->scn_visited_this_txg = 0; /* * Write out changes to the DDT that may be required as a * result of the blocks freed. This ensures that the DDT * is clean when a scrub/resilver runs. */ ddt_sync(spa, tx->tx_txg); } if (err != 0) return (err); if (dp->dp_free_dir != NULL && !scn->scn_async_destroying && zfs_free_leak_on_eio && (dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes != 0 || dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes != 0 || dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes != 0)) { /* * We have finished background destroying, but there is still * some space left in the dp_free_dir. Transfer this leaked * space to the dp_leak_dir. */ if (dp->dp_leak_dir == NULL) { rrw_enter(&dp->dp_config_rwlock, RW_WRITER, FTAG); (void) dsl_dir_create_sync(dp, dp->dp_root_dir, LEAK_DIR_NAME, tx); VERIFY0(dsl_pool_open_special_dir(dp, LEAK_DIR_NAME, &dp->dp_leak_dir)); rrw_exit(&dp->dp_config_rwlock, FTAG); } dsl_dir_diduse_space(dp->dp_leak_dir, DD_USED_HEAD, dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes, dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes, dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes, tx); dsl_dir_diduse_space(dp->dp_free_dir, DD_USED_HEAD, -dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes, -dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes, -dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes, tx); } if (dp->dp_free_dir != NULL && !scn->scn_async_destroying) { /* finished; verify that space accounting went to zero */ ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes); ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes); ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes); } EQUIV(bpobj_is_open(&dp->dp_obsolete_bpobj), 0 == zap_contains(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_OBSOLETE_BPOBJ)); if (err == 0 && bpobj_is_open(&dp->dp_obsolete_bpobj)) { ASSERT(spa_feature_is_active(dp->dp_spa, SPA_FEATURE_OBSOLETE_COUNTS)); scn->scn_is_bptree = B_FALSE; scn->scn_async_block_min_time_ms = zfs_obsolete_min_time_ms; err = bpobj_iterate(&dp->dp_obsolete_bpobj, dsl_scan_obsolete_block_cb, scn, tx); if (err != 0 && err != ERESTART) zfs_panic_recover("error %u from bpobj_iterate()", err); if (bpobj_is_empty(&dp->dp_obsolete_bpobj)) dsl_pool_destroy_obsolete_bpobj(dp, tx); } return (0); } /* * This is the primary entry point for scans that is called from syncing * context. Scans must happen entirely during syncing context so that we * cna guarantee that blocks we are currently scanning will not change out * from under us. While a scan is active, this funciton controls how quickly * transaction groups proceed, instead of the normal handling provided by * txg_sync_thread(). */ void dsl_scan_sync(dsl_pool_t *dp, dmu_tx_t *tx) { dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; int err = 0; state_sync_type_t sync_type = SYNC_OPTIONAL; + if (spa->spa_resilver_deferred && + !spa_feature_is_active(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER)) + spa_feature_incr(spa, SPA_FEATURE_RESILVER_DEFER, tx); + /* * Check for scn_restart_txg before checking spa_load_state, so * that we can restart an old-style scan while the pool is being - * imported (see dsl_scan_init). + * imported (see dsl_scan_init). We also restart scans if there + * is a deferred resilver and the user has manually disabled + * deferred resilvers via the tunable. */ - if (dsl_scan_restarting(scn, tx)) { + if (dsl_scan_restarting(scn, tx) || + (spa->spa_resilver_deferred && zfs_resilver_disable_defer)) { pool_scan_func_t func = POOL_SCAN_SCRUB; dsl_scan_done(scn, B_FALSE, tx); if (vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) func = POOL_SCAN_RESILVER; zfs_dbgmsg("restarting scan func=%u txg=%llu", func, (longlong_t)tx->tx_txg); dsl_scan_setup_sync(&func, tx); } /* * Only process scans in sync pass 1. */ if (spa_sync_pass(dp->dp_spa) > 1) return; /* * If the spa is shutting down, then stop scanning. This will * ensure that the scan does not dirty any new data during the * shutdown phase. */ if (spa_shutting_down(spa)) return; /* * If the scan is inactive due to a stalled async destroy, try again. */ if (!scn->scn_async_stalled && !dsl_scan_active(scn)) return; /* reset scan statistics */ scn->scn_visited_this_txg = 0; scn->scn_holes_this_txg = 0; scn->scn_lt_min_this_txg = 0; scn->scn_gt_max_this_txg = 0; scn->scn_ddt_contained_this_txg = 0; scn->scn_objsets_visited_this_txg = 0; scn->scn_avg_seg_size_this_txg = 0; scn->scn_segs_this_txg = 0; scn->scn_avg_zio_size_this_txg = 0; scn->scn_zios_this_txg = 0; scn->scn_suspending = B_FALSE; scn->scn_sync_start_time = gethrtime(); spa->spa_scrub_active = B_TRUE; /* * First process the async destroys. If we pause, don't do * any scrubbing or resilvering. This ensures that there are no * async destroys while we are scanning, so the scan code doesn't * have to worry about traversing it. It is also faster to free the * blocks than to scrub them. */ err = dsl_process_async_destroys(dp, tx); if (err != 0) return; if (!dsl_scan_is_running(scn) || dsl_scan_is_paused_scrub(scn)) return; /* * Wait a few txgs after importing to begin scanning so that * we can get the pool imported quickly. */ if (spa->spa_syncing_txg < spa->spa_first_txg + SCAN_IMPORT_WAIT_TXGS) return; /* + * zfs_scan_suspend_progress can be set to disable scan progress. + * We don't want to spin the txg_sync thread, so we add a delay + * here to simulate the time spent doing a scan. This is mostly + * useful for testing and debugging. + */ + if (zfs_scan_suspend_progress) { + uint64_t scan_time_ns = gethrtime() - scn->scn_sync_start_time; + int mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? + zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; + + while (zfs_scan_suspend_progress && + !txg_sync_waiting(scn->scn_dp) && + !spa_shutting_down(scn->scn_dp->dp_spa) && + NSEC2MSEC(scan_time_ns) < mintime) { + delay(hz); + scan_time_ns = gethrtime() - scn->scn_sync_start_time; + } + return; + } + + /* * It is possible to switch from unsorted to sorted at any time, * but afterwards the scan will remain sorted unless reloaded from * a checkpoint after a reboot. */ if (!zfs_scan_legacy) { scn->scn_is_sorted = B_TRUE; if (scn->scn_last_checkpoint == 0) scn->scn_last_checkpoint = ddi_get_lbolt(); } /* * For sorted scans, determine what kind of work we will be doing * this txg based on our memory limitations and whether or not we * need to perform a checkpoint. */ if (scn->scn_is_sorted) { /* * If we are over our checkpoint interval, set scn_clearing * so that we can begin checkpointing immediately. The * checkpoint allows us to save a consisent bookmark * representing how much data we have scrubbed so far. * Otherwise, use the memory limit to determine if we should * scan for metadata or start issue scrub IOs. We accumulate * metadata until we hit our hard memory limit at which point * we issue scrub IOs until we are at our soft memory limit. */ if (scn->scn_checkpointing || ddi_get_lbolt() - scn->scn_last_checkpoint > SEC_TO_TICK(zfs_scan_checkpoint_intval)) { if (!scn->scn_checkpointing) zfs_dbgmsg("begin scan checkpoint"); scn->scn_checkpointing = B_TRUE; scn->scn_clearing = B_TRUE; } else { boolean_t should_clear = dsl_scan_should_clear(scn); if (should_clear && !scn->scn_clearing) { zfs_dbgmsg("begin scan clearing"); scn->scn_clearing = B_TRUE; } else if (!should_clear && scn->scn_clearing) { zfs_dbgmsg("finish scan clearing"); scn->scn_clearing = B_FALSE; } } } else { ASSERT0(scn->scn_checkpointing); ASSERT0(scn->scn_clearing); } if (!scn->scn_clearing && scn->scn_done_txg == 0) { /* Need to scan metadata for more blocks to scrub */ dsl_scan_phys_t *scnp = &scn->scn_phys; taskqid_t prefetch_tqid; uint64_t bytes_per_leaf = zfs_scan_vdev_limit; uint64_t nr_leaves = dsl_scan_count_leaves(spa->spa_root_vdev); /* * Calculate the max number of in-flight bytes for pool-wide * scanning operations (minimum 1MB). Limits for the issuing * phase are done per top-level vdev and are handled separately. */ scn->scn_maxinflight_bytes = MAX(nr_leaves * bytes_per_leaf, 1ULL << 20); if (scnp->scn_ddt_bookmark.ddb_class <= scnp->scn_ddt_class_max) { ASSERT(ZB_IS_ZERO(&scnp->scn_bookmark)); zfs_dbgmsg("doing scan sync txg %llu; " "ddt bm=%llu/%llu/%llu/%llx", (longlong_t)tx->tx_txg, (longlong_t)scnp->scn_ddt_bookmark.ddb_class, (longlong_t)scnp->scn_ddt_bookmark.ddb_type, (longlong_t)scnp->scn_ddt_bookmark.ddb_checksum, (longlong_t)scnp->scn_ddt_bookmark.ddb_cursor); } else { zfs_dbgmsg("doing scan sync txg %llu; " "bm=%llu/%llu/%llu/%llu", (longlong_t)tx->tx_txg, (longlong_t)scnp->scn_bookmark.zb_objset, (longlong_t)scnp->scn_bookmark.zb_object, (longlong_t)scnp->scn_bookmark.zb_level, (longlong_t)scnp->scn_bookmark.zb_blkid); } scn->scn_zio_root = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_CANFAIL); scn->scn_prefetch_stop = B_FALSE; prefetch_tqid = taskq_dispatch(dp->dp_sync_taskq, dsl_scan_prefetch_thread, scn, TQ_SLEEP); ASSERT(prefetch_tqid != TASKQID_INVALID); dsl_pool_config_enter(dp, FTAG); dsl_scan_visit(scn, tx); dsl_pool_config_exit(dp, FTAG); mutex_enter(&dp->dp_spa->spa_scrub_lock); scn->scn_prefetch_stop = B_TRUE; cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&dp->dp_spa->spa_scrub_lock); taskq_wait_id(dp->dp_sync_taskq, prefetch_tqid); (void) zio_wait(scn->scn_zio_root); scn->scn_zio_root = NULL; zfs_dbgmsg("scan visited %llu blocks in %llums " "(%llu os's, %llu holes, %llu < mintxg, " "%llu in ddt, %llu > maxtxg)", (longlong_t)scn->scn_visited_this_txg, (longlong_t)NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), (longlong_t)scn->scn_objsets_visited_this_txg, (longlong_t)scn->scn_holes_this_txg, (longlong_t)scn->scn_lt_min_this_txg, (longlong_t)scn->scn_ddt_contained_this_txg, (longlong_t)scn->scn_gt_max_this_txg); if (!scn->scn_suspending) { ASSERT0(avl_numnodes(&scn->scn_queue)); scn->scn_done_txg = tx->tx_txg + 1; if (scn->scn_is_sorted) { scn->scn_checkpointing = B_TRUE; scn->scn_clearing = B_TRUE; } zfs_dbgmsg("scan complete txg %llu", (longlong_t)tx->tx_txg); } } else if (scn->scn_is_sorted && scn->scn_bytes_pending != 0) { + ASSERT(scn->scn_clearing); + /* need to issue scrubbing IOs from per-vdev queues */ scn->scn_zio_root = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_CANFAIL); scan_io_queues_run(scn); (void) zio_wait(scn->scn_zio_root); scn->scn_zio_root = NULL; /* calculate and dprintf the current memory usage */ (void) dsl_scan_should_clear(scn); dsl_scan_update_stats(scn); zfs_dbgmsg("scrubbed %llu blocks (%llu segs) in %llums " "(avg_block_size = %llu, avg_seg_size = %llu)", (longlong_t)scn->scn_zios_this_txg, (longlong_t)scn->scn_segs_this_txg, (longlong_t)NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), (longlong_t)scn->scn_avg_zio_size_this_txg, (longlong_t)scn->scn_avg_seg_size_this_txg); } else if (scn->scn_done_txg != 0 && scn->scn_done_txg <= tx->tx_txg) { /* Finished with everything. Mark the scrub as complete */ zfs_dbgmsg("scan issuing complete txg %llu", (longlong_t)tx->tx_txg); ASSERT3U(scn->scn_done_txg, !=, 0); ASSERT0(spa->spa_scrub_inflight); ASSERT0(scn->scn_bytes_pending); dsl_scan_done(scn, B_TRUE, tx); sync_type = SYNC_MANDATORY; } dsl_scan_sync_state(scn, tx, sync_type); } static void count_block(dsl_scan_t *scn, zfs_all_blkstats_t *zab, const blkptr_t *bp) { int i; /* * Don't count embedded bp's, since we already did the work of * scanning these when we scanned the containing block. */ if (BP_IS_EMBEDDED(bp)) return; /* * Update the spa's stats on how many bytes we have issued. * Sequential scrubs create a zio for each DVA of the bp. Each * of these will include all DVAs for repair purposes, but the * zio code will only try the first one unless there is an issue. * Therefore, we should only count the first DVA for these IOs. */ if (scn->scn_is_sorted) { atomic_add_64(&scn->scn_dp->dp_spa->spa_scan_pass_issued, DVA_GET_ASIZE(&bp->blk_dva[0])); } else { spa_t *spa = scn->scn_dp->dp_spa; for (i = 0; i < BP_GET_NDVAS(bp); i++) { atomic_add_64(&spa->spa_scan_pass_issued, DVA_GET_ASIZE(&bp->blk_dva[i])); } } /* * If we resume after a reboot, zab will be NULL; don't record * incomplete stats in that case. */ if (zab == NULL) return; mutex_enter(&zab->zab_lock); for (i = 0; i < 4; i++) { int l = (i < 2) ? BP_GET_LEVEL(bp) : DN_MAX_LEVELS; int t = (i & 1) ? BP_GET_TYPE(bp) : DMU_OT_TOTAL; if (t & DMU_OT_NEWTYPE) t = DMU_OT_OTHER; zfs_blkstat_t *zb = &zab->zab_type[l][t]; int equal; zb->zb_count++; zb->zb_asize += BP_GET_ASIZE(bp); zb->zb_lsize += BP_GET_LSIZE(bp); zb->zb_psize += BP_GET_PSIZE(bp); 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_2_of_2_samevdev++; 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 == 1) zb->zb_ditto_2_of_3_samevdev++; else if (equal == 3) zb->zb_ditto_3_of_3_samevdev++; break; } } mutex_exit(&zab->zab_lock); } static void scan_io_queue_insert_impl(dsl_scan_io_queue_t *queue, scan_io_t *sio) { avl_index_t idx; int64_t asize = SIO_GET_ASIZE(sio); dsl_scan_t *scn = queue->q_scn; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); if (avl_find(&queue->q_sios_by_addr, sio, &idx) != NULL) { /* block is already scheduled for reading */ atomic_add_64(&scn->scn_bytes_pending, -asize); sio_free(sio); return; } avl_insert(&queue->q_sios_by_addr, sio, idx); queue->q_sio_memused += SIO_GET_MUSED(sio); range_tree_add(queue->q_exts_by_addr, SIO_GET_OFFSET(sio), asize); } /* * Given all the info we got from our metadata scanning process, we * construct a scan_io_t and insert it into the scan sorting queue. The * I/O must already be suitable for us to process. This is controlled * by dsl_scan_enqueue(). */ static void scan_io_queue_insert(dsl_scan_io_queue_t *queue, const blkptr_t *bp, int dva_i, int zio_flags, const zbookmark_phys_t *zb) { dsl_scan_t *scn = queue->q_scn; scan_io_t *sio = sio_alloc(BP_GET_NDVAS(bp)); ASSERT0(BP_IS_GANG(bp)); ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); bp2sio(bp, sio, dva_i); sio->sio_flags = zio_flags; sio->sio_zb = *zb; /* * Increment the bytes pending counter now so that we can't * get an integer underflow in case the worker processes the * zio before we get to incrementing this counter. */ atomic_add_64(&scn->scn_bytes_pending, SIO_GET_ASIZE(sio)); scan_io_queue_insert_impl(queue, sio); } /* * Given a set of I/O parameters as discovered by the metadata traversal * process, attempts to place the I/O into the sorted queues (if allowed), * or immediately executes the I/O. */ static void dsl_scan_enqueue(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb) { spa_t *spa = dp->dp_spa; ASSERT(!BP_IS_EMBEDDED(bp)); /* * Gang blocks are hard to issue sequentially, so we just issue them * here immediately instead of queuing them. */ if (!dp->dp_scan->scn_is_sorted || BP_IS_GANG(bp)) { scan_exec_io(dp, bp, zio_flags, zb, NULL); return; } for (int i = 0; i < BP_GET_NDVAS(bp); i++) { dva_t dva; vdev_t *vdev; dva = bp->blk_dva[i]; vdev = vdev_lookup_top(spa, DVA_GET_VDEV(&dva)); ASSERT(vdev != NULL); mutex_enter(&vdev->vdev_scan_io_queue_lock); if (vdev->vdev_scan_io_queue == NULL) vdev->vdev_scan_io_queue = scan_io_queue_create(vdev); ASSERT(dp->dp_scan != NULL); scan_io_queue_insert(vdev->vdev_scan_io_queue, bp, i, zio_flags, zb); mutex_exit(&vdev->vdev_scan_io_queue_lock); } } static int dsl_scan_scrub_cb(dsl_pool_t *dp, const blkptr_t *bp, const zbookmark_phys_t *zb) { dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; uint64_t phys_birth = BP_PHYSICAL_BIRTH(bp); size_t psize = BP_GET_PSIZE(bp); boolean_t needs_io; int zio_flags = ZIO_FLAG_SCAN_THREAD | ZIO_FLAG_RAW | ZIO_FLAG_CANFAIL; int d; if (phys_birth <= scn->scn_phys.scn_min_txg || phys_birth >= scn->scn_phys.scn_max_txg) { count_block(scn, dp->dp_blkstats, bp); return (0); } /* Embedded BP's have phys_birth==0, so we reject them above. */ ASSERT(!BP_IS_EMBEDDED(bp)); ASSERT(DSL_SCAN_IS_SCRUB_RESILVER(scn)); if (scn->scn_phys.scn_func == POOL_SCAN_SCRUB) { zio_flags |= ZIO_FLAG_SCRUB; needs_io = B_TRUE; } else { ASSERT3U(scn->scn_phys.scn_func, ==, POOL_SCAN_RESILVER); zio_flags |= ZIO_FLAG_RESILVER; needs_io = B_FALSE; } /* If it's an intent log block, failure is expected. */ if (zb->zb_level == ZB_ZIL_LEVEL) zio_flags |= ZIO_FLAG_SPECULATIVE; for (d = 0; d < BP_GET_NDVAS(bp); d++) { const dva_t *dva = &bp->blk_dva[d]; /* * Keep track of how much data we've examined so that * zpool(1M) status can make useful progress reports. */ scn->scn_phys.scn_examined += DVA_GET_ASIZE(dva); spa->spa_scan_pass_exam += DVA_GET_ASIZE(dva); /* if it's a resilver, this may not be in the target range */ if (!needs_io) needs_io = dsl_scan_need_resilver(spa, dva, psize, phys_birth); } if (needs_io && !zfs_no_scrub_io) { dsl_scan_enqueue(dp, bp, zio_flags, zb); } else { count_block(scn, dp->dp_blkstats, bp); } /* do not relocate this block */ return (0); } static void dsl_scan_scrub_done(zio_t *zio) { spa_t *spa = zio->io_spa; blkptr_t *bp = zio->io_bp; dsl_scan_io_queue_t *queue = zio->io_private; abd_free(zio->io_abd); if (queue == NULL) { mutex_enter(&spa->spa_scrub_lock); ASSERT3U(spa->spa_scrub_inflight, >=, BP_GET_PSIZE(bp)); spa->spa_scrub_inflight -= BP_GET_PSIZE(bp); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); } else { mutex_enter(&queue->q_vd->vdev_scan_io_queue_lock); ASSERT3U(queue->q_inflight_bytes, >=, BP_GET_PSIZE(bp)); queue->q_inflight_bytes -= BP_GET_PSIZE(bp); cv_broadcast(&queue->q_zio_cv); mutex_exit(&queue->q_vd->vdev_scan_io_queue_lock); } if (zio->io_error && (zio->io_error != ECKSUM || !(zio->io_flags & ZIO_FLAG_SPECULATIVE))) { atomic_inc_64(&spa->spa_dsl_pool->dp_scan->scn_phys.scn_errors); } } /* * Given a scanning zio's information, executes the zio. The zio need * not necessarily be only sortable, this function simply executes the * zio, no matter what it is. The optional queue argument allows the * caller to specify that they want per top level vdev IO rate limiting * instead of the legacy global limiting. */ static void scan_exec_io(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb, dsl_scan_io_queue_t *queue) { spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; size_t size = BP_GET_PSIZE(bp); abd_t *data = abd_alloc_for_io(size, B_FALSE); if (queue == NULL) { mutex_enter(&spa->spa_scrub_lock); while (spa->spa_scrub_inflight >= scn->scn_maxinflight_bytes) cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock); spa->spa_scrub_inflight += BP_GET_PSIZE(bp); mutex_exit(&spa->spa_scrub_lock); } else { kmutex_t *q_lock = &queue->q_vd->vdev_scan_io_queue_lock; mutex_enter(q_lock); while (queue->q_inflight_bytes >= queue->q_maxinflight_bytes) cv_wait(&queue->q_zio_cv, q_lock); queue->q_inflight_bytes += BP_GET_PSIZE(bp); mutex_exit(q_lock); } count_block(dp->dp_scan, dp->dp_blkstats, bp); zio_nowait(zio_read(dp->dp_scan->scn_zio_root, spa, bp, data, size, dsl_scan_scrub_done, queue, ZIO_PRIORITY_SCRUB, zio_flags, zb)); } /* * This is the primary extent sorting algorithm. We balance two parameters: * 1) how many bytes of I/O are in an extent * 2) how well the extent is filled with I/O (as a fraction of its total size) * Since we allow extents to have gaps between their constituent I/Os, it's * possible to have a fairly large extent that contains the same amount of * I/O bytes than a much smaller extent, which just packs the I/O more tightly. * The algorithm sorts based on a score calculated from the extent's size, * the relative fill volume (in %) and a "fill weight" parameter that controls * the split between whether we prefer larger extents or more well populated * extents: * * SCORE = FILL_IN_BYTES + (FILL_IN_PERCENT * FILL_IN_BYTES * FILL_WEIGHT) * * Example: * 1) assume extsz = 64 MiB * 2) assume fill = 32 MiB (extent is half full) * 3) assume fill_weight = 3 * 4) SCORE = 32M + (((32M * 100) / 64M) * 3 * 32M) / 100 * SCORE = 32M + (50 * 3 * 32M) / 100 * SCORE = 32M + (4800M / 100) * SCORE = 32M + 48M * ^ ^ * | +--- final total relative fill-based score * +--------- final total fill-based score * SCORE = 80M * * As can be seen, at fill_ratio=3, the algorithm is slightly biased towards * extents that are more completely filled (in a 3:2 ratio) vs just larger. * Note that as an optimization, we replace multiplication and division by * 100 with bitshifting by 7 (which effecitvely multiplies and divides by 128). */ static int ext_size_compare(const void *x, const void *y) { const range_seg_t *rsa = x, *rsb = y; uint64_t sa = rsa->rs_end - rsa->rs_start, sb = rsb->rs_end - rsb->rs_start; uint64_t score_a, score_b; score_a = rsa->rs_fill + ((((rsa->rs_fill << 7) / sa) * fill_weight * rsa->rs_fill) >> 7); score_b = rsb->rs_fill + ((((rsb->rs_fill << 7) / sb) * fill_weight * rsb->rs_fill) >> 7); if (score_a > score_b) return (-1); if (score_a == score_b) { if (rsa->rs_start < rsb->rs_start) return (-1); if (rsa->rs_start == rsb->rs_start) return (0); return (1); } return (1); } /* * Comparator for the q_sios_by_addr tree. Sorting is simply performed * based on LBA-order (from lowest to highest). */ static int sio_addr_compare(const void *x, const void *y) { const scan_io_t *a = x, *b = y; return (AVL_CMP(SIO_GET_OFFSET(a), SIO_GET_OFFSET(b))); } /* IO queues are created on demand when they are needed. */ static dsl_scan_io_queue_t * scan_io_queue_create(vdev_t *vd) { dsl_scan_t *scn = vd->vdev_spa->spa_dsl_pool->dp_scan; dsl_scan_io_queue_t *q = kmem_zalloc(sizeof (*q), KM_SLEEP); q->q_scn = scn; q->q_vd = vd; q->q_sio_memused = 0; cv_init(&q->q_zio_cv, NULL, CV_DEFAULT, NULL); q->q_exts_by_addr = range_tree_create_impl(&rt_avl_ops, &q->q_exts_by_size, ext_size_compare, zfs_scan_max_ext_gap); avl_create(&q->q_sios_by_addr, sio_addr_compare, sizeof (scan_io_t), offsetof(scan_io_t, sio_nodes.sio_addr_node)); return (q); } /* * Destroys a scan queue and all segments and scan_io_t's contained in it. * No further execution of I/O occurs, anything pending in the queue is * simply freed without being executed. */ void dsl_scan_io_queue_destroy(dsl_scan_io_queue_t *queue) { dsl_scan_t *scn = queue->q_scn; scan_io_t *sio; void *cookie = NULL; int64_t bytes_dequeued = 0; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); while ((sio = avl_destroy_nodes(&queue->q_sios_by_addr, &cookie)) != NULL) { ASSERT(range_tree_contains(queue->q_exts_by_addr, SIO_GET_OFFSET(sio), SIO_GET_ASIZE(sio))); bytes_dequeued += SIO_GET_ASIZE(sio); queue->q_sio_memused -= SIO_GET_MUSED(sio); sio_free(sio); } ASSERT0(queue->q_sio_memused); atomic_add_64(&scn->scn_bytes_pending, -bytes_dequeued); range_tree_vacate(queue->q_exts_by_addr, NULL, queue); range_tree_destroy(queue->q_exts_by_addr); avl_destroy(&queue->q_sios_by_addr); cv_destroy(&queue->q_zio_cv); kmem_free(queue, sizeof (*queue)); } /* * Properly transfers a dsl_scan_queue_t from `svd' to `tvd'. This is * called on behalf of vdev_top_transfer when creating or destroying * a mirror vdev due to zpool attach/detach. */ void dsl_scan_io_queue_vdev_xfer(vdev_t *svd, vdev_t *tvd) { mutex_enter(&svd->vdev_scan_io_queue_lock); mutex_enter(&tvd->vdev_scan_io_queue_lock); VERIFY3P(tvd->vdev_scan_io_queue, ==, NULL); tvd->vdev_scan_io_queue = svd->vdev_scan_io_queue; svd->vdev_scan_io_queue = NULL; if (tvd->vdev_scan_io_queue != NULL) tvd->vdev_scan_io_queue->q_vd = tvd; mutex_exit(&tvd->vdev_scan_io_queue_lock); mutex_exit(&svd->vdev_scan_io_queue_lock); } static void scan_io_queues_destroy(dsl_scan_t *scn) { vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *tvd = rvd->vdev_child[i]; mutex_enter(&tvd->vdev_scan_io_queue_lock); if (tvd->vdev_scan_io_queue != NULL) dsl_scan_io_queue_destroy(tvd->vdev_scan_io_queue); tvd->vdev_scan_io_queue = NULL; mutex_exit(&tvd->vdev_scan_io_queue_lock); } } static void dsl_scan_freed_dva(spa_t *spa, const blkptr_t *bp, int dva_i) { dsl_pool_t *dp = spa->spa_dsl_pool; dsl_scan_t *scn = dp->dp_scan; vdev_t *vdev; kmutex_t *q_lock; dsl_scan_io_queue_t *queue; scan_io_t *srch_sio, *sio; avl_index_t idx; uint64_t start, size; vdev = vdev_lookup_top(spa, DVA_GET_VDEV(&bp->blk_dva[dva_i])); ASSERT(vdev != NULL); q_lock = &vdev->vdev_scan_io_queue_lock; queue = vdev->vdev_scan_io_queue; mutex_enter(q_lock); if (queue == NULL) { mutex_exit(q_lock); return; } srch_sio = sio_alloc(BP_GET_NDVAS(bp)); bp2sio(bp, srch_sio, dva_i); start = SIO_GET_OFFSET(srch_sio); size = SIO_GET_ASIZE(srch_sio); /* * We can find the zio in two states: * 1) Cold, just sitting in the queue of zio's to be issued at * some point in the future. In this case, all we do is * remove the zio from the q_sios_by_addr tree, decrement * its data volume from the containing range_seg_t and * resort the q_exts_by_size tree to reflect that the * range_seg_t has lost some of its 'fill'. We don't shorten * the range_seg_t - this is usually rare enough not to be * worth the extra hassle of trying keep track of precise * extent boundaries. * 2) Hot, where the zio is currently in-flight in * dsl_scan_issue_ios. In this case, we can't simply * reach in and stop the in-flight zio's, so we instead * block the caller. Eventually, dsl_scan_issue_ios will * be done with issuing the zio's it gathered and will * signal us. */ sio = avl_find(&queue->q_sios_by_addr, srch_sio, &idx); sio_free(srch_sio); if (sio != NULL) { int64_t asize = SIO_GET_ASIZE(sio); blkptr_t tmpbp; /* Got it while it was cold in the queue */ ASSERT3U(start, ==, SIO_GET_OFFSET(sio)); ASSERT3U(size, ==, asize); avl_remove(&queue->q_sios_by_addr, sio); queue->q_sio_memused -= SIO_GET_MUSED(sio); ASSERT(range_tree_contains(queue->q_exts_by_addr, start, size)); range_tree_remove_fill(queue->q_exts_by_addr, start, size); /* * We only update scn_bytes_pending in the cold path, * otherwise it will already have been accounted for as * part of the zio's execution. */ atomic_add_64(&scn->scn_bytes_pending, -asize); /* count the block as though we issued it */ sio2bp(sio, &tmpbp); count_block(scn, dp->dp_blkstats, &tmpbp); sio_free(sio); } mutex_exit(q_lock); } /* * Callback invoked when a zio_free() zio is executing. This needs to be * intercepted to prevent the zio from deallocating a particular portion * of disk space and it then getting reallocated and written to, while we * still have it queued up for processing. */ void dsl_scan_freed(spa_t *spa, const blkptr_t *bp) { dsl_pool_t *dp = spa->spa_dsl_pool; dsl_scan_t *scn = dp->dp_scan; ASSERT(!BP_IS_EMBEDDED(bp)); ASSERT(scn != NULL); if (!dsl_scan_is_running(scn)) return; for (int i = 0; i < BP_GET_NDVAS(bp); i++) dsl_scan_freed_dva(spa, bp, i); } Index: vendor-sys/illumos/dist/uts/common/fs/zfs/spa.c =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/spa.c (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/spa.c (revision 354961) @@ -1,8505 +1,8517 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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) 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) 2014 Integros [integros.com] * Copyright 2016 Toomas Soome * Copyright 2018 Joyent, Inc. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2017 Datto Inc. * Copyright 2018 OmniOS Community Edition (OmniOSce) Association. */ /* * 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 #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef _KERNEL #include #include #include #include #include #include #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_NULL, /* don't create a taskq */ ZTI_NMODES } zti_modes_t; #define ZTI_P(n, q) { ZTI_MODE_FIXED, (n), (q) } #define ZTI_BATCH { ZTI_MODE_BATCH, 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] = { "issue", "issue_high", "intr", "intr_high" }; /* * 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 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. * * 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. */ 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_P(12, 8), ZTI_NULL }, /* READ */ { ZTI_BATCH, ZTI_N(5), ZTI_N(8), ZTI_N(5) }, /* WRITE */ { ZTI_P(12, 8), 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 */ }; 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, char **ereport); static void spa_vdev_resilver_done(spa_t *spa); uint_t zio_taskq_batch_pct = 75; /* 1 thread per cpu in pset */ id_t zio_taskq_psrset_bind = PS_NONE; boolean_t zio_taskq_sysdc = B_TRUE; /* use SDC scheduling class */ uint_t zio_taskq_basedc = 80; /* base duty cycle */ boolean_t spa_create_process = B_TRUE; /* no process ==> no sysdc */ extern int zfs_sync_pass_deferred_free; /* * 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; /* * 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. */ boolean_t 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. */ boolean_t zfs_pause_spa_sync = B_FALSE; /* * ========================================================================== * 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, char *strval, uint64_t intval, zprop_source_t src) { const char *propname = zpool_prop_to_name(prop); nvlist_t *propval; VERIFY(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_uint64(propval, ZPROP_SOURCE, src) == 0); if (strval != NULL) VERIFY(nvlist_add_string(propval, ZPROP_VALUE, strval) == 0); else VERIFY(nvlist_add_uint64(propval, ZPROP_VALUE, intval) == 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; 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)); size = metaslab_class_get_space(mc); size += metaslab_class_get_space(spa_special_class(spa)); size += metaslab_class_get_space(spa_dedup_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) == FREAD), 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_HEALTH, NULL, rvd->vdev_state, src); version = spa_version(spa); if (version == zpool_prop_default_numeric(ZPOOL_PROP_VERSION)) src = ZPROP_SRC_DEFAULT; else src = ZPROP_SRC_LOCAL; spa_prop_add_list(*nvp, ZPOOL_PROP_VERSION, NULL, version, 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_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; int err; VERIFY(nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP) == 0); 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) { mutex_exit(&spa->spa_props_lock); return (0); } /* * 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) 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_pool_t *dp; dsl_dataset_t *ds = NULL; dp = spa_get_dsl(spa); dsl_pool_config_enter(dp, FTAG); err = dsl_dataset_hold_obj(dp, za.za_first_integer, FTAG, &ds); if (err != 0) { dsl_pool_config_exit(dp, FTAG); break; } strval = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP); dsl_dataset_name(ds, strval); dsl_dataset_rele(ds, FTAG); dsl_pool_config_exit(dp, 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; } spa_prop_add_list(*nvp, prop, strval, 0, src); kmem_free(strval, za.za_num_integers); break; default: break; } } zap_cursor_fini(&zc); mutex_exit(&spa->spa_props_lock); out: 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; 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: if (!zpool_prop_feature(propname)) { error = SET_ERROR(EINVAL); break; } /* * Sanitize the input. */ 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; 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: 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 && !spa_get_hostid()) 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; uint64_t propval; 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, and its property settings * must be supported. */ if (dmu_objset_type(os) != DMU_OST_ZFS) { error = SET_ERROR(ENOTSUP); } else if ((error = dsl_prop_get_int_ds(dmu_objset_ds(os), zfs_prop_to_name(ZFS_PROP_COMPRESSION), &propval)) == 0 && !BOOTFS_COMPRESS_VALID(propval)) { 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_WAIT || 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++) { /* * The kernel doesn't have an easy isprint() * check. For this kernel check, we merely * check ASCII apart from DEL. Fix this if * there is an easy-to-use kernel isprint(). */ if (*check >= 0x7f) { error = SET_ERROR(EINVAL); break; } } if (strlen(strval) > ZPROP_MAX_COMMENT) error = E2BIG; break; case ZPOOL_PROP_DEDUPDITTO: if (spa_version(spa) < SPA_VERSION_DEDUP) error = SET_ERROR(ENOTSUP); else error = nvpair_value_uint64(elem, &intval); if (error == 0 && intval != 0 && intval < ZIO_DEDUPDITTO_MIN) error = SET_ERROR(EINVAL); break; } if (error) break; } 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) { 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_VERSION || prop == ZPOOL_PROP_INVAL) { uint64_t ver; 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; } } /*ARGSUSED*/ static int spa_change_guid_check(void *arg, dmu_tx_t *tx) { uint64_t *newguid = 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", oldguid, *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) { spa_write_cachefile(spa, B_FALSE, 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 (AVL_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)); bcopy(&spa->spa_errlist_last, last, sizeof (avl_tree_t)); bcopy(&spa->spa_errlist_scrub, 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]; char name[32]; uint_t flags = 0; boolean_t batch = B_FALSE; if (mode == ZTI_MODE_NULL) { tqs->stqs_count = 0; tqs->stqs_taskq = NULL; return; } ASSERT3U(count, >, 0); tqs->stqs_count = count; tqs->stqs_taskq = kmem_alloc(count * sizeof (taskq_t *), KM_SLEEP); switch (mode) { case ZTI_MODE_FIXED: ASSERT3U(value, >=, 1); value = MAX(value, 1); break; case ZTI_MODE_BATCH: batch = B_TRUE; flags |= TASKQ_THREADS_CPU_PCT; value = zio_taskq_batch_pct; break; default: panic("unrecognized mode for %s_%s taskq (%u:%u) in " "spa_activate()", zio_type_name[t], zio_taskq_types[q], mode, value); break; } for (uint_t i = 0; i < count; i++) { taskq_t *tq; 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; 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 lower priority * than the other taskqs. */ if (t == ZIO_TYPE_WRITE && q == ZIO_TASKQ_ISSUE) pri--; 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) { ASSERT0(tqs->stqs_count); 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[gethrtime() % tqs->stqs_count]; } taskq_dispatch_ent(tq, func, arg, flags, ent); } 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); } } } #ifdef _KERNEL static void spa_thread(void *arg) { 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, int mode) { ASSERT(spa->spa_state == POOL_STATE_UNINITIALIZED); spa->spa_state = POOL_STATE_ACTIVE; spa->spa_mode = mode; spa->spa_normal_class = metaslab_class_create(spa, zfs_metaslab_ops); spa->spa_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; /* 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 } } 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)); } /* * 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); 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); 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_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); 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; } } /* * 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. */ static 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); } /* * Opposite of spa_load(). */ static void spa_unload(spa_t *spa) { int i; ASSERT(MUTEX_HELD(&spa_namespace_lock)); spa_load_note(spa, "UNLOADING"); /* * Stop async tasks. */ spa_async_suspend(spa); if (spa->spa_root_vdev) { vdev_initialize_stop_all(spa->spa_root_vdev, VDEV_INITIALIZE_ACTIVE); } /* * Stop syncing. */ if (spa->spa_sync_on) { txg_sync_stop(spa->spa_dsl_pool); spa->spa_sync_on = B_FALSE; } /* * Even though vdev_free() also calls vdev_metaslab_fini, we need * to call it earlier, before we wait for async i/o to complete. * This ensures that there is no async metaslab prefetching, by * calling taskq_wait(mg_taskq). */ if (spa->spa_root_vdev != NULL) { spa_config_enter(spa, SCL_ALL, spa, RW_WRITER); for (int c = 0; c < spa->spa_root_vdev->vdev_children; c++) vdev_metaslab_fini(spa->spa_root_vdev->vdev_child[c]); spa_config_exit(spa, SCL_ALL, spa); } 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; } 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; } 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); /* * Drop and purge level 2 cache */ spa_l2cache_drop(spa); for (i = 0; i < spa->spa_spares.sav_count; i++) vdev_free(spa->spa_spares.sav_vdevs[i]); if (spa->spa_spares.sav_vdevs) { 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; for (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]); } if (spa->spa_l2cache.sav_vdevs) { 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; } 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. */ 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); } if (spa->spa_spares.sav_vdevs) kmem_free(spa->spa_spares.sav_vdevs, spa->spa_spares.sav_count * sizeof (void *)); if (spa->spa_spares.sav_config == NULL) nspares = 0; else VERIFY(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0); 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_alloc(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. */ VERIFY(nvlist_remove(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, DATA_TYPE_NVLIST_ARRAY) == 0); 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); VERIFY(nvlist_add_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, spares, spa->spa_spares.sav_count) == 0); 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; 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); if (sav->sav_config != NULL) { VERIFY(nvlist_lookup_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0); newvdevs = kmem_alloc(nl2cache * sizeof (void *), KM_SLEEP); } else { nl2cache = 0; newvdevs = NULL; } oldvdevs = sav->sav_vdevs; oldnvdevs = sav->sav_count; sav->sav_vdevs = NULL; sav->sav_count = 0; /* * Process new nvlist of vdevs. */ for (i = 0; i < nl2cache; i++) { VERIFY(nvlist_lookup_uint64(l2cache[i], ZPOOL_CONFIG_GUID, &guid) == 0); 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); } } /* * Purge vdevs that were dropped */ 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); } } if (oldvdevs) kmem_free(oldvdevs, oldnvdevs * sizeof (void *)); if (sav->sav_config == NULL) goto out; sav->sav_vdevs = newvdevs; sav->sav_count = (int)nl2cache; /* * Recompute the stashed list of l2cache devices, with status * information this time. */ VERIFY(nvlist_remove(sav->sav_config, ZPOOL_CONFIG_L2CACHE, DATA_TYPE_NVLIST_ARRAY) == 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); VERIFY(nvlist_add_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE, l2cache, sav->sav_count) == 0); out: 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 != 0) return (error); nvsize = *(uint64_t *)db->db_data; dmu_buf_rele(db, FTAG); packed = kmem_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); kmem_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); VERIFY(nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP) == 0); 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, 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) { 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); } 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)); if (!spa_has_slogs(spa)) return (B_FALSE); for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; metaslab_group_t *mg = tvd->vdev_mg; if (tvd->vdev_islog) { metaslab_group_passivate(mg); slog_found = B_TRUE; } } return (slog_found); } 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]; metaslab_group_t *mg = tvd->vdev_mg; if (tvd->vdev_islog) metaslab_group_activate(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 { 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_ios--; cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); } /* * Maximum number of concurrent scrub i/os to create while verifying * a pool while importing it. */ int spa_load_verify_maxinflight = 10000; boolean_t spa_load_verify_metadata = B_TRUE; boolean_t spa_load_verify_data = B_TRUE; /*ARGSUSED*/ 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) { if (bp == NULL || BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp)) return (0); /* * 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); if (!BP_IS_METADATA(bp) && !spa_load_verify_data) return (0); zio_t *rio = arg; size_t size = BP_GET_PSIZE(bp); mutex_enter(&spa->spa_scrub_lock); while (spa->spa_load_verify_ios >= spa_load_verify_maxinflight) cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock); spa->spa_load_verify_ios++; 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); } /* ARGSUSED */ int verify_dataset_name_len(dsl_pool_t *dp, dsl_dataset_t *ds, 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) 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); 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, spa_load_verify_cb, rio); } (void) zio_wait(rio); 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; VERIFY(nvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_LOAD_TIME, spa->spa_load_txg_ts) == 0); VERIFY(nvlist_add_int64(spa->spa_load_info, ZPOOL_CONFIG_REWIND_TIME, loss) == 0); VERIFY(nvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_LOAD_DATA_ERRORS, sle.sle_data_count) == 0); } 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)); } 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); ASSERT3P(spa->spa_checkpoint_discard_zthr, ==, NULL); spa->spa_checkpoint_discard_zthr = zthr_create(spa_checkpoint_discard_thread_check, spa_checkpoint_discard_thread, spa); } /* * 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) { char *ereport = FM_EREPORT_ZFS_POOL; int error; spa->spa_load_state = state; 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) { zfs_ereport_post(ereport, spa, NULL, NULL, 0, 0); } } spa->spa_load_state = error ? SPA_LOAD_ERROR : SPA_LOAD_NONE; spa->spa_ena = 0; return (error); } /* * 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 (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); } /* * 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()) 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=%u", import_delay, MMP_FAIL_INT(ub), MMP_INTERVAL(ub), 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=%u", import_delay, MMP_INTERVAL(ub), ub->ub_mmp_delay, import_intervals); } else if (MMP_VALID(ub)) { /* * zfs-0.7 compatability 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=%u leaves=%u", import_delay, ub->ub_mmp_delay, 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=%u", import_delay, multihost_interval, 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 * spa_get_random(250) / 1000; import_expire = gethrtime() + import_delay; while (gethrtime() < import_expire) { 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", txg, ub->ub_txg, timestamp, ub->ub_timestamp, mmp_config, 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) { char *hostname = ""; 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; 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: http://illumos.org/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; char *comment; /* * 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); (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 & FWRITE)) { /* * 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)); } 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() == 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); VERIFY(nvlist_dup(features, &spa->spa_label_features, 0) == 0); } 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; VERIFY(nvlist_alloc(&unsup_feat, NV_UNIQUE_NAME, KM_SLEEP) == 0); 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))) { VERIFY(nvlist_add_string(unsup_feat, nvpair_name(nvp), "") == 0); } } if (!nvlist_empty(unsup_feat)) { VERIFY(nvlist_add_nvlist(spa->spa_load_info, ZPOOL_CONFIG_UNSUP_FEAT, unsup_feat) == 0); 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 */ VERIFY(spa_config_parse(spa, &mrvd, nv, NULL, 0, VDEV_ALLOC_LOAD) == 0); /* * 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)); } 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 secret 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 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) { 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; 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_DEDUPDITTO, &spa->spa_dedup_ditto); 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() == 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)); } /* * 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); 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_verify_logs(spa_t *spa, spa_import_type_t type, 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 asychronously 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) { int 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); /* * 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 = spa_get_random(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, 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 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); /* * 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 all DTLs to see if anything needs resilvering. */ 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"); 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); spa_config_exit(spa, SCL_CONFIG, FTAG); } spa_load_note(spa, "LOADED"); return (0); } static int spa_load_retry(spa_t *spa, spa_load_state_t state) { int 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); 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); 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); 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; 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, 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; *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_owner(&spa_namespace_lock) != curthread) { 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; 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); 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) { VERIFY(nvlist_dup(spa->spa_config, config, KM_SLEEP) == 0); VERIFY(nvlist_add_nvlist(*config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info) == 0); } 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) { VERIFY(nvlist_add_nvlist(*config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info) == 0); } if (locked) { spa->spa_last_open_failed = 0; spa->spa_last_ubsync_txg = 0; spa->spa_load_txg = 0; mutex_exit(&spa_namespace_lock); } *spapp = spa; return (0); } int spa_open_rewind(const char *name, spa_t **spapp, 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, 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; VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); VERIFY(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0); if (nspares != 0) { VERIFY(nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, spares, nspares) == 0); VERIFY(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0); /* * 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++) { VERIFY(nvlist_lookup_uint64(spares[i], ZPOOL_CONFIG_GUID, &guid) == 0); if (spa_spare_exists(guid, &pool, NULL) && pool != 0ULL) { VERIFY(nvlist_lookup_uint64_array( spares[i], ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc) == 0); vs->vs_state = VDEV_STATE_CANT_OPEN; vs->vs_aux = VDEV_AUX_SPARED; } } } } /* * 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; VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); VERIFY(nvlist_lookup_nvlist_array(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0); if (nl2cache != 0) { VERIFY(nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, l2cache, nl2cache) == 0); VERIFY(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0); /* * Update level 2 cache device stats. */ for (i = 0; i < nl2cache; i++) { VERIFY(nvlist_lookup_uint64(l2cache[i], ZPOOL_CONFIG_GUID, &guid) == 0); 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); VERIFY(nvlist_lookup_uint64_array(l2cache[i], ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc) == 0); vdev_get_stats(vd, vs); } } } static void spa_add_feature_stats(spa_t *spa, nvlist_t *config) { nvlist_t *features; zap_cursor_t zc; zap_attribute_t za; ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER)); VERIFY(nvlist_alloc(&features, NV_UNIQUE_NAME, KM_SLEEP) == 0); 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); VERIFY3U(0, ==, 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); VERIFY3U(0, ==, nvlist_add_uint64(features, za.za_name, za.za_first_integer)); } zap_cursor_fini(&zc); } VERIFY(nvlist_add_nvlist(config, ZPOOL_CONFIG_FEATURE_STATS, features) == 0); nvlist_free(features); } 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; VERIFY(nvlist_add_uint64_array(*config, ZPOOL_CONFIG_LOADED_TIME, loadtimes, 2) == 0); VERIFY(nvlist_add_uint64(*config, ZPOOL_CONFIG_ERRCOUNT, spa_get_errlog_size(spa)) == 0); if (spa_suspended(spa)) { VERIFY(nvlist_add_uint64(*config, ZPOOL_CONFIG_SUSPENDED, spa->spa_failmode) == 0); VERIFY(nvlist_add_uint64(*config, ZPOOL_CONFIG_SUSPENDED_REASON, spa->spa_suspended) == 0); } 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) { VERIFY(nvlist_add_uint64(dev[i], ZPOOL_CONFIG_GUID, vd->vdev_guid) == 0); } 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 concatentating with the * current dev list. */ VERIFY(nvlist_lookup_nvlist_array(sav->sav_config, config, &olddevs, &oldndevs) == 0); newdevs = kmem_alloc(sizeof (void *) * (ndevs + oldndevs), KM_SLEEP); for (i = 0; i < oldndevs; i++) VERIFY(nvlist_dup(olddevs[i], &newdevs[i], KM_SLEEP) == 0); for (i = 0; i < ndevs; i++) VERIFY(nvlist_dup(devs[i], &newdevs[i + oldndevs], KM_SLEEP) == 0); VERIFY(nvlist_remove(sav->sav_config, config, DATA_TYPE_NVLIST_ARRAY) == 0); VERIFY(nvlist_add_nvlist_array(sav->sav_config, config, newdevs, ndevs + oldndevs) == 0); for (i = 0; i < oldndevs + ndevs; i++) nvlist_free(newdevs[i]); kmem_free(newdevs, (oldndevs + ndevs) * sizeof (void *)); } else { /* * Generate a new dev list. */ VERIFY(nvlist_alloc(&sav->sav_config, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_nvlist_array(sav->sav_config, config, devs, ndevs) == 0); } } /* * 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); } } /* * Pool Creation */ int spa_create(const char *pool, nvlist_t *nvroot, nvlist_t *props, nvlist_t *zplprops) { spa_t *spa; 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; boolean_t has_features; char *poolname; nvlist_t *nvl; if (props == NULL || nvlist_lookup_string(props, zpool_prop_to_name(ZPOOL_PROP_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; 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; } 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 = 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) { VERIFY(nvlist_alloc(&spa->spa_spares.sav_config, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, spares, nspares) == 0); 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) { VERIFY(nvlist_alloc(&spa->spa_l2cache.sav_config, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_nvlist_array(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, l2cache, nl2cache) == 0); 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, txg); spa->spa_meta_objset = dp->dp_meta_objset; spa->spa_is_initializing = B_FALSE; /* * Create DDTs (dedup tables). */ ddt_create(spa); spa_update_dspace(spa); tx = dmu_tx_create_assigned(dp, txg); /* * 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 (spa_version(spa) >= SPA_VERSION_FEATURES) spa_feature_create_zap_objects(spa, tx); 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)); /* * Create the pool's history object. */ if (version >= SPA_VERSION_ZPOOL_HISTORY) spa_history_create_obj(spa, tx); /* * 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); if (props != NULL) { spa_configfile_set(spa, props, B_FALSE); spa_sync_props(props, tx); } dmu_tx_commit(tx); spa->spa_sync_on = B_TRUE; txg_sync_start(spa->spa_dsl_pool); mmp_thread_start(spa); /* * We explicitly wait for the first transaction to complete so that our * bean counters are appropriately updated. */ txg_wait_synced(spa->spa_dsl_pool, txg); spa_spawn_aux_threads(spa); spa_write_cachefile(spa, B_FALSE, B_TRUE); spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_CREATE); spa_history_log_version(spa, "create"); /* * 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; mutex_exit(&spa_namespace_lock); return (0); } #ifdef _KERNEL /* * Get the root pool information from the root disk, then import the root pool * during the system boot up time. */ extern int vdev_disk_read_rootlabel(char *, char *, nvlist_t **); static nvlist_t * spa_generate_rootconf(char *devpath, char *devid, uint64_t *guid) { nvlist_t *config; nvlist_t *nvtop, *nvroot; uint64_t pgid; if (vdev_disk_read_rootlabel(devpath, devid, &config) != 0) return (NULL); /* * Add this top-level vdev to the child array. */ VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvtop) == 0); VERIFY(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &pgid) == 0); VERIFY(nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID, guid) == 0); /* * Put this pool's top-level vdevs into a root vdev. */ VERIFY(nvlist_alloc(&nvroot, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_string(nvroot, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT) == 0); VERIFY(nvlist_add_uint64(nvroot, ZPOOL_CONFIG_ID, 0ULL) == 0); VERIFY(nvlist_add_uint64(nvroot, ZPOOL_CONFIG_GUID, pgid) == 0); VERIFY(nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &nvtop, 1) == 0); /* * Replace the existing vdev_tree with the new root vdev in * this pool's configuration (remove the old, add the new). */ VERIFY(nvlist_add_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, nvroot) == 0); nvlist_free(nvroot); return (config); } /* * Walk the vdev tree and see if we can find a device with "better" * configuration. A configuration is "better" if the label on that * device has a more recent txg. */ static void spa_alt_rootvdev(vdev_t *vd, vdev_t **avd, uint64_t *txg) { for (int c = 0; c < vd->vdev_children; c++) spa_alt_rootvdev(vd->vdev_child[c], avd, txg); if (vd->vdev_ops->vdev_op_leaf) { nvlist_t *label; uint64_t label_txg; if (vdev_disk_read_rootlabel(vd->vdev_physpath, vd->vdev_devid, &label) != 0) return; VERIFY(nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_TXG, &label_txg) == 0); /* * Do we have a better boot device? */ if (label_txg > *txg) { *txg = label_txg; *avd = vd; } nvlist_free(label); } } /* * Import a root pool. * * For x86. devpath_list will consist of devid and/or physpath name of * the vdev (e.g. "id1,sd@SSEAGATE..." or "/pci@1f,0/ide@d/disk@0,0:a"). * The GRUB "findroot" command will return the vdev we should boot. * * For Sparc, devpath_list consists the physpath name of the booting device * no matter the rootpool is a single device pool or a mirrored pool. * e.g. * "/pci@1f,0/ide@d/disk@0,0:a" */ int spa_import_rootpool(char *devpath, char *devid) { spa_t *spa; vdev_t *rvd, *bvd, *avd = NULL; nvlist_t *config, *nvtop; uint64_t guid, txg; char *pname; int error; /* * Read the label from the boot device and generate a configuration. */ config = spa_generate_rootconf(devpath, devid, &guid); #if defined(_OBP) && defined(_KERNEL) if (config == NULL) { if (strstr(devpath, "/iscsi/ssd") != NULL) { /* iscsi boot */ get_iscsi_bootpath_phy(devpath); config = spa_generate_rootconf(devpath, devid, &guid); } } #endif if (config == NULL) { cmn_err(CE_NOTE, "Cannot read the pool label from '%s'", devpath); return (SET_ERROR(EIO)); } VERIFY(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &pname) == 0); VERIFY(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_TXG, &txg) == 0); mutex_enter(&spa_namespace_lock); if ((spa = spa_lookup(pname)) != NULL) { /* * Remove the existing root pool from the namespace so that we * can replace it with the correct config we just read in. */ spa_remove(spa); } spa = spa_add(pname, config, NULL); spa->spa_is_root = B_TRUE; spa->spa_import_flags = ZFS_IMPORT_VERBATIM; if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &spa->spa_ubsync.ub_version) != 0) spa->spa_ubsync.ub_version = SPA_VERSION_INITIAL; /* * Build up a vdev tree based on the boot device's label config. */ VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvtop) == 0); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); error = spa_config_parse(spa, &rvd, nvtop, NULL, 0, VDEV_ALLOC_ROOTPOOL); spa_config_exit(spa, SCL_ALL, FTAG); if (error) { mutex_exit(&spa_namespace_lock); nvlist_free(config); cmn_err(CE_NOTE, "Can not parse the config for pool '%s'", pname); return (error); } /* * Get the boot vdev. */ if ((bvd = vdev_lookup_by_guid(rvd, guid)) == NULL) { cmn_err(CE_NOTE, "Can not find the boot vdev for guid %llu", (u_longlong_t)guid); error = SET_ERROR(ENOENT); goto out; } /* * Determine if there is a better boot device. */ avd = bvd; spa_alt_rootvdev(rvd, &avd, &txg); if (avd != bvd) { cmn_err(CE_NOTE, "The boot device is 'degraded'. Please " "try booting from '%s'", avd->vdev_path); error = SET_ERROR(EINVAL); goto out; } /* * If the boot device is part of a spare vdev then ensure that * we're booting off the active spare. */ if (bvd->vdev_parent->vdev_ops == &vdev_spare_ops && !bvd->vdev_isspare) { cmn_err(CE_NOTE, "The boot device is currently spared. Please " "try booting from '%s'", bvd->vdev_parent-> vdev_child[bvd->vdev_parent->vdev_children - 1]->vdev_path); error = SET_ERROR(EINVAL); goto out; } error = 0; out: spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); vdev_free(rvd); spa_config_exit(spa, SCL_ALL, FTAG); mutex_exit(&spa_namespace_lock); nvlist_free(config); return (error); } #endif /* * Import a non-root pool into the system. */ int spa_import(const char *pool, nvlist_t *config, nvlist_t *props, uint64_t flags) { spa_t *spa; char *altroot = NULL; spa_load_state_t state = SPA_LOAD_IMPORT; zpool_load_policy_t policy; uint64_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 = FREAD; 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); 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). */ VERIFY(nvlist_add_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info) == 0); 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); } VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); if (error == 0) error = spa_validate_aux(spa, nvroot, -1ULL, VDEV_ALLOC_SPARE); if (error == 0) error = spa_validate_aux(spa, nvroot, -1ULL, VDEV_ALLOC_L2CACHE); 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) VERIFY(nvlist_remove(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, DATA_TYPE_NVLIST_ARRAY) == 0); else VERIFY(nvlist_alloc(&spa->spa_spares.sav_config, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, spares, nspares) == 0); 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) VERIFY(nvlist_remove(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, DATA_TYPE_NVLIST_ARRAY) == 0); else VERIFY(nvlist_alloc(&spa->spa_l2cache.sav_config, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_nvlist_array(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, l2cache, nl2cache) == 0); 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"); spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT); mutex_exit(&spa_namespace_lock); return (0); } nvlist_t * spa_tryimport(nvlist_t *tryconfig) { nvlist_t *config = NULL; 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, FREAD); /* * 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); VERIFY(nvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME, poolname) == 0); VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE, state) == 0); VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_TIMESTAMP, spa->spa_uberblock.ub_timestamp) == 0); VERIFY(nvlist_add_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info) == 0); /* * 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 = 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); } VERIFY(nvlist_add_string(config, ZPOOL_CONFIG_BOOTFS, dsname) == 0); 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(char *pool, int new_state, nvlist_t **oldconfig, boolean_t force, boolean_t hardforce) { spa_t *spa; if (oldconfig) *oldconfig = NULL; if (!(spa_mode_global & FWRITE)) return (SET_ERROR(EROFS)); mutex_enter(&spa_namespace_lock); if ((spa = spa_lookup(pool)) == NULL) { mutex_exit(&spa_namespace_lock); return (SET_ERROR(ENOENT)); } /* * 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); mutex_enter(&spa_namespace_lock); spa_close(spa, FTAG); /* * The pool will be in core if it's openable, * in which case we can modify its state. */ if (spa->spa_state != POOL_STATE_UNINITIALIZED && spa->spa_sync_on) { /* * Objsets may be open only because they're dirty, so we * have to force it to sync before checking spa_refcnt. */ 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 && new_state != POOL_STATE_UNINITIALIZED)) { spa_async_resume(spa); mutex_exit(&spa_namespace_lock); return (SET_ERROR(EBUSY)); } /* * 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)) { spa_async_resume(spa); mutex_exit(&spa_namespace_lock); return (SET_ERROR(EXDEV)); } /* * We're about to export or destroy this pool. Make sure * we stop all initializtion 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. */ if (spa->spa_root_vdev != NULL) { vdev_initialize_stop_all(spa->spa_root_vdev, VDEV_INITIALIZE_ACTIVE); } /* * 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; spa->spa_final_txg = spa_last_synced_txg(spa) + TXG_DEFER_SIZE + 1; vdev_config_dirty(spa->spa_root_vdev); spa_config_exit(spa, SCL_ALL, FTAG); } } spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_DESTROY); if (spa->spa_state != POOL_STATE_UNINITIALIZED) { spa_unload(spa); spa_deactivate(spa); } if (oldconfig && spa->spa_config) VERIFY(nvlist_dup(spa->spa_config, oldconfig, 0) == 0); if (new_state != POOL_STATE_UNINITIALIZED) { if (!hardforce) spa_write_cachefile(spa, B_TRUE, B_TRUE); spa_remove(spa); } mutex_exit(&spa_namespace_lock); return (0); } /* * Destroy a storage pool. */ int spa_destroy(char *pool) { return (spa_export_common(pool, POOL_STATE_DESTROYED, NULL, B_FALSE, B_FALSE)); } /* * Export a storage pool. */ int spa_export(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(char *pool) { return (spa_export_common(pool, POOL_STATE_UNINITIALIZED, NULL, B_FALSE, B_FALSE)); } /* * ========================================================================== * Device manipulation * ========================================================================== */ /* * Add a device to a storage pool. */ int spa_vdev_add(spa_t *spa, nvlist_t *nvroot) { uint64_t txg, id; 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)); /* * 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 toplevels. */ 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 */ if (tvd->vdev_ops == &vdev_raidz_ops) { 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++) { /* * Set the vdev id to the first hole, if one exists. */ for (id = 0; id < rvd->vdev_children; id++) { if (rvd->vdev_child[id]->vdev_ishole) { vdev_free(rvd->vdev_child[id]); break; } } tvd = vd->vdev_child[c]; vdev_remove_child(vd, tvd); tvd->vdev_id = id; 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 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. */ int spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot, int replacing) { 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 (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 ((error = 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)); /* * Spares can't replace logs */ if (oldvd->vdev_top->vdev_islog && newvd->vdev_isspare) 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, EDOM)); /* * 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) sprintf(oldvd->vdev_path, "%s/%s", newvd->vdev_path, "old"); if (oldvd->vdev_devid != NULL) { spa_strfree(oldvd->vdev_devid); oldvd->vdev_devid = NULL; } } /* mark the device being resilvered */ newvd->vdev_resilver_txg = txg; /* * 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); 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 to restart in the future. We do this to * ensure that dmu_sync-ed blocks have been stitched into the - * respective datasets. + * respective datasets. We do not do this if resilvers have been + * deferred. */ - dsl_resilver_restart(spa->spa_dsl_pool, dtl_max_txg); + if (dsl_scan_resilvering(spa_get_dsl(spa)) && + spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER)) + vdev_set_deferred_resilver(spa, newvd); + else + dsl_resilver_restart(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. */ int spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid, int replace_done) { uint64_t txg; int error; vdev_t *rvd = 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_enter(spa); 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 spare, then it implies * that the spare should become a real disk, and be removed from the * active spare list for the pool. */ if (pvd->vdev_ops == &vdev_spare_ops && vd->vdev_id == 0 && pvd->vdev_child[pvd->vdev_children - 1]->vdev_isspare) 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! */ error = 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); 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); /* 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); } int spa_vdev_initialize(spa_t *spa, uint64_t guid, uint64_t cmd_type) { /* * 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); 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); mutex_exit(&spa_namespace_lock); 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); mutex_exit(&spa_namespace_lock); return (SET_ERROR(EINVAL)); } else if (!vdev_writeable(vd)) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); mutex_exit(&spa_namespace_lock); 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_DO && (vd->vdev_initialize_thread != NULL || vd->vdev_top->vdev_removing)) { mutex_exit(&vd->vdev_initialize_lock); mutex_exit(&spa_namespace_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); mutex_exit(&spa_namespace_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); mutex_exit(&spa_namespace_lock); return (SET_ERROR(ESRCH)); } switch (cmd_type) { case POOL_INITIALIZE_DO: vdev_initialize(vd); break; case POOL_INITIALIZE_CANCEL: vdev_initialize_stop(vd, VDEV_INITIALIZE_CANCELED); break; case POOL_INITIALIZE_SUSPEND: vdev_initialize_stop(vd, VDEV_INITIALIZE_SUSPENDED); break; default: panic("invalid cmd_type %llu", (unsigned long long)cmd_type); } mutex_exit(&vd->vdev_initialize_lock); /* Sync out the initializing state */ txg_wait_synced(spa->spa_dsl_pool, 0); mutex_exit(&spa_namespace_lock); return (0); } /* * Split a set of devices from their mirrors, and create a new pool from them. */ int spa_vdev_split_mirror(spa_t *spa, 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; 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 || !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; } } /* 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])) { error = SET_ERROR(EBUSY); break; } /* we need certain info from the top level */ VERIFY(nvlist_add_uint64(child[c], ZPOOL_CONFIG_METASLAB_ARRAY, vml[c]->vdev_top->vdev_ms_array) == 0); VERIFY(nvlist_add_uint64(child[c], ZPOOL_CONFIG_METASLAB_SHIFT, vml[c]->vdev_top->vdev_ms_shift) == 0); VERIFY(nvlist_add_uint64(child[c], ZPOOL_CONFIG_ASIZE, vml[c]->vdev_top->vdev_asize) == 0); VERIFY(nvlist_add_uint64(child[c], ZPOOL_CONFIG_ASHIFT, vml[c]->vdev_top->vdev_ashift) == 0); /* 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. */ VERIFY(nvlist_alloc(&nvl, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_uint64_array(nvl, ZPOOL_CONFIG_SPLIT_LIST, glist, children) == 0); kmem_free(glist, children * sizeof (uint64_t)); mutex_enter(&spa->spa_props_lock); VERIFY(nvlist_add_nvlist(spa->spa_config, ZPOOL_CONFIG_SPLIT, nvl) == 0); mutex_exit(&spa->spa_props_lock); spa->spa_config_splitting = nvl; vdev_config_dirty(spa->spa_root_vdev); /* configure and create the new pool */ VERIFY(nvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME, newname) == 0); VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE, exp ? POOL_STATE_EXPORTED : POOL_STATE_ACTIVE) == 0); VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_VERSION, spa_version(spa)) == 0); VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_TXG, spa->spa_config_txg) == 0); VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_GUID, spa_generate_guid(NULL)) == 0); 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); for (c = 0; c < children; c++) { if (vml[c] != NULL) { /* * Temporarily stop the initializing activity. We set * the state to ACTIVE so that we know to resume * the initializing once the split has completed. */ mutex_enter(&vml[c]->vdev_initialize_lock); vdev_initialize_stop(vml[c], VDEV_INITIALIZE_ACTIVE); mutex_exit(&vml[c]->vdev_initialize_lock); } } newspa->spa_config_source = SPA_CONFIG_SRC_SPLIT; /* 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) { VERIFY(nvlist_alloc(&newspa->spa_config_splitting, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_uint64(newspa->spa_config_splitting, ZPOOL_CONFIG_SPLIT_GUID, spa_guid(spa)) == 0); 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) { 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)); 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 disks as necessary */ spa_async_request(spa, SPA_ASYNC_INITIALIZE_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); } /* * Update the stored path or FRU for this vdev. */ 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); } 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) { sysevent_id_t eid; nvlist_t *attr; char *physpath; 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; physpath = kmem_zalloc(MAXPATHLEN, KM_SLEEP); (void) snprintf(physpath, MAXPATHLEN, "/devices%s", vd->vdev_physpath); VERIFY(nvlist_alloc(&attr, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_string(attr, DEV_PHYS_PATH, physpath) == 0); (void) ddi_log_sysevent(zfs_dip, SUNW_VENDOR, EC_DEV_STATUS, ESC_DEV_DLE, attr, &eid, DDI_SLEEP); nvlist_free(attr); kmem_free(physpath, MAXPATHLEN); } static 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)); 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)); 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), new_space, 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) spa_vdev_resilver_done(spa); /* * Kick off a resilver. */ - if (tasks & SPA_ASYNC_RESILVER) - dsl_resilver_restart(spa->spa_dsl_pool, 0); + if (tasks & SPA_ASYNC_RESILVER && + (!dsl_scan_resilvering(dp) || + !spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER))) + dsl_resilver_restart(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); } /* * 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); } 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); } 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) < (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 && rootdir != 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); } /* * ========================================================================== * SPA syncing routines * ========================================================================== */ static int bpobj_enqueue_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { bpobj_t *bpo = arg; bpobj_enqueue(bpo, bp, tx); return (0); } static int spa_free_sync_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { zio_t *zio = arg; zio_nowait(zio_free_sync(zio, zio->io_spa, dmu_tx_get_txg(tx), bp, zio->io_flags)); return (0); } /* * 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; zio_t *zio = zio_root(spa, NULL, NULL, 0); VERIFY3U(bpobj_iterate(&spa->spa_deferred_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 = kmem_alloc(bufsize, KM_SLEEP); VERIFY(nvlist_pack(nv, &packed, &nvsize, NV_ENCODE_XDR, KM_SLEEP) == 0); bzero(packed + nvsize, bufsize - nvsize); dmu_write(spa->spa_meta_objset, obj, 0, bufsize, packed, tx); kmem_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); } VERIFY(nvlist_alloc(&nvroot, NV_UNIQUE_NAME, KM_SLEEP) == 0); if (sav->sav_count == 0) { VERIFY(nvlist_add_nvlist_array(nvroot, config, NULL, 0) == 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); VERIFY(nvlist_add_nvlist_array(nvroot, config, list, sav->sav_count) == 0); 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_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", 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; char *strval, *fname; zpool_prop_t prop; const char *propname; zprop_type_t proptype; spa_feature_t fid; switch (prop = zpool_name_to_prop(nvpair_name(elem))) { case ZPOOL_PROP_INVAL: /* * We checked this earlier in spa_prop_validate(). */ ASSERT(zpool_prop_feature(nvpair_name(elem))); fname = strchr(nvpair_name(elem), '@') + 1; VERIFY0(zfeature_lookup_name(fname, &fid)); spa_feature_enable(spa, fid, tx); spa_history_log_internal(spa, "set", tx, "%s=enabled", nvpair_name(elem)); break; case ZPOOL_PROP_VERSION: intval = fnvpair_value_uint64(elem); /* * The version is synced seperatly 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-persisitent * 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. It's unnecessary * to do this for pool creation since the vdev's * configuratoin has already been dirtied. */ if (tx->tx_txg != TXG_INITIAL) vdev_config_dirty(spa->spa_root_vdev); spa_history_log_internal(spa, "set", tx, "%s=%s", nvpair_name(elem), strval); break; 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 (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", 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(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", nvpair_name(elem), intval); } else { ASSERT(0); /* not allowed */ } 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_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; case ZPOOL_PROP_DEDUPDITTO: spa->spa_dedup_ditto = intval; break; default: break; } } } 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 = vd->vdev_indirect_mapping; vdev_indirect_births_t *vib = vd->vdev_indirect_births; if (vd->vdev_ops == &vdev_indirect_ops) { ASSERT(vim != NULL); ASSERT(vib != NULL); } if (vdev_obsolete_sm_object(vd) != 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(vdev_obsolete_sm_object(vd), ==, 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 < spa->spa_alloc_count; i++) ASSERT0(zfs_refcount_count( &(mg->mg_alloc_queue_depth[i]))); mg->mg_max_alloc_queue_depth = max_queue_depth; for (int i = 0; i < spa->spa_alloc_count; i++) { mg->mg_cur_max_alloc_queue_depth[i] = 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_alloc_slots[i])); ASSERT0(zfs_refcount_count(&special->mc_alloc_slots[i])); ASSERT0(zfs_refcount_count(&dedup->mc_alloc_slots[i])); normal->mc_alloc_max_slots[i] = slots_per_allocator; special->mc_alloc_max_slots[i] = slots_per_allocator; dedup->mc_alloc_max_slots[i] = 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_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_cb, &spa->spa_deferred_bpobj, tx); } ddt_sync(spa, txg); dsl_scan_sync(dp, tx); svr_sync(spa, tx); spa_sync_upgrades(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 = spa_get_random(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); /* * 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_alloc_locks[i]); VERIFY0(avl_numnodes(&spa->spa_alloc_trees[i])); mutex_exit(&spa->spa_alloc_locks[i]); } /* * 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 (list_head(&spa->spa_state_dirty_list) != NULL) { /* * 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(); VERIFY(cyclic_reprogram(spa->spa_deadman_cycid, spa->spa_sync_starttime + 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); VERIFY(cyclic_reprogram(spa->spa_deadman_cycid, CY_INFINITY)); /* * 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_alloc_locks[i]); VERIFY0(avl_numnodes(&spa->spa_alloc_trees[i])); mutex_exit(&spa->spa_alloc_locks[i]); } /* * Update usable space statistics. */ while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, TXG_CLEAN(txg))) != NULL) vdev_sync_done(vd, txg); spa_update_dspace(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); } boolean_t spa_has_spare(spa_t *spa, uint64_t guid) { int i; uint64_t spareguid; spa_aux_vdev_t *sav = &spa->spa_spares; 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, &spareguid) == 0 && spareguid == guid) return (B_TRUE); } return (B_FALSE); } /* * 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); } sysevent_t * spa_event_create(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name) { sysevent_t *ev = NULL; #ifdef _KERNEL sysevent_attr_list_t *attr = NULL; sysevent_value_t value; ev = sysevent_alloc(EC_ZFS, (char *)name, SUNW_KERN_PUB "zfs", SE_SLEEP); ASSERT(ev != NULL); value.value_type = SE_DATA_TYPE_STRING; value.value.sv_string = spa_name(spa); if (sysevent_add_attr(&attr, ZFS_EV_POOL_NAME, &value, SE_SLEEP) != 0) goto done; value.value_type = SE_DATA_TYPE_UINT64; value.value.sv_uint64 = spa_guid(spa); if (sysevent_add_attr(&attr, ZFS_EV_POOL_GUID, &value, SE_SLEEP) != 0) goto done; if (vd) { value.value_type = SE_DATA_TYPE_UINT64; value.value.sv_uint64 = vd->vdev_guid; if (sysevent_add_attr(&attr, ZFS_EV_VDEV_GUID, &value, SE_SLEEP) != 0) goto done; if (vd->vdev_path) { value.value_type = SE_DATA_TYPE_STRING; value.value.sv_string = vd->vdev_path; if (sysevent_add_attr(&attr, ZFS_EV_VDEV_PATH, &value, SE_SLEEP) != 0) goto done; } } if (hist_nvl != NULL) { fnvlist_merge((nvlist_t *)attr, hist_nvl); } if (sysevent_attach_attributes(ev, attr) != 0) goto done; attr = NULL; done: if (attr) sysevent_free_attr(attr); #endif return (ev); } void spa_event_post(sysevent_t *ev) { #ifdef _KERNEL sysevent_id_t eid; (void) log_sysevent(ev, SE_SLEEP, &eid); sysevent_free(ev); #endif } void spa_event_discard(sysevent_t *ev) { #ifdef _KERNEL sysevent_free(ev); #endif } /* * Post a sysevent corresponding to the given event. 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 and history nvl. 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)); } Index: vendor-sys/illumos/dist/uts/common/fs/zfs/sys/spa_impl.h =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/sys/spa_impl.h (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/sys/spa_impl.h (revision 354961) @@ -1,413 +1,420 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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 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) 2017 Datto Inc. * Copyright (c) 2017, Intel Corporation. */ #ifndef _SYS_SPA_IMPL_H #define _SYS_SPA_IMPL_H #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif typedef struct spa_error_entry { zbookmark_phys_t se_bookmark; char *se_name; avl_node_t se_avl; } spa_error_entry_t; typedef struct spa_history_phys { uint64_t sh_pool_create_len; /* ending offset of zpool create */ uint64_t sh_phys_max_off; /* physical EOF */ uint64_t sh_bof; /* logical BOF */ uint64_t sh_eof; /* logical EOF */ uint64_t sh_records_lost; /* num of records overwritten */ } spa_history_phys_t; /* * All members must be uint64_t, for byteswap purposes. */ typedef struct spa_removing_phys { uint64_t sr_state; /* dsl_scan_state_t */ /* * The vdev ID that we most recently attempted to remove, * or -1 if no removal has been attempted. */ uint64_t sr_removing_vdev; /* * The vdev ID that we most recently successfully removed, * or -1 if no devices have been removed. */ uint64_t sr_prev_indirect_vdev; uint64_t sr_start_time; uint64_t sr_end_time; /* * Note that we can not use the space map's or indirect mapping's * accounting as a substitute for these values, because we need to * count frees of not-yet-copied data as though it did the copy. * Otherwise, we could get into a situation where copied > to_copy, * or we complete before copied == to_copy. */ uint64_t sr_to_copy; /* bytes that need to be copied */ uint64_t sr_copied; /* bytes that have been copied or freed */ } spa_removing_phys_t; /* * This struct is stored as an entry in the DMU_POOL_DIRECTORY_OBJECT * (with key DMU_POOL_CONDENSING_INDIRECT). It is present if a condense * of an indirect vdev's mapping object is in progress. */ typedef struct spa_condensing_indirect_phys { /* * The vdev ID of the indirect vdev whose indirect mapping is * being condensed. */ uint64_t scip_vdev; /* * The vdev's old obsolete spacemap. This spacemap's contents are * being integrated into the new mapping. */ uint64_t scip_prev_obsolete_sm_object; /* * The new mapping object that is being created. */ uint64_t scip_next_mapping_object; } spa_condensing_indirect_phys_t; struct spa_aux_vdev { uint64_t sav_object; /* MOS object for device list */ nvlist_t *sav_config; /* cached device config */ vdev_t **sav_vdevs; /* devices */ int sav_count; /* number devices */ boolean_t sav_sync; /* sync the device list */ nvlist_t **sav_pending; /* pending device additions */ uint_t sav_npending; /* # pending devices */ }; typedef struct spa_config_lock { kmutex_t scl_lock; kthread_t *scl_writer; int scl_write_wanted; kcondvar_t scl_cv; zfs_refcount_t scl_count; } spa_config_lock_t; typedef struct spa_config_dirent { list_node_t scd_link; char *scd_path; } spa_config_dirent_t; typedef enum zio_taskq_type { ZIO_TASKQ_ISSUE = 0, ZIO_TASKQ_ISSUE_HIGH, ZIO_TASKQ_INTERRUPT, ZIO_TASKQ_INTERRUPT_HIGH, ZIO_TASKQ_TYPES } zio_taskq_type_t; /* * State machine for the zpool-poolname process. The states transitions * are done as follows: * * From To Routine * PROC_NONE -> PROC_CREATED spa_activate() * PROC_CREATED -> PROC_ACTIVE spa_thread() * PROC_ACTIVE -> PROC_DEACTIVATE spa_deactivate() * PROC_DEACTIVATE -> PROC_GONE spa_thread() * PROC_GONE -> PROC_NONE spa_deactivate() */ typedef enum spa_proc_state { SPA_PROC_NONE, /* spa_proc = &p0, no process created */ SPA_PROC_CREATED, /* spa_activate() has proc, is waiting */ SPA_PROC_ACTIVE, /* taskqs created, spa_proc set */ SPA_PROC_DEACTIVATE, /* spa_deactivate() requests process exit */ SPA_PROC_GONE /* spa_thread() is exiting, spa_proc = &p0 */ } spa_proc_state_t; typedef struct spa_taskqs { uint_t stqs_count; taskq_t **stqs_taskq; } spa_taskqs_t; typedef enum spa_all_vdev_zap_action { AVZ_ACTION_NONE = 0, AVZ_ACTION_DESTROY, /* Destroy all per-vdev ZAPs and the AVZ. */ AVZ_ACTION_REBUILD, /* Populate the new AVZ, see spa_avz_rebuild */ AVZ_ACTION_INITIALIZE } spa_avz_action_t; typedef enum spa_config_source { SPA_CONFIG_SRC_NONE = 0, SPA_CONFIG_SRC_SCAN, /* scan of path (default: /dev/dsk) */ SPA_CONFIG_SRC_CACHEFILE, /* any cachefile */ SPA_CONFIG_SRC_TRYIMPORT, /* returned from call to tryimport */ SPA_CONFIG_SRC_SPLIT, /* new pool in a pool split */ SPA_CONFIG_SRC_MOS /* MOS, but not always from right txg */ } spa_config_source_t; struct spa { /* * Fields protected by spa_namespace_lock. */ char spa_name[ZFS_MAX_DATASET_NAME_LEN]; /* pool name */ char *spa_comment; /* comment */ avl_node_t spa_avl; /* node in spa_namespace_avl */ nvlist_t *spa_config; /* last synced config */ nvlist_t *spa_config_syncing; /* currently syncing config */ nvlist_t *spa_config_splitting; /* config for splitting */ nvlist_t *spa_load_info; /* info and errors from load */ uint64_t spa_config_txg; /* txg of last config change */ int spa_sync_pass; /* iterate-to-convergence */ pool_state_t spa_state; /* pool state */ int spa_inject_ref; /* injection references */ uint8_t spa_sync_on; /* sync threads are running */ spa_load_state_t spa_load_state; /* current load operation */ boolean_t spa_indirect_vdevs_loaded; /* mappings loaded? */ boolean_t spa_trust_config; /* do we trust vdev tree? */ spa_config_source_t spa_config_source; /* where config comes from? */ uint64_t spa_import_flags; /* import specific flags */ spa_taskqs_t spa_zio_taskq[ZIO_TYPES][ZIO_TASKQ_TYPES]; dsl_pool_t *spa_dsl_pool; boolean_t spa_is_initializing; /* true while opening pool */ metaslab_class_t *spa_normal_class; /* normal data class */ metaslab_class_t *spa_log_class; /* intent log data class */ metaslab_class_t *spa_special_class; /* special allocation class */ metaslab_class_t *spa_dedup_class; /* dedup allocation class */ uint64_t spa_first_txg; /* first txg after spa_open() */ uint64_t spa_final_txg; /* txg of export/destroy */ uint64_t spa_freeze_txg; /* freeze pool at this txg */ uint64_t spa_load_max_txg; /* best initial ub_txg */ uint64_t spa_claim_max_txg; /* highest claimed birth txg */ timespec_t spa_loaded_ts; /* 1st successful open time */ objset_t *spa_meta_objset; /* copy of dp->dp_meta_objset */ kmutex_t spa_evicting_os_lock; /* Evicting objset list lock */ list_t spa_evicting_os_list; /* Objsets being evicted. */ kcondvar_t spa_evicting_os_cv; /* Objset Eviction Completion */ txg_list_t spa_vdev_txg_list; /* per-txg dirty vdev list */ vdev_t *spa_root_vdev; /* top-level vdev container */ int spa_min_ashift; /* of vdevs in normal class */ int spa_max_ashift; /* of vdevs in normal class */ uint64_t spa_config_guid; /* config pool guid */ uint64_t spa_load_guid; /* spa_load initialized guid */ uint64_t spa_last_synced_guid; /* last synced guid */ list_t spa_config_dirty_list; /* vdevs with dirty config */ list_t spa_state_dirty_list; /* vdevs with dirty state */ /* * spa_alloc_locks and spa_alloc_trees are arrays, whose lengths are * stored in spa_alloc_count. There is one tree and one lock for each * allocator, to help improve allocation performance in write-heavy * workloads. */ kmutex_t *spa_alloc_locks; avl_tree_t *spa_alloc_trees; int spa_alloc_count; spa_aux_vdev_t spa_spares; /* hot spares */ spa_aux_vdev_t spa_l2cache; /* L2ARC cache devices */ nvlist_t *spa_label_features; /* Features for reading MOS */ uint64_t spa_config_object; /* MOS object for pool config */ uint64_t spa_config_generation; /* config generation number */ uint64_t spa_syncing_txg; /* txg currently syncing */ bpobj_t spa_deferred_bpobj; /* deferred-free bplist */ bplist_t spa_free_bplist[TXG_SIZE]; /* bplist of stuff to free */ zio_cksum_salt_t spa_cksum_salt; /* secret salt for cksum */ /* checksum context templates */ kmutex_t spa_cksum_tmpls_lock; void *spa_cksum_tmpls[ZIO_CHECKSUM_FUNCTIONS]; uberblock_t spa_ubsync; /* last synced uberblock */ uberblock_t spa_uberblock; /* current uberblock */ boolean_t spa_extreme_rewind; /* rewind past deferred frees */ kmutex_t spa_scrub_lock; /* resilver/scrub lock */ uint64_t spa_scrub_inflight; /* in-flight scrub bytes */ uint64_t spa_load_verify_ios; /* in-flight verification IOs */ kcondvar_t spa_scrub_io_cv; /* scrub I/O completion */ uint8_t spa_scrub_active; /* active or suspended? */ uint8_t spa_scrub_type; /* type of scrub we're doing */ uint8_t spa_scrub_finished; /* indicator to rotate logs */ uint8_t spa_scrub_started; /* started since last boot */ uint8_t spa_scrub_reopen; /* scrub doing vdev_reopen */ uint64_t spa_scan_pass_start; /* start time per pass/reboot */ uint64_t spa_scan_pass_scrub_pause; /* scrub pause time */ uint64_t spa_scan_pass_scrub_spent_paused; /* total paused */ uint64_t spa_scan_pass_exam; /* examined bytes per pass */ uint64_t spa_scan_pass_issued; /* issued bytes per pass */ + + /* + * We are in the middle of a resilver, and another resilver + * is needed once this one completes. This is set iff any + * vdev_resilver_deferred is set. + */ + boolean_t spa_resilver_deferred; kmutex_t spa_async_lock; /* protect async state */ kthread_t *spa_async_thread; /* thread doing async task */ int spa_async_suspended; /* async tasks suspended */ kcondvar_t spa_async_cv; /* wait for thread_exit() */ uint16_t spa_async_tasks; /* async task mask */ uint64_t spa_missing_tvds; /* unopenable tvds on load */ uint64_t spa_missing_tvds_allowed; /* allow loading spa? */ spa_removing_phys_t spa_removing_phys; spa_vdev_removal_t *spa_vdev_removal; spa_condensing_indirect_phys_t spa_condensing_indirect_phys; spa_condensing_indirect_t *spa_condensing_indirect; zthr_t *spa_condense_zthr; /* zthr doing condense. */ uint64_t spa_checkpoint_txg; /* the txg of the checkpoint */ spa_checkpoint_info_t spa_checkpoint_info; /* checkpoint accounting */ zthr_t *spa_checkpoint_discard_zthr; char *spa_root; /* alternate root directory */ uint64_t spa_ena; /* spa-wide ereport ENA */ int spa_last_open_failed; /* error if last open failed */ uint64_t spa_last_ubsync_txg; /* "best" uberblock txg */ uint64_t spa_last_ubsync_txg_ts; /* timestamp from that ub */ uint64_t spa_load_txg; /* ub txg that loaded */ uint64_t spa_load_txg_ts; /* timestamp from that ub */ uint64_t spa_load_meta_errors; /* verify metadata err count */ uint64_t spa_load_data_errors; /* verify data err count */ uint64_t spa_verify_min_txg; /* start txg of verify scrub */ kmutex_t spa_errlog_lock; /* error log lock */ uint64_t spa_errlog_last; /* last error log object */ uint64_t spa_errlog_scrub; /* scrub error log object */ kmutex_t spa_errlist_lock; /* error list/ereport lock */ avl_tree_t spa_errlist_last; /* last error list */ avl_tree_t spa_errlist_scrub; /* scrub error list */ uint64_t spa_deflate; /* should we deflate? */ uint64_t spa_history; /* history object */ kmutex_t spa_history_lock; /* history lock */ vdev_t *spa_pending_vdev; /* pending vdev additions */ kmutex_t spa_props_lock; /* property lock */ uint64_t spa_pool_props_object; /* object for properties */ uint64_t spa_bootfs; /* default boot filesystem */ uint64_t spa_failmode; /* failure mode for the pool */ uint64_t spa_delegation; /* delegation on/off */ list_t spa_config_list; /* previous cache file(s) */ /* per-CPU array of root of async I/O: */ zio_t **spa_async_zio_root; zio_t *spa_suspend_zio_root; /* root of all suspended I/O */ zio_t *spa_txg_zio[TXG_SIZE]; /* spa_sync() waits for this */ kmutex_t spa_suspend_lock; /* protects suspend_zio_root */ kcondvar_t spa_suspend_cv; /* notification of resume */ zio_suspend_reason_t spa_suspended; /* pool is suspended */ uint8_t spa_claiming; /* pool is doing zil_claim() */ boolean_t spa_is_root; /* pool is root */ int spa_minref; /* num refs when first opened */ int spa_mode; /* FREAD | FWRITE */ spa_log_state_t spa_log_state; /* log state */ uint64_t spa_autoexpand; /* lun expansion on/off */ uint64_t spa_bootsize; /* efi system partition size */ ddt_t *spa_ddt[ZIO_CHECKSUM_FUNCTIONS]; /* in-core DDTs */ uint64_t spa_ddt_stat_object; /* DDT statistics */ uint64_t spa_dedup_ditto; /* dedup ditto threshold */ uint64_t spa_dedup_checksum; /* default dedup checksum */ uint64_t spa_dspace; /* dspace in normal class */ kmutex_t spa_vdev_top_lock; /* dueling offline/remove */ kmutex_t spa_proc_lock; /* protects spa_proc* */ kcondvar_t spa_proc_cv; /* spa_proc_state transitions */ spa_proc_state_t spa_proc_state; /* see definition */ struct proc *spa_proc; /* "zpool-poolname" process */ uint64_t spa_did; /* if procp != p0, did of t1 */ boolean_t spa_autoreplace; /* autoreplace set in open */ int spa_vdev_locks; /* locks grabbed */ uint64_t spa_creation_version; /* version at pool creation */ uint64_t spa_prev_software_version; /* See ub_software_version */ uint64_t spa_feat_for_write_obj; /* required to write to pool */ uint64_t spa_feat_for_read_obj; /* required to read from pool */ uint64_t spa_feat_desc_obj; /* Feature descriptions */ uint64_t spa_feat_enabled_txg_obj; /* Feature enabled txg */ /* cache feature refcounts */ uint64_t spa_feat_refcount_cache[SPA_FEATURES]; cyclic_id_t spa_deadman_cycid; /* cyclic id */ uint64_t spa_deadman_calls; /* number of deadman calls */ hrtime_t spa_sync_starttime; /* starting time fo spa_sync */ uint64_t spa_deadman_synctime; /* deadman expiration timer */ uint64_t spa_all_vdev_zaps; /* ZAP of per-vd ZAP obj #s */ spa_avz_action_t spa_avz_action; /* destroy/rebuild AVZ? */ /* * spa_iokstat_lock protects spa_iokstat and * spa_queue_stats[]. */ kmutex_t spa_iokstat_lock; struct kstat *spa_iokstat; /* kstat of io to this pool */ struct { int spa_active; int spa_queued; } spa_queue_stats[ZIO_PRIORITY_NUM_QUEUEABLE]; /* arc_memory_throttle() parameters during low memory condition */ uint64_t spa_lowmem_page_load; /* memory load during txg */ uint64_t spa_lowmem_last_txg; /* txg window start */ hrtime_t spa_ccw_fail_time; /* Conf cache write fail time */ uint64_t spa_multihost; /* multihost aware (mmp) */ mmp_thread_t spa_mmp; /* multihost mmp thread */ list_t spa_leaf_list; /* list of leaf vdevs */ uint64_t spa_leaf_list_gen; /* track leaf_list changes */ /* * spa_refcount & spa_config_lock must be the last elements * because refcount_t changes size based on compilation options. * because zfs_refcount_t changes size based on compilation options. * In order for the MDB module to function correctly, the other * fields must remain in the same location. */ spa_config_lock_t spa_config_lock[SCL_LOCKS]; /* config changes */ zfs_refcount_t spa_refcount; /* number of opens */ }; extern const char *spa_config_path; extern 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); extern void spa_load_spares(spa_t *spa); extern void spa_load_l2cache(spa_t *spa); #ifdef __cplusplus } #endif #endif /* _SYS_SPA_IMPL_H */ Index: vendor-sys/illumos/dist/uts/common/fs/zfs/sys/vdev.h =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/sys/vdev.h (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/sys/vdev.h (revision 354961) @@ -1,190 +1,192 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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, 2017 by Delphix. All rights reserved. * Copyright (c) 2017, Intel Corporation. */ #ifndef _SYS_VDEV_H #define _SYS_VDEV_H #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif typedef enum vdev_dtl_type { DTL_MISSING, /* 0% replication: no copies of the data */ DTL_PARTIAL, /* less than 100% replication: some copies missing */ DTL_SCRUB, /* unable to fully repair during scrub/resilver */ DTL_OUTAGE, /* temporarily missing (used to attempt detach) */ DTL_TYPES } vdev_dtl_type_t; extern boolean_t zfs_nocacheflush; extern void vdev_dbgmsg(vdev_t *vd, const char *fmt, ...); extern void vdev_dbgmsg_print_tree(vdev_t *, int); extern int vdev_open(vdev_t *); extern void vdev_open_children(vdev_t *); extern boolean_t vdev_uses_zvols(vdev_t *); extern int vdev_validate(vdev_t *); extern int vdev_copy_path_strict(vdev_t *, vdev_t *); extern void vdev_copy_path_relaxed(vdev_t *, vdev_t *); extern void vdev_close(vdev_t *); extern int vdev_create(vdev_t *, uint64_t txg, boolean_t isreplace); extern void vdev_reopen(vdev_t *); extern int vdev_validate_aux(vdev_t *vd); extern zio_t *vdev_probe(vdev_t *vd, zio_t *pio); extern boolean_t vdev_is_concrete(vdev_t *vd); extern boolean_t vdev_is_bootable(vdev_t *vd); extern vdev_t *vdev_lookup_top(spa_t *spa, uint64_t vdev); extern vdev_t *vdev_lookup_by_guid(vdev_t *vd, uint64_t guid); extern int vdev_count_leaves(spa_t *spa); extern void vdev_dtl_dirty(vdev_t *vd, vdev_dtl_type_t d, uint64_t txg, uint64_t size); extern boolean_t vdev_dtl_contains(vdev_t *vd, vdev_dtl_type_t d, uint64_t txg, uint64_t size); extern boolean_t vdev_dtl_empty(vdev_t *vd, vdev_dtl_type_t d); extern boolean_t vdev_dtl_need_resilver(vdev_t *vd, uint64_t off, size_t size); extern void vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg, int scrub_done); extern boolean_t vdev_dtl_required(vdev_t *vd); extern boolean_t vdev_resilver_needed(vdev_t *vd, uint64_t *minp, uint64_t *maxp); extern void vdev_destroy_unlink_zap(vdev_t *vd, uint64_t zapobj, dmu_tx_t *tx); extern uint64_t vdev_create_link_zap(vdev_t *vd, dmu_tx_t *tx); extern void vdev_construct_zaps(vdev_t *vd, dmu_tx_t *tx); extern void vdev_destroy_spacemaps(vdev_t *vd, dmu_tx_t *tx); extern void vdev_indirect_mark_obsolete(vdev_t *vd, uint64_t offset, uint64_t size); extern void spa_vdev_indirect_mark_obsolete(spa_t *spa, uint64_t vdev, uint64_t offset, uint64_t size, dmu_tx_t *tx); extern void vdev_hold(vdev_t *); extern void vdev_rele(vdev_t *); extern int vdev_metaslab_init(vdev_t *vd, uint64_t txg); extern void vdev_metaslab_fini(vdev_t *vd); extern void vdev_metaslab_set_size(vdev_t *); extern void vdev_expand(vdev_t *vd, uint64_t txg); extern void vdev_split(vdev_t *vd); extern void vdev_deadman(vdev_t *vd); extern void vdev_get_stats(vdev_t *vd, vdev_stat_t *vs); extern void vdev_clear_stats(vdev_t *vd); extern void vdev_stat_update(zio_t *zio, uint64_t psize); extern void vdev_scan_stat_init(vdev_t *vd); extern void vdev_propagate_state(vdev_t *vd); extern void vdev_set_state(vdev_t *vd, boolean_t isopen, vdev_state_t state, vdev_aux_t aux); extern boolean_t vdev_children_are_offline(vdev_t *vd); extern void vdev_space_update(vdev_t *vd, int64_t alloc_delta, int64_t defer_delta, int64_t space_delta); extern int64_t vdev_deflated_space(vdev_t *vd, int64_t space); extern uint64_t vdev_psize_to_asize(vdev_t *vd, uint64_t psize); extern int vdev_fault(spa_t *spa, uint64_t guid, vdev_aux_t aux); extern int vdev_degrade(spa_t *spa, uint64_t guid, vdev_aux_t aux); extern int vdev_online(spa_t *spa, uint64_t guid, uint64_t flags, vdev_state_t *); extern int vdev_offline(spa_t *spa, uint64_t guid, uint64_t flags); extern void vdev_clear(spa_t *spa, vdev_t *vd); extern boolean_t vdev_is_dead(vdev_t *vd); extern boolean_t vdev_readable(vdev_t *vd); extern boolean_t vdev_writeable(vdev_t *vd); extern boolean_t vdev_allocatable(vdev_t *vd); extern boolean_t vdev_accessible(vdev_t *vd, zio_t *zio); extern boolean_t vdev_is_spacemap_addressable(vdev_t *vd); extern void vdev_cache_init(vdev_t *vd); extern void vdev_cache_fini(vdev_t *vd); extern boolean_t vdev_cache_read(zio_t *zio); extern void vdev_cache_write(zio_t *zio); extern void vdev_cache_purge(vdev_t *vd); extern void vdev_queue_init(vdev_t *vd); extern void vdev_queue_fini(vdev_t *vd); extern zio_t *vdev_queue_io(zio_t *zio); extern void vdev_queue_io_done(zio_t *zio); extern void vdev_queue_change_io_priority(zio_t *zio, zio_priority_t priority); extern int vdev_queue_length(vdev_t *vd); extern uint64_t vdev_queue_last_offset(vdev_t *vd); extern void vdev_config_dirty(vdev_t *vd); extern void vdev_config_clean(vdev_t *vd); extern int vdev_config_sync(vdev_t **svd, int svdcount, uint64_t txg); extern void vdev_state_dirty(vdev_t *vd); extern void vdev_state_clean(vdev_t *vd); +extern void vdev_set_deferred_resilver(spa_t *spa, vdev_t *vd); + typedef enum vdev_config_flag { VDEV_CONFIG_SPARE = 1 << 0, VDEV_CONFIG_L2CACHE = 1 << 1, VDEV_CONFIG_REMOVING = 1 << 2, VDEV_CONFIG_MOS = 1 << 3, VDEV_CONFIG_MISSING = 1 << 4 } vdev_config_flag_t; extern void vdev_top_config_generate(spa_t *spa, nvlist_t *config); extern nvlist_t *vdev_config_generate(spa_t *spa, vdev_t *vd, boolean_t getstats, vdev_config_flag_t flags); /* * Label routines */ struct uberblock; extern uint64_t vdev_label_offset(uint64_t psize, int l, uint64_t offset); extern int vdev_label_number(uint64_t psise, uint64_t offset); extern nvlist_t *vdev_label_read_config(vdev_t *vd, uint64_t txg); extern void vdev_uberblock_load(vdev_t *, struct uberblock *, nvlist_t **); extern 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); typedef enum { VDEV_LABEL_CREATE, /* create/add a new device */ VDEV_LABEL_REPLACE, /* replace an existing device */ VDEV_LABEL_SPARE, /* add a new hot spare */ VDEV_LABEL_REMOVE, /* remove an existing device */ VDEV_LABEL_L2CACHE, /* add an L2ARC cache device */ VDEV_LABEL_SPLIT /* generating new label for split-off dev */ } vdev_labeltype_t; extern int vdev_label_init(vdev_t *vd, uint64_t txg, vdev_labeltype_t reason); #ifdef __cplusplus } #endif #endif /* _SYS_VDEV_H */ Index: vendor-sys/illumos/dist/uts/common/fs/zfs/sys/vdev_impl.h =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/sys/vdev_impl.h (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/sys/vdev_impl.h (revision 354961) @@ -1,523 +1,524 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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) 2017, Intel Corporation. */ #ifndef _SYS_VDEV_IMPL_H #define _SYS_VDEV_IMPL_H #include #include #include #include #include #include #include #include #include #include #include #include #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 int zfs_vdev_queue_depth_pct; extern int zfs_vdev_def_queue_depth; extern uint32_t zfs_vdev_async_write_max_active; /* * Virtual device operations */ typedef int vdev_open_func_t(vdev_t *vd, uint64_t *size, uint64_t *max_size, uint64_t *ashift); typedef void vdev_close_func_t(vdev_t *vd); typedef uint64_t vdev_asize_func_t(vdev_t *vd, uint64_t psize); 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, uint64_t, size_t); 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_seg_t *in, range_seg_t *res); typedef struct vdev_ops { vdev_open_func_t *vdev_op_open; vdev_close_func_t *vdev_op_close; vdev_asize_func_t *vdev_op_asize; 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; /* * For translating ranges from non-leaf vdevs (e.g. raidz) to leaves. * Used when initializing vdevs. Isn't used by leaf ops. */ vdev_xlation_func_t *vdev_op_xlate; 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; uint64_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; uint64_t vq_last_offset; hrtime_t vq_io_complete_ts; /* time last i/o completed */ 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 */ 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 */ vnode_t *vdev_name_vp; /* vnode for pathname */ vnode_t *vdev_devid_vp; /* vnode for devid */ 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 */ boolean_t vdev_expanding; /* expand the vdev? */ boolean_t vdev_reopening; /* reopen in progress? */ boolean_t vdev_nonrot; /* true if solid state */ int vdev_open_error; /* error on last open */ kthread_t *vdev_open_thread; /* thread opening children */ uint64_t vdev_crtxg; /* txg when top-level was added */ /* * 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_t **vdev_ms; /* metaslab array */ 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_removing; /* device is being removed? */ 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 */ boolean_t vdev_initialize_exit_wanted; vdev_initializing_state_t vdev_initialize_state; 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; time_t vdev_initialize_action_time; /* start and end time */ /* for limiting outstanding I/Os */ kmutex_t vdev_initialize_io_lock; kcondvar_t vdev_initialize_io_cv; uint64_t vdev_initialize_inflight; /* * 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_nparity; /* number of parity devices for raidz */ char *vdev_path; /* vdev path (if any) */ char *vdev_devid; /* vdev devid (if any) */ char *vdev_physpath; /* vdev device path (if any) */ 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_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_resilver_deferred; /* resilver deferred */ 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 */ 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 */ }; #define VDEV_RAIDZ_MAXPARITY 3 #define VDEV_PAD_SIZE (8 << 10) /* 2 padding areas (vl_pad1 and vl_pad2) 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 struct vdev_label { char vl_pad1[VDEV_PAD_SIZE]; /* 8K */ char vl_pad2[VDEV_PAD_SIZE]; /* 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_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_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_seg_t *in, range_seg_t *out); extern uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize); extern uint64_t vdev_get_min_asize(vdev_t *vd); extern void vdev_set_min_asize(vdev_t *vd); /* * Global variables */ extern int vdev_standard_sm_blksz; /* zdb uses this tunable, so it must be declared here to make lint happy. */ extern int zfs_vdev_cache_size; /* * 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); extern boolean_t vdev_obsolete_counts_are_precise(vdev_t *vd); /* * Other miscellaneous functions */ int vdev_checkpoint_sm_object(vdev_t *vd); /* * The vdev_buf_t is used to translate between zio_t and buf_t, and back again. */ typedef struct vdev_buf { buf_t vb_buf; /* buffer that describes the io */ zio_t *vb_io; /* pointer back to the original zio_t */ } vdev_buf_t; #ifdef __cplusplus } #endif #endif /* _SYS_VDEV_IMPL_H */ Index: vendor-sys/illumos/dist/uts/common/fs/zfs/vdev.c =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/vdev.c (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/vdev.c (revision 354961) @@ -1,4333 +1,4367 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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 2017 Nexenta Systems, Inc. * Copyright (c) 2014 Integros [integros.com] * Copyright 2016 Toomas Soome * Copyright 2019 Joyent, Inc. * Copyright (c) 2017, Intel Corporation. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* * Virtual device management. */ static vdev_ops_t *vdev_ops_table[] = { &vdev_root_ops, &vdev_raidz_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 }; /* maximum scrub/resilver I/O queue per leaf vdev */ int zfs_scrub_limit = 10; /* default target for number of metaslabs per top-level vdev */ int zfs_vdev_default_ms_count = 200; /* minimum number of metaslabs per top-level vdev */ int zfs_vdev_min_ms_count = 16; /* practical upper limit of total metaslabs per top-level vdev */ int zfs_vdev_ms_count_limit = 1ULL << 17; /* lower limit for metaslab size (512M) */ int zfs_vdev_default_ms_shift = 29; /* upper limit for metaslab size (16G) */ int zfs_vdev_max_ms_shift = 34; boolean_t 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 vdev_dtl_sm_blksz = (1 << 12); /* * 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 vdev_standard_sm_blksz = (1 << 17); int zfs_ashift_min; /*PRINTFLIKE2*/ 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 %u: %s", indent, "", 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), "", (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); } /* * Given a vdev type, return the appropriate ops vector. */ static vdev_ops_t * vdev_getops(const char *type) { vdev_ops_t *ops, **opspp; for (opspp = vdev_ops_table; (ops = *opspp) != NULL; opspp++) if (strcmp(ops->vdev_op_type, type) == 0) break; return (ops); } /* * Derive the enumerated alloction bias from string input. * String origin is either the per-vdev zap or zpool(1M). */ 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); } /* ARGSUSED */ void vdev_default_xlate(vdev_t *vd, const range_seg_t *in, range_seg_t *res) { res->rs_start = in->rs_start; res->rs_end = in->rs_end; } /* * 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); } /* * 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)); /* * The allocatable space for a raidz vdev is N * sizeof(smallest child), * so each child must provide at least 1/Nth of its asize. */ if (pvd->vdev_ops == &vdev_raidz_ops) return ((pvd->vdev_min_asize + pvd->vdev_children - 1) / pvd->vdev_children); return (pvd->vdev_min_asize); } 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]); } 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) { return (vdev_count_leaves_impl(spa->spa_root_vdev)); } void vdev_add_child(vdev_t *pvd, vdev_t *cvd) { size_t oldsize, newsize; uint64_t id = cvd->vdev_id; vdev_t **newchild; spa_t *spa = cvd->vdev_spa; ASSERT(spa_config_held(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_zalloc(newsize, KM_SLEEP); if (pvd->vdev_child != NULL) { bcopy(pvd->vdev_child, newchild, 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); for (int c = newc = 0; c < oldc; c++) if (pvd->vdev_child[c]) newc++; newchild = kmem_alloc(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++; } } 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, NULL); list_link_init(&vd->vdev_leaf_node); mutex_init(&vd->vdev_dtl_lock, NULL, MUTEX_DEFAULT, 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); for (int t = 0; t < DTL_TYPES; t++) { vd->vdev_dtl[t] = range_tree_create(NULL, NULL); } 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; char *type; uint64_t guid = 0, islog, nparity; vdev_t *vd; vdev_indirect_config_t *vic; 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)); /* * Set the nparity property for RAID-Z vdevs. */ nparity = -1ULL; if (ops == &vdev_raidz_ops) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY, &nparity) == 0) { if (nparity == 0 || nparity > VDEV_RAIDZ_MAXPARITY) return (SET_ERROR(EINVAL)); /* * Previous versions could only support 1 or 2 parity * device. */ if (nparity > 1 && spa_version(spa) < SPA_VERSION_RAIDZ2) return (SET_ERROR(ENOTSUP)); if (nparity > 2 && spa_version(spa) < SPA_VERSION_RAIDZ3) return (SET_ERROR(ENOTSUP)); } else { /* * We require the parity to be specified for SPAs that * support multiple parity levels. */ if (spa_version(spa) >= SPA_VERSION_RAIDZ2) return (SET_ERROR(EINVAL)); /* * Otherwise, we default to 1 parity device for RAID-Z. */ nparity = 1; } } else { nparity = 0; } ASSERT(nparity != -1ULL); /* * If creating a top-level vdev, check for allocation classes input */ if (top_level && alloctype == VDEV_ALLOC_ADD) { char *bias; 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 (alloc_bias != VDEV_BIAS_LOG && spa->spa_load_state != SPA_LOAD_CREATE && !spa_feature_is_enabled(spa, SPA_FEATURE_ALLOCATION_CLASSES)) { return (SET_ERROR(ENOTSUP)); } } } vd = vdev_alloc_common(spa, id, guid, ops); vic = &vd->vdev_indirect_config; vd->vdev_islog = islog; vd->vdev_nparity = nparity; if (top_level && alloc_bias != VDEV_BIAS_NONE) vd->vdev_alloc_bias = alloc_bias; if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &vd->vdev_path) == 0) vd->vdev_path = spa_strdup(vd->vdev_path); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &vd->vdev_devid) == 0) vd->vdev_devid = spa_strdup(vd->vdev_devid); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PHYS_PATH, &vd->vdev_physpath) == 0) vd->vdev_physpath = spa_strdup(vd->vdev_physpath); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_FRU, &vd->vdev_fru) == 0) vd->vdev_fru = spa_strdup(vd->vdev_fru); /* * 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; 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 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_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); + if (nvlist_exists(nv, ZPOOL_CONFIG_RESILVER_DEFER)) + vdev_set_deferred_resilver(spa, vd); + /* * 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. Local vdevs will * remain in the faulted state. */ if (spa_load_state(spa) == SPA_LOAD_OPEN) { (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) { 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; } } } /* * 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); /* * 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); ASSERT(vd->vdev_initialize_thread == NULL); /* * Discard allocation state. */ if (vd->vdev_mg != NULL) { vdev_metaslab_fini(vd); metaslab_group_destroy(vd->vdev_mg); } 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_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); 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_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); tvd->vdev_mg = svd->vdev_mg; tvd->vdev_ms = svd->vdev_ms; svd->vdev_mg = NULL; svd->vdev_ms = NULL; if (tvd->vdev_mg != NULL) tvd->vdev_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_removing); tvd->vdev_removing = svd->vdev_removing; 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_removing = 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. */ 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_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; 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; } 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); } static 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); /* * The spa ashift values currently only reflect the * general vdev classes. Class destination is late * binding so ashift checking 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; } } } int vdev_metaslab_init(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; uint64_t m; 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 = kmem_zalloc(newc * sizeof (*mspp), KM_SLEEP); if (expanding) { bcopy(vd->vdev_ms, mspp, oldc * sizeof (*mspp)); kmem_free(vd->vdev_ms, oldc * sizeof (*mspp)); } vd->vdev_ms = mspp; vd->vdev_ms_count = newc; for (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(mos, 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); } } #ifndef _KERNEL /* * To accomodate zdb_leak_init() fake indirect * metaslabs, we allocate a metaslab group for * indirect vdevs which normally don't have one. */ if (vd->vdev_mg == NULL) { ASSERT0(vdev_is_concrete(vd)); vdev_metaslab_group_create(vd); } #endif 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); } } if (txg == 0) spa_config_enter(spa, SCL_ALLOC, FTAG, RW_WRITER); /* * If the vdev is being removed we don't activate * the metaslabs since we want to ensure that no new * allocations are performed on this device. */ if (!expanding && !vd->vdev_removing) { metaslab_group_activate(vd->vdev_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); 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); } kmem_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]); } ASSERT0(vd->vdev_ms_count); } 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; 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"); zfs_ereport_post(FM_EREPORT_ZFS_PROBE_FAILURE, spa, vd, NULL, 0, 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); zio_link_t *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_pad2)), 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_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; } boolean_t vdev_uses_zvols(vdev_t *vd) { if (vd->vdev_path && strncmp(vd->vdev_path, ZVOL_DIR, strlen(ZVOL_DIR)) == 0) return (B_TRUE); for (int c = 0; c < vd->vdev_children; c++) if (vdev_uses_zvols(vd->vdev_child[c])) return (B_TRUE); return (B_FALSE); } void vdev_open_children(vdev_t *vd) { taskq_t *tq; int children = vd->vdev_children; /* * in order to handle pools on top of zvols, do the opens * in a single thread so that the same thread holds the * spa_namespace_lock */ if (vdev_uses_zvols(vd)) { retry_sync: for (int c = 0; c < children; c++) vd->vdev_child[c]->vdev_open_error = vdev_open(vd->vdev_child[c]); } else { tq = taskq_create("vdev_open", children, minclsyspri, children, children, TASKQ_PREPOPULATE); if (tq == NULL) goto retry_sync; for (int c = 0; c < children; c++) VERIFY(taskq_dispatch(tq, vdev_open_child, vd->vdev_child[c], TQ_SLEEP) != TASKQID_INVALID); taskq_destroy(tq); } vd->vdev_nonrot = B_TRUE; for (int c = 0; c < children; c++) vd->vdev_nonrot &= vd->vdev_child[c]->vdev_nonrot; } /* * 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); } } /* * 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 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, &ashift); /* * 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, 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; } 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)); } if (vd->vdev_asize == 0) { /* * This is the first-ever open, so use the computed values. * For testing purposes, a higher ashift can be requested. */ vd->vdev_asize = asize; vd->vdev_max_asize = max_asize; vd->vdev_ashift = MAX(ashift, vd->vdev_ashift); vd->vdev_ashift = MAX(zfs_ashift_min, vd->vdev_ashift); } else { /* * Detect if the alignment requirement has increased. * We don't want to make the pool unavailable, just * issue a warning instead. */ if (ashift > vd->vdev_top->vdev_ashift && vd->vdev_ops->vdev_op_leaf) { cmn_err(CE_WARN, "Disk, '%s', has a block alignment that is " "larger than the pool's alignment\n", vd->vdev_path); } 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 min and max ashift values for normal data devices. * * DJB - TBD these should perhaps be tracked per allocation class * (e.g. spa_min_ashift is used to round up post compression buffers) */ if (vd->vdev_top == vd && vd->vdev_ashift != 0 && vd->vdev_alloc_bias == VDEV_BIAS_NONE && 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; } /* * If a leaf vdev has a DTL, and seems healthy, then kick off a * resilver. 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 && - vdev_resilver_needed(vd, NULL, NULL)) - spa_async_request(spa, SPA_ASYNC_RESILVER); + vdev_resilver_needed(vd, NULL, NULL)) { + if (dsl_scan_resilvering(spa->spa_dsl_pool) && + spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER)) + vdev_set_deferred_resilver(spa, vd); + else + spa_async_request(spa, SPA_ASYNC_RESILVER); + } return (0); } /* * 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; nvlist_t *label; uint64_t guid = 0, aux_guid = 0, top_guid; uint64_t state; nvlist_t *nvl; uint64_t txg; if (vdev_validate_skip) return (0); for (uint64_t c = 0; c < vd->vdev_children; c++) if (vdev_validate(vd->vdev_child[c]) != 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_TRUE, 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) { 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); } } /* * 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) { spa_t *spa = vd->vdev_spa; vdev_t *pvd = vd->vdev_parent; ASSERT(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(vd); } void vdev_rele(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_is_root(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(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 && !l2arc_vdev_present(vd)) l2arc_add_vdev(spa, vd); } 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 : 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; /* BEGIN CSTYLED */ /* * 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. */ /* END CSTYLED */ 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. * Ignore the DTLs and try all devices. This avoids a recursive * mutex enter on the vdev_dtl_lock, and also makes us try hard * when loading the pool (relying on the checksum to ensure that * we get the right data -- note that we while loading, we are * only reading the MOS, which is always checksummed). */ if (vd->vdev_spa->spa_load_state != SPA_LOAD_NONE) return (B_FALSE); 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); } /* * Returns B_TRUE if vdev determines offset needs to be resilvered. */ boolean_t vdev_dtl_need_resilver(vdev_t *vd, uint64_t offset, size_t psize) { 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, offset, psize)); } /* * Returns the lowest txg in the DTL range. */ static uint64_t vdev_dtl_min(vdev_t *vd) { range_seg_t *rs; ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock)); ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0); ASSERT0(vd->vdev_children); rs = avl_first(&vd->vdev_dtl[DTL_MISSING]->rt_root); return (rs->rs_start - 1); } /* * Returns the highest txg in the DTL. */ static uint64_t vdev_dtl_max(vdev_t *vd) { range_seg_t *rs; ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock)); ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0); ASSERT0(vd->vdev_children); rs = avl_last(&vd->vdev_dtl[DTL_MISSING]->rt_root); return (rs->rs_end); } /* * 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) { spa_t *spa = vd->vdev_spa; dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan; ASSERT0(scn->scn_phys.scn_errors); ASSERT0(vd->vdev_children); if (vd->vdev_state < VDEV_STATE_DEGRADED) return (B_FALSE); + if (vd->vdev_resilver_deferred) + return (B_FALSE); + if (vd->vdev_resilver_txg == 0 || range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) 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(scn->scn_phys.scn_min_txg, <=, vdev_dtl_min(vd)); ASSERT3U(scn->scn_phys.scn_min_txg, <, vd->vdev_resilver_txg); ASSERT3U(vd->vdev_resilver_txg, <=, scn->scn_phys.scn_max_txg); return (B_TRUE); } return (B_FALSE); } /* * Reassess DTLs after a config change or scrub completion. */ void vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg, int scrub_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); 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; mutex_enter(&vd->vdev_dtl_lock); /* * If we've completed a scan 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 (scrub_txg != 0 && (spa->spa_scrub_started || (scn != NULL && scn->scn_phys.scn_errors == 0)) && vdev_dtl_should_excise(vd)) { /* * We completed a scrub 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); } 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 and no longer has any * DTLs then reset its resilvering flag. */ if (vd->vdev_resilver_txg != 0 && range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]) && range_tree_is_empty(vd->vdev_dtl[DTL_OUTAGE])) vd->vdev_resilver_txg = 0; 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 (vd->vdev_nparity != 0) minref = vd->vdev_nparity + 1; /* RAID-Z */ 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); } int vdev_dtl_load(vdev_t *vd) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; int error = 0; if (vd->vdev_ops->vdev_op_leaf && vd->vdev_dtl_object != 0) { ASSERT(vdev_is_concrete(vd)); 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); mutex_enter(&vd->vdev_dtl_lock); error = space_map_load(vd->vdev_dtl_sm, vd->vdev_dtl[DTL_MISSING], SM_ALLOC); mutex_exit(&vd->vdev_dtl_lock); 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); } } for (uint64_t i = 0; i < vd->vdev_children; i++) { vdev_construct_zaps(vd->vdev_child[i], tx); } } 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, 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, NULL); 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, 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); required = !vdev_dtl_empty(tvd, DTL_OUTAGE); vd->vdev_cant_read = cant_read; vdev_dtl_reassess(tvd, 0, 0, B_FALSE); if (!required && zio_injection_enabled) required = !!zio_handle_device_injection(vd, NULL, 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. * Returns the spacemap object, or 0 if it wasn't in the ZAP * or the ZAP doesn't exist yet. */ int vdev_checkpoint_sm_object(vdev_t *vd) { ASSERT0(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER)); if (vd->vdev_top_zap == 0) { return (0); } uint64_t sm_obj = 0; int err = zap_lookup(spa_meta_objset(vd->vdev_spa), vd->vdev_top_zap, VDEV_TOP_ZAP_POOL_CHECKPOINT_SM, sizeof (uint64_t), 1, &sm_obj); ASSERT(err == 0 || err == ENOENT); return (sm_obj); } int vdev_load(vdev_t *vd) { int error = 0; /* * Recursively load all children. */ for (int c = 0; c < vd->vdev_children; c++) { error = vdev_load(vd->vdev_child[c]); 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]; if (zap_lookup(spa->spa_meta_objset, vd->vdev_top_zap, VDEV_TOP_ZAP_ALLOCATION_BIAS, 1, sizeof (bias_str), bias_str) == 0) { ASSERT(vd->vdev_alloc_bias == VDEV_BIAS_NONE); vd->vdev_alloc_bias = vdev_derive_alloc_bias(bias_str); } } /* * 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 = vdev_checkpoint_sm_object(vd); if (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; } } /* * 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 = vdev_obsolete_sm_object(vd); if (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); } } 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); } /* * 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)); 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); } 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, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); tvd = vd->vdev_top; /* * 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, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, 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)); } /* * 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, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, 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); } 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); if (wasoffline || (oldstate < VDEV_STATE_DEGRADED && vd->vdev_state >= VDEV_STATE_DEGRADED)) spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_ONLINE); 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, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) 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, 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. */ 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) { 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, 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; for (int c = 0; c < vd->vdev_children; c++) vdev_clear(spa, vd->vdev_child[c]); /* * It makes no sense to "clear" an indirect vdev. */ if (!vdev_is_concrete(vd)) 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 reponse 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; 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 (vd->vdev_aux == NULL && !vdev_is_dead(vd)) - spa_async_request(spa, SPA_ASYNC_RESILVER); + if (vd->vdev_aux == NULL && !vdev_is_dead(vd)) { + if (dsl_scan_resilvering(spa->spa_dsl_pool) && + spa_feature_is_enabled(spa, + SPA_FEATURE_RESILVER_DEFER)) + vdev_set_deferred_resilver(spa, vd); + else + spa_async_request(spa, SPA_ASYNC_RESILVER); + } 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; } 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); } 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. */ void vdev_get_stats(vdev_t *vd, vdev_stat_t *vs) { spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; vdev_t *tvd = vd->vdev_top; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); mutex_enter(&vd->vdev_stat_lock); bcopy(&vd->vdev_stat, vs, 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_rsize += VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE; /* * Report intializing 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 expandable space on top-level, non-auxillary 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 - spa->spa_bootsize, 1ULL << tvd->vdev_ms_shift); } if (vd->vdev_aux == NULL && vd == vd->vdev_top && vdev_is_concrete(vd)) { vs->vs_fragmentation = (vd->vdev_mg != NULL) ? vd->vdev_mg->mg_fragmentation : 0; } + if (vd->vdev_ops->vdev_op_leaf) + vs->vs_resilver_deferred = vd->vdev_resilver_deferred; /* * 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 == rvd) { for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *cvd = rvd->vdev_child[c]; vdev_stat_t *cvs = &cvd->vdev_stat; for (int t = 0; t < 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; } } mutex_exit(&vd->vdev_stat_lock); } 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; vdev_stat_t *vs = &vd->vdev_stat; 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) { if (flags & ZIO_FLAG_SCAN_THREAD) { dsl_scan_phys_t *scn_phys = &spa->spa_dsl_pool->dp_scan->scn_phys; uint64_t *processed = &scn_phys->scn_processed; /* XXX cleanup? */ if (vd->vdev_ops->vdev_op_leaf) atomic_add_64(processed, psize); vs->vs_scan_processed += psize; } if (flags & ZIO_FLAG_SELF_HEAL) vs->vs_self_healed += psize; } vs->vs_ops[type]++; vs->vs_bytes[type] += psize; 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; mutex_enter(&vd->vdev_stat_lock); if (type == ZIO_TYPE_READ && !vdev_is_dead(vd)) { if (zio->io_error == ECKSUM) vs->vs_checksum_errors++; else vs->vs_read_errors++; } if (type == ZIO_TYPE_WRITE && !vdev_is_dead(vd)) vs->vs_write_errors++; mutex_exit(&vd->vdev_stat_lock); if (spa->spa_load_state == SPA_LOAD_NONE && 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 and the root vdev. */ void vdev_space_update(vdev_t *vd, int64_t alloc_delta, int64_t defer_delta, int64_t space_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 * childrens', 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) { 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 we have brought this vdev back into service, we need * to notify fmd so that it can gracefully repair any outstanding * cases due to a missing device. We do this in all cases, even those * that probably don't correlate to a repaired fault. This is sure to * catch all cases, and we let the zfs-retire agent sort it out. If * this is a transient state it's OK, as the retire agent will * double-check the state of the vdev before repairing it. */ if (state == VDEV_STATE_HEALTHY && vd->vdev_ops->vdev_op_leaf && vd->vdev_prevstate != state) zfs_post_state_change(spa, 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; default: class = FM_EREPORT_ZFS_DEVICE_UNKNOWN; } zfs_ereport_post(class, spa, vd, NULL, save_state, 0); } /* Erase any notion of persistent removed state */ vd->vdev_removed = B_FALSE; } else { vd->vdev_removed = B_FALSE; } 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. * In addition, only a single top-level vdev is allowed. */ boolean_t vdev_is_bootable(vdev_t *vd) { if (!vd->vdev_ops->vdev_op_leaf) { char *vdev_type = vd->vdev_ops->vdev_op_type; if (strcmp(vdev_type, VDEV_TYPE_ROOT) == 0 && vd->vdev_children > 1) { return (B_FALSE); } else if (strcmp(vdev_type, VDEV_TYPE_MISSING) == 0 || strcmp(vdev_type, VDEV_TYPE_INDIRECT) == 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; vdev_remove_child(pvd, vd); vdev_compact_children(pvd); 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) { for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; vdev_deadman(cvd); } 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; /* * Look at the head of all the pending queues, * if any I/O has been outstanding for longer than * the spa_deadman_synctime we panic the system. */ fio = avl_first(&vq->vq_active_tree); delta = gethrtime() - fio->io_timestamp; if (delta > spa_deadman_synctime(spa)) { vdev_dbgmsg(vd, "SLOW IO: zio timestamp " "%lluns, delta %lluns, last io %lluns", fio->io_timestamp, (u_longlong_t)delta, vq->vq_io_complete_ts); fm_panic("I/O to pool '%s' appears to be " "hung.", spa_name(spa)); } } mutex_exit(&vq->vq_lock); } +} + +void +vdev_set_deferred_resilver(spa_t *spa, vdev_t *vd) +{ + for (uint64_t i = 0; i < vd->vdev_children; i++) + vdev_set_deferred_resilver(spa, vd->vdev_child[i]); + + if (!vd->vdev_ops->vdev_op_leaf || !vdev_writeable(vd) || + range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) { + return; + } + + vd->vdev_resilver_deferred = B_TRUE; + spa->spa_resilver_deferred = B_TRUE; } Index: vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_indirect.c =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_indirect.c (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_indirect.c (revision 354961) @@ -1,1828 +1,1835 @@ /* * 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. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* * 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. */ boolean_t 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. */ int zfs_indirect_condense_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. */ 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. */ 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. */ int zfs_condense_indirect_commit_entry_delay_ticks = 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. */ int zfs_reconstruct_indirect_combinations_max = 256; /* * Enable to simulate damaged segments and validate reconstruction. * Used by ztest */ 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 */ } 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[1]; /* variable-length */ } 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 = { vdev_indirect_map_free, zio_vsd_default_cksum_report }; /* * 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. */ if (vd->vdev_obsolete_sm == NULL) { ASSERT0(vdev_obsolete_sm_object(vd)); return (B_FALSE); } ASSERT(vd->vdev_obsolete_sm != NULL); ASSERT3U(vdev_obsolete_sm_object(vd), ==, 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_indirect_condense_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)", vd->vdev_id, dmu_tx_get_txg(tx), vic->vic_mapping_object, new_count, 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 = 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, 0, ZFS_SPACE_CHECK_NONE, 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. */ delay(zfs_condense_indirect_commit_entry_delay_ticks); } mapi++; } } /* ARGSUSED */ static boolean_t spa_condense_indirect_thread_check(void *arg, zthr_t *zthr) { spa_t *spa = arg; return (spa->spa_condensing_indirect != NULL); } /* ARGSUSED */ 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 = vdev_obsolete_sm_object(vd); ASSERT(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", vd->vdev_id, 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 = &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)); if (vdev_obsolete_sm_object(vd) == 0) { uint64_t obsolete_sm_object = space_map_alloc(spa->spa_meta_objset, 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)); ASSERT3U(vdev_obsolete_sm_object(vd), !=, 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(vdev_obsolete_sm_object(vd), ==, 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(spa_condense_indirect_thread_check, spa_condense_indirect_thread, spa); } /* * Gets the obsolete spacemap object from the vdev's ZAP. * Returns the spacemap object, or 0 if it wasn't in the ZAP or the ZAP doesn't * exist yet. */ int vdev_obsolete_sm_object(vdev_t *vd) { ASSERT0(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER)); if (vd->vdev_top_zap == 0) { return (0); } uint64_t sm_obj = 0; int err = zap_lookup(vd->vdev_spa->spa_meta_objset, vd->vdev_top_zap, VDEV_TOP_ZAP_INDIRECT_OBSOLETE_SM, sizeof (sm_obj), 1, &sm_obj); ASSERT(err == 0 || err == ENOENT); return (sm_obj); } boolean_t vdev_obsolete_counts_are_precise(vdev_t *vd) { ASSERT0(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER)); if (vd->vdev_top_zap == 0) { return (B_FALSE); } uint64_t val = 0; int err = zap_lookup(vd->vdev_spa->spa_meta_objset, vd->vdev_top_zap, VDEV_TOP_ZAP_OBSOLETE_COUNTS_ARE_PRECISE, sizeof (val), 1, &val); ASSERT(err == 0 || err == ENOENT); return (val != 0); } /* ARGSUSED */ static void vdev_indirect_close(vdev_t *vd) { } /* ARGSUSED */ static int vdev_indirect_open(vdev_t *vd, uint64_t *psize, uint64_t *max_psize, uint64_t *ashift) { *psize = *max_psize = vd->vdev_asize + VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE; *ashift = vd->vdev_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; 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. */ 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); bcopy(first_mapping, duplicate_mappings, 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); vdev_indirect_mapping_t *vim = v->vdev_indirect_mapping; ASSERT3P(vim, !=, 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_put(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; /* * Note, we may read from a child whose DTL * indicates that the data may not be present here. * While this might result in a few i/os that will * likely return incorrect data, it simplifies the * code since we can treat scrub and resilver * identically. (The incorrect data will be * detected and ignored when we verify the * checksum.) */ 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 = 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); 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_flags & (ZIO_FLAG_SCRUB | ZIO_FLAG_RESILVER)) { + 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 { /* - * 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(). + * 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, 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 }; void *bad_buf = abd_borrow_buf_copy(ic->ic_data, is->is_size); abd_t *good_abd = is->is_good_child->ic_data; void *good_buf = abd_borrow_buf_copy(good_abd, is->is_size); zfs_ereport_post_checksum(zio->io_spa, vd, zio, is->is_target_offset, is->is_size, good_buf, bad_buf, &zbc); abd_return_buf(ic->ic_data, bad_buf, is->is_size); abd_return_buf(good_abd, good_buf, is->is_size); } /* * 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. Note that we do this without regard for the DTL's, * 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; enum zio_flag flags = ZIO_FLAG_IO_REPAIR; if (!(zio->io_flags & (ZIO_FLAG_SCRUB | ZIO_FLAG_RESILVER))) flags |= ZIO_FLAG_SELF_HEAL; 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)); 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); zfs_ereport_post_checksum(zio->io_spa, vd, 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 = spa_get_random(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) { /* 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); } } } /* * Set each is_good_child to a randomly-selected child which * is known to contain validated data. */ int 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 16 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, ic->ic_data->abd_size); } iv->iv_attempts_max *= 2; if (iv->iv_attempts_max > (1ULL << 16)) { 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 TRUE when reconstruction is known to be possible. */ if (zfs_reconstruct_indirect_damage_fraction != 0 && spa_get_random(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, is->is_size) == 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_open = vdev_indirect_open, .vdev_op_close = vdev_indirect_close, .vdev_op_asize = vdev_default_asize, .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_type = VDEV_TYPE_INDIRECT, /* name of this vdev type */ .vdev_op_leaf = B_FALSE /* leaf vdev */ }; Index: vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_label.c =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_label.c (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_label.c (revision 354961) @@ -1,1499 +1,1505 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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) 2017, Intel Corporation. * Copyright 2019 Joyent, Inc. */ /* * 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 #include #include #include #include #include #include #include #include #include #include #include #include #include /* * 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)); } 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)); } } /* * 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_fru != NULL) fnvlist_add_string(nv, ZPOOL_CONFIG_FRU, vd->vdev_fru); if (vd->vdev_nparity != 0) { ASSERT(strcmp(vd->vdev_ops->vdev_op_type, VDEV_TYPE_RAIDZ) == 0); /* * Make sure someone hasn't managed to sneak a fancy new vdev * into a crufty old storage pool. */ ASSERT(vd->vdev_nparity == 1 || (vd->vdev_nparity <= 2 && spa_version(spa) >= SPA_VERSION_RAIDZ2) || (vd->vdev_nparity <= 3 && spa_version(spa) >= SPA_VERSION_RAIDZ3)); /* * Note that we'll add the nparity tag even on storage pools * that only support a single parity device -- older software * will just ignore it. */ fnvlist_add_uint64(nv, ZPOOL_CONFIG_NPARITY, vd->vdev_nparity); } 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_removing) { 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 (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_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_stat_t vs; vdev_get_stats(vd, &vs); fnvlist_add_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t *)&vs, sizeof (vs) / sizeof (uint64_t)); root_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 (1ULL << (i + 1) < vdev_removal_max_span) { 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 / zfs_remove_max_segment; 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; int c, idx; ASSERT(!vd->vdev_ishole); child = kmem_alloc(vd->vdev_children * sizeof (nvlist_t *), KM_SLEEP); for (c = 0, idx = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; /* * If we're generating an nvlist of removing * vdevs then skip over any device which is * not being removed. */ if ((flags & VDEV_CONFIG_REMOVING) && !cvd->vdev_removing) continue; child[idx++] = vdev_config_generate(spa, cvd, getstats, flags); } if (idx) { fnvlist_add_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, child, idx); } for (c = 0; c < idx; 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_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); 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) fnvlist_add_string(nv, ZPOOL_CONFIG_AUX_STATE, aux); 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; abd_t *vp_abd; zio_t *zio; 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(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); if (!vdev_readable(vd)) return (NULL); vp_abd = abd_alloc_linear(sizeof (vdev_phys_t), B_TRUE); vp = abd_to_buf(vp_abd); retry: for (int l = 0; l < VDEV_LABELS; l++) { nvlist_t *label = NULL; zio = zio_root(spa, NULL, NULL, flags); vdev_label_read(zio, vd, l, vp_abd, offsetof(vdev_label_t, vl_vdev_phys), sizeof (vdev_phys_t), NULL, NULL, flags); if (zio_wait(zio) == 0 && nvlist_unpack(vp->vp_nvlist, sizeof (vp->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; 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); } abd_free(vp_abd); 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. We also check if it is an l2cache device. */ 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: case VDEV_LABEL_L2CACHE: 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)); } } /* * Check to see if this is an l2cache device. */ if (spa_l2cache_exists(device_guid, NULL)) return (B_TRUE); /* * 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) == FREAD) 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 *pad2; uberblock_t *ub; abd_t *ub_abd; zio_t *zio; char *buf; size_t buflen; int error; uint64_t spare_guid, l2cache_guid; 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 (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. */ pad2 = abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE); abd_zero(pad2, 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, pad2, offsetof(vdev_label_t, vl_pad2), 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(pad2); 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); } /* * ========================================================================== * 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 = AVL_CMP(ub1->ub_txg, ub2->ub_txg); if (likely(cmp)) return (cmp); cmp = AVL_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. zfsonlinux >= 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 (AVL_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)) { 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); bzero(ub, 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); } /* * 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; 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[] */ 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; /* * 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) == 0) { 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); } 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 = kmem_zalloc(sizeof (uint64_t), KM_SLEEP); ASSERT(!vd->vdev_ishole); 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); } Index: vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_removal.c =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_removal.c (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/vdev_removal.c (revision 354961) @@ -1,2158 +1,2158 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* * This file contains the necessary logic to remove vdevs from a * storage pool. Currently, the only devices that can be removed * are log, cache, and spare devices; and top level vdevs from a pool * w/o raidz. (Note that members of a mirror can also be removed * by the detach operation.) * * Log vdevs are removed by evacuating them and then turning the vdev * into a hole vdev while holding spa config locks. * * Top level vdevs are removed and converted into an indirect vdev via * a multi-step process: * * - Disable allocations from this device (spa_vdev_remove_top). * * - From a new thread (spa_vdev_remove_thread), copy data from * the removing vdev to a different vdev. The copy happens in open * context (spa_vdev_copy_impl) and issues a sync task * (vdev_mapping_sync) so the sync thread can update the partial * indirect mappings in core and on disk. * * - If a free happens during a removal, it is freed from the * removing vdev, and if it has already been copied, from the new * location as well (free_from_removing_vdev). * * - After the removal is completed, the copy thread converts the vdev * into an indirect vdev (vdev_remove_complete) before instructing * the sync thread to destroy the space maps and finish the removal * (spa_finish_removal). */ typedef struct vdev_copy_arg { metaslab_t *vca_msp; uint64_t vca_outstanding_bytes; kcondvar_t vca_cv; kmutex_t vca_lock; } vdev_copy_arg_t; /* * The maximum amount of memory we can use for outstanding i/o while * doing a device removal. This determines how much i/o we can have * in flight concurrently. */ int zfs_remove_max_copy_bytes = 64 * 1024 * 1024; /* * The largest contiguous segment that we will attempt to allocate when * removing a device. This can be no larger than SPA_MAXBLOCKSIZE. If * there is a performance problem with attempting to allocate large blocks, * consider decreasing this. * * Note: we will issue I/Os of up to this size. The mpt driver does not * respond well to I/Os larger than 1MB, so we set this to 1MB. (When * mpt processes an I/O larger than 1MB, it needs to do an allocation of * 2 physically contiguous pages; if this allocation fails, mpt will drop * the I/O and hang the device.) */ int zfs_remove_max_segment = 1024 * 1024; /* * Allow a remap segment to span free chunks of at most this size. The main * impact of a larger span is that we will read and write larger, more * contiguous chunks, with more "unnecessary" data -- trading off bandwidth * for iops. The value here was chosen to align with * 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). * * Additionally, a higher span will have the following relatively minor * effects: * - the mapping will be smaller, since one entry can cover more allocated * segments * - more of the fragmentation in the removing device will be preserved * - we'll do larger allocations, which may fail and fall back on smaller * allocations */ int vdev_removal_max_span = 32 * 1024; /* * This is used by the test suite so that it can ensure that certain * actions happen while in the middle of a removal. */ -uint64_t zfs_remove_max_bytes_pause = UINT64_MAX; +int zfs_removal_suspend_progress = 0; #define VDEV_REMOVAL_ZAP_OBJS "lzap" static void spa_vdev_remove_thread(void *arg); static void spa_sync_removing_state(spa_t *spa, dmu_tx_t *tx) { VERIFY0(zap_update(spa->spa_dsl_pool->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_REMOVING, sizeof (uint64_t), sizeof (spa->spa_removing_phys) / sizeof (uint64_t), &spa->spa_removing_phys, tx)); } static nvlist_t * spa_nvlist_lookup_by_guid(nvlist_t **nvpp, int count, uint64_t target_guid) { for (int i = 0; i < count; i++) { uint64_t guid = fnvlist_lookup_uint64(nvpp[i], ZPOOL_CONFIG_GUID); if (guid == target_guid) return (nvpp[i]); } return (NULL); } static void spa_vdev_remove_aux(nvlist_t *config, char *name, nvlist_t **dev, int count, nvlist_t *dev_to_remove) { nvlist_t **newdev = NULL; if (count > 1) newdev = kmem_alloc((count - 1) * sizeof (void *), KM_SLEEP); for (int i = 0, j = 0; i < count; i++) { if (dev[i] == dev_to_remove) continue; VERIFY(nvlist_dup(dev[i], &newdev[j++], KM_SLEEP) == 0); } VERIFY(nvlist_remove(config, name, DATA_TYPE_NVLIST_ARRAY) == 0); VERIFY(nvlist_add_nvlist_array(config, name, newdev, count - 1) == 0); for (int i = 0; i < count - 1; i++) nvlist_free(newdev[i]); if (count > 1) kmem_free(newdev, (count - 1) * sizeof (void *)); } static spa_vdev_removal_t * spa_vdev_removal_create(vdev_t *vd) { spa_vdev_removal_t *svr = kmem_zalloc(sizeof (*svr), KM_SLEEP); mutex_init(&svr->svr_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&svr->svr_cv, NULL, CV_DEFAULT, NULL); svr->svr_allocd_segs = range_tree_create(NULL, NULL); svr->svr_vdev_id = vd->vdev_id; for (int i = 0; i < TXG_SIZE; i++) { svr->svr_frees[i] = range_tree_create(NULL, NULL); list_create(&svr->svr_new_segments[i], sizeof (vdev_indirect_mapping_entry_t), offsetof(vdev_indirect_mapping_entry_t, vime_node)); } return (svr); } void spa_vdev_removal_destroy(spa_vdev_removal_t *svr) { for (int i = 0; i < TXG_SIZE; i++) { ASSERT0(svr->svr_bytes_done[i]); ASSERT0(svr->svr_max_offset_to_sync[i]); range_tree_destroy(svr->svr_frees[i]); list_destroy(&svr->svr_new_segments[i]); } range_tree_destroy(svr->svr_allocd_segs); mutex_destroy(&svr->svr_lock); cv_destroy(&svr->svr_cv); kmem_free(svr, sizeof (*svr)); } /* * This is called as a synctask in the txg in which we will mark this vdev * as removing (in the config stored in the MOS). * * It begins the evacuation of a toplevel vdev by: * - initializing the spa_removing_phys which tracks this removal * - computing the amount of space to remove for accounting purposes * - dirtying all dbufs in the spa_config_object * - creating the spa_vdev_removal * - starting the spa_vdev_remove_thread */ static void vdev_remove_initiate_sync(void *arg, dmu_tx_t *tx) { int vdev_id = (uintptr_t)arg; spa_t *spa = dmu_tx_pool(tx)->dp_spa; vdev_t *vd = vdev_lookup_top(spa, vdev_id); vdev_indirect_config_t *vic = &vd->vdev_indirect_config; objset_t *mos = spa->spa_dsl_pool->dp_meta_objset; spa_vdev_removal_t *svr = NULL; uint64_t txg = dmu_tx_get_txg(tx); ASSERT3P(vd->vdev_ops, !=, &vdev_raidz_ops); svr = spa_vdev_removal_create(vd); ASSERT(vd->vdev_removing); ASSERT3P(vd->vdev_indirect_mapping, ==, NULL); spa_feature_incr(spa, SPA_FEATURE_DEVICE_REMOVAL, tx); if (spa_feature_is_enabled(spa, SPA_FEATURE_OBSOLETE_COUNTS)) { /* * By activating the OBSOLETE_COUNTS feature, we prevent * the pool from being downgraded and ensure that the * refcounts are precise. */ spa_feature_incr(spa, SPA_FEATURE_OBSOLETE_COUNTS, tx); uint64_t one = 1; VERIFY0(zap_add(spa->spa_meta_objset, vd->vdev_top_zap, VDEV_TOP_ZAP_OBSOLETE_COUNTS_ARE_PRECISE, sizeof (one), 1, &one, tx)); ASSERT3U(vdev_obsolete_counts_are_precise(vd), !=, 0); } vic->vic_mapping_object = vdev_indirect_mapping_alloc(mos, tx); vd->vdev_indirect_mapping = vdev_indirect_mapping_open(mos, vic->vic_mapping_object); vic->vic_births_object = vdev_indirect_births_alloc(mos, tx); vd->vdev_indirect_births = vdev_indirect_births_open(mos, vic->vic_births_object); spa->spa_removing_phys.sr_removing_vdev = vd->vdev_id; spa->spa_removing_phys.sr_start_time = gethrestime_sec(); spa->spa_removing_phys.sr_end_time = 0; spa->spa_removing_phys.sr_state = DSS_SCANNING; spa->spa_removing_phys.sr_to_copy = 0; spa->spa_removing_phys.sr_copied = 0; /* * Note: We can't use vdev_stat's vs_alloc for sr_to_copy, because * there may be space in the defer tree, which is free, but still * counted in vs_alloc. */ for (uint64_t i = 0; i < vd->vdev_ms_count; i++) { metaslab_t *ms = vd->vdev_ms[i]; if (ms->ms_sm == NULL) continue; spa->spa_removing_phys.sr_to_copy += metaslab_allocated_space(ms); /* * Space which we are freeing this txg does not need to * be copied. */ spa->spa_removing_phys.sr_to_copy -= range_tree_space(ms->ms_freeing); ASSERT0(range_tree_space(ms->ms_freed)); for (int t = 0; t < TXG_SIZE; t++) ASSERT0(range_tree_space(ms->ms_allocating[t])); } /* * Sync tasks are called before metaslab_sync(), so there should * be no already-synced metaslabs in the TXG_CLEAN list. */ ASSERT3P(txg_list_head(&vd->vdev_ms_list, TXG_CLEAN(txg)), ==, NULL); spa_sync_removing_state(spa, tx); /* * All blocks that we need to read the most recent mapping must be * stored on concrete vdevs. Therefore, we must dirty anything that * is read before spa_remove_init(). Specifically, the * spa_config_object. (Note that although we already modified the * spa_config_object in spa_sync_removing_state, that may not have * modified all blocks of the object.) */ dmu_object_info_t doi; VERIFY0(dmu_object_info(mos, DMU_POOL_DIRECTORY_OBJECT, &doi)); for (uint64_t offset = 0; offset < doi.doi_max_offset; ) { dmu_buf_t *dbuf; VERIFY0(dmu_buf_hold(mos, DMU_POOL_DIRECTORY_OBJECT, offset, FTAG, &dbuf, 0)); dmu_buf_will_dirty(dbuf, tx); offset += dbuf->db_size; dmu_buf_rele(dbuf, FTAG); } /* * Now that we've allocated the im_object, dirty the vdev to ensure * that the object gets written to the config on disk. */ vdev_config_dirty(vd); zfs_dbgmsg("starting removal thread for vdev %llu (%p) in txg %llu " "im_obj=%llu", vd->vdev_id, vd, dmu_tx_get_txg(tx), vic->vic_mapping_object); spa_history_log_internal(spa, "vdev remove started", tx, "%s vdev %llu %s", spa_name(spa), vd->vdev_id, (vd->vdev_path != NULL) ? vd->vdev_path : "-"); /* * Setting spa_vdev_removal causes subsequent frees to call * free_from_removing_vdev(). Note that we don't need any locking * because we are the sync thread, and metaslab_free_impl() is only * called from syncing context (potentially from a zio taskq thread, * but in any case only when there are outstanding free i/os, which * there are not). */ ASSERT3P(spa->spa_vdev_removal, ==, NULL); spa->spa_vdev_removal = svr; svr->svr_thread = thread_create(NULL, 0, spa_vdev_remove_thread, spa, 0, &p0, TS_RUN, minclsyspri); } /* * When we are opening a pool, we must read the mapping for each * indirect vdev in order from most recently removed to least * recently removed. We do this because the blocks for the mapping * of older indirect vdevs may be stored on more recently removed vdevs. * In order to read each indirect mapping object, we must have * initialized all more recently removed vdevs. */ int spa_remove_init(spa_t *spa) { int error; error = zap_lookup(spa->spa_dsl_pool->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_REMOVING, sizeof (uint64_t), sizeof (spa->spa_removing_phys) / sizeof (uint64_t), &spa->spa_removing_phys); if (error == ENOENT) { 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; return (0); } else if (error != 0) { return (error); } if (spa->spa_removing_phys.sr_state == DSS_SCANNING) { /* * We are currently removing a vdev. Create and * initialize a spa_vdev_removal_t from the bonus * buffer of the removing vdevs vdev_im_object, and * initialize its partial mapping. */ spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); vdev_t *vd = vdev_lookup_top(spa, spa->spa_removing_phys.sr_removing_vdev); if (vd == NULL) { spa_config_exit(spa, SCL_STATE, FTAG); return (EINVAL); } vdev_indirect_config_t *vic = &vd->vdev_indirect_config; ASSERT(vdev_is_concrete(vd)); spa_vdev_removal_t *svr = spa_vdev_removal_create(vd); ASSERT3U(svr->svr_vdev_id, ==, vd->vdev_id); ASSERT(vd->vdev_removing); vd->vdev_indirect_mapping = vdev_indirect_mapping_open( spa->spa_meta_objset, vic->vic_mapping_object); vd->vdev_indirect_births = vdev_indirect_births_open( spa->spa_meta_objset, vic->vic_births_object); spa_config_exit(spa, SCL_STATE, FTAG); spa->spa_vdev_removal = svr; } spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); uint64_t indirect_vdev_id = spa->spa_removing_phys.sr_prev_indirect_vdev; while (indirect_vdev_id != UINT64_MAX) { vdev_t *vd = vdev_lookup_top(spa, indirect_vdev_id); vdev_indirect_config_t *vic = &vd->vdev_indirect_config; ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops); vd->vdev_indirect_mapping = vdev_indirect_mapping_open( spa->spa_meta_objset, vic->vic_mapping_object); vd->vdev_indirect_births = vdev_indirect_births_open( spa->spa_meta_objset, vic->vic_births_object); indirect_vdev_id = vic->vic_prev_indirect_vdev; } spa_config_exit(spa, SCL_STATE, FTAG); /* * Now that we've loaded all the indirect mappings, we can allow * reads from other blocks (e.g. via predictive prefetch). */ spa->spa_indirect_vdevs_loaded = B_TRUE; return (0); } void spa_restart_removal(spa_t *spa) { spa_vdev_removal_t *svr = spa->spa_vdev_removal; if (svr == NULL) return; /* * In general when this function is called there is no * removal thread running. The only scenario where this * is not true is during spa_import() where this function * is called twice [once from spa_import_impl() and * spa_async_resume()]. Thus, in the scenario where we * import a pool that has an ongoing removal we don't * want to spawn a second thread. */ if (svr->svr_thread != NULL) return; if (!spa_writeable(spa)) return; zfs_dbgmsg("restarting removal of %llu", svr->svr_vdev_id); svr->svr_thread = thread_create(NULL, 0, spa_vdev_remove_thread, spa, 0, &p0, TS_RUN, minclsyspri); } /* * Process freeing from a device which is in the middle of being removed. * We must handle this carefully so that we attempt to copy freed data, * and we correctly free already-copied data. */ void free_from_removing_vdev(vdev_t *vd, uint64_t offset, uint64_t size) { spa_t *spa = vd->vdev_spa; spa_vdev_removal_t *svr = spa->spa_vdev_removal; vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping; uint64_t txg = spa_syncing_txg(spa); uint64_t max_offset_yet = 0; ASSERT(vd->vdev_indirect_config.vic_mapping_object != 0); ASSERT3U(vd->vdev_indirect_config.vic_mapping_object, ==, vdev_indirect_mapping_object(vim)); ASSERT3U(vd->vdev_id, ==, svr->svr_vdev_id); mutex_enter(&svr->svr_lock); /* * Remove the segment from the removing vdev's spacemap. This * ensures that we will not attempt to copy this space (if the * removal thread has not yet visited it), and also ensures * that we know what is actually allocated on the new vdevs * (needed if we cancel the removal). * * Note: we must do the metaslab_free_concrete() with the svr_lock * held, so that the remove_thread can not load this metaslab and then * visit this offset between the time that we metaslab_free_concrete() * and when we check to see if it has been visited. * * Note: The checkpoint flag is set to false as having/taking * a checkpoint and removing a device can't happen at the same * time. */ ASSERT(!spa_has_checkpoint(spa)); metaslab_free_concrete(vd, offset, size, B_FALSE); uint64_t synced_size = 0; uint64_t synced_offset = 0; uint64_t max_offset_synced = vdev_indirect_mapping_max_offset(vim); if (offset < max_offset_synced) { /* * The mapping for this offset is already on disk. * Free from the new location. * * Note that we use svr_max_synced_offset because it is * updated atomically with respect to the in-core mapping. * By contrast, vim_max_offset is not. * * This block may be split between a synced entry and an * in-flight or unvisited entry. Only process the synced * portion of it here. */ synced_size = MIN(size, max_offset_synced - offset); synced_offset = offset; ASSERT3U(max_offset_yet, <=, max_offset_synced); max_offset_yet = max_offset_synced; DTRACE_PROBE3(remove__free__synced, spa_t *, spa, uint64_t, offset, uint64_t, synced_size); size -= synced_size; offset += synced_size; } /* * Look at all in-flight txgs starting from the currently syncing one * and see if a section of this free is being copied. By starting from * this txg and iterating forward, we might find that this region * was copied in two different txgs and handle it appropriately. */ for (int i = 0; i < TXG_CONCURRENT_STATES; i++) { int txgoff = (txg + i) & TXG_MASK; if (size > 0 && offset < svr->svr_max_offset_to_sync[txgoff]) { /* * The mapping for this offset is in flight, and * will be synced in txg+i. */ uint64_t inflight_size = MIN(size, svr->svr_max_offset_to_sync[txgoff] - offset); DTRACE_PROBE4(remove__free__inflight, spa_t *, spa, uint64_t, offset, uint64_t, inflight_size, uint64_t, txg + i); /* * We copy data in order of increasing offset. * Therefore the max_offset_to_sync[] must increase * (or be zero, indicating that nothing is being * copied in that txg). */ if (svr->svr_max_offset_to_sync[txgoff] != 0) { ASSERT3U(svr->svr_max_offset_to_sync[txgoff], >=, max_offset_yet); max_offset_yet = svr->svr_max_offset_to_sync[txgoff]; } /* * We've already committed to copying this segment: * we have allocated space elsewhere in the pool for * it and have an IO outstanding to copy the data. We * cannot free the space before the copy has * completed, or else the copy IO might overwrite any * new data. To free that space, we record the * segment in the appropriate svr_frees tree and free * the mapped space later, in the txg where we have * completed the copy and synced the mapping (see * vdev_mapping_sync). */ range_tree_add(svr->svr_frees[txgoff], offset, inflight_size); size -= inflight_size; offset += inflight_size; /* * This space is already accounted for as being * done, because it is being copied in txg+i. * However, if i!=0, then it is being copied in * a future txg. If we crash after this txg * syncs but before txg+i syncs, then the space * will be free. Therefore we must account * for the space being done in *this* txg * (when it is freed) rather than the future txg * (when it will be copied). */ ASSERT3U(svr->svr_bytes_done[txgoff], >=, inflight_size); svr->svr_bytes_done[txgoff] -= inflight_size; svr->svr_bytes_done[txg & TXG_MASK] += inflight_size; } } ASSERT0(svr->svr_max_offset_to_sync[TXG_CLEAN(txg) & TXG_MASK]); if (size > 0) { /* * The copy thread has not yet visited this offset. Ensure * that it doesn't. */ DTRACE_PROBE3(remove__free__unvisited, spa_t *, spa, uint64_t, offset, uint64_t, size); if (svr->svr_allocd_segs != NULL) range_tree_clear(svr->svr_allocd_segs, offset, size); /* * Since we now do not need to copy this data, for * accounting purposes we have done our job and can count * it as completed. */ svr->svr_bytes_done[txg & TXG_MASK] += size; } mutex_exit(&svr->svr_lock); /* * Now that we have dropped svr_lock, process the synced portion * of this free. */ if (synced_size > 0) { vdev_indirect_mark_obsolete(vd, synced_offset, synced_size); /* * Note: this can only be called from syncing context, * and the vdev_indirect_mapping is only changed from the * sync thread, so we don't need svr_lock while doing * metaslab_free_impl_cb. */ boolean_t checkpoint = B_FALSE; vdev_indirect_ops.vdev_op_remap(vd, synced_offset, synced_size, metaslab_free_impl_cb, &checkpoint); } } /* * Stop an active removal and update the spa_removing phys. */ static void spa_finish_removal(spa_t *spa, dsl_scan_state_t state, dmu_tx_t *tx) { spa_vdev_removal_t *svr = spa->spa_vdev_removal; ASSERT3U(dmu_tx_get_txg(tx), ==, spa_syncing_txg(spa)); /* Ensure the removal thread has completed before we free the svr. */ spa_vdev_remove_suspend(spa); ASSERT(state == DSS_FINISHED || state == DSS_CANCELED); if (state == DSS_FINISHED) { spa_removing_phys_t *srp = &spa->spa_removing_phys; vdev_t *vd = vdev_lookup_top(spa, svr->svr_vdev_id); vdev_indirect_config_t *vic = &vd->vdev_indirect_config; if (srp->sr_prev_indirect_vdev != UINT64_MAX) { vdev_t *pvd = vdev_lookup_top(spa, srp->sr_prev_indirect_vdev); ASSERT3P(pvd->vdev_ops, ==, &vdev_indirect_ops); } vic->vic_prev_indirect_vdev = srp->sr_prev_indirect_vdev; srp->sr_prev_indirect_vdev = vd->vdev_id; } spa->spa_removing_phys.sr_state = state; spa->spa_removing_phys.sr_end_time = gethrestime_sec(); spa->spa_vdev_removal = NULL; spa_vdev_removal_destroy(svr); spa_sync_removing_state(spa, tx); vdev_config_dirty(spa->spa_root_vdev); } static void free_mapped_segment_cb(void *arg, uint64_t offset, uint64_t size) { vdev_t *vd = arg; vdev_indirect_mark_obsolete(vd, offset, size); boolean_t checkpoint = B_FALSE; vdev_indirect_ops.vdev_op_remap(vd, offset, size, metaslab_free_impl_cb, &checkpoint); } /* * On behalf of the removal thread, syncs an incremental bit more of * the indirect mapping to disk and updates the in-memory mapping. * Called as a sync task in every txg that the removal thread makes progress. */ static void vdev_mapping_sync(void *arg, dmu_tx_t *tx) { spa_vdev_removal_t *svr = arg; spa_t *spa = dmu_tx_pool(tx)->dp_spa; vdev_t *vd = vdev_lookup_top(spa, svr->svr_vdev_id); vdev_indirect_config_t *vic = &vd->vdev_indirect_config; uint64_t txg = dmu_tx_get_txg(tx); vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping; ASSERT(vic->vic_mapping_object != 0); ASSERT3U(txg, ==, spa_syncing_txg(spa)); vdev_indirect_mapping_add_entries(vim, &svr->svr_new_segments[txg & TXG_MASK], tx); vdev_indirect_births_add_entry(vd->vdev_indirect_births, vdev_indirect_mapping_max_offset(vim), dmu_tx_get_txg(tx), tx); /* * Free the copied data for anything that was freed while the * mapping entries were in flight. */ mutex_enter(&svr->svr_lock); range_tree_vacate(svr->svr_frees[txg & TXG_MASK], free_mapped_segment_cb, vd); ASSERT3U(svr->svr_max_offset_to_sync[txg & TXG_MASK], >=, vdev_indirect_mapping_max_offset(vim)); svr->svr_max_offset_to_sync[txg & TXG_MASK] = 0; mutex_exit(&svr->svr_lock); spa_sync_removing_state(spa, tx); } typedef struct vdev_copy_segment_arg { spa_t *vcsa_spa; dva_t *vcsa_dest_dva; uint64_t vcsa_txg; range_tree_t *vcsa_obsolete_segs; } vdev_copy_segment_arg_t; static void unalloc_seg(void *arg, uint64_t start, uint64_t size) { vdev_copy_segment_arg_t *vcsa = arg; spa_t *spa = vcsa->vcsa_spa; blkptr_t bp = { 0 }; BP_SET_BIRTH(&bp, TXG_INITIAL, TXG_INITIAL); BP_SET_LSIZE(&bp, size); BP_SET_PSIZE(&bp, size); 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); DVA_SET_VDEV(&bp.blk_dva[0], DVA_GET_VDEV(vcsa->vcsa_dest_dva)); DVA_SET_OFFSET(&bp.blk_dva[0], DVA_GET_OFFSET(vcsa->vcsa_dest_dva) + start); DVA_SET_ASIZE(&bp.blk_dva[0], size); zio_free(spa, vcsa->vcsa_txg, &bp); } /* * All reads and writes associated with a call to spa_vdev_copy_segment() * are done. */ static void spa_vdev_copy_segment_done(zio_t *zio) { vdev_copy_segment_arg_t *vcsa = zio->io_private; range_tree_vacate(vcsa->vcsa_obsolete_segs, unalloc_seg, vcsa); range_tree_destroy(vcsa->vcsa_obsolete_segs); kmem_free(vcsa, sizeof (*vcsa)); spa_config_exit(zio->io_spa, SCL_STATE, zio->io_spa); } /* * The write of the new location is done. */ static void spa_vdev_copy_segment_write_done(zio_t *zio) { vdev_copy_arg_t *vca = zio->io_private; abd_free(zio->io_abd); mutex_enter(&vca->vca_lock); vca->vca_outstanding_bytes -= zio->io_size; cv_signal(&vca->vca_cv); mutex_exit(&vca->vca_lock); } /* * The read of the old location is done. The parent zio is the write to * the new location. Allow it to start. */ static void spa_vdev_copy_segment_read_done(zio_t *zio) { zio_nowait(zio_unique_parent(zio)); } /* * If the old and new vdevs are mirrors, we will read both sides of the old * mirror, and write each copy to the corresponding side of the new mirror. * If the old and new vdevs have a different number of children, we will do * this as best as possible. Since we aren't verifying checksums, this * ensures that as long as there's a good copy of the data, we'll have a * good copy after the removal, even if there's silent damage to one side * of the mirror. If we're removing a mirror that has some silent damage, * we'll have exactly the same damage in the new location (assuming that * the new location is also a mirror). * * We accomplish this by creating a tree of zio_t's, with as many writes as * there are "children" of the new vdev (a non-redundant vdev counts as one * child, a 2-way mirror has 2 children, etc). Each write has an associated * read from a child of the old vdev. Typically there will be the same * number of children of the old and new vdevs. However, if there are more * children of the new vdev, some child(ren) of the old vdev will be issued * multiple reads. If there are more children of the old vdev, some copies * will be dropped. * * For example, the tree of zio_t's for a 2-way mirror is: * * null * / \ * write(new vdev, child 0) write(new vdev, child 1) * | | * read(old vdev, child 0) read(old vdev, child 1) * * Child zio's complete before their parents complete. However, zio's * created with zio_vdev_child_io() may be issued before their children * complete. In this case we need to make sure that the children (reads) * complete before the parents (writes) are *issued*. We do this by not * calling zio_nowait() on each write until its corresponding read has * completed. * * The spa_config_lock must be held while zio's created by * zio_vdev_child_io() are in progress, to ensure that the vdev tree does * not change (e.g. due to a concurrent "zpool attach/detach"). The "null" * zio is needed to release the spa_config_lock after all the reads and * writes complete. (Note that we can't grab the config lock for each read, * because it is not reentrant - we could deadlock with a thread waiting * for a write lock.) */ static void spa_vdev_copy_one_child(vdev_copy_arg_t *vca, zio_t *nzio, vdev_t *source_vd, uint64_t source_offset, vdev_t *dest_child_vd, uint64_t dest_offset, int dest_id, uint64_t size) { ASSERT3U(spa_config_held(nzio->io_spa, SCL_ALL, RW_READER), !=, 0); mutex_enter(&vca->vca_lock); vca->vca_outstanding_bytes += size; mutex_exit(&vca->vca_lock); abd_t *abd = abd_alloc_for_io(size, B_FALSE); vdev_t *source_child_vd; if (source_vd->vdev_ops == &vdev_mirror_ops && dest_id != -1) { /* * Source and dest are both mirrors. Copy from the same * child id as we are copying to (wrapping around if there * are more dest children than source children). */ source_child_vd = source_vd->vdev_child[dest_id % source_vd->vdev_children]; } else { source_child_vd = source_vd; } zio_t *write_zio = zio_vdev_child_io(nzio, NULL, dest_child_vd, dest_offset, abd, size, ZIO_TYPE_WRITE, ZIO_PRIORITY_REMOVAL, ZIO_FLAG_CANFAIL, spa_vdev_copy_segment_write_done, vca); zio_nowait(zio_vdev_child_io(write_zio, NULL, source_child_vd, source_offset, abd, size, ZIO_TYPE_READ, ZIO_PRIORITY_REMOVAL, ZIO_FLAG_CANFAIL, spa_vdev_copy_segment_read_done, vca)); } /* * Allocate a new location for this segment, and create the zio_t's to * read from the old location and write to the new location. */ static int spa_vdev_copy_segment(vdev_t *vd, range_tree_t *segs, uint64_t maxalloc, uint64_t txg, vdev_copy_arg_t *vca, zio_alloc_list_t *zal) { metaslab_group_t *mg = vd->vdev_mg; spa_t *spa = vd->vdev_spa; spa_vdev_removal_t *svr = spa->spa_vdev_removal; vdev_indirect_mapping_entry_t *entry; dva_t dst = { 0 }; uint64_t start = range_tree_min(segs); ASSERT3U(maxalloc, <=, SPA_MAXBLOCKSIZE); uint64_t size = range_tree_span(segs); if (range_tree_span(segs) > maxalloc) { /* * We can't allocate all the segments. Prefer to end * the allocation at the end of a segment, thus avoiding * additional split blocks. */ range_seg_t search; avl_index_t where; search.rs_start = start + maxalloc; search.rs_end = search.rs_start; range_seg_t *rs = avl_find(&segs->rt_root, &search, &where); if (rs == NULL) { rs = avl_nearest(&segs->rt_root, where, AVL_BEFORE); } else { rs = AVL_PREV(&segs->rt_root, rs); } if (rs != NULL) { size = rs->rs_end - start; } else { /* * There are no segments that end before maxalloc. * I.e. the first segment is larger than maxalloc, * so we must split it. */ size = maxalloc; } } ASSERT3U(size, <=, maxalloc); /* * An allocation class might not have any remaining vdevs or space */ metaslab_class_t *mc = mg->mg_class; if (mc != spa_normal_class(spa) && mc->mc_groups <= 1) mc = spa_normal_class(spa); int error = metaslab_alloc_dva(spa, mc, size, &dst, 0, NULL, txg, 0, zal, 0); if (error == ENOSPC && mc != spa_normal_class(spa)) { error = metaslab_alloc_dva(spa, spa_normal_class(spa), size, &dst, 0, NULL, txg, 0, zal, 0); } if (error != 0) return (error); /* * Determine the ranges that are not actually needed. Offsets are * relative to the start of the range to be copied (i.e. relative to the * local variable "start"). */ range_tree_t *obsolete_segs = range_tree_create(NULL, NULL); range_seg_t *rs = avl_first(&segs->rt_root); ASSERT3U(rs->rs_start, ==, start); uint64_t prev_seg_end = rs->rs_end; while ((rs = AVL_NEXT(&segs->rt_root, rs)) != NULL) { if (rs->rs_start >= start + size) { break; } else { range_tree_add(obsolete_segs, prev_seg_end - start, rs->rs_start - prev_seg_end); } prev_seg_end = rs->rs_end; } /* We don't end in the middle of an obsolete range */ ASSERT3U(start + size, <=, prev_seg_end); range_tree_clear(segs, start, size); /* * We can't have any padding of the allocated size, otherwise we will * misunderstand what's allocated, and the size of the mapping. * The caller ensures this will be true by passing in a size that is * aligned to the worst (highest) ashift in the pool. */ ASSERT3U(DVA_GET_ASIZE(&dst), ==, size); entry = kmem_zalloc(sizeof (vdev_indirect_mapping_entry_t), KM_SLEEP); DVA_MAPPING_SET_SRC_OFFSET(&entry->vime_mapping, start); entry->vime_mapping.vimep_dst = dst; if (spa_feature_is_enabled(spa, SPA_FEATURE_OBSOLETE_COUNTS)) { entry->vime_obsolete_count = range_tree_space(obsolete_segs); } vdev_copy_segment_arg_t *vcsa = kmem_zalloc(sizeof (*vcsa), KM_SLEEP); vcsa->vcsa_dest_dva = &entry->vime_mapping.vimep_dst; vcsa->vcsa_obsolete_segs = obsolete_segs; vcsa->vcsa_spa = spa; vcsa->vcsa_txg = txg; /* * See comment before spa_vdev_copy_one_child(). */ spa_config_enter(spa, SCL_STATE, spa, RW_READER); zio_t *nzio = zio_null(spa->spa_txg_zio[txg & TXG_MASK], spa, NULL, spa_vdev_copy_segment_done, vcsa, 0); vdev_t *dest_vd = vdev_lookup_top(spa, DVA_GET_VDEV(&dst)); if (dest_vd->vdev_ops == &vdev_mirror_ops) { for (int i = 0; i < dest_vd->vdev_children; i++) { vdev_t *child = dest_vd->vdev_child[i]; spa_vdev_copy_one_child(vca, nzio, vd, start, child, DVA_GET_OFFSET(&dst), i, size); } } else { spa_vdev_copy_one_child(vca, nzio, vd, start, dest_vd, DVA_GET_OFFSET(&dst), -1, size); } zio_nowait(nzio); list_insert_tail(&svr->svr_new_segments[txg & TXG_MASK], entry); ASSERT3U(start + size, <=, vd->vdev_ms_count << vd->vdev_ms_shift); vdev_dirty(vd, 0, NULL, txg); return (0); } /* * Complete the removal of a toplevel vdev. This is called as a * synctask in the same txg that we will sync out the new config (to the * MOS object) which indicates that this vdev is indirect. */ static void vdev_remove_complete_sync(void *arg, dmu_tx_t *tx) { spa_vdev_removal_t *svr = arg; spa_t *spa = dmu_tx_pool(tx)->dp_spa; vdev_t *vd = vdev_lookup_top(spa, svr->svr_vdev_id); ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops); for (int i = 0; i < TXG_SIZE; i++) { ASSERT0(svr->svr_bytes_done[i]); } ASSERT3U(spa->spa_removing_phys.sr_copied, ==, spa->spa_removing_phys.sr_to_copy); vdev_destroy_spacemaps(vd, tx); /* destroy leaf zaps, if any */ ASSERT3P(svr->svr_zaplist, !=, NULL); for (nvpair_t *pair = nvlist_next_nvpair(svr->svr_zaplist, NULL); pair != NULL; pair = nvlist_next_nvpair(svr->svr_zaplist, pair)) { vdev_destroy_unlink_zap(vd, fnvpair_value_uint64(pair), tx); } fnvlist_free(svr->svr_zaplist); spa_finish_removal(dmu_tx_pool(tx)->dp_spa, DSS_FINISHED, tx); /* vd->vdev_path is not available here */ spa_history_log_internal(spa, "vdev remove completed", tx, "%s vdev %llu", spa_name(spa), vd->vdev_id); } static void vdev_remove_enlist_zaps(vdev_t *vd, nvlist_t *zlist) { ASSERT3P(zlist, !=, NULL); ASSERT3P(vd->vdev_ops, !=, &vdev_raidz_ops); if (vd->vdev_leaf_zap != 0) { char zkey[32]; (void) snprintf(zkey, sizeof (zkey), "%s-%"PRIu64, VDEV_REMOVAL_ZAP_OBJS, vd->vdev_leaf_zap); fnvlist_add_uint64(zlist, zkey, vd->vdev_leaf_zap); } for (uint64_t id = 0; id < vd->vdev_children; id++) { vdev_remove_enlist_zaps(vd->vdev_child[id], zlist); } } static void vdev_remove_replace_with_indirect(vdev_t *vd, uint64_t txg) { vdev_t *ivd; dmu_tx_t *tx; spa_t *spa = vd->vdev_spa; spa_vdev_removal_t *svr = spa->spa_vdev_removal; /* * First, build a list of leaf zaps to be destroyed. * This is passed to the sync context thread, * which does the actual unlinking. */ svr->svr_zaplist = fnvlist_alloc(); vdev_remove_enlist_zaps(vd, svr->svr_zaplist); ivd = vdev_add_parent(vd, &vdev_indirect_ops); ivd->vdev_removing = 0; vd->vdev_leaf_zap = 0; vdev_remove_child(ivd, vd); vdev_compact_children(ivd); ASSERT(!list_link_active(&vd->vdev_state_dirty_node)); tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); dsl_sync_task_nowait(spa->spa_dsl_pool, vdev_remove_complete_sync, svr, 0, ZFS_SPACE_CHECK_NONE, tx); dmu_tx_commit(tx); /* * Indicate that this thread has exited. * After this, we can not use svr. */ mutex_enter(&svr->svr_lock); svr->svr_thread = NULL; cv_broadcast(&svr->svr_cv); mutex_exit(&svr->svr_lock); } /* * Complete the removal of a toplevel vdev. This is called in open * context by the removal thread after we have copied all vdev's data. */ static void vdev_remove_complete(spa_t *spa) { uint64_t txg; /* * Wait for any deferred frees to be synced before we call * vdev_metaslab_fini() */ txg_wait_synced(spa->spa_dsl_pool, 0); txg = spa_vdev_enter(spa); vdev_t *vd = vdev_lookup_top(spa, spa->spa_vdev_removal->svr_vdev_id); ASSERT3P(vd->vdev_initialize_thread, ==, NULL); sysevent_t *ev = spa_event_create(spa, vd, NULL, ESC_ZFS_VDEV_REMOVE_DEV); zfs_dbgmsg("finishing device removal for vdev %llu in txg %llu", vd->vdev_id, txg); /* * Discard allocation state. */ if (vd->vdev_mg != NULL) { vdev_metaslab_fini(vd); metaslab_group_destroy(vd->vdev_mg); vd->vdev_mg = NULL; } ASSERT0(vd->vdev_stat.vs_space); ASSERT0(vd->vdev_stat.vs_dspace); vdev_remove_replace_with_indirect(vd, txg); /* * We now release the locks, allowing spa_sync to run and finish the * removal via vdev_remove_complete_sync in syncing context. * * Note that we hold on to the vdev_t that has been replaced. Since * it isn't part of the vdev tree any longer, it can't be concurrently * manipulated, even while we don't have the config lock. */ (void) spa_vdev_exit(spa, NULL, txg, 0); /* * Top ZAP should have been transferred to the indirect vdev in * vdev_remove_replace_with_indirect. */ ASSERT0(vd->vdev_top_zap); /* * Leaf ZAP should have been moved in vdev_remove_replace_with_indirect. */ ASSERT0(vd->vdev_leaf_zap); txg = spa_vdev_enter(spa); (void) vdev_label_init(vd, 0, VDEV_LABEL_REMOVE); /* * Request to update the config and the config cachefile. */ vdev_config_dirty(spa->spa_root_vdev); (void) spa_vdev_exit(spa, vd, txg, 0); spa_event_post(ev); } /* * Evacuates a segment of size at most max_alloc from the vdev * via repeated calls to spa_vdev_copy_segment. If an allocation * fails, the pool is probably too fragmented to handle such a * large size, so decrease max_alloc so that the caller will not try * this size again this txg. */ static void spa_vdev_copy_impl(vdev_t *vd, spa_vdev_removal_t *svr, vdev_copy_arg_t *vca, uint64_t *max_alloc, dmu_tx_t *tx) { uint64_t txg = dmu_tx_get_txg(tx); spa_t *spa = dmu_tx_pool(tx)->dp_spa; mutex_enter(&svr->svr_lock); /* * Determine how big of a chunk to copy. We can allocate up * to max_alloc bytes, and we can span up to vdev_removal_max_span * bytes of unallocated space at a time. "segs" will track the * allocated segments that we are copying. We may also be copying * free segments (of up to vdev_removal_max_span bytes). */ range_tree_t *segs = range_tree_create(NULL, NULL); for (;;) { range_seg_t *rs = avl_first(&svr->svr_allocd_segs->rt_root); if (rs == NULL) break; uint64_t seg_length; if (range_tree_is_empty(segs)) { /* need to truncate the first seg based on max_alloc */ seg_length = MIN(rs->rs_end - rs->rs_start, *max_alloc); } else { if (rs->rs_start - range_tree_max(segs) > vdev_removal_max_span) { /* * Including this segment would cause us to * copy a larger unneeded chunk than is allowed. */ break; } else if (rs->rs_end - range_tree_min(segs) > *max_alloc) { /* * This additional segment would extend past * max_alloc. Rather than splitting this * segment, leave it for the next mapping. */ break; } else { seg_length = rs->rs_end - rs->rs_start; } } range_tree_add(segs, rs->rs_start, seg_length); range_tree_remove(svr->svr_allocd_segs, rs->rs_start, seg_length); } if (range_tree_is_empty(segs)) { mutex_exit(&svr->svr_lock); range_tree_destroy(segs); return; } if (svr->svr_max_offset_to_sync[txg & TXG_MASK] == 0) { dsl_sync_task_nowait(dmu_tx_pool(tx), vdev_mapping_sync, svr, 0, ZFS_SPACE_CHECK_NONE, tx); } svr->svr_max_offset_to_sync[txg & TXG_MASK] = range_tree_max(segs); /* * Note: this is the amount of *allocated* space * that we are taking care of each txg. */ svr->svr_bytes_done[txg & TXG_MASK] += range_tree_space(segs); mutex_exit(&svr->svr_lock); zio_alloc_list_t zal; metaslab_trace_init(&zal); uint64_t thismax = SPA_MAXBLOCKSIZE; while (!range_tree_is_empty(segs)) { int error = spa_vdev_copy_segment(vd, segs, thismax, txg, vca, &zal); if (error == ENOSPC) { /* * Cut our segment in half, and don't try this * segment size again this txg. Note that the * allocation size must be aligned to the highest * ashift in the pool, so that the allocation will * not be padded out to a multiple of the ashift, * which could cause us to think that this mapping * is larger than we intended. */ ASSERT3U(spa->spa_max_ashift, >=, SPA_MINBLOCKSHIFT); ASSERT3U(spa->spa_max_ashift, ==, spa->spa_min_ashift); uint64_t attempted = MIN(range_tree_span(segs), thismax); thismax = P2ROUNDUP(attempted / 2, 1 << spa->spa_max_ashift); /* * The minimum-size allocation can not fail. */ ASSERT3U(attempted, >, 1 << spa->spa_max_ashift); *max_alloc = attempted - (1 << spa->spa_max_ashift); } else { ASSERT0(error); /* * We've performed an allocation, so reset the * alloc trace list. */ metaslab_trace_fini(&zal); metaslab_trace_init(&zal); } } metaslab_trace_fini(&zal); range_tree_destroy(segs); } /* * The removal thread operates in open context. It iterates over all * allocated space in the vdev, by loading each metaslab's spacemap. * For each contiguous segment of allocated space (capping the segment * size at SPA_MAXBLOCKSIZE), we: * - Allocate space for it on another vdev. * - Create a new mapping from the old location to the new location * (as a record in svr_new_segments). * - Initiate a logical read zio to get the data off the removing disk. * - In the read zio's done callback, initiate a logical write zio to * write it to the new vdev. * Note that all of this will take effect when a particular TXG syncs. * The sync thread ensures that all the phys reads and writes for the syncing * TXG have completed (see spa_txg_zio) and writes the new mappings to disk * (see vdev_mapping_sync()). */ static void spa_vdev_remove_thread(void *arg) { spa_t *spa = arg; spa_vdev_removal_t *svr = spa->spa_vdev_removal; vdev_copy_arg_t vca; uint64_t max_alloc = zfs_remove_max_segment; uint64_t last_txg = 0; spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); vdev_t *vd = vdev_lookup_top(spa, svr->svr_vdev_id); vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping; uint64_t start_offset = vdev_indirect_mapping_max_offset(vim); ASSERT3P(vd->vdev_ops, !=, &vdev_indirect_ops); ASSERT(vdev_is_concrete(vd)); ASSERT(vd->vdev_removing); ASSERT(vd->vdev_indirect_config.vic_mapping_object != 0); ASSERT(vim != NULL); mutex_init(&vca.vca_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&vca.vca_cv, NULL, CV_DEFAULT, NULL); vca.vca_outstanding_bytes = 0; mutex_enter(&svr->svr_lock); /* * Start from vim_max_offset so we pick up where we left off * if we are restarting the removal after opening the pool. */ uint64_t msi; for (msi = start_offset >> vd->vdev_ms_shift; msi < vd->vdev_ms_count && !svr->svr_thread_exit; msi++) { metaslab_t *msp = vd->vdev_ms[msi]; ASSERT3U(msi, <=, vd->vdev_ms_count); ASSERT0(range_tree_space(svr->svr_allocd_segs)); mutex_enter(&msp->ms_sync_lock); mutex_enter(&msp->ms_lock); /* * Assert nothing in flight -- ms_*tree is empty. */ for (int i = 0; i < TXG_SIZE; i++) { ASSERT0(range_tree_space(msp->ms_allocating[i])); } /* * If the metaslab has ever been allocated from (ms_sm!=NULL), * read the allocated segments from the space map object * into svr_allocd_segs. Since we do this while holding * svr_lock and ms_sync_lock, concurrent frees (which * would have modified the space map) will wait for us * to finish loading the spacemap, and then take the * appropriate action (see free_from_removing_vdev()). */ if (msp->ms_sm != NULL) { VERIFY0(space_map_load(msp->ms_sm, svr->svr_allocd_segs, SM_ALLOC)); range_tree_walk(msp->ms_freeing, range_tree_remove, svr->svr_allocd_segs); /* * When we are resuming from a paused removal (i.e. * when importing a pool with a removal in progress), * discard any state that we have already processed. */ range_tree_clear(svr->svr_allocd_segs, 0, start_offset); } mutex_exit(&msp->ms_lock); mutex_exit(&msp->ms_sync_lock); vca.vca_msp = msp; zfs_dbgmsg("copying %llu segments for metaslab %llu", avl_numnodes(&svr->svr_allocd_segs->rt_root), msp->ms_id); while (!svr->svr_thread_exit && !range_tree_is_empty(svr->svr_allocd_segs)) { mutex_exit(&svr->svr_lock); /* * We need to periodically drop the config lock so that * writers can get in. Additionally, we can't wait * for a txg to sync while holding a config lock * (since a waiting writer could cause a 3-way deadlock * with the sync thread, which also gets a config * lock for reader). So we can't hold the config lock * while calling dmu_tx_assign(). */ spa_config_exit(spa, SCL_CONFIG, FTAG); /* * This delay will pause the removal around the point - * specified by zfs_remove_max_bytes_pause. We do this + * specified by zfs_removal_suspend_progress. We do this * solely from the test suite or during debugging. */ uint64_t bytes_copied = spa->spa_removing_phys.sr_copied; for (int i = 0; i < TXG_SIZE; i++) bytes_copied += svr->svr_bytes_done[i]; - while (zfs_remove_max_bytes_pause <= bytes_copied && + while (zfs_removal_suspend_progress && !svr->svr_thread_exit) delay(hz); mutex_enter(&vca.vca_lock); while (vca.vca_outstanding_bytes > zfs_remove_max_copy_bytes) { cv_wait(&vca.vca_cv, &vca.vca_lock); } mutex_exit(&vca.vca_lock); dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir); VERIFY0(dmu_tx_assign(tx, TXG_WAIT)); uint64_t txg = dmu_tx_get_txg(tx); /* * Reacquire the vdev_config lock. The vdev_t * that we're removing may have changed, e.g. due * to a vdev_attach or vdev_detach. */ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); vd = vdev_lookup_top(spa, svr->svr_vdev_id); if (txg != last_txg) max_alloc = zfs_remove_max_segment; last_txg = txg; spa_vdev_copy_impl(vd, svr, &vca, &max_alloc, tx); dmu_tx_commit(tx); mutex_enter(&svr->svr_lock); } } mutex_exit(&svr->svr_lock); spa_config_exit(spa, SCL_CONFIG, FTAG); /* * Wait for all copies to finish before cleaning up the vca. */ txg_wait_synced(spa->spa_dsl_pool, 0); ASSERT0(vca.vca_outstanding_bytes); mutex_destroy(&vca.vca_lock); cv_destroy(&vca.vca_cv); if (svr->svr_thread_exit) { mutex_enter(&svr->svr_lock); range_tree_vacate(svr->svr_allocd_segs, NULL, NULL); svr->svr_thread = NULL; cv_broadcast(&svr->svr_cv); mutex_exit(&svr->svr_lock); } else { ASSERT0(range_tree_space(svr->svr_allocd_segs)); vdev_remove_complete(spa); } } void spa_vdev_remove_suspend(spa_t *spa) { spa_vdev_removal_t *svr = spa->spa_vdev_removal; if (svr == NULL) return; mutex_enter(&svr->svr_lock); svr->svr_thread_exit = B_TRUE; while (svr->svr_thread != NULL) cv_wait(&svr->svr_cv, &svr->svr_lock); svr->svr_thread_exit = B_FALSE; mutex_exit(&svr->svr_lock); } /* ARGSUSED */ static int spa_vdev_remove_cancel_check(void *arg, dmu_tx_t *tx) { spa_t *spa = dmu_tx_pool(tx)->dp_spa; if (spa->spa_vdev_removal == NULL) return (ENOTACTIVE); return (0); } /* * Cancel a removal by freeing all entries from the partial mapping * and marking the vdev as no longer being removing. */ /* ARGSUSED */ static void spa_vdev_remove_cancel_sync(void *arg, dmu_tx_t *tx) { spa_t *spa = dmu_tx_pool(tx)->dp_spa; spa_vdev_removal_t *svr = spa->spa_vdev_removal; vdev_t *vd = vdev_lookup_top(spa, svr->svr_vdev_id); vdev_indirect_config_t *vic = &vd->vdev_indirect_config; vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping; objset_t *mos = spa->spa_meta_objset; ASSERT3P(svr->svr_thread, ==, NULL); spa_feature_decr(spa, SPA_FEATURE_DEVICE_REMOVAL, tx); if (vdev_obsolete_counts_are_precise(vd)) { spa_feature_decr(spa, SPA_FEATURE_OBSOLETE_COUNTS, tx); VERIFY0(zap_remove(spa->spa_meta_objset, vd->vdev_top_zap, VDEV_TOP_ZAP_OBSOLETE_COUNTS_ARE_PRECISE, tx)); } if (vdev_obsolete_sm_object(vd) != 0) { ASSERT(vd->vdev_obsolete_sm != NULL); ASSERT3U(vdev_obsolete_sm_object(vd), ==, space_map_object(vd->vdev_obsolete_sm)); space_map_free(vd->vdev_obsolete_sm, tx); VERIFY0(zap_remove(spa->spa_meta_objset, vd->vdev_top_zap, VDEV_TOP_ZAP_INDIRECT_OBSOLETE_SM, tx)); space_map_close(vd->vdev_obsolete_sm); vd->vdev_obsolete_sm = NULL; spa_feature_decr(spa, SPA_FEATURE_OBSOLETE_COUNTS, tx); } for (int i = 0; i < TXG_SIZE; i++) { ASSERT(list_is_empty(&svr->svr_new_segments[i])); ASSERT3U(svr->svr_max_offset_to_sync[i], <=, vdev_indirect_mapping_max_offset(vim)); } for (uint64_t msi = 0; msi < vd->vdev_ms_count; msi++) { metaslab_t *msp = vd->vdev_ms[msi]; if (msp->ms_start >= vdev_indirect_mapping_max_offset(vim)) break; ASSERT0(range_tree_space(svr->svr_allocd_segs)); mutex_enter(&msp->ms_lock); /* * Assert nothing in flight -- ms_*tree is empty. */ for (int i = 0; i < TXG_SIZE; i++) ASSERT0(range_tree_space(msp->ms_allocating[i])); for (int i = 0; i < TXG_DEFER_SIZE; i++) ASSERT0(range_tree_space(msp->ms_defer[i])); ASSERT0(range_tree_space(msp->ms_freed)); if (msp->ms_sm != NULL) { mutex_enter(&svr->svr_lock); VERIFY0(space_map_load(msp->ms_sm, svr->svr_allocd_segs, SM_ALLOC)); range_tree_walk(msp->ms_freeing, range_tree_remove, svr->svr_allocd_segs); /* * Clear everything past what has been synced, * because we have not allocated mappings for it yet. */ uint64_t syncd = vdev_indirect_mapping_max_offset(vim); uint64_t sm_end = msp->ms_sm->sm_start + msp->ms_sm->sm_size; if (sm_end > syncd) range_tree_clear(svr->svr_allocd_segs, syncd, sm_end - syncd); mutex_exit(&svr->svr_lock); } mutex_exit(&msp->ms_lock); mutex_enter(&svr->svr_lock); range_tree_vacate(svr->svr_allocd_segs, free_mapped_segment_cb, vd); mutex_exit(&svr->svr_lock); } /* * Note: this must happen after we invoke free_mapped_segment_cb, * because it adds to the obsolete_segments. */ range_tree_vacate(vd->vdev_obsolete_segments, NULL, NULL); ASSERT3U(vic->vic_mapping_object, ==, vdev_indirect_mapping_object(vd->vdev_indirect_mapping)); vdev_indirect_mapping_close(vd->vdev_indirect_mapping); vd->vdev_indirect_mapping = NULL; vdev_indirect_mapping_free(mos, vic->vic_mapping_object, tx); vic->vic_mapping_object = 0; ASSERT3U(vic->vic_births_object, ==, vdev_indirect_births_object(vd->vdev_indirect_births)); vdev_indirect_births_close(vd->vdev_indirect_births); vd->vdev_indirect_births = NULL; vdev_indirect_births_free(mos, vic->vic_births_object, tx); vic->vic_births_object = 0; /* * We may have processed some frees from the removing vdev in this * txg, thus increasing svr_bytes_done; discard that here to * satisfy the assertions in spa_vdev_removal_destroy(). * Note that future txg's can not have any bytes_done, because * future TXG's are only modified from open context, and we have * already shut down the copying thread. */ svr->svr_bytes_done[dmu_tx_get_txg(tx) & TXG_MASK] = 0; spa_finish_removal(spa, DSS_CANCELED, tx); vd->vdev_removing = B_FALSE; vdev_config_dirty(vd); zfs_dbgmsg("canceled device removal for vdev %llu in %llu", vd->vdev_id, dmu_tx_get_txg(tx)); spa_history_log_internal(spa, "vdev remove canceled", tx, "%s vdev %llu %s", spa_name(spa), vd->vdev_id, (vd->vdev_path != NULL) ? vd->vdev_path : "-"); } int spa_vdev_remove_cancel(spa_t *spa) { spa_vdev_remove_suspend(spa); if (spa->spa_vdev_removal == NULL) return (ENOTACTIVE); uint64_t vdid = spa->spa_vdev_removal->svr_vdev_id; int error = dsl_sync_task(spa->spa_name, spa_vdev_remove_cancel_check, spa_vdev_remove_cancel_sync, NULL, 0, ZFS_SPACE_CHECK_EXTRA_RESERVED); if (error == 0) { spa_config_enter(spa, SCL_ALLOC | SCL_VDEV, FTAG, RW_WRITER); vdev_t *vd = vdev_lookup_top(spa, vdid); metaslab_group_activate(vd->vdev_mg); spa_config_exit(spa, SCL_ALLOC | SCL_VDEV, FTAG); } return (error); } void svr_sync(spa_t *spa, dmu_tx_t *tx) { spa_vdev_removal_t *svr = spa->spa_vdev_removal; int txgoff = dmu_tx_get_txg(tx) & TXG_MASK; if (svr == NULL) return; /* * This check is necessary so that we do not dirty the * DIRECTORY_OBJECT via spa_sync_removing_state() when there * is nothing to do. Dirtying it every time would prevent us * from syncing-to-convergence. */ if (svr->svr_bytes_done[txgoff] == 0) return; /* * Update progress accounting. */ spa->spa_removing_phys.sr_copied += svr->svr_bytes_done[txgoff]; svr->svr_bytes_done[txgoff] = 0; spa_sync_removing_state(spa, tx); } static void vdev_remove_make_hole_and_free(vdev_t *vd) { uint64_t id = vd->vdev_id; spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; boolean_t last_vdev = (id == (rvd->vdev_children - 1)); ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); vdev_free(vd); if (last_vdev) { vdev_compact_children(rvd); } else { vd = vdev_alloc_common(spa, id, 0, &vdev_hole_ops); vdev_add_child(rvd, vd); } vdev_config_dirty(rvd); /* * Reassess the health of our root vdev. */ vdev_reopen(rvd); } /* * Remove a log device. The config lock is held for the specified TXG. */ static int spa_vdev_remove_log(vdev_t *vd, uint64_t *txg) { metaslab_group_t *mg = vd->vdev_mg; spa_t *spa = vd->vdev_spa; int error = 0; ASSERT(vd->vdev_islog); ASSERT(vd == vd->vdev_top); ASSERT(MUTEX_HELD(&spa_namespace_lock)); /* * Stop allocating from this vdev. */ metaslab_group_passivate(mg); /* * Wait for the youngest allocations and frees to sync, * and then wait for the deferral of those frees to finish. */ spa_vdev_config_exit(spa, NULL, *txg + TXG_CONCURRENT_STATES + TXG_DEFER_SIZE, 0, FTAG); /* * Evacuate the device. We don't hold the config lock as * writer since we need to do I/O but we do keep the * spa_namespace_lock held. Once this completes the device * should no longer have any blocks allocated on it. */ ASSERT(MUTEX_HELD(&spa_namespace_lock)); if (vd->vdev_stat.vs_alloc != 0) error = spa_reset_logs(spa); *txg = spa_vdev_config_enter(spa); if (error != 0) { metaslab_group_activate(mg); return (error); } ASSERT0(vd->vdev_stat.vs_alloc); /* * The evacuation succeeded. Remove any remaining MOS metadata * associated with this vdev, and wait for these changes to sync. */ vd->vdev_removing = B_TRUE; vdev_dirty_leaves(vd, VDD_DTL, *txg); vdev_config_dirty(vd); vdev_metaslab_fini(vd); spa_history_log_internal(spa, "vdev remove", NULL, "%s vdev %llu (log) %s", spa_name(spa), vd->vdev_id, (vd->vdev_path != NULL) ? vd->vdev_path : "-"); /* Make sure these changes are sync'ed */ spa_vdev_config_exit(spa, NULL, *txg, 0, FTAG); /* Stop initializing */ (void) vdev_initialize_stop_all(vd, VDEV_INITIALIZE_CANCELED); *txg = spa_vdev_config_enter(spa); sysevent_t *ev = spa_event_create(spa, vd, NULL, ESC_ZFS_VDEV_REMOVE_DEV); ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); /* The top ZAP should have been destroyed by vdev_remove_empty. */ ASSERT0(vd->vdev_top_zap); /* The leaf ZAP should have been destroyed by vdev_dtl_sync. */ ASSERT0(vd->vdev_leaf_zap); (void) vdev_label_init(vd, 0, VDEV_LABEL_REMOVE); if (list_link_active(&vd->vdev_state_dirty_node)) vdev_state_clean(vd); if (list_link_active(&vd->vdev_config_dirty_node)) vdev_config_clean(vd); ASSERT0(vd->vdev_stat.vs_alloc); /* * Clean up the vdev namespace. */ vdev_remove_make_hole_and_free(vd); if (ev != NULL) spa_event_post(ev); return (0); } static int spa_vdev_remove_top_check(vdev_t *vd) { spa_t *spa = vd->vdev_spa; if (vd != vd->vdev_top) return (SET_ERROR(ENOTSUP)); if (!spa_feature_is_enabled(spa, SPA_FEATURE_DEVICE_REMOVAL)) return (SET_ERROR(ENOTSUP)); /* available space in the pool's normal class */ uint64_t available = dsl_dir_space_available( spa->spa_dsl_pool->dp_root_dir, NULL, 0, B_TRUE); metaslab_class_t *mc = vd->vdev_mg->mg_class; /* * When removing a vdev from an allocation class that has * remaining vdevs, include available space from the class. */ if (mc != spa_normal_class(spa) && mc->mc_groups > 1) { uint64_t class_avail = metaslab_class_get_space(mc) - metaslab_class_get_alloc(mc); /* add class space, adjusted for overhead */ available += (class_avail * 94) / 100; } /* * There has to be enough free space to remove the * device and leave double the "slop" space (i.e. we * must leave at least 3% of the pool free, in addition to * the normal slop space). */ if (available < vd->vdev_stat.vs_dspace + spa_get_slop_space(spa)) { return (SET_ERROR(ENOSPC)); } /* * There can not be a removal in progress. */ if (spa->spa_removing_phys.sr_state == DSS_SCANNING) return (SET_ERROR(EBUSY)); /* * The device must have all its data. */ if (!vdev_dtl_empty(vd, DTL_MISSING) || !vdev_dtl_empty(vd, DTL_OUTAGE)) return (SET_ERROR(EBUSY)); /* * The device must be healthy. */ if (!vdev_readable(vd)) return (SET_ERROR(EIO)); /* * All vdevs in normal class must have the same ashift. */ if (spa->spa_max_ashift != spa->spa_min_ashift) { return (SET_ERROR(EINVAL)); } /* * All vdevs in normal class must have the same ashift * and not be raidz. */ vdev_t *rvd = spa->spa_root_vdev; int num_indirect = 0; for (uint64_t id = 0; id < rvd->vdev_children; id++) { vdev_t *cvd = rvd->vdev_child[id]; if (cvd->vdev_ashift != 0 && !cvd->vdev_islog) ASSERT3U(cvd->vdev_ashift, ==, spa->spa_max_ashift); if (cvd->vdev_ops == &vdev_indirect_ops) num_indirect++; if (!vdev_is_concrete(cvd)) continue; if (cvd->vdev_ops == &vdev_raidz_ops) return (SET_ERROR(EINVAL)); /* * Need the mirror to be mirror of leaf vdevs only */ if (cvd->vdev_ops == &vdev_mirror_ops) { for (uint64_t cid = 0; cid < cvd->vdev_children; cid++) { vdev_t *tmp = cvd->vdev_child[cid]; if (!tmp->vdev_ops->vdev_op_leaf) return (SET_ERROR(EINVAL)); } } } return (0); } /* * Initiate removal of a top-level vdev, reducing the total space in the pool. * The config lock is held for the specified TXG. Once initiated, * evacuation of all allocated space (copying it to other vdevs) happens * in the background (see spa_vdev_remove_thread()), and can be canceled * (see spa_vdev_remove_cancel()). If successful, the vdev will * be transformed to an indirect vdev (see spa_vdev_remove_complete()). */ static int spa_vdev_remove_top(vdev_t *vd, uint64_t *txg) { spa_t *spa = vd->vdev_spa; int error; /* * Check for errors up-front, so that we don't waste time * passivating the metaslab group and clearing the ZIL if there * are errors. */ error = spa_vdev_remove_top_check(vd); if (error != 0) return (error); /* * Stop allocating from this vdev. Note that we must check * that this is not the only device in the pool before * passivating, otherwise we will not be able to make * progress because we can't allocate from any vdevs. * The above check for sufficient free space serves this * purpose. */ metaslab_group_t *mg = vd->vdev_mg; metaslab_group_passivate(mg); /* * Wait for the youngest allocations and frees to sync, * and then wait for the deferral of those frees to finish. */ spa_vdev_config_exit(spa, NULL, *txg + TXG_CONCURRENT_STATES + TXG_DEFER_SIZE, 0, FTAG); /* * We must ensure that no "stubby" log blocks are allocated * on the device to be removed. These blocks could be * written at any time, including while we are in the middle * of copying them. */ error = spa_reset_logs(spa); /* * We stop any initializing that is currently in progress but leave * the state as "active". This will allow the initializing to resume * if the removal is canceled sometime later. */ vdev_initialize_stop_all(vd, VDEV_INITIALIZE_ACTIVE); *txg = spa_vdev_config_enter(spa); /* * Things might have changed while the config lock was dropped * (e.g. space usage). Check for errors again. */ if (error == 0) error = spa_vdev_remove_top_check(vd); if (error != 0) { metaslab_group_activate(mg); spa_async_request(spa, SPA_ASYNC_INITIALIZE_RESTART); return (error); } vd->vdev_removing = B_TRUE; vdev_dirty_leaves(vd, VDD_DTL, *txg); vdev_config_dirty(vd); dmu_tx_t *tx = dmu_tx_create_assigned(spa->spa_dsl_pool, *txg); dsl_sync_task_nowait(spa->spa_dsl_pool, vdev_remove_initiate_sync, (void *)(uintptr_t)vd->vdev_id, 0, ZFS_SPACE_CHECK_NONE, tx); dmu_tx_commit(tx); return (0); } /* * Remove a device from the pool. * * Removing a device from the vdev namespace requires several steps * and can take a significant amount of time. As a result we use * the spa_vdev_config_[enter/exit] functions which allow us to * grab and release the spa_config_lock while still holding the namespace * lock. During each step the configuration is synced out. */ int spa_vdev_remove(spa_t *spa, uint64_t guid, boolean_t unspare) { vdev_t *vd; nvlist_t **spares, **l2cache, *nv; uint64_t txg = 0; uint_t nspares, nl2cache; int error = 0; boolean_t locked = MUTEX_HELD(&spa_namespace_lock); sysevent_t *ev = NULL; ASSERT(spa_writeable(spa)); if (!locked) 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; if (!locked) return (spa_vdev_exit(spa, NULL, txg, error)); return (error); } vd = spa_lookup_by_guid(spa, guid, B_FALSE); if (spa->spa_spares.sav_vdevs != NULL && nvlist_lookup_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0 && (nv = spa_nvlist_lookup_by_guid(spares, nspares, guid)) != NULL) { /* * Only remove the hot spare if it's not currently in use * in this pool. */ if (vd == NULL || unspare) { char *nvstr = fnvlist_lookup_string(nv, ZPOOL_CONFIG_PATH); spa_history_log_internal(spa, "vdev remove", NULL, "%s vdev (%s) %s", spa_name(spa), VDEV_TYPE_SPARE, nvstr); if (vd == NULL) vd = spa_lookup_by_guid(spa, guid, B_TRUE); ev = spa_event_create(spa, vd, NULL, ESC_ZFS_VDEV_REMOVE_AUX); spa_vdev_remove_aux(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, spares, nspares, nv); spa_load_spares(spa); spa->spa_spares.sav_sync = B_TRUE; } else { error = SET_ERROR(EBUSY); } } else if (spa->spa_l2cache.sav_vdevs != NULL && nvlist_lookup_nvlist_array(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0 && (nv = spa_nvlist_lookup_by_guid(l2cache, nl2cache, guid)) != NULL) { char *nvstr = fnvlist_lookup_string(nv, ZPOOL_CONFIG_PATH); spa_history_log_internal(spa, "vdev remove", NULL, "%s vdev (%s) %s", spa_name(spa), VDEV_TYPE_L2CACHE, nvstr); /* * Cache devices can always be removed. */ vd = spa_lookup_by_guid(spa, guid, B_TRUE); ev = spa_event_create(spa, vd, NULL, ESC_ZFS_VDEV_REMOVE_AUX); spa_vdev_remove_aux(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, l2cache, nl2cache, nv); spa_load_l2cache(spa); spa->spa_l2cache.sav_sync = B_TRUE; } else if (vd != NULL && vd->vdev_islog) { ASSERT(!locked); error = spa_vdev_remove_log(vd, &txg); } else if (vd != NULL) { ASSERT(!locked); error = spa_vdev_remove_top(vd, &txg); } else { /* * There is no vdev of any kind with the specified guid. */ error = SET_ERROR(ENOENT); } if (!locked) error = spa_vdev_exit(spa, NULL, txg, error); if (ev != NULL) { if (error != 0) { spa_event_discard(ev); } else { spa_event_post(ev); } } return (error); } int spa_removal_get_stats(spa_t *spa, pool_removal_stat_t *prs) { prs->prs_state = spa->spa_removing_phys.sr_state; if (prs->prs_state == DSS_NONE) return (SET_ERROR(ENOENT)); prs->prs_removing_vdev = spa->spa_removing_phys.sr_removing_vdev; prs->prs_start_time = spa->spa_removing_phys.sr_start_time; prs->prs_end_time = spa->spa_removing_phys.sr_end_time; prs->prs_to_copy = spa->spa_removing_phys.sr_to_copy; prs->prs_copied = spa->spa_removing_phys.sr_copied; if (spa->spa_vdev_removal != NULL) { for (int i = 0; i < TXG_SIZE; i++) { prs->prs_copied += spa->spa_vdev_removal->svr_bytes_done[i]; } } prs->prs_mapping_memory = 0; uint64_t indirect_vdev_id = spa->spa_removing_phys.sr_prev_indirect_vdev; while (indirect_vdev_id != -1) { vdev_t *vd = spa->spa_root_vdev->vdev_child[indirect_vdev_id]; vdev_indirect_config_t *vic = &vd->vdev_indirect_config; vdev_indirect_mapping_t *vim = vd->vdev_indirect_mapping; ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops); prs->prs_mapping_memory += vdev_indirect_mapping_size(vim); indirect_vdev_id = vic->vic_prev_indirect_vdev; } return (0); } Index: vendor-sys/illumos/dist/uts/common/fs/zfs/zil.c =================================================================== --- vendor-sys/illumos/dist/uts/common/fs/zfs/zil.c (revision 354960) +++ vendor-sys/illumos/dist/uts/common/fs/zfs/zil.c (revision 354961) @@ -1,3418 +1,3418 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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] */ /* Portions Copyright 2010 Robert Milkowski */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* * 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 * instantianted (e.g. if the dataset is a normal fileystem, 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. */ int zfs_commit_timeout_pct = 5; /* * 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. */ boolean_t zil_nocacheflush = B_FALSE; /* * 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. */ uint64_t zil_slog_bulk = 768 * 1024; static kmem_cache_t *zil_lwb_cache; static kmem_cache_t *zil_zcw_cache; static void zil_async_to_sync(zilog_t *zilog, uint64_t foid); #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 = AVL_CMP(DVA_GET_VDEV(dva1), DVA_GET_VDEV(dva2)); if (likely(cmp)) return (cmp); return (AVL_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; zc->zc_word[ZIL_ZC_GUID_0] = spa_get_random(-1ULL); zc->zc_word[ZIL_ZC_GUID_1] = spa_get_random(-1ULL); zc->zc_word[ZIL_ZC_OBJSET] = dmu_objset_id(zilog->zl_os); zc->zc_word[ZIL_ZC_SEQ] = 1ULL; } /* * Read a log block and make sure it's valid. */ static int zil_read_log_block(zilog_t *zilog, const blkptr_t *bp, blkptr_t *nbp, void *dst, char **end) { enum zio_flag 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; 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 (bcmp(&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); bcopy(lr, dst, 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 (bcmp(&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); bcopy(lr, dst, 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) { enum zio_flag 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) bzero(wbuf, 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; 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) bcopy(abuf->b_data, wbuf, arc_buf_size(abuf)); arc_buf_destroy(abuf, &abuf); } return (error); } /* * 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) { 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; 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; if (blk_seq > claim_blk_seq) break; if ((error = parse_blk_func(zilog, &blk, arg, txg)) != 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, &blk, &next_blk, lrbuf, &end); if (error != 0) 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; if ((error = parse_lr_func(zilog, lr, arg, txg)) != 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; ASSERT(!claimed || !(zh->zh_flags & ZIL_CLAIM_LR_SEQ_VALID) || (max_blk_seq == claim_blk_seq && max_lr_seq == claim_lr_seq)); zil_bp_tree_fini(zilog); zio_buf_free(lrbuf, SPA_OLD_MAXBLOCKSIZE); return (error); } /* ARGSUSED */ static int zil_clear_log_block(zilog_t *zilog, blkptr_t *bp, void *tx, uint64_t first_txg) { 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); } /* ARGSUSED */ static int zil_noop_log_record(zilog_t *zilog, lr_t *lrc, void *tx, uint64_t first_txg) { return (0); } static int zil_claim_log_block(zilog_t *zilog, 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_log_record(zilog_t *zilog, lr_t *lrc, void *tx, uint64_t first_txg) { lr_write_t *lr = (lr_write_t *)lrc; int error; if (lrc->lrc_txtype != TX_WRITE) return (0); /* * 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)) != 0) return (error); return (zil_claim_log_block(zilog, &lr->lr_blkptr, tx, first_txg)); } /* ARGSUSED */ static int zil_free_log_block(zilog_t *zilog, blkptr_t *bp, void *tx, uint64_t claim_txg) { zio_free(zilog->zl_spa, dmu_tx_get_txg(tx), bp); return (0); } static int zil_free_log_record(zilog_t *zilog, lr_t *lrc, void *tx, uint64_t claim_txg) { lr_write_t *lr = (lr_write_t *)lrc; blkptr_t *bp = &lr->lr_blkptr; /* * If we previously claimed it, we need to free it. */ if (claim_txg != 0 && lrc->lrc_txtype == TX_WRITE && 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_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 (AVL_CMP(v1, v2)); } static lwb_t * zil_alloc_lwb(zilog_t *zilog, blkptr_t *bp, boolean_t slog, uint64_t txg) { lwb_t *lwb; lwb = kmem_cache_alloc(zil_lwb_cache, KM_SLEEP); lwb->lwb_zilog = zilog; lwb->lwb_blk = *bp; 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_tx = NULL; lwb->lwb_issued_timestamp = 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)); 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)); 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(). */ 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. */ boolean_t 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(). */ 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); } /* * 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 slog = FALSE; /* * 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 endianess */ 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->os_dsl_dataset->ds_object, txg, &blk, NULL, ZIL_MIN_BLKSZ, &slog); 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); /* * If we just allocated the first log block, commit our transaction * and wait for zil_sync() to stuff the block poiner into zh_log. * (zh is part of the MOS, so we cannot modify it in open context.) */ if (tx != NULL) { dmu_tx_commit(tx); txg_wait_synced(zilog->zl_dmu_pool, txg); } ASSERT(bcmp(&blk, &zh->zh_log, sizeof (blk)) == 0); 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. */ void 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; 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) { 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); } 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); } 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, 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); } BP_ZERO(&zh->zh_log); dsl_dataset_dirty(dmu_objset_ds(os), tx); dmu_objset_disown(os, 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); 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; dsl_dataset_dirty(dmu_objset_ds(os), tx); } ASSERT3U(first_txg, ==, (spa_last_synced_txg(zilog->zl_spa) + 1)); dmu_objset_disown(os, 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. */ /* ARGSUSED */ int zil_check_log_chain(dsl_pool_t *dp, dsl_dataset_t *ds, void *tx) { 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)); 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; dmu_tx_t *tx = lwb->lwb_tx; zil_commit_waiter_t *zcw; spa_config_exit(zilog->zl_spa, SCL_STATE, lwb); zio_buf_free(lwb->lwb_buf, lwb->lwb_sz); mutex_enter(&zilog->zl_lock); /* * Ensure the lwb buffer pointer is cleared before releasing the * txg. 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; lwb->lwb_tx = NULL; ASSERT3U(lwb->lwb_issued_timestamp, >, 0); zilog->zl_last_lwb_latency = gethrtime() - lwb->lwb_issued_timestamp; 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 ((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; 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); /* * Now that we've written this log block, we have a stable pointer * to the next block in the chain, so it's OK to let the txg in * which we allocated the next block sync. */ dmu_tx_commit(tx); } /* * 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_put(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; 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. */ 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) 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 defered, + * 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]); 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_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_DONT_PROPAGATE, &zb); ASSERT3P(lwb->lwb_write_zio, !=, NULL); lwb->lwb_state = LWB_STATE_OPENED; mutex_enter(&zilog->zl_lock); 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. */ uint64_t zil_block_buckets[] = { 4096, /* non TX_WRITE */ 8192+4096, /* data base */ 32*1024 + 4096, /* NFS writes */ UINT64_MAX }; /* * 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. * Note that if the allocation of nlwb synced before we wrote * the block that points at it (lwb), we'd leak it if we crashed. * Therefore, we don't do dmu_tx_commit() until zil_lwb_write_done(). * We dirty the dataset to ensure that zil_sync() will be called * to clean up in the event of allocation failure or I/O failure. */ 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); lwb->lwb_tx = tx; /* * 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 * guesssing 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]; i++) continue; zil_blksz = zil_block_buckets[i]; if (zil_blksz == UINT64_MAX) zil_blksz = SPA_OLD_MAXBLOCKSIZE; 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); /* pass the old blkptr in order to spread log blocks across devs */ error = zio_alloc_zil(spa, zilog->zl_os->os_dsl_dataset->ds_object, txg, bp, &lwb->lwb_blk, zil_blksz, &slog); 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); } 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 */ bzero(lwb->lwb_buf + lwb->lwb_nused, 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); /* * If there was an allocation failure then nlwb will be null which * forces a txg_wait_synced(). */ return (nlwb); } 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, lwb_sp, reclen, txg; 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); } else { dlen = 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; if (reclen > lwb_sp || (reclen + dlen > lwb_sp && lwb_sp < ZIL_MAX_WASTE_SPACE && (dlen % ZIL_MAX_LOG_DATA == 0 || lwb_sp < reclen + dlen % ZIL_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; ASSERT3U(reclen + MIN(dlen, sizeof (uint64_t)), <=, lwb_sp); } dnow = MIN(dlen, lwb_sp - reclen); lr_buf = lwb->lwb_buf + lwb->lwb_nused; bcopy(lrc, lr_buf, reclen); lrcb = (lr_t *)lr_buf; /* Like lrc, but inside lwb. */ lrwb = (lr_write_t *)lrcb; /* Like lrw, but inside lwb. */ /* * 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) { 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; } else { ASSERT(itx->itx_wr_state == WR_INDIRECT); dbuf = NULL; } /* * 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, lrwb, dbuf, lwb, lwb->lwb_write_zio); if (error == EIO) { txg_wait_synced(zilog->zl_dmu_pool, txg); return (lwb); } if (error != 0) { ASSERT(error == ENOENT || error == EEXIST || error == 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 lrsize) { itx_t *itx; lrsize = P2ROUNDUP_TYPED(lrsize, sizeof (uint64_t), size_t); itx = kmem_alloc(offsetof(itx_t, itx_lr) + lrsize, KM_SLEEP); itx->itx_lr.lrc_txtype = txtype; itx->itx_lr.lrc_reclen = lrsize; itx->itx_lr.lrc_seq = 0; /* defensive */ itx->itx_sync = B_TRUE; /* default is synchronous */ return (itx); } void zil_itx_destroy(itx_t *itx) { kmem_free(itx, offsetof(itx_t, itx_lr) + itx->itx_lr.lrc_reclen); } /* * Free up the sync and async itxs. The itxs_t has already been detached * so no locks are needed. */ static void zil_itxg_clean(itxs_t *itxs) { itx_t *itx; list_t *list; avl_tree_t *t; void *cookie; 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 (AVL_CMP(o1, o2)); } /* * Remove all async itx with the given oid. */ static 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; /* * Object ids can be re-instantiated in the next txg so * remove any async transactions to avoid future leaks. * This can happen if a fsync occurs on the re-instantiated * object for a WR_INDIRECT or WR_NEED_COPY write, which gets * the new file data and flushes a write record for the old object. */ if ((itx->itx_lr.lrc_txtype & ~TX_CI) == TX_REMOVE) zil_remove_async(zilog, itx->itx_oid); /* * 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", 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 comitted) 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); if (taskq_dispatch(zilog->zl_dmu_pool->dp_zil_clean_taskq, (void (*)(void *))zil_itxg_clean, clean_me, TQ_NOSLEEP) == 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. */ static 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)) { 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_waiters; lwb_t *lwb; itx_t *itx; 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_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 { ASSERT3S(lwb->lwb_state, !=, LWB_STATE_ISSUED); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_WRITE_DONE); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_FLUSH_DONE); } while (itx = list_head(&zilog->zl_itx_commit_list)) { 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); } 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 occuring, 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); } else if (lrc->lrc_txtype == TX_COMMIT) { ASSERT3P(lwb, ==, NULL); zil_commit_waiter_link_nolwb( itx->itx_private, &nolwb_waiters); } } list_remove(&zilog->zl_itx_commit_list, itx); 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)) { zil_commit_waiter_skip(zcw); list_remove(&nolwb_waiters, zcw); } } 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 its "closed", then no itxs have been committed to * it, so there's no point in issuing its zio (i.e. * it's "empty"). * * If its "open" state, 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 occuring * 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 occuring 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 occuring 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 it's 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. */ } } /* * 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 commited 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_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 aquired 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 informaiton * 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. */ clock_t timeleft = cv_timedwait_hires(&zcw->zcw_cv, &zcw->zcw_lock, wakeup, USEC2NSEC(1), CALLOUT_FLAG_ABSOLUTE); if (timeleft >= 0 || 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 signalled 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() { 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)); 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 signalled, 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 it 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 signalled * -- 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 signalled, in the zio completion * callback of the lwb's zio[*]. * * * Actually, the waiters are signalled 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 depenency 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 sematics 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) { /* * 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; 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; ASSERT(list_head(&zilog->zl_lwb_list) == NULL); bzero(zh, sizeof (zil_header_t)); bzero(zilog->zl_replayed_seq, 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; } } 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); } mutex_exit(&zilog->zl_lock); } /* ARGSUSED */ static int zil_lwb_cons(void *vbuf, void *unused, int kmflag) { lwb_t *lwb = vbuf; 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); } /* ARGSUSED */ static void zil_lwb_dest(void *vbuf, void *unused) { lwb_t *lwb = vbuf; mutex_destroy(&lwb->lwb_vdev_lock); avl_destroy(&lwb->lwb_vdev_tree); list_destroy(&lwb->lwb_waiters); } 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); } void zil_fini(void) { kmem_cache_destroy(zil_zcw_cache); kmem_cache_destroy(zil_lwb_cache); } 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; mutex_init(&zilog->zl_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&zilog->zl_issuer_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); return (zilog); } void zil_free(zilog_t *zilog) { 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 (int 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); cv_destroy(&zilog->zl_cv_suspend); kmem_free(zilog, sizeof (zilog_t)); } /* * Open an intent log. */ zilog_t * zil_open(objset_t *os, zil_get_data_t *get_data) { 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; 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); /* * We need to use txg_wait_synced() to wait long enough for the * ZIL to be clean, and to wait for all pending lwbs to be * written out. */ if (txg != 0) txg_wait_synced(zilog->zl_dmu_pool, txg); if (zilog_is_dirty(zilog)) zfs_dbgmsg("zil (%p) is dirty, txg %llu", zilog, 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); 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 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); } 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 lwb's 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 (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 **zr_replay; void *zr_arg; boolean_t zr_byteswap; char *zr_lr; } zil_replay_arg_t; static int zil_replay_error(zilog_t *zilog, 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, 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. */ bcopy(lr, zr->zr_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); } /* ARGSUSED */ static int zil_incr_blks(zilog_t *zilog, blkptr_t *bp, void *arg, uint64_t claim_txg) { zilog->zl_replay_blks++; return (0); } /* * If this dataset has a non-empty intent log, replay it and destroy it. */ void zil_replay(objset_t *os, void *arg, zil_replay_func_t *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) { zil_destroy(zilog, B_TRUE); return; } zr.zr_replay = replay_func; zr.zr_arg = arg; zr.zr_byteswap = BP_SHOULD_BYTESWAP(&zh->zh_log); zr.zr_lr = kmem_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); kmem_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; } 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); } /* ARGSUSED */ int zil_reset(const char *osname, void *arg) { int error; error = zil_suspend(osname, NULL); if (error != 0) return (SET_ERROR(EEXIST)); return (0); } Index: vendor-sys/illumos/dist/uts/common/sys/fs/zfs.h =================================================================== --- vendor-sys/illumos/dist/uts/common/sys/fs/zfs.h (revision 354960) +++ vendor-sys/illumos/dist/uts/common/sys/fs/zfs.h (revision 354961) @@ -1,1199 +1,1201 @@ /* * 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 http://www.opensolaris.org/os/licensing. * 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, 2016 by Delphix. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright 2017 Joyent, Inc. * Copyright (c) 2017 Datto Inc. * Copyright (c) 2017, Intel Corporation. */ /* Portions Copyright 2010 Robert Milkowski */ #ifndef _SYS_FS_ZFS_H #define _SYS_FS_ZFS_H #include #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_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_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 usr/src/common/zfs/zfs_prop.c. */ typedef enum { ZPROP_CONT = -2, ZPROP_INVAL = -1, 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, /* not exposed to the user */ 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_FILESYSTEM_LIMIT, ZFS_PROP_SNAPSHOT_LIMIT, ZFS_PROP_FILESYSTEM_COUNT, ZFS_PROP_SNAPSHOT_COUNT, ZFS_PROP_REDUNDANT_METADATA, ZFS_PROP_PREV_SNAP, ZFS_PROP_RECEIVE_RESUME_TOKEN, ZFS_PROP_REMAPTXG, /* not exposed to the user */ ZFS_PROP_SPECIAL_SMALL_BLOCKS, ZFS_NUM_PROPS } zfs_prop_t; typedef enum { ZFS_PROP_USERUSED, ZFS_PROP_USERQUOTA, ZFS_PROP_GROUPUSED, ZFS_PROP_GROUPQUOTA, ZFS_NUM_USERQUOTA_PROPS } zfs_userquota_prop_t; extern const char *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. 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 { 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_COMMENT, ZPOOL_PROP_EXPANDSZ, ZPOOL_PROP_FREEING, ZPOOL_PROP_FRAGMENTATION, ZPOOL_PROP_LEAKED, ZPOOL_PROP_MAXBLOCKSIZE, ZPOOL_PROP_BOOTSIZE, ZPOOL_PROP_CHECKPOINT, ZPOOL_PROP_TNAME, ZPOOL_PROP_MAXDNODESIZE, ZPOOL_PROP_MULTIHOST, ZPOOL_NUM_PROPS } zpool_prop_t; /* Small enough to not hog a whole line of printout in zpool(1M). */ #define ZPROP_MAX_COMMENT 32 #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 /* * Dataset property functions shared between libzfs and kernel. */ const char *zfs_prop_default_string(zfs_prop_t); uint64_t zfs_prop_default_numeric(zfs_prop_t); boolean_t zfs_prop_readonly(zfs_prop_t); boolean_t zfs_prop_visible(zfs_prop_t prop); boolean_t zfs_prop_inheritable(zfs_prop_t); boolean_t zfs_prop_setonce(zfs_prop_t); const char *zfs_prop_to_name(zfs_prop_t); zfs_prop_t zfs_name_to_prop(const char *); boolean_t zfs_prop_user(const char *); boolean_t zfs_prop_userquota(const char *); boolean_t zfs_prop_written(const char *); int zfs_prop_index_to_string(zfs_prop_t, uint64_t, const char **); int zfs_prop_string_to_index(zfs_prop_t, const char *, uint64_t *); uint64_t zfs_prop_random_value(zfs_prop_t, uint64_t seed); boolean_t zfs_prop_valid_for_type(int, zfs_type_t); /* * Pool property functions shared between libzfs and kernel. */ zpool_prop_t zpool_name_to_prop(const char *); const char *zpool_prop_to_name(zpool_prop_t); const char *zpool_prop_default_string(zpool_prop_t); uint64_t zpool_prop_default_numeric(zpool_prop_t); boolean_t zpool_prop_readonly(zpool_prop_t); boolean_t zpool_prop_feature(const char *); boolean_t zpool_prop_unsupported(const char *name); int zpool_prop_index_to_string(zpool_prop_t, uint64_t, const char **); int zpool_prop_string_to_index(zpool_prop_t, const char *, uint64_t *); uint64_t zpool_prop_random_value(zpool_prop_t, 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_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_type_t; /* * 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 /* * When bumping up SPA_VERSION, make sure GRUB ZFS understands the on-disk * format change. Go to usr/src/grub/grub-0.97/stage2/{zfs-include/, fsys_zfs*}, * and do the appropriate changes. Also bump the version number in * usr/src/grub/capability. */ #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. * * When changing, be sure to teach GRUB how to read the new format! * See usr/src/grub/grub-0.97/stage2/{zfs-include/,fsys_zfs*} */ #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 /* 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 ":" * (e.g. "org.open-zfs:") 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_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 */ #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_REMOVING "removing" #define ZPOOL_CONFIG_RESILVER_TXG "resilver_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_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 */ /* * 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_DATA_ERRORS "verify_data_errors" #define ZPOOL_CONFIG_REWIND_TIME "seconds_of_rewind" #define VDEV_TYPE_ROOT "root" #define VDEV_TYPE_MIRROR "mirror" #define VDEV_TYPE_REPLACING "replacing" #define VDEV_TYPE_RAIDZ "raidz" #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" /* 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_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" #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" /* * This is needed in userland to report the minimum necessary device size. * * Note that the zfs test suite uses 64MB vdevs. */ #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 "/etc/zfs/zpool.cache" /* * 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 */ VDEV_AUX_SPLIT_POOL, /* vdev was split off into another pool */ VDEV_AUX_ACTIVE, /* vdev active on a different host */ VDEV_AUX_CHILDREN_OFFLINE /* all children are offline */ } 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; /* * Initialize functions. */ typedef enum pool_initialize_func { POOL_INITIALIZE_DO, POOL_INITIALIZE_CANCEL, POOL_INITIALIZE_SUSPEND, POOL_INITIALIZE_FUNCS } pool_initialize_func_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_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_issued; /* issued 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 scurb pass */ /* cumulative time scrub spent paused, needed for rate calculation */ uint64_t pss_pass_scrub_spent_paused; /* Sorted scrubbing new fields */ /* Stored on disk */ 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 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; typedef enum { VDEV_INITIALIZE_NONE, VDEV_INITIALIZE_ACTIVE, VDEV_INITIALIZE_CANCELED, VDEV_INITIALIZE_SUSPENDED, VDEV_INITIALIZE_COMPLETE } vdev_initializing_state_t; /* * Vdev statistics. Note: all fields should be 64-bit because this * is passed between kernel and userland as an nvlist uint64 array. */ 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[ZIO_TYPES]; /* operation count */ uint64_t vs_bytes[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_initialzing_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 */ } vdev_stat_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 elments 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_DISK_ROOT "/dev/dsk" #define ZFS_DISK_ROOTD ZFS_DISK_ROOT "/" #define ZFS_RDISK_ROOT "/dev/rdsk" #define ZFS_RDISK_ROOTD ZFS_RDISK_ROOT "/" /* general zvol path */ #define ZVOL_DIR "/dev/zvol" /* expansion */ #define ZVOL_PSEUDO_DEV "/devices/pseudo/zfs@0:" /* for dump and swap */ #define ZVOL_FULL_DEV_DIR ZVOL_DIR "/dsk/" #define ZVOL_FULL_RDEV_DIR ZVOL_DIR "/rdsk/" #define ZVOL_PROP_NAME "name" #define ZVOL_DEFAULT_BLOCKSIZE 8192 /* * /dev/zfs ioctl numbers. */ typedef enum zfs_ioc { ZFS_IOC_FIRST = ('Z' << 8), ZFS_IOC = ZFS_IOC_FIRST, ZFS_IOC_POOL_CREATE = ZFS_IOC_FIRST, ZFS_IOC_POOL_DESTROY, ZFS_IOC_POOL_IMPORT, ZFS_IOC_POOL_EXPORT, ZFS_IOC_POOL_CONFIGS, ZFS_IOC_POOL_STATS, ZFS_IOC_POOL_TRYIMPORT, ZFS_IOC_POOL_SCAN, ZFS_IOC_POOL_FREEZE, ZFS_IOC_POOL_UPGRADE, ZFS_IOC_POOL_GET_HISTORY, ZFS_IOC_VDEV_ADD, ZFS_IOC_VDEV_REMOVE, ZFS_IOC_VDEV_SET_STATE, ZFS_IOC_VDEV_ATTACH, ZFS_IOC_VDEV_DETACH, ZFS_IOC_VDEV_SETPATH, ZFS_IOC_VDEV_SETFRU, ZFS_IOC_OBJSET_STATS, ZFS_IOC_OBJSET_ZPLPROPS, ZFS_IOC_DATASET_LIST_NEXT, ZFS_IOC_SNAPSHOT_LIST_NEXT, ZFS_IOC_SET_PROP, ZFS_IOC_CREATE, ZFS_IOC_DESTROY, ZFS_IOC_ROLLBACK, ZFS_IOC_RENAME, ZFS_IOC_RECV, ZFS_IOC_SEND, ZFS_IOC_INJECT_FAULT, ZFS_IOC_CLEAR_FAULT, ZFS_IOC_INJECT_LIST_NEXT, ZFS_IOC_ERROR_LOG, ZFS_IOC_CLEAR, ZFS_IOC_PROMOTE, ZFS_IOC_SNAPSHOT, ZFS_IOC_DSOBJ_TO_DSNAME, ZFS_IOC_OBJ_TO_PATH, ZFS_IOC_POOL_SET_PROPS, ZFS_IOC_POOL_GET_PROPS, ZFS_IOC_SET_FSACL, ZFS_IOC_GET_FSACL, ZFS_IOC_SHARE, ZFS_IOC_INHERIT_PROP, ZFS_IOC_SMB_ACL, ZFS_IOC_USERSPACE_ONE, ZFS_IOC_USERSPACE_MANY, ZFS_IOC_USERSPACE_UPGRADE, ZFS_IOC_HOLD, ZFS_IOC_RELEASE, ZFS_IOC_GET_HOLDS, ZFS_IOC_OBJSET_RECVD_PROPS, ZFS_IOC_VDEV_SPLIT, ZFS_IOC_NEXT_OBJ, ZFS_IOC_DIFF, ZFS_IOC_TMP_SNAPSHOT, ZFS_IOC_OBJ_TO_STATS, ZFS_IOC_SPACE_WRITTEN, ZFS_IOC_SPACE_SNAPS, ZFS_IOC_DESTROY_SNAPS, ZFS_IOC_POOL_REGUID, ZFS_IOC_POOL_REOPEN, ZFS_IOC_SEND_PROGRESS, ZFS_IOC_LOG_HISTORY, ZFS_IOC_SEND_NEW, ZFS_IOC_SEND_SPACE, ZFS_IOC_CLONE, ZFS_IOC_BOOKMARK, ZFS_IOC_GET_BOOKMARKS, ZFS_IOC_DESTROY_BOOKMARKS, ZFS_IOC_CHANNEL_PROGRAM, ZFS_IOC_REMAP, ZFS_IOC_POOL_CHECKPOINT, ZFS_IOC_POOL_DISCARD_CHECKPOINT, ZFS_IOC_POOL_INITIALIZE, ZFS_IOC_POOL_SYNC, ZFS_IOC_LAST } zfs_ioc_t; /* * 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. */ 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_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; /* * 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_DSNAME "dsname" #define ZPOOL_HIST_DSID "dsid" #define ZPOOL_HIST_ERRNO "errno" /* * The following are names used when invoking ZFS_IOC_POOL_INITIALIZE. */ #define ZPOOL_INITIALIZE_COMMAND "initialize_command" #define ZPOOL_INITIALIZE_VDEVS "initialize_vdevs" /* * 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_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_CHECKPOINT 0x10 #define ZFS_IMPORT_TEMP_NAME 0x20 #define ZFS_IMPORT_SKIP_MMP 0x40 /* * 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_END * 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" #ifdef __cplusplus } #endif #endif /* _SYS_FS_ZFS_H */